ALBERT R. MANN LIBRARY AT CORNELL UNIVERSITY CORNELL UNIVERSITY LIBRARY 3 1924 057 348 65 r^ '^>. Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924057348165 THE NATURAL HISTORY OF PLANTS THEIE FORMS, GROWTH, REPRO'DUCTION, and DISTRIBUTION FROM THE GERMAN OF AJSFTON KERJ^ER von MARILAUN PEOPESSOR OP BOTANY IN THE UNITEBSITT OP VIENNA BY F. W. OLIVER, M.A., D.Sc. QUAIN PKOPESSOK OP BOTANY IN UNIVEESITY COLLEQE, LONDON WITH THE ASSISTANCE OF MAEIAN BUSK, B.Sc. and MAEY F. EWAET, B.Sc. WITH ABOUT 1000 ORIGINAL WOODCUT ILLUSTRATIONS AND SIXTEEN PLATES IN COLOURS HALF-VOLUME IV. NEW YORK HENRY HOLT AND COMPANY 1895 I\.<^(.5 5-1. CONTENTS OF HALF-VOLUME IV, LIST OF ILLUSTEATIONS. FAGB Plate XV. Frondose and Fruticose Lichens, - - - to face 694 „ XVI. Eucalyptus Grove and Grass-trees in Australia, - - 730 Illustrations in the Text — Fig. S57 to Fig. 482. The History of Species {Continued). 2. Alteration in the Form of Species (Continued)— The Influence of Mutilation on the Form of Plants, - - - - - 514 Alteration of Form by Parasitic Fungi, - - - 518 Alteration of Form by Gall-producing Insects, 527 The Genesis of New Forms as a result of Grossing, ----- 554 3. The Origin of Species — The Genesis of New Species, - ..... 576 Derivation of Existing Species, - ..... 595 The Subdivisions of the Vegetable Kingdom, ... . 600 4. The Distribution of Species — The Distribution of Species by Oifshoots, ... 790 The Dispersion of Species by Means of Fruits and Seeds, ... - 833 Limits of Distribution, ..... 878 Plant Communities and Floras, - ... 885 5. The Extinction of Species, - - 899 ERRATA. Vol. I. Plate I., fly-sheet, q, for Cress read Sv/mmer Savdry (Saturya hortensis). p. 336, line 15, for Daffodil read LiLy. p. 547, line 16, for 16-r read -16-1°. p. 663, line 27, tor repens read reptans. p. 685, line 24, delete cownter-. „ line 28, for clockwise read cownter-clochmse. Vol. II. p. 69, line 4, for Lycopodiaeece read LyeopodiaZes. p. 89, line 12, tor S13 << cmd SIS « read SW « (wwi S13 ". p. 120, line 7 from bottom, for acer read acris. p. 121, line 1, for Leguminous read Cruciferous. p. 280, line 2 from bottom, for Lentihulacew read Lentihulariaceai, p. 612, line 11 from the bottom, for calcination read caloijlcatiun. p. 633, line 1, after though insert moi. p. 641, line 12, for not read -net. p. 648, line 7, for iAe read some. „ line 16, ioiaUimetes read ahvnetes. p. 696, line 25 from bottom, for Antherocerotacece read Anthocerotacecs. p. 698, line 1, for Antheroeerotacece read Anthocerotacecs, p. 704, line 14, for Pteridophyte read Pterid^hyta. INFLUENCE OF THE SUBSTRATUM. 497 Under these circumstances it is a matter of indifference whether 10 per cent or only traces of lime or silica can be demonstrated in the soil, and the hypothesis that plant-species which grow on limestone fail to grow on slate because they are not able to supply their need of calcium, or that . the plants growing on slate cannot flourish on limestone mountains because they cannot obtain the necessary amount of silica, must be abandoned, as well as the assumption that these substances when absorbed as food serve as a stimulus to change of form. I strongly supported this latter hypothesis at the time, and thought I should be able to strengthen and confirm it by careful cultural experiments. Seeds of several species which demand lime were sown in soil containing hardly perceptible quanti- ties of lime, and the seedlings were watered with water devoid of calcium; in another place seeds of species demanding a silica-containing substratum were placed in soil which contained much limestone, and the seedlings were watered with lime- water. At first it seemed as if an alteration of form had actually taken place in some individuals. But this was a mistake, or rather, the alteration only consisted in the greater or less luxuriance of the foliage, lengthening or shortening of the stem, abundant or scanty development of flowers and the like. But no actual change of form which would be retained by their descendants could be obtained. The species of plants accustomed to lime, grown on a soil devoid of lime, presented a miserable appearance, with scanty flowers which ripened only a few seeds, whilst the silica-demanding species grown on lime-containing soil soon withered and died without flowering at all. The change of form, indeed the actual inter- change I had anticipated between the closely allied species which grow on the two rocky substrata in a state of nature, did not occur at all. If we still take the case of siliceous and calcareous plants, and regard the soil as the source of free inorganic substances which influence the plants, we are forced to assume that greater quantities of one substance will be injurious to one or other of them. The absorbent cells have the capacity of choosing between the substances at their disposal, but this capacity has a definite limit in every species. The cells can absorb as much as they require from a very weak solution of common salt, soda, gypsum, calcium bicarbonate, &c., but a concentrated solution of these salts may injure and destroy their structure and function. If it is allowed to act for any length of time on the cells whose function is to absorb inorganic nutriment, the death of the whole plant will inevitably result. If the Moss which grows on blocks of granite is 'watered with a saturated solution of gypsum; if the soil into which our Meadow-grasses send their roots is watered with a saturated solution of common salt; or if the humus in which the plants of an upland moor grow is mixed with sodium carbonate or calcium bicarbonate, the plants invariably perish, and the same mineral substances, which in a very weak solution are needful, or at any rate harmless, become poisonous when the solutions are concentrated. The fact that one species of plant prefers this and another that mineral substance (see vol. i. p. 73), however, renders it probable that the injurious effect of materials in large quantity in the soil varies, that a large quantity of VOL. II. 82 498 DEPENDENCE OF PLANT FORM ON SOIL AND CLIMATE. common salt would be injurious to one species, and an abundance of sodium or potassium salts to another. From the present standpoint of our knowledge concerning the absorption of inorganic materials by plants, therefore, Unger's classification, especially the expressions silica-demanding and silica-preferring, is no longer suitable, and it would be more to the purpose to speak of plants which are injured by lime, potash, &c. The difference in the vegetation on the closely adjoining limestone and slate mountains met with in so many places in the Alps, and so well seen in the neigh- bourhood of Kitzbiihel, where the climatic influences on the two ranges are identical, can be accounted for most satisfactorily in the following way. Plant-species which •demand or prefer a siliceous soil are absent from limestone mountains wherever their roots would be exposed to more free lime than is beneficial; if present they would be weakened, and thus vanquished in the struggle with their fellows, to whom the larger quantity of lime is harmless, and they would eventually perish. These plants flourish luxuriantly, however, on slate mountains, because there the «oil does not contain an injurious amount of lime. The absence of species, demand- ing or preferring lime, from slate mountains can be explained in the same way. When seeds are brought thither by the wind from the neighbouring limestone mountains and germination commences, their further development is visibly retarded; they dwindle wherever there is not much lime, and are overgrown «,nd suppressed by the siliceous species which flourish there so luxuriantly. The brown or black mass formed by the decomposition of dead plant residues, known as humus, plays a very important part in the contrasting vegetation on limestone and slate mountains. To obtain a true idea of its significance it must first be pointed out that three distinct stages can be distinguished in the development of a -continuous and intricate plant-covering. To the first stage belong the plants which settle down on the bare earth content with a substratum wholly devoid of humus; in the course of time they conquer the -most barren rock, the barest boulders, and the dreariest shifting sands. The species of this group belong chiefly to the Lichens, Mosses, Grasses, Pinks, Crucifers, House-leeks, Saxifrages, and Composites, whose spores, seeds, and fruits are exceptionally well adapted for wind distribution, and can be transferred with ease to the steepest slopes and the most uncompromising crags. The second stage includes plants which require a moderate amount of soil mixed with humus; they establish themselves on the ground pre- pared by the first settlers, wresting it from them and talking possession, and then suppressing and overgrowing them entirely. These plants belong to very different famihes, whose distribution and establishment are effected in very many ways to be described subsequently. The third stage of development consists of plants for which the abundant humus stored up successively by the plants of the second stage is absolutely indispensable. Bog-moss, Lycopodiums, Sedges, and Heaths form the chief part of this stage. In the course of years the amount of inorganic materials in the soil which supports the plants of the third stage continuously diminishes. Plants which require a large quantity of inorganic salts languish, and are, looreover, INFLUENCE OF THE SUBSTRATUM. 499 overcome by saprophytes which find a suitable habitat there and flourish in abundance. The decayed portions of Saprophytes contain relatively little inorganic material. No trace of lime (in particular) is to be found in their ash. In this way a superficial layer of humus is formed which actually excludes a large number of plants. The next deeper layer may contain a considerable quantity of inorganic salts, but they are valueless to plants rooted in the upper (humus) layer, as they cannot penetrate it. It has been shown by experiment that pure humus possesses the power of holding back materials which are soluble in water. It possesses this property to such an extent that if salt solutions are filtered through a layer of humus the water which escapes below is almost pure. It is therefore impossible for inorganic substances from the deeper layers of the soil, much less from the underlying rock, to reach the surface layer of humus in solution by diffusion; and if some mineral ingredients are not introduced by irrigation or flooding, the upper layer of soil consists of pure humus on which only saprophytic plants can flourish. The formation of such layers of humus occurs much more easily and quickly on slate mountains than on limestone, because in the former the rock and the products of its decomposition retain water much better, and a uniform saturation promotes the development of humus, and also because on slaty soil the second stage of the development of the plant-covering consists of plants which require very few inor- ganic food -substances, and accordingly very few inorganic materials are yielded by the humus, which originates at the cost of the decaying portions of these plants. But a thick stratum of pure humus may also arise in course of time on lime- stone mountains. Only the soil must be uniformly moist in that spot, and neither sand nor mud must be deposited on it. If these conditions are fulfilled a deep humus will gradually spread itself over limestone rocks and debris in the third stage of development, the superficial layer of which will contain no trace of lime, but will afford an excellent soil for silica-loving plants (i.e. for those to which lime is injurious). The isolated occurrence of so-called siliceous or slate- plants on limestone mountains, even in the middle of a patch of plants which are characteristic of a limestone soil, may be naturally explained in this fashion. The water which moistens the rock and soaks the soil has, apart from its mechanical action, the important function of opening up mineral substances and of forming solutions from which the absorbent plant-cells may take their choice. The atmospheric water which penetrates into the earth from above is especially valuable as a solvent on account of the carbonic acid gas it contains. It is immeasurably more valuable to every part of the soil which is riddled by the roots of living plants than the soil-water, so poor in carbonic acid, which collects on impervious strata of the soil and soaks upwards through the superficial layers. The power of the soil to retain water depends mainly on the extent of breaking up undergone by the rock whose disintegration has formed the soil and upon the amount of clay which has arisen from this disintegration. But the amount of humus which in course of time has mixed with the disintegration and the decom- position products of the underlying rock is also an important factor, and thus very 500 DEPENDENCE OF PLANT FOKM ON SOIL AND CLIMATE, complex conditions arise which render the estimation of the soil's capacity for retaining water very difficult. If permeable sandy soil, poor in humus, is deprived of ground water and is dependent for its moisture solely on the atmosphere, the plants growing in it will be retarded in their development if rain and dew are absent for any length of time, and their outward appearance will be altered by this restriction of growth. Annual plants subjected to a lack of moisture in the soil' just at the time when their growth should be at its maximum, show best how far these alterations will go. The stem-structures remain short, the foliage-leaves shrink to their smallest extent, and no lateral shoots are developed. Only a few, or perhaps only one, of the flower-buds mature; it is small, opens comparatively very early, and the whole plant has a dwarfed aspect. Annual plants of the Poppy {Papaver Bhceas, somniferum), Pheasant's Eye (Adonis cestivalis, fiammea), Corn- cockle (A grostevima Githago), Corn&ower (Centaiirea Gyanus),and common Groundsel (Senecio vulgaris) grown on a dry soil differ from plants grown in the same place, but in a damp year, to such an extent in the size of all their parts that at first sight, they might be mistaken for other species. A clay soil which retains water is less exposed to danger of too great drjmess, but if it is not mixed with humus, and therefore loosened, it has the disadvantage that the water it contains cannot take- up the inorganic foods quickly enough and in sufficient quantity for the require- ments of the plants. This drawback explains the surprising fact that plants grown on heavy wet clay soils have a dwarfed appearance exactly like plants growing on dry sandy soil. In regions liable to flooding by streams and rivers where not. infrequently sandy and clay soils, in all degrees of porosity and admixed with humus in all possible proportions, are to be met with within a few yards of one another, certain species of plants are to be found growing near together in all imagin- able degrees of size, e.g. Aster Tripolium, Bidens cernua and tripartita, Polygonum lapathifolium, Rumex maritimus, Veronica Anagallis. In places where the- seedlings cannot find enough free mineral foods, in spite of the abundant moisture in the soil, the stem rises to some 3-8 cm.; in places which favour the absorption of food, to some 50-80 cm. We will describe only one species, Veronica Anagallis,. more in detail. Plants of this species are found with stems 3-5 cm. high and 0'5 mm. thick, with foliage-leaves 6-12 mm. long and 5-6 mm. broad when fully developed. The number of flowers in one inflorescence is about 4-5, the calyx and ripe capsular fruit measure 3 mm. in length. Contrasting with these are plants with stem 30-50 cm. high and 7-8 mm. thick, whose fully -formed leaves are 80 mm. long and 35 mm. broad. There are 40-50 flowers in each inflorescence, and the calyx and ripe capsule measure 4-5 mm. in length. Generally speaking these plants are about ten times as large as the others. If the soils which give rise to such surprising differences in size are examined it will be noticed that the dwarfed specimens are rooted in a heavy soil devoid of humus, while the large luxuriant plants flourish in a clay soil which is mixed with plenty of humus, and is therefore very open. Obviously the plants could not obtain from the heavy clay soil what they required for the structure of a vigorous plant, even although the INFLUENCE OF THE MEDIUM. 501 ground was well moistened and warmed; but this they could obtain in abundance from the saturated clay soil containing the humus. It has been already stated that the ground water is less favourable for vegetation than rain and dew on account of its paucity of carbonic acid. But the moistening of the ground by water which wells up from below brings other evils in its train. By this means the soil is over-saturated for a long time, a condition which the roots of most land-plants will not tolerate. When it remains stationary for a long while potassium and sodium salts, and, under certain conditions, humous acids pass into it from the wet earth in quantities anything but advantageous to the plants. Vege- tation, therefore, exhibits a scanty growth in places where the ground water influences the stratum of soil penetrated by rodts, and it usually consists of comparatively few species. In low-lying regions, where the ground water rises to the surface, we have the formation of lakes and ponds with variable water-level. Sometimes the plants growing in such places are quite submerged, while at other times their stem and leaves are above water. Land plants do not take kindly to this. Most of them cannot survive very long immersion; they become suffocated, die, and decompose under water in a few days. Only a few species have the remarkable power of growing equally well below or above water, and these are, of course, extremely interesting on account of their form. In accordance with the great contrast presented by the external conditions of life to which these species are temporarily exposed we have a fundamental change both in their outward appearance and in the internal structure of their several organs. In order that the stem and leaves should be held in the best position by the flowing water, the mechanical tissue in ;submerged varieties of these species is much reduced (see vol. i. pp. 424 and 665). They are also devoid of the contrivances which usually regulate transpiration, since no evaporation occurs under water. Stems grown under water consequently appear limp and flaccid when taken out of it; their leaves, when compared with those growing in the air, are much weaker and more delicate. They have no gloss, but are brighter green in colour, and in the air they collapse and dry up in a very ■short time. A vertical section through the leaf shows that the number of cells between the upper and lower epidermis is much reduced, and that the cells are shortened in a direction perpendicular to the leaf surface. The foliage-leaves of Veronica Beccahunga, when grown under water, are hardly one-third as thick as those grown in the air, and between the upper and lower epidermis there are only 4-5 layers of short cells, while in corresponding leaves of aerial plants there are 10-12 cell-layers and a distinct division into palisade and spongy parenchyma (see vol. i. p. 279). The shape of the leaf is also much changed under water. In Veronica Beccahunga the difference in aerial and submerged leaves is very slight, consisting only in the shortening of the petiole and in the marginal teeth becoming less marked. In Veronica Anagallis, likewise, the alteration in shape is incon- ■siderable, but in many others it is very noticeable, and we shall return to it when speaking of the influence of light. 502 DEPENDENCE OF PLANT FORM ON SOIL AND CLIMATE. Plants rooted in the mud of a river-bed, the stems and leaves of which are surrounded by rapidly-flowing water, must possess corresponding strength if they are not to be torn. In comparing two plants of the same species, the one growing in the still water of a deep lake, the other in a rapidly-flowing stream, it will be noticed that the walls of the superficial cells of the latter have become strongly thickened, and that strong bundles of bast-fibres have developed in the cortex of the stem, while in the former only the weakest traces of bast-fibres can be seen. The extraordinary length of stem, petiole, and leaf-blade is also very surprising in plants which grow in rapid water. The Pondweed Potamogeton fiuitans, the Rushes Juncus lamprocarpus and swpinus, the Grasses Agrostis stolonifera and Olyceria fiuitans are very instructive examples. A plant of the last-named Grass growing on damp soil on the edge of a stream over the water had linear, bluntly-pointed leaves, whose sheaths were on the average 15 cm. long, the blades 23 cm. long and 8-5 mm. broad. After this plant had been sub- merged under rapidly-flowing water in the following year, leaves unfolded, which tapered gradually to a point, with a sheath having a mean length of 47 cm., and blades 73 cm. long but only 5 mm. broad. The blades produced in running water were three times as long and actually rather narrower than in the air. There was no difference in the number of strands traversing the blade, but they were nearer to one another than in the aerial leaves. The Arrow-head (Sagittaria sagittifolia), which usually grows on the muddy bottom of shallow lakes, raising its leaves above the still water, has gained its name from the likeness of its leaf-blade to an arrow. If it is planted in the bed of a rapid stream so that the leaves during their development are exposed to a vigorous current, the leaf-blade is almost entirely suppressed. What still remains has the form of a spade, but not infrequently all trace of lamina is wanting. The petiole, however, lengthens to 70 cm., and forms a limp, flat, pale-green ribbon 1-2 cm. broad, which might easily be mistaken at flrst sight for the leaf of Vcdlisneria. Another remarkable change which is effected by submerging growing plants is the non-development of the epidermal structures called hairs, so that the leaves and stems of submerged plants always appear smooth. The suppression of hair- structures is very noticeable in the aquatic variety of Polygonum amphibium. In aerial plants of this species the leaves have short petioles, are lanceolate in shape, and are covered thickly with short hairs, which are rough to the touch; while the aquatic plants have long-stalked, broadly-linear leaves completely smooth on both sides. The humidity of the atmosphere has a marked effect on the form of land plants. Transpiration, which is so deeply concerned in all the vital processes, is carried on very slowly in air which is almost or quite saturated with water- vapour. If plants of a species which usually grows in dry air come into a humid atmosphere, they must be furnished with means for aiding evaporation. On the other hand, plants which grow in dry air must be protected against excessive transpiration. The aids and protective measures were so minutely described INFLUENCE OF TEMPERATURE. 503 in vol i. pp. 284 and 307, that it is needless to repeat them here; but it should be noted that the capacity of plants to construct their tissue as need requires, either for aiding transpiration or for protection against excessive evaporation, is very limited. It must also be pointed out that it is very difficult to distinguish clearly between the direct effect of the humidity of the air and the effects of other influences. Heat and light, as well as the amount of moisture in the soil, are intimately connected with the humidity of the air, but the relations are difficult to estimate. To a certain extent they are interchangeable, and therefore, in most instances, it is impossible to say which external influence is the cause of any particular alteration in the tissue concerned in transpiration. For the answer to the chief question, whether it is possible for a change in the conditions of life to cause an alteration of form in the sense of an adaptation, it is really a matter of indifference which influence causes the visible effect. Only here, as in so many other cases, matters are simplified if a certain partiality is permitted in experiments for solving these difficult questions, and if the interwoven influences of soil and climate are treated separately. The effect of heat on growing plants was discussed at vol. i. p. 523. It only remains to say here that the formation of starch and other reserve-foods, as well as the formation of sugar in fruits, is largely connected with heat. Fruits of the same species which ripen under a higher temperature differ greatly in the amount of sugar they contain from those ripening at a lower temperature. It is generally accepted that the size also of the stem, foliage, flowers, and fruit is influenced by heat. The changes which occur when plants in flower, after being for some time in a very warm room are transferred into a cooler room, the other conditions remaining the same, are in particular now recognized. When a large-flowered bulbous plant, e.g. the Belladonna Lily {Amaryllis Belladonna), is transferred to a cold greenhouse after opening its first flowers in a warm one, the flowers it here develops at a lower temperature are almost a third smaller than those produced in the warm house. But when the first flowers open in the cold, and the later ones in a warm atmosphere, the former remain small and the latter are larger in size. It is important to emphasize this circumstance in order that the phenomenon here exhibited may not be mistaken for another, in case we should be led to think that the flowers of a plant which first unfold are larger than those which succeed them even when there has not been the slightest alteration in the conditions of light, heat, humidity, &c. It is particularly instructive, when examining the effect of heat on the form of a species, to compare plants grown in water of different temperatures but under conditions otherwise similar. In mountainous districts the springs on the same mountain slope have a different temperature according to their elevation, and yet the same species of plants may be found growing in springs at the foot and higb up on the mountain. Let us take as examples plants of Gardamine amara, Myosotis palustris, Pedicularis palustris, and Veronica Beccabunga. These species grow at the foot of the Patscherkofel, near Innsbruck, in the bed 504 DEPENDENCE OF PLANT FOEM ON SOIL AND CLIMATE. of Streams with a mean temperature of 10-2 ° C, but they also flourish in a stream above the tree-line, at a height of 1921 metres above the sea-level, known as the '' Kreuzbrunnen ". Comparing plants of the same species growing under the influence of these different temperatures, the following differences are to be noted:— Plants of Veronica Beccabunga growing in spring water at a temperature of 10-2° C. were 20-50 em. high, and displayed 4-6 internodes between the bottom in which they were rooted and the level of the first inflorescences. The internodes of the stem were 60-120 mm. long and 5 mm. thick; the leaves springing from the middle of the plant were 40-60 mm. long, 20-25 mm. broad, and each of the flower racemes had 12-16 flowers. Plants growing in the spring water at a temperature of 4'2° C. were 10-15 cm. high with 4-6 internodes between the ground and the level of the first infiorescences. The internodes were 15-30 mm. long and 10-12 mm. thick, and each inflorescence had 12-16 flowers. Gardamine •amara, Myosotis paluatris, and Pedicularis palustris behaved similarly. There seemed to be no alteration in the form of the leaves and flowers; the corollas assumed a rather deeper tint in the Kreuzbrurmen; Myosotis palustris, which was 20 cm. high at the foot of the Patscherkofel, was 4-5 cm. high in the Kreuz- brunnen, and closely resembled the Eritrichium nanum of the Southern Alps in the deep blue of its corollas. Gardamine amara, in the same cold spring, in addition to the shortening of its internodes and diminution of its foliage -leaves, displayed a red colour on the outside of its white petals which was not present in plants at lower levels. The powerful influence of light on the development of plants was discussed at vol. i. p. 371. The question now before us is how far bright and subdued light are able to alter the size, form, and colour of plants. The following is a general review of what has been ascertained in the matter from experiments and direct observa- tion of nature. When plants of a species develop in subdued light they always have higher stems and longer leaves than when grown in bright light, provided, of course, that the conditions of moisture and temperature have been as far as possible identical. This difference is especially noticeable in comparing two plants of a species, one of which has developed in the dim light of a greenhouse in the short days of winter, the other in an unshaded place in the open country during the summer when the light lasts for 16-17 hours every day. The former has a lank thin stem, delicate yellowish-green leaves, and either none of its flowers unfold or else they have a weak appearance and their corollas are pale and flaccid. The illuminated plant has, on the other hand, a compact vigorous stem, dark green leaves, and unfolds a multitude of bright-hued flowers. One only of the large number of experiments which have been performed for the purpose of determining this matter definitely will be mentioned here— one indeed which shows how far the form of the flowers also may be affected. Seeds of a biennial Saxifrage, Saxifraga controversa, which were sown in several flower-pots filled with similar soil, produced numerous young plants. A pot with six of these young plants was taken in the autumn into the hot-house; another, likewise containing INFLUENCE OF LIGHT. 505 six young plants, passed the winter under a thick coat of snow in the open. At the beginning of December the six plants in the hot-house sent up from the centre of their small leaf-rosettes slender stalks 10 cm. high, whose upper internodes were 22 mm. long and 1 mm. thick. The stem-leaves were yellowish, entire, elongated, 6-7 mm. long and 2 mm. broad; calyx-tube 4 mm. long, 13 mm. broad; calyx-teeth 2 mm. long, 1'5 mm. broad; petals 3'5 mm. long, 2 mm. broad; stamens 1 mm. long. It was noted that lateral axes only developed in the axils of the upper stem-leaves, and that the buds of the lateral shoots in the lower leaf-axils atrophied. In the following May strong stems 6 cm. high were sent up from the leaf -rosettes of the plants which had wintered under the deep snow in the open; their upper internodes were 12 mm. long and 2 mm. thick. The stem-leaves were somewhat broadened in front with dentate margin, red in colour, 5 mm. long and 3 mm. broad. The measurements of the parts of the flowers were: — Calyx-tube, 2 mm. long, and 2 mm. broad; calyx-teeth, 1'5 mm. long, 1 mm. broad; petals, 2"3 mm. long, and 2 mm. broad; stamens, 1 mm. long. From the axils of the stem-leaves flower-bearing shoots developed, which, like the parts of the main stem exposed to the sun, were coloured red. Here then the alterations which certainly are due to the various light influences consist not only in the lengthening and shortening of the stem- and foliage-leaves, but the flowers are correspondingly changed. The petals of the flowers which opened at the New Year when the days were shortest were not only relatively but" actually narrower than those which belonged to flowers which opened in the early summer when the days were longest. It has already been stated that the elongation of the leaves and the division of the leaf-lamina into long narrow segments in submerged leaves is associated with the diminution undergone by the light in passing through the water (see vol. i. p. 665). The elongation of submerged leaves is very well seen in the water Star- wort (Gallitriche) and Mare's-tail (Hippuris). In the latter the linear submerged leaves are thirty times as long as they are broad, while the length of the aerial leaves is only 7-9 times their width. In Rori/pa amphibia the leaves which develop under water are deeply cleft compared with those produced in the air. The aerial leaves of this Crucifer are linear-lanceolate, about ten times as long as broad, with finely toothed margin. Under water the leaves have an elliptical shape, are 2-3 times as long as broad, and the lamina is cleft almost down to the midrib in narrow segments 2-3 cm. long, like a comb or feather. The aerial leaves of the whorled Waterwort {Elatine Alsinastrum) are grouped in whorls of three. They have an ovate shape, and their margins are finely notched. Each is traversed by 3-5 veins. The leaves developed under water are divided almost their whole length into 3-4 narrow linear segments, and each whorl looks as if it were composed of twelve leaves. Each segment is smooth round the edge, and traversed only by one central vein. The diflerence between the aerial and sub- merged leaves of the white-flowered Crowfoots (belonging to the Batrachium section of the genus Banunculus) is even more surprising. Plants of these Crow- 506 DEPENDENCE OF PLANT FORM ON SOIL AND CLIMATE. foots which have developed on muddy but not inundated ground display three- or five-cleft leaves whose segments are light green in colour, shiny, and almost fleshy, and spread out flat. "When these plants are grown under water the leaves appear quite different; they become divided into numerous thread-like or hair- shaped segments which have a dark-green colour, and the polished surface has entirely disappeared. The shade afibrded by stones, loose earth, undergrowth, and neighbouring bushes and shrubs acts on growing stems, foliage-leaves, and flowers just in the same way as the light-subduing layer of water. In a place near my country house which was formerly used for storing wood and dry twigs, but which had remained unused for a long time, the Creeping Thistle (Cirsium arvense) had established itself and formed an intricate growth. The crowded stems attained a height of 80 cm. at the time of flowering and fruit ripening. In the winter of 1885 wood was again stored there in piles 150 cm. high. When, early in the following summer, the new shoots of the Thistle began to spring up they were obliged to content themselves with growing through the dark chinks between the blocks of wood. Many were thus forced to bend and twist, and finally came against some insur- mountable obstacle so that they dwindled in the crevices of the wood-stack without ever reaching the light. Others again which were able to find a fairly straight road through the crevices grew up until they reached the surface of the wood-heap, they then continued to grow 50 cm. higher and unfolded large foliage-leaves on this upper portion. They also developed branches with flower-heads, and from a distance it looked as if a group of Thistles had grown on the top of the wood- stack. The stems had attained a height of 2 metres. The lower internodes were twice as long as usual, the foliage-leaves which sprang from the stalk inside the dark crevices were small, yellowish green, and the buds in their axils did not develop. The Cow-berry (Vaccinium Vitis-IdcBo) behaves similarly when its shoots are obliged to grow up to the light through dead tree-trunks. Shoots which force their way in the dark between the bark and the wood of the trunk may reach the height of a metre, while neighbouring ones, springing directly from the soil of the forest are only 15 cm. high. The shoots inside the bark have a reddish colour, and they bear small pale scales instead of dark-green foliage-leaves. From the creeping stems of the White Clover (Trifolvwm repens) spring erect petioles terminating in three leaflets, and an erect angular stem bearing a flower- head. In sunny places, especially where no neighbouring plants cast a shade, the petioles reach a length of 8 cm., and the stem of 10 cm. But if dense bushes overshade the Clover, the petiole and stem elongate until the leaflets and capitulum they bear reach the light. Under these conditions petioles 28 cm. long have been found, and stems attaining a height, of 55 cm. An extraordinary elongation also occurs in the radical leaves of the Dandelion (Taraxacum officvnale) in places where high Grasses and thick bushes shade the moist soil. In the open the leaves reach a length of 20 cm., but in the shade they become twice or three times as long. The lower part of the leaf lengthens most, the free end is comparatively INFLUENCE OF LIGHT. 507 very little altered, and in the central portion the only change is that the lobes and teeth become shorter and less clearly marked. In order to ascertain the effect of covering plants with earth, numerous bulbs of a species of Tulip (Tulipa Gesneriana) were planted at the same depth in one garden bed, and in another some corms of the Spring Crocus (Crocus vernus). Earth was heaped over these bulbs and corms in successive heights of 5, 10, 16, 20, 25, 30, 35, 40, 45, and 50 cm. Naturally the leaf -tips and flower-buds were first seen in the places where the bulbs were only covered with 5 cm. of soil; in both beds the development was delayed — in the other cases in proportion to the height of the soil above the bulbs. Some flower-buds of the Crocus appeared above the 20 cm. of soil, one of the Tulip above the 30 cm. Numerous leaf -tips of the Crocus appeared above the 35 cm., and a few of the Tulip above the 40 cm. of soil. The perianth-tube, the peduncle and the foliage-leaves were almost twice as long as those which had developed under only 5 cm. of soil. The flowers were smaller, and unfolded just above the soil; the leaves were narrower and pale yellow in colour as far up as they were covered with the soil. Neither the Crocus nor the Tulip raised their leaves higher than 40 cm. Apparently the reserve-materials stored in the corm and bulb-scales were not sufficient for a further elongation. The stems and leaves of the Crocus and Tulip thus exhibit alterations similar to those observed in the sprouts of Potato-tubers in a dark cellar. We should expect that if moisture and lack of light produce elongation of shoots and various alterations in leaves, a brilliant illumination would have the opposite effect on growing plants. This is indeed the fact. Plants which have been for a year in the shade and have been placed at the beginning of their development in the following year in the sun display shorter internodes and firmer leaves; they blossom more abundantly, the flowers are of a deeper hue, and in many cases a covering of hairs is formed over the green portions. It is not necessary to mention how far transpiration, which is much more active in the sun than in the shade, is concerned in this; these alterations are certainly produced in the end by sunlight. The effect of brilliant illumination is best seen by comparing plants grown from similar seeds at different elevations, but under identical conditions in other respects. The results obtained in my experimental garden near the summit of the Blaser in the Tyrol, at a height of 2195 m., during the years 1875-1880 illustrate this very fully, and I will briefly recount them here. The seeds of some annual plants were sown in September. The beds were covered with a layer of snow a metre thick throughout the winter. The germination of the seeds took place in the following year soon after the snow melted between the 10th and 25th June. The seedlings therefore developed during the time when the sun was highest and the days longest, and the young plants were exposed to a temperature not lower but rather higher than that enjoyed by plants from similar seeds which began to develop in the experimental beds of the Vienna Botanic Garden in March, when the daylight lasted about 12 hours. The seedlings of several species (e.g. Gilea tricolor, Hyos- cyamus albus, Plantago Psyllium, Silene Oallica, Trifolium incarnatum) were 508 DEPENDENCE OF PLANT FORM ON SOIL AND CLIMATE. killed by the isolated frosts which occurred in each of the six years of the experi- ment, not only in the last week of June, but during July and August; but others j {e.g. Agrostemma Githago, Gentaurea Gyanus, Iberis amara, Lepidium sativum, ,|. Satv/reja hortensis, Senecio vulgaris, Turgenia latifolia, Veronica polita, Viola = arvensis) only underwent a short temporary stoppage of growth from this cause, and opened their flowers at the end of August and beginning of September. In the plants of some species (e.g. Senecio vulgaris, Veronica polita, Viola arvensis) i ripe seeds capable of germinating were formed in September. The flowering speci- mens, in comparison with those which had grown during the short days of the spring exposed to numerous night-frosts in the Vienna gardens, displayed extremely shortened intemodes. The number of intemodes was also lessened, or rather, fewer were developed. For example, where 10 intemodes developed in an experi- mental plant in Vienna, in the Alpine garden a corresponding plant would only have 5-6. The same was true of the development of the flowers. While in a plant of Viola arvensis in Vienna the axillary buds of the first six foliage-leaves were suppressed and flowers were not produced until the seventh and eighth leaves, flowers grew from the third and fourth axillary buds in the same species of plant grown in the Alpine experimental garden. The number of flowers on a plant was less, the petals were smaller on the average, and, generally speaking, the annual plants in the Alpine garden had the same appearance as those grown in the plain on dry, sandy soil described on p. 600. It has already been stated on p. 453 that some of the species which are annuals in the valley and on the plain do not die in the autumn in the Alpine garden, but remain alive through the winter and in the following year develop new shoots from the stem. To describe the alterations undergone by biennial species in Alpine regions we will take Libanotis montana (an Umbellifer) as an example. Its stem in the Alpine garden was 16-24 cm. high and developed 5 intemodes which were 2-5 cm. long. From the axils of the 5 green stem-leaves sprang lateral shoots which did not branch but terminated in a single umbel, so that the plant only bore 5 umbels altogether. The plants grown from similar seeds in the Vienna Botanic Garden exhibited a stem more than a metre high with 10 intemodes each 10-20 cm. long. No lateral shoots were produced from the axils of the lower stem-leaves. Those from the axils of the middle and upper leaves were branched and bore several umbels. On an average a plant had about 20 umbels altogether. Over 300 species of perennial plants were grown in the Alpine experimental garden. Only 32 of them blossomed, however. Those whose flowers usually pre- cede the foliage-leaves were in full blossom at the beginning of July, the others, which had to develop a leafy stem before their flowers appeared at the top or in the axils of the leaves of this stem, did not flower until the end of August and beginning of September. Three species of the latter kind will be more particularly treated of here; one species whose stem bears only a single leaf and is terminated by a single flower (Parnassia palustris), one whose stem is beset with decussate leaves and terminates in a loose inflorescence composed of small cymes (Lychnis INFLUENCE OF ELEVATION. SOQ' Viscaria), and a third whose stem bears alternate leaves and whose flowers are grouped in capitula (Pyrethrum corymbosum). The Grass of Parnassus (Parnassia palustris) from the Alpine garden, when compared with plants grown in the experimental beds of the Vienna Botanic Garden, showed the following measurements: — Vienna Botanic Garden. Experimental Garden on the Blaser, Height of stem 20-27 cm. 3'3 cm. long, 2'4 cm. broad. 2-8-3-4 cm. 5-9 cm. I'D cm. long, 06 cm. broad. 1-8-2-0 cm. Diameter of flower In the Alpine regions, therefore, the plant was only ^ or :| as high and the leaves only \-\ as large as in Vienna, whilst the flowers in the Alpine region had a much smaller diameter than in Vienna. Comparing the hermaphrodite plants of Lychnis Viscaria in the experimental garden of the Blaser with those of the same species at Vienna, we obtain the following: — Vienna Botanic Garden, Experimental Garden on the Blaser.' Height of the stem, including ) the axis of the inflorescence \ Dimensions of lower leaves Inflorescence Calyx 400-500 mm. 80 mm. long, 4 mm. broad. 80 „ „ 50 „ „ 15 „ „ 6-5 „ „ 1^ jj )) " » )) 8 mm. long. 230-240 mm. • 50 mm. long, 3 mm. broad. 60^ „ „ 40 „ „ l"*o )j )j 5 „' „ 8 )> )) 6'8 „ „ 7 mm. long. Lamina of petals Plants from the Alpine garden, therefore, when compared with those from the Vienna Botanic Garden, exhibit smaller measurements of stem, leaves, and flowers. The following points were also noted: the number of internodes in plants from the Vienna Garden was 9, of which 5 were on the axis of the inflorescence; each cyme consisted of 3-6 flowers, and the whole inflorescence bore 33-40 flowers. Plants from the Alpine Garden had only 6-7 internodes, of which 3 belonged to the inflorescence; the cymes composing the inflorescence were only occasionally 3-flowered; in most of them only the central flower developed, the two lateral ones being suppressed. The whole inflorescence included only 5-11 flowers. Plants of Pyrethrum corymbosum, from the Alpine Garden, compared with those from the Vienna Botanic Garden (all raised from one batch of seeds) showed the following differences: — Vienna Botanic Garden. Experimental Garden on the Blaser. TTfifrht of tlie stem 950 mm. 170 mm. long, 50 mm. broad. 26 mm. 8 mm. long, 4 mm. broad. 250 mm. 45-50 mm. long, 20 mm. broad. 20 mm. 7 mm. long, 3 mm. broad. Diameter of the capitulum,.... 510 DEPENDENCE OF PLANT FOKM ON SOIL AND CLIMATE. In this case, again, plants from the Alpine Garden, when compared with those of the Vienna Botanic Garden, had smaller stems, leaves, and flowers. The lobes of the foliage-leaves from the middle of the stem of plants from the Alpine Garden were pinnate, and the pinn^ were either entire or beset with two small teeth on each side, near the apex. The stem had ten foliage-leaves, the four uppermost of which were much reduced and served as scale-leaves for the lateral shoots arising from their axils. These lateral axes were not branched, and each bore only a single capitulum. There were five capitula altogether. On plants from the Vienna Botanic Garden the lobes of the foliage-leaves from the middle of the stem were more divided, and the pinnae were beset on each side near the top with 3-5 teeth. The stem bore 25-27 foliage-leaves, of which the 6-8 upper ones were much reduced in size, and functioned as scale-leaves for the lateral shoots in their axils. These lateral shoots were branched, and each branch terminated in a capitular kifiorescence. The total number of capitula was 20-30. From these examples it will be seen that all the parts of plants grown in the Alpine experimental garden were much hindered in their growth. The foliage- and floral-leaves were smaller, the stems shorter, the number of intemodes, foliage -leaves, inflorescences, and flowers was diminished. The flowers were relatively nearer the earth, and this was due not only to the diminished number and length of the internodes of the stem, but principally to the fact that the flowers sprang from the axils of the lower stem-leaves. Plants growing in Alpine regions derive a great advantage from these altera- tions, which are chiefly produced during their development by the influence of the long and bright daylight of June, July, and August. If these plants had to produce the same under-structure as their fellows in the Vienna Botanic Garden, 2015 metres lower down, much time would be lost, and the earliest flowers would hardly open before October, at a time when the winter snow is already beginning to fall. But since the number of internodes is restricted, and flowers are developed from the lower stories, it is possible for the plants to blossom at the end of August and beginning of September, and perhaps to ripen their fruits — one of the chief aims of the plant's existence. To this modiflcation in their mode of development is also due in part the repeatedly -mentioned fact that many alpine plants blossom earlier than those in lower regions. But in order to avoid misunderstanding, it must be expressly stated that in not one of the thirty-two perennial, nor in the biennial and annual species which blossomed in the Alpine experimental garden, was the early flowering hereditary; con- sequently these plants must be carefully distinguished from the so-called asyn- gamic species, which will be spoken of in one of the last chapters in this book. The relation of light to the colouring matters of plants has been repeatedly the subject of careful investigation. All observers agree that the amount of the pigment known as anthocyanin increases and diminishes with the stronger or weaker sunlight enjoyed by the parts of the plant in question, and that the yellow colouring matter of flowers holds a similar relation. Chlorophyll, however, COLOUR AT DIFFERENT ELEVATIONS. 511 is actually destroyed by bright light in plants which are not properly screened, and the green tissue is then blanched and assumes a yellow tint. Since the intensity of the sun's rays increases with the elevation in mountain districts (see vol. i. p. 525), we should expect that this effect of light would be shown particularly well in plants of high elevation. And this is certainly the case. The flowers of species grown in the Alpine garden on the Blaser at a height of 2195 metres above the sea exhibited, as a rule, brilliant floral tints, and some were decidedly darker than the flowers grown in the Vienna Botanic Garden. Agrostemma Oithago, Campanula pusilla, Bianthus inodorus (sylvestris), Gypso- phila repens, Lotus comiculatus, 8aponaria ocymoides, Satureja hortensis, Taraxacum, officinale, Vicia Gracca, and Vicia sepium are good examples of this. Several species, which produced pure white petals in the Vienna gardens, e.g. Lihanotis montana, had petals coloured reddish-violet by anthocyanin on their under sides in the Alpine garden. The glumes of all the Grasses which were green, or only just tinged with violet at a low level became a dark brownish- violet in the Alpine garden. The abundant formation of anthocyanin in the green tissue of the foliage-leaves and sepals, and in the stem, was particularly apparent. The leaves of the Stonecrops, Sedwm acre, album, and sexangulare became purple-red, those of Bracocephalum Ruyschianum and Leucanthemum, vulgare violet, those of Lychnis Viscaria and Satureja hortensis a brownish- red, and the foliage-leaves of Bergenia crassifolia and Potentilla Tiroliensis, even in August, had the scarlet-red colour which they usually assume in sunny spots in the valley in late autumn. I must not omit to mention that, according to some of my zoological friends, many animals, especially spiders and snails, which have been transferred from the plains to the mountain-heights, assume a darker tint in alpine regions. A considerable number of plant species, especially those which grow in the valley in shaded or half -shaded places, as, for example, Arabia procurrens, Bigitalis ochroleuca, Geum urbanum, Orobus vernus, Valeriana Phu, and V. simplicifolia, Viola cucullata, developed more or less yellowish leaves in the Alpine garden, where they were exposed to the full sunlight. It was mentioned in vol. i. p. 393, that the Flax (Linum usitatissimum), which flourishes in mountain valleys at a height of 1500 metres, where its chlorophyll is uninjured, nevertheless turns yellow in the Alpine garden at a height of 2195 metres; From this general review of the modiflcations in plant-form obtained by culture-experiments, a series of important conclusions may naturally be drawn. In the first place we must point out that two kinds of characters are to be observed in plants, those which are the result of certain conditions and properties of soil and climate, and those which appear independently of these external influences. This distinction is so important that we shall illustrate it by two examples. The white Water-lily, Nymphcea alba, develops scale-leaves of ovate or lanceolate shape with no separation into petiole and lamina. The foliage-leaves, however, 512 DEPENDENCE OF PLANT FOEM ON SOIL AND CLIMATE. have a rounded petiole and a disc-shaped lamina. These characters are always present whether the seed which produced the plant germinates in a deep lake or in the mud of a marshy meadow. In the marshy meadow the scale-leaves remain short, and the walls of their epidermal cells thicken in a remarkable way; the petioles of the aerial foliage-leaves become about a span long, and, in order to increase their resistance to bending, a strong layer of bast arises, the thickness of these bast-layers amounting to 017 mm. The walls of the epidermal cells are thickened, 5-9 layers of collenchymatous cells are formed under the epidermis with walls -007 mm. thick, and the air-spaces in the centre of the leaf- stalk are much narrowed. But if this species of Water-lily grows under water, the scale-leaves elongate into long and flaccid ribbons, and the petioles of the foliage-leaves continue to grow until their blades are raised to the surface of the water. According to its depth they attain a length of 30, 40, 50-100 cm. Resistance to bending is but little required by the petioles, which are surrounded by water, and the bast is therefore only slightly developed. The strings of bast which traverse the leaf-stalk are only O'll mm. thick, the walls of the epidermal cells are only half as thick as in the aerial leaves, only 3-6 layers of collenchyma are developed below the epidermis and the air-spaces in the centre of the leaf -stalk have a diameter of over half a millimetre. These petioles are consequently flexible, and cannot support the leaf -blade if taken out of the water. The general form of the scale- and foliage-leaves, the segmentation of the latter into petiole and blade, the configuration of the blade and the distribution of the bundles in it are all the result of internal forces due to the specific constitution of the protoplasm; but the thickness of the epidermal cells, the strength of the mechanical tissue, and the length of the leaf-stalk, are determined by the depth of the water-covering. The same thing is seen in the flowers; their structure depends upon the specific constitution of the protoplasm, but the size of the petals is determined by the temperature of the water. The Meadow-grass Poa annua has a rapid growth; its haulms and leaf -sheaths are round, the leaf -lamina is traversed by seven strands, the lower branches of the inflorescence are single or paired but never whorled, and the spikelets of the panicle are much compressed and egg-shaped in outline. These characters are unalterable and are observed in Poa annua under all conditions. But when the haulms growing in the gardens in the plain project beyond the short upmost leaf the spikelets become 6-7-flowered, and have a pale green colour. When the plants become perennial in alpine regions the haulms bend towards the ground and remain so short that they do not reach above the highest foliage-leaf; the spikelets develop only 3-4 flowers, and their glumes are dark violet on the surface and brownish-yellow at the edge; thus these modifications are in relation to peculiarities of situation (in the plain and alpine regions) as eflPect to cause, and are to be ascribed to the influences of heat, light, and moisture, which act in various ways according to the situation. These alteratio^s are always to the advantage of the plant. They make the ARE THESE CHANGES IN FORM PERMANENT? 513 individual more resistant, support and protect its organs, and render it possible for the separate parts to perform their work in spite of the necessarily altered conditions. They seem to have the task of keeping the plant alive under very different vital conditions, of promoting growth and the formation of offshoots and fruit with the smallest possible expenditure, and they may therefore be regarded as adaptations to the particular conditions of soil and climate. The capacity for adaptation is of course founded in the specific constitution of the protoplasm, and is very different in different species. One species may adapt itself by appropriate alterations to the influence of bright light, submersion under water, a dry atmosphere, &c., while another cannot do so. If the protoplasm of the Flax {Linwm usitatissimwm) could manufacture as much anthocyanin in its green tissue as the Summer Savory {Satureja hortensis) it would blossom and ripen its fruits in alpine regions as this plant does, and would not succumb to the effect of the strong light. If the protoplasm of the Common Bent-grass (Agrostis vulgaris} were able to continue its constructive activity under water it would not perish as soon as it is submerged, but would maintain itself like the stoloniferous species {Agrostis stolonifera) by green stalks and leaves adapted to an aquatic habitat. In short, the adaptability of each species is restricted within definite limits which depend upon the specific constitution of the protoplasm and cannot be overstepped. It is a matter of great import in the history of species whether modifications in form effected by change of soil and climate are transmitted to the descendants, and whether they can be inherited. This of course can only be ascertained by experiments, and by experiments in which all possible sources of error have been eliminated. This last remark is made advisedly, for the sources of error in such experiments are very numerous. I will briefly indicate two which interfered with some experiments I carried out in the years 1863 and 1864. It is not enough to be careful that the seeds sown in the prepared experimental beds are all from the same plant; care must also be taken to see that they are not the result of a hybrid cross-fertilization. Some seeds taken in 1863 from a plant of Bianthus alpinus growing in the Botanic Garden at Innsbruck, and sown in different soil in two experimental beds, produced plants in soil free from lime, which, in their external appearance, agreed with Bianthus deltoides. It seemed as if Bianthus alpinus, a lover of limestone rock, had become transformed into Bianthus deltoides when grown without lime. The seeds of the plant so like Bianthus deltoides were again sown in soil without lime, but the resulting plants no longer resembled this species; they showed themselves to be constant in their characteristics. The whole experiment with Bianthus alpinus was then repeated, but this time the plants on the clay soil without lime did not change, and I was obliged to conclude that the plant I had regarded as a stage in the transformation of Bianthus alpinus into Bianthus deltoides was a hybrid of these two species. In order to be certain about this a crossing between the two species was effected artificially. From the resulting seed plants were actually grown which were exactly like those I had regarded as transformations, and there was no longer any doubt that some of the VOL. II, 83 514 THE INFLUENCE OF MUTILATION ON THE FOEM OF PLANTS. stigmas of the Bianthus alpinus which had yielded the seeds for the first experi- ment had been pollinated by insects with the pollen of Bianthus deltoides. Mistakes often arise also from the fact that the young stages of many plants are very different from the fully-grown specimens. Young Birches grown from the seeds of Betula verrucosa bear leaves which are simply serrated, thickly covered with hairs, and soft to the touch. They are deceptively like the leaves of adult plants of Betula alba or pubescens. The leaves of the adult Betula verrucosa have quite a different form; they are doubly serrated, smooth, and harsh to the touch. These latter are the only form of leaf described in Botanical books for Betula verrucosa. Anyone sowing the seeds from a grown tree, and watching them grow up, with leaves of a different shape and surface, might easily think an actual fundamental change had occurred, and might be tempted to regard the transforma- tion as the direct effect of a change in external influences. It is perhaps superfluous to state that due regard was paid to these possible sources of error in the later series of cultural experiments, carried out during six years in the Alpine garden on the Blaser (2195 metres), and for comparison in my Villa Marilaun in the high-lying Tyrolese Gschnitzthal (1215 metres), in the Botanic Garden at Innsbruck (569 metres), and in the Botanic Garden of the Vienna University (180 metres); in no instance was any permanent or hereditary modifi- . cation in form or colour observed. Seeds of a plant grown in the valley when sown in the Alpine region produced plants which exhibited the modifications described above. They were also mani- fested by the descendants of these plants but only as long as they grew in the same place as their parents. As soon as the seeds formed in the Alpine region were again sown in the beds of the Innsbruck or Vienna Botanic Gardens the plants raised from them immediately resumed the form and colour usual to that position. The modifications of form and colour produced by change of soil and climate are therefore not retained in the descendants; the characteristics which appear as the expression of these changes are not permanent, and the individuals are to be there- fore regarded as varieties, of which Linnseus says in his PhUosophia Botanica: " Varietates tot sunt, quot differentes plantse ex ejusdem speciei semine sunt pro- ductse. Varietas est Planta mutata a caussa accidentali: Climate, Solo, Galore, Ventis, &c., reducitur itaque in Solo mutato." THE INFLUENCE OF MUTILATION ON THE FORM OF PLANTS. When Birches and Firs grow up side by side in a wood-clearing, the crowns of the Birches will overtop the Firs in some twenty years' time, and this will seriously interfere with the growth of the latter. With every blast of wind the whip-like branches of the Birch strike against the upper shoots of the Firs, so that these gradually wither and die off. A lateral branch of a Fir tree altering its direction of growth and replacing the dead leader will, in its turn, soon be scourged to death. The top of the Fir is permanently mutilated, and the injury THE INFLUENCE OF MUTILATION ON THE FORM OF PLANTS. 515 can be recognized years after by the flattened form of the crown, so different from the usual appearance, when the offending Birches have perhaps long dis- appeared. Many other trees wage the same war with one another, the result in each case being the mutilation and alteration of the form of the summit of one of the trees. The Maple, for example, is either put quite hors de combat by the long thorny branches of a neighbouring Gleditschia (Gleditschia triacanthos) or else the crown becomes lop-sided owing to the destruction of the branches on the side facing the Gleditschia. The way in which the appearance of Firs, Larches, Beeches, and Ling is altered by the attacks of ruminants, especially goats, was described in vol. i. p. 445, and we may add here that Pines and Junipers are mutilated in the same manner. The consequence is that lateral branches, which would not otherwise develop, grow out in the following year from the base of the twigs which have been bitten off Apparently no other alteration takes place in these plants. But when huge boughs are broken off close to the ground by storms and the weight of snow, when the tree-trunks of the forest are sacrificed to the wood-cutter's hatchet, and the stems of seedling trees and shrubs in the meadow to the mower's scythe, when all the young shoots are frozen by a night's frost in spring, or when all the leaves are devoured by caterpillars and the branches are left bare as in winter — then the consequences are much more serious. In these cases new shoots make their appearance either from " eyes " in the stem or from the reserve-buds of the branches and twigs, or by buds produced by the roots below the ground. The leaves of these shoots, or suckers, as they are called, differ very much from those of the branches which have been broken, eaten, cut, or frozen off. The leaves from the crown of the Aspen (Populus tremula) are stiff and smooth in their adult condition; the circular blade is borne on a long petiole, and its margin is coarsely notched and undulated. The lateral veins traversing the blade are lost in a network near the edge in which no strong curved ribs occur. The leaves of a sucker from the base of a mutilated stem, or from the root, are soft and thickly covered on both sides with downy hairs; the heart-shaped blade is borne on a short stalk, and the margin is beset with numerous upwardly-directed notched teeth. The lateral veins of the blade merge near the edge of the leaf into a network, in which strong curved ribs are plainly visible. The leaves from the crown of the Oak (Quercus pedunculata) are deeply lobed and furnished with two so-called auricles at the base; those of the suckers are quite entire or very slightly lobed, with no auricles at the base. The leaves of the sucker of the common Beech (Fagus sylvatica) are more or less plainly serrated at the edge, while those of the topmost branches of the tree are quite entire. In the Black Mulberry {Morus nigra), and in the Paper Mulberry (Broussonetia papyrifera), the leaves of the sucker have a sinuous margin and are more or less deeply lobed, but those of the tree-top are heart-shaped with notched margins and no lobes. The leaves of the sucker of the Birch (Betula verrucosa) are simply serrated, with velvety hairs; those on the crown of the tree are doubly serrated and 516 THE INFLUENCE OF MUTILATION ON THE FORM OF PLANTS. smooth. The leaves on the suckers of the Round-eared Willow (Salix aurita) are broadly ovate, fairly smooth, and the veins in the blade form a wide-meshed reticulum; the leaves on non-mutilated branches are widened in the upper third, strongly wrinkled, and covered with grey hairs, whilst the reticulum of the veins is narrow-meshed. In Salix rosmarinifolia, the leaves of the suckers are twice or three times as broad as those of the normal branches, and they are smooth, while those of ordinary branches are covered with silky hairs, and gleam like silver. Hundreds of trees and shrubs might be mentioned in which there is a distinct diiference between the foliage of the suckers and of the normal branches of the crown. But these few examples will suffice, and we will only mention the Norway Maple {Acer platanoides), because the difference in the foliage-leaves can be seen from the illustrations in vol. i. The leaves of the summit (see vol. i. fig. 106, p. 416, and fig. 109, p. 419) are borne on long petioles, the blade is 5-7 lobed, and the lobes are short and beset with several pointed, tapering teeth. The leaves of the suckers in this same Norway Maple are short-stalked, the blade is slightly 3-lobed, and each lobe is triangular and without the elongated pointed teeth. They exactly resemble the first foliage-leaves shown in vol. i. p. 9, fig. 1 ^. This is also true of the leaves on the suckers of other woody plants. The shoots developed from reserve buds, " eyes ", and the like, repeat to a certain extent the beginning of the leafy stem, so that the phenomenon is only an exhibition of the usual metamorphosis of the foliage-leaves. The difference between the older and younger, i.e. lower and upper foliage-leaves, only seems strange because the twa kinds of leaf-forms are not usually seen simultaneously on one and the same plant. By the time the crown of a tree has developed, the first (oldest) leaves which adorned the young sapling have long disappeared. Many descriptive Botanists, as a rule, only consider the foliage-leaves of the fully -grown trees and bushes; some of them have hardly ever seen the first leaves of the commonest trees, and when they do happen to come across them they regard them as an extraordinary phenomenon, declare the shoots bearing them to be " bud variations ", and draw bold and bewildering hypotheses from their appearance. This alteration in form, however, has nothing to do with the formation of varieties, nor is it dependent either upon the infiuence of the soil or upon the effect of climate. Moreover, the form of leaf characteristic of the sucker is not possessed by the secondary shoots which arise from the suckers; these are adorned with the same foliage which occurs on the topmost branches of the tree. Alterations in the scale-leaves as well as in the foliage are brought about by mutilation of the branches. When the upper portions of Willow boughs with their foliage-buds are cut off, leaving the lower portions with the buds of the flower-catkins on them, the small pale scales at the base of the catkins change into green foliage-leaves; the axis bearing these leaves elongates, and the catkins then form the termination of a leafy shoot. Many Willows, e.g. Salix cinerea and S grandifolia, by this metamorphosis assume a very unusual appearance. In the following year the branches bearing the flower- catkins, if they are THE INFLUENCE OF MUTILATION ON THE FORM OF PLANTS. 517 not mutilated afresh, will again put out short catkin-stalks with small ' pale scales. Mutilation of herbaceous plants is caused by herbivorous animals, viz. insects and mammals, and on a large scale by man when he mows the meadows and cuts the crops and makes other necessary invasions on the natural vegetation in the interests of husbandry. The alterations caused by these mutilations of the foliage-leaf region are in the main the same as in woody plants. From the remaining stumps of the stem lateral shoots arise whose first leaves are like the first leaves of the seedling. Usually they are less divided and have fewer hairs than the leaves on shoots of normal plants, and on this account they have a very diflferent character. In the floral region the effects of mutilation are twofold — first the peduncles or the lateral axes which are terminated by inflorescences elongate, and then the flowers become smaller. For example, when a vigorous stalk of the Ox-eye Daisy {Chrysanthemwrn Leucanthemum) bearing a capitulum is cut off" close to the ground, long lank lateral stems develop from the axils of the lowest remaining leaves, each one ending in a capitulum. The main stem is now seen to be branched at its base, which is never the case in normal plants. If about half the stalk of the common Foxglove is cut off" in the spring long flower-racemes will arise from the axils of the leaves just below the cut, but the flowers will be only half as large as those which would have developed on the uncut main stem. The stem of Althcea pallida rises a metre above the ground if its development is not hindered, and forms fascicles of short-stalked flowers in the axils of the upper leaves. If the stem is broken ofl" lateral axes develop from the axils of the remaining leaves, and bear little long-stalked flowers. Particularly good examples are furnished by the annual weeds Delphinium Ajacis, Nigella arvensis, Stellera Passerina, and the like, which grow up amongst cereals. Their main stems are broken off" when the corn is cut, and they then develop comparatively long branches with small flowers from the remaining stumps. If only single flower- buds, and not the whole inflorescences, are removed from a herbaceous plant whose main stem terminates in a long raceme, so that each flower is cut away in turn from below upwards just before it opens, the rachis of the raceme elongates enormously and flower-buds are developed at its end which would certainly not have unfolded had there been no mutilation. In the Red Foxglove, for example, the rachis of the raceme which has been damaged in this way will grow to twice its ordinary length, and twice as many flowers will be developed. The last and highest flowers in such racemes, however, are only half the size of those which arise on normal racemes. We must now consider certain perennial meadow plants which when mown down are stimulated by the mutilation to develop flower-stalks in the same year, which would, in the normal course of things, not have flowered till the year follow- ing. In Alpine valleys it is a very common thing for the flowers of the spring plants AneTnone vernalis, Geraniiim sylvaticum, Gentiana verna, PolygonuTn Bistorta, Primula elatior and P. farinosa, Trollius Europceus, &c., to appear in 518 ALTERATION OF FORM BY PARASITIC FUNGI. the autumn in meadows which have been mown in the spring. The flowers appear- ing under these circumstances are remarkable for their small size. Their diameter is at least a third smaller than that of the spring flowers. In conclusion we may refer to the gardener's artifice which has already been described (p. 453) of pro- ducing perennial plants with woody stems from an annual Mignonette plant by mutilation. We might also mention the dwarf shrubs and trees produced by combined mutilation and grafting, especially the strange-looking little Ivy treea obtained by grafting a flowering branch of Ivy on an erect stem a span high, and the dwarf Conifers so much in favour with the Japanese. Gardeners and descriptive Botanists have frequently determined and described mutilated plants as other species, hybrids, or varieties. They are neither the one nor the other. The peculiar appearance of the altered members resulting from mutilation is exactly determined beforehand in each species; it is due to the specific constitution of the species, and thus is part of its being. It is not produced by the external influences ■ which lead to the formation of varieties, but is brought about by inherent necessity quite independent of the influences of climate and soil. ALTERATION OF FORM BY PARASITIC FUNGI. A considerable number of the trees and shrubs of Central and Southern Europe bear bristling, much-branched structures on some of their boughs which, from a distance, look like large birds' nests or brooms, and which have been popularly termed "witches' brooms". They are the outward and visible signs of a disease from which the plants in question suflTer, and, as their name testifies, their origin was thought to be connected with witches. Traditionally witches have the power of "wishing'' harm to mankind, animals, and plants; and superstitious people, at the sight of these peculiar pathological structures on the trees, may have started the idea that the disease was caused by witches that they might have brooms ready at hand for their midnight ride on the Brocken. Other plant diseases have been ascribed to unusual conditions of weather, especially to long-continued rain or great drought. It is not long since the discovery was made that most of the diseases attacking trees, shrubs, and herbs are caused by Fungi, and that atmos- pheric conditions are only concerned in the matter in so far as they hinder or favour the establishment and development of these parasites. All the Fungi in question are parasites. They penetrate into the tissues of the host-plant and sooner or later cause the death of the aflfected part, and frequently of the entire host-plant. The living protoplasm in the cells and tissues of the host which is influenced by the parasite undergoes fundamental changes in its com- position. Some of the cells are drained, their living protoplasm being consumed, so to speak, and these cells are obviously marked for destruction. Others are not killed, but changed. The metamorphosis occurs, in the first place, in the consti- tution of the living protoplasts which have not yet completed their development, the change much resembling that known as fermentation in fluid substances TYPES OF FUNGAL GALLS. 619 (c/. vol. i. p. 508). In fermentation the chemical composition of the fluid is altered, its chemical compounds are shaken, decomposed, and split up and new compounds are formed by the action of the living Yeast cells. The same thing happens here in the interior of the living plant in its turgid, meristematic tissue — that is to say, in a group of protoplasts which still have the power of growing at the expense of materials supplied them, of increasing in size, and of multiplying by division. But these cells no longer behave as — in the absence of the parasite — they would have done. Profoundly modified under the influence of the parasite, but yet not killed, these cells, by their continued division, form tissues and organs of new and unusual form; in other words, that part of the host which is invaded but not killed by the parasite will continue to grow and increase in size, and in consequence of the change which its protoplasm has experienced will assume a different outward form. These altered tissue-bodies produced by parasitic Fungi are called gall-structures. They are usually characterized by an excessive growth known as hypertrophy, as well as by their altered shape. The hypertrophy is without doubt caused by a stimulus proceeding from the parasite. We may conclude that the significance of the increased growth lies in the abundant supply of nourishment thus placed at the disposal of the parasite, since the large quantity of food-material brought for the excessive development of the hypertrophied growth connotes a large supply for consumption by the parasite. In many cases, however, the hypertrophied tissue merely forms a wall protecting the host against the further depredations of the intruder. It then contains no nourishment for the use of the parasite, being built up chiefly of corky cells, which the latter cannot consume or destroy. Such a tissue might be compared to the so-called callus which grows up in plants in parts de- prived of epidermis after an injury, or in other wounds, and gradually covers them over with a protective layer. The formation of the gall is often restricted to only a STnall portion of the afflicted plant; in other cases whole leaves and branches, and sometimes even ex- tensive shoots, become modified in shape. To get a general idea of the four types of hypertrophied growths it will be best to take them one after the other in the order mentioned, commencing with the simplest. The simplest of these galls consist of a few degenerate and metamorphosed cells in the centre of an extensive and unaltered tissue. They are produced chiefly by parasites of the genera Rozella, Synchytrium, Exdbasidiwm, and Gymnosporan- gium. Rozella septigena, one of the Chytridiese, develops swarm-spores which attack the various species of the fungal genus Saprolegnia. They settle on the tubular branches of the Saprolegnia at a place where it was just about to divide and to produce swarm-spores of its own. In consequence of the invasion of the parasite this does not take place, but the tubular cells which would have formed a Saprolegnia-STporangiwca. divide instead into short barrel-shaped cells, each of which becomes a sporangium of Rozella septigena. In addition to this the infected cells develop lateral outpushings which swell up spherically, and each contains a resting- 520 ALTERATION OF FORM BY PARASITIC FUNGI. spore of the parasite. Parasitic species of Synchytrium cause a vesicular enlarge- ment of single cells of the epidermis in the leaves of phanerogamic host-plants. The not uncommon species Synchytrium Anemones and 8. Taraxaci produce only a slight overarching, and the enlargement of the cells is hardly more than four times, often only twice the usual size. But, by the influence of Synchytrium Myosotidis, hypertrophied epidermal cells rise up from the leaves of the Forget- me-not (Myosotis) in the form of comparatively large, club-shaped, bottle-like, or egg-shaped bladders of golden or reddish yellow colour, and each contains the parasite, or rather its spores. The parts of the leaf attacked by Synchytrium Myosotidis are also much thickened, the paHsade cells and the air-containing lacunae of the spongy parenchyma (of. vol. i. p. 279) disappear, and the tissue consists entirely of large similarly-shaped cells which fit close to one another, leaving no spaces between. In the gall caused by Synchytrium pilijicum on Potentilla Tormentilla the much-enlarged cells in which the parasite settles are overgrown by the adjoining hypertrophied cells, some of which rise up in the form of hairs, and the whole new structure resembles a hairy wart. A curious gall is produced by Exohasidium Vaccinii on a sharply-defined portion of the fohage-leaves of the Alpine Rose (Rhododendron hirsutum and ferrugineum). A spherical spongy body rises from a restricted portion of the leaf, usually from the under side of the somewhat projecting midrib, sometimes only as large as a pea, sometimes as big as a cherry, and occasionally even attaining the dimensions of a small apple. It is yellow, but rosy-cheeked like an apple on the side turned to the sunlight, and it reminds one of this fruit by its succulent tissue and sweet taste. Indeed, these galls are sometimes called "Alpine Rose- apples ". Their surface is covered with a bloom which is caused by the numerous spores developed there and does not consist of wax like the bloom on an apple rind. The neck joining the gall to the leaf is not more than 1-2 mm. across, and, what is still more remarkable, except for this sharply-defined place of connection the infected leaf is unaltered. Galls produced by the Gymnosporangia on the leaves of the Mountain Ash, Pear-tree, Rock-medlar, and other Pomese exhibit strange forms. One of them, caused by Gymnosporangium conicum, on the foliage of the Rock-medlar (Aronia Totundifolia), is represented in fig. 357 \ It resembles a tubercle furnished with horns projecting from the lower surface of the leaf. Microscopic examination shows that the knob consists of the strangely metamorphosed spongy parenchyma of the leaf. The intercellular spaces which normally contain air are quite filled with the mycelial threads, and in the projecting portion of the tubercle, which is very hard and almost cartilaginous, tubes are inserted which terminate blindly below, where the spores of the parasite are developed, whilst above they are open and fringed, tlius allowing the spores to escape. These tubes look like horns to the naked eye. Usually several galls occur together on the same leaf. They are conspicuous at some distance on account of their colour. The chlorophyll is destroyed wherever the mycelium of the parasite extends and a reddish-yellow FUNGUS-GALLS ON STEMS. 521 colour takes its place, so that orange spots appear on the surface of the foliage, contrasting vividly with the green of the unaltered portions of the leaf. Galls rising from sharply defined parts of the stem are comparatively rare. One of the most remarkable is produced on the stems of a Laurel (Laurus Canariensis) by the parasitic Exobasidium Lauri. When it appears above the bark it looks like an aerial root, but rapidly grows into a branched spongy body 8-12 cm. long similar in appearance to one of the Fungi belonging to the family Clavariese (c/. fig. 195^, p. 21). The galls produced by Entyloma Aschersonii and Magnusii on the Composites Helichrysum arenarium, and Gnaphaliuvi luteo-albv/m Fig. 357.— rungus-galls. 1 Gall on the stem of the Juniper (Juniperus communis) produced by Gymnosporangium clavctriaiforme. 2 (jall on the leaves of Aronia rotundifolia produced by Gymnospora/ngium conicum. take the form of outgrowths, varying from the size of a pea to that of a walnut, developed from special spots on the root. Whether the spherical tubercles growing on the root-fibres of many Leguminosse, especially those of the Bird's-foot Trefoil (Lotus corniculatus), the Fenugreek (Trigonella fcemrni-grcecum), Lady's-Fingers (Anthyllis Vulneraria), Lupin (Lupinus variabilis), and the Liquorice (Glycyrrhiza glabra) are to be regarded as true galls caused by the Bacteria -like organisms invariably to be found in their interior is questionable. According to the most recent investigations they are the outward expression of a case of symbiosis and not of pure parasitism. Gall developments which involve whole roots or rootlets are found on the Alder (Alnus glutinosa), and on the Cabbage XBrasaica oleracea). The gall which is produced on Alder roots by Frankia Alni attains the size of a walnut and has a 522 ALTERATION OF FORM BY PARASITIC FUNGI. curious gnarled appearance; all the fibres of the root-branch thicken in a club-like or tuberous manner and become twisted and entangled with one another. The so- called " Fingers and Toes", caused by the Myxomycete (Plasmodiophora Brassicce), is a gall-like hypertrophy on the root of Brassica oleracea, which not uncommonly grows to the size of a man's head. Many woody plants have galls which alter the internal structure as well as the outward appearance of large tracts of the stem. The parasites settle in the corti- cal parenchyma, producing hypertrophy there, and afterwards the most varied distortions and alterations in the wood of that region of the stem. The trunk, branch, or twig becomes much swollen or knotted and the cortex rent and torn. Resin or a gummy mucilage sometimes runs out of the rifts in the galL As such a parasite exercises its metamorphosing faculty for several years, the canker (as it may be termed) increases in size continually. Sporangia of varied form and colour appear annually on the affected places, and again disappear when they have shed their spores. The part of the stem or branch above the cankerous cushion dwindles and dies off sooner or later. It rarely happens that the tree or shrub is able to rid itself of the parasite. Occasionally a growth of wood and cork from the adjoining healthy part walls in the cankerous spot so that the parasite is destroyed. The gall produced by Gymnosporangium clavariosforme on the trunks and branches of the common Juniper (Juniperus communis) is an example of this form (see fig. 357 1). From the hypertrophy there project in the early spring golden- yellow tongues (shown in the figure) consisting of masses of spores embedded in mucilage. Other similar growths are produced on species of Juniper by Oymrvo- sporangium conicum, G. Sabince, and G. tremelloides, but it would take too long to describe their differences in detail. It is important to mention, however, that each of these parasites has two stages of development, living on different hosts, the hypertrophies as well as the associated spore-producing organs of the parasite being different in the two cases. The "-fficidium stage" produces carti- laginous swellings (see p. 520) in definite spots on the foliage of various Pomese (Aronia, Cratcegus, Pyrus, Sorhus), the " Teleutospore stage" thickenings and tuberous outgrowths on the trunks of Junipers (Juniperus communis, excelsa, Sabina), and these parasites can travel from one host to the other in turn. (The two stages on different hosts are shown in fig. 357; these are not of the same fungus, but of nearly allied ones, and illustrate the point mentioned.) The parasite Peziza Willkommii attacks the trunks and branches of the Larch (Larix JSuropcea), and produces the well-known Larch-disease or " Larch-canker". The parasite having gained access at some point on the stem or branch first pene- trates the cortical parenchyijaa, and affects the cambium so as to prevent the further development of wood in that place. The development of the wood on the opposite side of the stem, i.e. the formation of annual rings, may proceed for several years, and in this way the attacked spot on the trunk takes the form of a depression, which is rendered the more conspicuous should the wood and cortex surrounding the parasite have undergone a greater thickening than usual. In ENTIRE LEAVES AFFECTED BY FUNGI. 523 time the patch becomes a sunken, blistered hole from which resin flows; and every year the fructifications appear above the cortex in the form of numerous little cup-like structures which are white outside and scarlet-red in the concavity. As the disease progresses the infected patch gradually spreads, and infected trunks and branches can be easily distinguished at a distance, Towards the end of summer the needles on the twigs above the canker turn yellow, while those on the healthy branches are still a beautiful green. This premature discoloration is a sure sign of the speedy death of the whole bough. A similar canker is produced on the Fig. 358.— Various Galls iGall on the bract-scales o£ the pistillate flowers of the Gray Alder (Alnus incana) produced by Exoaseus Alni-incance. 2 Inflorescence of Valerimiella carinata. s The same inflorescence with galls produced by a gall-mite. * leaf rosette of the House-leek (Sempervivwm MHum).' « Leaf rosette of the same plant which has been attacked by the fungus Endo- phyllum Sempervivi and has become hypertrophied. Silver Fir (Abies pectinata) by ^cidium elatinum, but instead of being only on one side of the branch, as in the Larch, it forms a uniform swelling all round it. Cankers of this kind are produced by a Bacterial organism {Bacillus amylovorus) on fruit-trees (Apple, Pear, &c.), and on various trees belonging to the Amentiferae (Beeches, Hornbeams, Oaks, &c.) by the Fungus Nectria ditissima. When whole leaves undergo hypertrophy of the kind we have particularly remarkable changes of form. For example, the normal leaves forming the rosettes of the House-leek (Sempervivum hirtum; see fig. 358 *) are broadly obovate in form, being little more than twice as long as they are broad. The leaves of the same plant after they have been attacked by the parasitic Endophyllum Semper- 524 ALTERATION OF FORM BY PARASITIC FUNGI. m,vi (see fig. 358 =) are seven times as long as broad and linear in shape. They stand erect, and are of a much paler colour than the healthy leaves. The Wood Anemone (Anemone nemorosa) affords another example (see fig. 259, p. 229). It spreads by creeping stems under the surface of the ground, and forms small colonies in light thickets and in meadows. The plants consist partly of flowering lateral shoots, and partly of foliage-leaves, which emerge above the ground from the creeping underground stem. In normal leaves the erect petioles are all the same length, and the leaflets are extended at about the same level. But when the ^oidium stage of Puccinia fusca has settled on them this becomes altered. The Hades of the infected leaves tower over their healthy neighbours in consequence of the elongation of their petioles, whilst their leaflets are smaller and less divided. The length of the petiole in normal leaves is some 12-13 cm., in hypertrophied leaves 15-18 cm.; but the size of the altered segments, compared with those of normal leaves, is as 5 : 7. Similar changes are observed in leaves of SoldaneUa alpi/na when attacked by Puccinia Soldanellce. The petioles of the infected leaves are 2-4 times as long as the normal ones, the blade is smaller and hollowed like a spoon instead of being flat, and the colour is an ochreous yellow instead of a dark green. The same alterations in the length of the petiole, and, in the size and colouring of the- leaf -lamina, are produced in the leaves of Alchemilla vulgaris by Uromyces Alchemilloe and in those of Pkytewma orbi- culare by Uromyces Phyteu/matum. To this class belongs also the so-called " curl " disease of Peach and Almond trees, produced by Exoascus deformans, and rendered conspicuous by the considerable enlargement, undulation, and bladder- •like expansion of the infected leaf-surface, which acquires generally a very brilliant coloration. Floral-leaves are comparatively seldom metamorphosed by Fungal parasites. In the Alder {Alnus glutinosa and incana) the bracts of the pistillate flowers are changed by Exoascus Alni-incance (=:E. amentorum) into elongated purple-red spatulate lobes much twisted and bent (see fig. 358 1); Peronospora violacea some- times causes the stamens to change into petal-like structures in the flowers of Knautia arvensis, so that they then seem to be " double "; Ustilago Maydis causes a growth of tissue in the pistillate flowers of the Maize, the result being that instead of grains irregular cushion-like structures 7 cm. in diameter are produced. Taphrina aurea, which settles on the pistillate flowers of Poplar (Populus alba and tremula) causes the ovaries to form golden-yellow capsules more than twice the usual size. The galls produced by Exoascus Pruni on the ovaries of wild Plum, Bullace, Sloe, and Bird Cherry (Prunus domestica, insititia, spinosa, Padus) belong also to this class. The tissue of the ovary increases in size, but not in the same way as in fruit forma- tion. The resulting body is flattened on two sides, brittle and yellow; the seed inside is abortive, and a hollow space is left in its stead. The gall produced from the ovary of Prunus domestica has the form of a rather curved pocket, which looks as if it had been powdered outside with flour at the time the spores ripen. These hypertrophies, which are popularly termed "pocket-plums", "bladder-plums", ENTIRE SHOOTS AFFECTED BY FUNGI. 52& &c., fall off the trees at the end of May. They are eaten in many districts, but have an insipid, sweetish taste. Galls consisting of whole shoots, both the stem and its leaves being altered by the parasite, are found principally on trees and shrubs, and only rarely on herbaceous plants. Examples of the latter, however, are furnished by the metamorphosed shoots of the Shepherd's Purse (Capsella Bursa -pastoris) produced by Gystopus candidus and Peronospora parasitica. Here the leaves, especially the floral-leaves, as well as the ground-tissue of the stem undergo pronounced hypertrophy. The petals, which measure only 2 mm. in length in a healthy plant, may become even 15 mm. long; the sepals also elongate, become fleshy and brittle, and are distorted and crumpled in all manner of ways. Only six stamens are developed in normal flowers, but in hypertrophied specimens there are often eight. The metamorphosis produced by TJromyces Pisi in one of the Spurges, Ewphorhia Cyparissias, is even more remarkable. The stem elongates far beyond its usual dimensions, and the leaves, which are crowded together on normal shoots, are thus separated by con- siderable intervals. The distance between two adjoining successive leaves in the healthy Euphorbia Cyparissias is only 0'5 mm., but in the hypertrophied specimens it becomes 2-3 mm. Infected shoots on an average are twice as high as healthy ones. The foliage-leaves, which are thin, flexible, linear, and twelve times as long as they are broad in the healthy plant, become, in the infected specimens, thick, brittle, elliptical, and only 2-3 times as long as they are broad. The bluish-green colour of the normal plant is changed into a yellow-ochre tint, and this contributes not a little to the odd appearance of the plant. Affected plants are not uncommon in Switzerland; a locality in which this disease has been very prevalent in recent years being Saas-F^e in the Saas-thal. The metamorphoses produced on the shoots of Periwinkles (Vinca herbacea, major, and minor) by the Uredospore-stage of Puccinium Vincce and on shoots of Cirsium arvense by the Teleutospore-stageof Puccinium suaveolens are very like those of the Euphorbia just mentioned, since the stem becomes much elongated and the leaves shorter, broader, yellow, and brittle. When flowers are developed on these aflfected shoots, they are more or less abortive and sickly, and no fruits or fertile seeds arise therefrom. Frequently the shoots • blossom prematurely. For example, we can at once detect by its elongated rosette- leaves when Privnula Clusiana and mimima are infected by TJromyces Primuloe i/ntegrifolicB, and it may be observed when this is the case that the shoots do not wait until the next spring to develop the flowers laid down in the summer, as usual, but open them in the autumn of the same year instead. The Cowberry (Vaccinium Vitis-Idoea) is especially worthy of notice among low woody plants, because two kinds of parasite attack its shoots. Melampsora Goeppertiana, in the Teleutospore-stage, causes a marked, gouty thickening in the cortical parenchyma, which is converted into a spongy tissue; at first it is flesh- coloured, but soon assumes a chestnut-brown tint. The stems elongate very much and grow vertically upwards; and when several of them close together are thus attacked they present a besom -like appearance. The foliage-leaves are much 526 ALTERATION OF FORM BY PARASITIC FUNGI. farther apart than in the healthy plant on account of this stretching of the stem. The lower leaves of the shoot are transformed into small fringed scales, and the upper ones are so much shortened that their outline becomes almost circular. The second parasite to which the Cowberry shoot is subject is Exobasidium Vaceinu (a near ally of the already mentioned Exobasidium Lauri, p. 521). The stem becomes pale rose-red colour, and rather thickened and spongy, but it does not elongate much more than usual; the leaves become blistered and curiously convex on the Pig. 359. — A "Witches Bxotm un the Siher Fir, produced by ^tidiiati daUnum under surface. The substance of the infected leaves becomes brittle and loses its chlorophyll. A red tint appears in place of the green, especially on the upper surface of the leaf, whilst the lower surface, on which the spores develop, looks as if it had been dusted over with flour. Usually the buds develop prematurely on these shoots, i.e. the buds which, under ordinary circumstances, would not develop until the next year push out and form new shoots shortly after they have been laid down. The axes of these shoots, however, remain short; their leaves are closely crowded, red in colour, and sessile. From a distance the premature shoots look like large double red flowers inserted in the dark green of the non-infected Cowberry bush. The shoots which develop prematurely on the shrubs of the Bog Whortleberry (Vaccinium uliginosum) by the action of Exobasidium Vaccinii are often met with in alpine regions, and are even more noticeable on account of their fiery-red ALTERATION OF FORM BY GALL-PRODUCING INSECTS. 527 colour. The Bearberry (Arctostaphylos Uva-ursi), Ledum palustre, and the Marsh Andromeda (Andromeda polifolia) are subject to similar metamorphoses at the hands of Exobasidium Vaccinii, so that Vacoinium Vitis-Idcea may be regarded as typical of them. When the shoots of the larger shrubs or trees are metamorphosed by parasitic •Fungi attacking their branches, we have the formation of the structures popularly termed Witches' brooms, which were mentioned at the beginning of this chapter. The stimulus necessary for their formation is afforded in different plants by different parasites; on Barberry bushes {Berber is vulgaris) by ^cidium Magel- hmnicum (to be distinguished from the common ^. berberidis), on the Gray Alder {Alnus incana) by Eocoascus epiphyllus, on the Hornbeam (Garpinus Betulus) by Exoascus Carpini, on the Bullaee (Prunus insititia) by Exoasous insititicB, on other species of the genus Prunus by Exoascus Gerasi, on the Birch (Betula verrucosa) by Exoasous turgidus, on the Weymouth Pine (Pinus Strobus) by Peridermium Strobi, and on the Silver Fir (Abies pectinata) by ^cidium elatinum.. Witches' brooms also occur on the Mastic tree (Pistacia LentisGus), and on Beeches, Pines, Larches, Spruce Firs, &e., although hitherto we ■ have not been able to ascertain definitely what parasitic Fungi are the cause in these cases. The Witches' broom of the Silver Fir has been selected and figured (see fig. 359) as a type of these peculiar structures. It always grows on one of the horizontally projecting lateral branches of the Fir, and raises its erect or curved twigs from the upper side, resembling, as it were, an epiphyte growing on the bark of the horizontal bough. The twigs are grouped in whorls and not in two rows, as usually happens in the lateral shoots of the Silver Fir. They are all shortened and thickened, and remarkably soft and pliable, because the cortical parenchyma has become spongy and the wood is only slightly developed. The buds, which in healthy tissue are egg-shaped, are almost spherical here. As in other instances of hypertrophied plant-members, we have a precocious development, a so-called "prolepsis", in these Witches' brooms. The buds swell earlier and unfold earlier than those of healthy twigs. The leaves remain short, yellow, somewhat crumpled, and fall off when a year old, while those of normal twigs are long, linear, straight, dark green on the upper side, and remain in position from 6-8 years. The growth of the twig is restricted; it dies off in a few years, and then, inserted on the dark green branches of the Silver Fir, remain the dry, bristling brooms, whose appearance has stimulated the imagination of the peasantry and given rise to the superstitions alluded to at the beginning of this chapter. ALTEEATION OF FORM BY GALL-PEODUOING INSECTS. Certain members of the Arachnoidea, Diptera, and Hymenoptera, which attack and penetrate the tissues of living plants and incite the formation of peculiar excrescences, are known as gall-mites, gall-gnats, and gall-wasps. The 528 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. growths, like small rosy-cheeked apples, which occur on the foliage of Oaks, popularly known as "oak-apples", are amongst the best known. The terms "gall" and "gall-apple" were used by writers in the sixteenth century, and (like the Old English word galle, the French galle, and the Italian galla) are derived from the Latin word galla, used for these' outgrowths by Pliny in his Natural History. The sixteenth-century writers distinguish between "gall-nuts" and "gall-apples", meaning by the former the small hard outgrowths on the leaves of Beech-trees. Afterwards the word gall was used for all the outgrowths produced by animals on green living plants. More than that — the hypertrophies described in the preceding chapter, produced in green host-plants by the various families of Fungi, are also included under the term. It has been proposed recently to substitute the word cecidium for gall, and to distinguish the excrescences as myco-cecidia, nemato-ceeidia, phyto-cecidia, diptero-cecidia, &c., according as they owe their origin to Fungi, Thread-worms (Nematodes), Gall-mites (Phytoptus), Gnats (Biptera), &c. A systematic classification of this sort, on the lines of the classification of animals, might be of use to Zoologists, but to the Botanist its value is only secondary. He must, as in other similar cases, keep to morphology as the primary ground of classification, aiid has to arrange the structures according to their agreement in development. Moreover, in a general review, it is necessary to consider whether a whole group of plant-organs or one alone undergoes metamor- phosis; and the starting-point of the outgrowth must also be ascertained; i.e. whether it is the foliage-leaves, floral-leaves, stems, or root -structures, &c., which are the head- quarters of the excrescence. When the gall originating as the nest or temporary habitation of a single animal or colony of animals is limited to a single plant organ it is said to be simple; if, on the other hand, several plant organs are concerned in its production it is said to be compound. Si/mple galls may, for convenience of description, be divided into (1) Felt- galls, (2) Mantle-galls, and (3) Solid galls. The Felt-galls are chiefly due to hypertrophied epidermal cells growing out into hairy coverings of various sorts and shapes; Mantle and Solid galls, however, are rather more complicated. In both cases insects are present in swellings of various descriptions, but there is this essential distinction: — The Mantle-gall is a hollow structure which, though it may arise in various ways and assume a multiplicity of forms, always has a portion of the surface of the affected organ for its lining — in other words, it is a chamber formed by hypertrophied growth around the place occupied by the insect. In the Solid gall, on the other hand, some spot is pierced by an insect and the eggs deposited vn the tissues (not on the surface), the punctured spot forms a swelling with the larva inside, but the lining of the chamber is in no sense a portion or development of the original surface of the organ affected. Again, whilst in most mantle-galls the cavity of the gall is in open communication with the outside, and the insect can escape by this aperture (though this is not invariably the case), in the solid gall there is not such opening, and the insect FELT-GALLS. 529 has to bore its way out. Needless to say, of both these types there are numerous modifications, but they fall into the two classes (of mantle and solid galls) according to their mode of development. The majority of felt-galls are produced by gall-mites. They form cottony or felted growths on limited and sharply defined areas of green leaves and stems, the surface of which is otherwise smooth, or possesses but few hairs. Some- times they have the form of small tufts, bands, or stripes, sometimes of large spots with irregular contour. In most instances the felt is situated on the under side of the foliage-leaf, and the gall-mite usually prefers the projecting veins to the green surface. In the Lime, Alder, Hornbeam, and Horse -Chestnut, the mites usually establish themselves in the angles formed by the lateral strands where they arise from the midrib, the projecting veins forming the framework for the felted hairs. In the Bramble (Rubus) and the Burnet (Poterium) it sometimes happens ithat the felt is continued down from the lamina to the leaf-stalk, and occasionally the green cortex of the succulent twig is covered with felted bands and spots. In some Brambles and Cinquefoils the sepals become furred by the action of gall-mites, the usual consequence being that the outline also becomes distorted. A swelling or slight hollowing of the green leaf-tissue very frequently accompanies the formation of felted galls, in which case the hairy covering is only visible on the concave side whilst the other remains smooth. This is most remarkable in the foliage of the Avens (Geum), Vine (Vitis), and Walnut-tree (Juglans), where a dozen white or brown-felted pit-like depressions are sometimes to be seen on the under side of a single leaf. The colour of the felted hairs is white in the leaves of Beeches, Limes, Bird Cherry, Brambles, Cinquefoils and Burnets, green in the common Maple, yellow in the Spindle-tree (Euonymus verrucosus), sulphur-yellow in Alnus orientalis and Black Poplar {Populus nigra), carmine red at first and then violet in Alnus viridis and in the Birches (Betula alba, carpatica, &c.), and brown in the Avens {Geum macrophyllum), Horse- Chestnut {^sculus Hippocastanum), and in the Aspen (Populus tremula). The felted galls which are light in their young stages usually take on a brown tint afterwards. Microscopic investigation has shown that in the formation of felted galls, the epidermal cells, originally tabular in shape and closely fitting, swell out and become transformed into bent and twisted tubes generally shaped like a club or retort, the stimulus being afforded by a minute gall-mite (Phytoptus). These cells look like short hairs to the naked eye, and as they stand side by side in large numbers the covering has a velvety or felted appearance. The mites which produce the felt, deposit their eggs in the juicy hair-shaped cells, and their young live on the materials contained in them. It should be mentioned that formerly these velvety and felted coverings were regarded as Fungi, and were described as distinct genera under the names Erineumn and Phylleriuvi (e.g. the gall known as Erineum quercinumi on the leaves of Quereus Gerris). To this group belongs also the gall occurring on the Wood Meadow Grass (Poa nemoralis) con- sisting of cells which resemble root-hairs, which is produced by the gnat Hormo- VOL. II. 84 530 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. myia Poce. The hair-shaped cells are epidermal, and spring from the stem above the nodes; they break through the leaf -sheath which proceeds from the adjacent node, and are arranged in two groups, which grow in opposite directions, so as to wrap round the stem from the two sides. The whole hairy mass looks as if it had been parted into two. At first the hairs are white; later they become light brown, and when the gall is fully developed they have the form of brown felted strands, wound round the stems and firmly inclosing the larva of the gnat in question. A large number of simple galls are grouped together under the name of Mantle- galls. The insects which give rise to them spend their lives on the surface of the leaves, where they multiply and attach their eggs to the epidermis. A growth is excited in certain layers of the cell-tissue by the stimulus which the animals exercise on their place of settlement. Cavities are thus formed which serve as dwellings for the animals and their brood, and which surround them like a pro- tecting mantle. Mantle-galls may be divided according to their structure into scroll-, pocket-, and covering -galls. Scroll-galls are caused by gall-mites, leaf -lice, tree-hoppers, and flies, and usually occur on the blades, rarely on the petioles of the leaves. The surface inhabited by these animals, which, in the ordinary course of things would have spread out flatly, grows more luxuriantly on one side than on the other, and the result is the formation of a scroll, i.e. of a chamber in which the animals are hidden. It is always the side on which the animals live which becomes concave, and the leaf is usually curled up lengthwise. In the Alpine Rose (Rhododendron), Crane's-bill {Geranium sanguineum), and Orache (Atriplex hastata, oblongifolia, &c.), it is the upper side of the leaf which is tenanted by the insects, and is therefore the one to roll up; it is the lower side, however, in the Buckthorn (Rhamnus caihartica) and the non-climbing species of Honeysuckle {Lonicera alpigena, &c.). In many instances the whole leaf-lamina is rolled up, but more frequently the alteration is restricted to the edge of the leaf when the margin appears to be bordered with a swollen hollow cushion often corrugated or undulating. In the Alpine Rose (Rhododendron ferrugineum and hirsutum) both halves of the leaf -blade are rolled round (see figs. 360 ^ and 360 ^), but usually the rolling is so slight that the gall has the form of a boat or hollow trough. Some- times an alteration in the shape of the leaf accompanies the rolling. For example, the foUage of the Abele (Populus alba) on which Pachypappa vesicalis establishes itself when the leaves are very young, exhibits in addition to the rolling a deep hollowing of the blade. Instead of the short blunt lobes, long pointed segments are formed, which stand side by side when they are rolled up, and cross over one another in many ways so that the mantle-gall on the hollow side is shut in by a veritable lattice-work. The parts of the tissue brought into contact by the rolling do not fuse together, and therefore the cavity in which the gall-producing insects live is always in open communication with the exterior. In most cases the tissues concerned are thickened, brittle, more or less devoid of chlorophyll, and yellow in colour. Not infrequently a red pigment is formed in them, so that the outside of the gall has a yellowish-red colour. The scroll-gall produced by the hemipterous FORMS OF MANTLE-GALLS. 531 Trioza Rhamni on the margin of Buckthorn (Rhamnus cathartica) leaves is very- hard and thickened like cartilage. In many plants the epidermal cells lining the gall elongate into hairy structures, as in the felt-galls previously described. Their juicy contents are used as food by the young gall-mites. This is the case, for , example, in the Alpine Rose (Rhododendron ferrugineum, cf. fig. 360 ^). Poclcet- are closely allied to the scroll-like forms. The tissue of the leaf -lamina or Fig. 360.— Galls. ^ Covering-galls on the petiole of the Black Poplar {Populvs nigra) produced by Pemphigus spirotheca. 2 Scroll-galls on the leaves of an Alpine Rose (Ehododendronjerrugineum) produced by gall-mites, s Transverse section of one of these galls. * and 5 Bud -galls on the branchlets of the Wild Thyme (Thymus Serpyllvm) produced by gall-mltes. « Blister-like galls on the leaf of the Red Currant (Ribes rubrum) produced by Myzus ribis. ' ' Part of the leaf seen from below. 8 Vertical section of a portion of this gall. 9 Solid gall on the leaf of the Gray Willow (Salix incana) produced by Nematus pedunculi. 10 The same gall cut open. " Part of the wall of this gall in vertical section. 1. 2, 4, 6, and 9 natural size ; » and 6 X 4; » and ' x 8; » and " x 60. petiole and sometimes that of the cortex in young twigs is subjected to a stimulus where the animals (gall-mites, leaf -lice, diptera) settle, with the result that a hollow protuberance arises whose excavated cavity serves as a temporary dwelling for the insects. The protuberances exhibit a great variety of form and shape, and they differ considerably in their internal structure. The following are the most notice- able forms. First, the plaited galls. They form deep, plaited, sometimes twisted channels in the leaf -tissue which open on the upper side by a narrow hole, and 532 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. project like weals on the lower surface of the leaf. The growing tissue which forms the floor of the channel is yellow and often lined with short hairs. The channel usually follows the course of the larger veins of the lamina, and some- times actually traverses one. Plaited galls are produced by gall-mites. The best known are those on the foliage of Garpinus Betulus, Clematis Flammula and C recta, and Ribes alpinum. Wrinkled galls come next to the plaited form. The protuberance is here limited to the green tissue shut in by strong rib-like projecting veins, and is only shallow; the upper side of the leaf has bulgings and protuberances and the lower pits and cavities. The protuberances are always developed in numbers close together, so that the leaf looks very much wrinkled in that region. Examples of this form are furnished by the wrinkled galls on the Elm (Ulmus campestris; see fig. 361 *) produced by the leaf- louse Schizoneura Ulmi, and on the Red Currant (Ribes rubrvmi; see figs. SQO^''''^) by another leaf-louse, Myzus ribis. In the latter several wrinkles are usually united into large blister-like protuberances, red on the outside, and covered with jointed cellular structures bearing glands which look to the naked eye like short hairs. This form, though resembling certain felted galls, is distinguished from them by the different form of the hairs arising in consequence of the stimulation. In the Mouse-ear Hawkweed (Hieracium Pilosella) leaf-fleas (Psyllodes) produce minute protuberances with narrow mouths, which stand out from the lower side turned towards the ground like small warts, and when they occur close together give a corrugated appearance to the leaf. Hollow protuberances of this sort arising upon restricted areas of the leaf-surface, and growing very actively, give rise to bag or sack-like structures attached by a very narrow neck. From their resemblance to a head such galls are sometimes termed capitate galls (Cephalonion). In others, where the outgrowth is fairly thick- walled and in form horn-like, the designation horn gall (Geratonion) has been given. Between these forms numerous intermediate forms exist which may be compared to pockets, bags, nails, &c. Many of these galls project from both the upper and lower side of the leaf, as though a nail had been driven through it — hence the last-mentioned name. The capitate-gall of the Sloe {Prunus spinosa), caused by a gall-mite, projects almost as much from the under as from the upper side, whilst the similar gall on the foliage of the Bird Cherry (Prunus Padus) rises on the upper side as a long pocket, but below only projects like a small wart. Many capitate and horn-like galls are developed only on one side, and here again there is a very great variety. When the protuberances are due to mites the cavity always opens on the lower side of the leaf. Both the inner wall and mouth of the cavity are covered with hairs, and sometimes the aperture seems to be actually plugged up by them. In the bag-shaped protuberances produced by the leaf-louse Tetraneura Ulmi on Elm leaves, a relatively large slit is formed just at the narrow part of the bag at the moment when the insects leave the cavity (see fig. 361 ^). The external surface of the protuberances caused by mites on the foliage of Alders (Alnus), Maples (Acer), and Limes (Tilia) is smooth, in those of the Bird Cherry (Prunus Padus), FORMS OF MANTLE-GALLS. 533 and Wayfaring Tree (Viburnum Lantana) ciliated, whilst in the inflated galls of the Elm caused by the white woolly leaf-louse (Schizoneura lanuginosa), it is covered with fine hairs like velvet. The capitate galls on the foliage of Maples, Alders, and Limes, of the Guelder-rose and Strawberry, are scattered abundantly over the whole lamina; in the Sloe they stand out chiefly from the margin of the leaf, and in Elms they occur singly or in groups on its central portions. The size of these galls depends upon their distribution. Those which rise in hundreds from the Fig. 361.— Galls. >-* Solid galls on a Rose-leaf ; i of Mhodites Boste, 2 of Ekodites jEglanterice, s of Rhodites sjnnosissimce. * Wrinkled galls on an Elm-leaf (JTlmus caynpestris) produced by Schizoneura Ulmi. s Purse galls on the same leaf, produced by Tetranevra Ulmi. « Covering gall on the same leaf, produced by Tetratieura alba, f Solid galls on the leaf of the Purple Willow {Salia purpurea), produced by Nematus gallarum. 8 Solid galls on the leaves of the same Willow, produced by Nematus vematffr. same lamina have a diameter of 1-3 mm., while those which occur singly or in small groups, often attain a diameter of 2-3 cm. Contrasting with these embossed or pocket-galls are the covering galls, forming a third type of mantle-gall. In these, as in the embossed forms, the insects pro- ducing the galls live in their cavities, but the course of development is quite different in the two cases. The tissue round the place where an animal has settled or where an egg has been fastened to the epidermis in this type begins to grow, rising up in the form of a fleshy mound or wall which continues to grow until the animal is wholly roofed in. The cavity in this case does not arise from an excava- tion (as in the embossed or "pocket" type), but from an overarching of the tissue. The 534 ALTERATION OF FORM BY GALL- PRODUCING INSECTS. external appearance of these galls is very varied. One of the simplest forms occurs on the leaves of the Ash (Fraxinus excelsior, see fig. 362^), where it is produced by the gall-gnat Biplosis botularia. The insect having laid its eggs in the chan- nelled depressions above the leaf -veins, fleshy cushions arise on either side of the groove which meet above and roof them over. The cushions of tissue forming the roof do not fuse; their succulent edges merely meet, and when the time comes for the gall-gnats to leave their temporary abode the tissue dries up and shrivels, Galls. .eoTion of one the.e galls. V c!ZrJn' onTellortZlt^eiZm n' '''°''^'" **™'^- ' ^""^''^Imal s one Of these galls out through with the optr ulut t^^^ C«<*""»^*« «-*• away ; x 3. The remaining figures natural size. attached, and the same with the operculum falling leaving a gaping slit as shown in fig. 362 \ The same thing happens on the leaves or rather leaf-vems of the Stinging-nettle {Urtica dioica) and of the Alder (Alnus gluUnosa), where the galls are produced by gall-gnats {Gecidomyi^ vHic<^, alni), and on the mxdnb of Elm leaves (Ulmus canupestris; see fig. 361«), where the galls are produced by a leaf-louse {Tetraneura alba). The so-called turpentine gall-apples (Carobe di Giude; see fig. 362*), which SOLID GALLS. 535 are caused by leaf-lice on various species of Pistacia, also belong to the covering class. The rudiment of a foliage-leaf, which in the normal course of events would have developed into a pinnate leaf with dark-green elliptical leaflets, grows out into a pod-like structure not unlike a locust-bean (fruit of Geratonia Siliqua). These galls are longitudinally grooved, and it can be seen more or less distinctly that the furrows correspond to the edges of the leaflets, only here the leaflets have become wrapped in, very much thickened and elongated, and fused with one another. In the cavity inclosed by the fused leaflets lives a colony of leaf-lice {Pemphigus comicularius) which have developed under the protection of the gall. When it is time for them to leave the cavity the top of the pod opens by the separation and bending back of the tips of the fused leaflets which form the wall of the cavity (see fig. 362*). The Chinese galls of commerce, produced also by Aphides (on Rhus semialata), develop much in the same way. They are hollow, irregularly pear-shaped structures with thin walls covered externally with a gray down. Two other covering galls which deserve special mention on account of their form arise on the petioles of the Poplar, particularly on the species Populus nigra, pyramidalis, and dilatata. The one, caused by a leaf -louse, Pemphigus hursarius (see fig. 362 ^), consists of a smooth expansion, red in colour externally, on the upper side of the grooved petiole. If the local swelling be cut through it is seen to be hollow, the cavity in which the leaf -lice live being shut in by thick fleshy walls. The fleshy tissue of the walls is formed by a growth of the cells round the place where the gall-producing insect has settled. A hole is formed at a point remote from the petiole (where the growing tissue met and formed a dome) as soon as the time comes for the inhabitants to make their exit. This is bordered by thick lips as shown in fig. 362 ^ The other gall which appears on Poplar petioles, pro- duced by Pemphigus spi/rotheca, is formed by the thickening of the edges of the grooved petiole, which rise up as fleshy cushions and meet above the depression. At the same time the petiole undergoes a spiral twisting, and a gall is thus pro- duced whose cavity is spirally twisted like the interior of a snail's shell. The thickened edges of the petiole do not fuse; at first they fit close to one another, but later on they separate, and a spiral hole out of which the white, downy leaf- louse can creep is the result (see fig. 360^, p. 531). We will now leave the mantle-galls and pass on to a consideration of the solid or tubercular galls. These are of the nature of swellings of limited size on single plant-organs, and are produced by insects which pierce the plant-tissue and lay their eggs in the wound. In this way either the epidermis of the chosen spot alone is injured, or the egg is inserted into the deeper-lying tissues. In both cases an active cell-division is incited in the neighbourhood of the injury. If, however, the egg has only been deposited in the epidermis, the larva which arises from it must penetrate into the interior of the now swollen tissue; when the egg is laid at once deep down this farther penetration on the part of the larva is of course unnecessary. The cavity in which the larvse dwell may be called the larval chamber, and this sort of gall can be classified according to the number of chambers which it contains. 536 ALTERATION OF FOEM BY GALL-PEODUCING INSECTS. whether only one or several (./. figs. 363 ^ and 3630- A great variety is met with in the structure of the wall of the larval chamber. It always has a layer of juicy, thin-walled cells immediately surrounding the egg, known as the medulla or pith of the gall, and an outer layer which surrounds the inner like a skin or bark (see fig. 360 1"). In most instances a third layer is inserted between them which consists of very hard cells forming a protective layer. It should also be noted that the layers of the wall of the gall separate in many instances, so that it is possible to distinguish an "inner" and an "outer gall". The gall-pith furnishes the larva with food when it emerges from the egg, and for this purpose the cells are stored with nourishing substances. The development of the pith takes place with great rapidity, and begins as soon as the egg has been laid in the tissue. The larva when hatched finds the inner wall of the chamber which has been fitted for its temporary abode always provided with the necessary food, and it immediately attacks and devours the juicy tissue with great avidity. The cells which are demolished, wonderful to relate, are replaced almost at once. The cells of the gaU-pith remain capable of division as long as the larva in the chamber requires food, and the surface cells which have been devoured in the gall-chamber are soon replaced by new cells from below, just as grass which has been mown down or cropped by cattle in a meadow sends up new stems and leaves. The spheroidal gall arising on the leaves of Salix incana (cf. fig. 360^) has only one chamber, and here the larva lives at the expense of the starch and other food-materials contained in the extremely thin-waUed cells which constitute the gall-pith (fig. 360^^). The larva traverses the chamber in a circle, beginning the destruction of the ceUs at a certain place and eating on as it continues its peregrination (fig. 360^°). New cells have already been formed for its nourishment by the time it again reaches the place from which it started. The hard and cortical layers are modified in very many ways as protective measures against the drying up of the gall in summer on the one hand, and against the attacks of birds and larger animals on the other. For the latter purpose the cortical layer is often fashioned like the pericarps of fruits which have to protect the seeds {of. p. 442). This also explains the bitter substances, hard skin, furry coat, bristling processes, and numerous other protective structures which are developed in and on galls just as on pericarps, and which contribute not a little to the remarkable similarity between galls and fruits. Many peculiar developments on the surface of these fruit-like galls cannot indeed be explained in this way, but, as in so many other cases, we conclude that they must afford some other advantage concerning which our understanding is still at fault. The external similarity between fruits and solid galls affords us useful points for classifying the latter into groups, which we may name berry-like, plum-like, apple-like, nut-like, capsule-like, &c. The currant gall produced by Spathegaster baccarum on the male catkins of the Oak has not only the form and size of a Eed Currant berry, but is also succulent and coloured red, and when several of these galls are formed on the same inflorescence it looks at first sight just as if racemes SOLID GALLS. 53T of red currants had been borne by some chance or other on Oak twigs. The galls produced by the Beech-gall gnat (Hormomyia fagi) on the foliage of the Beech resemble small plums, being surrounded by a hard layer which consists of a stone kernel and a layer of cells which might be compared to the fleshy part of a plum. The galls caused by gall -wasps of the genus Aulax on the nutlets of many Labiatse, especially on Nepeta Pannonica and Salvia officinalis also assume the form of stone-fruits. The insect lays its eggs in one of the four nutlets developed at the base of each flower; and within a week this grows into a smooth greenish- yellow ball which has the external appearance of an unripe cherry. A section through it shows that it possesses also the same structure as a cherry, plum, or stone-fruit in general. The succulent outer layer surrounds a hard stony kernel, but in the cavity of the kernel there lies the white larva of the gall-producer instead of the seed. These galls fall off just like fruits in July, and lie on the ground during the winter; and the mature insect does not bite an opening in the wall of the gall through which it can emerge until the following year. It has been already remarked at the beginning of this section how strong is the resemblance between apple-fruits and the spherical oak-galls, known as oak-apples, which are produced by various Cynipedes (see fig. 364^), together with the small red-cheeked galls produced by Rhodites Eglanterics and Nematus gallarum (see figs. 361^ and 361^) on Rose and Willow leaves respectively. Pith-galls which resemble certain dry fruits are very common. Those produced on the green cortex of young Oak twigs by Aphilothrix Sieboldi (see fig. 364 ^) remind one of the fruits of species of Metrosideros, those produced by Neuroterus lanuginosus and Spathegaster tricolor on the leaves of the Turkey Oak (Quercus Cerris; see figs. 364^^ and 364^*) have a decided similarity to the indehiscent fruits of the Wood- ruff and of the Goose-grass (Asperula odorata and Galium Aparine). The "spangle" galls produced on Oak-leaves by the gall -wasps Neuroterus fumi- pennis and numismatis resemble the fruits of Omphalodes (see figs. 364-'^ and 364^*), and the galls on the leaves of Duvaua longifolia produced by an insect Cecidoses Eremita have the form of a capsule which opens by an operculum (see figs. 362^ and 362^). Like fruits these galls may appear in all imaginable con- ditions with smooth, warted, or rugged surfaces, or covered with woolly or velvety hairs, with bristles or spines, fringes or claws, or even with moss-like outgrowths. The galls with moss-like covering occurring on the Wild Rose have been known from remote times as Bedeguars. They are caused by the Rose-gall wasp {Rhodites Rosce), which deposits its pointed, sometimes hooked eggs early in the spring in the substance of an undeveloped leaf while it is still folded up in the bud. The growth of the leaf becomes altered, the first sign being the development of numerous hairs. The larvae, when they creep out of the eggs, penetrate deeper into the tissue of the leaf, and it swells out into a solid gall containing as many chambers as there are larvse. Hairs and fringes continue to form on the exterior till those curious structures are formed which were said to have the power of inducing a peaceful sleep when laid under the pillow. Usually the stalks of the 538 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. young bud-leaves are pierced and then the upper portion of the leaf becomes atrophied. More rarely is the egg laid in the epidermis of one of the leaflets, in which case the leaves attain their normal size and only this particular leaflet is decorated with little bedeguars, as shown in fig. 361 1. When the petioles of three young leaf-rudiments are pierced simultaneously, as often happens, three single galls are produced close together on a shortened axis, and the whole structure may then attain the size of a pine-cone. The portion of meristematic tissue which is pierced? by the insect when it deposits its eggs sometimes remains an open passage; but more often a corky tissue is formed at the wounded spot which quite closes the chamber wherein the larva dwells. Under these circumstances the insect when it emerges must itself make an exit-passage from the gall, and this it does by biting a hole through it with its mandibles (see fig. 364 ^). The gall- wasps (Cynipedes) invariably leave the chamber which has hitherto served them both as a safe habitation and as an inexhaustible storehouse in this way. This does not occur, however, in some of those solid galls which owe their origin to gall-gnats of the genera Eormomyia, Diplosis, and Gecidomyia, for example, in those on the leaf-blade and petiole of the Aspen (Populus tremula) produced by Diplosis treTnuloB and on the leaves of Willows (Salix Gaprea, cinerea, grandifolia) by Hormomyia Caprem. Here the exit-passage is formed during the development of the pith. The gall consists, as in most other solid galls, of a pith, a hard layer, and an epidermis, but the enormously developed pith and the hard layer do not quite entirely surround the small larval chamber, they leave a small aperture on the part of the gall which is most arched. As long as the epidermis stretches over this place the mouth of the passage is of course not evident, but when the time comes for the insect to quit the chamber a gaping slit is spontaneously formed in the tense epidermis. In many instances the insect or the pupa as it pushes forward may break through the thin skin. A peculiar closure which might be compared to a lid is formed in the common solid galls which are produced so abundantly on Beech leaves by HorTnomyia fagi and which have been already alluded to. Just as the pupa of many Lepidoptera projects out of the hole in the cocoon which the caterpillar has spun for it far enough to allow the insect to fly away uninjured when it emerges, so that of Hormomyia fagi presses through the lid- like closure at the base of the gall, and the winged insect comes out leaving the chrysalis-case behind it. The opening of some solid galls, which resemble operculate capsules, and which may be termed capsule-galls, is especially remarkable and requires a more de- tailed description. As long as the larva or grub can remain and obtain food in the larval chamber the gall is completely closed, but when the time approaches for it to move its quarters and to enter the pupal stage in the ground a circular line of separation is formed in the tissue, and the part of the wall within the circle comes away as a lid. The process is seen very prettily in the gall produced on the leaves of the Turkey Oak {Quercus Gerris) by the gall-gnat Gecidomyia cerris (see fig. 362 '). In its closed condition the gall is a firm rounded chamber DEHISCENCE OF CERTAIN GALLS. 639 SO embedded in the leaf that it projects on the upper side as a small pointed cone, and on the lower side as a disc covered with a thick coating of hairs. In the autumn a circular piece like a lid becomes detached from the lower side of the chamber. It corresponds exactly with the extent of the hairy disc, and is so sharply defined that it looks as if it had been cut out with a knife (see figs. 362 8 and 362 ^). The operculum falls oflF, and the larva which had emerged from the egg and which has lived all the summer in the gall-chamber tumbles out and makes its way into the ground, where it begins to spin. By the next spring it has entered the pupal stage, and the gall-gnat creeps out of the chrysalis about May. Still more peculiar are the galls produced by Cecidoses Eremita on the green 7 Fig. 363.— Solid Galls. ' Capsule-like galls on a leaf of the Broad-leaved Lime (Tilia grandifolia) produced by Bormomyia Riaumuriana. 2 Longi- tudinal section through one of the galls, showing the maggot in the interior ; x 2. s Longitudinal section through a capsule gall from which the inner gall is just being extruded ; x 2. * Outer gall after the extrusion of the inner gall; x 2. 5 Inner gall at the moment when the operculum is thrown off ; x 2. * Capsule-galls on the leaf of a Brazilian species of Cdastrus. 1 Longitudinal section through one of these galls ; x 2. ' The same after the inneT gall has fallen out ; x 2. 1 and B natural size. cortical tissue of young twigs of Duvaua longifolia, a South American represen- tative of the Anacardiacese (see figs. 362^ and 362^). The gall is quite spherical and very hard, and its large cavity conceals the caterpillar which has been hatched from the egg. When the time draws near for the formation of the pupa, a plug with a projecting rim is developed on the side of the gall furthest from its point of attachment. When the plug is pushed out a circular hole is left which leads into the gall-chamber through which the caterpillar escapes from its dwelling. People who have not seen these galls with their own eyes might almost think this description was the work of imagination. And yet there are still more wonderful forms in this class of gall-structures. On the foliage of the Lime {Tilia gra/ndifolia) a growth arises round the eggs of the gall-gnat Horinomyia Reau- muriana which at first has the form of a flat lens inserted in the green tissue of the blade, but which gradually enlarges until it projects from the upper side like a 540 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. blunt cone and from the lower as a hemispherical wart. The gall-chamber is inhabited by the maggot of the gall-gnat. The top of the conical part loses its colour in July and becomes yellow and brown, and a rim is formed around its summit. On cutting a vertical section through the gall at this stage it is seen that the tissue forming the wall of the chamber consists of two parts (see figs. 363 ^). The inner layer, which contains the maggot, is surrounded by an outer one which gradually passes into the green substance of the leaf and extends up to the rim just mentioned. The whole structure has separated into an "outer" and an "inner" gall, the inner gall resembling an egg lying in an egg-cup (c/. fig. 363 ^). During the summer the inner gall separates completely from the outer and is actually thrown off by it. For the accomplishment of this the tissue of the outer gall swells up very much, so that it exercises a pressure on the inner gall which is shaped not unlike a cone, somewhat narrower below than at the top (see fig. 363^). The extruded inner gall falls on the ground below the Lime-tree and assumes a dark- brown colour; the outer gall remains as a little crater embedded in the leaf -blade and ultimately shrivels up (c/. figs. 363^ and 363*). The detached inner gall is smooth at the blunt and previously upper extremity, and striated at the other; it is not unlike a detached composite-fruit. The gall-gnat within feeds for a little time longer on the succulent lining, and then rests through the winter; in the spring it makes its escape. To do this it bites a ring-shaped groove below the conical top of the gall and presses against the roof, which, owing to the breaking of the tissues around the ring, comes away like a lid (see fig. 363°). A similar state of affairs prevails in a gall formed on the foliage of a Brazilian species of Celastrus (see figs. 363 *■ '• ^), but here the inner gall (which comes away) has several chambers, and the outer gall has the form of a cup set in the green blade. The place of origin of all these solid galls depends of course upon the insects producing them. These are usually very fastidious about the place where they will lay their eggs, and it is truly astonishing with what care they search out spots difficult of access, and at once favourably situated as regards food supply and likely to afford a safe habitation for their offspring during the larval stages. The small gall- wasp Blastophaga grossorum lays its eggs in the ovaries of the " gall- flowers " in the interior of the figs of Ficus Carica (see p. 160 and figs. 240 " and 240 1^ p. 157). The gall-wasps Andricus amenti and Neuroterus Schlechtendali deposit them in the stamens of the Turkey Oak; the gall-wasp Gynips caput- medusce lays hers in the side of the bract-scales which surround the pistillate flowers of the Oak (Quercus sessiliflora and pubescens), and so produces a gall with innumerable stiff-pointed . fringes entangled with one another which ward off the attacks of other animals (see fig. 364 1"). Countless gall-producing insects deposit their eggs on the lower side of foliage leaves, some preferring the lamina, others the veins. Andricus curvator prefers the margin of Oak leaves, Diphsis tremulce the petiole of the Aspen at its junction with the blade. Several gaU-wasps, as, for example, Andricus cestivalis and Andricus grossularice, seek out the floral recep- tacle in the male catkins of the Turkey Oak for the deposition of their eggs, whilst ALTERATION OF FOEM BY GALL-PRODUCING INSECTS. 541 several Cynipedes, e.g. Aphilotrix Sieboldi (see fig. 364 1) lay their eggs in the green cortex of the young twigs. Solid galls are very rare on roots, but they do occur (~:rj-n Fig. 364.— "Various Oak-galls. 1 Solid galls on the cortex produced by Aphilothrix Sieboldi. 2 Bud-gall from, a foliage-bud produced by Cynips Hartigii. s Solid galls on an Oak twig produced by Cynips Eollari. * One of these galls cut in half, s Bud-galls from foliage-buds produced by Cynips lucida. « One of these galls cut in half. 7 Leafy bud-galls produced by AphUotkrix gemmce. 8 Bud-galls from foliage-buds produced by Cynips polycera. s Longitudinal section through one of these galls. 10 Gall on the pericarp of Quereus pubescens produced by Cynips caput-medusoB. n-i* Spangle galls on a leaf of the Turkey Oak {Quercus Cerris) ; 11 produced by Neurotents lanuginosus ; 12 by Neuroterus nwmismatis ; is by Neuroterus fumipennis ; 14 by Spathegaster tricolor. in this situation in the oak, being produced by the gall-wasps Aphilothrix radicis and Biorhiza aptera. 542 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. When several organs of a plant imtaediately adjacent to one another are con- cerned in the production of a gall it is said to be compound. Compound galls are for the most part produced from buds, and they are all comprehended under the general name of Bud-galls. They are extraordinarily varied in their characters, some being merely abbreviated axes clothed with scale-like leaves, in others only the base of the shoot is involved and above the gall it continues its growth quite normally, whilst in others again the axial portion of the structure is much swoUen, and the leaves hardly represented at all. It is difficult to give any satisfactory classification of these bud-galls; still, for the sake of arranging our facts, we may distinguish these types, viz.: — the ordinary bud-gall, the cucJcoo-gall, and cluster- gall. Ordinary bud-galls involve several, often all, the members of a shoot. The axis of the shoot is always deformed and abnormally thickened. The swollen portion contains in its interior one or several larval chambers surrounded by a pith-Uke layer. Two varieties of ordinary bud-gall may be distinguished. The first is leafless; no leaves are present, or, more correctly, they are transformed into tubercles, pegs, and knobs which merge insensibly into the swollen axis which contains the larval chamber. The second possesses leaves, the gall being covered with scale-like bracts or more or less fully developed green foliage-leaves. Amongst the leafless bud-galls the most interesting are those which are armed with special means of protection agains.t the attacks of animals on the watch for the larvse of the gall- wasps. The gall shown in figs. 364 ^ and 364 *, produced by Gynips polycera on the leaf -buds of Quercus pubescens and sessilifiora, which to a certain extent afiects a whole lateral shoot, has the form of a young Medlar fruit, and on it may be seen 3-5 metamor- phosed leaf-structures projecting as stiff'-pointed pegs which gradually pass into the tissue of the shoot axis. This gall is one-chambered, and the tissue of the wall has separated into an outer layer and an inner spherical pithy gall. The gall shown in fig. 364 ^ is produced by the gall-wasp Cynips Hartigii which lays an egg in the middle of the leaf -bud of the Oak {Quercus sessilifiora). The bud does not develop into a leafy shoot, but into a small one-chambered gall with large tooth-like or club-Hke processes which represent metamorphosed leaves. The thickened angular ends of these projections fit closely to one another so as to form a sort of second outer coat to the gall-chamber through which hostile ichneumon-flies cannot penetrate. ' The gall much resembles the cone-fruit of a Cypress in the arrangement and form of its superficial processes. The galls produced from the buds of various Oaks {Quercus pendulina, sessilifiora, pubescens) by the gall- wasp Cynips lucida are still more peculiar (see figs. 364^ and 364^). They contain several larval chambers with abundant pithy tissue, whilst innumerable slender processes resembling limed twigs in being very sticky on the capitate thickened end project from their exterior. Ichneumon-flies and other animals hostile to the gall-producers take good care not to come into contact with these spikes which are to be regarded probably as trans- formed leaves springing from the swollen axis. Among the galls produced, from leaf -buds belonging to this group there are some in which the leaves are merely indicated as tubercles. This is the case, for example, in the many-chambered, BUD-GALLS. 543 spongy gall, red-cheeked on the sunny side but pale elsewhere, which is produced on the tips of the branches of the Oak by the gall-wasp Dryoterus terminalis, and looks very like a potato in shape. The leaves are only represented by small ill- defined knobs and ridges, just as in the potato. To this class of galls belongs also that to which the term " nut " is popularly applied, and even in commerce, the name has been transferred from this to the whole of the first group of compound galls (bud-galls). The "nut" is produced on the Oak by Gynips calicis as an angular and irregularly-grooved gall which originates at the end of a flower axis, and the cupule formed of several bract-scales as well as the ovaries are concerned in the growth. This class of galls also includes the irregular blunt swellings on Aspen twigs (Populus tremula), which are caused by the larva of a beetle (Saperda populnea), and in addition the many-chambered woody "canker cushions" as large as a nut which are produced on the branches of Willows by Nematus medullaris. The gall shown in fig. 364 ^, which arises on various Oaks {Quercus pedunculata, sessiliflora, pubescens) by the action of the gall- wasp, Aphilothrix geTumce, may be selected as a type of leafy bud-galls. It resembles the cone of a Hop or Larch, and is developed from a foliage-bud. It has a much-abbreviated swollen axis, whose tissue separates into an inner and outer gall, beset with numerous dry, brown lan- ceolate hairy scales having the form of bract-scales. Bud-galls which are covered with green foliage-leaves are produced by the gall-wasp Andricus inflator on the Oak, but they are more commonly met with on herbaceous plants, e.g. by Urophora cardui on Girsium arvense, by Diastrophus Scabiosoe on several Knapweeds {Gen- taurea alpestris, G. Badensis, G. Scabiosa), by Aulax Hieracii on various Hawkweeds (Hieraoiwm murorum, sylvaticv/m, tridentatum, &c.). Usually the foliage-leaves are stunted, and not infrequently the blades of some of them are quite obliterated, so that the gall in that region is only furnished with scaly leaf-sheaths. A Sage growing in the Isle of Crete so often bears leafy bud-galls resembling a small Quince-apple, produced by a species of Aulax, that Linnaeus called it Salvia pomi- fera. The stem of this Sage is swollen out like a ball, and the spherical mass, covered with a gray felt of hairs on the exterior, is surmounted at the top with a group of small wrinkled leaves, which look like the persistent calyx of a Quince- apple. The best known and most widely distributed of these forms, found on the Hawkweeds named above, consist of knob-like swellings of the stem. The larval chamber is situated inside the enlarged pith, the ring of vascular bundles, which has undergone much shifting, forms the protective layer, and the cortex of the afiected region of the stem forms the cortical layer of the gall. The epidermis is densely covered with hairs. Leaving the galls which consist of modified foliage-buds, we pass on to such as consist of metamorphosed flower-buds. They arise from flower-buds in which small gall-gnats have laid their eggs. The larva hatched from the egg lives in the cavity of the ovary, or. in one of its loculi when there are several, and this space, therefore, becomes the larval chamber. The corolla, which envelops the ovary in the flower- bud, remains closed, like a cap on the top of the larval chamber. The calyx becomes 544 ALTERATION OF FOEM BY GALL-PRODUCING INSECTS. inflated, enlarged, and sometimes fleshy. The whole gall resembles a bud or small bulb; it is not unlike one of those bulbils which so often arise instead of flowers on the flowering axis of certain species of Allium. They occur especially on the Bird's-foot Trefoil (Lotus corniculatus), where they are produced by the gall-gnat Gecidomyia Loti, on the various species of Mullein {Verhascum Austriacum, nigrum, Lychnitis, &c.) by Gecidomyia Verbasci, on several species of Germander (Teucrium Tnontanum, Scordium, &c.), caused by Lactomelopus Teucrii, and on the Kampion (Phyteuma orbiculare), where they are produced by Gecidomyia phyteumatis. Closely allied to these bud-galls are those remarkable gall-structures which are commonly known in Austria as "cuckoo-buds". The cuckoo is supposed to be concerned in their formation, just as it is in that of the frothy saliva-like masses deposited by the Gicada on the Cuckoo-flower (GardaTnine pratensis). The name " cuckoo-galls " may be employed for the whole of this sub-group. They are char- acterized by their pale whitish colour, soft spongy tissue, and especially by the fact that they only involve the base of the shoot, while the upper end can continue its growth unaltered. In this respect they may be compared to a Pine-apple fruit, where the axis rises above the fleshy collective fruit (cf. p. 436) as a green leafy tuft, which does not lose its growing power even with the ripening of the fruit. The history of the development of cuckoo-galls is probably like that of covering galls; and the main distinction lies in the fact that in the former the gall is pro- duced not merely from a single organ or some part of it, but from a whole group of adjoining plant-members. The best known and most widely distributed gall of this group is produced by the pine-apple aphis Ghermes abiefis on the twigs of the Spruce Fir (Abies eaxelsa, see flg. 362 1, p. 534). Early in the spring, before the foliage-leaves have begun to unfold, the parthenogenetie females, the foundresses of the colony, attach themselves each to the base of a young leaf and lay a mass of eggs at the spot to which they have adhered. The larvse, hatching, penetrate the surrounding parts of the shoot with their beaks; the shoot swells, as do the bases of the needles, and a growth, the Spruce gall or Pine-apple gall results. The gall somewhat resembles a small Fir-cone about an inch long, with the surface divided into small convex areas, each bearing a short needle-like projection in the middle; these are the deformed needles, which, becoming swollen, touch each other on the outside of the gall. They are separate inside, so that the gall contains a series of cavities or chambers. In these chambers the larvae live in numbers, either entering the chambers during the growth of the gall or being inclosed by the swelling of the surrounding needles— this point is not certainly determined. They remain in the small cavities so formed and feed, cast their skins, and multiply there. In August the gall begins to dry up, each of the small cavities opens by a slit in front of the green needle-point surmounting the cushion (see fig. 362 ^ p. 534), and the winged insects now leave the place in which they have passed the spring and summer. Cuckoo-galls are met with almost as frequently on Stellatse, viz. on various species of Bedstraw (Galium Austriacum, boreale, uliginosum, &c.) and Woodruff BUD-GALLS. 545 {Asperula galioides, tinctoria, &c.) as on Fir-trees. The infected parts of the shoot remain stunted, and white spongy cushion-shaped growths, which are somewhat grooved, arise at the bases of the leaves. Since the growing tissues of neighbouring leaves touch one another the grooves or channels form small cavities in which live the larvae of the gall-producing gnats (Gecidomyia Oalii and Asperulce). In the common Bedstraw {Galium, MoUugo) these spongy growths arise, not from the bases of the leaves, but from the green cortex of the stem round the insertion of the leaves and lateral branches. They rise up as cushions and lobes, and several join together to form a sort of dome, under which the larvae of the gall-gnat dwell. The foliage-leaves are scarcely altered in form, and when lateral twigs arise from the place they also are unchanged. It not infrequently happens that short lateral axes terminated by flowers spring up quite unmodified above the spongy white cuckoo-gall. Cuckoo-galls also occur on Cruciferse, viz. on Barbarcea vulgaris. Nasturtium palustre, sylvestre, and Sisymbrium Sophia. They are produced by Gecidomyia Sisymbrii, and originate principally at the bases of the flower-stalks half-way up the inflorescences. They look like spongy white bodies which surround the pedicels like the brim of a hat. As the growths from neighbouring pedicels meet together they roof over chambers which serve as habitations for the larvae of the gaU-gnats. Viewed from outside the galls appear like irregular white bodies inserted in the inflorescence, which remind one of the fruit of the white Mulberry- tree. The term cluster-gall is reserved for that type of bud-gall in which the axis is much restricted or stunted and covered with densely crowded leaf -structures; it is in the chinks and recesses between the crowded leaves of these galls that the insects concerned pass the whole or a portion of their lives. The animals which cause the galls belong to very dififerent classes. Gnats, leaf -fleas, leaf -lice, and mites are the commonest varieties. The gnats only live in the galls during the egg and larval stages, but the others pass their whole life there. They invariably settle on the end of a shoot while it is still undeveloped in the bud. The axis of the shoot remains more or less stunted in consequence of the influence the animals exercise on it and its leaves undergo fundamental alterations. The blade or sheath of the leaf is deepened and hollowed to afford sufficient space to the animals which have established themselves between them, and as these parts of the leaves touch one another recesses are formed not unlike those which are developed in fir-cones for the growing seeds. The sheathing part of the leaf is often rather thickened, and its succulent cells serve as food for the animals living in the gall; in other instances the hollowed leaf -blades are thickly covered with hairs, and this coat then has the same significance with regard to the insects as the felt of hairs on isolated leaves already described. Very different forms of galls are produced according as to whether the free ends of the leaves turn back or remain in contact, and whether the axis from which the leaves spring is more or less contracted. Sometimes they remind one of open rosettes, sometimes of closed balls, bunches and tufts, sometimes of pig-tails and witches' brooms. vol.. II. 85 546 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. Clustered galls may be divided into two classes, those which develop in the region of the flowers and those in the foliage region respectively. The most noticeable and best known forms of the galls occurring in the foliage region on rudimentary leafy shoots are the following: — First, those peculiar structures on the tops of WiUow twigs {Salix aurita, Gaprea, grandifolia, &c.) which are popu- larly termed "Rose Willows". They are caused by the gall-gnat Ceeidomyia rosaria. The leaf-bud from which they arise keeps its axis quite short and develops on it numerous green leaves arranged like the petals of a double rose. The lowest leaves of the "rose" differ but slightly from the normal foliage of the particular species of Willow. Usually there seems to be only a shortening and broadening of the petiole and leaf -sheath, the green blade being almost un- altered. In the upper inner leaves, however, the sheath-like part of the leaf is much increased in size, and nearer the centre of the "rose" the leaves become scale-like. The leaf-blade has entirely disappeared, and the end of the contracted axis possesses only the remains of leaf-sheaths. It is worth noting that the number of leaves in a Rose Willow is always greater than would be found on an unaltered shoot of the same species. For example, if the number of leaves on the one-year-old shoot of the Sallow (Salix Gaprea) is 25, the number in a "rose" on the same species would be at least twice as large. This can only be explained by supposing that a "prolepsis" has occurred, i.e. that not only the shoot laid down for the current year has developed, but also one originating from a bud of this shoot, which, under normal conditions, would not have developed until the following year. When autumn comes the rosette-shaped gaUs on the Willow bushes show up conspicuously at a distance because the leaves forming them do not fall off like the rest, but remain behind as brown dried structures at the ends of the branches. They are also found associated with the catkins. The rosette-shaped galls produced by the gall-gnat Gecidomyia cratcegi at the tips of Hawthorn twigs (GratcBgus Oxyacantha and monogyna) also claim atten- tion. They are full of bristles and resemble tiny birds' nests. The stimulus of the gnat larvse excites a deeper and more frequent segmentation in the leaves and stipules. Narrower points and fringes which are much bent and which resemble the antlers of reindeers replace the broad lobes. Also soft spines with capitate ends rise up from the green cortex of the twigs and from the tissue of the leaf-blade, especially above the vascular bundles, and 3-5 of them often fuse together into cock's-comb-like structures. These bristling rosettes on Hawthorn branches also remain long after the time the ordinary foliage falls off. In marked contrast to the rosette-like cluster-galls are others whose leaves .all fold together in a ball something like the leaves of a cabbage, the whole gall having a button-like appearance. The outer leaves are round and hoUowed on their upper side, and they usually fold together like mussel-shells. The inner leaves have a similar form, but they are much smaller and more concave, and they have become succulent and paler in colour. The galls produced by Geci- domyia genisticola on Genista tinctoria and those which Gecidomyia Veronicoe BTJD-GALLS. 547 gives rise to on Veronica chamcedrys, and which gall-mites produce on the Wild Thyme (Thymus Serpyllum; see figs. 360* and 360 ^ p. 531), form white buttons on the ends of the shoots which show up conspicuously from the dark green of the surrounding foliage. The white colour is due to the fact that the outer leaves, which fold together like mussel-shells, are thickly covered on the outside with white hairs. Gecidomyia Artemisice produces on the branches of Artemisia campestns a closed cluster-gall which is cased in white wool like a shroud. On the other hand, the large, button-shaped, closed cluster-galls which are produced by Gecidomyia rosaria on Willows (Salix purpurea, &c.) and by a gall-mite on the spikes of the Brome-grass (Bromus) are green and smooth, or at least they have not more than the usual number of hairs. On the shoots of the Yew (Taxus haccata), the Flax (Linum usitatissimum), JEvphorbia Gyparissias, the Moss Campion (Silene acaulis), and several Ericas (Erica arborea, carnea, &c.) the influence of various gnats (Gecidomyia Taxi, EuphorbicB, JSricce, scoparice, &c.) produces galls with linear erect leaves crowded together into tufts. The base of the crowded leaves and the axis of the gall are usually rather thickened, so that it looks as if the linear leaves were set on a rounded button, and this is particularly marked in Euphorbia Gyparissias. This division includes the gall formations occurring on Juniper twigs (Juniperus communis), which are caused by the gall -gnat Lasioptera junipervna. The acieular leaves of the Juniper are arranged in whorls of three on normal shoots. By reason of the influence of the gall-gnat Hormomyia juniperina the whorls at the top of the twig become so changed that the last but one represents a cup bordered with three teeth in consequence of the broadening of the needles, while the terminal whorl is metamorphosed into a dwelling surrounded by three short leaflets. This gall closely resembles the cone of the Arbor Vitse (Thuja occidentalis, orientalis, and plicata) in form. An insect, Livia Juncorum, produces galls on various Rushes (Ju/ncus), espe- cially Juncus alpinus and lamprocarpus, which look like knots or tassels. The axis of the shoot is contracted, the sheathing portions of the leaves which cover one another are much widened, and the colour is pale except where it is reddened by exposure to the sun; their appearance is like the outer covering or top of a tassel. The stunted green blades which spring from the sheathing portions are thread-like and arranged as the loose strands of the tassel. Not infrequently short lateral shoots arise in the axils of some of the leaves, and then the whole structure looks like a bunch of tassels. Closely allied to these cluster-galls on the stems of Eushes are such as re- semble tufts and witches' brooms, produced by mites on the branches of hairy Willows, especially on the white Willow (Salix alba). Instead of the long leafy Willow rod which would have emerged under ordinary circumstances from a foliage-bud, a confused mass of twigs with short leaf-scales is developed which at first seems a perfect mystery. By careful examination it is seen that the axis of the shoot laid down in. the bud has remained stunted, and that lateral 548 ALTERATION OF FORM BY GALL-PEODUCING INSECTS. shoots have developed from the axils of its leaves. These lateral shoots again develop lateral axes in the axils of their leaves, and so on to the third, fourth, and fifth degree. Thus, in the course of a month, shoots have unfolded, which, except for the influence of the gall-mites, would not have followed one another for three, fotir, five, or even six years, and therefore these galls afford us another instance of what has been termed "prolepsis" or precocious development of structures which would not yet arise. Of course all the axes of these shoots are dwarfed and the leaves which clothe them are diminished in size. The shortening and diminishing increase gradually, so that the axes and leaves of the fourth and fifth degree are much smaller than those of the second and third. The last lateral shoots remain bud-like, and their small scaly leaves fold over one another like the bracts in the involucre of a Composite. The "witches' brooms" which are caused by gall-mites on Lilac (Syringa vulgaris) and Privet (Ligustrum vul- gare) bushes are similar in nature to these closed galls on the Willows. Frequently the metamorphosis of the leaves on the axes of the third, fourth, and fifth degree includes those' of the floral region, and such cases form to some extent a bond of union between cluster-galls on foliage and on floral regions, respectively. One of the most remarkable changes exhibited by the gall-structures just men- tioned, viz. the abbreviation of the axis, is of course not to be noted in cluster- gaUs in the floral region. The part of the axis which forms the floral receptacle does not grow into an elongated shoot, but always remains short, and the floral- leaves it bears stand close to one another, forming whorls in whose niches and recesses numerous small animals can hide. But these animals effect other very marked alterations by their stimulus. In some flowers, instead of the norma,' red, blue, white, or yellow petals, green leaflets appear which resemble foliage- leaves in character, and then we say that the flowers have become "green" or " leafy ". In other plants the stamens are transformed into petals, and the flowers are said to be " double ". Finally, it may happen that the carpels which are usually united together to form a syncarpous ovary stand on the receptacle as distinct structures, and that to a certain extent their union has been dissolved. In these cases we speak of " antholysis '' (c/. p. 80). The influence of gall-mites also produces metamorphosed flowers which may be both green and double, and in which the pistil may have separated into its individual carpels. The best flowers for observing these metamorphoses in all imaginable degrees are the small-flowered species of the Ghickweed genus {Cerastium macrocarpum, triviale, «fec.), several Caryophyllacese {Lychnis Viscaria, Saponaria officinalis, &c.), Cruciferi» (Gardamine uliginosa, Oamelina sativa, Lepidium Draha), Gentians (Gentiana acaulis, rhcetica), Speedwells (Veronica officinalis, saxatUis} and Milfoils (Achillea Millefolium, nana). In Speedwells the petals come to resemble leaves. The bunches, rosettes, and balls of small green leaves replacing the flowers are set close together on the rachis of the inflorescence and form green racemes and tufts, sometimes even small witches' brooms. In Veronica saxatilis the rachis of the raceme, the pedicels, and the bracts are covered with hairs, which COMPOUND GALLS IN FLOWERS. 549 is not the case when the plants are free from the mites; the foliage-leaves in the neighbourhood of the raceme are also lobed and deeply indented, which again is not the case in uninfected plants of this species. In the capitula of the above- named Milfoils the peripheral ray-florets as well as the central tubular ones become leaf -like, and this gives rise to the most peculiar forms. A capitulum is often sepa- rated into several stalked sub-capitula, the flowers being metamorphosed into green funnels with jagged mouths, and into small flat-lobed and toothed foliage-leaves, whilst short, green, scale-like leaflets rise from the midribs of these leaves repre- senting the metamorphosed stamens. A very remarkable "doubling" produced by gall-gnats is to be observed in flowers of the Alpine Rose {Rhododendron ferru- gvneum). The stamens and carpels are changed into red petals by their influence. Since Rhododendron flowers have ten stamens and five carpels, there should be only fifteen red leaflets in the centre of each, but as a matter of fact there are double and treble as many, and there has been not merely a metamorphosis but also a multiplication of the leaves. The flowers of some plants which belong to the Valerians, especially of the Corn-salad {Valerianella carinata), of which a small umbellate cyme is shown in fig. 358 ^ p. 523, are doubled by the influence of a gall-mite, but without any multipUcation of the petals. The doubling is restricted to the transformation of the stamens into a whorl of petals. But we also have another peculiar alteration. The petals become enlarged to more than fifty times their normal size, and change into fleshy lobes which are fused with one another into a disc. As all these lobes bend back, and become concave below, cavities are formed under the flowers in which the gall -mites can dwell (see fig. 358 ^ p. 523). The axis of the inflorescence and the stalks of single flowers are often thickened and fleshy in these cluster-galls, and they are also stunted and bent in the most varied manner. If several neighbouring pedicels fuse together, structures like cocks'-combs, or like a compressed and flattened branch, arise; to these the term fasciation is applied. Sometimes when numerous pedicels arranged in the form of umbels fuse together we have structures like coral-colonies, or irregular clumps which are beset with green flowers usually much reduced in size. This is the case in the fasciations of the Ash (Fraxinus excelsior and Ornus), which are caused by a gall-mite (Phytoptus), and which occur so abundantly that it looks as if the tops of the trees had been sown with them. The enumeration, here, of various forms of galls commenced with the incon- spicuous felt-galls on the under side of isolated foliage-leaves, and it ends with the complex cluster-gall, in which hundreds of flower-stalks and leaves are frequently concerned. Of course, only types of the individual groups which have been mentioned in this long series could be instanced, and we cannot make any attempt to describe all the gall-structures at present known, about 1600 in number. Whether the extension of gall-researches in tropical regions will yield new forms which stand outside the pale of the classification given it is difficult to say. Apparently this will not be the case. Perhaps thousands of hitherto unknown galls might be added 550 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. to the list, but we should expect that they would fall under one or other of the above-mentioned groups. Gall-structures have a peculiar significance for the section of the Natv/ral His- tory of Plants which deals with the question of the origin of species, since they show most distinctly how fundamental deviations from the original plan of con- struction may occur in the adult condition of a portion of a plant. We must be careful to remember always that the immense variety of structures which we call galls would not have existed except for the effect produced on the plants by mites, leaf -lice, gnats, wasps, &c. The foliage of Rhododendron would not have been rolled up, but quite flat, if gaU-mites had not been present; the branches of Pistacia Lentiscus would have borne pinnate foliage with shining dark-green leaflets and not fleshy-red pods if they had not been attacked by leaf -lice ; the leaf- bud of Quercus pubescens would have developed into a long leafy shoot instead of a body like a medlar if the gall-wasp Gynips polycera had not provoked the change; the foliage of Veronica saamtilis would not have been lobed like a hand, but would have had a slightly sinuous margin; and the upper leaves of Thymus would have been green, spoon-shaped, and smooth on the surface instead of circular and covered with white hairs, had no gall-mites settled on them. The flowers of Rhododendron ferrugineum, Lychnis Viscaria, Veronica, Ga/rdamine, &c., would not have "doubled", and the stamens would not have changed into petals, if they had not been under the influence of gall-mites. Of course, the influence of the animals can only produce these effects on parts of plants which are in an embryonic condition. Mature stems and leaves may be eaten and destroyed by insects, but they can no longer be metamorphosed. But the undifferentiated rudiments upon which the influence is effective are, so to speak, formless. Leaves, stems, and fruits arise from tissue-masses having the form of tubercles and cushions, and each tubercle or cushion originates from a few cells which give no indication of what is to develop from them. Nevertheless, experi- ence teaches us that the plan of construction for the plant-member proceeding from these primitive forms is definitely laid down from the beginning for each species, and the idea that the plan of construction is rooted in the specific constitution of the protoplasm of the plant— i.e. in the cell or cells which form the primitive stage or rudiment of the developing leaf, stem, &c., is confirmed. If an alteration in this plan of construction is produced by these animals, it can only be by some alteration of the specific constitution of the protoplasm. How the alteration is effected is just the puzzle which is at present occupying the attention of naturalists. Once it was thought that the formation of galls was the result of injuries caused in the growing tissue by the ovipositor or sucking organ of insects, but recent investigations have not confirmed this view. The cells actually injured by the insect in laying its eggs perish, and consequently lose the ' power of metamorphosis or of producing modified daughter-cells. Cork, closing over the wound, is always formed from the adjoining living tissue, but foi a long THE ORIGIN OP GALLS. 551 time no gall is produced. The eggs deposited in the tissue, or attached to it, are also incapable of directly inciting gall-formation. There is no marked alteration in the neighbourhood until the grub or larva leaves the egg and excretes a fluid substance. Then growing cells of the most varied description are formed adjacent to the larva, and these rapidly assume the peculiar forms which have just been described. This, of course, applies also to cases where the larva has been hatched from the egg at some distance from the spot and has had to seek out a tissue suitable for its dwelling, as also to instances where adult gall-mites and leaf-lice choose out a suitable place for the deposition of their eggs and then secrete a fluid round them when they lay them. If the animal dies, the growth and renewal of the tissue immediately ceases. The cells round the dead body turn brown and die, so that we may conclude the formation of the gall to be due solely to the substance excreted by living animals. Those who investigate galls consider that it is chiefly the acrid "saliva" excreted by the larvae to liquefy their food which acts on the cell-tissue of the dwelling they have selected, but there is no doubt that other excretions may also take part. The chemical composition of this substance is unknown, but we shall hardly be wrong if we include it in the group of nitrogenous compounds called enzymes which were discussed at vol. i. p. 464. Enzymes have the power of altering and decomposing substances, even through the cell-waU, and in this way we can account very simply for a whole series of otherwise inexplicable phenomena in the formation of galls. Moreover, urea or closely-allied nitrogenous compounds may be excreted, so that there is nothing to be said against the view that some of the substances diffuse into the interior of the plant-cells. It is at least certain that the fluid substances excreted by the gall-producing animals, in whatever way they influence the protoplasm in the plant-cells, do not kill it, but actually stimulate it to an extraordinary new activity directly demonstrated by the pro- duction of tissues with a definite external form. Observation shows that these tissues are formed and fashioned differently from what they would have been without the influence of this substance. It follows, therefore, that the substances excreted by the animals have the capacity of affecting in some way the specific constitution of the protoplasm which deter- mines the species in the plant-cells influenced by them. It is specially interesting to note in this connection that it not only is the protoplasm of the cells directly acted on by the excretion which is stimulated to an altered form of constructive activity, but that this stimulus is transmitted from cell to cell in ever- widening circles. The spruce-fir aphis Chermes abietis attaches itself firmly by its beak to the scale of a Fir bud, and can directly influence only a few cells of the young shoot hidden in the bud. Nevertheless thousands of cells on this shoot soon begin to assume an altered form, a proceeding which reminds us strongly of the action of a ferment (cf. vol. i. p. 505), and also brings to our mind the influence exerted by the spermatoplasm on the ovary. The spermatoplasm is only directly concerned with a few cells in the ovule, but these propagate the influence on all sides to 552 ALTERATION OF FORM BY GALL-PRODUCING INSECTS. the carpels and to the receptacle, and sometimes even to the flower-stalk. All these parts would not have developed as they have done had it not been that the minute quantity of spermatoplasm of a pollen-grain had united with a minute cell in the ovule. It will he convenient to consider here the already mentioned similarity between galls and fruits. If the leaf -rudiments in the bud of a Pistacia shrub are not affected by leaf -lice they develop into shining green pinnate foliage-leaves; but if the protoplasm in some of the cells has been altered by the excretions of Pemphigus cornicularius this same rudiment will assume the form of a carpel, and become fashioned into a hollow body deceptively like a pod. The fact that the Pistacia shrub bears plum-fruits and not pods makes it still more remarkable, for the structure arising from the effect of the animal's excretion, when mature, is not like the fruit of the Pistacia, as we should naturally have expected, but like that of a completely different plant species, viz. the Carob (Geratonia Siliqua). The same is true of the metamorphosis caused by the excretion of a gall-gnat {Lasioptera juniperina) on the uppermost leaves of the Juniper (Juniperus communis) which assume a form very like the fruit of the Arbor Vitse (Thuja), and many other instances might be mentioned in which galls are produced in certain species of plants by animal excretions, looking outwardly very like the pods, capsules, nuts, drupes, and berries of other species. This resemblance to certain fruits is rendered the more pronounced by the development upon the galls of pig- ments, wax-like excretions, and hairy coverings, but of course they contain no seeds in their interior — only the larvae of the animals whose excretions produce the changes of form. The wonderful thing is that the metamorphosis of the growing tissue into a fruit-like body is always of the greatest advantage to the animal which has settled in it, since the tissue serves not only for dwelling and food but also for protection against unfavourable weather and against the attack of foes. It is also a fact of great importance that different animals produce differently shaped galls on the same plant. The Bedeguars produced by Rhodites Roscb, the pea-like galls produced by Rhodites eglanterice, and the clustered protuberances produced by Rhodites spinosissimce may all occur side by side on the same rose- leaf (see figs. 361 1- 2. s, p. 533). On the same elm-leaf Schizonev/ra Ulmi produces a wrinkled gall, Tetraneura Ulmi a pocket-gall, and Tetraneura alba a covering gall (see figs. 361*' 5, 6, p_ 533) Tj^g spherical gall of Nematus gallarum and the bladder- like gall of Nematus vesicator occur close together on the foliage of the Purple Willow (see figs. 361 ^ and 361 ^), and one sees Oak -leaves on which the small spangle-galls of four different gall-wasps, viz. Neuroterus lanuginosus, numis- matis, fmnipennis, and Spathegaster tricolor are all present together (see figs. 36411.12.13.". p. 541). It has been shown that some Oaks, for example, Quercus pedunculata, may bear as many as 20-30 different forms of gall produced by as many kinds of gall-wasps. The characteristic shape, colour, and hair-covering of these forms of gall is so constant- that we can state with certainty what gall-wasps have given rise to them. These facts force us to the conclusion that the fluids GENERAL CONSIDERATIONS, 553 excreted by different gall-producing insects are specifically distinct. It is only in this way that we can account for the fact that the same vegetable protoplasm is incited in one case to produce a fleshy covering gall, in another a hollow pocket, and in a third a closed gall-apple as dwellings for the particular insects concerned. It should also be mentioned that the same species of insect produces very similar but slightly different galls on different plants. For example, the gall produced by Nematus pedunculi on the lower side of the white-haired leaves of Salix incana is covered with a white felt of hairs, that which the same gall-gnat pro- duces on the smooth leaves of Salix purpurea is smooth; the gall produced by Bhodites Bosce on the light green leaves of Rosa canina is pale yellow and some- what reddened on that side turned towards the sun; that on the violet leaves of Rosa rubrifolia produced by the same insect-species is dark violet, &c. These distinctions, though only insignificant, show how certain external characteristics founded in the specific constitution of the protoplasm of different plant-species find expression even in the gall-structures. These facts confirm the view that the fluids excreted by different species of insects, as well as the protoplasm of each plant species, have a peculiar composition. It is then quite obvious that the alteration which the protoplasm of a species of plant undergoes under the influence of a specific fluid will be subject to definite laws. The protoplasm of the particular plant-cell receives by reason of the alteration, as it were, a new definite constitution with tendencies not the same as before; but since this constitution determines the outer form of the tissue derived from these cells, the tissue itself will become shaped into a particular specific form. These conclusions are of importance with respect to the question of the origin of new species, inasmuch as they throw some light on the processes which lead to the origin of new forms. We can now say that the alteration in the form of a plant only occurs if the constitution of the protoplasm which forms the starting- point of the plant is itself first altered. The structures known as galls have not the power of maintaining and multi- plying themselves, but when their task is ended they perish. In other words, the progeny arising from the seeds of a plant beset with galls exhibits none of the alterations shown by the members or shoots of the parent plant. If, for example, an Oak which is covered with galls is propagated by seeds, the offspring show no trace of the structural alterations exhibited by the branches, foliage, or flowers of the mother-plant. The only change which is perhaps sometimes retained in the offspring is the metamorphosis of the stamens into petals, which has long been known as doubling, and perhaps also the formation of fasciations, &c. in the floral region, as in Cabbages (where it is known as a Cauliflower). Few attempts have yet been made to investigate this matter. My own knowledge of the subject is restricted to some observations made on the Speedwell Veronica officinalis. Plants of Veronica officinalis which in consequence of the settlement of gaU-mites on them produced double flowers in 1877 in the garden of my country house were planted close beside others free from gall-mites and with normal flowers. In the 554 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. following year the gall-mites settled on the latter also, and the greater part of their flowers then became double. The same result was obtained after living gall-mites were transferred by me to isolated plants of Verondca officinalis with single flowers. These in the following year also bore some double flowers. Fruits with ripe seeds were only produced from the flowers which had remained single amongst the double ones; and the plants which grew up from these seeds always bore single flowers only. The gall-mites disappeared for some unascertained reason — probably they died in the winter. V&ronica officinalis has only two stamens in each flower, and in the double flowers both these and the two carpels are changed into petals so that of course we could not expect fruit and seeds from them. It would not be impossible, however, that flowers of other plant families which are provided with a large number of stamens might behave differently. It might happen, for example, that only some of the stamens would be changed into petals by the gall-mites, and that the carpels would remain capable of fertilization. If on such plants fruits and seeds capable of germination should ripen, the latter might perhaps produce plants with completely and half double flowers. This would be explained by supposing that the alteration undergone by the protoplasm of the cells in the outer part of the flower had extended to the inner, especially to the ovules and seeds, and further to the plants proceeding from these seeds. I would there- fore not undertake to state that the Stocks (Matthiola annua and incana), the Wallflower (Gheiranthus Cheiri), the Pinks (Dianthus Caryophyllus, plrnnarius, &c.), the Poppies (Papaver Bhceas and somniferum), various Ranunculacese (Delphinium, Pceonia, Ranunculus), and many other plants which have long been cultivated in gardens with semi-double flowers, and which produce such flowers when propagated by seeds, had not gained this characteristic in the first place by the influence of gall-mites. It is less probable, though not beyond the range of possibility, that by the grafting of Hawthorn branches whose uppermost leaves have been deeply segmented by the influence of the gall-gnat Cecidomyia Cratcegi, a Hawthorn bush might be produced which would exhibit these deep segmentations and slits on all its foliage. However, these last remarks are the merest suppositions; at present we have not the data on which to base any definite conclusions. THE GENESIS OF NEW FOEMS AS A EESULT OF CEOSSING. The aim of agriculturalists has always been so to cultivate their land as to rear plants likely to grow luxuriantly, to bear good fruit, and thus to afford an abun- dant harvest in return for their pains. Gardeners similarly have made it their endeavour to produce from wild plants races whose flowers are superior to those of the ancestral stock in form, colour, and scent; and the results of their labours are the delight and admiration of all lovers of beauty. In both cases the idea has been to perfect and "ennoble", and the means adopted have been successful to a degree calculated to amaze anyone who studies the history of cultivated plants with attention. The methods which led to these results have not always been deHber- LARGE PRODUCTION OF NEW GARDEN FORMS. 555 ately adopted, nor have they depended on scientific researches. On the contrary, chance observations made by growers in the course of their deahngs with vegetable life as it occurs in nature have been the means of suggesting the first unaided attempts to make crops more productive, fruits and vegetables more palatable, and flowering-plants more pleasing to the eye. The most important method adopted has been the artificial crossing of the species which are brought under cultivation. When we consider that, from time immemorial, Chinese and Japanese gardeners have produced Asters, Chrysanthe- mums, Camellias, Pinks, Peonies, and Roses, of which the majority are the results of crossing, we may assume with certainty that the practice of dusting flowers of one species with the pollen of another species first came into use in those countries. It is true that in Europe the contrivance was known to rose-growers at the time of the Roman Empire, but it was not employed on an extensive scale till the seven- teenth century, when the fashion for breeding TuHps and Auriculas became the rage. The gardeners of that day still made a great secret of their mode of pro- cedure, and it was not till the latter half of the eighteenth century that the production of new forms of plants by the aid of artificial crossing was carried on at all generally. For some decades the rearing of these new forms, which are called hybrids, has been one of the most important parts of a gardener's duties, and we shall not exaggerate if we put the number of hybrids hitherto produced in gardens in the course of the nineteenth century at 10,000. Many hybrids which were great favourites only a short time ago have disappeared from our gardens and have been replaced by others. As in so many other matters, the fashion changes; new forms are in constant request, and horticulturists endeavour to meet the demand by introducing wild plants from the most various regions and crossing them with those already under cultivation. It is now no longer uncommon for gardeners, in advertising some plant which has been brought from distant parts, to recommend it to the trade, not on the ground of its own beauty, but because it possesses flowers of an exceptional colour or leaves of a peculiar cut, and will there- fore, in all probability, if crossed with other species, yield handsome new hybrids. Rose-growers always welcome the discovery of any instance of variation in the Wild Rose as an important event, because, by crossing this Rose with others, they are able to produce a large number of new forms, and there is always the chance that one or other of them may fimd favour with the public. On an average, 60 newly- bred Roses come into the market yearly; in the year 1889 the number even amounted to 115! A Rose cultivator at Meidling, near Vienna, grows in his garden nearly 4200 different kinds of Rose, and yet he is still far from possessing all the forms which have been produced in recent times (chiefly by French growers) by crossing one with another. According to his estimate, the number of Tea and Indian Roses alone is nearly 1400, and the total number of all the difierent Roses which the trade has produced up to the present day amounts to 6400. The plant-forms which are called into existence by the operation of crossing are, in the case of Roses, reproduced largely by means of brood-bodies (cuttings and .556 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. layers) as well as by budding and grafting (see vol. i. p. 213); but the first origin of the new forms is always to be traced to crossing. This statement applies also to many other plants of which gardeners have taken possession, and especially to cases where propagation by seed requires more time and trouble than multiplication of brood-bodies. The kinds of Tulip, Gladiolus, and Lily produced by crossing are propagated most easily by means of bulbs, and the tuberous Begonias, Dahlias, and Gesneraceas by tubers, whilst Pinks, Pelargoniums, Cactuses, and many others are most rapidly reproduced by cuttings. Moreover, these methods ensure the preser- vation of the peculiarities of the new forms unchanged, and such perpetuation of characteristics would be much more difficult to achieve if the plants were propa- gated by means of seeds. On the other hand, a number of new forms which have originated as the results of crosses effected in gardens, such as those of Petunia, Portulaca, and Viola, are reproduced with less trouble and greater rapidity by seeds, and that method is in such cases preferred to the cultivation of brood-bodies. The statement that new forms of plants are bred originally in gardens by any ■other method than that of crossing is incorrect; it is sometimes made in ignorance, but sometimes also with the intention of deceiving. In former times gardeners believed that, in order to produce new forms, it was sufficient to plant different species in close proximity to one another. The idea was that if the seeds of such plants were taken and sown in good soil, there would always be found amongst the seedlings a few forms differing from the parent; these were to be selected for especial care in cultivation, and were to be treated as starting-points of new forms. The gardeners who acted on this assumption had not, it is true, themselves crossed the flowers; and if this was all they meant, there was no falsehood in the state- ment. The operation of crossing was, however, performed without their knowledge by hive- and humble-bees and other insects, and the planting together of the •different species was only of advantage inasmuch as it facilitated the conveyance of pollen from one species to the stigmas of another. A celebrated grower of the old school once assured me, in all seriousness, that he did not himself cross the plants he reared, but that he had repeatedly observed that early in the morning, soon after a flower opened, it put forth infinitesimally flne threads which radiated in all directions and reached across to the flowers of other plants, forming in a short time a web like that of a spider! I would not have mentioned this statement were it not for the importance of pointing out the unreliable character of so many of the statements made by gardeners, especially in the past; and I repeat that the person responsible for the above communication is a well-known and much-esteemed horti- culturalist. Gross inventions such as the above would, of course, be at once seen through and rejected by any thoughtful man; nevertheless, in some instances, reports of growers, likewise untrue or inaccurate, but not bearing the stamp of improbability so plainly upon the face of them, have been credited and have even found their way into books, particularly into those whose authors have omitted to confirm the reports by watching the garden-experiments from beginning to end themselves. The statements are then not infrequently quoted as "results obtained THE PRODUCTION OF HYBRIDS. 557 by trustworthy experiments made by gardeners" and relied upon for the foundation of "laws based upon facts"; theories are then built up upon them, and are copied from one book to another. It becomes very difficult afterwards to get rid of such propositions, especially if they afford support to the hypotheses of distinguished savants. An instructive example of the kind of thing referred to is afforded by the following statement which for a long time obtained currency in works on Botany: "Hybrids resulting from the crossing of two species exhibit two forms, in each case, according as the pollen employed in generating them belongs to the one or to the other species." There are, of course, two ways in which a pair of species,, which may be represented by the letters A and B, may be crossed. In one case the pollen from A is transferred to the stigma in B, whilst in the other case the pollen is taken from B and transferred to the stigma in A. Now, it has been asserted that it is possible to determine from the form of the hybrid which of the two parent-species supplied the pollen. The hybrid is said to resemble that parent more closely from which the pollen was derived so far as its flowers are concerned, but to show greater affinity to the fertilized stock in respect of its foliage. This is, however, absolutely untrue. All careful experiments, made without prejudice have pointed to the conclusion that it makes no difference to the forms, either of the leaves or of the flowers, whether the pollen has been taken from the one parent-species or from the other. Besides this, in most instances the approximation of the hybrid- form to one or the other stock affects all parts of the hybrid plant in an equal degree, and not the flowers or the foliage only. In the rare cases where a hybrid has flowers more like those of one parent and leaves more like those of the other, it is just as likely for one stock as the other to have yielded the pollen used for the cross. But it would be a thankless task to attempt to correct all the hasty, careless, and erroneous statements, past and present, and it will probably be of greater utility to give a general survey of what has been ascertained concerning the form and temperament of hybrids by impartial observers who have taken into account all the concurrent circumstances, and have allowed for the sources of error inci- dental to the experiments. The formation of a hybrid plant presupposes two stock-plants as parents which have different properties and characters. A cross must take place between the two — that is to say, the stigma of the one must be dusted with pollen from the other; the pollen must put forth pollen-tubes and an effectual union between spermato- plasm and ooplasm must be effected. For the sake of brevity, the plant from which the pollen, and therefore also the spermatoplasm, is derived is called the paternal stock or father-plant, and that which has its stigma dusted and its ooplasm ferti- lized is called the maternal stock or mother-plant. As a plant's external marks and characteristics, which are perceptible to our senses, are an expression of its internal organization and of the specific constitution of its protoplasm, it may be assumed that the plant-individual which owes its existence to the union of two protoplasts; 558 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. of different constitutions possesses marks and attributes, some of which are charac- teristic of the father and others of the mother. As a matter of fact, the individuals which growers call by the various names of hybrids, mongrels, and bastards answer .to this supposition. Some of their attributes and external characters are derived from the paternal, some from the maternal stock. If the relative admixture of the two stocks were to be determined in the case of a hybrid by summing up the characteristics transmitted from each source, the share of each parent would in many cases be found to be a half. In such a case it is usual to say that the hybrid is a mean between its parents. Examples are afforded by Oeurn hyhridum, which originates from crossing Geum, montanum and G. rivale, Hieracium stolonijlorum, derived from Hieracium aurantiacum and H. pilosellceforme, and Nuphar inter- medium, a cross between Nuphar luteum and N. pumilum. But it also happens sometimes that the proportion of characters inherited by a hybrid from one of the parent-stocks amounts to about two-thirds, leaving only one-third as the propor- tion inherited from the other, and in that case the hybrid is said to exhibit an approximation to one of the parent-species. The Saxifrage hybrids are very instructive examples of this class. The stigmas in a flower of Saxifraga aizoides were dusted with the pollen of S. ccesia. Fertilization was successfully effected, and a capsule containing fertile seeds came to maturity. The plants reared from these seeds resembled in some cases the intermediate form described by Botanists under the name of Saxifraga patens; others approximated more closely to the paternal stock, and a third group to the maternal stock. A similar result was obtained by crossing a flower of Saxifraga aizoides with the pollen of S. mutata. From seeds of the same capsule were produced two different kinds of hybrid, of which one was intermediate between the parents (Saxifraga Hausmanni), whilst the other {Saxifraga inclinata) approximated more closely to S. mutata. These experiments point to the conclusion that the share which each parent has in the form of a hybrid depends upon the quantity of its spermatoplasm or ooplasm, as the case may be, used in the production of the seedling relatively to the quantity derived from the other parent; and in the case of hybrids issuing from the same fruit we are forced to suppose that the variability of the degree in which a particular form is inherited is caused by the existence of quantitative differences between the several spermatoplasmic and ooplasmic nuclei (or cells) which coalesce in pairs in the interior of the same ovary. Important evidence in support of this supposition is derived from experiments made in connection with the crossing of various composites of the Thistle genus (Cirsium). In these plants each ovary contains a single ovule only, and therefore each fruit can only produce a single individual. On the other hand, each capitulum is composed of a large number of florets, and when a head is in full bloom nearly a hundred mature stigmas project from it in close proximity to one another. If pollen from another species is transferred by means of a paint-brush on to these stigmas, wholesale crossing, so to speak, ensues; and it may be confidently expected that a proportion of these simultaneous crosses will be effectual. The harvest of fruit DIVERSITY OF HYBRIDS FROM SAME PARENTS. 559 from the capitula selected for the experiment was not, it is true, very plentiful, but some fruits invariably came to maturity. If these one-seeded fruits, all of which have originated at the same time and under similar conditions, are sown, the indi- viduals produced from them are but seldom like one another. The attributes and external marks of the paternal and maternal stocks respectively are in some repre- sented in the proportion of about 2 : 1, in others in the proportion of 1 : 2, and in a third class in practically equal proportions; cases are even known where four, five, or more distinct forms of hybrid have developed from the fruits of a single capitulum. I found the most striking diiferences amongst the hybrids obtained by crossing a capitulum of Girsium oleraceum with the pollen from a capitulum of Cirsium heterophyllum. Those produced by crossing a head of Cirsium Pannoni- cum with pollen from a head of Girsiwm Erisithales were scarcely less conspicuously diverse. As it must be presumed that the spermatoplasm and ooplasm — or rather the spermatoplasmic and ooplasmic nuclei — which have been formed in the adjacent florets of a capitulum are alike in composition and ultimate structure, the variety of the hybrids springing from such florets must depend solely on the relation between the masses of ooplasm and spermatoplasm respectively which coalesce in a flower, and the ratio between these masses must be in the one case 1 : 2, in a second 1:1, and in a third 2:1. Of course these ratios give but an approximate measure of the degree in which each parent has participated in the generation of the inter- mediate forms. Where five kinds of intermediate forms occur the series would be approximately represented by the ratios 1:4, 2:3, 1:1, 3:2, and 4:1. The first time flowers of a head of Cirsium Erisithales were crossed with pollen from Cirsium palustre they produced two different forms of hybrid, one of which approximated to the paternal, the other to the maternal stock; but no form exactly midway between the two parent-stocks was obtained. A second trial of the same experiment resulted in the production of a single form which did occupy this inter- mediate position. These observations show that there is no definite law governing the form of hybrids; one might even say that irregularity is here the rule. On one occasion all the individuals which are the result of a cross between two species are alike, on another occasion they constitute more or less numerous links in a chain of intermediate forms. Hybrids which do not stand midway between the parent-species but approximate to one or other of them are called goneoclinic (7'»'e'>s= parent, kUpw-I lean). That such hybrids may arise from a first cross is established beyond question by the experiments above recorded; but there is also another process whereby they are produced, and that is the crossing of a hybrid with one of its parent-stocks. These crosses are effected in great numbers, and all observers agree that in general the results are better in such cases than where two species are crossed; that is to say, if a hybrid's stigmas are dusted with the pollen of one of the parent-species a larger number of fertile seeds may be looked for than if the plants crossed were of distinct species. The individuals resulting from the cross of a hybrid with one of the parent- stocks also occupy, as might be expected, a position as regards marks and attributes 560 THE GENESIS OF NEW FOEMS AS A RESULT OF CROSSING. intermediate between their progenitors; here again, however, forms are not always alike, and sometimes several intermediate forms make their appearance. If the hybrid derived from Girsium Erisithales and C. Pannonicwm, which inherits equally from both parents, be crossed with pollen from Girsium Erisithales, the resulting individuals have a stronger resemblance to the latter species and are instances of goneoclinic hybrids. But when Girsium Erisithales and C. Pannonicum are crossed for a first time hybrids are also produced which are not exactly midway between the parents but are more like G. Erisithales. These naturally resemble the goneoclinic hybrid derived from crossing the offspring of G. Erisithales and G, Pannonicum with G. Erisithales, and if one were not in a position to follow the history of the origin of the hybrid in question its characteristics alone would not furnish sufficient data for a judgment as to the mode of production of the goneo- clinic hybrid. Hybrids which are the offspring indirectly of three different species are called ternary hybrids. The word "indirectly" must be specially emphasized here, lest the reader should fall into the error of supposing it to be possible that pollen-tubes from two or more species should simultaneously fertilize a single ovule. Such an occurrence never happens, not even if a mixture of pollen belonging to two or more different species be placed upon the stigma of a particular flower. On the other hand, it has been ascertained by numerous experiments that if the hybrid-offspring of two species is crossed with the pollen of a third species, or vice versa, another hybrid is produced. Thus, for instance, if the hybrid of Linaria genistifolia and L. purpurea is crossed with the pollen of L. striata the result is a ternary hybrid. In one experiment the stigmas in a capitulum of Girsivmi Linkianum (the hybrid offspring of G. Erisithales and G. Pannonicum) were dusted with pollen taken from G. palustre. A considerable number of fruits ripened, and the hybrids which arose from them were ternary hybrids, exhibiting marks and characteristics of G. Erisithales, G. Pannonicum, and G. palustre respectively. These hybrids, more- over, were not all alike; some of them bore striking resemblance to Girsium aquilo- na/re, the hybrid produced by crossing G. palustre and G. Pannonicum, and exhibited very few of the peculiarities of G. Erisithales, whilst other individuals were ex- tremely like Girsium ochroleucum, the hybrid obtained by crossing G. Erisithales and G. palustre, and only showed slight indications of its relation to G. Pannoni- cum. Growers of garden-flowers have achieved great success in producing orna- mental ternary hybrids in a number of genera (Achimenes, Begonia, Bianthus, Gladiolus, &c.). Ternary hybrids of various Willows are also met with in gardens; one of these is obtained by crossing 8alix Gremsensis, a hybrid of Salix Gaprea and S. daphnoides, with S. viminalis, another by crossing Salix Wichuroi (the hybrid-product of S. incana and S. purpurea) with S. cinerea, and so forth. Willows have been also used for the prosecution of still further experiments. The crossing of two hybrids of different parentage on both sides resulted in the genesis of hybrids combining four species of Willow. Indeed, six different species of Willow have on one occasion been combined by successive crossing — Wichura having COMBINATION OF PARENT-CHARACTERS. 561 succeeded in producing in Breslau a compound hybrid in which were united Salix Gaprea, S. daphnoides, S. Lapponum, S. purpurea, 8. Silesiaca, and S. viminalis. It need hardly be said that the characteristics of the six ancestral species in such a case as that of the last-mentioned hybrid are not easily identified. Even where a hybrid is the offspring of a single cross between two species it is not always easy to determine its origin from its external appearance, and in the absence of any knowledge of the history of its production. The characteristics of the parent-stocks are not combined in all hybrids according to a single definite rule. Sometimes the combination seems to amount to a complete fusion, so that the form produced might be compared to an alloy of two metals. Very often a new form is generated which combines in a definite geometrical ratio the characteristics of the parents in respect of the position and direction as well as the shape and size of its separate parts. In that case there is said to be a union of the parental characters. The structural characters of both stocks are represented unmodified, but are so closely bound together as to suggest a composite crystal founded upon two different crystalline forms. Just as in definite combinations of crystals the faces of one component form are dominant, and determine the general aspect in one case and those of the other component form in another case, so in many hybrid plants sometimes the attributes of the one parent, sometimes those of the other, are most conspicuously reproduced. Other hybrids again are analogous to combinations in which both crystalline forms are equally represented. Again, in addition to the above classes of hybrids wherein the parental characters are either completely merged together or intimately united there are many, cases where those characteristics are present almost un- changed, and subsist side by side like the particles of a rock. The most common case of this mixture or juxtaposition of properties occurs where the hybrid displays hairs, glands, or prickles of two forms interspersed together, one of which is identical with the form of the structure in question exhibited by the maternal stock, whilst the other has been inherited unchanged from the paternal stock. Or, one part of the hybrid's flower may be coloured like one parent and another like the other parent. Hybrids are also known in which the foliage is almost indistinguishable from that of one parental stock whilst the flowers are like those of the other, so that at first sight a hybrid of the kind looks as if it were a plant of the former species with flowers of the second species affixed to it for a joke. On closer inspection some slight differences may be perceived between the leaves and flowers of the hybrid and those of the parent species respectively, but this does not alter the fact that hybrids exist whose leaves resemble far more closely those of one parent, whilst their flowers are more like those of the other. Probably it was the occurrence of such a hybrid which suggested the proposition referred to on p. 557 that in the product of a cross between two species the flowers reveal the paternal and the foliage the maternal stock. But this statement is incorrect, as was said before, for some hybrids approximate to the maternal stock in respect of their flowers, and to the paternal stock in respect of their leaves. Of the three ways in which the parental characters may be combined in a Vol. II. 86 662 THE GENESIS OF NEW FOEMS AS A RESULT OF CROSSING. hybrid— i.e. by fusion, by union, or by mixture— one alone sometimes prevails in all the parts; but usually, on the contrary, there is an incalculable degree o£ variation in this connection. There are, for example. Rose hybrids in which the outline of the leaves exhibits a union, the colour of the flowers a fusion, and the hairs a mix- ture of the corresponding characters in the parents. A brief account will now be given of a few examples to illustrate the manner in which the combination of parental characters in a hybrid is manifested throughout the structure of the plant and particularly in the forms of the stems, leaves, and flowers, and in those of thorns, bristles, hairs, and other epidermal appendages. The species of Willow known as Salix Caprea grows in the form of a little tree, with thick, straight, erect branches, each of which bears about 25 leaves; Salix repens, •on the other hand, is a low shrub with a procumbent stem and slender, rod-like branches ascending in curves from it, and each bearing about 40 leaves. The hybrid of these two is a small tree with a bent stem and ascending branches, which in length, thickness, curvature, and direction are intermediate between the Salix Caprea and Salix repens, and which are furnished with some 30 leaves apiece. Again, the foliage-leaves of Prunella vulgaris have their margins entire, whilst those of P. laciniata are deeply cut, and their hybrid Prunella intermedia has lobed leaves. The leaves of Potentilla sterilis (or P. Fragariastrum) possess three leaflets, each leaflet being furnished on either side with from four to five serrate indentations. The leaves of Potentilla micrantha likewise possess three leaflets, but each leaflet has from seven to ten serrate teeth on each side. In the hybrid of these two species the leaflets have from six to eight indentations on each side. The shape of the leaf is, as is well known, intimately connected with the course, ramification, and disposition of the bundles called nerves. Now, if the net-work of strands in the leaves of the parent-species is compared with that in the leaves of the hybrid, it is astonishing to find how the union of the two systems may be traced in the minutest details. No group of plants lends itself better to this sort of investigation than the Willows. Even if a single leaf of the hybrid offspring of two species of Willow is submitted for inspection, it is possible in most cases to determine, from the number and distribution of the nerves, the identities of both its parents. Salix purpurea produces one hybrid when crossed with Salix grandi- folia, and a second when crossed with Salix Caprea. Salix grandifolia has twice as many lateral nerves in each leaf as Salix Caprea, and this difference is repro- duced in the corresponding hybrids, whose leaves in all other respects resemble one another closely. The involucral leaves of Composites are well known to be extremely diverse in shape, and systematic Botanists have always attributed great importance in the discrimination of species to the size, shape, and margination of these leaves and to the peeuHar appendages at their apices. Now, the hybrids of Composites not infrequently have involucral leaves which differ widely from the forms character- istic of the parent-stocks. Thus, for instance, each leaf of the involucre in a capitulum of Centaurea rwpestris terminates in a long yellow prickle, whilst the COMBINATION OF PARENT-CHARACTERS. 563 corresponding structure in Centaurea Scabiosa is bordered by a broad, membranous, fringe -like edge of a dark -brown colour. In Centaurea sordida (Grafiana), the hybrid offspring of these two species, each involucral leaf is edged with a narrow, light-brown membranous and fringed border, and terminates in a short yellowish prickle. A very instructive example of the union of parental characters affecting all the different parts of the floral region is afforded also by the Labiate Marrubium remotum, which is produced by crossing Marrubium peregrinum and M. vulgare. The small tuft-like inflorescences in the leaf-axils of Marrubium peregrinum include from 10 to 18 flowers, those of M. vulgare from 4 to 5, and those of the hybrid M. remotum from 5 to 10. The calyx of M. peregrinum is grey and covered with felted hairs, and its margin is provided with five large subulate teeth which terminate in straight points. The calyx of M. vulgare is green and sparsely clothed with hairs, and its edge has ten small teeth which terminate in stiff reflexed points like hooks. Five of these teeth are rather longer than the rest. The calyx of M. remotum is greyish -green and clothed with a loose felt; its edge is furnished with five big subulate teeth which terminate in stiff out-curved points, and have from two to five very small teeth interspersed between them. The three lobes of the under-lip of the corolla are almost of equal length in Marrubium peregrinum, whilst in M. vulgare the middle lobe is three times as long as the two lateral lobes. In the hybrid Marrubium. remotv/m the middle lobe of the lower lip is half as long again as the lateral lobes. An excellent example is also afforded by Dianthus (Enipontanus, a hybrid Pink resulting from a cross between Dianthus alpinus and B. superbus. In B. alpinus the bract-like scales at the base of the calyx are almost as long as the tube of the calyx itself, whilst in B. superbus their relative length is only a quarter or a third; in the hybrid B. (Enipontanus these bracts are half as long as the tube. The petals of D. alpinus have broad lp.minse beset at the margin with short triangular teeth, those of B. superbus have their laminae slit up into a number of narrow strips, and those of B. (Enipontanus have deeply-incised laminae, the margins being divided into linear segments. The dimensions of the various parts of the flower in a hybrid also exhibit in most cases a combination of the corresponding parental character- istics. Thus, for instance, the perianth of the Orchid Gymnadenia conopsea has a long spur — that is to say, the segment of the perianth known as the labellum is produced backwards into a saccate protuberance supposed to resemble a spur, and this portion of the petal in Gymnadenia conopsea is 15 mm. in length. In Nigri- tella nigra, on the other hand, the spur is very short, measuring about 2 mm. The hybrid of these two Orchids, Nigritella suaveolens, has a spur varying from 5 to 7 mm. in length. In Willow hybrids the number of stamens in each flower of the hybrid is invariably between the corresponding numbers in the two parent- species. For example, the number of stamens in a flower of Salix alba is 2, in Salix pentandra 5-12, and in their hybrid Salix Ehrhartiana 3-4. The cellular structures produced from the epidermis of the stem and leaves which are differentiated as hairs, bristles, scales, glands, &c., and are classed together 564 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. under the name of investments (^Tidwrnentum), are very constant characters in most species of plants. The occurrence of stellate hairs, in particular, is looked upon by systematic Botanists as an important point in assisting them to distinguish between similar species, and so also is the presence of glandular hairs composed of simple rows of cells, and terminating in globular bladders full of ethereal oUs. Hybrids exhibit the most varied combinations of the indumenta of their parents. In the majority of cases the characteristics of the two stocks in this respect are mixed, but less frequently are they united, and in the latter case the shape, size, and number of hairs, bristles, scales, and glands are intermediate between those of the same appen- dages in the two parent-species. The Lungwort genus (PulmoTMria), which has a special tendency to hybridization, includes only a few species, but each one may be recognized by the nature of its indumentum. Thus, Pulmonaria officinalis is dis- tinguished by the thousands of short unicellular prickly hairs, scarcely perceptible to the naked eye, which are interspersed amongst the long scattered bristles on the upper surfaces of the leaves. In Pulmonaria angustifolia the leaves are destitute of these minute prickles, but bear on their upper surface a more abundant quantity of straight appressed bristles of equal length. The leaves of the hybrid derived from the two preceding species, viz. Pulmonaria hybrida, are richly supplied with long bristles, and interspersed amongst these may be seen a large number of shorter bristles which are about two or three times as long as the prickly hairs of Pul- monaria officinalis. A very instructive example is also aiforded by the hybrid Rhododendron intermedium, which is easily produced by crossing the two Alpine- Eoses (Rhododendron ferrugineum and Rhododendron hirsutum). The upper faces of the leaves of R. ferrugineum are dark-green, smooth, and shining, whilst their backs are rusty and dull owing to the presence of a dense crowd of tiny scales. The margins are not ciliate. The leaves of R. hirsutum are light-green and beset with scattered whitish glands (see vol. i. p. 232, figs. 54 ^ and 54 ^), and their margins are fringed with long hairs. In Rhododendron intermedium both kinds of epider- mal appendage are displayed side by side. The under surface of the leaf is furnished with brown scales, though not so profusely as in Rhododendron ferrugineum,, and its edge is fringed with hairs, but not so thickly as in R. hirsutum. The same sort of thing occurs in Eoses, Cinquefoils, Blackberries, Drabas, Hawkweeds, and many other plants. Where one parent Eose bears only non-glandular and the other only glandular hairs the hybrid is sure to be clothed with a mixture of the two kinds of hairs. Several species of Cinquefoil (Potentilla) have stellate or tufted hairs, whilst others are entirely free from them and bear none but simple hairs on their leaves. In hybrids derived from two of these species — one with compound and the other with simple hairs — stellate or fasciculated hairs are invariably intermixed with a large number of simple hairs. A few species of the perennial Whitlow-grasses (Braba), which are indigenous to mountainous districts in Central Europe, have rectilinear anvil-shaped hairs, whilst others have three- or four-rayed stellate hairs. In the hybrids which spring from these different species rectilinear and stellate hairs grow together on the same leaf. If the hairs of two parent-species are of the MINUTE STRUCTURE OF HYBRIDS. 565 same form but of unequal length, those of their hybrid offspring have a length which corresponds approximately to the mean between the lengths in the parent- species. Thus the length of the hairs on the backs of the leaves is 0-3 mm. in Salix aurita, 1-2 mm. in Salix repens, and 0-6 mm. in their hybrid Salix plicata. The hairs in Sahx Caprea measure 0-8 mm., in Salix viminalis 0-3 mm., and in SaUx acuminata, their offspring, 0'5 mm. Whenever one stock is glabrous and the other hairy, one may be quite sure that the corresponding parts of their hybrid will be furnished with hairs, but less profusely than the parent-species from which that particular characteristic is derived. This is the case, for instance, with Primula Sturii, the hybrid produced by crossing the glabrous Primula minivia with Pri- mula villosa, which has glandular hairs. The leaves of the latter are thickly covered with these hairs, which vary from 0'7 mm. to 1 mm. in length, and Primula Sturii has scattered glandular hairs which measure 0'3 mm. The hybrids obtained by crossing the Purple Willow (Salix purpurea) with the Common .Osier (Salix viminalis) are distinguished by Botanists into two sections, one of which — Salix rubra — -approximates to the Purple Willow and the other — Salix elceagnifolia — to the Common Osier. The leaves of the Purple Willow when mature are glabrous at the back, those of the Common Osier have small glistening hairs lying appressed to their under surfaces, parallel to the lateral nerves, and measuring 03 mm. There are about 1800 of these hairs on a square millimetre. The hairs of the hybrid Salix elceagnifolia are of the same length as those of ;Si. viminalis, but there are only about 800 of them to the square millimetre, whilst the hairs of the hybrid Salix rubra are somewhat shorter, and there are only 400 to the square millimetre. Recently the discovery has been made by Wettstein that the form and dis- position of the cells and tissues in hybrids is also a combination of the corresponding characteristics in the parent-species. The various species of the Pine genus (Pinus) may be distinguished with certainty by the anatomical structure of their needle-shaped leaves, in particular by the thickness of the epidermal cells, the number of the stone-cells lying beneath the epidermis, and the number of the resin- ducts. In the hybrids the anatomical characters of the parents in these respects are united, and the result is indeed often an exact arithmetic mean between the two. Thus a needle of the Scotch Pine (Pinus sylvestris) contains from 6 to 10 resin-ducts, that of the Mountain Pine, Pinus Mughus (montana), contains from 3 to 5, and that of the hybrid offspring of the two from 5 to 7 such ducts. The Junipers (Juni/perus) afford a similar instance. In their case the leaves are dis- tinguished by the various thickness and length of the layer of sclerotic-cells which covers the back of each leaf, by the width of the resin-duct running through the middle of the leaf, and by the number of the cells encasing that duct. In the hybrids, such as Juniperus Kanitzii, which is produced by crossing Juniperus communis and J. sabinoides, there is evidently a union of the parental attributes in the corresponding cellular structui'es in the leaves. It has also been shown by Hildebrand that in the Wood-Sorrel (Oxalis) hybrids also the anatomical characters of the parents are united, but by far the most comprehensive study which has 566 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. been made in recent times into the minute structure of plant-hybrids is by Mac- farlane. He selected a number of hybrids, and worked through their anatomy from base to apex in the most painstaking manner. He dealt with roots, stems, leaves, and the various portions of the flower, always comparing their various tissues (both as regards size, form, and distribution) with those of their parent- forms. And his result is to confirm what has been written above, though it is obvious he had never read these pages. Amongst the more interesting of his results may be mentioned those on starch-grains. Of course in a great many of the parent-forms uniting to form a hybrid there is no recognizable difference in the size or structure of the starch-grains. But in the genus Hedychium (belonging to the family Zingiberacese) exceptions to this rule were found. Thus, those of Hedychium Sadlerianv/m are intermediate in form and size between those of its two parents, H. Oardnerianum and H. coronarium; and those of a hybrid between H. elatwm and H. coronarium exhibited similar intermediate characters. It is important to note also that the aromatic substances and colouring matters produced in the cells of a hybrid are inherited partly from the maternal, and partly from the paternal stock. As we have several times already had occasion to mention, the various species of the Rose genus may be recognized at once by their peculiar scent. The perfume of -Rosa Centifolia is the one which in particular is understood by the rose-scent, but it is very difierent from that of Rosa al/pina, and the latter in its turn is unlike any of the scents emitted by Rosa arvensis, R. Gallica, R. Indica, fee. Rosa Nastercma has a scent strongly resembling that of Pinks, whilst Rosa lutea and R. punica are notorious for their disagreeable smell. Now the hybrid Roses emit odours in which the scents of the parent-species are merged together in a great variety of ways. Usually the scent of one stock predominates, and there is only a suggestion of the other. Sometimes, however, an entirely new scent is evolved from the fusion of the two, as is the case, for instance (according to Mac- farlane), in Hedychium Sadlerianum, the hybrid between H. Oardnerianum and H. coronarium; and, again, in other cases, one of the component odours is inten- sified and the other is extinguished. The same statement applies to the aromatic substances to which the scent of the foUage is due. The hybrids of Rosa glutinosa, Rosa rubiginosa, and R. rugosa, with Rosa Gallica and R. Centifolia, are very interesting in this connection. The aromatic substances which are con- tained in fruits and excite our nerves of taste are also inherited, partly from the maternal and partly from the paternal stock. Owing, however, to the difficulty of naming the various sensations of smell and taste it is of little use to discuss the subject more fully. As regards the colouring-matters reproduced in hybrids the first point to notice is that in cases where the foliage is of diflFerent shades of green in the parent-species the leaves of the hybrid exhibit a shade intermediate between the two. Conspicuous instances of this are afforded by the hybrid Willows derived from Salix nigricans and S. purpv/rea. In both these species the foliage becomes black when it withers, and this characteristic is transmitted, though not in its full strength, to the hybrids COLOURS IN HYBRIDS. 567" which S. nigricans and S. purpurea form with other Willows whose foliage turns brown when it dries up. The colour of the flowers in hybrids is usually the result of a fusion of the colours in the parent-species; less frequently it is a mixtu/re of the original colours. The cases of fusion occur especially amongst the hybrids of Orchids, Louseworts, Anemones, Pulsatillas, Medicagos, and Mulleins. If the tone of the red or blue petals in one parent-species is dull and in the other bright, the same colour reappears in the hybrid, but of a medium tone. Thus the colour of the petals in Gymnadenia conopsea is rose-red and in Nigritella nigra dark blood- red, whilst in their hybrid, Nigritella suaveolens, it is bright carmine. The corolla of Pedicularis incarnata is of a subdued carmine tint, and that of P. recutita of a dark reddish-brown, whilst their hybrid, P. atrorubens, has a dark purple corolla. Where the floral colour of one parent-species is white and that of the other a full yellow, red, or blue, the hybrid's flower usually exhibits a pale yellow, red, or blue coloration. The flowers of Anemone nemorosa are white, those of A. ranunculoides golden-yellow, and those of their hybrid, A. intermedia, sulphur-yellow. The colour of the flowers in a hybrid whose parents have yellow and violet, or blue flowers, respectively, is very remarkable. Medicago media, which is the hybrid offspring of the yellow-flowered Medicago falcata, and the blue-purple flowered M. sativa very often has green corollas. The hybrids {Verbascv/m commutatu/m, V. rubigi- nosum, V. Schmidtii, V. versiflorum, &c.) obtained by crossing the yellow-flowered Mulleins with Verbascum, phoeniceu/m, whose flowers are a conspicuous purple, all display a bright brown tint in their corollas. The colour in question is just the same as that which is produced by mixing gamboge with the purple prepared from carmine and indigo. Quite a different tint is exhibited by the corolla of Verbascum pseudophoeniceum, the hybrid generated by crossing V. Blattaria and V. phceni- ceum. One of the parent-species (F. Blattaria) in this case has pale yellow and the other (V. phoeniceuTn) violet-purple corollas, and in the hybrid (V. pseudophoeni- cev/m) the corolla is pale crimson. Nor are cases wanting in which hybrids have been produced from forms with red and blue flowers respectively. The brilliant scarlet-flowered Delphinium nudicaule has been crossed in the Edinburgh Botanic Garden with the dark blue-flowered D. cashmirianumi, the hybrid product being of a lurid purple-red hue. Darwin obtained by crossing the red and blue Pimpernels (Anagallis) a progeny some of which were blue, some red, and some intermediate in colour. As a final instance of this colour- fusion may be cited the hybrid Pitcher- plant Nepenthes Mastersiana. This hybrid is produced from N. sanguinea, the pitchers of which are of large size and vary in colour from greenish-scarlet to crimson, and of N. Kharsiana, which bears long narrow pitchers, varying from yellowish -green to dull red-green. The hybrid (says Macfarlane) presents a corresponding latitude in colour effect, though on the average it is greenish-crimson. The hybrids which originate from crosses between Primula Awricula, whose flower is all of one colour, and Primula Garniolica, P. hirsuta, P. Oenensis, P. villosa, &c., which have bi-coloured flowers, are also of great interest in this con- nection. P. pubescens, the hybrid produced by crossing P. Auricula and P. hirsuta. 568 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. is the stock from which the garden Auriculas are derived. The colour of the corolla in P. Auricula is a uniform golden-yellow excepting that at the throat, i.e. at the junction of the tube with the expanded limb, there is a floury efflorescence which, like that covering the calyx, pedicels, and bracts, is due to a peculiar modification of the epidermis. The corolla of P. hirsuta is bi-coloured; the segments of the limb are violet-red, whilst the throat is white. The two tints are sharply marked oiF from one another, and in consequence a white five-rayed star is seen in the middle of the flower. There is in this case no trace of a floury efflorescence. In the hybrid offspring of these two Primulas both the violet-red of the limb and the white of the throat are blended with yellow; the former exhibits a touch of brown, and in the middle of the flower is a pale-yellow star. It is much less common for those floral colours which are inherited by a hybrid from the parent-species to be displayed in juxtaposition than in a blended condition. Since the time of the Eoman Empire gardeners have crossed the red-flowered Rosa Gallica and R. Bamascena with the white-flowered Rosa alba and obtained thereby hybrids in which the petals are striped and spotted longitudinally with red and white (so-called " York and Lancaster " roses). Similar cases occur amongst hybrid Calceolarias, Pinks, Petunias, and Wood-Sorrels, and instances of Tulip and Iris hybrids are also known where the perianths exhibit the two different colours of the parent-species side by side in streaks and patches. A hybrid of Iris Florentina and /. Kochii is especially deserving of notice. The perianth in /. Florentina is milk- white and that of /. Kochii is dark violet. The hybrid of these two species was flrst obtained in May, 1871, in the Botanic Garden at Innsbruck; one of the indivi- dual plants thus produced had two of the outer and one of the inner members of the perianth shaped like those of I. Kochii and of a deep violet colour, and one of the outer and two of the inner members shaped like those of I. Florentina and milky- white in colour. This arrangement of colours re-appeared year after year until in 1877 a single flower, in which the lower white members had some dark violet streaks widening out from the middle to the edge of the perianth also made its appearance. A second plant of the same hybrid developed flowers which only differed from those of I. Florentina in that a few of the white petals had dark violet streaks widening out towards the circumference. An equally noteworthy case is that of a hybrid reared in the Botanic Garden at Florence from /. Oer- manica and I. sambucina, of which a specimen was sent to me in 1872. One inferior and two superior perianth-members displayed on one half of their surfaces the colour and pattern peculiar to I. sambucina, and on the other half those charac- teristic of /. Oermanica. The rest of the perianth could not be distinguished except by its smaller size from that of Iris Germanica. It must not be supposed, however, that the presence of variegated stripes, patches, or speckles on petals is always an indication of hybridity. Viola poly- chroma, a very common Alpine species, not infrequently produces simultaneously two, three, or four open flowers, every one of which presents a different mixture of tints, and amongst plants of this species covering only a small patch of ground it GRAFT-HYBRIDS. 569 would be easy to find 100 blossoms with corollas differing from one another in the distribution of their colours and in the arrangements of the spots and streaks upon them. Similar phenomena are exhibited by Iris pumila and Polygala amarella. The flowers in Polygala amarella are equally blue and white or sprinkled with blue and white, and it is also no rare thing for plants to bear white flowers interspersed with a few which are sprinkled with blue. In the same way several species of Anthyllis, Euphrasia, Oaleopsis, Linaria, Melittis, Ophrys, Orchis, Saxifraga, &c., exhibit considerable variation in the colours and markings of their petals, which yet is not to be attributed either to hybridization or to the influence of soil or climate. Reference must also be made here to the large number of species (already mentioned on p. 194) in which the floral coloration is by turns blue and white, red and white, blue and red, yellow and white, and so on. Heterochromatism, i.e. the change in the coloration and marking of petals, serves, in fact, in some plants as a specific character. Contrasting with these heterochromatic species are those with homochromatic flowers, which, as far as experience has shown, invari- ably present the same colour and pattern, and only exhibit a slight variation in the depth of the colour when subjected to the influence of light of varying degrees of intensity. Iris Kochii and I. Florentina, Primula Awricula, and P. hirsuta, together with other pairs of species referred to above as the progenitors of hybrids of special interest, belong to the category of plants possessing homochromatic flowers, and it is obvious that in the hybrid offspring of such plants the floral coloration would be an important sign of identity. This will be the most convenient place in which to introduce a few words concerning the Bizzaria of Italian gardeners, and also concerning so-called graft- hybrids. The name of Bizzaria has been given by the Italians to an extremely curious Orange. Gallesio (1839) states that this Orange-tree produces at the same time foliage, flowers, and fruit identical with the Bitter Orange {Citrus Aurantivm) and with the Citron of Florence (Citrus medico), and likewise compound fruit, with the two kinds either blended together, both externally and internally, or segregated in various ways. In the fruits of the Bizzaria which I have seen, five longitudinal stripes of the colour of a Citron were interpolated in the fruit of the Orange. Other fruits were, on the whole, like Oranges, excepting as regarded an eighth of their mass, which in form, colour, and taste resembled a Citron, and was also peculiar for its extreme convexity. This anomalous segment stretched in the form of a light-coloured cushion from one pole of the spherical fruit to the other. Growers maintain that the Bizzaria is the result of a cross between Citrus medica and Citrus Aurantiurn, though the gardener who, in 1644, in Florence, raised this tree, declared it was a seedling which had been grafted, and after the graft had perished the stock sprouted and produced the Bizzaria (according to which account it would be a graft-hybrid). In other similar cases of Citrus hybrids, how- ever, such as the Bergamot Orange, alleged to be a hybrid of the ordinary Lemon and the Bitter Orange, one finds the characteristics of the parent-species do not 570 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. reappear in juxtaposition (as in the Bizzaria), but are united or fused together. Whether the case of Bergamot Pears, which are striped green and yellow, and that of the half dark- and half light-coloured grapes, of which a few occur occasionally in otherwise ordinary bunches of the fruit, are to be looked upon as parallel phenomena to that of the Bizzaria must remain uncertain until it has been ascer- tained to what particular crosses of the various species of Pyrus and Yitis the innumerable Pear-trees and Vines now cultivated owe their origin. Over and over again gardeners have asserted that hybrids may also be pro- duced by budding and grafting, and in order to distinguish plants so arising from those which are the result of a cross {i.e. from true hybrids), they are called graft- hybrids. One of these plants, a Laburnum named Cytisus Adami, which exhibits a curious mixture of the characteristics of Cytisus Laburnum (the ordinary yellow Laburnum) and Cytisus purpureus in the same individual, has been the subject of lively discussion in scientific circles. It is indeed difficult to imagine anything more curious than a plant of Cytisus Adami. Most of the flowers derive their characters equally from both parent-forms; the calyx is not so thickly clad with silky hairs as in C. Laburnum nor so smooth as in C. pu/rpureus, and the corollas are of a dirty-red colour, compounded of the purple of C. purpureus and the yellow of C. Laburnum. But the curious thing is that on many of the racemes a few blossoms of different appearance are interspersed amongst these red flowers, some having yellow corollas and silky-haired calices as in C. Laburnum,, and others, still more remarkable, having half their petals like C. purpu/reus and half like C. Laburnum, or a third of their petals like C. purpureus and two-thirds like C. Laburnum, or some one of many other combinations. According to Schnittspahns, this anomalous form of Cytisus was first produced at Vitry, near Paris, in the year 1826, by a grower named Adam, who inserted a bud of C. pur- pureus into a stock of C. Laburnum. The shoot which sprang from the bud was not a pure branch of C purpureus, but had characteristics derived both from G. purpureus and from C. Laburnum. Buds for propagating C. Adami were sent from Vitry to gardens all over Europe, and were in some cases inserted into stocks of C. Laburnum, and in other eases into stocks of C. Jacquinianus and C. alpinus. In many cases gardeners grafted buds of C. purpureus in addition to those of G. Adami on to the same stocks, and thus produced shrubs of most extraordinary appearance. Of the branches some resembled C. Laburnum, C. Jacquinianus, or G. alpinus, others Cytisus Adami, and others again C. pmpureus; and amongst the racemes were many which bore the ordinary fiowers of C. Adami, interspersed with a few blossoms of C. Laburnum, and others in whose flowers a mixture of the properties of C. Laburnum and C. purpureus was apparent. The fact of main interest, however, is that cuttings from Adam's original plant (the alleged graft- hybrid of C. Laburnum and C. purpureus) should bear not only flowers of an intermediate type (as might be looked for in a hybrid), but that on certain branches the flowers break back (or revert) to the pure form of one or other of the parents, or that a single flower should exhibit on one half the characters of one parent and GRAFT-HYBRIDS. 571 on the other those of the other parent. Thus the alleged graft-hybrid bears three distinct sorts of flowers, and often parti-coloured combinations of the two parent forms. The anatomical details of the tissues of the Adami-iorms have been examined and compared with those of the two parent-forms by Macfarlane. It appears that the tissues show a remarkable mingling of the two parent-forms. In some the one, in others the other parent-form predominates. Though in the flowers (i.e. the pure Adami-Aowera) the mingling is quite consistent with its being a well-balanced seed-hybrid, in the vegetative regions the strikingly diversified intermixture of tissues is unlike that met with in any seed-hybrid hitherto examined. It should be mentioned that where the ^c^ami-plant bears Labv/rnum or piorpureus shoots and flowers the anatomical characters of these shoots is identical with the normal G. Laburnum and C. purpureus, respectively. Finally the Adami-S.owevB never ripen seeds (the ovules being malformed), though when the parent-forms occur upon it they ripen fruit and seed. As a general rule the relations of the graft to the substratum (or stock) are very different from those manifested in the case of Adam's experiment. The shoot de- veloped from the ingrafted bud makes the same use of the substratum in which it is imbedded as a parasite makes of its host-plant (see vol. i. p. 213). It procures from the substratum a supply of " crude sap ", and this material is absorbed and worked up by the protoplasts of the cells of the graft in the same way as the liquid substances of the soil which are sucked up by roots. It must be premised that those cells of the graft which take up the crude nutrient sap are adapted to their work in very much the same way as are the suction-cells of roots, that is to say, they are able to exercise a selective power, and only admit such substances as are good for the species to which the scion belongs. Any influence that the substratum might have on the graft could scarcely be other than such as would be exercised by soils of various composition. At the most we should expect variations in shape and colour, which have no permanence, and are not retained by the scion's posterity. As a matter of fact, if, for instance, cuttings are taken from an Apricot-tree and grafted on to various other Amygdalese, or are transferred from a Pear-tree to Quinces, White-thorns, and other Pomacese, they do not exhibit the slightest alteration in fruit, flowers, or foliage after entering into organic union with the stock. Again, when hybrid Koses produced by crossing are propagated by budding and grafting, the result is the same whatever species of Wild-rose is taken for the substratum or stock. In all the thousands of cases of propagation by these means none has been observed in which the stock has had any essential influence upon the form of the scion. In 1876 and 1877 certain experiments were made in the Botanic Garden at Innsbruck on the genus Iris. They were suggested by the fact of the production of the hybrids of that genus already referred to, and consisted in grafting buds from the root-stock of one species of Iris on to that of another species of the same genus. The experiment was attended with perfect success, but the shoots and flowers developed from the ingrafted buds showed no trace of any influence on the 572 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. part of the substratum. Buds of Iris Kochii grafted on /. Florentina produced unaltered plants of /. Kochii, and buds of /. Florentina grafted on /. Kochvi developed simply plants of /. Florentina. In the Botanical Garden at Vienna there is a male Ginkgo-tree {Ginkgo biloba) which, more than a hundred years ago, was the subject of an important experiment. When the tree was still quite small the bud of a female tree was grafted upon it by Jacquin, and a lateral branch was developed from this bud. What we have now is a mighty tree with a number of branches bearing male flowers, and one large branch bearing female flowers. The notable thing about the tree is that the grafted branch follows a course of develop- ment which is obviously different from that of the stock. Every year in the spring it puts forth foliage about fourteen days later than the male branches, and in the autumn its leaves are still green long after the rest have turned yellow and, for the most part, fallen off". From this instance we may infer that the shoots developed from the grafted bud adhere with the greatest tenacity even to individual charac- teristics, and do not suffer the substratum to aflect them even in respect of their annual development. These facts have of recent years led many people to the opinion that the genesis of graft-hybrids is simply a gardener's story, and that even the most famous of the supposed graft-hybrids— Cj/feus Adami — does not owe its origin to budding, but to a cross between Gytiaus Laburnum, and G. pv/rpureus. Still, in view of the curious mixing of the parent-characters in Gytisus Adami, as revealed by Mac- farlane's investigations, it would perhaps be well to suspend our judgment. It is true that even in true seed-hybrids {e.g. the Iris hybrids mentioned on p. 568) a mixing (not a fusion) of the parental characters of the flowers was observed. Fresh observations in this field are wanted, directed especially with a view to showing whether or not the sum-total of the characters of Gytisus Adami are absolutely unique amongst hybrid-plants of whatsoever origin. A further instance of the same nature may be mentioned, as it has been the subject of careful scientific investigation and experiment. In 1876 a Jerusalem Artichoke {Hdianthus tuberosus) was grafted upon a Sunflower {Helianthus annuus) in the neighbourhood of Bristol, and it was alleged that as a consequence the Sunflower stock had acquired from the Artichoke the property of producing tubers on its subterranean portions. Quite lately a series of buds of the Artichoke were grafted on Sunflower stalks by Vochting, and the results carefully followed. It was not found that the properties of the one were in any instance transmitted to the other, although scion and stock grew together in perfect harmony. The importance of this subject is so great that I cannot refrain from trespassing for a moment on the domain of Zoology in order to refer to a case which shows that the animal world also sometimes affords instances of the characteristics of both parents being manifested in juxtaposition in their hybrid-offspring instead of being merged together or united in close combination. Tetrao medius is well known to be a hybrid produced by a crossing between the Black grouse {Tetrao tetnx) and the Capercailzie {Tetrao Urogallus). This hybrid is so common in Tyrol that the GRAFT-HYBRIDS. 573 poulterers in Innsbruck receive for sale on an average six specimens every year from huntsmen in the immediate neighbourhood. The plumage of some individual examples of Tetrao medius is curiously striped with alternate groups of feathers inherited from T. tetrix and T. JJrogallus respectively. In 1879 a huntsman brought me from the remotest part of the Gschnitzthal in Tyrol a hen of Tetrao medius whose plumage exhibited a mixture of the feathers of T. tetrix and T. JJrogallus, irregularly distributed in stripes and patches all over the body. The case of this hybrid affords valuable confirmation of the results of the experiments made on Iris hybrids, and there can no longer be any doubt of the fact that there are hybrids generated by crossing in which the parental characters reappear in juxtaposition. In spite of all this, however, I should not like to deny the possibility of the existence of graft-hybrids, for there are certain considerations which tend to a con- trary conclusion. In most cases the relation to the substratum of those cells of the ingrafted shoot or bud which take the crude nutrient sap from the stem of th& stock-plant is just the same as that of a parasite's suckers; they are clearly marked off from the cells of the substratum and are not influenced thereby either in their shape or in their ultimate structure, whilst, conversely, no essential modification is undergone by the substratum through the presence of the graft. There is nothing, however, to exclude the possibility of a fusion between the protoplasmic contents of adjacent cells taking place at the spot where stock and graft unite, and the con- sequent development of a tissue which is composed of cells arising from a division of the cells containing the mixed protoplasms, and which unites the characteristic features of the tissues belonging to the stock and to the graft respectively. In fact, something of the kind has been observed in the case of the parasitic Balanophorese (see vol. i. p. 194). Now supposing such an intermediate tissue were to be formed at the junction between a graft and its substratum, one or more shoots might spring from it and they would doubtless combine the characteristics of the two species employed as stock and scion. In relation to the genesis of new forms of plants in nature, the question of the possibility of the existence of graft-hybrids is of secondary importance ; but it is of no small moment in connection with the comprehension of the processes involved in hybridization; for, the researches suggested by this problem have led to the con- clusion that the marks and attributes of a particular species which are percep- tible to our senses are an outward sign corresponding to the ultimate structure and molecular composition of a specific protoplasm, and that wherever the special characters of two species are united in a single plant-form, that form is built up from protoplasm which owes its origin to a combination of the protoplasms of two parent-species. It is only by adhering to this train of thought that one is able to understand how it is that, also in the matter of chronological development, the vital manifesta- tion connected with the shape, anatomical structure, scent, and colour occupy in hybrids a position intermediate between the corresponding manifestations in the parental species. In the Botanic Garden at Vienna there has been for many years 574 THE GENESIS OF NEW FORMS AS A RESULT OF CROSSING. a Buckthorn-shrub, named Rhamnus hybrida, which sprang from a cross between Rhamnus aVpina and Rhamnus Alatemus. One of the parent-species, R. alpina, has deciduous foliage, i.e. leaves which are green in the summer and wither and drop in the autumn ; the other, R. Alatemus, has evergreen leaves, which last through the winter and remain on the branches for two years. The hybrid, R. hybrida, possesses leaves which do not fall off in the autumn, nor yet last fresh and green for two years, but which maintain their verdure through one winter and fall in the spring when new shoots are sprouting from the buds. The behaviour of hybrids as regards their season of flowering is also very remarkable. From 1863 to 1874 I kept notes concerning the flowering of some flf ty different kinds of Willow, growing in the Botanic Garden at Innsbruck, and each year made an entry of the day on which the first flower opened in each plant, whether a pure species or a hybrid. Earliest Date of Flowering of a number of Willows growing in the Botanic Garden at Innsbruck. (The date given is the average for 12 years.) .Salix Cremsensis March 17 Salix Caprea March 16 Salix daphnoides March 18 „ Mauternesis )J 23 „ Caprea „ 16 „ purpurea April 7 „ attenuata J) 25 „ Caprea - „ 16 „ grandifolia - March 27 „ "Wimmeri » 26 „ daphnoides ,., 18 „ incana April 17 „ Austriaca April 3 „ grandifolia - „ 27 „ purpurea " .1 „ Seringeana » 3 „ Caprea „ 16 „ incana „ 17 „ capnoides )) 5 „ cinerea - April 10 „ incana ,, 17 „ intermedia J) 6 „ grandifolia - March 27 „ incana „ 17 „ rubra - n 6 „ viminalis April 3 „ purpurea - „ 7 „ Kerneri }j 10 „ viminalis „ 3 „ incana „ 17 „ Oenipontana » 12 „ purpurea „ 7 „ incana „ 17 „ auritoides jj 14 „ purpurea „ 7 „ aurita „ 19 „ Fenzliana )) 21 „ retusa » 21 „ glabra - „ 21 „ retusoides )) 21 „ retusa „ 21 „ Jacquiuiana „ 21 „ alpigena - » 23 „ retusa „ 21 „ hastata „ 27 „ excelsior - )j 23 „ fragilis „ 13 „ alba - - „ 27 „ Ehrhartiana )» 29 „ alba - „ 27 „ pentandra - May 6 The name in the first column is that of a hybrid in each case, and the names on the same line in the second and third columns are those of its parent-stocks. The above table, which gives the means of the dates recorded in 12 years of the -first opening of the male flowers in 15 species and 17 hybrids produced from them by a variety of crosses, shows that the hybrids invariably flower on days between those on which the parent-species enter upon that stage of development. It will be observed that the two alpine Willows, Salix retusa and Salix Jacquiuiana, flowered on an average in the 12 years on the same day, and that their hybrid Salix retusoides kept also to that date. We have hitherto dealt with those of the marks, attributes, and vital phenomena manifested by hybrids which are derived partly from the one parent-species and partly from the other, and we must now pass to the consideration of such character- istics as cannot be attributed to inheritance from those species. There is, in the ffrst place, the fact that the majority of the hybrids produced from crosses develop CERTAIN CHARACTEBS OF HYBRIDS. 575 with striking rapidity and exuberance; they not uncommonly flower the very first year after they are sown, whereas the seedlings of the parent-species may not attain to the flowering stage for two or three years; and in respect of the size of the foliage, and still more that of the flowers, hybrids often exceed both parent-species. The latter circumstance is indeed one of the chief reasons why growers make such frequent use of the process of crossing. They are thus enabled to meet the demands of connoisseurs, who prefer to have plants with conspicuous flowers in their gardens. The augmentation in the size of the flowers usually ceases after the first, or at any rate after the second year. Subsequently, the flowers of hybrids become smaller again. On this account gardeners are in the habit of producing such hybrids as are especially valued for their large flowers over and over again by the original method. Of the large number of observations recorded on this subject we will here mention one as an example. Isoloma Decaisneanwm of the order Gesneraceas is the product of a cross between Isoloma Tydoewm and IsoloTna sciadocalyx. The seeds obtained after crossing germinate early, and the seedlings develop rapidly into exceptionally robust plants. The foliage-leaves are three times as large and the flowers twice as large as those of the parent-species; in addition, the flowers are much more numerous than on the parent-plants, and in consequence the hybrid has a much more imposing and showy appearance than either of its progenitors. Many plants which grow on soil rich in humus in mountainous regions, such as the Lungworts (Pulinonaria) and Primulas (Prvmula), do not thrive particularly well in gardens, and certain species die after a short time even when cultivated with, the greatest care. Yet the hybrids of such species flourish wonderfully well under similar circumstances. They blossom luxuriantly, and may be kept in a state of vigorous growth for many years. Examples of this are afforded by Prvmula pubescens and Primula Venzoi. One of the parent-stocks of P. pubescens, viz. Primula hirsuta, can only be reared if the soil used is expressly prepared for itself, and if several other special precautions are taken, whereas the hybrid, P. pubescens, grows exuberantly in ordinary garden-soil. The case of P. Venzoi, the hybrid-offspring of Primula tyrolensis and Primula Wulfeniana, is still more remarkable. Although both the parent species are reared with difficulty, even when the greatest care is bestowed upon their cultivation, Primula Venzoi will flourish with extreme luxuriance if planted close to them in the same soil and under the same external conditions. Another phenomenon sometimes exhibited by hybrids is a change in the dis- tribution of the sexes. It often happens, for instance, that hybrids produce pseudo-hermaphrodite female flowers and pseudo-hermaphrodite male flowers (of. p. 294), even where both parent-species have true hermaphrodite flowers. In Willow hybrids a partial transformation of male into female flowers, and vice versd, has been not infrequently observed, and we then have moncEcious catkins bearing flowers, half of which are female and half male. This change also occurs in true species, but only as a rare exception, whilst in the case of hybrids it is by no means uncommon. 576 THE GENESIS OF NEW SPECIES. Hybrids also exhibit the phenomenon known as the "doubling" of flowers, which depends upon the transformation of stamens into petals, independently of the action of tiny gall-mites, which are the frequent cause of doubling in other plants (c/. p. 548). Several hybrid Roses, Pinks, and Camellias are only known with double flowers. It is diflBcult to explain the fact, repeatedly conflrmed by observation, of the appearance in hybrids of characters which are not present in either parent-species, or rather which cannot be traced to inheritance from either of those species. Thus it sometimes happens that individual plants of a hybrid develop sinuate foliage- leaves with wavy outline, though in both of the parent-species the leaves are either entire or only slightly toothed. The hybrid Salvia sylvestris occasionally exhibits deeply sinuate radical leaves, whilst Salvia nemorosa and Salvia pratensis, the two species to which it owes its origin, never do so. Another instance of the same kind is that of a Stock, the hybrid of Matthiola incana and Matthiola Maderensis. Neither the one nor the other parent-species has sinuate leaves, yet here and there plants of the hybrid display foliage with the margins so deeply cut as to remind one at flrst sight of Matthiola sinuata. Again, in Primula pubescens the leaves are sometimes more deeply sinuate than in either Primula Auricula or Priiriula hirsuta. In hybrids of the Foxglove genus {Digitalis), flowers not infrequently make their appearance wherein the corolla is produced underneath into a spur as in the Toad-flax (Linaria). One hybrid produced by crossing two species of Water Lily, NymphcBa Lotus and Nymphcea dentata, displayed dark violet lines on its sepals which are not to be seen in either parent-species. Reference must also be made to the comparative frequency with which hybrids bearing white flowers spring from species with blue, violet, red, or yellow blossoms whose non-hybrid offspring only produce colourless flowers on very rare occasions. Lastly, we may mention the fact that as from species so also from hybrids varieties may be formed; but they have no permanence amongst the descendants of a race, passing into other varieties whenever it undergoes the restrictive influence of a change in external conditions. 3. THE ORIGIN OF SPECIES. Genesis of New Species.— Derivation of Existing Species.— The Sub-divisions of the Vegetable Kingdom. THE GENESIS OF NEW SPECIES. It is now more than forty years since I discovered, on an island on the Danube not far from the little town of Durenstein, a Willow which had till then remained unknown to Botanists. The plant in question was growing on the island in com- pany with a number of other Willow-trees and Willow-shrubs belonging to the THE GENESIS OP NEW SPECIES. 57T species known as Salix incana and Salix daphnoides respectively, and it was apparently intermediate in form between those two species. The hairs, the system of ramification, the foliage, and the flowers resembled those of S. incana in some respects and those of S. daphnoides in others, and a single glance would have led any unbiassed observer to conjecture that he had to deal with the product of a cross between these two species. This discovery ,1 made in one of the first years of my career as a student of Botany, chanced upon a time when Botanists were beginning to take a particularly keen interest in all cases of intermediate forms observed growing in a state of nature. Some of the leading men at that time refused to believe in the existence of any wild hybrids, and were of opinion that the supposed cases were varieties of species whose presence was to be explained by a tendency in the plant itself to change its form. They also held the view that all plants between which one or more- intermediate forms had been found to exist were to be included in a single species, and, in accordance with this, they not infrequently treated three, four, or more kinds- of plant previously classed as distinct species as being really "varieties" of a single species, because forms obviously intermediate between them, i.e. so-called "transitional forms", had been discovered. This practice was carried so far that several systematic Botanists of that day included in one species 5, 10, and even 15 distinct Hawk- weeds which had been previously described as separate species, the reason for the change being that they were all linked together by transitional forms. Another school of Botanists, on the other hand, recognized in most of the so-called transi- tional forms the results of natural crossing, but they did not deny the existence in plants of a capacity to form varieties in the Linnean sense in response to changes of soil or climate. To my mind even at that time there could be no doubt which of the two oppos- ing theories concerning the genesis, significance, and position of intermediate forms- was to be preferred. The discovery of the hybrid Willow referred to led to my paying particular attention to plant-hybrids, and in the course of the last forty years I have made extensive series of experiments to clear up many obscure points,, and to correct the prejudices which then prevailed. One misconception as to the nature and significance of hybrids, which had great- weight and found expression in the name of " bastard " assigned to them, consisted in the idea that they were contrary to nature. The German word "bastart" is defined by Grimm as a base and useless species. This prejudice was carried so far that Kant positively denied their independent existence, and believed they must neces- sarily die out with the first generation. Connected with this notion was another, according to which hybrids were destitute of the power of producing fertile seeds and propagating their kind sexually. It probably arose from observation of the hybrids of the Mullein genus (Verbascum), which in Central Europe are so common ^The little paper recording the finding of this "Willow, with some additional remarks, by Anton Kerner, was- published in 1852 (Vienna, Zool. Bot. Ver. Verhandl. II., 1852). This seems to have been Kerner's second definite contribution to science; what would appear to be his first is printed in the same publication a few- months previously. Vol. II. 87 578 THE GENESIS OF NEW SPECIES. and conspicuous that they were accepted as the results of crosses between two species even by those amongst the earlier Botanists who were most disinclined to recognize the existence of plant-hybrids. Now, these Mullein hybrids do not for the most part mature any seeds. The pistil itself is usually incompletely developed, and even if one or other of the fruit-capsules does develop, the ovules in it are abortive and infertile. Nevertheless it would be erroneous to say that no Mullein hybrid has ever produced seeds capable of germination. Two such hybrids were artificially generated in my garden, viz.: — Verbascum rubiginosum, by crossing Verbascum Austriacum with the pollen of Verbascum phoeniceum, and Verbascum pseudophceniceum, by crossing Verbascum Blattaria with V. phceniceum. The former of these hybrids, it is true, never produced seeds capable of germination; but in the case of the second, although most of the capsules were empty and -abortive, a few containing fertile seeds ripened; so that even the hybrids of Mulleins are not invariably sterile. Anyone who will look beyond the limited range of that particular genus will be convinced that in respect of their capacity for sexual reproduction hybrids do not •differ essentially from plants which have been admitted to be "true", permanent species by Botanists of all periods. For the case of a few of these true species, as, for instance, Gochlearia ArTnoracia, Grambe tataria, LiliuTu bulbiferwm, Lysimachia Num/mularia, Rubus odorus and R. Nutkaensis, it has long been known that if the stigmas are dusted with pollen from the stamens in the same flowers very few seeds, if any, are set, whilst pollen from other flowers is obviously preferred by them. On the other hand, there are true species whose flowers are pseudo-hermaphrodite, i.e. they have the appearance of being bisexual, but are reaUy unisexual (c/. p. 294). In one individual we find that the ovaries are fully developed, as also the stamens, but that no pollen capable of fertilizing ovules is produced in the anthers; in -another plant the ovaries are imperfectly developed, whilst the anthers are filled with effective pollen. For seeds to be set in such circumstances two individuals at least are requisite, and pollen from a plant bearing pseudo-hermaphrodite male flowers must be transferred to the stigmas of the pseudo-hermaphrodite female flowers. Now hybrids with pseudo-hermaphrodite flowers also exist, and in their •case, as in that of true species, two kinds of individual are requisite to produce aeeds capable of germination. Supposing, however, in such a case that the two kinds of plant necessary for reproduction do not grow close together or do not flower simultaneously, or that one of them is altogether absent — a contingency which must often occur — fertilization cannot be effected, and consequently no seeds can be formed. It is scarcely necessary to amplify the proposition that dioecious hybrids behave in this respect in the same manner as true species, and that pollination and fruit-formation may in them be impeded likewise by dichogamy or by hetero- stylism. In many hybrids, again, as in true species, the relative positions of stamens and pistil, the height of the stigma, the length of the filaments, and other conditions of the kind are not conducive to autogamy, and consequently no trans- ference of pollen from the anthers to the stigmas in the same flower can take place THE FERTILITY OF HYBRIDS. 679 either at the beginning or at the end of the period o£ bloom. Such hybrids are dependent upon foreign pollen, and if none is brought at the right time by wind or insects no pollination or fertilization takes place, and no seeds are developed. Even these brief references to recently-discovered phenomena connected with fertilization are sufficient to indicate that the suppression of the function of fruit- formation in hybrids is due in most cases to the same causes as operate on true ■species. Innumerable experiments have proved that if at the proper time pollen of the right sort is placed upon the stigmas of hybrids fertile seeds are developed as in the case of true species. We must now consider a statement which for long held its place in Botanical works owing to the positive manner in which it was enunciated by an eminent authority. The proposition in question, whilst admitting the fertility of hybrids, asserted that it was conditional on the stigmas being supplied with pollen from one or other of the parent-stocks, and that no fruit was formed as a result of autogamy. This assumption rested partly on certain series of experiments performed on garden- plants by the Botanist Koelreuter in the second half of the eighteenth century.^ By crossing two species of Tobacco-plant (Nicotiana rustica and Nicotiana pani- culata) Koelreuter produced a hybrid which in its characteristics was an exact mean between the two parent-species. The stigmas in the flowers of this hybrid were then dusted with pollen from one of the parents, and the result of this second cross was another hybrid, the characters of which resembled those of the species which supplied the pollen more closely than was the case with the first hybrid. The same treatment was applied to the second hybrid, and thus, after three generations, a plant was evolved completely resembling the male progenitor. The first hybrid had, therefore, in a sense, "reverted" to that particular parent-species. The "rever- sion " of the hybrid to the other parent-species was similarly procured after three generations. Such a result could naturally not have ensued if the action of the pollen of the parental stocks on the hybrid had not been entirely efiectual. It is accordingly quite true that hybrids are fertile when the pollen used to fertilize them is taken from either of the parent-species, but the further assertion that they are sterile if their own pollen is employed for the purpose is incorrect, at any rate as a universal proposition. Koelreuter's own conscientious experiments show con- clusively that it is possible for hybrids to bring fruits to maturity as a result of autogamy, and that, as a matter of fact, the majority do develop such fruits. We may also refer to the large number of plants with ornamental flowers, such as Begonias, Pansies, and Pinks, which every year in our gardens produce seeds auto- ^amously, and are reproduced in great numbers by means of those seeds (see p. 556). Some interesting experiments have also been made on Medicago media, the hybrid of Medicago falcata and M. sativa. This plant, which is, in many places, cultivated on a large scale for fodder, is propagated continuously by seeds which are in a very ^Joseph Gottlieb Koelreuter (1733-1806) was the first to investigate the question of hybridization scientifically and thoroughly. His work, which rants witli the best of modern times, is contained in his Vorlaufige Nachricht von eiiiigeih das Oeschlecht d&r Pfianzen hetreffenden Versudien und Beohachtttngen (published 1761-1766) ; a convenient reprint was issued in 1893. 5 go THE GENESIS OF NEW SPECIES. subordinate degree, if at all, dependent on the parent-species for their production. Nay more, it would be prejudicial were one of those species to supply the pollen seeing that the fertility of this hybrid is diminished thereby. We have here a case of a hybrid manifesting enhanced fertility as compared with one of its parents; tor Medicago falcata is one of those Papilionaceae in which autogamy is attended with very small result. It has been ascertained that in cases where the flowers of that species are thrown upon their own resources for pollen, out of every 30 flowers only two or three fruits containing seeds capable of germination are matured. If flowers of Medicago falcata are crossed with pollen belonging to another species, such as Medicago sativa, a much larger number of fruits is produced, and it is stated that the number of seeds is almost doubled. Medicago media usually sets from six to eight pods in each raceme if the flowers depend upon their own pollen for fertihzation. If, however, pollen is brought to them from Medicago falcata the degree of fertility is strikingly reduced; the flowers so crossed either remain quite empty or develop pods whose seeds are not capable of germination. Medicago media is therefore an instance of a hybrid which is actually injured by being crossed with one of the parent-species, but is successfully reproduced by autogamy. The upshot of all these experiments is that the results of autogamy are no different in hybrids from what they are in species. We may infer from the same experiments in what way reversions of hybrids to their parent-species should be regarded. Botanists possessed with the idea that every hybrid was the product of some process contravening the laws of nature imagined that this anomaly must be counteracted, and that this was effected by a tendency existing in the hybrid's descendants to approximate sometimes to one and sometimes to the other parent-species, so that in the course of a few generations they would completely revert to the form of a true species. As the accounts given by gardeners seemed to confirm the fact of the occurrence of such reversions, no doubt was raised as to the soundness of this view. The reports of gardeners on the subject were, however, founded on inaccurate observation, want of knowledge, and self-deception. In former years the phenomena of pollen-transport in the case of pseudo-hermaphrodite, dichogamous, and heterostyled flowers, and many other things connected therewith, were not appreciated at their full value; to most gardeners they were altogether unknown, and it was only in the rarest instances that any measures were taken to preserve the species and hybrids reared in gardens from extraneous pollen. The majority of growers had no suspicion that the fruit formed on a particular cultivated hybrid might be due to the effect of pollen con- veyed by the wind or by insects from one of the parent-species flourishing in the vicinity, and if they noticed that the seedlings arising from that fruit exhibited characteristics approximating to either of the parent-species they were in the habit of calling the phenomenon a reversion. But if a gardener takes care that the flowers of a hybrid under cultivation are only supplied with pollen developed in flowers of their own kind whilst that of other species is excluded, the plants which arise from the seeds of the hybrid exhibit the characters of that hybrid unaltered. Thus, SPECIES AND RACES. 581 hybrids prove true to seed, to use a gardener's expression, and there is no truth in the assertion that they have an innate tendency to revert to one of the parent- forms. At one time the attempt was made to distinguish two sorts of hybrids — those arising between species, which were regarded as sterile, and those arising between races, which wore regarded as fertile. By "races" are understood forms which, whilst not differentiated by characters of sufficient importance to rank as species, are yet reproduced by seed and transmit their characters to their offspring. They seem to stand midway between what are called varieties and sub-species. Forms arising by the crossing of species were termed hybrids, those arising by the crossing of races " blendlings ". But in this matter Botanists argued in a hopeless circle. Firstly, it was said that if races were crossed the intermediate forms were fertile; whilst those springing from species were sterile; and, secondly, the distinction between races and species was defined as consisting in the fact of the fertility of the intermediate forms produced by crossing races, as compared with the infertility of those derived from crosses between species. A distinction founded on such reasoning as this is, of course, destitute of any value or meaning. What, then, is the difference between races and species ? There are certain forms which have a similar physiognomy, an agreement in certain striking particulars. They are bound together by these common characteristics into a single group, and it must be supposed that they are nearly allied in respect of their origin also. But no more than affinity can be predicated by characters which, though perhaps less striking than the others, are yet trans- mitted unmodified to descendants and prove themselves to be constant attributes. It has been sought to apply the term "races" to nearly akin forms of the kind. But the degree of variation has nothing to do with the conception of a species; the essential point is that the characters wherein the variation is manifested are trans- mitted unchanged to the descendants, and this happens as a fact in all the cases to which the name of race has been affixed. The use of the term would obviously imply quite a different connotation of the name of species from that which Linnaeus, with logical exactitude, attached to it. According to him a species was not an assemblage of individuals of the same form, but an assemblage of individuals of different forms, constituting a group of units and not itself the unit of the system. If, like the French system, we were to distinguish the groups of nearly allied species as " petites espfeces " from those exhibiting more marked differences and less nearly akin to one another, which would be known as "grandes especes ", that would involve quite sufficient recognition of the difference which exists in various degrees between members of the two categories in question; but the introduction of the word "race" side by side with the word "species" suggests the idea of some line of demarcation between the two such as does not in reality exist. Again, if there is no definite boundary between race and species the separation of blendlings from hybrids also fails, and with it the proposition that only those hybrids are fertile which are the offspring of races. In respect of fertility, then, there is no difference between hybrids and species. 582 THE GENESIS OF NEW SPECIES. In the one as in the other we find those floral contrivances for bringing about cross- fertilization in the first place and autogamy in the second, of which an account was given in the first part of this volume; in the one case as in the other cross-fertiliza- tion often takes place as a result of those contrivances, and both categories include forms which are incapable of self-fertilization, and only develop fruits and fertile seeds in consequence of geitonogamy or xenogamy. Seeing that it has been ascer- tained also that, provided the pollen from other species is excluded, hybrids transmit their form unchanged to their posterity, and that the substitution of brood-bodies for fruits as a means of reproduction and the enhancement of the development of those bodies in the event of there being no fruit, are phenomena common to species also, we come to the conclusion that no line of demarcation between hybrids and species exists in respect of the function of propagation. The consideration of all these facts led me years ago to raise the question whether hybrids could originate species,^ and to answer it in the affirmative. Looked at from this point of view, the hybrids which have been and are being produced in nature acquire a special significance, and it becomes important to form a correct notion as to their existence, behaviour, and distribution in localities where the life of plants is untrammelled and undisturbed. Only the vegetation of Europe has been thoroughly studied in this connection, yet this alone afibrds a fund of information, and we may take it for granted that what is true for Europe will apply likewise to the other quarters of the globe. "We shall be rather below than above the mark if we estimate at a thousand the number of wild hybrids belonging to the Flora of Europe which have been brought to light during the last forty years. Of these only a small proportion are of the class of Cryptogams, but this circumstance is due to the fact that it is only lately that Botanists have paid any attention to hybrid Cryptogams. Future researches will no doubt establish the hybrid nature of many so-called "transitional forms". Amongst Mosses in particular, several hybrids arising from species which grow in ditches and marshy places (Hypnum aduncum, H. fluitans, H. lycopodioides, &c.) have been discovered. A few hybrids of the genera OHhotrichum, Grimmia, Fhyscomitrium, and Funaria have also been identified. Fern hybrids are known in the genera Aspidium, Asplenium, Ceterach, Polypodium, and Scolopendrium. Scolopendrium hybridum, which was observed in Istria, is especially remarkable as being the result of a cross between two species possessing widely different forms and included in diflterent genera. The parent-species of this hybrid are. firstly, Scolopendrium officinarum, which is glabrous and grows in clefts in damp, shady rocks and walls; and, secondly, Ceterach officinarum, which has the under surfaces of its fronds thickly covered with brown scales and flourishes in the crevices of dry walls exposed to the sun. Amongst the Horse-tails we may mention Equisetum ^nvMdatwm, a rather common hybrid, which owes its existence to the crossing of Equisetum arvense and E. limosum. In the division of the Coniferse no less than seven hybrids have been recently 1 Oesterreioh. botanisohe Zeitsohrift xxi. p. 34 (1871). NATURAL HYBRIDS. 583. identified, and this fact is of no little significance when taken in connection with the circumstance that Europe only possesses 41 species of Conifers. Juniperu» Kanitzii, the hybrid ofispring of Juniperus communis and /. sabinoides, is a very instructive instance owing to the great diversity in the form of the two parent- species. Comparatively few hybrids occur amongst Grasses. The majority belong to the genus Calamagrostis. Strangely enough, most hybrid grasses have arisen from crosses between species of different genera, as, for instance, Festuca and Lolium, Triticum and Mymus, Triticum and ^gilops. The hybrid derived from ^gilops ovata and Triticum sativum, and known by the name of JSgilops triticoides, and the hybrid ^gilops speltceformis, obtained by crossing ^gilops triticoides with Triticum sativum, have been the subject of lively discussion in their time, and have contri- buted not a little to clearing up our ideas concerning hybrids. As a set-off" to the Grasses, the groups comprising Reeds, Rushes, and Sedges include a comparatively large number of hybrids. For example, in the genus Carex instances have been discovered in the most widely different localities. Amongst Lilifloreae and Iridese only a few wild hybrids have been found, but on the other hand a large number occur amongst Orchidaceae all over Europe. It is striking how many of these Orchid hybrids spring from species which are placed by Botanists in different genera. Hybrids are known, for instance, which are derived respectively from Aceras and Orchis, from Anacamptis and Orchis, from Godo- glossv/m and Orchis, from Gym,nadenia and Orchis, from Himantoglossum and Orchis, from Serapias and Orchis, from Gymnadenia and Nigritella, and from Epipactis and Gephalanthera. The hybrid Epipactis speciosa, lately discovered in the Erlafthal of Lower Austria, is the result of a cross between Epipactis rubigi- nosa and Gephalanthera alba, and is of special interest on account of its manifesting characters strongly resembling those of species indigenous to regions at a great distance from the place where the hybrid occurs, for at first sight Epipactis speciosa might easily be taken for Epipactis gigantea, which is a native of North America, or for the Japanese species named Epipactis Thunbergii. Hybrids are comparatively numerous amongst the Pond-weed group {Potamo- geton). These are aquatic plants which discharge their pollen in the form of clouds of dust, and at the season of pollination raise their flowers above the surface of the water. Owing to their being completely protogynous (see p. 310), autogamy is out of the question. The crossing of pairs of species is especially promoted by the circumstance that the different species flower in definite succession, so that always just at the time that one species is terminating its period of bloom another is coming into flower. Plants which have their flowers in catkins (amentaceous), such as Oaks, Birches, Alders, Poplars, and Willows, produce an uncommonly large number of hybrids. In Willows pollination is effected by insects, in the other genera by the wind. This gives occasion for us to raise, in connection with this group, the question whether hybrids originate more frequently from entomophilous or from anemophilous plants. 584 THE GENESIS OF NEW SPECIES. The fact that many more than a hundred Willow hybrids are known looks as if the agency of insects were the more favourable to hybridization. At the same time we must bear in mind that the number of species of Willow in Northern and Central Europe is very large, whilst the number of different Birches, Alders, and Oaks is small. Taken in relation to the number of species, the hybrids belonging to the plants last mentioned are no less numerous than those of Willows, and it appears, therefore, that hybrids occur as frequently amongst plants whose pollen is dispersed in the form of dust as amongst plants with adhesive pollen which is transported by insects. The prevalence of hybrids amongst Pond-weeds points to the same conclu- sion. A comparison between the Docks and Polygonums even indicates that in anemophilous plants, such as the Docks, hybrids come into existence more readily than in the case of entomophilous plants as represented by the Polygonums, for in scarcely any genus is the number of hybrids so great in proportion to the number of species as in the Dock genus, and the ratio is certainly higher than it is with the Polygonums. As regards the Caryophyllacese it is remarkable that Dianthus has many hybrids and Silene few, although these two genera resemble one another in the distribution of their sexes and in being chiefly visited by lepidopterous insects. In the genus Viola hybrids are extremely common. It has been shown that many Violas which were formerly designated as " transitional forms " are in reality hybrids, and thus the grounds upon which systematic Botanists treated whole series of species as one only are removed. As with the Violas in Central Europe so also with their allies the Cistuses in the South, hybrids are numerous; several kinds of Gistus hybrid occur so commonly that they were described as species by the Botanists of former times. It is noticeable in the Cruciferse that no hybrids are formed in nature between the numerous annual species of this family. There are also only a few hybrids known which are derived from the perennial species. The genera Roripa and Braba are, however, exceptional in this respect. The case of Ranunculacese is similar. In the comprehensive genera Aconitum, Helleborus, and Ranunculus, only a few hybrids have been identified with certainty, whilst in Anemone and Fidsatilla there are almost as many hybrids as species. The case of the hybrid Water-lily Nuphar intermedium will be discussed later on. Many hybrids are known in the genera Tilia, Hypericum,, Malva, Rhammbs, Pistacia, Acer, Ewphorhia, and Epilobium, the last alone including fifty different kinds. This makes it all the more remarkable that so varied a family as the Umbelliferse yields very few hybrids. Of the numerous hybrids -belonging to the Saxifragacese we may mention as specially noteworthy those derived from species which differ exceedingly from one another in form and size. One cannot easily imagine two plants in the limits of a single genus which present a greater contrast to one another in respect of flowers, leaves, and general mode of growth than is exhibited in the following cases: — Saxifraga ccesia and ;Sf. mutata, S. Aizoon and 8. cuneifoUa, S. aizoides and 8. squarrosa, and yet hybrids have sprung from the crossing of these species. NATURAL HTBKIDS. 585 About two hundred hybrids, for the most part fertile, have come into existence in nature within the family of the Kosacese. The genera Oeum, Potentilla, Rubus, Rosa, and Sorbus are inexhaustible in the formation of hybrids. On the other hand, the Papilionaceae, a family allied to the Rosacese, is peculiarly wanting in hybrids. This discrepancy between nearly-related families in respect of the number of their hybrids is again noticeable in the class of plants known as the Monopetalse. The Labiatae, e.g. the genera Ajuqa, Prunella, Galamintha, Lamium, Marrubium, Mentha, Salvia, and Stachys include numbers of hybrids, whilst the Boraginacese have very few. Of the latter only one or two hybrids are known, and these belong to the genera Pulmonaria and Symphytum. The Scrophulariacese and Rhinanthacese have long been noted for the great variety of their hybrids; the genera Pedicularis and Verbascum, in particular exhibit a great wealth of forms, Pedicularis in alpine regions and Verbascum on the shores of the Mediterranean. Verbascum alone includes more than fifty. In the case of Gentianacese, also, the products of a great number of difierent crosses are found in the Alps, the chief parent-species being the long-stalked Oentiana lutea, 0. Pannonica, 0. punctata, and 0. purpurea. Pri- mulaceae, whether growing on the upper or the lower levels of mountains, are famous for the abundance of their hybrids. In the genera Androsace, Primula, and Soldanella the number of hybrids identified by Botanists surpasses the number of the species from which they have sprung. The Ericaceae, although comparatively poorly represented in Europe, exhibit several hybrids derived from the genera Erica, Rhododendron, and Vacci/nium. Hybrids are also specially numerous in the family of the Rubiacese, the genus Galium being the most prolific in this respect. But the greatest number of all is to be found amongst the Compositse. More than two hundred hybrids have been identified in the genera Achillea, Carduus, Centaur ea, Girsium, Hieracium, Inula, and Lappa alone. Of hybrid Compositse the following are particularly worthy of notice: — Erigeron Hulsenii, which is the result of a cross between Erigeron Canadense, an annual species brought into Europe from America, and Erigeron acer, a perennial species indigenous to Europe, and, secondly, the Cudweed hybrids, Filago mixta, F. neglecta, F. subspicata, &c., which spring from crosses between annual species, and are, therefore, exceptional, for annual species in general rarely produce hybrids. In many cases only a few individual instances of these natural hybrids have been observed; in some, indeed, one single specimen alone has as yet been discovered. The majority, however, grow by the hundred or by the thousand in the localities proper to them, and many are represented by immense numbers of plants, and possess a wide area of distribution. Salvia betonicifolia, a hybrid derived from Salvia nem,orosa and S. nutans, occurs as commonly as its progenitors in places on the grassland in the central parts of Siebenbilrgen (Hungary); Marrubium remotum, a hybrid of Marrubiwm peregrinum and M. vulgare, is to be found everywhere on the plains of South-Eastern Europe, especially in the flat country by the Theiss and 586 THE GENESIS OF NEW SPECIES. the lower Danube; Roripa anceps, the hybrid of J.oripa amphibia and B. sylvestris, is met with all over the district forming the basin of the Baltic; Primula digenea, a hybrid derived from Primula vulgaris and P. elatior, grows in its thousands in the upland meadows of the Eastern Alps; Betula alpestris, produced by the crossing of Betula alba and B. nana, is met with in great abundance in the Jura, in Scandi- navia, and in the North of Russia, and here and there entire copses are composed of it; Nigritella suaveolens, a hybrid produced from OyTnnadenia conopsea and Nigritella nigra, is so common in the Central Alps, for instance, in the Pusterthal of Tyrol, that hundreds of specimens are sometimes encountered in a single meadow; Primula Salisburgensis, the hybrid of Pritnula glutinosa and P. mimima, is repre- sented by a host of plants on the Tyrolese Alps, as, for instance, on the Muttenjoch and the neighbouring mountains which separate the Gschnitzthal from the Obem- bergthal. Several hundreds of other cases of the kind might be mentioned, but the limits of this book will not allow me to do more than give the following names selected from the long list that lies before me : Hybrid. Parents. Asplenium Germanicwm Asplenium RutSi. nemorosa or S. pratensis, and have been found by experiment to be fertile in a proportion of more than 60 per cent. Salvia sylvestris has therefore scattered itself all over this dry meadow-land, and manifests all the characteristics essential to our conception of a species. A third example is Nv/phar intermedium, a hybrid derived from Nuphar luteum and Nuphar pumilum. This plant grows in lakes in the Black Forest and in the Vosges. It is also scattered over North Germany, and occurs with increasing frequency in Central and Northern Russia and in Sweden. It has been found as far north as Lapland. At the northern extremity of this large area of distribution Nv/phar intermedium is more abundant than the species from which it is derived; indeed in many places it occurs in their absence, and in fact passes beyond the northern limits of their area of distribution. In these situations there is, of course, no possibility of the hybrid's crossing with either of the parent- species or of the formation of goneoclinic hybrids. Nuphar intermedium subsists independently there, multiplies without change of form, and has in fact established itself as a species. This phenomenon is explained as follows: The northern limit of 590 THE GENESIS OF NEW SPECIES. the distribution of each of the three Water-lilies in question is determined by their not being able to ripen fruits beyond that limit. Of the three, Niiphar luteuTn flowers the latest, and therefore its fruits are also the latest to rip -^n, whence it follows that it is the first to fall behind; that is to say, it reaches the northern limit of distribution sooner than the others, and ceases to ripen fruit in regions where the others are still able to do so. But Nuphar pumilum and If. intermedium are also different from one another in this respect. In Norbotten and Lapland Nuphar intermedium ripens its fruits a little earlier than N. pumilum, and it is conse- quently able to extend rather further north than N. pumilum,. The further north plants go, the shorter becomes the time allotted for the performance of their annual work; and those which ripen their fruits early have a great advantage over those which ripen later. Concerning Nuphar intermedium,, it has also been ascertained that the individual plants produced in nature are more fruitful than those reared in gardens from artificial crosses. In the case of plants obtained in this manner in the Botanic Gardens at Konigsberg each capsule contained from 15 to 18 fertile seeds, whilst capsules ripened in the small lakes of the Black Forest contained from 38 to 63, and others taken from plants growing in Lapland con- tained from 41 to 72 such seeds. From these data we may infer, in the first place, that N. intermedium is most prolific in situations beyond the range of the parent- species; and, secondly, that it would be wrong to suppose that because a hybrid may be comparatively infertile or actually sterile in a particular locality, such infertility is a characteristic of the plant wherever it may occur. As may be gathered from the above account of these three examples, the advan- tage which a hybrid may possess over the parent-species, whereby it is enabled to subsist and multiply side by side with those species, is not always of the same kind. In one case it is the more vivid coloration of the flowers, in another the fact of the hybrid being better adapted to a particular state of the ground, whilst in the third the earlier ripening of the fruits, which enables the hybrid to stand a more rigorous climate, gives the requisite advantage. These do not, of course, exhaust, by a long way, the possible sources of superiority, and there are many instances of hybrids which thrive better than the parent-species when the climate becomes milder, moister, or drier, as the case may be. It is obvious that of all the different advantages which may come into play those connected with climatic con- ditions are the most important, and the genesis of hybrids is probably most fre- quently due to the operation of this kind of advantage. Far too little significance has been attached to the fact that the greater number of hybrids are not found in districts where the parent-species grow together with equal luxuriance, but occur where one or other of those species is meagrely repre- sented, owing to the climate not being favourable to its distribution. Again large numbers of hybrids are found in parts where the boundaries of several species coincide. In Europe such regions exist in the strips of land where the advance- posts of the Floras of the Baltic and Black Sea, and the Floras of the Baltic and the Mediterranean, respectively, encounter one another, and particularly in the THE DISTRIBUTION OF HYBRIDS. 591 lower limits of the Alpine Flora. Another characteristic of these borderlands is the fact that the separate hybrids growing on them are almost always represented by a large number of individuals, and the consequence is that there is no great chance of their crossing with the parent-species and gradually reverting, in succes- sive generations, to those species. If isolated individuals belonging to a particular hybrid grow in the midst of thousands of plants of the parent-stocks, their stigmas will, in all probability, be dusted with pollen from the latter species. This proba- bility diminishes, however, as the number of individuals of the parent- species flowering in the same locality as the hybrid diminishes; if that number is small the hybrid is thrown mainly upon its own resources for fertilization, and, provided it is self -fertile, there is nothing to prevent its multiplying and becoming dispersed. . Connected with the above is the further fact that in the neighbourhood of a hybrid which has become a species there is sometimes no trace of one of the parent- species, it having completely died out. The slightest change in climatic conditions may cause the plants of a particular stock to succumb at the confines of the stock's area of distribution, where they are only present in small numbers, and are anything but strong; and when this happens we find the other parent-species alone growing side by side with the hybrid, and even that species is possibly not so well adapted as the hybrid to the altered conditions. Of cases in point we will select two from the East of Europe and two from the West. When Epilobium alsinefolium, and Epilobium palustre are crossed a hybrid is obtained which, when fertilized with its own pollen, yields a large quantity of fertile seeds. The plants reared from these seeds exhibit the same characteristics as the plant from which the seeds were taken. This hybrid grows together with the parent-species in the Riesengebirge, and has received the name of Epilobium scaturiginum. It is likewise a native of the Bihar Gebirge, on the confines of Hungary and Transylvania, and is of very common occurrence in the springs and rivulets in the vicinity of the Hochkamm (a mountain of this chain). Yet, of the two parent-species, only one, viz. Epilobium palustre, grows amongst these moun- tains. Again, Prunella hybrida is a hybrid springing from Prunella laciniata and Prunella vulgaris. It is widely distributed in the Wienerwald district, and in some places is commoner than the parent-species, whilst in Moravia and Bohemia it occurs in places where one of the latter, viz. Prunella laciniata, is entirely absent. A third instance is afibrded by Primula brevistyla, called also Primula variabilis, a hybrid derived from Primula vulgaris and P. officinalis. This plant is true to seed, and is met with everywhere throughout almost the whole of Europe in company with both parent-species. In some districts of France it is found also in places where one or other of the latter species does not grow at all, and even where both are absent. Our fourth example, Linaria stricta, is the hybrid of Linaria striata and Linaria vulgaris. It occurs in many places in the West of Europe, together with its progenitors, but in the South of France, in the neighbourhood of Montpellier, it is found growing with Linaria striata alone, whilst the other parent- stock, L. vulgaris, is never found in the district. 592 THE GENESIS OF NEW SPECIES. We shall have another opportunity of describing the way in which the lines of demarcation of the ranges of entire floras become displaced in consequence of the changes which the climate of a region is liable to undergo in course of time. These displacements of floral regions are, as a rule, the result of very slow and inconspicu- ous migrations on the part of the plants constituting the floras in question. The direction of migration is invariably towards the places whose climatic conditions agree best with the organization of the plants, and is, in the case of any one species, either an advance or a retreat, according to the nature of the circumstances which impel the species to migrate. The difierent plants of a flora do not all migrate in a host together. Some species abandon their former home entirely and establish themselves in a new locality more or less remote from it; others leave a few of their kind behind in the old settlement at isolated spots which happen to be in peculiarly favourable situations, and many succumb to the effects of the new con- ditions or to the hardships incidental to the migration, and so die out. These changes in the range of floras are naturally accompanied by all sorts of alterations in the social relationships of the plants concerned especially with regard to the co-existence of hybrids and their progenitors. It may happen that one or both parent-species are left behind, whilst the hybrid advances, or the hybrid may remain behind, whilst one of the parent-species advances; or, again, one of the parent-stocks or both may die out. The facts concerning these local displacements explain the phenomenon that species which, from their characteristics, may be looked upon as hybrids of two other species, oocupy in each case a district which is separated, and often at a considerable distance, from the areas inhabited by the species supposed to be their progenitors. The characteristics of the kind of Sorrel named Rumex Patientia lead one to the conclusion that it is a hybrid derived from Rumex aquaticus and Rumex crispus. It is found, however, growing wild in Hungary and in Bosnia in parts where neither Rumex aquaticus nor R. crispus occurs at all. In Herzegovina there grows fairly commonly a Micromeria which has been named by one of my friends Micromeria Eerneri. So far as its charac- teristics are concerned it must be considered to be a hybrid of Micromeria grceca and Micromeria Juliana; yet neither of these two species grows in Herzegovina at the present time, and they are not met with at any nearer spot than the part of Dalmatia which stretches westward from Herzegovina, and belongs to the area of distribution of the Mediterranean flora. In the little upland valleys of Planail and Plawen, which run down from the mountains of the Oetzthal into the valley of the Adige, there grows a Pulsatilla named Putsatilla nutans. If it occurred in company with Pulsatilla vulgaris and Pulsatilla montana, all Botanists would be unanimous m lookmg upon it as the product of a cross between those two species. Yet Pulsa- tilla vulgaris and P. montana do not grow in the high valleys in question but are flrst met with at a distance of many miles from them, the former in the Unterinnthal and the latter m the Vintschgau (a portion of the Adige valley) Inasmuch as the last-mentioned cases have to do with processes which have taken place long ago they partly belong to the next chapter, where the genesis of THE GENESIS OF NEW SPECIES. 69S species in the past will be discussed. They here bring to a natural conclusion a. series of examples adduced to show in what manner a genesis of new species may ensue in the present, and may have taken place in the past. No sharp line of demarcation is to be found between different epochs in this connection any more than in the case of any of the other phenomena which, in the aggregate, constitute the history of species. Now that it has been shown how new species arise from hybrids, or, in other words, from the crossing of species in pairs, the question presents itself whether, in addition to this one method, there are not also others leading to the same result. In answering this question we must bear in mind that every permanent change in external form which is inherited by a plant's descendants must be preceded by a change in the constitution of the protoplasm, and that so far as investigation has elicited the facts, the centre of the change is located exclusively in a particular protoplast which lies hidden in the ovary and there receives the sperm atoplasm. The stimulus which causes the change in this protoplast can only proceed from the spermatoplasm, and every speculation concerning the formation of new species must therefore be associated with the question whether in the intercrossing of plants of one species and in autogamy the protoplasm in the course of its journey to the ooplasm may, as a result of its exposure to new external conditions, undergo modi- fications of so fundamental a kind that its influence on the ooplasm is subject to- corresponding variations. In the first place, it might be imagined that the pollinated stigmas do not always act in the same way upon the spermatoplasm of the pollen- cell. Reference has already been made to the fact that a stigma may sometimes be- almost simultaneously dusted with the pollen of very different plants (see p. 404), but that it has the power of exercising a selection, and that in every case only one kind of pollen is induced to put forth tubes by which a real fertilization is accom- plished. The other kinds of pollen upon the stigma are not known to have a direct effect upon the ovule. But that there is some interaction between them and the protoplasm in the cells of the stigma is evidenced by the fact that they swell up wherever they are in contact, and (as has been shown, p. 414) are frequently found developing pollen-tubes. Now it is possible that the reciprocal action of the contents of these pollen-cells and the contents of the stigmatic cells may produce some change in the latter, which is transmitted to the contents of those other pollen-tubes which are to enter into combination with the ooplasm. Such modification might conceiv- ably affect the nature of the stimulus imparted to the ooplasm, and this alteration in the stimulus might be manifested in a change in the form of the individual arising from the fertilized ooplasm. The likelihood of all these possibilities and assumptions being satisfied is extremely small, but as no researches have yet been instituted into the matter, it cannot be dismissed with an unconditional negative. In artificial crosses between different species of Cirsium it has often been noticed that pollen-cells taken from a single capitulum vary in their effects upon the stigmas- of a second capitulum, inasmuch as the seeds produced by the different florets,, though all fertilized with the same kind of pollen, yield dissimilar plants wherk Vol. II. 88 594 THE GENESIS OF NEW SPECIES. they germinate. The variation is limited, it is true, to the different degrees in which the seedHngs resemble one or other of the parent-plants. If the cross is between two plants of the same species no such variation can occur, seeing that the plants crossed are alike in form. But there is still the question whether differences in the age, size, and luxuriance of growth of the individuals which cross may not have some influence on the result. So far as my experiments show, these differences have no effect on the genesis of new forms, and have no prospect of becoming per- manent characters in the offspring. A poor stunted plant growing on dry soil may produce seeds which, on being planted in a good moisf soil give rise, under favour- able conditions, to well-developed plants capable of flowering luxuriantly. As is well known, the first flowers of an inflorescence are always much larger than those which subsequently open at the apices of the spike or raceme, or on the ultimate ramifications of the cyme as the case may be. Now, if the large earliest flowers are crossed one with another, and likewise the small latest flowers, and the seeds so ■obtained in each case are kept separate but reared under similar conditions, the plants produced from them do not differ in the slightest degree from one another, but in their turn bear flowers, of which the first are the largest and the last the smallest. Notwithstanding these results, however, I should not like, without further investigation, to deny the possibility of the specific constitution of the spermatoplasm undergoing some change as a result of external influences in the course of its development, whether during its imprisonment in anthers or antheridia or on its way to the ooplasm, or to say such change might not cause the descendants of the plants concerned to differ in form from the individual from which they sprang. It has been established beyond all doubt that modifications of form directly induced by conditions of soil or climate are not hereditary, and that every change of form which persists in the descendants is only brought about as the result of a process of fertilization, or, in other words, that new species can only arise through fertilization. Herein lies. also the solution of the marvellous phenomenon known as the alternation of generations, and of the question why plants in general flower and undergo fertilization. To these processes is due the genesis of new species. The propagation of plants, their multiplication and dispersal, may also be effected by means of brood-bodies, and as a matter of fact these processes are continuously operating on a vast scale. But the plants reproduced by brood-bodies retain the form of their ancestors unaltered, and no new forms arise in this way. Suppose that a locality is occupied exclusively by plants which multiply by brood-bodies only and do not change their form, and that in consequence of a change in the climate such species as are not adapted to the new conditions abandon their homes, or else languish and die out, the probability is that many of the vacated spots will remain unoccupied owing to there being no recruits in the neighbourhood, or from out its confines, that are better adapted to the new conditions. If, on the other hand, the area in question is inhabited by plants which reproduce sexually and which, by crossing one with another, produce descendants of diverse forms, there is DERIVATION OF EXISTING SPECIES. 595 every probability that amongst the assemblage of new forms some will be better adapted to the new conditions when a change of climate occurs than those of the old species which are driven out thereby, and that these new forms will therefore be able to take the place of the latter. It is only from this standpoint that we can properly understand the phenomena of the alternation of generations, the separation of the sexes, dichogamy, and all the rest of the wonderful floral contrivances, the object of which is to facilitate the crossing of two species during the first stage of flowering and only to allow of crossing between plants of one species, or of geitonogamy, autogamy, or cleisto- gamy in the event of no inter-specific crossing taking place. As a result of these contrivances, numberless new forms are continually being generated which are respectively adapted to all the most various conditions of soil and climate. So long as no change in climatic conditions takes place, the majority of these forms have very little chance of surviving and of naturalizing themselves as species amongst the plants already established in the same locality. But when, in conse- quence of a change of climate, the ranks of the species in possession of the ground are thinned through *the abdication of many of those best adapted to the condi- tions of life previously existing, the real significance of the new forms which have arisen as a result of the sexual process is manifested in the fact of those which are best adapted to the new conditions taking possession of the spots vacated and settling down there as new species. DERIVATION OF EXISTING SPECIES. The plants preserved as fossils in former ages are not only the forerunners but the ancestors of the existing vegetation of to-day. There was no general rejuvenescence and extinction of organisms coincident with the beginning and end of the several "periods" of the history of the earth. The changes in the organic world, like those in the inorganic crust of the earth, were accomplished gradually by slow degrees, and the organisms of the present day are a continuation of, and have been slowly evolved from, those of former ages. So far, there is little diflerence of opinion amongst naturalists; but as to the causes of the differences in form between the vegetation of the present and the past, the most various theories are held. Nor is this surprising, seeing how largely our conclusions are based on conjectures. And when the flood-gates of speculation are rolled back it is not always that the proven is clearly distinguished from the unproven. An import is attached to isolated facts which they do not merit, and — most mischievous of all — the existence of wide lacunse in our knowledge is con- cealed, or these lacunas are dexterously bridged over with unmeaning, high-sounding words and hollow phrases which, while astonishing us for the moment, leave us chastened and confounded. The confirmed mistrust aroused by these extravagances which obtains concerning all that bears on the derivation of species demands that we should devote a brief consideration to the prevailing theories, and especially to 596 DERIVATION OF EXISTING SPECIES. such as bear upon the conversion of species of former times into those of the exist- ing vegetation. A change in the conditions of life has, according to a widely-spread view, been the immediate cause of a change in the vegetation. The altered conditions of life provoke new wants in the plant, and these new requirements have led to a trans- formation of their organs. Stimulated by use, the organs in question become enlarged and further developed; others, no longer of service, become smaller, atrophy, and disappear. It is the cumulative result of these small and almost imperceptible changes that in course of time becomes apparent. These structural changes are transmitted to the progeny, and with an increasing tenacity, the greater the number of generations which have been exposed to the altered condi- tions. This, the theory of adaptations, has provoked wide discussion and criticism. It is urged against it that, whether wild or cultivated plants be considered, it is only isolated or a few individuals, never the whole of the members of a species, which exhibit these variations and transmit them to their offspring. If these new characters are immediately due to the soil or climate, then all the individuals of a species, exposed to like conditions of growth (environment), should exhibit them and hand them on to their offspring. The permanence of the influence — and to this many naturalists and others attach great importance — is without significance in this matter. When a change is called forth — be it by an altered source of nourishment, by the influence of heat or cold, light or darkness, moisture or dry- ness — it must become apparent upon the growing plant, since a change in the plant stands to a change in the environment as effect to cause. If the cause ceases, so also does the effect, equally after the lapse of a year or a hundred years. But a much more potent criticism of the theory of adaptation is the result of a series of experiments which were carried out for the solution of these questions. From them we see that an altered environment calls forth certain changes in the plants submitted to it, but that these are not transmitted to the offspring, are not heredi- tary, and that the influences of soil and climate do not provoke a fundamental change in the constitution of the protoplasm. Influences of this sort can induce a diseased condition in a plant and can even kill it, but they cannot bring about a change which can be transmitted to the next generation. Though soil and climate play a most important part in the struggle of species and varieties for existence, and though the environment has a great influence on the origin of varieties and on the distribution and migration of plants — as the immediate stimulus to the origin of new and transmissible characters, and thus to the modification of species, change of environment is without significance. Another theory deaUng with the origin and modification of species is that known as the theory of progressive transformation by inherent forces. According to it, the impulse to change resides in the inherent tendency of all species to perfect themselves. This theory transcends all experience and depends on premises and draws conclusions essentially metaphysical in nature; it deals only in part with the results of scientific observation. It presupposes a creation of living VARIOUS THEORIES. _ 597 protoplasts endowed with the capacity to alter their constitution on their own initiative; and, further, that these alterations take place along predetermined lines in a direction leading from a lower to a higher platform; consequently the imperfect organism necessarily, in course of time, passes over into a highly developed, perfect one. Against this theory the following may be urged: The first assumption involves creation. The question is: Is it possible for a living protoplast to be formed from inorganic matter without the co-operation of already existing living beings? The question obviously concerns the present and future as well as the past, for what has happened once may again take place, for the forces of nature, according to the laws of the conservation of matter and energy, remain the same for all time. The discussion of this question resolves itself into this: whether a little bit of protoplasm can arise from inorganic matter, and after its origin can acquire the capacity of growing by the absorption of food from its environment, &c.: in a word, whether it can exhibit those changes and movements which we term life. When first organic compounds (formic acid, urea, sugar, &c.) were synthesized in chemical laboratories from inorganic substances like ammonia, carbonic acid, and water — compounds which formerly had only been produced as a result of the activity of living protoplasm — naturalists began to think that these things might take place in nature independent of already existing plants. It seemed possible that these substances might, under the uncontrolled forces of nature, unite and arrange themselves in the same manner as occurs within a vegetable cell. The tendency of matter to combine, which plays so important a part in nature, was pointed out, and especially the similarity between the structure of crystals and that of certain cells; the properties of finely-divided soil also were called to mind, how it absorbed gases, took up water in varying quantities, altered salt-solutions, separating certain of their constituents, and what was especially noteworthy, increased the capacity of many simple substances to combine. This was at a time when chief importance was attached to the chemical properties of protoplasm; it was thought that, once given the substance, it would form itself into cells like crystals. Of the ultimate structure of protoplasm and of the nucleus knowledge was as yet very incomplete. The tendency of that time was to explain all those phenomena which constitute life as , the resultant of the various forces which form inorganic bodies, and to deny the existence of any wide gulf between the living and non-living world. The experiments to produce living matter had all of them negative results. But this of course is no proof of its impossibility; for it can always be urged that wrong methods were followed, and improper conditions imposed. Nor, on the other hand, does it follow from the fact that hitherto living matter has never been known to originate independently of existing organisms, that its production is impossible. Since we cannot arrive at definite results by experiment, the in- vestigator must depend on other considerations. The second assumption of the theory of transformation from internal causes, that plants have the inherent capacity to modify their internal constitution and. 598 DERIVATION OF EXISTING SPECIES. similarly, their external form spontaneously, has been so fully met by the obser- vations recorded in the last chapter that it is hardly necessary to deal with it now at great length. I shall content myself with pointing out that it is impossible to give a natural explanation of such a phenomenon. Every variation presupposes a corresponding disturbance; for the acquirement of any new structural character the plan of construction must undergo some fundamental alteration. The naturalist is unable to grapple with the phrases "internal causes", "internal force", "force of transformation ", " tendency to differentiate ", " principle of progressive trans- formation", when attempting to explain variation in a natural manner upon mechanical principles. Nor is the likening of this transformation to the meta- morphosis which every individual passes through at various periods of its existence at all to the point, since metamorphosis repeats itself with great constancy in every species according to the plan of construction which is laid down in the specific constitution of the protoplasm. That the protoplasm of any species should, in the absence of any impulse or stimulus from outside, be able spontaneously to alter its plan of construction contradicts all our experience of the normal action of natural forces. Even should we conceive vital force, the dormant energy of the protoplasm, to be converted into an active form, it could only give rise to movements which have their origin in the specific constitution of the protoplasm. And now we come to the assumption that this inherent force of transformation is a progressive one, that it leads to a higher or more perfect development. But what is to be regarded as a higher development amongst plants? A tree with its brightly coloured flowers and luscious fruits seems more highly developed to the non-botanist than a low herb with inconspicuous flowers, or than the green filaments of a Spirogyra destitute of flowers. The supporters of the theory under discussion assert that the highest development is that which exhibits the greatest complexity of form, and in which division of labour is carried furthest. And in this assertion they do not essentially differ from the popular view. Complexity of form and division of labour are undoubtedly carried further in an Apple-tree than in the Spirogyra of the ponds and ditches. But it must not be forgotten that the differentiation of a plant-body into various tissues, the production of wood, bast, and cork in its stem, of cuticle, stomates, and hairs on its leaves, of various colouring-matters and aromatic substances in its petals, and of sweet juices in its fruits, stands in "harmonious relation to the environment of the plant m question. Change the conditions, and imagine the Apple-tree submerged in a pond; it is no longer in harmony with its surroundings, its complexity of tissues, its wood, stomates, &c., are not so well adapted for these conditions as are the Spiro- gyras and Water-weeds equipped with organs of another type. The size of a plant is often— in the popular estimate— the indication of its high organization. A big plant gives the impression of possessing a more perfect development than a small one. But this criterion leads to no satisfactory result; it is sufficient to instance the case of certain huge sea- weeds (Macrocystis) of the southern seas, which exceed our greatest forest trees in height. Many Thallophytes, only visible under VAEIOUS THEORIES. 69& the microscope, show a greater complexity of structure of their constituent cells than do many Flowering Plants; and, should especial importance be attached to this character. Diatoms and Desmids must be regarded as more highly organized than many small annual Composites. The idea of progressive development implies a recognition of that species of plant which is most highly developed and which stands upon the apex of the pyramid, or, at any rate, of the group of plants which has already reached the furthest point— is it the Aristolochiaceee, Cannaceae, Mag- noliacese, the Orchids, the Composites, the Ranunculacese, the Papilionaceee, or the Pomegranates? Any one who has studied carefully the structure of these plants knows well that it is impossible to make an estimate of this kind. In a book of Botany one group must be treated first and another last, but this does not necessarily imply that the last is the most highly developed; indeed the various writers of systematic works begin and end with the most various groups. Like the theory of adaptability, that of progressive transformations from inherent forces fails to give us a reasonable explanation of the variations which plants have undergone in process of time. A third theory, based on the observations of modem times, is as follows: That variations of form in the offspring arise through crossing, from the union of two dissimilar protoplasts. This theory, based on the union of unlike forms, has been fully sketched out in the last chapter. It assumes the existence in former times of a vegetation rich in forms — an assumption amply justified by the fossil remains which have been preserved. New forms arose, not by a progressive development such as has been alluded to, but by a transformation or metamorphosis of those already in existence. It was from the union of existent types that incipient new species were produced. By the periodic recurrence of changes in climatic con- ditions the areas of plant-distribution have received continual displacements, and it was then that these incipient species or varieties were put to the test. Those well- suited to the fresh conditions settled down into new species. They replaced their less well-adapted ancestors in the plant-community, and they played the same part as these had formerly done. A change indeed is brought about; but not (on the lines of the theory of adaptability) as a direct result of climatic influences, nor from an inherent tendency to progressive development. It arises rather from a change in the specific constitution of the protoplasm in consequence of the crossing of unlike forms. In basing the transformation of species on a crossing of this nature we are relieved the necessity of picturing lacunee in a vegetation as a result of climatic changes, or of any serious disturbance of the inter-relations of its various component forms. Bacteria and Moulds, Mosses and Lichens, Ferns, Grasses, Palms, and Coniferous Trees, have all of them a special function to fulfil in the great community of plants, and they are to a certain degree dependent on one another. Were one removed the whole would be affected, and it might well happen, did a given group come to speedy extinction, that the whole community of plants might suffer. But in every group at all times and in all places a reserve of new forms continually arises by crossing, so that this danger is averted. With climatic €00 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. changes, of the older, less fit forms some are extinguished, whilst young, new- forms step into their places. Thus we see also that the conversion of Mosses into Ferns, of Ferns into Conifers, and of Grasses into Pinks, &c., as assumed by the theory of progressive transformation, would be a positive disadvantage to plants as a community, and that its tendency would be in the direction of anything but real progress. It is important to recognize the fact that in the production of new forms by crossing, it is not especially such forms as are constituted to resist an anticipated change of climate that are produced. Of the forms which arise, some are fitted for a more inhospitable, others for a milder climate; but it cannot be said of any that they possess an assured future. Such only are able to maintain, propagate, and establish themselves, as are from their internal organization and external form in harmony with the prevailing climatic conditions of the moment. Those so constituted that they are unable to thrive under the given external conditions linger and become extinct; they are outstripped and overgrown by such as find the environment to their liking. Hence we speak of the struggle for existence. Plants in harmony with their surroundings are the victors, and they establish themselves upon the arena of this encounter. This, briefly, is Darwin's theory of Natural Selection, a theory which marks an advance upon all other theories of the origin of new species. Though many views may be held as to the precise manner of origin and transformation of forms, there can be no difference of opinion as to the significance of the struggle for existence and of the survival in this struggle of those forms best fitted by their organization to the circumstances of the environ- ment. THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. The fact that the savants of ancient times made no attempt to classify plants according to their structural characters is explained by their limited botanical knowledge. Their interest was restricted to such plants as were in use as drugs, poisons, and charms, to vegetables, fruits, and cereals, finally, to such as were of value for decorative purposes and as symbols of religious observances. Nor was the number of these plants considerable. Some five hundred forms were known to Theophrastus (300 B.C.), whilst Pliny (23-79 a.d.) records about twice that number. The characters of these few plants could be retained in the memory for the purposes of comparative investigations, and their recognition depended in large part upon the general impression gained in the ordinary intercourse with nature. Enumerations of plants were based far more on their medicinal or economic uses, on their hurtfulness and beneficence for mankind, than on any structural characters they might possess in common. Even in the herbals of the sixteenth century, containing, as they did, new descriptions and incomparable woodcuts, were the medicinal and economic properties of the various plants still especially emphasized; Botany was still almost exclusively the handmaiden of medicine and agriculture. OLDER SYSTEMS OF CLASSIFICATION. 601 The first botanical writer to break with these old traditions was Clusius (1526-1609); he described plants as he observed them, quite apart from their value to man. Clusius, though a Belgian, spent many years of his life at Vienna, and thoroughly explored the Flora of Austro-Hungary; previously he had investi- gated the plants of Spain and Portugal. To England he paid more than one visit, and received many exotic plants from Sir Francis Drake, the voyager. In his Bariorum Plantarum Historia, published originally in 1576, we find the first attempt to classify plants according to their similar characters. In separate books he deals with trees, shrubs, and under-shrubs, bulbous plants, sweet-smelling flowers, scentless flowers, poisonous, narcotic, and acrid plants, with plants having a milky juice, and with Umbellifers, Ferns, Grasses, Leguminosse, and certain Crypto- gamie plants. In those days some 4000 plants were distinguished by Botanists, and the want of some system of classification was gradually felt. The groups of Clusius and his contemporaries were inadequate, and the system of Cesalpino (1519-1603), published in the first book of his Be Plantis Libri XVI (1583), failed to obtain the recognition it undoubtedly deserved — perhaps because it was only sketched out in outline and lacked a full and detailed rendering. Cesalpino was the first to convert observation into real scientific research; he drew attention to the more hidden organs of plants, to the position of the seeds, the number and mode of insertion of the cotyledons, &c., to the presence or absence of flowers. It is to Tournefort (1656-1708), a Frenchman, that we owe the first complete review of known plants in synoptical form. In his Institutiones Rei HerbaricB (published 1700) 10,146 species of plants are distinguished and arranged in 698 genera, which again are assembled under 22 classes. Classes 1-15 include herbs and under-shrubs, 16 and 17 flowerless plants (Cryptogams), and 18-22 shrubs and trees. The herbs, shrubs, and trees are distinguished by the form of their flowers, especial importance being attached to the presence of calyx and corolla, to the regularity or irregularity of the flower, and to the petals— whether they are free or united with one another. Not long afterwards Linnseus produced a classification of plants based on the distribution of the sexes, and especially upon the number of the stamens in the flowers. The terms species and variety, genus and class, were more clearly and intelligibly defined than heretofore, and his 1050 genera were included under the 24 classes already enumerated (p. 288). The Linnean classification, known as the Sexual System, enjoyed an unprecedented recognition. It constituted a well-arranged summary of a great mass of scattered observations, and made it possible for species to be identified by means of concise descriptions. It was not the fault of this accomplished and renowned naturalist if a greater importance were attached to his system than he himself ever intended. Linnaeus never regarded these 24 classes as real and natural branches of the vegetable kingdom, and expressly says so; it was constructed for convenience of reference and identification of species. A real natural system, founded on the true affinities of plants as indicated by their structural characters, he regarded 602 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. as the highest aim of botanical endeavour. He never completed a natural system, leaving only a fragment (published 1738). The credit of actually founding a natural system of plants is usually attributed to Bernard de Jussieu (1699-1777) and his nephew Antoine Laurent de Jussieu (1748-1836). For many years this system only found expression in the laying out of the beds in the Botanic Garden of Trianon (at Versailles); it first became generally known some thirty years after its inception, when the younger de Jussieu published his Genera Plantarum (1789). A hundred families of plants are distinguished and grouped under fifteen classes, which, in their turn, fall under three main co-ordinated divisions (Acotyledones, Monocotyledones, Dicoty- ledones). The three main divisions^ are founded upon the structure of the embryo at germination. In the Acotyledones the embryo consists of but a single cell and is destitute of cotyledons, in the Monocotyledones it is multicellular and provided with one cotyledon, whilst in Dicotyledones there are two cotyledons. The Acotyledones are equivalent to the Cryptogamia of Linnaeus (his 24th class, cf. p. 290) and constitute the 1st class of the new system. The Monocotyledones fall into three classes according to the relative position of the stamens to the ovary (Monohypogynae, Monoperigynae, Monoepigynee). The Dicotyledones are first subdivided into three groups according to the structure of the perianth, viz., into those destitute of petals (Apetalse); those with distinct calyx and corolla, the petals being united (Monopetalae); and those also having calyx and corolla, with all the petals free from one another (Polypetalse). Each of these groups is subdivided into three classes, based on the relative position of stamens to ovary (in the case of the Monopetalse of corolla to ovary). Since in the Dicotyledones with unisexual flowers it was impossible to indicate the relative position of stamens and ovaries, a special class (Diclines irregulares) was set aside for them. The institution of this last class does not mark an advance towards a natural system; whilst the limitations of the other classes in respect of the relative positions of stamens to ovary is cumbrous and unnatural, still they are less artificial than those of the Linnean Sexual System. The distinctive features of the system of de Jussieu are the broad characters upon which the families are based — the whole structure of the plant being taken into consideration — and especially the recogni- tion of Monocotyledons and Dicotyledons as equivalent groups of Flowering lA. L. DE JUSSIBU'S SYSTEM OP 1789. Acotyledones j ' ("Stamina hypogyna II Monocotyledones J perigyna .,..,..... HI i epigyna !..."!,!!!!.!!!!!jv.' ' C Stamina epigyna V Apetalse J perigyna vf L „ ^yp°sy"=' '.^.'.'.'....\'.'Z'^^.'.''.''^.'yvn. jboroUa hypogyna VIII. n,Vntvl»^„n»<, /Monopetate J perigyna IX. Dicotyledones <^ i- ^ g_i„_ ( Antheris oonnatls X. V rsj •( distinctis XI. T, , ,, C Stamina epigyna ' XH. Polypetalse J hypogyna XHI T,. ,. • , '^ perigyna .'.'..'.'.'.'.'.'.'.'..'. XIV.' Diclmes irregulares XV OLDER SYSTEMS OF CLASSIFICATION. 603 Plants. A. p. De Candolle^ (1778-1841) in his Theorie EUmentaire de la Botanique, ou Exposition des Principes de la Classification naturelle (published 1813), dis- tinguished between cellular and vascular plants (Oellulares and Vasculares). The former are constructed of cells alone, whilst in the latter vessels also are met with. The cellular plants were divided into those without leaves (Oellu- lares aphyllae) and those provided with leaves (Oellulares foliacese). The vascular plants were divided according to anatomical views current at the time into those in which the vascular bundles were scattered through the stem and were supposed to originate from within (Endogense), and into those in which the vascular bundles were arranged in a ring and were added to from without (Exogense). The group Endogense included the Vascular Cryptogams (Endogense cryptogamse), forms destitute of flowers, and the Monocotyledons of de Jussieu (Endogenae phanerogamse). The Exogense, the equivalent of de Jussieu's Dicoty- ledones, were divided into those with a simple perianth (Monochlamydese), and those with a distinct calyx and corolla (Diplochlamydese). The latter are further subdivided into three groups: the OoroUiflorse, in which the petals are united into a continuous corolla; the Calyciflorse, in which the petals are inserted upon the calyx; and the Thalamiflorse, in which the petals are free and inserted upon the floral receptacle. Although De CandoUe based his system upon characters essentially difierent from those used by de Jussieu, and although in both systems there are many deviations in the limitations of the classes and families, there is on the whole an agreement in many essential particulars. Especially may we note the recognition of Monocotyledons and Dicotyledons (though under different names) as the two contrasting main divisions of Flowering Plants. And further, that the Cellular and Vascular Cryptogams are sharply distinguished from one another. The main groups, the Cellular and Vascular Cryptogams, the Mono- cotyledons and Dicotyledons, are met with (under various names) from this time onwards in all later schemes of classification; and, so far as we can tell, appear to constitute so many natural groups — groups, that is, of which the members are all more nearly allied by descent to one another than to the members of the other groups. Following De CandoUe many Botanists elaborated schemes of classification during the first half of the nineteenth century; these included Reichenbach, Oken, Agardh, Martins, Brongniart, Bartling, Endlicher, Lindley, and many others. To the non-botanist, recognizing the fact that there can be but one real natural system 1 A. P. DE CANDOLLB'S SYSTEM. I. VASCULAR OR COTYLBDONOUS PLANTS. 1. ExoGENa) OR Dicotyledons. A. Perianth double (calyx and corolla). ThalamiflorsB (petals distinct, inserted on the receptacle). Calyoiflorae (petals free and inserted on the oalyz). CoroUiSorffi (petals united together). B. Monochlamydese (perianth simple). VASCULAR OR COTYLBDONOUS PLANTS (continued). 2. ENDBGBNiE OB MONOCOTYLEDONS. A. Phanerogams (=true Monocotyledons). B. Cryptogams (= Vascular Cryptogams and II. CELLULAR OR ACOTYLBDONOUS PLANTS. A. Foliaoeas (leaf y= Mosses and Liverworts). B. Aphyllse (not having leaves=Thallophytes.) 604 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. of plants, this great variety of specialist opinion is somewhat surprising, and tends to shake his confidence in all botanical systems. But it must be remembered that in the development of a natural system the imagination plays a much more im- portant part than in the elaboration of an artificial one, nor can prevailing currents of thought, or the particular habit of mind of the observer, be without their in- fluence. Very prominently does this appear in the case of the Botanists who came under the influence of what has been termed nature-philosophy during the early portion of this century. Thus Reichenbach and Oken proposed systems which can only appear to us absurd; but it would be wearisome and useless to follow their absurdities in detail. The system of classification proposed by Endlicher^ (1805-1849), and published in his Genera Plantarum secundum or dines Naturales disposita (published 1836- 1840) is based on the systems of de Jussieu and De CandoUe. In it 6838 genera, arranged in 277 families or orders, are included. Here for the first time are the Coniferse and Gnetaeese distinguished as a special group, and designated as Gymno- sperms. Here also is that group of cellular plants known as the Thallophyta care- fully distinguished into three series, the Algae, Lichens, and Fungi. But we still find these groups treated as equivalent to the Horse-tails, Ferns, Lycopods, &c.; so also with the Gymnosperms, they are not treated as a distinct subdivision, but only as a class of Dicotyledons (Acramphibrya), the other classes of which are the Apetala, Gamopetala ( = Monopetalse), and Dialypetala ( = Polypetalse). In fairly recent times Bentham and Hooker,^ in their well-known Genera Plant- arum (published 1862-1883), follow essentially the systems of De CandoUe and Endlicher. They assemble all flowering plants (100,220 species) into 8417 genera, and these under 210 families or orders. Of these orders 3 belong to the Gymno- sperms, 35 to the Monocotyledons, 36 to the Monochlamydese, 46 to the Gamopetalae, and 90 to the Polypetalje. 1 ENDLICHER'S SYSTEM. I. THALLOPHYTA (no opposition of stem and root). II, GORUOVEYTA—Continwd. Peotophtta. Alg^. Lichenes. Hysteeophyta. Fungi. II. CORMOPHYTA (opposition of stem and root). AOKOBKYA (stem growing at the point only). Anophyta (Liverworts and Mosses). Protophyta (Vascular Cryptogams and Cycads). Hysterophyta (certain parasites, Bala- nophoreee, RafflesiacesE, &c.). Amphibeya (stem growing at the circum- ference). Includes Monocotyledons. AcEAMPHlBRYA (stem growing at both point and circumference). Gymnosperma (oTuIes naked, fertilized directly from the micropyle). Apetala (Perianth 0, rudim. or simple). Gamopetala (Perianth double, petals united). Dialypetala (Perianth double, petals free). 2 THE MAIN DIVISIONS OF BENTHAM AND HOOKER'S SYSTEM (Foe Flowering Plants only). DlCOTYLEDONES. DlOOTYLEDONE8-Co»««™d. Polypetalse. Monochlamydese, GymnospermsB, MONOOOTYLEDONES. CLASSIFICATION OF THALLOPHYTES. 605 Though the grouping of Dicotyledons (according to the characters of the perianth) into Monochlamydese ( = Apetalse), Gamopetalae ( = Monopetalse), and Poly- petalae ( = Dialypetal^), is very generally recognized to be an unnatural one, it is no easy task to replace it by a better one. The families belonging to the great group Dicotyledons show the most multifarious relations to one another. An arrangement, based on the assumption that these families have been developed one from another, is not discoverable; whilst an arrangement in linear series is as unnatural as one resembling a tree with its branches. Very appropriate was Linnasus's comparison of the limitations of these families with the dovetailing of the frontiers of countries on a map. One family stands in touch with two, another with three, others again with four or more allied families. This contact or rela- tionship occurs on the most various sides. Some families are extremely large and comprehensive, others relatively small, and, as it were, jammed in between them; whilst others resemble scattered islands off the coast of a continent. Well worthy of consideration is the system of Alexander Braun,i published in 1864 in Ascherson's Flora der Provinz Brandenburg. Though the division of Dicotyledons into Apetalee, Sympetalse, and Eleutheropetalse suggests at first sight the classifications of de Jussieu and Endlicher, there is a diiFerence, and an important one. A large number of the families included by the older Botanists in the Apetalee are here placed in the Eleutheropetalae. With the Eleutheropetalse are ranked those plants " in which calyx and corolla are typically present, the latter consisting of separate petals ". They are ranged in 24 Alliances or Cohorts — Hydropeltidinse, Polycarpicae, Rhceadinse, Parietales, Passiflorinse, Guttiferse, Lamprophyllee, Hes- perides, Frangulinse, .^sculinse, Terebinthinse, Gruinales, Columniferse, Urticinse, Iricoccse, Caryophyllinse, Saxifraginse, Juliflorse, Umbelliflorse, Myrtiflorse, Thy- melseinse, Santalinse, Rosiflorse, Leguminosse. In recent times we have the systems of Eichler and Engler. They follow the lines laid down by Alexander Braun, but, carrying his method further, the group Apetalse (or Monochlamydeae) is entirely abolished, its members being referred in part to the Sympetalae, and in part to the Eleutheropetalse ( = Archichlamydese of Engler). During the last fifty years our knowledge of the Cryptogams constituting the group Thallophyta (founded by Endlicher, cf. foot-note, p. 604) has increased by leaps and bounds. Several attempts have been made to bring together the results of the various researches upon this group, and to utilize them for classificatory 1 ALEXANDER BRAUN'S SYSTEM. I. BRYOPHYTA. 1. Thallodea (Algse, Lichens, and Fungi). 2. Thallophyllodea (Charaoese, Mosses, and Liverworts). II. CORMOPHYTA. 1. Phyllopterides (Perns and Equisetums). 2. Masohalopterides (Lyoopods). 3. Hydropterides (Rhizocarps). III. ANTHOPHYTA (Flowering Plants). A. Gtmnospeem^ (seeds exposed). 1. Frondosse (Cycads). 2. Acerosae (Conifers). B. Angiospeem^ (seeds in an ovary). 1. Monocotyledones. 2. Dicotyledones. a. Apetalee. 6. Sympetalee. c. Eleutheropetalse. 606 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. purposes. The old division into Algae, Fungi, and Lichens, based on the presence or absence of chlorophyll and on the mode of life of the forms in question, has been by many authors abandoned. Cohn in 1872 divided the Thallophytes, according to their methods of reproduction, into seven groups: Schizospore^, Zygosporese, Basidio- sporesB, AscosporesB, Tetrasporese, Zoosporese, and Oosporeae. Sachs, in 1874, follow- ing somewhat similar lines, made four classes: Protophyta, Zygosporese, Oosporese, and CarposporesB, in each of which groups both chlorophyll-containing (Algae) and colourless forms (Fungi) occur, as may be seen by a perusal of the classification quoted below.^ Goebel (1882) returns in part to the older method and distinguishes between Algse and Fungi; but with these as groups of equal systematic importance he ranks the Myxomycetes, Diatomacese, and Schizophyta. The Algse he divides into Chlorophycese (Green Algse), Phseophycese (Brown Algse), and Ehodophycese (Red Sea-weeds); and the Fungi into Chytridiacese, Ustilaginese, Phycomycetes, Ascomycetes, .^cidiomycetes, Basidiomycetes. Warming (1884) returns completely to the older method, dividing the Thallophytes into Algse and Fungi, ranking the Myxomycetes with the Fungi, and the Diatoms and Schizophycese with the Algse. The balance of opinion at the present time, largely swayed by the views and researches of Brefeld upon the Fungi, favours a grouping of the bulk of Thallo- phytes into Algse and Fungi. Brefeld regards the various families of Fungi as more intimately related amongst themselves than are these families to corresponding families of Algse. That Fungi have arisen from Algse at some remote period, and have then amongst themselves undergone development along various lines, is very generally held; but the view that the difierent families of Fungi stand in near relationship to the several algal groups — as indicated, for instance, in the system of Sachs, {of. foot-note below) — is not at present the prevalent one. And amongst the Algse, also, the attempt to classify the various forms into families according to the relative simplicity or complexity of their organs of reproduction (as Sachs suggested) no longer finds general favour. Amongst the Algse we find a number of extensive 1 SACHS'S CLASSIFICATION OF THALLOPHYTES. I. Protophyta. Containing Cklorophyll. Ifot containing Chlorophyll. Cyanophycese. SoMzomycetes. PalmellaoesB (in part). Saccharomycetes. II. Ztgospoke^. Conjugating Cells Motile. Pandorinese Myxomycetes. (Hydrodiotyese). Conjugating Cells Stationary. ConjugatsB. 1 Zygomycetes. III. OOSPOEE^, Sphseroplea. Vauoheria (CloUaste,,), f Saprolegniea,. „, . (. Peronosporese. VolvocinesB. (EdogonesB. Fuooide^. IV. Cabpospoke^f, Coleochteteaa. Ascomycetes (including Lichens) FloridesB. .ffioidiomycetes (Uredinese). Characeae. Basidiomycetes. THE EVOLUTION OF PLANTS. 607 groups, within the limits of each of which all stages of complexity in reproduction occur. The attempt to string together forms agreeing in reproductive methods makes it necessary to break up groups which on general grounds seem to be natural families. And as it is a natural system that we are striving after, systems like that of Sachs (which may be compared to the artificial sexual system of Linnseus) must be abandoned. That the publication of the Sachsian system in his widely-read "Text-book" has done great service to Botany there can be no doubt; it has stimu- lated thought and observation, and has led more speedily than would otherwise have been the case to the establishment of broad and probably sound views as to the relations of the Thallophytes. However, the Myxomycetes, approaching as they do certain groups of the animal kingdom, are kept apart from the rest of the Thallophytes in the most recent system. The classification of plants according to their similarity of structure — species into genera, genera into families or orders, families into alliances or cohorts, these into classes, and classes into two chief branches or phyla, the Cryptogams and Phanerogams — leads to the presumption that these two chief branches have arisen from a common stock, have diverged from a common stem. A consideration of all animal and plant forms similarly leads us to the belief that the main stems of the Animal and Vegetable Kingdoms, respectively, meet at their points of origin. By studying systems of classification drawn up on paper and restricted to two dimensions of space, we involuntarily conceive the classes and orders of the vegetable kingdom, as a tree which continually branches, finally ending in thousands of twigs which represent the various species. Such is, rightly or wrongly, the conception of all Botanists who have concerned themselves with the construction of a natural system. They only difier in so far that some regard the Thallophytes as standing at the base, and derive from these the Liverworts and Mosses, from these the Ferns, &c., and so on to the Gymnosperms and Angiosperms; whilst others make a subdivision of the main trunk at once into Cryptogams and Phanerogams, each of these continually branching according to the various classes and families. Others again, whilst conceiving the whole vegetable kingdom as having a common origin, regard this as the centre of a sphere, and that the several phyla and classes radiate out from this, producing numerous branches and twigs at the surface of the sphere. Each of these hypotheses presupposes, in the first instance, the existence (or spon- taneous generation) of a few Thallophytes of extremely simple structure which have become differentiated, i.e. given rise to more complex offspring which form the beginnings of the branches of the tree. To this kind of development of a tree- structure, the terms Phylogenesis or Phylogeny (from (pv\ii, a tribe; and yei^^da, to produce) is given. Obviously, not only the original forms possess the capacity of differentiating, but their offspring also, and so on through the entire tree. But views are divided ac to whether this continued differentiation follows a predeter- mined plan, is due to definite inherent forces, or whether it may not be restricted in this sense and due to other and external causes. 608 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. That a natural system has been evolved along lines resembling the ramifications of a tree, which commenced with simply organized structures and terminated with the most complex ones, finds a certain confirmation when the history of develop- ment of the individual (ontogeny, from fflv, 6vtos, being; and yepi'da, to produce) is compared with its position upon the phylogenetic tree. As we know, the greatest of all flowering trees begins its existence as a single protoplasmic mass. This surrounds itself with a cell- wall and increases in size and complexity at the expense of nutriment derived from its environment. Gradually cells and tissues arise and the young organism becomes segmented into an axis and appendages. These, again, assume the most varied forms corresponding to a subdivision of labour. Ultimately the uppermost and last-produced members of the plant are transformed into flowers and fruits. It is thought that, just as a plant is gradually differentiated in this way so have all plants undergone a similar transformation, step by step. The egg-cell, the starting-point of the individual, may be compared to a Myxomycete, the cell- complex which arises from the egg-cell after fertilization to a Thallophyte, the segmented axis and appendages of the seedling to a Vascular Cryptogam, and the complete plant-body, finally, to a Flowering Plant. This comparison, like so many others which captivated the imaginative faculties in the days of the speculations of nature-philosophy, has found many adherents; indeed, it has served as dogma and guiding light in many investigations. But it is difficult to harmonize it with other theories well-proved by experience. The main burden of this comparison (known as the " Recapitulation-theory ") is that the vegetable kingdom as a whole has undergone a developmental history and transformation resembling that of a single member of the group of Flowering Plants. But first it must be asked, what is the meaning of metamorphosis in the individual, and what object has been attained by it? Though the actual processes taking place in the living protoplasm in metamorphosis are unknown, this much seems certain: That these changes occur along lines sufficiently well indicated; that the fashioning of the successive stages of any given species is accomplished according to a definite plan; that external influences, such as soil and climate, do not permanently affect this plan; and that consequently the plan of construction of these successively appearing stages is laid down in the protoplasm itself. The ultimate object of metamorphosis in plants is the production of fruit; and with the formation of an ovary the metamorphosis ends, the fertilized egg-cell beginning the metamorphosis all over again, i.e. it is the starting-point of a series of trans- formations along identical lines. And this applies equally to the Apple-tree and to the Palm, to the Pine, the Horse-tail, the Moss, the Bladder-wrack, Stonewort, to the Mould, and to the simplest of green Algse. Only in the last-named forms are the intermediate stages fewer than in the first-named. But it will hardly be sug- gested that the latter have not on this account attained the end in view. Simple plants whose fertilization and fruit-formation is accomplished under water, whether it be on inundated land, in the mud of a river, or at the bottom of a lake, attain this end without a complex metamorphosis of petals and stamens; whilst many THE EVOLUTION OF PLANTS. 60& of the denizens of the ocean have no need of segmentation into stem and leaf. Indeed, such a metamorphosis for this purpose would be a disadvantage, anything but a progressive development. Similarly is it inconceivable, from all we know of the relations between external conditions and the form of an organism, that a Fern (for instance), unable to accomplish its fertilization in dew or rain-water, should in consequence depart from its usual habit and strike out a new line of metamorphosis. Thus we may conclude ^ that the development of the individual (Ontogeny) cannot be regarded as an epitome of the ancestral history or line of descent of that individual, and, further, that ontogeny gives no support to the assumption of a ramifying phylogenetic tree starting with simple forms and ending with complex much-differentiated ones. The results of developmental investigations showing a marked similarity in the form of organs serving similar purposes in the most different groups of plants have been brought forward in support of the assumption that Flowering Plants have arisen from simple Cellular Plants by a series of progressive transformations. Though these organs are in some groups of more simple, in others of more complex structure, their similarity is unmistakable; it is upon this that the view is widely based that organisms exhibiting similar organs have been derived from one another. But this inference is inadmissible. The similarity in question finds a simpler ex- planation as the expression of the attainment of a common object. Thus fertiliza- tion consists in the coming together and uniting of two portions of protoplasm which have originated at a distance from one another; the similarity of ways and means in attaining this object are obvious. In one case water is employed as a means, in another the air. There is, truly, a difference in detail, but the general similarity remains. From this general similarity all we are justified in inferring is that the organisms in question all reproduce themselves by fertilization, not that they have a common origin. This conclusion leads to the question whether, in view of the diversity of the organs of fertilization, several distinct stems of plants may not have co-existed all along. We know from observation and experiment that new forms do not as a rule arise from offshoots, but from fruits. New groups of plants might thus (so it was said) arise from existing ones solely by the sexual method. Complex Thallophytes might arise from the fruits of simpler ones. Ferns from the fruits of Mosses, and so on. Assumptions of this kind belong to a period at which the phenomena of fertilization and fruit-production, especially in the Cellular Plants, were only very imperfectly understood. No Botanist nowadays would suggest the possibility of a Spirogyra or a Ulothrix, or an (Edogonium, or a Stonewort, arising from the fruit of a Vaucheria. It might well happen that a new species could arise by the crossing of two dichogamous species of Vaucheria, but this new form would be essentially a Vaucheria, and its methods of fertilization would be essen- tially similar to that of its parent-forms. It is likewise as impossible for the male protoplasm of a Vaucheria to enter the chamber in which the conjugation of the ^ See editorial note prefacing this volume. VOi. II.' 89 610 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. sexual cells of Spirogyra takes place (c/. vol. i. Plate I. fig. 1), or for this spermato- plasm to fuse with the gametes of Ulothrix (cf. p. 49), as for the gametes of Ulothrix to enter and fuse with the egg-cell of a Vaucheria; no new group of plants could have arisen in this way. We may conclude then that plants belonging to groups with a marked diversity in their sexual characters have not arisen from one another, but belong to stems which have co-existed as distinct types from the first. Nor must we omit to notice the observations which have been made in regard to symbiosis and the inter-relations of green and non-green plants. It has already been pointed out (vol. i. pp. 263, 264) how that the continuation of plant-life, and indeed all life, is dependent on the activity which saprophytic plants exhibit in the decomposition of dead organisms. Green plants could not exist independent of colourless saprophytes, nor these latter without green plants. This must ever have been the case, and it must so remain. We may then draw another conclusion, viz. that those colour- less plants which by their activity bring about the decomposi- tion of green plants cannot have arisen from green plants, but that from the beginning they have belonged to a distinct stem. It is to Palaeontology that we must look for the most trustworthy solution of the question as to whether numerous plant-stems have existed side by side from the first, or whether the groups which at present co-exist have in process of time branched forth from a single stem. Were it a fact that those forms which show a far-reaching division of labour, and a complex structure of organs, which we term " higher plants ", have arisen from others of very simple mode of life and possessing a simple structure and which are known as " lower plants ", then should we expect the earth to have been covered formerly by lower plants alone. And then, fol- lowing this epoch, would have come a time when plants would have existed which might have served as the starting-points of the later-appearing distinct groups. W^e should expect to find in those strata of the earth's crust regarded by geologists as the oldest of all nothing but the remains of very simple Thallophytes, then, following these. Wracks, Red Sea-weeds, and Lichens, and after these Stoneworts {Chara), Mosses, or some other type of plant which, having given rise respectively to Stoneworts and Mosses, has, after this differentiation, become extinct. From the graphite, which is looked upon as the oldest trace of vegetable life on the earth, unfortunately we obtain no conclusive evidence on this matter. From its presence on slate mountains together with crystalline limestone and pyrites rig. 365.— Spirophyton from the Upper Devonian. THE VEGETATION OF FORMER TIMES. 611 we might conclude that it originated from plants adhering to the limestone reefs formed by animals or from sea-plants which lived on the borders of these limestone cliffs. Where graphite is found in greatest quantity one is tempted to think it might have been derived from peat moors. As we have said, all these are merely suppositions, for since the carbon, lime, and silicates have become crystalline all the points for the determination of the families to which the graphite-forming plants belonged are lacking. It might be noted here, by the way, that although graphite does indeed furnish the oldest traces of plant-life on the earth this does not prove that the plants which gave rise to it were necessarily the iirst which Fig. 366.— Biella helicophylla growing under water. Enlarged. existed there. It is doubtful whether the rock which is associated with graphite formed the first hard crust of the earth. Much more probably this rock was com- posed of other broken rocks just as it has itself been again demolished, furnishing the material for new strata. The shapes of vegetable remains from palaeozoic formations are fairly easily recognizable. Those which were formerly regarded as fossil Sea-wracks have indeed been more recently interpreted as the trails of worms and medusee, but some of them are without doubt the remains of Sea-wracks. The only other known lowly plant which at that time had an aquatic habit is the curious Spirophyton, the so-called Cock's-tail Alga (see accompanying fig. 365). This, though some regard it as of purely inorganic origin, may perhaps be regarded as a submerged Liverwort; at any rate it is not without resemblance to Riella Reuteri, which at the present time lives in the Lake of Geneva, and to the Algerian Riella helicophylla (see fig. 366). No trace is to be found of Thallo- 612 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. phytes which may have existed on land, but we have huge tree-like Vascular Cryptogams with trunks, fronds, and leaves which are to be placed side by side with our present-day Bquisetums, Ferns, and Lycopods. Cycads and Conifers also are not absent from the Coal Measures. No Angiospermous flowering plants have hitherto been demonstrated in these strata, but it would be foolish to regard this fact as a proof that neither Thallophytes nor Angiosperms flourished at this period. That which has been preserved to us from this time certainly forms but a small fraction of the old vegetation, and is restricted apparently to the flora of peat-moors which were just as poor in species and just as monotonous as they are to-day. The plants which at the present time predominate on the moors are still the Equisetums, Ferns, Lycopods, and Conifers, and, in tropical regions, the Cycads; only a few species from each group, but standing in thousands side by side and aggregated into dense communities. Anyone who has worked out the history of these moors knows that the soil must have been prepared for these plants by other growths. Equisetum limosum, Aspidium Thelypteris, Lycopodium inundatum, &c., do not flourish in soil poor in humus; in order to obtain their requisite food and to develop they require soil which is saturated with the dead remains of earlier settlers. Experience tells us that the plants which appear as the first inhabitants belong to widely difierent groups (see vol. i. p. 268). Now if we hold to the view that the formation of peat-moors in long- past ages occurred just as in the present day, we must assume that the colonies of Equisetums, Ferns, Lycopods, and Cycads were preceded by other plants which, as the first settlers, prepared the soil. We cannot indeed determine from the surviving remains to which groups these first settlers belonged; but, looking back on the history of our present peat-moors, it seems not improbable that among them were both Thallophytes and Angiospermous flowering plants. The fact that the fossil remains of Equisetums, Lycopods, and Cycads, which spread so widely over the peat-moors of palaeozoic times, have reached us in such good condition is explained by the presence of humus-acids, which are formed universally in the peat (see vol. i. p. 263). There are four conditions which render it possible for a plant to be preserved as a fossil: humus-acids form the first; the second is the resin which exudes from the pine-wood and forms amber; the third is mud and sand brought by floods; and the fourtb the silicification and calcination of the cell-wall or the formation of a lime incrustation which is precipitated from calcareous water on to the various parts of the plant. It is certain that these four conditions have always been efiective, but it is doubtful whether all the fossils formed in the fourth manner at all periods have remained. For many older strata have long been destroyed and used in the building up of younger layers, and many risings and sinkings of these strata have taken place. It would indeed be difficult to find a single place on the earth's surface which has not been repeatedly above and under the sea. Much that might lead us to definite conclusions at present lies inaccessible to us, covered with immense masses of water at the bottom of the sea, and the view has actually been suggested from studies made on the few accessible THE VEGETATION OF FORMER TIMES. 613 and closely investigated spots on the earth's surface that the fossil remains found there are not more than a minute fragment of the vegetation of periods long elapsed. With these remarks we might mention that it is not beyond the range of possi- bility that, in addition to the Vascular Cryptogams, Cycads, and Cdnifers growing on peat-moors, plants of other habitats, especially those of fresh and salt water, or perhaps of sand-dunes and river-banks, might be found in the strata of paleozoic times. But no one would doubt that among these would be Angiospermous Phane- rogams, and this throws some light on plant remains which have come to us from the mesozoic period. For example, in the upper layers of the chalk we find, in addition to the plants of peat-moors, the inhabitants of a luxuriant forest-flora of Angiospermous flowering plants. There are Planes, Birches, Beeches, Oaks, Poplars, Willows, Fig and Laurel trees. Maples, Ivy and other Araliacese, Bread-fruit trees, Tulip-trees and Magnolias, Cherry-trees, and Leguminosse of the division Csesal- pinese. Palms, Bushes, and Grasses. If we do not believe in the theory that these Angiosperms were first created in the mesozoic period, and still less in the greater marvel that they have sprung from the Vascular Cryptogams, Cycads, and Conifers, we are forced to the conclusion that they too must have existed as far back as the palaeozoic time. It is to be specially noted that not the slightest trace of inter- mediate or transitional forms which might connect the aforesaid Angiospermic Phanerogams with the Gymnosperms or with the Vascular Cryptogams has been found. One leaf is immediately recognized as belonging to a Tulip-tree, a second to Maple, a third to a Fig-tree, a fourth to a Palm, &c., but no plant has been discovered anywhere which would perhaps form a connecting link between the Palms or Figs and the Conifers or Vascular Cryptogams. Even a cursory glance at the plant-forms named shows that they were members of mixed forests. It may be assumed, however, that other plant communities peopled the earth at the same time as these forests. The rocky terraces and boulders, as well as the flat dry land, were certainly not destitute of vegetation. Nor is it surprising that no fossil remains of the inhabitants of these places have remained. The under-shrubs and herbs of a dry soil decompose immediately after their death, and leave behind only formless humus, which mixes with the soil. Just as little fossil remains will reach posterity of the Lichens and Mosses, Pinks and Composites, Saxifrages and succulent plants which inhabit the rocks on the dry mountain- slopes at the present day, as of the Tulips and Irises, Umbelliferse and Saltworts of the steppe-flora; and a great mistake would be made if, millions of years afterwards, it were reasoned from the lack of fossil remains of these, plants that they could not have existed in our time. It would be just as wrong for us to argue from the absence of such plants in the strata of earlier periods that they had never existed in those times. The same thing applies to most fresh-water and marine Algse, and to the numberless saprophytes which eifect the destruction of dead animal and vegetable- bodies above and under water, and thus maintain the eternal cycle of life as a whole. Of the first-mentioned the only fossil remains which can be recognized 614 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. are those of Diatoms, whose cell-wall is transformed into an imperishable siliceous frustule, together with those Florideae which provide themselves like corals with a calcareous skeleton, and some tough Sea- wracks. It is, however, a very significant fact that the innumerable fossil Diatoms which come to us in so-called tripoli-powder and Diatom-earth, and the many calcareous Floridese which come down to us as Nullipore banks are deceptively like those living at the present day, that these groups have remained unaltered for eons, and that no form has been discovered in any of the older strata which could be regarded as a link with another group. No fossil remains are known beyond doubt to exist of the numerous aquatic plants with delicate cell-walls which perish as rapidly as they develop, of the SphcereUa species which give a red colour to rain-water and to the snow-field, of the micro- scopic Desmids, of the green filaments of Spirogyra, of the remarkable green tubes of Vaucheria pictured in plate I., &c. Some woody Polyporus species of Fungi have reached us, but in forms which look very like those at present growing on old tree-trunks. Some species of Moulds have been preserved in amber. I have before me a piece of amber in which insects are imbedded; from one of them spreads a web of mycelial threads which doubtless belonged to some mould-like Fungus such as to-day attacks various insects. The myceliums of various Fungi, also, are found penetrating the tissues of many of the fossil cryptogamic stems of the Coal Measures. This fact is very instructive, since it shows that in the tertiary period, and in much earlier periods also, the relation of saprophytic plants to the dead bodies of animals and plants were the same as they are to-day. All these results taken together prove that delicate Thallophytes whose cells do not become siliceous or cal- careous, or which are not inclosed in resin, cannot be preserved in a fossil condition. But no one would conclude from this that the groups to which such delicate growths belong were not represented in earlier periods. On comparing the past and present of the Vegetable Kingdom from these points of view, more especially with regard to the question whether existing groups stood side by side in earlier periods also, or whether, in the course of time, they have sprung from a single individual or from a few spontaneously-generated indi- viduals, we are obliged to decide in favour of the former. The so-called "higher" plants are not derived from the so-called " lower "; the groups of higher and lower plants co-existed from the beginning side by side. But variations within the limits of each group have always taken place. New species, i.e. new groups of species, arose in consequence of the crossing of the species already in existence. Of these the species which were best suited to the climatic conditions of the time being survived. But the variation in the formation of new species never went so far as to do away with the characteristics of the group. We immediately recognize in the fossil Laurel-trees, Magnolias, Oaks, Palms, Grasses, Pines, Equisetums, Ferns, Lycopods, FloridesB, Diatoms, and Moulds the ancestors of the now existing species! This would be impossible if the group-characteristics had disappeared in the modifi- cations which the species have undergone. When I now attempt the task of stating in detail what has been furnished by THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. 615 theoretical considerations, and of bringing forward the various groups which have from the beginning existed side by side, distinguishing them by the enumeration of their peculiar characteristics, I am not blind to the enormous difficulties of the undertaking. Although Palaeontology, Morphology, and Physiology afford valuable results, they are not enough, and neither of the three sciences gives sufficient data for the complete solution of the problem. One of the greatest obstacles is the afore- said incompleteness of the geological record. From the existing remains we may indeed conclude on the whole that numerous groups stood side by side in the mesozoic and paleozoic periods, but the evidence of many groups which exist at the present day without transitional forms is wanting, and when we assume their existence we make use indeed of a justifiable hypothesis, but have no proof what- ever. The danger, on the other hand, of establishing homologies from the similarity of an organ which is observed in groups of species now living has already been mentioned (see p. 609). Up to a certain point all organs which have similar work to perform agree with one another. This agreement is the more pronounced the greater the similarity of the conditions under which the organs have to do the work. Species of very different groups which live under water exhibit many characteristics in common; plants whose pollen-grains are transported by the wind show a great agreement in the structure and position of the parts of the flower. In the same way the form of flower- visiting insects necessitates a number of similar characteristics in the flowers visited. For example, we might instance the sweep- ing hairs on the style of the Proteacese and of the Composites, as well as certain developments which are met with in the flowers of Aroids, which are visited by small flies, and also in the Aristolochiaceee. In spite of this consideration, however, the similarity in the structure and form of organs, both of those serving for pro- pagation and for nourishment and growth, must obviously be kept to the forefront; similarity must always be an important factor in the limitation of groups. As we have in the preceding chapter established the fact that each species is built up by protoplasm with a specific constitution, the question might be pro- pounded whether each plant-group has not something in common in this respect. Many observations argue differently for this view. It has been repeatedly stated that the Moulds, Oscillatoriese, Sea-wracks, Stoneworts, &c. give off a scent which, although it differs according to the species, is yet very similar upon the whole, and that one is justified in assuming a specific constitution of the protoplasm in each of these groups on this account. Moreover, the scent which the Mosses exhale is found in no other group of plants. The same is true of Ferns. The delicate fronds of the tropical Filmy Ferns exhale the same peculiar scent as the larger Ferns of our forests. The Coniferse, Umbelliferse, Labiatese, Leguminosae, and Cruciferae exhibit similar conditions. Is it not also a striking phenomenon that the parasitic Fungus Cronartium asdepiadeum should settle on Cynanchum Vincetoxicum, as well as on Gentiana asclepiadea, i.e. upon two plants which the Botanist certainly places in different families, but which he regards as belonging to the same alliance ? To these facts many others might be added, especially with regard to the choice of ^16 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. vegetable food by animals. But our knowledge in this respect is so fragmentary and uncertain that for the present we cannot make use of these conditions in the limitations of the groups. The capacity for sexual union is of the utmost importance in defining plant groups. Species which can unite sexually belong undoubtedly to the same group. Nothing can be urged against this principle, and if it could be universally applied, the division of the groups would be settled. But in this matter there are very many pros and cons. The converse of the proposition requires consideration. It will not do to say that all plants which cannot unite sexually belong to different groups. It has been shown that crossings can be successfully effected in Orchids which all Botanists regard as members of different genera, but, on the other hand, it is demonstrated that crossings between very similar species of the Umbellifer family lead to no fruit formation. No one, however, would conclude from this that these Umbellif ers belonged to different groups. On reflecting in what a small number of flowering plants the fertilizing process has hitherto been observed, and remembering that the fertilization of many Thallophytes is still totally unknown, the hope of being able to utihze these conditions in limiting the groups becomes very much lessened. In the review of the various groups of the vegetable kingdom which follows^ no attempt is made to present the groups in the form of an ideal natural system. So far as the Thallophytes, Bryophytes, Pteridophyte's, and Gymnosperms are con- cerned, there is a very general consensus of opinion amongst Botanists, and the serial arrangement here followed is in harmony with it. But as regards the Angio- spermous flowering plants, and in particular the Dicotyledons, it is as yet too early in a book of this nature to embody all the most recent suggestions as to the affinities of the various families. Attention was drawn on p. 605 to the system of Alexander Braun, and it was pointed out that he was the first to try and break up the large and unsatisfactory class Monochlamydese or Apetalse, and to relegate its families in part to their true position. This attempt has been very fully carried out by Eichler (1883), and by Engler (1892); these two Botanists admitting only two classes of Dicotyledons (Choripetalse or Archichlamydese and Sympetalse). But as yet many of their placings of individual families are but tentative, and we may well wait a few years for a system on these lines to settle down into more or less permanent form. An instance of too hasty rearrangement of a natural system to meet recently discovered facts may be quoted here. In 1891 Treub discovered that Gasuarina possessed chala- zogamic fertilization, and in 1892 Engler (following Treub) separated Gasuarina from all other Angiosperms as the sole genus in a new class Chalazogamse. Since then it has been found (see p. 413) that chalazogamic fertilization is much more general than was at first supposed, and that in the group Amentacese it is widely spread, though by no means of universal occurrence. To break up the Amentacese in the drastic manner involved, if the class Chalazogamse be maintained, seems a most ^ Of. editorial note at commencement of this volume. THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. 617 undesirable and unnatural thing to do; it will on the whole be best to abolish a special class of Chalazogams, and, if thought necessary, to rearrange the families which constitute the Amentacese, but not to sever them from one another. For the Angiosperms we shall in the main follow the arrangement of the well-known Genera Plantarum of Bentham and Hooker, though we reserve our freedom to deal with certain families as seems well to us. The vegetable kingdom we divide first into four main divisions or phyla: (1) The Myxothallophyta, including the Myxomycetes only, a group standing apart from (2) the Thallophyta, which include the various classes of Algse and Fungi. Then follows (3) the Archegoniatse, forms possessing archegonia and fertilized by motile spermatozoids, and including the Liverworts and Mosses, and the series of the Ferns (Phyla (1) (2) and (3) constitute what are usually referred to as "Cryptogams"), and (4) the Phanerogamia or flowering plants, fertilized by means of pollen-tubes. They fall into two sub-phyla, Gymnosperms and Angiosperms, and the latter into two classes. Monocotyledons and Dicotyledons. Finally we have the 3 sub-classes of Dicotyledons — Monochlamydeae, Monopetalse, and Polypetalse. Here, in outline, is the system : — (1) MYXOTHALLOPHYTA, containing 1 class only. (2) THALLOPHYTA, containing 5 classes. I. Schizophyta. II. Dinoflagdlata. III. Badllariales, IV. Oamophycece. V. Fungi. (3) AKCHIGONIAT^, containing 2 classes. I. Bryophyta. II. Pteridophyta. (4) PHANEROGAMS, containing 2 sub-phyla. A. GTMSrOSPERJIiE. B. Angiosperms, containing 2 classes. I. Monocotyledones. II. Bicotyledones, containing 3 sub-classes. a. Monochlamydese. b. Monopetalse. c. Polypetalse. Each class (or sub-class) is further divided into a number of cohorts or alliances, and each of these alliances into orders or families. The alliances will be taken one by one in the following pages, their main characters generally indicated, and the families which they comprise enumerated. It will not be possible within the limits of this book to deal with the several families in at all a comprehensive manner, though the endeavour will be made to point out structural and other characters of interest, and where certain genera or groups of genera have a special interest these will be alluded to. No attempt is made to observe any due sense of proportion in treating of the different alliances. Thus a small alliance containing but few mem- bers of especial interest will receive more detailed consideration than one vastly 618 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. larger, the numerous representatives of which are unrelieved in their monotony. This method, truly, is an unconventional one, but in view of the restrictions of space, perhaps better suited to our purpose than any other. Phylum 1.— MYXOTHALLOPHYTA. Organisms destitute of chlorophyll, whose vegetative state consists of a mass of naked proto- plasm ( Plasmodium). Ee production by spores, from which arise swarm -spores or myx- amoebse, which unite again into plasmodia. Alliance I. — Myxomycetes, Slime-Fungi. For the most part saprophytic upon dead organic and especially vegetable sub- stances; they occur chiefly on accumulations of the dead parts of plants — leaves. Fig. 367.— Myxomycetes. ' ^ TZ^^ST'fLfJnZmnfT,'',,- ' ^ ''""'' ^PO^^ngium; x 6. > Dendritic mass of sporangia of Spumana alba on CrZuriumminu^ZT'Z °',t°'f »!" "f 'r"™; x^- '^ S™"? »' 'Porangia of the same. » and 7 Sporaiigia of l,ratenummmutum,'x2i. o Spovungia, ot Arcyrm punicea. » A single sporangium; x 10. lo Part of the neWike capilhtium Of the same; xl60. n ft„otiflcation of Lycogala epiden,in.m on a piece of wood >. LeTea^ frazil Plasmodium on the right; several sporangia on the left. J^eocarpm jmguis, a tan, rotting wood, and the like; they are rarely parasitic. Their life-history is as follows-:— On the germination of the spores the membrane bursts, and a slimy, nucleated mass of protoplasm escapes, which either swims in water by means of a THALLOPHYTA. GID' single flagellum, or creeps about on a damp substratum ; these motile bodies are the myxamoebce. These amoebsB increase at the expense of absorbed nutriment, and undergo repeated division. Ultimately they fuse together into masses of naked protoplasm— the plasmodium-stage— which creep about until spore-formation sets in. The plasmodium is transformed into the reproductive stage, numerous sporangia arising all over its surface. This stage in different cases shows the most varied structure, as may be seen by reference to fig. 367. Either the whole plasmodium is transformed into a single sporangium, as in Lycogala epidendrum (fig. 367 "), or a number of sporangia arise. A portion of the protoplasm becomes hardened to form the wall of the sporangium, whilst the contents is resolved into a mass of dust-like spores. In addition there arises in the majority of forms a sort of internal skeleton, the capillitiimi, which may consist either of a number of elongated tubes with characteristic thickenings on their walls, or these tubes may be united together intO' a continuous network (fig. 367^). On the bursting of the sporangium the spores are scattered and carried away by currents of air. They germinate when they reach a moist substratum, and the life-cycle is passed through anew. Under un- favourable conditions a plasmodium may become encysted, forming a transitory resting-stage. If placed in w-ater, the plasmodium escapes from the cyst, and continues its life-history even after a lapse of several months. The substance of the membrane, whether of the sporangial wall, spores, or capillitium, does not seem to consist of cellulose, but rather of a congealed protoplasm. We see, then, that the life-history of a Myxomycete is divided into a nutritive stage consisting of naked, membraneless, protoplasmic masses, and a sporangial, spore-producing stage. In respect of their nutritive stage the Myxomycetes very nearly resemble certain groups of the Protozoa, and on the strength of this resemblance they are regarded by many Botanists and Zoologists as belonging rather to the animal than to the- vegetable kingdom. In their manner of reproduction they certainly show analogies to many of the Fungi however. Plasmodiophora Brassicce is a parasitic Myxomycete which attacks the roots of the Cabbage, causing the disease known as " Fingers and toes " (cf. p. 622). Fossil Myxomycetes are not known. About 450 species have been distinguished. Phylum 2.— THALLOPHYTA. A large and very heterogeneous collection of plant-forms is included under this term. The word (Greek eaXUs, and 0ut6i') literally means plants with undifferentiated shoots, and includes practically all plants standing below the Mosses and Liverworts in complexity of organization. It is impossible to characterize positively a group, or rather a collection of groups, which shows so wide a range of organization as we find among the Thallophytes. They are often characterized negatively as plants whose bodies show no distinction between axis and appendages (stem and leaves). To such a plant-body the name thallus is given. But though this definition holds good for the great majority of the Thallophytes, yet there are forms (e.g. Bryopsis, €20 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. the verticillate Siphonese, Braparnaldia, many of the Brown Sea- weeds) which do show a distinction between axis and appendages, though the distinction is not usually so marked as that between the stem and leaves of a higher plant. Simi- larly, though for the most part possessing a relatively low organization, some Thallophytes (especially the higher Fungi and certain of the Brown Sea-weeds) show considerable diiferentiation of structure. The Thallophyta consist of many divergent and more or less sharply characterized groups. These include the Green Algse (Ghlorophycece, which includes the alliances Protococcoideae, Siphonese, Con- fervoideae, Conjugatse, and Charales), the Brown Sea- weeds (PhcBophycece), the Red Sea-weeds (Rhodophycece or Morideoe), the Blue-green Algse (Gyanophycece or ■Schizophycece), the Bacteria (Sohizomycetes, here included with the last-named group in one class, the Schizophyta), the Diatoms {Bacillariales), and the Fungi. Some Botanists consider that the Myxomycetes (or Mycetozoa), and the Dino- flagellata (Peridinece) are also plants. These last groups are almost certainly branches of the Protozoa; and though possessing certain plant-like characters (as indicated above), they show no near affinity with other plant groups. Whether they should be considered as plants or animals must therefore remain a matter of opinion. The same may be said of certain other organisms regarded by zoologists as Flagellate Protozoa, but showing undoubted affinities with the lower forms of Green Algae. There exists, in fact, an unbroken series of forms, connecting un- doubted Protozoa, having mouths by which they eat solid food, with undoubted green plants depending entirely on soluble inorganic food. The dividing line "between animals and plants is here obviously an artificial one, and is naturally drawn by difierent authorities at different points in the series. The name Thallophyta, then, is given to all plant-forms below a certain grade of organization, and includes many separate and widely divergent lines of descent. The Thallophytes may be pictured as the shrubby growth around the base of the phylogenetic tree representing the plant kingdom. The lower part of the main trunk of the tree, that is to say, the line of descent by which the higher plants have originated, is probably represented by certain of the Green Algse. Class I.— SCHIZOPHYTA. For the most part exceedingly small organisms, which propagate entirely by asexual methods. They consist of isolated cells, cell-filaments, surfaces, or masses. Though till recently regarded as without nuclei, these bodies have been found in a number of forms, and this view is being abandoned. They include both coloured and colourless forms; but the coloured forms never exhibit pure chlorophyll. THALLOPHYTA. 621 Alliance II. — Cyanophycese, the Blue-green Algee. Families: Ghroococcacece, Nostocacece. Includes pigmented forms in which in addition to chlorophyll phycocyanin is present, giving the cells a bluish, violet, or reddish tint. They occur in water or in moist places, and their cells may be united together into aggregates of various kinds. The cell-walls are usually mucilaginous, so that the cells or filaments cling together in colonies, or they are inclosed in special sheaths. The simpler forms included under the Chroococcacese are unicellular; the products of their division may either remain united into colonies or become quite free from one another. The rest are filamentous, and are included under the Nostocacese, whose filaments, can become segmented into small portions which move away by a peculiar motion not yet fully understood (c/. vol. i. p. 40). At times also certain cells become resting-spores and can endure climatic vicissitudes. They are widely dispersed over the globe, and are met with in cold glacier-streams and have been found living in hot springs at a temperature of even 85° C. Some 800 living species are distinguished. Ghroococcacece. — Includes the unicellular forms. Glceocapsa (vol. i., Plate I. n, o) forms little mucilaginous colonies, often found on the moist window-panes of hot- houses. Merismopedia forms films on stagnant water, and Glathrocystis like certain of the Nostocacese (alluded to below) arises in quantity in water. A form probably referable to this group (Bermogloea Limi) developed in 1874 in such quantities off the Adriatic coasts as to seriously interfere with the fishing industry. A commission was appointed to investigate the matter, but in six weeks the Bermogloea vanished as suddenly as it had appeared. Nostocacece are, for the most part, filamentous, though in some forms the cells may become isolated. Nostoc itself is common, and takes the form of irregular gelatinous colonies, which contain numerous interwoven necklace-like filaments. In some districts, owing to its sudden appearance after rain, it has received the name of " Falling Stars". This explains the allusion in the following lines from Dryden's (Edipus: — " The tapers of the gods. The sun and moon, run down like waxen globes ; The shooting stars end all in purple jellies, And chaos is at hand "■ A species common in China, i^. edule, is used as a thickening for soup, and an allied form, Hormosiphon arcticus, abounds in the Arctic regions upon floating ice. Anabcena Flos-aquce, Aphanizomenon Flos-aquce, &c., appear in fresh and brackish water — sometimes in enormous quantities, and to considerable depths. The Tricho- desmium, UrythrcBum^— another of these " flowers of the sea " — referred to at vol. i. p. 389, belongs also to this group. Very little is really known about the life- histories of these interesting plants, which so frequently appear in great quantities at or near the surface of the water and then as mysteriously disappear. But now '622 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. that systematic observations are being made of the organisms which occur at the surface {e.g. at the Biological station on the Ploner See, Schleswig-Holstein) we may- hope that these lacunae in our knowledge may be filled up. Recent investigations (by Klebahn) upon several of these " flowers of the sea " (Gloiotrichia echinulata, Anabcena Flos-aqiue, Aphanizomenon Flos-aquce, TrichodesrrduTn, &c.) seem to indicate that they possess special organs of flotation designated " gas- vacuoles ". It would appear that these natant forms have in consequence a smaller specific gravity than the surrounding water, and if the surface be quite unruffled tend to float, whilst any disturbance, such as waves, &c., is sufiicient to cause their ■distribution through the upper layers of the water. Whether these " flowers " pass another stage deep down in the water is not fully ascertained. Their spores, so far as they have been observed, do not seem to possess " gas-vacuoles ", and sink to the bottom. The phenomenon here indicated is not unlike that occurring in the Protozoon Arcella, the protoplasm of which is able by secreting a bubble of gas to rise to the surface, and, by absorbing it, to cause the organism to sink. The Oscillarias consist of filaments of disc-like cells ; they exhibit curious gliding movements, which have been already alluded to (cf. vol. i. p. 40). Eivularia is .distinguished by the fact that its filaments are whip-like, ending in a fine point, whilst in Scytonema this distinction of base and apex is not found. They generally occur in more or less mucilaginous masses. A number of the Schizophycese are associated with certain Fungi to form Lichens (cf. later, and vol. i. p. 244); and certain of them occur embedded in the tissues of other plants. Thus species of Nostoc are met with in certain Liverworts (Anthoceros) and in the roots of Cycas; and Anabcena in special cavities in the leaves of the Rhizocarp Azolla. It is not known what may be the exact physio- logical relations between these Blue-green Algse and the plants they inhabit— whether they are parasitic or symbiotic. Alliance III. — Schizomycetes, the Bacteria. On the whole, in the matter of their structure and aggregation, present many characters in common with the Blue-green Algse. They are, however, destitute of the characteristic pigment of that group, and pass their lives as parasites and saprophytes, obtaining their food from ready-formed organic matter. Nuclei have been distinguished in some few forms, and the cells are inclosed in a membrane which, though often mucilaginous, does not consist of cellulose. In size the cells reach very small dimensions, and may be regarded as the smallest of all plant- forms. A number of different forms are distinguished; the Cocci, consisting of minute round cells; short rod-like forms, the Bacteria; longer rod-like forms, the BacilH; various spiral forms known as Spirillum, Vibrio,, and Spirochaete; filamen- tous forms, Leptothrix and Crenothrix. A very common mode of occurrence of the various forms is in gelatinous masses, to which the name Zooglcea has been applied (cf. fig. 368^). Propagation takes place vegetatively by cell-division so long as the THALLOPHYTA. 623 conditions remain favourable for further growth. When the substratum is exhausted, &c., spores are formed which can remain for long periods — until, indeed, circumstances are again favourable for renewed development. These spores may- arise either inside the bacterial cells ( = endospores), the protoplasm contracting somewhat and forming a new wall around itself, or the cells may become trans- formed entirely into spores ( = arthospores), the wall becoming specially thickened. In no case is there any sexual process. Whilst many Bacteria are only known under certain forms and are regarded as species of the genera Micrococcus, Bac- terium, Bacillus, Spirillum, &c., others are known which, in the course of their development, pass through several such forms, and are termed pleomorphic. That all Bacteria are thus pleomorphic seems improbable, though the discovery that pleomorphism existed at one time led to the wildest generalizations. Large numbers of Bacteria display an active movement which, though formerly attri- buted to various contractions of their bodies, are now known to be due to cilia. These cilia may be borne in tufts of 6 or 6 at the two ends of the organism, as in Spirillum Undula, or they may be solitary at one extremity, as in the Cholera Bacterium (Spirochcete choleroe asiaticce), or they may be scattered over the surface of the organism, as in the Hay-bacillus, Bacillus subtilis. It is due to the extreme fineness of these cilia that they were not recognized long ago. Though the forms under which Bacteria occur are relatively few, their mode of life and special activity is exceedingly varied. The interest attaching to Bacteria rests largely on their eiFects on the substratum from which they draw their food. Taking, first, the saprophytes. These split up their substratum into simple sub- stances. In some cases there is a complete oxidation, with production of carbon dioxide and water; in others this is only partial, as in some of the cases of fermentation, e.g. when alcohol is oxidized into acetic acid by the activity of the vinegar organisms Bacillus and Micrococcus aceti (cf. figs. 368^ and 368*). Or there may be a decomposition unaccompanied by simple oxidation, as in many cases of fermentation, e.g. as when sugar is split into alcohol and carbon dioxide. Often these operations are accompanied by the development of a foul-smelling gas, when we speak of putrefaction. The number of saprophytic Bacteria which excite characteristic splittings in their substrata is considerable. In addition to those already quotfed, we may mention Bacillus Am,ylobacter, the organism of butyric acid ferpaentation; Bacillus lacticus, which causes milk to become sour; Leuconostoc mesenterioides, which has the power of converting large quantities of sugar into a gelatinous mass in a very short space of time. Again, in a number of forms the production of a special colouring matter is associated with the activity of the organisms, as is the case with Micrococcus prodigiosus (cf. fig. 368 ^), the "blood- portent" which makes its appearance on various starchy food-stuflfs, and Beggiatoa roseo-persicina, found on decaying vegetable matter in water, and known as "peach-mud". Many Bacteria are parasitic in the bodies of animals, and some among them are harmless. This is the case with Sarcina ventriculi (fig. 368^"), known only in the human alimentary canal in the form of packets of cells. Harm- 624 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. less also are a number of Bacteria found on the mucous membrane of the mouth. On the other hand, many ai-e associated with definite diiSeases. Spirochcete Ober- Tneieri (fig. 368 ^) is found in the blood in great quantities during relapsing fever; Bacillus anthracis (figs. 368'' and 368^) causes anthrax in cattle, &c.; and a great many other diseases — diphtheria, cholera (figs. 368 ^ and 368 ^), tuberculosis, leprosy, &c. — are associated with the activity of specific bacterial organisms. Nor must we omit to mention the numerous forms which occur in the soil, some of which are concerned in the process of nitrification, i.e. which oxidize ammonia into nitric acid, thus rendering this source of nitrogen available to higher plants, whilst others m ® Fig. 368.— Bacteria. 1 The "Wood-portent," Mio'ococms prodigiosus. 2 Zoogloea-form of same. ' Bacterium aceti. * The same more highly magnified. » S-pirochmte eholerce asiatiae. 6 The same more highly magnified. » Bacillus anthracis and red blood- corpuscles. « The same more highly magnified. 9 Spirochcete Ohermeieri and red blood-corpuscles, m Sarcina ventrieuU 1. 2, ', 5, 7, 9 X 300; l» X 800; 4, 6, e x2000. actually fix free nitrogen, as is the case with the organism occurring in the root- tubercles of many leguminous plants (cf. p. 521). There is no doubt this organism (Rhizobium, as it has been called) can store up free nitrogen, and that leguminous plants, when associated with it, obtain nitrogen not to be accounted for as combined nitrogen obtained from the soil. Curious also is the activity of the sulphur and iron Bacteria. The former (e.g. Beggiatoa alba) have the power of reducing the sulphates contained in the waters which they inhabit and of storing up sulphur- grains in their protoplasm; whilst the latter (e.g. Grenothrix Kuhniana), not uncommon in water-pipes, where they often develop in enormous quantities, store up iron in the gelatinous sheaths of their filaments. That Bacteria existed in former times, and were then, as now, the agents of decomposition, seems probable in view of the conditioii revealed by residues of dead THALLOPHYTA. 625 plants in the coal measures. It is thought that Bacillus Amylobacter has been identified in a silicified state. Of living Bacteria a very large number have been distinguished. Class II.— DINOFLAGELLATA, Peridine^. Alliance IV. This compact group of unicellular organisms is, as has been said above, a branch of the Flagellate Protozoa. They have therefore no very near affinities with other plant groups, though the nutrition of many is thoroughly plant-like. They would come nearest to the motile (Flagellate) forms of Protococcoidese (see p. 628). The great characteristic of the group is the possession of two flagella, one directed longitudinally and attached to the anterior end of the body, the other transversely directed and often situated in a circular transverse groove. There is sometimes also a longitudinal furrow. It is this transverse flagellum which is specially concerned in movement. There is often a cell membrane of cellulose, and the cell possesses green or brownish chromatophores containing chlorophyll and a single large nucleus. Reproduction is effected by binary fission, usually during a resting stage of the cell. Most of the forms are marine, and some are the cause of sea-phosphorescence. The group is divided into two sections — the Adinida without, and the JDinifera with a transverse furrow. Ceratium and Peridiniv/m are two well-known genera. Class III.— BACILLAEIALES. Alliance V. Family: Diatomacece, Diatoms. These are a large group of unicellular plants which grow both in fresh and salt water and upon moist soil. As a rule they occur together in large numbers. The protoplasm is coloured brown by a brown pigment, diatom,in, which masks the chlorophyll which is also present. The colouring matter is restricted to special chromatophores, which may be few or numerous. The cell- wall is incrusted with silica, and is a very characteristic feature of the Diatom. The wall consists of two halves or valves (frustules) which fit into one another like the lid on to a pill-box. These valves are smooth or variously sculptured, dotted, ribbed, &c., and enjoy a wide popularity as microscopic objects on account of the beauty and delicacy of their tracery. Some idea of the variety and form of Diatom-cells may be obtained from the accompanying figure 369. In the colonial forms the cells are attached to the substratum directly (fig. 369 ^) or by means of branching filaments (fig. 369 "). Others are attached to one another in zigzag chains or continuous ribbons (figs. 869 ^^ and 369 ^^). Others, again, are embedded in mucilage. Many of the forms VOL. II. 90 €26 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. exhibit a curious creeping movement, which is explained as being due to an external sheath or to filaments of protoplasm; the median line (or "raphe") shown by certain forms (e.g. Navicula, fig. 369 *) is interpreted as a narrow slit at which this external protoplasm is extruded. Diatoms propagate by continuous longitudi- nal division; the valves are slightly separated, and division takes place parallel to the faces of the valves. Each daughter-cell thus possesses one of the valves of the mother-cell, and they complete their integument by secreting another on the side away from it. The new valve is always slightly smaller than the other one and Kg. 369. — Diatoms, • Several Individuals of Synedra Ulna attached to a cell of an aquatic plant. 2 A single individual of Synedra Uh\a more highly magnified. » and < Navieula, Liber, seen from the side and from in front. = and « Similar views of Ifavieula tumida. ^ TriceratiuTn Favus. 8 Ccumpylodistms spiralis. ^ PleurosigTna angulatv/m. 10 and 11 GrammatopJwra ser- pentina. 12 and 18 Two views of Gomphonema capitatwm. i* Gomphoiiema capUatum on branched stalks which are attached to some algal filament is Diatom vidgare ; the cells hang together into a zigzag hand, i^ and 1? FragUlaria virescens, showing an individual from two aspects and a row of six joined together into a ribbon, is Cocconema Cistula. 19 Two individuals of Cocconerrui Cistula inclosed in a mucilaginous envelope preliminary to auxospore-formation. 20 xhe two auxospores have elongated, and the old cases are seen to right and left; there is no conjugation in this instance. All highly magnified. fits under its rim, so that the Diatoms arising in this way become smaller and smaller till a certain minimum is reached. When this diminution has gone on for a certain period an enlargement is attained by the formation of what are termed auxospores. The contents of the cell gather themselves together, they become free from the valves, enlarge, and then put on new valves. In other cases an actual conjugation takes place, two individuals uniting into an auxospore; or each may divide into two daughter-cells, which fuse in pairs, forming two auxospores. Fig. 369^^ shows two cells of Cocconema Cistula embedded in mucilage previous to auxospore-formation; in fig. 369^" each cell has escaped from its valves, and has doubled its original length. In this instance, however, there is no accompanying conjugation. Diatoms are very widely distributed over the globe, and occur in quantities in THALLOPHYTA. 627 the surface layers of the sea; some 30 species have been found amongst the inor- ganic dust of the snow-field. There are some 2000 species of living forms. Considerable deposits of Diatoms occur in various parts of the world; of these the most remarkable is that of Richmond, Virginia, U.S.A. It is said to extend for many miles and to be 40 ft. deep. They are found in secondary, tertiary, and more recent rocks. Siliceous marl consists entirely of the tests of Diatoms. A block of such a Diatom-deposit some two cubic feet in bulk from a fresh-water lake in Australia is exhibited in the Botanical Department of the British Museum; the number of Diatoms contained in it (there are 21 different species) is estimated as exceeding 12 billions (12 x 10^^). Cla.ss IV.— GAMOPHYCE^. Sub-class I. — Chlobophycb^, the Green Algce. There can hardly be a more fascinating group of plants than this, whether to ■ the strictly scientific botanist or to the more catholic lover of nature. In the first place, the Green Algae are among the most widely diffused of plant-forms. They grow practically in every place where enough moisture, together with light and air, are to be had. Between tide-marks on almost every coast, floating on the surface of the deep sea, covering damp earth, walls, palings, and tree trunks, sticking to the surface of leaves in the moist atmosphere of tropical forests and jungles, and inhabiting almost every river, brook, pond, ditch, or casual pool of rain-water in all quarters of the globe, are members of this ubiquitous group to be found. Nor are they wanting from more extraordinary situations. In Switzerland, Norway, and other countries where snow is more or less permanent, the bright red patches on its surface, known as "red snow", are formed by the microscopic Alga {Sphcerella niva- lis) shown in Plate L of the first volume. Other forms, not so far removed from Sphcerella, live in the intercellular spaces of higher plants, such as the Ivy-leaved Duckweed (Lemna trisulca), the Moneywort (Lysimachia num/mnlaria), and others. Yet other Algae are found inhabiting the jelly of certain fresh-water sponges in the East Indies, where they seem to live in a regular symbiosis with their hosts. But perhaps the most curious dwelling-places of all are the hollow hairs of the Three-toed Sloth (Bradypus tridactylus), which are inhabited by an Alga called Trichophilus. In the second place, many of the Green Algae (and among these some of the very commonest) are the most beautiful forms of life imaginable, and the main features of their structure can be made out with quite low powers of the microscope. But perhaps the greatest claim on our interest is made by the fact that we must look among the Green Algae not only for indications of the origin of all plant life, and of the forms from which the whole of the higher plant world arose, but also for hints towards the solution of some of the most difficult and fundamental problems with which Biologists are occupied, queytions as to the real nature and origin of sexual reproduction, as to the distinction between gametes and asexual reproduc- 628 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. tive cells, as to the physiological conditions which determine their formation, and finally, questions as to the division of labour in the protoplast itself, the function of the nucleus and its relation to the protoplasm, and how far the latter can be resolved into separate, but mutually co-operative parts. The framework of our knowledge on the former group of questions has been, and is being, built up largely from a study of the Green Algse; of the latter — and even more fundamental— we know, as yet, very little, but the same group of plants is already beginning to yield important results, and we have every reason to expect an even richer harvest in the immediate future. We shall now proceed to a brief review of the groups into which the Chloro- phyeese may be divided, mention being made of the more interesting forms in each group. The sub-class Chlorophycese may be defined as follows: — Thallus of very various form, one- or many-celled, coloured green by chlorophyll which is contained in chromatophores of very various shape: the green colour rarely masked by other pigments. Reproduction by motile zoospores, and by gametes, which either resemble small zoospores, and are equal in size (isoplanogametes), or are diflferentiated into two categories; first, relatively small, active male gametes— called spermatozoids when they are very highly difierentiated — and secondly, relatively large, passive female gametes — known as eggs or oospheres when they are quite passive. The cell produced by the fusion of the bodies of two gametes is called the zygote, and gives rise to a new plant either mediately or immediately. When the gametes are sexually differentiated, the zygote (now called a fertilized egg or oospore) is pro- duced only by the fusion of a male with a female gamete. Other reproductive cells are known as aplanospores and aJcinetes. Aplamospores are formed by the protoplasm of a cell of the thallus rounding itself off and putting on a new cell-wall, or dividing into several parts, each of which acquires a separate cell-wall. Akinetes are simply single cells of the thallus, whose original walls thicken directly, the cells becoming separated from the rest of the thallus. These two categories of reproductive cells germinate at once to form new plants. The Chlorophycese comprehend the following alliances: — Protococcoidese, Siphonese, Confervoideae, Conjugatse, and Charales. Alliance VI. — Protococooideffi. Families: Chlamydomonadece, Volvocece, Pleurococcacece, Endosphceracece, Hydrodictyacece. Unicellular forms, actively swimming, floating, or fixed in habit, living either singly or united into colonies. Family Chlamydomonadece. This family consists of minute green (sometimes red) organisms which spend the greater part of their life actively swimming about in water. Several species of the genus Ghlamydomonas are very common in stagnant water. Each Ghlamydomonas-cell consists of a roundish mass of proto- THALLOPHYTA. 629 plasm covered by a delicate cell-wall of cellulose. At first sight the whole body appears green, but on careful examination it will be seen that the green colouring- matter is really confined to a layer on the surface of the body. This chlorophyll- layer or chromatophore is specially thickened at one end (the posterior end) of the body, and a small, clear, spherical body (the pyrenoid) is often to be seen inclosed in this thickened portion. Round the sides of the body the chromatophore is thinner, and towards the anterior end it stops, leaving a small area of colourless protoplasm to occupy the front extremity of the organism. This is in connection with another small mass of colourless protoplasm which occupies the hollow of the cup-shaped chromatophore. A minute red speck is visible at one side of the body, situated sometimes on the surface, sometimes at the limit between the chromatophore and the central colour- less protoplasm. The rotation of the Chlamydomonas about its antero-posterior axis, which accompanies the jerky, forward movement (cf. vol. i. p. 29) can be well traced by the aid of this eye-spot, which can be seen to be carried round and round as the organism turns on its axis. The means by which the Chlamydomonas moves are not at first sight apparent. But when the cell has come to rest, and especially if it has been killed and its protoplasm fixed with a dilute solution of iodine, a pair of very delicate protoplasmic filaments, called fiagella, can often be seen projecting from the colourless anterior spot of protoplasm. It is by the con- tinual lashing of these fiagella that the Ghla/mydomonas is pulled through the water. The iodine will also bring into view a small spherical nucleus, stained dark brown, situated in the colourless central protoplasm, and will stain the neighbour- hood of the pyrenoid a dark blue. This last phenomenon is due to the formation of starch by the protoplasm round the pyrenoid. The exact part played by the pyrenoid is very obscure, but there can be no doubt that it influences in some way the formation or deposition of starch by the protoplasm. The movements of Chlamydomonas are often in direct relation to light, the organisms moving towards a source of light of moderate intensity. Thus if a glass vessel filled with water containing GhlaTnydomonas be placed in a sunny window, there will be a general movement of the swarming cells to the sunny side of the vessel, causing the water to become much greener on that side. There is some evi- dence that it is the eye-spot which enables the organism to perceive the direction from which light is proceeding. Nearly all the motile Protococcoidese, as well as zoospores and gametes which are sensitive to light, possess an eye-spot, and it has been thought that the eye-spot-pigment (a substance called hcematochrome) may act in the same way as the visual purple in the retina of the eye. If specimens of Chlamydomonas be kept for a day or two in water, some of them will often be found to have divided to form daughter individuals, which still remain inclosed within the cellulose membrane of the mother. This division is preceded by the drawing in of the fiagella. The protoplasm of the body then withdraws itself from the wall, and divides transversely to form two roundish masses. Each of the latter may either at once put on a cell-wall and develop fiagella, or it may divide 630 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. again, so that four daughter individuals instead of two are formed. Eventually the daughters escape from the membrane of the mother, leaving it quite empty. The process of reproduction by simple division of all the protoplasm of an indi- vidual's body into parts, each of which forms the body of a daughter individual, is an example of almost the simplest type of reproduction known. It is true that in the lower Protozoa, which have no rigid cellulose membrane, we find an even simpler type. Since the entire organism consists of protoplasm, there is nothing left of the parent individual after division has taken place. The body of the parent simply becomes the body of the offspring. In the type of Chlamydomonas we have the dead cellulose membrane representing all that is left of the body of the parent. In many of the higher Algas, and in all plants above the level of Thallophytes, only a part of the protoplasm of the plant-body is used in the formation of the reproductive cells. The rest must then eventually die. But in these lower forms, where all the protoplasm of the body is used in the production of new individuals, death, as a necessary event, can hardly be said to occur. Gametes are formed in Chlamydomonas in exactly the same way as daughter individuals. They are, however, smaller and have no cell- wall. In one species at least the gametes are of two sizes. Of the smaller (microgametes) eight are produced from a parent individual, while only two of the larger size (megagametes) are formed from the parent cell. In the process of conjugation a microgamete and megagamete come into contact at their anterior colourless ends, the flagella are drawn in, and a thick cellulose membrane is secreted round the bodies of both. The protoplasm of the microgamete then passes over into the space inclosed by the part of the mem- brane belonging to the megagamete, and completely fuses with the protoplasm of the latter. A wall is then formed, cutting off the empty shell of the microgamete. The contents of the zygote eventually divides to form two or four new individuals which escape from its membrane. In the conjugation of most species where there is no distinction in size between the gametes, a cell- wall is only acquired after the foundation of the zygote. This is the regular course of events in the conjugation of the motile gametes of Green Algse. The genus Sphcerella resembles Chlamydomonas in the fundamental points of its structure. The main distinction is the existence of a considerable space separating the membrane from the main body of the protoplasm of Sphoerella. This space is bridged by fine strands of protoplasm, which radiate from the central mass and end in fine branches under the membrane. The anterior colourless protoplasm is drawn out into a beak, and to the extremity of this the two flagella are attached. The flagella often pass through two very delicate cellulose tubes, which in the common species, S. pluvialis, diverge from the extremity of the beak, and end on either side at the membrane. In another species {S. Bwtschlii) the beak runs right up to the membrane, and the flagella-tubes, which are short and slightly curved, lie on the outer surface of the membrane. Sphcerella pluvialis is a very well-known microscopic object, being extremely common in pools of rain-water. Its protoplasm often contaiiis a good deal of the red THALLOPHYTA. 631 pigment hsematochrome, which may almost completely mask its green colour. It was from this circumstance that it received its name of Hoematococcus pluvialis. Hsematochrome is even more constantly present in Sphcerella nivalis, the "red snow", whose appearance and life-history have already been described (vol. i. p. 39). S. Butschlii has its hsematochrome concentrated in an eye-spot like that of Ghlamy- domonas. The VolvocecB differ from the Chlamydomonadece in consisting of motile colonies of cells, the members of each colony being united in a common investment. The bodies of the individuals composing the colony are also joined in some genera by protoplasmic processes. The body of each individual is identical, in the fundamen- tal points, of its structure, with that of a Ghlamydomonas or of a Sphcerella. The Volvocece present us with an interesting series of forms, showing a gradually increas- ing sexual diiFerentiation of gametes, and, in the higher forms, an interesting sub- ordination of the individual to the colony as a whole. Gonium is a form in which the colony consists of (usually) sixteen Ghlamydo- moTias-like cells arranged in a flat plate, which swims in a line at right angles to its surface, the flagella of the central cells of the disc projecting forwards, those of the peripheral cells obliquely outwards and forwards. All the cells are inclosed in a general mucilaginous envelope, and are joined to one another by protoplasmic processes. Reproduction is effected by the division of the constituent cells of the colony in two planes at right angles to one another and to the plane of the colony, so that each mother colony produces sixteen daughter colonies, whose discs of cells all lie in the same plane. Meanwhile, the mother cells are separated from one another by the gradual liquefaction of the general mucilaginous envelope, and thus the daughter colonies become independent. Formation of isogametes also takes place, but is not thoroughly understood. Stephanosphoera is a very beautiful form, occurring especially in pools of rain- water collected in rock hollows in hilly districts. It is often found in company with Sphcerella pluvialis. The colony consists of a ring of (usually) eight Sphcerella-like cells arranged in the equatorial plane of a spherical or ovoid cellulose membrane. When reproduction is about to occur, the constituent cells draw in the protoplas- mic processes by which they are attached to the general membrane; each secretes a membrane of its own, and then its protoplasm divides in two planes to form eight (sometimes seven) daughter-cells. When these have acquired flagella they begin to swarm, and eventually escape by bursting the membrane of the mother colony. Gametes are formed in the same way, but usually by more divisions, as many as thirty-two being sometimes produced from a single cell. In most cases all the cells of a colony divide at once to form gametes, but this is not invariably the case. Each bundle of gametes produced from a single cell breaks up, and all the gametes begin to swarm within the colony. The gametes are spindle-shaped, each with two flagella and an eye-spot. They conjugate in pairs, usually inside the general mem- brane, but conjugation never takes place between two gametes derived from the g32 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. same mother cell. The actual process of pairing has been fully described, and it may be taken as a type of the course of events, as it has been observed an all Algae with isoplanogametes whose pairing has been fully investigated. Out of the crowd of gametes swarming in all directions, two approach and stroke each other with their flagella; in some cases the two separate and both become again lost m the crowd, but, when pairing is going to take place, they become firmly fixed together by their colourless anterior ends. The long axes of their bodies may then lie m one straight line, or may diverge at a wide angle. A rotation of each of the painng gametes about its fixed anterior end now always occurs, the bodies becoming gradu- ally approximated, so that their long axes come to lie nearly parallel. Fusion of the protoplasm follows, beginning at the already joined anterior ends, and progressing rapidly backwards till a single mass of protoplasm is formed. The four flagella still move actively, and the Zygozoospore, as this active type of zygote is often called, escapes from the general membrane of the mother colony, becomes spherical by shortening of its long axis, loses its flagella, and puts on a cell- wall. The protoplasm soon loses its green colour, becoming reddish, and the zygote enters on a resting period. Eventually its protoplasm divides, producing zoospores, each of which gives rise to a new colony. The division of the cells of a colony to form zoospores or gametes begins in the evening, and is finished soon after sunrise. In dull weather, however, its comple- tion is delayed far into the day. This dependence of the formation of zoospores and gametes upon the influence of light, if not invariable, is found very widely among the Green Algse. Pandorina is a fairly common form in ponds, &c. The colony consists of sixteen wedge-shaped cells arranged in a sphere, and covered by a general invest- ment, which is of considerable density at its external surface. The apex of each wedge is directed towards the centre of the sphere, and there is little space left between the adjacent cells. The formation of daughter-colonies is similar to that obtaining in Gonium and StephanosphcBra. The young colonies escape by liquefac- tion of the investing membrane. Colonies of gametes are formed in the same way, but often consist of eight instead of sixteen cells, and the acquirement of flagella and liquefaction of the mother membrane takes place more slowly. Event- ually the membrane of each gamete-colony also becomes liquefied, and most of the gametes swarm out into the water. A great number of mother colonies of diflferent sizes simultaneously take part in this production of gametes, so that the water becomes filled with masses of swarming gametes of very variable dimensions. No distinct size-categories are, however, to be distinguished. Conjugation now takes place between pairs of gametes either of the same or of difierent sizes; with this exception, that the largest gametes do not fuse with one another. They are rela- tively inactive, sometimes, indeed, remaining fixed in their colonies, and are sought out and paired with by the smaller and more active individuals. Here, then, we have a most interesting stage in the evolution of sex. The largest, relatively passive, gametes may fairly be called female, while the different sizes of smaller THALLOPHYTA. 633 gametes, though sexually differentiated inter se, since conjugation is apparently possible between any two, whatever their relative size, may perhaps be considered as male in relation to the largest. It is probable that fully-differentiated male and female gametes arose from forms such as we find in Pandorina, by the suppression of the intermediate sizes, the smaller and more active gametes taking on the function of actively seeking out the larger passive individuals, which on their side contribute practically the whole of the stock of food required by the zygote in germination. In correspondence with this we find the chromatophore (chlorophyll-corpuscle), which may be considered as the specially food-producing organ of the algal cell, much reduced and eventually functionless or absent altogether in the more highly differentiated male gametes (spermatozoids). Eudorina has a colony of sixteen or thirty -two almost spherical cells consider- ably separated from one another, and inclosed in a general investment like that of Pandorina. In the production of daughter-colonies Eudorina resembles the lower forms of the volvocine series, but in the sexual differentiation of the gametes there is a decided advance upon that obtaining in Pandorina. The perfectly passive female gametes (oospheres) hardly differ from the ordinary cells of a vegetative colony, while the active male gametes (spermatozoids) are formed in bundles of sixty-four by successive divisions of similar cells. Here, then, we find the marked difference in size between the two categories of gametes brought about, as it very often is among the Algse (and, indeed, among many other plants and animals), by a marked difference in the number of divisions occurring in their respective mother- cells. The present case in which strictly comparable cells on the one hand directly give rise to eggs, and on the other divide to form sixty-four spermatozoids each, is rather extreme, but we have already met with a similar case in a species of GhlaTnydomonas. Each spermatozoid of Eudorina is club-shaped, with a colourless pointed anterior end bearing two flagella and possessing an eye-spot, and a yellowish thick posterior extremity representing the (reduced) chlorophyllous portion of the typical volvocine cell. The spermatozoid bundle (male colony) escapes from its mother-cell- membrane, and swarms as a whole towards a female colony. On reaching the latter the spermatozoids get their flagella, become entangled in the thick mucilage, and rapidly separating from one another, worm their way into the female colony. Some succeed in fusing with the individual female gametes, and each zygote thus formed will eventually give rise to a new Eudorina colony. A form recently discovered almost at the same time in three different States of Xorth America, and known as Pleodorina, shows an important difference from the types we have hitherto been considering. Each spherical colony consists of about 28 cells, but not all of these are capable of producing daughter-colonies. This >ower is confined to those cells which occupy the posterior half or two-thirds of the sphere (it should be explained that the colony moves forward in relation to a iefinite axis). The smaller anteriorly-placed cells are thus purely vegetative in 634 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. function, and necessarily die after the reproductive cells have given rise to daughter- colonies. This is the first time we have met with such natural death among the Algee, and it is very clearly seen to be connected with the separation of the assimila- tive and reproductive functions. The formation of gametes in Pleodorina has unfortunately not yet been observed. The remarkable and beautiful organism called Volvox has been known for more than two centuries, and has long been among the most favourite of microscopic objects. The purely scientific interest which it has aroused has been as great as the aesthetic admiration which it has excited. Long and animated controversies have raged on the question as to whether it was to be regarded as an animal or a plant, as an individual or as a colony of individuals. And although these ques- tions have now lost much of their actuality through the gradual recognition by naturalists that we have absolutely no criteria by which they can be settled, there have arisen problems which promise quite as much interest and excitement for the future. Volvox is much larger than the forms hitherto described. The colony is spherical, and possesses a single layer of cells on its surface. In V. Globator the sphere may be over a millimetre in diameter, but more usually its diameter is only some three-quarters of a millimetre. A particularly large specimen may possess as many as 22,000 cells (10,000 is a more usual number). Many of the cells, however, as in Pleodorina, are simply vegetative, and take no part in the reproduction of the colony. Usually, in fact, only a very small minority of the cells are repro- ductive. The two species of Volvox differ from each other in a great number of points, but we have only space for a very brief description of some of the most interesting. The cells of Volvox Globator are all united together by very stout processes. Each cell, which is inclosed in a separate cell- wall, possesses all the ordinary features of the Chlamydomonas type. Daughter-colonies are developed from special cells, usually eight in number, called parthenogonidia. They are always formed in the posterior part of the mother-colony, early becoming larger than the ordinary vegetative cells. Each divides repeatedly, and forms a hollow sphere of closely-packed cells, which, after the last division, mostly acquire the characters of the adult vegetative cells, the remainder gradually increasing in size to form the reproductive cells. The young colonies then escape from the mother, apparently by pushing themselves against and making rents in its posterior wall. Subsequently the cell-membranes swell a good deal, separating the cell-bodies from one another, and the colony attains its adult size. In other cases gametes may be formed in a young colony. About five cells (androgonidia), strongly resembling the parthenogonidia, divide to form discs or hollow spheres of a hundred or more spermatozoids. The spermatozoids resemble in a general way those of Eudorina, but are peculiar in having the pair of flagella inserted laterally at the base of the colouriess beak and near the eye-spot. In the THALLOPHYTA. 635 same colonies, but a little later than the androgonidia, about thirty much larger spherical cells, the oospheres, are developed, and these are duly fertilized, but by spermatozoids derived from another colony. The zygote has a sculptured exine. Parthenogonidia are not found in those colonies which produce gametes. Volvox aureus ( = F. minor), the commoner form, is usually much smaller than V. Globator, and has rounded cells more widely separated and connected by very delicate processes. But perhaps its most striking characteristic is the very great variability in the number and distribution of the reproductive cells. The partheno- gonidia, which vary in number from one to sixteen, may either occur alone or in one colony with androgonidia or oospheres, or both. Most of the sexual colonies are dioecious, though this is not always the case. The colonies containing androgo- nidia unaccompanied by other reproductive cells often develop very numerous (up to 1100) spermatozoid bundles, the androgonidia forming one-third of all the cells of the colony. The spermatozoids differ from those of V. Globator by their larger size, by their terminal flagella at the end of a shorter beak, and by the possession of a well-developed leaf-green chromatophore. We must, therefore, consider V. aureus as not so highly developed, in some respects at least, as V. Globator. A Volvox-colony always swims in the direction of a given axis passing through its body, and at the same time rotates to the right or left about an axis which is inclined obliquely to the antero-posterior axis. The eye-spots of the vegetative cells are much better developed in the anterior half of the colony, and are always situated on the side of the cell nearest the anterior pole. These facts tend to support the view of the function of eye-spots in general suggested above. Volvox stands at the head of the series of colonial (coenobe-forming) organisms which we have been tracing, a series diverging from a Ghlamydomonas- or Sphce- rella-\ike type, and whose successive forms gradually increase in size, complexity, and sexual differentiation. Volvox itself has been well spoken of as " the culmina- tion of Nature's attempt to evolve a higher organism out of a ccenobe ". It was an attempt which failed, or rather which could not be carried any further than Volvox itself. A delicate, easily-ruptured Volvox-ST^heve could certainly not continue to exist if it were much more than a millimetre in diameter. As it is, the wall is often split, and all sorts of smaller organisms get inside, resulting in the more or less speedy collapse of the Volvox-colony. . But there are other series diverging from the Chlamydomonadege, and some at least of them have followed lines on which it was possible for higher and more varied plant-forms to be developed. At the first stage along one of these lines of descent we find ourselves among forms in which the dominant phase of the life-history falls in a resting stage, either fixed or freely floating in the water. From this resting stage motile forms (zoo- spores), corresponding with the free-swimming Ghlamydomonas individuals, are directly developed. These zoospores, after a short period of swarming, come to rest, often fixing themselves by their anterior end to some solid object. With little or no change in the constitution and appearance of the cell the main portion of the €36 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. life cycle is passed in this fixed condition, and cell divisions take place, the products eventually again developing flagella and being set free as zoospores. The genera Ghlorangium and Physocytium are examples of the simplest form of this type of life-history. Forms with a rather more complicated structure in the fixed stage are found in the genera Mischococcus (a common form on the surface of threads of the higher Algse), Uuglenopsis (a newly-discovered American plant), and their allies- In these the protoplasm of the zoospore, after fixing itself and putting on a delicate cell- wall, pushes out the surface of its membrane away from the substratum, thus forming a tube of gradually increasing length, the apex of which is always occupied by the protoplasm. Division of the protoplasm and subsequent pushing out of the wall of the tube in different directions by the daughter-cells results in a branching of the hollow stalk, and in this way quite a considerable branching plant-body may be produced. Eventually some or all of the cells occupying the apices of the various branches of the tube acquire flagella and escape into the water as zoospores, which again settle on solid objects and give rise to new plants. Other forms in which the cell derived from a zoospore multiplies by division, the products eventually again giving rise to zoospores, are Schizochlamys, Botryococcus, Bictyosphwriv/m and Tetraspora. In these, however, the immotile phase is not fixed, but forms floating colonies of various conformation. Into this topic we cannot enter further, except to remark that Tetraspora forms flat colonies of cells arranged in one plane and held together by the swollen mucilaginous cell- walls. Cell division takes place in planes at right angles to that of the colony. This type of colony is specially interesting, as it suggests the form of thallus found in Ulvacece, which in turn appears to lead on to the higher forms Confervoidece. Pleurococcacece. — More or less closely allied to the above-mentioned genera are others which do not form zoospores at all. These types with no motile phase in their life-cycle may be conveniently classed together as Pleurococcacece. The type- genus Pleurococcus contains some of the most widely-distributed algal forms known. P. vulgaris forms the bulk of the green coating of damp earth, tree trunks, palings, &c., in all regions of the globe. It consists of roundish cells, dividing in three direc- tions in space and thus forming solid masses of cells hanging together in multiples of two, and often flattened by lateral contact. Each cell contains several parietal chro- matophores which may, however, fuse together to form a single one. Eesting aki- netes are formed by the cells ceasing to divide, becoming spherical, and thickening their walls. At the same time oil appears in the protoplasm. It is probably mainly in this phase that Pleurococcus gets distributed by the wind from one place of growth to another. Owing to the resemblance of the akinetes of some of the confervoid Algse to those of Pleurococcus, it has often been stated, and indeed is still held by some algologists that Pleurococcus itself is merely a growth-phase of these higher Algffl. But recent culture-experiments leave little room for doubt that Pleurococcus is a perfectly autonomous form, although it may often be associated with pleurococcoid stages of other Algse. Eremosphcera is a pretty form, common in fresh water, with single floating spherical cells. Each cell contains numerous separate chlorophyll- THALLOPHYTA. 637 bodies, embedded in a parietal layer of protoplasm, and a nucleus suspended by protoplasmic strands in the centre of the cell. Multiplication is effected by division of the protoplasm into two daughter-cells which escape by rupture of the mother- cell membrane. Scenedesmus is another motionless floating fresh-water form. It consists of oblong cells united into groups of two, four, or eight, which lie side by side, palisade fashion. Some or all of the cells often possess straight or horn-like projections of their walls, which give the cell groups a very characteristic appear- ance. The single solid chromatophore occupies nearly the whole cell cavity. Ghlo- rella is a genus whose cells are symbiotic with Radiolaria (yellow cells). Other forms live in a similar relation with certain Coelenterates and Platyhelminths. Several help to form lichens. Since the various genera of Pleurococcacece differ thus very widely in the form and structure of their cells, and indeed are only united by the negative character of the absence of zoospores, it is almost certain that they cannot be considered as forming a natural group. The various genera are very probably allied to different neighbouring groups from which they have been derived by the suppression of the habit of forming zoospores. The Undosphceraceoe are a small and very natural group of unicellular Algse, characterized by their habit of living in the intercellular spaces of various higher plants. They possess motile zoospores, or gametes, or both, but the motionless cells produced from these do not undergo vegetative divisions. Very possibly they represent a separate line of descent from the Chlamydomonadese, a line of descent in which the motionless cell has become the dominant phase in the life-cycle, and has been specially adapted to the new conditions of life, but differs from the motion- less cells of the " Tetrasporaceas " in directly forming zoospores without undergoing purely vegetative divisions. Two forms of Endosphseraceae may be taken as illustrations of this type of life- history. ChlorochytriuTn Lemnce inhabits the intercellular spaces immediately under the epidermis of the leaves of Lemna trisulca (the Ivy-leaved Duckweed). Each plant consists of a single, thick- walled, oval cell with a parietal chromatophore containing numerous pyrenoids and a large central vacuole. Very numerous pear-shaped isogametes are formed by successive divisions of the protoplasm of the cell. Then a layer of substance outside the mass of gametes (probably the ectoplasm of the cell) begins to swell strongly, and bursts not only the cell- wall but also the super- incumbent tissue of the Duckweed leaf, forming a sphere of mucilage in which the gametes begin to swarm and to conjugate in pairs. Spherical zygozoospores are thus produced; these escape from the mucilage, and after some free swarming in the surrounding water, settle on the boundary between two epidermal cells of a Duckweed leaf, draw in their flagella, put on a cell-membrane, and form a definite parietal chlorophyll-body with a single pyrenoid. After two or three days a delicate, colourless tube is put out, which forces its way between the two epidermal cells of the leaf, and reaches an intercellular space. The contents of the zygote slowly pass over into the apex of this tube, which gradually increases in size and assumes the 638 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. characters of a young vegetative cell, the original zygote-wall remaining on the surface of the leaf as a mere cellulose knob. The generations rapidly succeed one another during the summer months, the last-formed cells of the season becoming packed with starch grains and passing the winter in this state. These resting cells can withstand desiccation, in case the pond in which the duckweed lives becomes dried up. Phyllobium dimorphum forms large immotile cells between the tracheids of the vascular bundles in the leaves of the creeping Moneywort (Lysimachia nummularia). This plant lives in damp woods and other shady places. The Rhine plain in the neighbourhood of Strasburg, where Phyllobium was first found in the leaves of the Moneywort, is usually flooded during the month of June, partly by the rising of the river, and partly by the thunderstorms which usually occur about that time of the year. The Phyllobium-cells take this opportunity to form their gametes, which are of two distinct sizes, each cell producing gametes of one size only. After the escape of the gametes into the surrounding water conjugation occurs. The zygozoospores produced have only two flagella, the body and flagella of each microgamete being completely lost in the megagamete, just as the body of a spermatozoid is completely lost in the substance of the egg. After coming to rest on the surface of a Lysimachia leaf, and acquiring cell-membranes, the zygotes put out delicate tubes which enter the stomata of the leaf. If a leaf is infected by a few zygotes only, the tubes formed reach the vascular bundles, and forcing their way between the elements of the wood, grow forward in the bundles, branching when they branch, and attaining to a considerable length. Eventually, towards the end of the summer, the protoplasmic contents of each tube becoming concentrated in one spot, this part of the tube swells and is cut off from the remainder by the formation of transverse partitions. The swollen part of the tube thus forms a large ■cell which rests during the winter, and in the next summer will produce gametes. If, on the other hand, the leaf is infected by a large number of zygotes, most of the tubes never get any further than the intercellular spaces immediately under the stomata. In this position they form small resting cells in large numbers. These eventually form zoospores, which apparently behave, on germination, just like the zygozoospores. The dimorphism of the resting cells of Phyllobium thus depends directly on the amount of space at the disposal of the germ tubes. This conclusion can be confirmed by cultivating the germ tubes apart from the leaves of the host. The purpose of the germ tubes of Chlorochytrium, Phyllobium, and their allies in penetrating the leaves of their hosts, seems to be simply that they may gain the ^vantage of a quiet protected place for their development. Just in the same way Diatoms and other unicellular forms often live comfortably in the empty cells of Algse, the intercellular spaces of the Bog-moss (Sphagnum), and similar situations. Only in the case of these Endosphsereas the association of the Alga with its habitat is invariable and adaptive, not merely casual and unrelated. But the Endosphserea are not parasites in any sense. They take no food from their " hosts " nor do they ■exercise any appreciable influence on the latter. This is sufficiently proved by the THALLOPHYTA. 639 fact that Lemna trisulca lives quite happily and can flower when infested with Chlorochytrium, and that the germ tubes o£ Phyllobium dimorphum usually enter dead leaves of the Moneywort. Another form which always enters the living leaves of a river-weed, continues its course of development whether the leaves die or re- main alive. It is not, however, difficult to imagine how a form like Phyllohiwm, living as it does in the vascular bundles of its host, might acquire a parasitic habit by tapping the food supplies. As a matter of fact certain confervoid Algse are known whose presence results in the death of the leaves they inhabit, though pro- bably not by direct appropriation of the food of the host. Resembling the Endosphcerece in possessing motionless cells which form zoospores but do not undergo vegetative divisions, are certain common fresh- water forms of which Characium and SciadiuTn may be mentioned. A plant of Sciadium ori- ginally consists of a single cylindrical cell whose contents breaks up into zoospores. These zoospores have acquired the peculiar habit of settling on the rim of the mother-cell, instead of seeking out fresh spots for their development. Each zoospore produces a single cell like the mother, so that a whorl of cells of the new genera- tion is formed on the top of the original cell. This process may be repeated for two or three generations, after which the zoospores will settle on some other object and start fresh " plants ". The HydrodictyacecB are a group of Algse which form immotile colonies. The cells of these colonies resemble the single cells of the forms we have just been con- sidering in producing zoospores or gametes, but undergoing no vegetative divisions. The colony is formed by the joining together in a definite way of the group of zoospores formed in a single cell of the mother-colony. Each of these zoospores then develops into an adult vegetative cell. The recently discovered genus Euastropsis (so called from its likeness to the Desmid Euastrum) is the simplest type of the family. It consists of two mitre- shaped cells joined to one another by their bases. Each cell contains a parietal chromatophore with a single pyrenoid, and a single nucleus. The contents breaks up by successive divisions into 2-32 zoospores, which escape from the cell surrounded by a general membrane. After oscillating for about a quarter of an hour, the zoo- spores become attached in pairs by their anterior ends. Each pair then takes on the characters of the two-celled colony. Pediastrum (fig. 370 ^) consists of a disc of cells, of which the marginal ones are often drawn out into lobes or processes. The chromatophore is parietal with a single pyrenoid; there are numerous nuclei. The formation of zoospores is like that of Euastropsis, but their movement is more lively, and eventually all the zoospores formed in a single cell join together to form a new Pediastrurn-colonY (figs. 370 " and 370 *). Gametes are formed in the same way as the zoospores, but are smaller and more numerous. They escape from the investing membrane, swim freely in the water, and fuse in pairs to form zygotes. From these zygotes new Pediastrum- colonies are produced indirectly, probably by a method like that obtaining in Hydrodictyon. 640 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Hydrodictyon, the Water-net (figs. 370 ^ and 370^) is a beautiful organism forming net-like colonies of cylindrical cells, which are joined end to end, formmg the sides of the polygonal meshes. Each cell may be as much as 1 centimetre in length. A thin layer of protoplasm containing numerous small nuclei lines the wall and incloses a large central vacuole. The chromatophore, or chlorophyll-layer of the protoplasm, contains many pyrenoids, each surrounded by a sheath of starch grains. Fine-grained starch is also scattered through the substance of the chro- Fig. 370.— HydrodietyacesB. 1 The Water-net (Hydrodictyon utriffuloswm), nat. size. 2 a portion of the same magnified 50 diameters. 5 *, and s Formation of zoospores in a cell of Hydrodictyon, showing their union together, and escape as a young net ; x 300. <* Pediastrum granulatum; development and escape of zoospores, the lightly-dotted chambers already vacated. 7 and 8 Zoospores after their escape arranged as a new Pediastrutn plant ; x 240. matophore. This stroma-starch appears in great quantity when growth is checked and assimilation remains active, disappearing again if assimilation is stopped. The pyrenoid-starch, on the other hand, seems to be withdrawn from the ordinary meta- bolism of the cell, since it is formed round each pyrenoid early in the life of the cell, and remains there under all circumstances, unless the cell is on the point of actual starvation in the dark, till the onset of reproduction. When zoospores are about to be formed the pyrenoids together with their starch disappear, and abund- ant stroma-starch appears. At the same time the nuclei multiply a good deal by division, and eventually the whole of the protoplasm divides to form a great num- THALLOPHYTA. 641 ber of zoospores. These zoospores, however, never become free like those of Pedi- astrum, but remain joined together by strands of protoplasm, and after a certain amount of shifting backwards and forwards, come to rest with their ends in contact. Each then gradually assumes the characters of a Hydrodictyon-cell, the young colony eventually escaping from the mother-cell (figs. 370 ^- *■ ^). Gametes are formed in the same way as zoospores, but are smaller and more numerous. The spherical zygote gradually increases in size, and its contents breaks up into 2-5 large zoo- spores, which develop into large cells with pointed processes, the so-called poly- hedra. In the interior of each polyhedron an embryonic Hydrodictyon-net is developed from swarm-spores, and in the cells of this ordinary Hydrodictyon colonies are found. It has been shown experimentally that any Hydrodictyon not above a certain size and age is capable of producing either zoospores or gametes, and that the stimulus to the formation of one or the other is given by external conditions. Thus bright light, fresh water rich in inorganic nutritive salts, and fairly high temperatures, are favourable to the production of zoospores, while the reverse of these conditions, and especially the presence of organic substances, such as sugar, tend to make the cells of a net produce gametes. The conditions favourable to- zoospore -formation are also of course, favourable to active vegetative growth, and no doubt the abundant formation of new protoplasm is a necessary preliminary to the production of zoospores. A slight check to the processes of assimilation and growth is apparently necessary in order to give play to the zoospore-forming forces. Thus, experimentally, a change from a strong solution of nutritive salts to fresh water will induce the formation of zoospores in nets which would simply have gone on growing if left in the nutritive solution. A similar check is probably given by the waning light in many Algae in which zoospores are produced at night. For the production of gametes, on the other hand, an actual reversal of the conditions favourable to growth is necessary. In nature this probably happens when by very active growth the whole of the water of a pool is filled with nets, the inorganic food and oxygen are exhausted, and the normal chemical processes of the cell receive a check. The formation of gametes and zygotes under these conditions is obviously adaptive, since the zygote can, although it need not, rest during several months till the conditions are quite altered. We may therefore conclude that, whereas zoospores are especially designed to multiply and distribute the species, zygotes are intended to preserve it under unfavourable conditions. It is probable that the production of large zoospores and polyhedra is a necessary part of the life-cycle following the germination of the zygote, and cannot be altered by the incidence of difierent conditions. Alliance VII. — SiphonesB. Thallus consisting of a tube, often much branched, and containing many nuclei. This tube is the production of a single cell, but in the more complicated forms is often shut ofF into compartments by transverse septa. Reproduction by zoospores Vol. II. 91 ^42 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. and planogametes, or spermatozoids and eggs: in very many forms no reproductive cells are known. The higher forms of Siphoneae often produce plant-bodies of very definite and characteristic external form, and of considerable size. In some cases these simulate the external form of various higher plants. Families: BotrydiacecB, Phyllosiphonacece, VaucheriacecB, Bryopsidacece, CaulerpacecB, Godiacece, Valoniacece, Verticellatce. Botrydiacece. — Botrydium granulatum is a little plant found growing espe- cially on loam at the damp edges of ponds and ditches. It consists of a club-shaped or balloon-shaped green shoot-portion, about 1-4 millimetres in diameter, continuous with a simple or branched tubular colourless root-portion which is embedded in the substratum. The entire plant consists of a single cell, that is to say, its cavity is •continuous throughout. The wall is lined with a thin layer of protoplasni, which contains many nuclei, and, in the shoot, a net-like chlorophyll-layer. BotrydivMi can reproduce itself in very various ways, according to the incidence of external conditions. The simplest form of propagation is by budding, which takes place under conditions favourable to the ordinary vegetation of the plant. The shoot-portion of a small vegetative plant sends out a process which swells to the size of the mother shoot, puts out a colourless root, and is then constricted off to form a separate plant. But if the plants are covered with water they cannot go on growing comfortably, and accordingly the protoplasm breaks up to form a number of zoospores, each with a single flagellum and two lateral chromatophores. The mass of zoospores is subjected to considerable pressure by the swelling up of a ring-like area of the wall, and the tension becomes so great as to rupture the wall in the centre of the ring and expel the mass of zoospores into the water. On damp soil the zoospores come to rest, and germinate to form new plants. If a zoospore cannot escape from the water it enters on a resting stage, which gives rise to a new plant directly it finds itself on damp soil. Further, if a young plant is exposed to bright sunlight, its protoplasm breaks up into a number of spherical cells, each of which puts on a cell-wall. If now these spherical cells (gametangia) are placed in water, the contents of each breaks up into spindle-shaped, biflagellate gametes, which conjugate in pairs to form zygotes. These zygotes can rest for a longer or shorter time, but if placed on damp earth they at once germinate to form new plants. If, on the other hand, the gametangia are placed in water after being kept for two years they give- rise to biflagellate cells which rather resemble the gametes, but which on damp soil germinate directly to form vegetative plants. Finally, if the gametangia are at once placed on damp soil, their contents does not break up, but the whole gametangium germinates and produces a plant. Supposing a large Botrydium-plant, with a balloon-shaped shoot, be exposed to insolation (bright light) its contents, instead of forming gametangia, travels down into the root, and the protoplasm there divides to form rows of root-cells, each with an independent cell-wall. If a root containing root-cells be placed in water, the protoplasm of each cell breaks up into a number of zoospores: if the root be placed on damp earth each THALLOPHTTA. 643 root-cell sends out a tube which grows into a young vegetative plant : if left in the soil, the latter remaining damp, each root-cell germinates to form a peculiar thick- walled plant called a hypnosporangium which can withstand desiccation, and which in water gives rise to a number of zoospores. The above facts may be briefly expressed by saying that any stage in the life- history of Botrydium tends (1) on damp soil to form vegetative plants, (2) in water to form zoospores or gametes, (3) in dry conditions to form resting cells. There can, then, be no doubt that here, as in the case of Hydrodictyon, we are justified in saying that the form which any given set of conditions tends to produce is adapted to meet those conditions. PhyllosiphonacecB. — Phyllosiphon Arisari is an Alga living in the intercellular spaces of the leaves of Arisarum vulgare in Southern France and Italy. Its thallus consists of a much-branched tube, the parietal protoplasm containing many nuclei and small disc-shaped chromatophores. Unlike the Endosphoerem, the pre- sence of Phyllosiphon has a considerable effect on its host, causing blotches of dead tissue to appear in the leaves. As the Alga can presumably assimilate quite well for itself, this destruction of tissue is probably caused by withdrawal of water from the cells of the host. The only kind of reproductive cells which Phyllosiphon is known to produce are aplanospores (non-motile spores). These are formed by the division of the whole of the protoplasm of the thallus. They are extruded by the swelling of the inner membrane of the tube which bursts the outer membrane just under a stoma, shooting out a jet of mucilage in which the aplanospores are em- bedded. The latter germinate directly, the germ-tubes entering the leaf between two epidermal cells. VaucheriacecB. — This family includes only the well-known and widely distri- buted genus Vaucheria. Different species of Vaucheria grow in brackish and fresh water, both running and stagnant, or in the air in damp situations. The thallus •consists of relatively coarse branched tubes, quite visible to the naked eye. The interior of the tube is lined by a layer of protoplasm containing numerous disc- shaped chlorophyll-grains and many nuclei. The Vaucheria-'pl&ut is fixed to its substratum by short-branched, colourless processes, but, except in connection with the formation of reproductive cells, transverse septa are not formed in the tubes. The gametes of Vaucheria are formed in special organs, known as antheridia and oogonia. The distinction of sex is very strongly marked, the male gametes or spermatozoids being very small oval cells, each with two laterally inserted flagella, while the female gametes or eggs are very large and quite motionless. The antheridia are often spirally curved branches of the main tube, a transverse wall separating the upper part of the spiral, the antheridium proper, from the lower part, which is continuous with the cavity of the vegetative tube. Sometimes, however, the antheridium is straight and club-shaped, and in other cases it may be separated from the main tube by an intermediate cell. The thirty-five species of Vaucheria are classified according to the characters of their antheridia. The oogonia are 644 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. ovoid or spherical, and usually possess a lateral beak. While the antheridium pro- duces a large number of the small spermatozoids, the oogonium gives rise to a single large egg. An account has already been given of the process of fertilization (see p. 68, and figs. 204 « and 204 », on p. 53). Vaucheria also produces peculiar zoospores, whose development and subsequent behaviour have been described on pp. 23 and 24 of vol. i. (see Plate I. a-d). The cUia with which the surface of the zoospore is clothed are arranged in pairs, and in the colourless external layer of protoplasm just below each pair of cilia is situated a single nucleus. This suggests that the zoospore of Vaucheria is to be regarded as really equivalent to a great many zoospores which have not separated during development, each nucleus, with its pair of cilia and a certain amount of chloro- phyll and protoplasm, representing an ordinary zoospore. It has been found that if, as often happens, the zoospore breaks into two during its struggles to escape from the end of its tube, the front part rounds itself off and swims away, behaving just like an entire zoospore. In some species of Vaucheria the zoospores are only partially clothed with cilia and come to rest soon after their escape. In yet others they have no cilia at all, and either escape by dissolution of the end of the tube, or germinate in situ. Here then we have a transition from the formation of active zoospores to the production of passive aplanospores. The occurrence and form of reproduction is here even more entirely under the control of conditions than is the case in Hydrodictyon. The age and size of the plant are no longer factors, since sexual organs and zoospores can be formed on quite short germ-tubes. Cultivation of the plant in a solution rich in inorganic food-salts always gives it a tendency to produce zoospores, but the immediate stimulus to their formation is given by a distinct change in the conditions, just as is the case in the Water-net. In Vaucheria this change is especially necessary, since each zoospore is formed in the apex of a tube, and apical growth must be stopped in order to allow free play to the zoospore-forming forces. The nature of the change, so far as regards the medium, is apparently im- material — it may be a change from running water to still water, or from a damp atmosphere to water, or in the temperature or concentration of the culture-solution, but it is a change from light to darkness which is especially effective. This is quite contrary to the case of Hydrodictyon. The presence of water and a temperature between 3° and 26° C. are absolutely necessary conditions of zoospore-formation. The former condition is obviously adaptive. The formation of sexual organs is specially induced by the replacement of inorganic salts by organic substances {e.g. sugar), just as in the case of the gametes of Hydrodictyon, but antheridia and oogonia, which are formed much more frequently and easily than are the Water-net gametes, often appear in the presence of salts, and when growing in a damp atmos- phere. Light and a temperature above 3° C. are absolutely necessary conditions. The most striking point in the physiology of the reproduction of Vaucheria is the prominent part played by the sexual organs. These are undoubtedly the princi- pal means of propagation the plant possesses, the zoospores, which usually fulfil this THALLOPHYTA. 645 rdle, having fallen to a subordinate position. Only in forms like Vaucheria clavata, adapted to life in rapidly -flowing water, have the zoospores a primary importance. Sexual reproduction is here difficult, and zoospores are always produced in abun- dance on the slightest change of conditions. Bi^ofsidacecB.- — This family consists of marine, mostly tropical, forms. The plant-body of Bryopsis has quite a definite form, and consists of a tube forming the main axis, fixed below by short root branches, and bearing above in acropetal suc- cession a series of branches, some of unlimited, some of limited growth. In these latter are formed swarming cells which are of two sizes, the smaller being yellowish, while the larger have each a green chromatophore. It seems very probable that these are anisogametes, but their conjugation has not been observed. Derbesia is a genus like Bryopsis in many respects, but with special zoosporangia which produce curious zoospores, each with an anterior crown of cilia. Gaulerpacece. — The genus Caulerpa contains nearly one hundred species, which present the most varied external forms, simulating those of many of the higher plants, such as Mosses, Ferns, Mare's-tails, Cactuses, Conifers, &c. Each plant, how- ever, consists simply of a single much-branched but uninterrupted tube, the branches taking the forms of roots, leafy shoots, &c. The tube is supported internally by a complicated system of " beams" of cellulose which run out from the walls. The Caulerpas live mainly in tropical and subtropical seas. They often grow together in large masses, forming great beds of sea-weed, their creeping stems or "rhizomes'' extending many yards. No reproductive cells have as yet been found in any of them, multiplication taking place apparently solely by the breaking off of parts of the thallus, which drift and fix themselves elsewhere. GodiacecB. — Under this name we may conveniently place together a group of forms specially characterized by a thallus consisting of richly-branched tubes, which are interwoven to form a mass of more or less solid character, which possesses in each genus a definite and characteristic external conformation. Thus, Penicillus has a long cylindrical "stalk" fixed below by "rhizoids" and bearing above a head of free dichotomously branching radiating filaments. The older parts of the stalk are strongly incrusted with calcium carbonate. Udotea has a stalk often creeping and branching, bearing flat fan-shaped fronds. Spherical bodies, the nature of which is unknown, are borne on short side branches of the tubes of which the frond is built up. HaliTneda possesses a thallus mainly composed of series of heart- or kidney-shaped segments, which give many of the species the appearance of an Opuntia. There is usually a considerable deposit of calcium carbonate covering the thallus. Eoundish structures, produced in grape- like bunches on the edges of the segments, liberate swarming cells whose behaviour has not been followed. Codium has no well difierentiated stalk or segments; the thallus is very various in form, and is difierentiated into a well-marked pith and cortex, the tubes being mainly longitudinal and loosely packed in the former, while the latter consists of club-shaped closely-packed branches arranged at right angles to the surface. In 646 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. certain branches of the latter swarmers of two sizes are produced. Analogy would lead us to suppose that at least the smaller of these are gametes, but their behaviour has not been observed. Some species of Oodium (e.g. C. tomentosum, with a dichoto- mously branched furry thallus) occur on our own coasts. The remaining Codiaces are largely tropical, but very widely distributed. Valoniacece.—We may include in this family an assemblage of genera whose thallus consists of a branching tube, usually forming transverse walls, but with no interweaving of the branches such as we get in Codiacese. The simplest type is found in Valonia, a form which at first consists of a single club-shaped cell, which produces a whorl of branches at its upper end. Each of these may again produce a whorl of branches of the second order. A group of very beautiful genera form leaf -like structures, the branching of the thallus taking place in one plane. Struvea consists originally of a single cell, which grows apically and becomes divided by transverse walls into a series of segments. Each of these segments bears a pair of branches coming off right and left, and each branch behaves like the main axis. The secondary and tertiary branches thus pro- duced come into contact, fixing themselves one to another by means of curious little rosette-like organs called tenacula, and the whole thus forms a net-like structure with larger or smaller meshes between the branches. The definite usually oval form of the " leaf ", of which the main axis forms the midrib, and the primary branches the principal veins, is due to the latter, after they have attained a certain length, ceasing to produce branches on the side towards the base of the thallus, and at the same time bending forwards and inwards to join the primary branch next in front. Struvea delicatula sometimes lives in the tissue of a Sponge belonging to the genus Halichondria. There is a reciprocal effect on the form of the two organisms, the Alga only taking on the characteristic Struvea-iorm when part of its thallus grows out clear of the body of the Sponge. In consequence of this it was for some time not suspected that this sponge-inhabiting Alga had any connection with Struvea. Neither organism seems to suffer from the association, and there is some evidence for regarding it as a case of true symbiosis. Anadyomene is another very beautiful form resembling Struvea in the construc- tion of its thallus, but with no meshes between the cells. It consists of two kinds- of cells, the more elongated form the "ribs'' of the thallus, the smaller and more rounded make up the intermediate tissue. The genus Boodlea, in which the branching takes place in more than one plane, forms a transitional form connecting these genera with Gladophora, which is usually regarded as belonging to the Confervoidese. Verticillatce. — In this group of the Siphonese the thallus consists of a long cylindrical undivided stalk, fixed below by rhizoids, and bearing above acropetal whorls of simple or branched appendages of limited growth. In some of these appendages gametes maybe produced; zoospores are apparently absent. It includes two sub-families, the Acetabularieae and Dasycladeee. Acetabulariew. — Fertile and sterile appendages distinct. Acetabularia mediter- THALLOPHFTA. 64T ranea: the lower part of the long cylindrical stalk is incrusted with calcium car- bonate, and fixed to its substratum by short irregularly branched rhizoids. The rhizoid-bearing portion is called the foot, and below it there is a thin-walled branched continuation of the stalk, called the basal division. Near the apex of the stalk are borne 1-4 whorls of polychotomously branched sterile appendages, which soon fall off. Above these is an umbrella-shaped whorl of simple appendages in lateral contact, whose cavities are not shut off from that of the stalk. The whole of the upper part of the plant dies off each autumn, only the foot and basal division remaining alive through the winter. In the spring a new shoot is produced. Apparently after several years the contents of each simple appendage of the umbrella (which may now be a centimetre or more in diameter) divides up into a number of oval bodies, each surrounded by a fairly thick wall and containing chlorophyll and starch. These are the gametangia. After their escape by the dissolution of the umbrella, the contents of each divides up to form a number of gametes. Considerable pressure, caused by swelling of the ectoplasm and osmotic tension in the vacuole of the game- tangium, bursts off a lid at one end, and the gametes escape. Conjugation only occurs between gametes derived from distiuct gametangia. Basydadece. — No distinction between fertile and sterile appendages. Dasydadus has a single stalk-cell fixed below like Acetabularia, but bearing very numerous whorls of appendages, which stand so close together as to give the entire plant a resemblance to a minute fox's brush. Each appendage bears a terminal whorl of branches, and in the middle of these is a shortly-stalked, nearly spherical gametangium. The gametes conjugate, but apparently only with those from certain other plants. This fact at first led to the supposition that we had here a physiological distinction of sex in gametes, which in external appearance are all alike. This is, however, quite an unjustifiable and unnecessary assumption. We have no right to predicate sexual differences between gametes which do not show any of the well-recognized characters of male and female reproductive cells. The tendency to avoid pairing with closely related gametes, which we may call exogamy, is quite a distinct phenomenon, not only among isogamous Alg», but also among many of the higher plants, where it coexists with strongly-marked sex. The pheno- mena of self -sterility is an extreme case of this. Neomeris and Gymopolia are two tropical and subtropical genera, whose thallus is very strongly incrusted with calcium carbonate. - The arrangement of the branches resembles that found in Dasydadus, but on the ends of the younger ones hairs are borne, which serve to protect the growing apex of the plant. In Gymo- polia, of which the main stalk branches, and the thallus attains a considerable size, these hairs are borne by simple branches produced on special constricted and uncal- cified zones of the stalk. The apices of the secondary branches are in both genera swollen up, and in close lateral contact, thus forming a continuous surface on the exterior of the plant. The calcium carbonate is deposited as a thick layer under- neath these swollen ends. ^48 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. A whole series of fossil forms from the chalk and tertiary deposits serve to con- nect the various existing types of these and allied genera. Alliance VIII. — Confervoideae. The Algse included under this alliance possess a type of thallus composed of dis- tinct and separate cells. These cells are united usually into linear series, which form branched or unbranched threads. In a few families, however, cell-division takes place in two, or even three, dimensions, resembling the Protococcoidese in the forma- tion of cell-surfaces or cell-masses. Zoospores are produced by nearly all confervoid forms. Aplanospores and akinetes are common. The gametes may be isogamous, or they may show marked sexual differentiation. Families: Ulvacece, UlotrichacecB, Cylindrocapsacece, (Edogoniacece, Clado- phoracecB, Gomontiacece, Sphceropleacece, GhoetophoracecE, TrentepoMiacece, Mycoi- deaceoB, Goleochcetacece. UlvacecB. — This family is usually regarded as the lowest of the confervoid series. It is characterized especially by forming cell -surfaces instead of filaments. Zoo- spores with four flagella and isogametes with two, as well as allimetes, are formed in the group. Monostroma very much resembles Tetraspora among the Protococcoidese, from which we may suppose the confervoid forms to have arisen in evolution. The thallus consists of a single layer of roundish or angular cells. In germination, the zygote divides to form a small hollow sphere, which splits, and extends itself to form a flat plate. At first fixed by rhizoids, the thallus later floats freely in the water. The chromatophore is a parietal plate, covering more or less of the cell-wall, and contains a single pyrenoid. Gametes (which may develop without conjugation), or zoospores may be formed in almost any cell of the thallus. Ulva differs from Monostroma in possessing a thallus of two layers of cells, those of each layer dividing independently of the other. The zygote germinates to form a fixed cell thread, which later on produces the two-layered thallus. Ulva latissima (the Green Laver or Sea-lettuce) is very common on the rocks of our coasts near high-tide mark. It forms large green wavy fronds firmly fixed to the substratum. It is sometimes used as an article of food. Enteromorpha is a large genus, several species of which are common on our coasts, and some in fresh water. The thallus forms a branched hollow tube, the wall of which is one cell thick. Any cell of the thallus may act as the apical cell of a branch. The apex of the branch is solid, but the cells soon round themselves off to form the wall of the tube. The gametes and zoospores resemble those of Monostroma and Ulva. Litterstedtia is a Cape and Australian form much like Ulva, but with a deeply- lobed thallus. Zoospores are produced only in the cells of the lobes. Ulotrichacece. — This family contains several genera common in fresh water, and some marine forms. The thallus consists of an unbranched filament of cells seldom much longer than they are broad. The chromatophore is single, parietal, and of THALLOPHYTA. 649 very various form. Zoospores are formed in most genera. Gametes, where known, are motile and isogamous. Aplanospores and akinetes are very commonly formed, under unfavourable conditions. Ulothrix (fig. 371), the best-known genus, possesses cells of very variable length. The chromatophore, which contains several pyrenoids, is an interrupted cylinder, and may or may not occupy the whole length of the cell. When the conditions are suddenly changed, zoospores or gametes are very readily formed, the former 1-4 the latter 4-32 in a cell. According to the size of the mother-cell and the number of divisions taking place, the size of the zoospores and gametes varies greatly, the '^t£30W^ Hg. S!l.— Ulothrix zonata. 1 Two filaments of this plant. ^ Escape of gametes In packets. ^ Spherical packet of gametes free from the filament. ^ Separa- tion of the gametes, b Gametes swimming about and pairing. 6 Products of pairing of gametes (zygotes) attached to substratum. 7-9 Zygote giving rise to zoospores, lo Two zoospores. ^ x 250 ; 2-10 x 400. (Partly after Dodel-Port.) only constant distinction between them being the number of flagella, which in the zoospores are four, in the gametes two (c/. figs. 371 ^^ and 371 ^). The zoospores or gametes escape from the mother-cell through a hole in the wall. They are surrounded by a bladder derived, probably, from ectoplasm. The swelling of this in the water helps to drag them out of the cell-cavity. The tension of the vacuole of the mother- cell, which is visible among the zoospores or gametes as a smaller bladder, also assists in pressing out the mass (figs. 371 ^' ^' *). The zoospore settles on some solid object, and, after putting out a short root-process from its colourless anterior end, produces a new cell-thread. Some of the gametes develop parthenogenetically, in which case they germinate just like zoospores, but produce usually smaller and weaker plants. The gametes which conjugate (see figs. 371 ^ and 371 ^) produce zygotes which sink to the bottom of the water, and after a period of rest grow into unicellular plants, each of which forms 2-14 zoospores (figs. 371 ^' ^' *> ^). These probably give rise tc ordinary plants. 650 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. The asexual generations of Ulothrix are produced during the autumn and winter, gametes being formed in spring, and the zygote resting through the summer. This course of events differs from that obtaining in most Algse with a similar life- history. Ulothrix seems especially adapted to life in cold water. Hormidium is a genus whose members grow largely on damp earth, tree trunks, &c. The chromatophore is solid, with radiating processes and a central pyrenoid. The cells of some species divide longitudinally, so as to form threads two cells broad. Various other genera are common in fresh water. Cylindrocapsacece. — Cylindrocapsa forms unbranched threads of short cells with very thick walls, each thread being fixed in the young state by a cellulose foot. The gametes show a considerable sexual differentiation. The males are pear-shaped, elongated, yellow, with two flagella at the anterior end. They are produced two in an antheridium, which is formed by the division of an ordinary cell into two or four. The protoplasm of an ordinary cell rounds itself off directly to become an egg, the wall swelling and bursting at one side to allow of the entrance of the spermatozoids. Parthenogenesis also occurs. (EdogoniacecB. — (Edogoniwm has a thallus consisting of an unbranched thread, of rather long cells, of which the basal one is fixed to some solid object. The chromatophore often forms a continuous parietal layer containing several pyrenoids. All the cells, except the basal one, are capable of division. When division is going to occur a rim of cellulose is formed inside the cell close to the upper transverse wall. After the nucleus has divided, and the new transverse wall is formed, the rim is opened, as it were, by a circular cut from without, and the tension of the cell causes a pulling-out of the substance of the rim. The result is the intercalation of a new cylindrical piece of cell -wall in the upper daughter-cell. The young transverse wall now moves up to the lower edge of the intercalated piece of mem- brane. The latter soon acquires the ordinary thickness of a side wall, but the segment of the old cell-wall above the spot at which the rim was formed remains projecting beyond the new piece like the eaves of a house beyond its side walls. After the cell has divided again, another projecting piece will be left in the same way, and the series of eave-like projections so formed are a very characteristic feature of the cells of an CBdog'onmm-thread. The zoospores of (Edogonium are formed singly in the cells of a thread. The entire cell-body, with the exception of the ectoplasm, rounds itself off and escapes from the cell by a split in the wall. An anterior circlet of ciUa surrounds the colourless "mouth place," of the zoospore. In germination, the zoospore fixes itself by the mouth place, sending out short fixing processes, puts on a cell-membrane, and then grows out to form a new thread. The gametes of (Edogonium are sexually differentiated. The oogonium is formed by the swelling-up of the uppermost daughter-cell after a division. The contents round off to form a single large oosphere. Either a round hole appears in the wall, or a circular split is formed at the upper end of the oogonium, the part of the filament above rotating through a few degrees so as to THALLOPHTTA. 651 leave an opening. A certain amount of the protoplasm of the oosphere is extruded at this spot, and forms a sort of canal of mucilage, through which a spermatozoid passes in fertilization. The spermatozoids are produced one or two in an anther- idium, which are short cells poor in chlorophyll, formed by the repeated division of certain cells of a thread. The spermatozoids resemble small zoospores. In some cases they are not produced directly from the cells of a thread, but the latter give rise to special zoospores called androspores, which escape and settle either on or in the immediate neighbourhood of an oogonium. The androspore then germinates, producing a small structure called a dwarf male. This consists of a very few cells, one or more of which become antheridia, and opening by a lid, gives rise to sper- matozoids, one of which pierces the canal of the oogonium, and fuses with the oosphere. The oospore, which is of a red or brown colour, produces four zoospores in germination. Bulbochcete is a genus resembling (Edogonium in its life -history, but consists of a branched thread, only the basal cell being capable of dividing. The cells bear characteristic hairs, swollen at the base, whence the name of the genus. Species of (Edogonium and Bulbochcete, which are both genera of considerable size, are found in our ponds and ditches. The next three families are distinguished from all other Confervoidece by pos- sessing more than one nucleus in each cell. Cladophoracece. — Gladophora is a very widely distributed genus, inhabiting both fresh and salt water. A great number of species have been described, but it is doubtful whether many of the forms are entitled to specific rank. The thallus has a very characteristic habit. It is fixed below by an elongated basal cell, and is profusely branched, sometimes forming a spherical mass. The elongated cells possess parietal chromatophores, which cover the whole cell-wall, and possess many pyrenoids. Sometimes the chlorophyll-layer is separable into distinct angular plates. There are many nuclei in the layer of protoplasm imme- diately within the chlorophyll. Many zoospores are produced in each cell. The nuclei divide a good deal, the pyrenoids disappear, and the protoplasm then divides into a number of separate masses, each of which forms a single zoospore with either four or two flagella. Gametes with two flagella are formed in many species quite like the zoospores. The zygote germinates directly to form a new Cladophora-'plant. The Cladophoracese show on the one hand a transition to the Siphonese, and on the other, through certain genera with unbranched thallus and few nuclei in each cell, to the Ulotrichacese. Oomontiacece. — Gomontia polyrhiza is an isolated form which perforates the shells of various marine molluscs, such as the whelk, the oyster, the mussel, &c. The thallus radiates on the surface of the shell, and sends branches into the sub- stance, gradually disintegrating it. Certain branches become zoosporangia or aplanosporangia; these lose their attachment to the thallus and form fresh rhizoids. The zoospores are pear-shaped, and germinate directly to form a new thallus. i652 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Sphceropleacece. — Sphceroplea annuUna is a curious Alga which appears occa- sionally on flooded fields or in other situations. Its thallus consists of simple threads of cells which are of very various length, sometimes enormously long. The side walls are thin, but the transverse walls are often thick, and both are liable to have curious thick and quite irregular projections of cellulose. The chromatophores form irregular rings at intervals, and contain many pjn-enoids. There are many nuclei, and several variable vacuoles in each cell. All the cells may produce sexual organs, the threads being either moncBcious or dioecious. The contents of the cells which become antheridia become yellow-red, and break up to form a great number of elongated, pointed spermatozoids. These escape through small holes in the wall. In the oogonia the protoplasm divides to form one or two series of spherical oospheres, each with a colourless spot. The oospore has three membranes, of which the outermost is folded so as to give the oospore a star-like appearance. In germination the oospore produces 1-8 zoospores, which have a green posterior and a pale-red anterior end. Each eventually stretches itself to form a spindle- shaped cell, and a multiplication of nuclei and pyrenoids takes place before trans- verse divisions occur and a new Sphceroplea-SiaTaent is produced. Parthenogenesis occurs, but apparently no zoospores are formed other than those produced ia the germination of the oospore. ChcetophoracecB. — This family contains forms with a branching thallus, the branches often ending in fine hairs. The chromatophore is parietal, with one or more pyrenoids. Zoospores with two or four, and gametes with two flagella are produced. Stigeoclonium, Drapamaldia, and Ghcetophora are three genera common in. fresh- water in this country, and all very slimy to the touch. In the first-named genus the thallus is fixed by means of a basal disc of cells, the sole; the branching is simple and irregular, the branches often ending in long multicellular hairs. Drapamaldia shows a marked distinction between axis and appendages. The axial cells are much larger, and at the same time poorer in chlorophyll than those of the branches. The branches come ofF in bunches, and often end in many-celled hairs. The chro- matophore possesses many pyrenoids varying in number according to the size of the cell. Ghcetophora possesses a thallus whose threads radiate and branch in all directions. The whole is surrounded by a mucilaginous investment of considerable firmness, sometimes almost leathery in consistence, so that a GhcBtophora--pl&nt has the appearance of a slimy green ball. There are several other genera belonging to this family, many of them being epiphytic or endophytic. Entoderma lives in the cell-membranes of the Brown Sea-weed Ectocarpus. TrentepohliacecB.— This family differs from the last in possessing no hairs, and in forming its zoospores in special zoosporangia. Trentepohlia is a fairly large genus containing forms mostly living in the air on damp stones and similar situations. The thallus consists of rounded thick-walled cells, and is dichotomously or irregularly branched, partly creeping, partly upright. THALLOPHYTA. 653 The chromatophores are many, disc-like, and angular, with no pyrenoids. The cell- contents is usually much marked with haematochrom. The zoosporangia and gamet- angia are usually terminal, often swollen cells. The gametes and zoospores are much alike. Trentepohlia lolithus, growing on damp stones, is known as the " Violet-stone " from possessing a scent which recalls that of violets. T. umbrina is often attacked by lichen-forming Fungi. T. spongophila inhabits the jelly of Ephydatia (Spongilla) fiuviatilis in a certain volcanic lake in Sumatra. The situation is very advantageous to the Alga, but the Sponge seems to suffer from the piercing of its tissue by the guest. Trichophilus is a nearly allied form which lives in the hollow hairs of the Three-Toed Sloth. MycoideacecB. — These are epiphytic or parasitic forms nearly allied to the two- preceding families, and probably derived from one of them. They form regular discs of cells often attached to the host by much-branched unicellular rhizoids. The disc grows by regular divisions of its marginal cells. Zoospores and sometimes- gametes are formed in all or some of the cells. Ghcetopeltis forms very regular discs of cells on fresh- water plants in Europe. Mycoidea (Mycoidea parasitica) forms discs of cells between the cuticle and epidermal cells of Camellias, Rhododendrons, &c. in the East Indies and South America. In this position it withdraws a good deal of water from the tissue of the leaf, and this leads to the dying of the leaf-cells in a gradually increasing area round the parasite. Eventually a hole is formed right through the leaf, and the Mycoidea-ihaMus occupying a position all round the area of dead tissue continues to increase. Zoospores are only formed in the wet season, at which time alone have they any chance of swarming and germinating on the surface of the leaf. In this position primary (embryonic) discs are formed, many of which die, and others are attacked by Fungi to form Lichens, but some succeed in sending processes through the cuticle and establishing themselves below. ColeoehcetacecB. — This family contains a single small genus, GoleochoBte, which forms radiating, dichotomously branching rows of cells, usually on the surface of other plants in fresh- water. If the cell-rows are in lateral contact, a close disc is- formed (C. scutata), if separate a looser one (0. soluta), or the branching may be rather irregular (C. divergens). The chromatophore is parietal and disc-shaped, and contains a single pyrenoid. Zoospores can be produced in all or only the end cells of the rows. A single one is formed from each cell. In germination a new plant is directly formed. Coleochcete is oogamous, the plants being either monoecious or dioecious. The oogonium is always formed from the end cell of a row. The cell swells and puts out a narrow tube which opens at the end, and extrudes a drop of mucilage. The protoplasm of the swollen basal part then rounds itself off. In the forms with a disc-shaped thallus, the antheridia are produced by the division into four of the members of certain cell-groups. Each daughter-cell (antheridium) then liberates a single spermatozoid. In the branching forms certain end cells form ilask-shaped swellings (antheridia), which are cut off from the mother-cell by transverse walls. 654 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Each antheridium then liberates a spermatozoid. Fertilization has not been observed, but there can be no doubt that a spermatozoid passes down the neck of the oogonium and fuses with the oosphere. Subsequently the neck of the oogonium breaks oif, the oospore puts on a cell- wall, and the neighbouring cells branch so as to form a closely investing cortex of cells round the oospore. The contents of these cortical cells turn red-brown; and after a resting stage the cortex comes off, the oospore having in the meanwhile divided to form a disc of cells, each of which gives rise to a zoospore, which escapes and germinates to form a new plant. The structure of the oogonium, and the formation of a cortex round the oospore, as well as the formation of " carpospores " by the latter, recall the simpler Red Sea-weeds, a group which some authorities consider to be derived from some form like Goleochcete. The indirect formation of new plants by the products of division of the zygote, a phenomenon we have already met with in Pandorina, Hydrodictyon, Ulothrix, (Edogonium, and other forms, is probably the means by which the sporophyte generation of Mosses and Liverworts arose. At first the zygote gave rise at once to spores, but later on a certain amount of sterile tissue was produced in addition, and this formed the body of the sporophyte. Alliance IX. — Conjugataft. Families: Besmidioidece, Spirogyracece, Zygnemacece, MougeotiacecB. This is a very sharply characterized alliance of Green Algse. It is indeed diflScult to determine its affinities. The forms belonging to it are especially charac- terized by never forming zoospores, and by possessing aplanogametes, i.e. gametes which, instead of escaping from the mother-cell and swarming freely, never leave the cavities. of the cells in which they are produced. When conjugation is about to occur the two cells (gametangia), the contents of which will form gametes, approach one another, and their walls come into contact, either directly, or by the putting out from one or both cells of a short cellulose tube. The area of wall at the place of contact breaks down, and the whole or part of the contents of each cell then fuses with the corresponding protoplasm of the other to form a zygote. The chromatophores of the Conjugate, though very various in the different families, are all very different from the types met with among the other Green Algae. I. Besmidioidece. Cell-contents and outline symmetrically arranged on each side of a given median plane which is often coincident with a more or less deep constriction.. Often unicellular. II. Zygnemoidece. Cells cylindrical, without median constriction, always form- ing threads. 1. Spirogyracew. Chromatophores one or more, parietal, spiral 2. Zygmmaaem. Chromatophores two, axile, roundish. 3. Mougeotiacece. Chromatophore single, axile, plate-like. THALLOPHYTA. 655 DesTmdioidecB. — The Desmids are a large family of fresh- water forms numbering over a thousand species. Perhaps their most favourite habitat in this country is the water which collects and stands between mosses and similar plants on imper- vious soils. Many of the Desmids are among the most beautiful of algal forms. The great characteristic of the Desmid-cell is its almost invariable division into two symmetrical halves, often separated by a circular constriction {of. fig. 372). The cell-membrane usually consists in fact of two distinct valves whose edges meet in this median plane. In cell-division these two valves are forced apart, a new cylindrical piece of menabrane being intercalated between them. A transverse wall Fig. S72. Desmids. 1 Micrasterias papillifera. 2 Mwrasterias morsa. * XanthidiuTn aculeaium. 5 Staurastrum i» Staur- 3 Cosmarmm polygonum, furcatum. ' EvMStrum dblongum. i Penium Brebisscmii. « Closterium Lunula. ' Xwnthidium octocorne. astrwm altemans (two views). " Cosmarium tetraophthalmum. " Aptogonum Besmidium. AU the figures magnifled about 200 times. is then formed at the equator, and each half of the new piece of wall gradually assumes the characters of the old half -cell to which it belongs. The actual form of the cell is very various; it is often lobed, and its wall sculptured in various ways. A good idea of the shapes of some of the commoner types can be obtained by an inspection of fig. 372; see also Plate I., i, k. The chromatophores are also extremely various in form. A common type is an axile rod bearing longitudinal plates which radiate in all directions. Each chroma- tophore contains one or more pyrenoids. The cell-membrane is usually perforated by series of regularly arranged, very minute pores which give exit to extremely delicate filaments of protoplasm. The projecting end of each filament is surrounded by a mass of mucilage, and these masses together form a complete sheath covering the entire cell, and sometimes even g56 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. groups of cells (cf. similar phenomenon in Diatoms, p. 626). Many Desmids possess the power of locomotion. Their movement is slow and creeping, and although the means by which it is brought about are by no means fully understood, it has been shown to have a connection with the excretion of mucilage, and to stand in relation with light and gravitation. Some forms (e.g. Glosterium and Penium, figs. 372 8 and 372 ">) have a curious rotating movement, one end being temporarily fixed while the other moves up towards the source of light. Conjugation takes place as already described, the conjugating cells either coming into direct contact, or putting out short tubes. The conjoined tubes are known as the conjugation-canal. The zygote may be formed either in this or m one of the conjugating cells. The membrane of the zygote consists of three layers, the outer one being sculptured in various ways. In germination, the outer, sculp- tured membrane is burst open, and the protoplasm divides to form 2-8 cells which gradually take on the characters of the adult individual. The cells of the thread-forming Desmids (cf. fig. 372 ^^) possess all the characters of the unicellular types, the daughter cells simply remaining together after division (which always occurs in one plane) has taken place. SpirogyraceoB.— This family contains the single genus Spirogyra, many species of which are amongst the commonest Algse in our ponds and ditches. They form green or yellowish-green slimy masses on the mud at the bottom or' floating at the surface of the water. Such a Spirogyra-ma,BB usually contains several species, although the bulk of it is often formed by one. The cells of the different species vary much in diameter, from the delicate S. tenuissima which is only about one hundredth of a millimetre across, to S. crassa, which forms coarse threads as much as one seventh of a millimetre thick. Roughly speaking, the broader the cell, the greater number of chromatophores it possesses. The smaller species possess only one spiral band in each cell (Plate I., 1), the largest as many as eight or nine. Each Spirogyra-cell is a cylinder, in most cases considerably longer than it is broad (though the relation of length to breadth is variable even in one and the same species), with a very delicate layer of protoplasm lining the wall, and a large central vacuole. Each chromatophore forms a band embedded in the protoplasm, and twisting spirally round and round the cell at an approximately constant angle. When there are more bands than one they cross each other at regular intervals, forming a beautiful lattice-work. In some species each band makes several com- plete turns in its course down the cell (fig. 373); in others, it may be inclined at such a small angle with the long axis that it makes less than a complete turn in the length of the cell. In S. orthospira the bands are practically parallel with the long axis, so that they do not follow a spiral course at all, but form straight bands. Each chromatophore may be isodiametric, or, on the other hand, it may form a flattened band. In the latter case its edges are usually irregularly scalloped. A single row of pyrenoids at larger or smaller intervals is found in each chromato- phore. THALLOPHTTA. 657 Each cell contains a single nucleus which occupies approximately the centre of the cell. In the larger species it is suspended in the middle of the vacuole by a number of branching threads, many of which run into the peripheral protoplasm opposite a pyrenoid. This can be particularly well seen in the large 8. crassa. It suggests that the nucleus plays some active part in the function of starch formation. In the smaller species when the breadth of the cell is perhaps not more than double the diameter of the nucleus, the latter often lies in the concavity of, and in contact with, a chromatophore. In this case the protoplasmic threads are not so obvious. AU the cells of a Spvrogyra-thxead are capable of division. After the nucleus has divided, a rim of cellulose is secreted in the equatorial plane of the cell. This is gradually added to from within, till a complete disc of cellulose is formed, sepa- rating the two daughter-cells. Curious folds are always formed on the transverse walls of some species. Each fold forms a circular rim near the periphery of the transverse wall and projecting into the cell cavity. These folds are sometimes completely evaginated, the transverse wall thus having its surface considerably in- creased and becoming strongly concave towards the cell-cavity. This happens especially when two cells are separating from one another, or, when a gamete is formed from a neighbour- ing cell. MultipKcation is often eifected by the break- ing up of a filament into segments consisting of a few cells each which go on dividing and form a new thread. Conjugation takes place in two ways; a zygote being produced either from gametes formed in two neighbouring cells of the same filament, or in two cells belonging to distinct filaments. In the former case a small swelling is formed opposite a septum, a small area of the septum breaks down, and the contents of one cell, rounding itself off from the walls, passes through the aperture thus formed and fuses with the contents of the other cell, also rounded ofi', to form a zygote, which immediately puts on a membrane, and enters upon a resting stage. In the second or "ladder-type" of conjugation (fig. 373, and Plate I., 1), two filaments come to lie side by side, and the contents of some or all of the cells of one thread round off, each cell- wall growing out into a short tube towards a cell of the other thread. Each cell of the other thread then either swells up towards this tube or puts out a similar tube, and the walls coming into contact are absorbed, an open canal (conjugation-canal) thus being formed between the two cells. The con- tents of the first cell then passes through the canal into the cavity of the second, the contents of which has also rounded off, and fusion occurs between the two gametes. The whole of the cells of two filaments frequently conjugate about the Vol. II. 92 Fig. 373. — Spirogyra. 1 Two filaments commencing to conjugate. 2 Formation of zygotes. 058 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. same time, and the series of conjugation-canals thus formed give the appearance of the rungs of a ladder. All the zygotes are formed in one filament, which we may consider as physiologically female, its gametes being relatively passive compared with those of the other (male), which initiate the process, and actively pass through the canals. The relative behaviour of the two threads shows that an influence is exerted by the male on the female cell, the former determining the outgrowth and direction of the tube belonging to the latter, as well as the rounding off of the female gamete. If the female tube is not put out opposite the male, the former bends round to meet the latter, and if the male cell dies in the middle of the process, the female tube goes on growing indefinitely, and the female gamete does not round itself off. The influence exerted is in all probabihty a chemical influence, a pheno- menon which seems to occur in connection with the process of the conjugation of gametes throughout the vegetable kingdom (c/. the remarks on pp. 68 and 413). The chromatophore of the germinating zygote is formed from that of the female gamete alone, the band belonging to the male gamete gradually disintegrating in the zygote. This is an interesting example of the reduction of the specially vegeta- tive portion of the male cell. ZygnemacecE.—The cell of Zygnema differs from that of Spirogyra (Plate I., m) in its remarkable and beautiful star-shaped chromatophores. There are two of these in each cell, occupying positions equidistant from the centre of the cell. Each possesses a rounded central portion, containing a single pyrenoid, from which thicker or thinner processes radiate in all directions. The nucleus forms a bridge between the two chromatophores. Conjugatioa takes place much as in Spirogyra, the zygote being formed either in the conjugation-canal or in one of the conjugating cells. Mougeotiacece. — This family is characterized by the possession of single axile plate-like chromatophores, and by the fact that part of the protoplasm of a con- jugating cell does not enter into the formation of the zygote. The chromatophore, which possesses a single row of pyrenoids, can alter its position according to the strength of the light to which it is exposed. In moderate light the plane of its surface is at right angles to the line of the incident rays, in stronger light it places itself in the same plane as these rays, so that they only fall upon its edge. In very strong light it contracts to form an irregular body in the centre of the cell. In conjugation the cells of two threads either put out tubes, and form spherical zygotes in the conjugation-canals (Mesocarpus-type), or the two cells bend towards one another, and form a four-sided zygote, one side of which occupies the centre of each cell. The wall of the zygote thus cuts off the two ends of the two cell cavities, so that the zygote appears as if it were surrounded by four empty cells {Stauro- spermum-type). These, however, soon break off. Oonatonema forms so-called aplanospores in the following way. A cell increases to double its former length, its contents (chromatophore, &c.) dividing into two parts. A swelling is formed in the middle of the cell, into which the two chromato- THALLOPHYTA. 659 phores and the bulk of the protoplasm move from each end. A wall is now formed round the contents of this swelling. Although the behaviour of the nucleus has not been followed, this certainly suggests a reduced process of conjugation, each half of the cell representing a gamete. Mougeotiopsis is a genus whose chromatophore possesses no pyrenoids. Alliance X. — Charales. Family : Gharacece, the Stoneworts. Are green submerged plants with segmented axes bearing whorls of leaf -like appendages at the nodes, upon which the antheridia and oogonia are borne. The antheridia are spherical and contain a large number of filaments, each cell of which produces a spermatozoid with two long cilia. The oogonium consists of an egg-cell inclosed in five spirally -twisted, tubular cells; on germination the egg-cell gives rise to a simple segmented filament (the pro-embryo) from which the adult form arises as a lateral bud. There are no swarm-spores. Vegetative propagation is by bulbils, detached branches, &c. This alliance, though placed here in sequence with the other alliances of Green Algae, is probably remote from them in actual affinity. The Charales form an isolated and anomalous group, and various views are enter- tained as to their true position. Members of this group occur very commonly in ditches, ponds, &c., and in brackish water. In the Norfolk Broads very extensive growths of these plants occur in the muddy bottom of the Broads, the living plants resting on the decom- posing remnants of former generations; in this way the bottom level is being gradually raised. Ghara fragilis (see fig. 874) is perhaps the commonest species of the group, and is cosmopolitan in its distribution. The plant is some 12 inches high, and consists of axis with whorled leaf -like appendages inserted at the nodes. The axis consists of a number of long cells (the intemodal cells) with which alternate the short nodal cells. The former remain undivided, whilst the latter originate the append- ages and also a number of tubes, which, growing both upwards and downwards, everywhere cover in the internodal cells, forming a sort of cortex. The " leaves " have a structure essentially similar to that of the stem; they are, however, of limited growth (fig. 374^). They bear at their nodes tiny leaflets and the reproductive organs. The oogonia and antheridia occur together in this species (figs. 374 ^ and 374^), the latter below the former. The antheridia are spherical orange-coloured bodies, consisting of eight shields or plates whose edges dovetail into one another; each bears a process (the manubrium) on its inner surface, and each of these manubria bears a tuft of filaments (fig. 374^), in every cell of which a coiled spermatozoid is produced bearing 2 long cilia at the tip (figs. 374® and 374 ^). The shields now dis- articulate and the spermatozoids escape. The oogonia (or amphigonia) remotely resemble the archegonia of Ferns (cf. fig. 346^, p. 472). Each contains a big oval egg- cell inclosed in a sheath of 5 tubes spirally wound around it. The tips of these tubes 660 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. form a crown surmounting the oogonium (fig. 374«). At fertilization the spermato- zoids penetrate between the cells of the crown, so reaching the egg-cell. The whole oogonium is soon detached and remains dormant through the winter. In the sprmg it germinates, pushing out a tube which becomes transversely segmented. This is the pro-embryo. From it, as a lateral bud, the adult sexual plant arises. This process resembles that of the Mosses, in which the leafy Moss-plant arises from 1 A portion of the plant. 2 a piece of the axis with appendages, upon which are inserted the sexual organs. 8 A single appendage, showing the flask-shaped archegonia and spherical antheridia, * A single antheridium. ^ A plate of an antheridium with manubrium and whip-like filaments of cells containing spermatozoids. ^ Several cells from one of the whip-like filaments : the cells in the middle contain each a spermatozoid ; the spermatozoid is escaping from the upper- most cell; the lowest cell is already vacated. ' A single spermatozoid. 8 Archegonium inclosing the egg-cell. 1 nat. size; 2x10; 3x16; *x35; 6x100; 6x300; 'X600; 8x60. the protonema as a lateral bud. But the comparison with Mosses must not be pushed too far, as in Char a there is no sporogonium. There are some 67 species of Chara, of which G. foetida is also very common. Many of them are covered with stiff hairs, and they are for the most part brittle owing to the incrustation of carbonate of lime (c/. vol. i. p. 260). The phenomenon of parthenogenesis in Chara crinita has already been described in detail (pp. 463, 464). In Gha/ra stelUgera ( = Tolypellopsis ulvoides) the nodes of the stem become THALLOPHYTA. 661 swelled out with starch, and assume a stellate form (starch-stars). They serve as organs of vegetative propagation. The other large genus of Characese is Nitella. It is especially distinguished by the fact of its stems and leaves being destitute of cortex. There are 67 species of Ghara and 77 of Nitella. Characeous fruits (Gyrogonites) are met with in large numbers in the lower chalk and in tertiary formations. Only rarely are fragments of the stems, &c., recognized. Alliance XI. — Phseophyceae. Families: Ectocarpacece, Sphacilariaeece, Gutleriaceoe, LaminariaceoB, Fucaceoe. Includes the whole series of the Brown Sea-weeds, essentially characterized by the fact that their chlorophyll-corpuscles include, in addition to chlorophyll, a brown pigment, Phycophsein, which masks the green colour of the chlorophyll. The forms included under this alliance are all multicellular, and range from simple threads of cells to large complex forms showing a differentiation into a root-like attachment- organ, stipe, and expanded leaf-like frond. In several of these larger forms the internal structure almost rivals that of Flowering Plants in complexity. Within the limits of the group we find sexual reproduction, in some cases by the fusion of equivalent motile gametes (c/. p. 50), in others of well-marked egg-cells or sper- matozoids. Fertilization and the complete life-history has been studied in relatively few cases. Ectocarpaceoe. — ^These are mostly filamentous and often branched. Attached to the branches are the sporangia and gametangia. From the former motile zoospores are liberated. From the latter similar bodies — the gametes — are liberated. These possess two cilia, attached laterally^ to the gametes. The process of fertilization has been followed in Ectocarpus silicvZosus. Certain of the gametes come to rest first, and these are approached by a number of other gametes, which swarm around them. Ultimately one of the swarming gametes fuses with the resting gamete and fertilizes it. This process has been thought to indicate the existence of a certain sexual differentiation amongst the gametes, the gamete which comes to rest first being the egg-cell. However, there is no demonstrable structural difference between them. Sphacela/riacecB. — The filaments consist of many layers of cells. Reproductive organs agreeing in the main with the last-named family. Gutleriaceoe. — Mostly branched, ribbon-like forms. The gametangia are arranged in tufts, and the gametes differ in size, but both possess two cilia. LaminariacecB. — Perhaps the most interesting family of the alliance. Many forms are known to liberate motile reproductive cells from various portions of their surface, but the fate of these bodies has not been ascertained. They are large marine forms, some of them attaining gigantic dimensions. Laminaria digitata, which grows in quantities near low-water mark on our coasts (where it forms a regular " Laminaria-zone "), has a tuft of powerful roots holding it to the rocks, a long 662 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. stipe, and a flat, expanded leaf, deeply lobed like a hand. It is often met with two metres in length. The leaf is of a leathery consistency, and the flattened or cylin- drical stalk has a wide parenchymatous cortex and central "medulla", in which run curious tubes (the "trumpet hyphse") which swell out at intervals, the swelling being traversed in a transverse direction by a delicate sieve-plate. The stem possesses a peripheral cambium-like zone, which adds each year a new zone of tissue to the cortex. The stems, which are sometimes found much thicker than one's Fig. 376.— LamlnariacesB, with perforated fronds. 1 Agarum Gmelini (after Biocreux). 2 ThallasiophyUum dathrum (after Poatels and Buprecht). Both much reduced. thumb, show in section a series of rings, reminding one of the annual rings of a dicotyledonous stem. A long ribbon-like form, L. saccharina, is also common on our shores. In other forms the frond is branched and often curiously appendaged at the base. In the two genera represented in the accompanying illustration (figs. 376 1 and 376 ^) Agarum (from the North Atlantic) has a simple, Thallasiophyllum (North Pacific) a compound frond. In both the fronds are perforated or fenestrated. Lessonia (Southern Pacific) attains to tree-like dimensions and is much branched; It has a stem as thick as a man's thigh. But the two most impressive genera are Macrocystis and Nereocystis. The former, which occurs throughout the southern oceans and on the western coast of N. America, consists of a long stalk, sometimes attammg a length of 300 metres, but in thickness not exceeding a penholder To THALLOPHYTA. 663 this stalk are attached a series of long ribbon-like leaves, each of which, just at its point of insertion upon the stem, swells into an air-bladder about the size of a pigeon's egg. Thus the stem, which is attached below, is buoyed up, and the long leaves depend into the water. In structure the stem is not unlike that of a Lami- naria; but it possesses in addition to the medulla, with its trumpet hyphse, a zone containing large sieve-tubes, which resemble those con- tained in the soft bast of a Flower- ing Plant (c/. vol. i. fig. 10*, p. 45, and fig. 125 ^ p. 469). Nereocystis, occur- ring on the W. coast of N.America, consists of a long stalk (attaining to a length of nearly 100 metres), at- tached at its lower extremity and ex- panding above into a huge retort- shaped air-sac, from the surface of which a number of fronds (6-10 metres in length) arise. Like Macrocystis, its stem contains well- marked sieve-tubes. It is used by the Aleutians as fish- ing-line. Of La- minariacese about 90 species have been distinguished (including 30 species of Laminaria). FucacecB. — Includes a number of the larger common sea- weeds. They are characterized — like the last family — by a segmentation into a well-marked shoot and organ of attachment. The former is usually flattened and branched, and often bears air-bladders. Reproduction is by means of spermatozoids and non-ciliated egg-cells, which arise in flask-shaped hollows (conceptacles) on definite portions of the shoot or frond. Asexual reproduction by detachment of fragments. ■a&iC: Fig. 37e.—Fucus vesiculosus. 1 Vertical section through a female conceptacle. 2 a single oogonium from the conceptacle surrounded by sterile hairs, s A detached oogonium containing 8 egg-cells ; the inner lamella of the wall is much swollen. * Liberation of the egg-cells. ix50; 2,3,4x160. (After Thuret.) 664 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. The Wrack-genus Fucus forms an exceedingly conspicuous feature of our sea- shore flora The shoot is flattened and ribbon-like, branching in one plane and attached to stones, &c., by a sucker-like disc. In many species there is a midrib, as also air-bladders. The tips of many of the branches are studded with little rugosi- ties — really indicating the presence of pear- shaped hollows, the con- ceptacles. From the lin- ing of these conceptacles project the oogonia and branched filaments bear- ing antheridia respective- ly. As a rule the male and female organs occur on distinct plants, though in some species the anther- idia and oogonia occur side by side in the same conceptacle. The struc- ture of the sexual cells and the act of fertiliza- tion (which occurs out- side the mouth of the conceptacle) have been fully described on pp. 51, 52; they are represented in flgs. 376 and 377. Sixteen species of Fucus are distinguished, and they occur for the most part in the seas of the northern hemisphere. Fucus vesiculosus and F. serratus are the com- monest. Several other genera are represented in Britain, Pelvetia, Ascophyllum, Cystoseira, Halidrys, and Himanthalia. The last-named genus is altogether peculiar, and consists of a top-shaped body at- tached by its pointed end, whilst from the upper surface of the "top" arise several ribbon -like outgrowths which branch repeatedly and attain to a length of several metres. It is upon these ribbons that the conceptacles are borne. Of exotic forms a few may be mentioned. Burvillcea from the southern seas resembles a thick, fleshy Laminaria in habit; from the summit of its thick stipe arise a Fig. 377 Fucus vesiGulosus, I Vertical section througli a male conceptacle, 2 A portion of one of the shrubby, branched hairs bearing antheridia. s Spermatozoids escaping from the anther* idia. * Spherical egg-cell with spermatozoids attached, 1 x 50 ; 2 x 160 ; 8, < X 360, (After Thuret,) THALLOPHYTA. 665 number of irregular fleshy lobes, which are produced into branching, whip-like filaments. Its tissues possess a curious honey-combed structure. It is stated that in Chili, &c., D. utilis is used as an article of food. Sargassum is distinguished by its high differentiation. It has cylindrical stalks bearing leaf -like appendages, and little stalked spherical air-bladders, and receptacles for the sexual organs. Some 150 species of this very varied genus are known, scattered over the warmer zones of the world. Particular interest attaches to the Gulf -weed (Sargassum bacciferum, fig. 378) which forms the chief component of the floating masses of Sargassum in certain regions of the Atlantic. The Sargasso Sea has received its name from the enormous amounts of this floating weed which are met with there. It occupies an area in the Atlantic perhaps equal to that of the continent of Europe. There are two main accumulations, the larger south-west of the Azores, the smaller situated between the Ber- mudas and Bahamas, whilst connect- ing them is a narrow belt. The exact nature of these accumulations is not ascertained. According to one view the Gulf -weed actually lives a pelagic life, growing and multiplying in this huge eddy in mid-ocean, and is thoroughly adapted to its special environment; whilst, on the con- tending hypothesis, the vegetation of the Sargasso Sea is purely ephem- eral, does not reproduce, and is con- stantly renewed by ocean currents, which bring with them countless fragments forcibly torn by tempests from the shores of Florida and the Bahamas. It is further alleged that the floating Gulf -weed is met with only in a condition more or less unhealthy (moribund) and in various states of decomposition. The weak point in the latter hypothesis is the lack of convincing evidence to show that 8. haccifermti grows attached in the region of the West Indies, &c., in quantity sufficient to supply the Sargasso Sea. Of another species, 8. vulgare, there is plenty, but this is not the prominent constituent of the Sargasso Sea— indeed a trained algologist, in passing recently through this sea, examined samples amounting to more than a ton, but it was only 8. bacciferum he found. Here, evidently, is still matter for the leisured naturalist. Over 300 species of Fucacese (including 150 Sargassums) have been distinguished. fig. 378.— A branch of the Gulf-weed, Sargassum lace\ferwm, with leaves and air-sacs. 666 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Alliance XII. — Dietyotales. Family: Dictyotaceoe. A small group of Brown Sea-weeds distinguished by the fact that both egg-cells and spermatozoids are destitute of cilia. The sexual cells are contained in club- shaped vesicles, inserted in tufts on the surface of the plants. Asexual reproductive cells ("tetraspores") are formed in sporangia in fours. They include the common Dictyota dichotoma and the beautiful iridescent fan-like Padina pavonia. Alliance XIII. — Floridese, Red Sea-weeds. Aquatic, for the most part marine, plants, which contain in addition to chlorophyll a red or purple pigment, phyco-erythrin; the pigment, as in the brown sea- weeds, is confined to definite corpuscles. Reproduction is by means of asexual spores (teti-aspores), and sexually by non-ciliated spermatia and procarps. With the exception of Batrachospermum, Lemanea, and one or two other genera, the Floridese are marine organisms and inhabit on the whole a deeper zone than any other sea- weeds. Several views prevail as to the significance of the red pigment. As has been already indicfited (vol. i. p. 390) the rays of light, useful in ordinary chlorophyll-assimilation, are soon absorbed, as white light traverses con- siderable strata of water. Such light as penetrates some distance from the surface is preponderatingly blue, and, as is now known, such rays are actually destructive of vegetable protoplasm. It may well be then that the red pigment serves to screen the protoplasm from the action of these rays, permitting the chlorophyll to make use of such of the red rays as filter to it; or — what is more probable — the red pigment is itself an assimilating pigment, either directly absorbing the blue rays and allowing the protoplasm of the chlorophyll-corpuscles to use their energy for building up complex food-materials, or indirectly (as indicated at vol. i. p. 390) by altering their wave-length they are made serviceable to the chlorophyll-corpuscles. The Floridese exhibit an enormous variety of form, and almost all of them are attached. There are the delicate cell-filaments of the Callithamnions, the corticated Polysvphonias and Geramiums so common on our coasts, the fleshy cylindrical Gracilarias and Pdlyides, the flat and lobed Ohondrus and Gigartina, the leathery Iridcea, and a host of others. One of the most beautiful of British genera is Delesseria, with its creeping stalk and crimson leaves with midribs and veins. In some species the leaves are entire, in others their margins are sinuous and lobed. Of all red sea- weeds perhaps the Australian Clavdea holds the palm for beauty with its large latticed, rose -pink fronds. Certain groups, Corallina, Melobesia, Litho- thamnion, &c., are encrusted with large amounts of carbonate of lime, and build regular banks and reefs under the sea. In all there are some 280 genera and 1800 species of Floridese. Reproduction by means of asexual spores is a common phenomenon in the group. These spores, though not invariably, are most frequently formed in clusters of four. THALLOPHYTA. 667 and are termed tetraspores. In the simpler forms they arise in little projecting sporangia, in other cases they are on specialized branches or embedded in the sub- stance of the frond. The sexual organs, however, are very remarkable, and diifer from those of other Thallophytes. The male cells arise from clusters of cells (antheridia) at the tips of branches or in groups upon the surface of the frond. Each antheridium liberates a single, non-ciliated male cell termed a sper7)iatium. The female organs or pro- carps consist of a filamentous receptive organ, the trichogyne, and a basal part, the carpogonium (cf. figs. 204 '' and 204 *, p. 53). Besides the carpogonium there are auxiliary cells, to be described directly. The process of fertilization has been followed in detail in Nemalion since the earlier portions of this volume passed through the press, and we now know that the suggestion of an osmotic fertilization in the Floridese {cf. p. 60) is erroneous. The spermatium becomes attached to the trichogyne, and at this point the wall of the trichogyne is absorbed. The nucleus of the spermatium enters the trichogyne and travels down it to the carpogonium at the base, where it fuses with the carpogonial nucleus. The carpogonium now divides, giving rise to a mass of spores (the carpospores). In other cases the process is less simple. As before there is a trichogyne and carpogonial cell, but associated with the latter a number of auxiliary cells. The carpogonium when it is fertilized does not give rise at once to a mass of spores, but enters into a process of conjugation with these auxiliary cells, and from them the carpospores arise. This conjugation may be an immediate fusion, or may be brought about by the instrumentality of special tubes. Thus we may suppose the fertilizing influence to be transmitted. In some cases (as in Dudresnaya) this may take place over considerable distances, from branch to branch. In the simpler forms the procarps are modified branches projecting freely into the water; in the more fleshy and ribbon-like forms they are sunk in hollows on the thallus and often associated with a large number of auxiliary cells. The trichogyne projects into the water through a small pore in the portion of the surface which roofs over the procarp and auxiliary cells. As a result of fertilization of the carpogonium and its conjugation with the auxiliary cells, a large mass of carpospores arises, which raises up the surface like a blister. Fossil remains of red sea- weeds occur under the name of NuUipores. These are the calcareous incrustations of the Corallinas, Melobesias, Lithothamnions, &c., men- tioned above. They occur in both the secondary and tertiary rocks. The Leitha limestone, largely used for building purposes in Vienna, comes from extensive nulli- pore banks in the Leitha Mountains, south-west of Vienna on the Hungarian frontier, and, just as in Paris many of the finest buildings are constructed of the consolidated calcareous remains of Foraminiferee, so in Vienna are the incrustations of certain red sea-weeds put to this purpose. 668 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. aass v.— rUNGI. Parasitic or saprophytic plants destitute of chlorophyll and for the most part possessing a mycelium. Sexual reproduction known and generally admitted in the Phycomycetes only. Asexual reproduction by means of spores and conidia. Sub-class I.— Phtcomtcetbs. Mycelium for most part unicellular, tubular, and branched. Sexual reproduction both by conjugation of equivalent cells and by egg-cells. \Co\wit Fig. 379. — Chytridiaceae and Ancylistacese. 1, 2j 3 Lagenidiu/m Habenhorstii, parasitic upon Spirogyra. *, s Polyphagits EugUms. * Rhizidiomycev apophysatits, parasitic on an oogonium of Saprolegnia. Alliance XIV. — Oomycetes Families: Peronosporeoe, SaprolegniaceoB, Ghytridiaceoe, Ancylistacece, EntoTnophthorecB. Mycelium often very slightly developed; asexual reproduction by means of swarm-spores; sexual reproduction by egg-cells. On the whole in this group we are dealing with Fungi which very nearly resemble the Algse of the Alliance Siphoneae (e.g. Vaucheria) both as regards the structure of their mycelium and mode of repro- duction. A loss of sexuality is to be noted in many members of this group. THALLOPHYTA. 669 Peronosporece.— Are mostly parasitic upon Flowering Plants, and the cause of many destructive diseases. They estabhsh themselves by means of a branching, tubular, non-septate mycelium which penetrates the intercellular system of the host- plant (c/. p. 56). They propagate asexually by means of unicellular sporangia borne on branched hyphse which project from the stomates, &c., of the host (c/. fig. 38V); these sporangia (or spores as they are sometimes termed) liberate on a moist sub- stratum a number of swarm-spores (figs. 381 ^-I^'*') which originate new plants. Fig. 380.— Swarm-spores in Saprolegniacese and Cbytridiacese. 1 AcMya prolifera. 2, s, 4 Successive stages of swarm-spore-formation in Achlya prolifera. ^ Chytridium OIUl, parasitic on the oogonium of the Alga (Edogonium ; development of swarm-spores, 6 Saprolegnia laetea. ? Development of swarm-spores in the same. (Partly after De Bary and Pringsheim.) ix20; 2, =,4x400; 6x300; «xlOO; i'x300. Sexual reproduction also takes place by the formation of oogonia and tube-like antheridia. The latter become attached to the former (fig. 381^), and, putting out fertilizing tubes which penetrate to the egg-cell within the oogonium, transmit their spermatoplasm. No spermatozoids are differentiated, but the spermatoplasm travels en masse. The fertilized egg-ceU enters on a resting stage, and when it germinates may either give rise to swarm-spores (e.g. Cystopus) or grow at once into a new plant (Pythium, Peronospora). To Phytophthora infestans is due the well- known Potato-disease. The Fungus attacks the foliage and reproduces abun- dantly asexually. Later, its mycelium penetrates to the tubers and passes into a dormant state there. Consequently when stored these potatoes go bad, and if ^70 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. used for planting are liable to reproduce the disease next year. Sexual reproduc- tion is as yet not certainly known to occur in the life-history of the Potato-disease Fungus Phytophthora omnivora and Pythium de Bm-yanum attack and destroy many young seedlings, causing them to "damp off". Various species of Peronospora are known which attack large numbers of cultivated plants. P. parasxUca works havoc amongst the Cruciferse; P. viticola ( = Plasmopara viticola, fig. 381) has been referred to as a deadly disease on the Vine; P. Vicim on various leguminous Fig. 381.— The False Vine-mildew, Peronospora viticoUi. > A bunch of grapes attacked by the False Vine-mildew. 2 Spores or conidia on branched hyphse projecting from a stoma of a Vine-leajt. « Fertihzation. * A single oonidium. = Swarm-spores arising within the conidium. « A single swarm-spore. 1 nat. size ; 2 X 80 ; »-6 x 350 ; = x 380. (>-« after De Bary.) crops; P. Hyoscyami on Tobacco-plants, besides which there are many others. The effects of Cystopus candidus have already been mentioned (p. 525); it is common upon cruciferous plants. About 100 species of Peronosporese have been distinguished. Saprolegniacece. — Are all aquatic and for the most part saprophytic; a few are parasitic on fish. In structure they much resemble Peronosporese. Swarm-spores are liberated in large numbers from the enlarged ends of the hyphse (fig. 380). Sexual organs arise much as in Peronospore^, but although the antheridia develop fertilizing tubes in several instances, no case has been observed in which an actual transfer of spermatoplasm occurs. As a rule several or many egg-cells are produced THALLOPHYTA. 671 in the oogonia (fig. 382), and these germinate parthenogenetically. Although the sexual organs are still preserved in this group their function has been lost, leading the way to their complete atrophy in many other groups of Fungi. Many members of the group occur upon the bodies of dead insects and fish {e.g. various species of Saprolegnia, Achlya, Aphanomyces). Aphanomyces phycophilus is parasitic upon Algse (e.g. Spirogyra) and Saprolegnia Ferax upon Salmon, &c. About 45 species have been distinguished. GhytridiacecB. — Small parasitic aquatic Fungi whose mycelium is almost entirely wanting; they produce characteristic spor- angia which liberate uniciliate swarm-spores. Polyphagus EuglericE develops a very delicate mycelium, of which the branches become attached to Euglena-cells (fig. 379^, two Euglenas are thus attacked), and from the central portion of the mycelium arises a sporangium from which uniciliate swarm- spores (fig. 379*) are liberated. These in turn germinate, develop threads, and entangle fresh Euglenas. Rhiziddomyces attacks the oogonia of Saprolegnias, sending a branching mycelium into their interior (fig. 379®) and producing a pear-shaped sporangium at the surface which liberates numerous swarm- spores. Ghytridium Olla (fig. 380*) pro- duces its sporangia on the oogonia of (Edo- gonvwm; its swarm-spores escape by the removal of a lid. Over 180 species are known. AncylistacecB. — Very nearly resemble the Chytridiacese in mode of life, but difier in exhibiting sexuality. Lagenidm/m Rahen- horstii attacks cells of Spirogyra, &c. The spores become attached to the Spiro- gyra-cells, and penetrate the wall (fig. 379^) by means of a tube which branches about within, forming a lobed, irregular body (fig. 379 ^), which may open at the outside, liberating swarm-spores (fig. 379^), or sexual organs may arise inside and fertilization take place. 14 species have been distinguished. EntomophthorecB. — A group of forms almost all of them parasitic on insects. They are adapted to non-aquatic life, and connect the Peronosporese with the Zygomycetes. The tubes of these parasites having eflTected an entrance into the body of an Fig. 382. — Achlya lignicola. 1 Oogonia witli antheridia and fertilizing tubes ; no fer- tilization liappenSi however. 2 An oogonium con- taining egg-cells which have put on cell-walls with- out heing fertilized, land 2x400. (After Sachs.) 672 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. insect, bud and sprout there with great activity. Entomophthora radicans commonly attacks caterpillars of the Cabbage-white (Pieris Brassicce). Having spread through its interior, it sends out tufts of hyphse on the ventral side (fig. 383 1), thus rooting the caterpillar to the substratum. It now develops hyphae all over the body wrapping up the caterpillar like a mummy (fig. 383 ^). At the tips of these hyphse conidia are abstricted and shot off" to some little distance (figs. 383 3- *. 6). A con- ^;^?*S« Fig. 383.— Eatomophthoreee: Entomophthora and Empusa. I A caterpillar of tlie Cabbage-white Butterfly attacked by EntoTnophthora radicans. 2 The same caterpillar fully invested by the Fungus. 8 Tufts of conidia-bearing hyphse from the baclc of the caterpillar. * Conidia separating from the tips of the hyphse. 6 Disarticulated conidia. 6 a Fly attacked by Empusa Muscce. f Hyphse of Empusa Muscor, from the tips of which conidia are being shot off. ' Conidium inclosed in sticky mucilage. 1, s, = nat. size; Sx80; *, ', 'x300; 8x630. (After Brefeld.) jugation of branches sometimes occurs, whilst in other cases fruits are formed parthenogenetically. Empusa Muscce produces a disease common amongst flies in the autumn. The Fungus having effected an entrance into the body of a fly gradually fills it up with its sprouts. In due time tubes penetrate the surface and develop conidia at their extremities (fig. 383 '^). These are shot off" as in the last case, and one may often see flies stuck to the window-pane in autumn surrounded by a halo of these conidia (fig. 383 ^). About 80 species of Entomophthorese are known. THALLOPHYTA, 673 Alliance XV. — Zygomycetes. Families : MucoracecB, Mortierellece. Are mould-like saprophytes with a much -branched, non-septate mycelium; sexual reproduction by conjugation; swarm-spores never met with. I'he com- mon Mucor Mucedo (fig. 384) may be regarded as typical of this group. Its mycelium establishes itself on the substratum, and develops long-stalked sporangia at various points on its sur- face (fig. 3841). In each sporangium (fig. 384 2) numer- ous spores are contained, and these can germinate, producing new mycelia on a suitable substratum. A conjugation of branches of the mycelium leading to the production of a fruit or zygospore (fig. 384 ^) occasionally takes place, but the sexual method of repro- duction is much commoner in other members of the group (of. pp. 53, 54). The zygospore (or zygote) is invested in a strongly thickened membrane and can remain dormant for a considerable period. The hyphse in many of the Mucors can break up into continuous chains of cells which disarticu- late and propagate the plant; these oflfshoots are known as chlamydospores or gemmae. It often happens amongst the Mucors that although the conjugating branches are produced, they do not conjugate but each produces a fruit parthenogenetically. These, in contradistinction to zygospores, are called " azygospores ". Or, as in Mucor tenuis, the "conjugating branches" no longer arise in pairs but isolated; these also form azygospores. Thus in this group, as in the Saprolegniacijse of the alliance Oomycetes, we note a tendency for sexuality to become obsolete (cf. p. 670). A good deal of variety exists in the Mucoracese in the arrangement of the sporangia. In Thamnidium the sporangial branch ends in a large sporangium, and in addition bears laterally a number of tiny sporangia (sporangioles) containing four spores each,, whilst in Chostocladium there is a further reduction, and the sporangioles contain Vol. II. 93 Fig. 384.— Jfucor. 1 Mycelium with stalked sporangia ; x 40. 2 a single sporang ium ; x 2 8 A zygospore produced by conjugation ; xl60. 674 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. but a single spore. Whilst the bulk of Mucoracese are saprophytes on animal ex- creta, &e., a certain number are parasitic on Mucor itself as well as other Fungi. The MortierellecB, though in general resembling Mucoracese, are distinguished by the fact that their zygospores become invested in a plexus of mycelial hyphse which form a sort of pericarp around the fruit. The base of the sporangial hypha also is invested in a sort of bird's-nest. This condition is of interest as it leads on to the more complex fruits of higher Fungi. The Zygomycetes include 125 species. Sub-Class II. — Mesomtcetbs. Mycelium multicellular; asexual reproduction alone is known by means of spores (not limited in number) developed in sporangia; or by conidia. They are regarded as occupying an intermediate position between the lower Fungi and the two large groups of higher Fungi, the Ascomycetes and Basidiomycetes respectively. Alliance XVI. — Hemiasei. Tube-like sporangia containing an unlimited number of spores. This alliance leads on to the Ascomycetes. Families : AscoidecB, Protomycete^, Thelebolece. These are mostly simple forms of parasites characterized by the indefinite number of spores contained in their sporangia. Some of them produce chlamydo- spores freely. Thelebolus is interesting in that its sporangium is inclosed in a cortex and may be compared with Mortierella of the Zygomycetes on the one hand, and with the corticated Ascomycetes (carpo-asci) on the other. There are about 20 species belonging to this alliance. Alliance XVII. — Hemibasidii. Parasites with a septate mycelium, which forms numerous chlamydospores. From these spores a promycelium is developed on which conidia (sporidia) are produced. No sporangia are formed. This alliance is thought to lead on to the Basidiomycetes. Families: Ustilaginacece, Tilletiacece. These are all parasitic forms, and are known as the Smuts. The mycelium grows in the living tissues of the host, and concludes its development with the production of chains of chlamydospores (cf. p. 673), which are provided with a thick membrane, and are usually dark in colour. It is in respect of this character that the name " Smut " has been given. Very often these chlamydospores are produced in connection with the fruiting organs of the host-plant (various Grasses, &c.). The THALLOPHYTA. 675 chlamydospores only germinate after they have been distributed, and in a very characteristic manner. A short tube is formed, and from this (the promycelium or basidium, of. p. 674) conidia ( = sporidia) are abstricted. In the Ustilaginaceae this basidium is septate and the conidia are abstricted laterally ; in the Tilletiacese the basidium is non-septate, and the conidia are abstricted as a crown at the apex. This is the main difference between the two families. The conidia, which are budded off from the promycelium, have the power of budding in a suitable soil with enor- mous facility, forming new conidia, and this may be continued for a considerable time. In this way the soil becomes thoroughly infected, and should a young seed- ling host-plant arise, it is almost certain to be penetrated and invaded by one of these germ-like conidia. Ustilaginaceae. — These are the Smut-fungi properly speaking. Ustilago segetvmi ( = Ustilago carho), the common Smut of Wheat, Barley, Oats, &c., has been very fully investigated. The cereals in question become infected when quite young by the tiny conidia, which put out delicate germinal tubes and penetrate the young growing tissues. Should the young plants escape infection at this stage, they are safe against the parasite, which cannot penetrate the hard, adult tissues. The tubules of the Ustilago penetrate from cell to cell, and take up their position at the growing point. Here they keep pace with the growth of the host, but the presence of the parasite is not manifest externally until the grain begins to ripen. As the grains begin to swell, the fungus increases rapidly, and occupies the greater portion of their substance with its mycelium. It thrives, of course, at the expense of the food which would otherwise have served in forming the embryo and reserve of food-material in the seed. Finally, the Fungus resolves itself into masses of black chlamydospores — the " smut " — which appear between the glumes of the ear. These chlamydospores, as already stated, after a period of rest, produce their promycelia and bud off conidia, which in turn bud off other conidia, and so the ground is kept infected. It was formerly a matter of mystery how the Fungus got into the host-plant, as traces of it were not recognized till it burst out in the ripening ear in the " smut " stage. It is only comparatively recently that the period of infection has been recognized, and the fact that the mycelium grew up, so to speak, with the host-plants was fuUy realized. Ustilago Maydis produces hypertrophied growths on the Maize, and has been alluded to in a former chapter (cf. p. 624). Ustilago violacea attacks the stamens of many Caryophyllacese, developing its chlamydospores in place of pollen in the anthers. Ustilago longissima is very common in the leaves of the aquatic Grasses Olyceria aquatica and G. fiuitans; it produces its chlamydospores as long, parallel, black lines. More than 60 Ustilagos have been distinguished ; a large number of them attack cereals and other grasses. Tilletiacece. — Have on the whole a life-history resembling the TJstilaginese ; their main point of difference residing in the fact that the promycelium is unsegmented, and produces its conidia in a crown at the apex. Tilletia Tritici attacks wheat. In Urocystis the spores are clustered into little balls, the accessory spores forming a 676 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. sort of cortex around the central spore which germinates as in Tilletia. Urocystis ViolcB is common on the leaf -stalks and blades of Violets. About 100 species of Tilletiacese have been distinguished. , Sub-class III.— Mtcomtcetes. Mycelium many-celled. Reproduction asexual, either by spores of limited num- ber in asci, or by conidia of limited number on basidia. Alliance XVIII. — Ascomycetes. Parasitic (or saprophytic) Fungi, producing spores in special tubular sporangia, the asci. These spores, termed asoospores, do not exceed 8 in number. In addition to the asci there are subordinate conidial fructifications. Families: Exoascaceoe, Perisporiacece, Pyrenomycetes, Discomycetes. As stated, this alliance is characterized by the possession of sporangial fructifi- cations, consisting of tubular asci containing as a rule 8 ascospores. A number of such asci are shown in fig. 388^ with sterile supporting hyphse, the paraphyses, between them. In addition to these ordinary and typical ascus-bearing fructifi- cations, secondary fructifications producing conidia or chlamydospores are largely met with; consequently many of these Fungi appear under several forms in addition to the ascus stage. The Ascomycetes are divided into families according to the 'characters of the ascus-fruit. In the Exoascacece the asci are borne freely and exposed on the mycelium; in the other three families in special receptacles. In the Perisporiacece the group of asci is contained in a nut-like or tuber-like body; in the Pyrenomycetes the asci are produced in special pear-shaped excavations in a solid tissue which open by a pore to the exterior; in the Discomycetes the receptacle forms an open plate or cup, or sometimes an irregular body covered with the layer of asci. Exoascacece. — This family comprises the parasitic genera Exoascus and Taph- rina, the gall-like deformations caused by which were so fully described on pp. 524 and 527. The tissues of the host-plants are penetrated by the mycelia of these forms, and the asci are produced over the surface of the parts attacked, generally bursting through the cuticle of the epidermis. Each ascus contains 8 spores, but in many species many more than this number are often found; this is due to the fact that the 8 original ascospores begin to bud whilst still within the ascus, producing a large number of secondary spores (conidia). Exoascus has a perennial mycelium, and to its species are due a large number of the " witches'-brooms " and other hypertrophies. E. Pruni produces the " pocket-plum " (cf. p. 524); E. Alni-incanai the curiously altered Alder catkins represented in fig. 358 ^ (p. 523); E. Carpini the birds'-nest-like witches'-brooms of the Hornbeam; E. deformans the "curl" of Peach-leaves. Taphrina is largely a leaf -parasite, and its mycelium is not peren- nial. T. carnea causes blisters on the leaves of the Birch. THALLOPHYTA. 677 About 50 species of Exoascacese have been distinguished. Perisporiacece. — Here are included all forms in which the asci are inclosed in fruit-like bodies, i.e. the Mildews, Moulds, and Truffles. The Mildews are chiefly leaf-parasites, and spread their mycelium over the surface of the foliage and send their suckers (or haustoria) into the epidermal cells (c/- -fig- 32 2, vol. i. p. 165). In due time they produce their ripe ascus-fruits like tiny black grains scattered over the surface of the leaf. Each of these fruits con- sists of a shell-like investment inclosing one or more asci, each of which contains 8 spores. Sphcerotheca is the simplest form, there being but a single ascus in its fruit. S. pannosa is the Rose-mildew, and S. Gastagnei the Hop-mildew, a very destructive parasite in Hop-growing districts. Erysiphe has several asci in its fruits, and includes the well-known E. TucJceri, the true Mildew of the Vine (to be distinguished from Peronospora viticola, figured on p. 670, which is the false Mildew). A tropical genus of leaf -parasites allied to our Mildews is Meliola, which is widely distributed. The Moulds include several exceedingly common saprophytes which make their appearance on the most various sorts of organic matter. The Blue Moulds, which occur on jam, bread, leather, &c., are probably the best known and most commonly recognized of all the smaller Fungi. These forms spread their mycelia over any suitable substratum, and penetrate it with their hyphse. Their usual fructification is not the ascus-fruit, but clusters of conidia, borne on erect hyphse, which stand out from the mycelium. Two common Moulds are represented in fig. 193, p. 18. Asper- gillus niger (figs. 193* and 193^) bears its conidia in spherical tufts on enlarged aerial hyphse. The swollen end of an aerial hypha is densely set with cylindrical cells, from which the conidia are abstricted one after another. Penicillium crusta- ceum (figs. 193 * and 193 ^) is very similar, but here the conidia are borne on a hypha which branches near its extremity like a compound umbel. Another form, Eurotium, is shown in fig. 385 ^, p. 679. The ascus-fruits of these Moulds are not very conspicuous, nor are they always very plentifully developed. They arise on the mycelium after the conidial stage is over, and when ripe are about the size of small shot. They commence by the entwining of certain hyphal branches {Peni- cillivmi, fig. 193 ^ p. 18; Eurotium, fig. 385 ^ p. 679) which have been regarded as representing male and female organs (cf. p. 60). That fertilization takes place is strenuously denied by many modem mycologists, and the sexual nature of the entwining hyphae is not universally recognized. Be this as it may, the result of the process in question (which also takes place in the Mildews) is the formation of a sinuous hypha, which becomes embedded in a dense cortical sheath which grows up from the mycelium close by the place of origin of the entwined hyphee. This is the young fruit-body; that of Eurotium is shown in section in fig. 385^". From the central hypha numerous asci, each containing 8 spores, are ultimately developed (figs. 193' and 385^^). The ripe ascus-fruit, which frequently takes several months to mature, consists of a hard outer shell containing numerous asci (one of these fruits is shown in fig. 385 '', in the right-hand bottom comer); it is 678 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. able to remain dormant over long periods. Aspergillus is sometimes parasitic, and is stated to promote a diseased condition of animal tissues known as mycosis. A. fumigatus is found growing spontaneously in the air-passages of birds and in the human lungs. TuberacecE. — These are subterranean saprophytes whose mycelia grow in humus. Their fructifications are solid, tuber-like bodies of various sizes. The Truffles (Tuber rufum, T. melanosporum, T. cBstivumi, &c.) are well-known, and several of the species are esteemed as delicacies. In section the Truffle-fruit shows a curious mottled appearance due to its irregular chambered character. The chambers are filled with hyphse which produce numerous oval asci, each containing 4 spores (of. fig. 387 ^ which represents an enlarged chamber); the spores are covered with delicate spines. The wall of the fructification consists of a hard parenchyma-like layer, and possesses a rough and warted exterior. As the fructification ripens the mycelium disappears, the Truffles being found detached in the soil. The spores are ultimately liberated by the decay of the fruit. Elaphomyces produces fruits about the size of a nut. The mycelium of species of this genus is concerned in the for- mation of the fungal investment or mantle of the roots of Pine-trees known as mycorhiza (c/. vol. i. pp. 249, 250); in the Amentacese, &c., neither the species nor group to which the Fungi forming this mantle belong has been ascertained. Pyrenomycetes. — An extremely large and varied group, including both parasitic and saprophytic forms. The essential character of the family is the presence of flask-shaped chambers with a pore at the apex in which the asci are produced. The chambers, the perithecia, may be either solitary upon the mycelium in the simpler forms or embedded in receptacles of most varied form (the stromata) in the more complex. Sections of perithecia are shown in figs. 385 ^ and 386 °. A great many Pyrenomycetes possess conidial as well as ascus- fructifications. The eonidia may arise in tufts from the surface of the mycelium or in urn-shaped cavities — not unlike perithecia — which have been termed pycnidia. As a rule, the conidial pre- cedes the ascal stage. This variety in fruiting has led to the recognition of several forms, which are only stages in the life-history of one Fungus. Consequently, as knowledge extends, many of these supposed species have to be suppressed. Among the simpler forms may be mentioned Podospora, which produces solitary sessile perithecia upon its mycelium. Polystigma rubrum, which occurs on the leaves of species of Cherry and Plum, produces a brilliant red spotting on the leaves. The mycelium permeates the internal tissues, and .during the summer the conidial receptacles or pycnidia are formed. Later, usually in the following spring, in the fallen leaves, the perithecia arise, and the ascospores now liberated infect the young foliage in the vicinity. Nectria cinnabarina, another fairly simple form, occurs as little red cushions on the branches of Horse-Chestnut, Elm, Sycamore, &c.; these cushions burst through the bark budding off eonidia (this is the bright red stage), whilst later, as tiny lobes, the perithecia arise upon them, each lobe containing a single perithecium (this is the dull crimson stage). The number of simple parasitic Pyrenomycetes is almost endless. THALLOPHYTA, 1 CordycepB Taylori, a pyrenomycetous Fiirgu which attacks caterpillars; the branched antler-like stroma has developed rfrom the sclerotlum, and its lower warted portion bears the perithecia. 2 Three perithecia; enlarged. 3 a perithecium in section. * Two asci containing filamentous spores. » Vertical section of a perithecium of Xylaria Hypoxylon. 6 Ascus of same. "^ Mycelium of Eurotium. bearing a conidial hypha (to riglit, above), a commencing fruit (to left), and a ripe ascus-fruit (to right, below). 8 a conidium of the same being abstricted. » Entangled hyphse from which a fruit arises ; the spiral central hypha has been interpreted as a female organ, the tubes growing up the side as male. 10 A young fruit of the same in section ; the asci arise later from the large coiled central hypha. 11 A single ascus of Eurotium. inatsize: s,8 5x50-90. *,6x600: 7x190: 8-iix250. 680 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Of complex forms with citromata we may mention Gordyceps, Xylaria, and Olaviceps. Gordyceps militaris and other species attack caterpillars. The germ- tube having once effected an entrance into the body of the caterpillar and estab- lished itself in the superficial layers begins to sprout vigorously, these sprouts being carried in the blood to all parts of the body. The sprouts now grow into hyphse, and gradually the whole caterpillar is replaced by a dense fungal tissue which rig. 386.— The Ergot of Eye, Claviceps purpurea. 8 Stied ^ZZVZ T t "'^^""P^- ' «<»"«'» «"=tog from the mycelium which develops around the ovary, hec'a atthe eXl ZZ T ?' ^'f ""T- ' ^""^itudiual .eotlon through the head of a stroma showing the peri- po es 8 Mam?nt„ Jll Ts"" ^^"""^^ '"" ^"^'^'"'^ *'"'"S *h« ''"''■ " ^'"^ ' ^scus Uberating its fllamentous spores. 8 Mameutous spores, '."uat. size; 2x200; 4x40; »x60; « and 'xTOO; 8x760. (Partly after Tulasne.) maintains outwardly the form and appearance of the caterpillar, although of animal substance but little traces are left. This fungal mass is known as a sclerotium, and it can remain dormant for some time. Ultimately a branching stroma arises from the sclerotium (Gordyceps Taylori, fig. 385 \ p. 679), or, in C. militaris, several club- shaped stromata. These remarkable stromata are covered with little papillae on their lower portion, and each of these papilla corresponds to a perithecium THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. way groups of cells may hang together in chains (fig. 387^) for a short time, but they disarticulate sooner or later. As the substratum becomes exhausted the cells exhibit a tendency to form spores in their interior (2-8), the cell being as it were transformed into an ascus (fig. 387''). The special physiological activity associated with this and several other species is alcoholic fermentation {cf. vol. i. p. 506), i.e. the splitting up of sugar into alcohol and carbonic acid. S. cereviseoB is used in brewing, S. ellipsoideus causes the fermentation in the juice of the Grape. 1 S. Tnycoderma forms a scum on wine and beer, and is of interest in that it produces mycelial tubes. In all there are some 40 species of Saccharomyces. Alliance XIX. — Basidiomycetes. Parasites and saprophytes, reproducing by means of conidia which arise on basidia in definite number. Besides these characteristic conidia there are subordinate fructifications. Families: Uredinece, Auriculariacece, TremellacecB/PilacracecB, Dacromycetes, HymenoTnycetes, Gasteromyeetes. The Basidiomycetes are an exceedingly large alliance, and include forms from the simplest to the most complex. They all agree, however, in the production of conidia from a definite hasidium, a character which gives its name to the group. As has been already pointed out (p. 674) there are two families in the Hemibasidii, i.e. the Ustilaginacese and the Tilletiacese. In both families a promycelium or basidium arises from the chlamydospore; in the former it is septate and conidia are cut off laterally, in the latter it is non-septate and the conidia are produced in a tuft at the extremity. In the Uredinese a basidium arises from each cell of the teleutospore (the probable equivalent of a chlamydospore), and this basidium is trans- versely septate, four cells being cut off at the end of the tube away from the spore. Each of these cells produces a little process, and from each process a conidium is abstricted. In all the other families of Basidiomycetes the teleutospore (or chlamy- dospore) is suppressed, and the basidia are directly continuous with the hyphae of the Fungus, Otherwise, the basidia of Auriculariaceee, Tremellace^, and Pilacraceae all belong to the type of the Uredinece, and are derivable from the Ustilago-iorm. On the other hand, the basidia of Hymenomycetes and Gasteromyeetes, and probably also of Dacromycetes, belong to the Tilletia type {cf. p. 676). The conidia do not arise laterally, but from four processes at the tip of an unsegmented basidium (see figs. 389^ and 390 ^). The very general restriction of the number of processes ansmg on the basidium to four is without doubt a feature of some importance, and Mg. 3S9. ' Portion of a lamella of an Agaricus with a basidial layer(from which conidia are heing budded off) on either side. 2 Shows three basidia, more high- ly magnified, from the basidial layer of the same fungus; conidia are being abstricted from the four processes (sterigmata). 'x200; 2x600. THALLOPHYTA. 685 is a further justification for the inclusion of all these families— at first sight so different — in a single alliance. Uredinece. — These are the Rust-fungi, parasites for the most part on the foliage of higher plants. They are outwardly manifest in the form of yellow or brown ^S^. Fig. 390.— Basidioraycetes. ^ Clavaria aurea. ^ Dcedalia querciTia. ^ Marasmius tenerrimus. ^ Marasmius perforates. ^ Craterellus clavatus. ^Amanita phalloides. 7 a portion of the basidial layer of the last-named Fungus showing the sterigmata and conidia. « Hydnum imbricatuTn. ^ Polyporua p&rennis. 7x260; the rest nat. size, spots and streaks, due to the spores, which are formed in masses on the surface. The mycelium inhabits the intercellular system of the host-plant, and draws its. nourishment from the living cells. The spores are regarded as chlamydospores, that is to say, localized, thick-walled segments of the hyphae having the properties of reproductive cells. These spores are met with in three forms in the Uredineee. The 686 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM Teleutospores (from reXeuT-^ an end), which are the type most universally present, arise towards the end of the vegetative period. In the majority of cases they remain dormant during the winter. On germinating they form a 4-celled basidium, from each cell of which a conidium is abstricted. Commonly the teleutospores are in rows of two, i.e. constitute spindle-shaped, 2-celled bodies, each cell (spore) of which can produce a basidium. Sometimes the teleutospores are in many-celled rows, occasionally solitary. The Uredospores (from uro, to burn) as a rule precede the teleutospores and germinate at once. They are oval and unicellular. The JEcidio- spores arise in chains in special receptacles (the secidium-fruits) which are at first closed. , Both uredospores and secidiospores difier from teleutospores in that they produce a mycelium at once on germination. They never form basidia and conidia. A fourth sort of reproductive organ, a secondary conidial stage, is often found associated with the secidium stage; this is the pycnidium (cf. p. 678), a hollow receptacle whose lining cells abstract tiny condia. The fate of these pycno-conidia and the part they play in the life-history of the Fungus is unknown. There is thus in the Uredineee great variety in the kinds of reproductive cells. Some species possess all of them and produce them one after another upon the same host-plant (e.g. Puccinia galii and P. primulce); others possess teleutospores only or teleutospores and one of the other types, whilst in others again all the forms are present, but they are not developed upon the same host-plant. This latter property of developing the different stages on two host-plants (known as hetercecism) is by no means uncommon amongst the Uredineae. It is remarkable enough to merit a short description here, though, of course, any exhaustive account of the family as a whole is impossible. The following are well-known cases of hetercecism. Goleo- sporium senecionis, which produces its uredospores and teleutospores on the Groundsel (Senecio vulgaris), its iBcidiospores on Pinus sylvestris (the latter stage formerly known as Peridermium pini); Puccinia graminis, which produces its uredospores and teleutospores on Wheat and other grasses, its Eecidiospores on the Barberry (this stage formerly known as jEcidium berberidis); Gymnosporangium juniperinum, which produces its teleutospores on the Juniper, its secidiospores on the Mountain Ash (Pyrus aucuparia). These three may serve as types of a very large number of similar forms. Puccinia graminis, the Rust of Wheat, is the most famous of all. The secidium-stage (accompanied by pycnidia) arises in spring on the foliage of the Barberry and the secidiospores here produced cannot germinate on the same host, but only on Wheat, Oats, Rye, or some other Grass. Here a new mycelium is established bearing first uredospores and later on teleutospores. The uredospores can germinate at once on other grasses, but the teleutospores remain dormant through the winter, and in spring give rise to basidia from which conidia are abstricted; these conidia can germinate only on the Barberry— not on grasses— and from the mycelium which arises from them fresh secidium-fruits are developed. In the absence of either host-plant (i.e. of the Barberry or of suitable Grasses) the life cycle of the Puccinia is of course interrupted. For the extermination of the Rust disease of cereals the removal of Barberry bushes is an obvious measure. In THALLOPHYTA. 687 parts of Europe, however (e.g. many upland valleys in Switzerland) this precaution is not taken and the ^oidium-covered Barberries and rusted crops may be seen standing side by side. It is true the rust does not destroy the wheat crop entirely but it seriously diminishes the yield of grain. Long before the life-history of the Rust-fungus had been scientifically traced the Barberry was known to have an evil influence upon cereals. So long ago as 1760 the state legislature of Massa- chusetts passed an Act^ compelling the inhabitants to extirpate all Barberry bushes. The main facts connected with the life-history of Gymnosporangiimi have already been mentioned (cf. p. 522), and the two stages are represented on p. 521. The projecting lobes on the Juniper (fig. 357 1, page 521) consist of masses of teleuto- spores embedded in mucilage. When wetted they swell up, the basidia are pro- duced and the conidia abstricted; the latter are then blown away and, should they alight upon the young foliage of a suitable member of the Pomaceae, penetrate the tissues and produce the JEcidium stage. The efiects of various other members of this group are referred to on pp. 524, 525. Between four and five hundred parasitic Uredinese have been distinguished. AuriculariacecB. — Include the well-known Jew's-ear Fungus (Auricularia sambucina) not infrequent on dead branches of the Elder. From its fertile surface basidia with conidia resembling those of the last family are produced, but the basidia are continuous with the hyphse of the Fungus, no chlamydospores being produced TremellacecB. — Gelatinous forms found on rotting tree-trunks. Their substance is curiously lobed and plaited; Tremella Mesenterica, which forms large gelatinous orange masses on dead branches, is the commonest of them. PilacracecB. — Include a single genus only, Pilacre; it grows on Beech-bark, and consists of a spherical head mounted on a stalk. It is of interest because its basidia (from which the conidia are abstricted) are inclosed in a loose layer of hyphse — the outward continuations of the hyphss upon which the basidia are borne — and it is thought to lead on the family of the Gasteromycetes, in which the basidia are entirely covered in. Dacromycetes. — Gelatinous forms resembling the Tremellacese. They approach the Hymenomycetes in that their basidia are destitute of septa. The processes from which the conidia are abstricted are very long. Dacromyces deliquescens is common as a red-coloured tough gelatinous mass on wooden palings. Hymenomycetes. — An extensive family characterized by the production of a ^ The Barberry Law op Massachusetts. — Anno Eegni Regis Georgii II. Vioesimo Octavo, Chap. X. (published January 13, 1755). An Act to prevent Damage to £lnglish Grain arising from Barberry Bushes. Whereas it has been found by experience, that the Blasting of Wheat and other English grain is often occasioned by Barberry Bushes, to the great loss and damage of the inhabitants of the Province : — Be it therefore enacted by the Governour, Council, and House of Representatives, that whoever, whether community or private person, hath any Barberry Bushes standing or growing in his or their Land, within any of the Towns in this Province, he or they shall cause the same to be extirpated or destroyed on or before the thirteenth Day of June, Anno Domini One Thousand Seven Hundred and Sixty. And so forth. (Prom Plow- right's British Uredinece.) THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. fertile surface (the hymenium) consisting of basidia with 4 processes (or sterig- niata) from each of which a conidium (or basidiospore) is abstricted. Typical basidia are represented in fig. 389 ^ p. 684, and fig. 390 =. The hymenium in this group is exposed at the time when the basidiospores are produced. As in the Ascomycetes, so here, there is an extensive range of forms from simple to complex. Simplest of all are the Exobasidece, mostly parasites on Ericaceae. Exobasidium Rhododendri, which causes the Alpine Rose apples, has already been mentioned (p. 520), as also E. Vaccinii and E. Lauri (p. 526). The hymenium in these simple forms is produced over the whole surface of the hypertrophy or Ulster caused by the Fungus. In the Telephorece a definite tissue bearing the hymenium is developed; this is termed the hymenophore or fructification. In the simple Gorticium this forms smooth sheets of waxy nature everywhere encrusting the substratum. The basidia occur over the free surface. Stereum', also very common on old trunks, forms leathery plates usually attached laterally or stalked. The hymenium is on the smooth under surface, whilst the upper surface of the fructification is more or less velvety. Craterellus (fig. 390 '), laterally attached, is hollowed out above; the under surface is the fertile one. In the Clavarice the fructification is club-shaped and fieshy, and covered externally by the hymenium — as in the sulphur-yellow Clavaria incequalis very common on grass lawns and pastures — or it is branched and coral-like as in Clavaria aurea (fig. 390 ^). In the Hydnece there is a well-marked distinction into a stalk and cap (known as the pileus); the hymenium is on the under surface, and is in the form of a number of crowded spines or teeth (e.g. Hydnum imbricatwm, fig. 390^). The Polyporeoe form a large and important group, characterized by the fact that the hymenium has the form of a number of pits, tubes, or meshes, usually on the under surface of the fructification. The simplest of these is the Dry-rot Fungus, Merulius lacrymans. The mycelium of this Fungus penetrates the substance of ill-preserved woodwork in houses, disintegrating it and reducing it to a brittle consistency. At places it produces fructifications, flat irregular bodies whose under surface, the hymenium, consists of a honeycombing of shallow .depressions. The property which these fructifications possess of excreting drops of water has given to this Fungus the name lacrymans. In Polyporus the hymenium has the form of numerous fine tubes lined with basidia. The fructification may be either bracket-like, as in Polyporus fomentarim (growing on the Beech-trunk to the right hand of Plate XIV.), or it may have the cap-like form of a mushroom mounted on a central stalk, e.g. P. perennis (fig. 390 ^) and the allied Boletus edulis (PL XIV. left-hand bottom corner). To these Polyporuses belong some of the most destructive diseases of timber, their myceliums penetrating the wood everywhere, softening and disintegrating it (e.g. Polyporus igniarius, P. fomentarius, P. sulphureus, and P. anwosws— otherwise known as Trametes raddciperda). The last-named P. annosus causes a well-known disease of coniferous timber, the wood coming out in white spots before it is finally disintegrated. It is of interest because the fructifications are produced on the roots of the trees attacked. Allied to Polyporus THALLOPHYTA. is DcBdalia (fig. 390 ^) a bracket-like form in which the hymenium takes the form of irregular branching slits on the under surface. The Agaricinece, which include the very numerous mushroom and toadstool Fungi, are for the most part umbrella- like in form, having a central stalk and expanded circular receptacle (the pileus). The under surface of the pileus is occupied by lamellae or gills which radiate from the insertion of the stalk to the margin of the pileus (cf. figs. 390^'*'^). The basidia forming the hymenium are set upon the sides of these gills. Fig. 389 ^ shows an enlarged section through a part of a gill. In the more complex forms- certain structures are present inclosing the young fructification, but they are ruptured as the pileus expands, and in adult fructifications the gills are freely exposed. Their remains may often be seen on the mature Fungus, as in Amanita (fig. 390^). Just below the pileus there is a membranous ring (the annulus); at an earlier period it was attached to the margin of the pileus covering in the hymenium, and forming what is termed the velum partiale. In the same Fungils- may be noted the remains of another sheath, the velum universale, which en- wrapped the entire fructification. This is shown in fig. 390 * as a ruptured sheath (the volva) at the base of the stalk, whilst portions of the covering which invested the pileus (and was continuous with the ruptured sheath alluded to) are to be seen as white felty patches on the scarlet pileus of Am,anita muscarius (cf. Plate XIV.)., The forms and varieties of the Agaricinese are far too numerous even for mention. Many of them are edible, notably the Mushroom, Agaricus campestris, and the yellow-coloured Cantharellus cibarius (allied to the Agaricinese, Plate XIV. on the left). Others again are poisonous, as, for instance, the scarlet Am,anita muscarius (Plate XIV.), which receives the name Tnuscarius from the fact that decoctions of this Fungus were formerly used for killing flies. Certain forms (Bussula and Lactarius) contain a latex of a white or yellow colour. A number are character- ized by producing sclerotium-like bodies {cf. p. 681). As a rule in the Agarics the fructifications arise directly from the mycelium, but in Goprinus stercorarius,- Lentinus, &c., tuber-like masses of fungal substance are formed, and it is from these that the fructifications arise. These sclerotia, often attaining large dimen- sions, have been found by travellers in various parts of the world, and the fructifi- cation which arises from them is not in all cases known. Several of them, formerly name Pabhyma, &c., are now known as belonging to the genus Lentinus. Very curious are the string-like sclerotia of Agaricus melleus which, from their root- like nature, were formerly termed "Rhizomorphs". They are found especially in Conifers, growing between the wood and bark, and having a ribbon-like form; from them cylindrical branches may arise which penetrate the soil and attack the root of some other tree. Ultimately the mushroom-like fructifications arise from these rhizomorphs. A few lichens derived from the Hymenomycetes are treated at p. 695. Gasteromycetes. — These are characterized by the fact that the basidia arise in closed chambers, which collectively constitute the gleba, and that this is covered by a continuous cortex or peridium,. They include the Pufi"-balls, Earth-stars, Stink- VOL. II. 94 690 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. horns, &c. They are all of them more or less subterranean saprophytes, appearing above' the surface to discharge their spores. The arrangements for the accomplish- ment of this purpose are very varied. The PufF-balls include the genera Lycoper- don, Bovista, Scleroderma, &c. When young a large portion of the interior has a chambered structure (the gleba), and in these chambers the spores are budded off. At maturity the fructification appears above the surface of the ground, and the whole of the substance of the walls of the chambers breaks down, except for certain branching threads (the ca/pillitium, see fig. 391 % which persists along with the Fig. 391.— G-aateromycetes. ' I/yeoperdim constellatwm. 2 Tvloatoma mammosum. ' Capillltlum and spores of Tulostoma. * Oeaster multifidiis. ^ deastet fomicatiis. « Cyathus striatus. » Longitudinal section of same. » Clathrus cancellatus. *x80; ' slightly enlarged; the rest nat. size. minute spores {Scleroderma has no capillitium). The latter escape by the peridium becoming perforated. Lycoperdon (see fig. 391 ^) differs from Bovista in having a sterile basal portion, which is sometimes considerably elongated. In Tulostoma (fig. 391 ^) the outer layer of the peridium bursts and the sterile basal portion elongates considerably, hoisting up the gleba inclosed in an inner peridium. The Giant Puff-ball {Lycoperdon giganteum) sometimes attains huge dimensions — occasionally a metre in diameter. Allied to the Puff-balls is Geaster, the Earth-star (figs. 391 * and 391 ®). In this genus the outer peridium splits into segments and folds back, the inner peridium becoming perforated and liberating the spores. Earth-stars are met with now and then, but they are not usually very common. In Cyathus (figs. 391^ and 391'') we have a form resembling a little bird's-nest THALLOPHYTA. 691 containing eggs. Here the chambers, instead of being numerous and deliquescent as in Lycoperdon, are few and provided with thick, hard walls; they arise in a matrix which occupies the whole body of the Fungus, and when mature they become isolated by the disappearance of the matrix. Each chamber is attached by a string to the wall of the peridium (c/. fig. 391 '). The opening at the top arises by the coming away of a membrane which previously closes it in. Perhaps the most remarkable group of Gasteromycetes is the Phalloidece, which includes the common Stink-horn Fungus and other forms. Whilst immature they are egg- like, but at ripening the investment bursts and the remarkable gleba is hoisted up. Clathrus (shown in fig. 391 ^) has its gleba spread over a hollow spherical lattice- work; the gleba is red in colour and the appearance of the Fungus very striking; it is rare in this country. The Stink-horn, Phallus impudicus, is less rare. When the gelatinous investment bursts, a spongy, spindle-shaped stalk expands and raises up the green, cap-like gleba. The Phalloidess depend upon insects for the dispersal of their spores. Flies are attracted by the bright coloration and foul smell of these Fungi, and they lick up the mucilage into which the gleba deliquesces with great avidity. Perhaps ■ the tropical Bictyophora phalloidea is the most remarkable of the group. It resembles a Phallus, but unfolds around itself a delicate white network which hangs expanded from below the gleba-cap like a crinoline. It is thought that this appendage — like the white corolla of a night-flowering plant — renders the Fungus additionally conspicuous after dark. The Phalloidese, in re- ference to their marked attraction for insects, are sometimes spoken of as the " Flowering Fungi ". ADDITIONAL GEOUP OF FUNGI. LICHENES. In our review of the various alliances and families of Fungi the fact has been from time to time noted that certain members of various groups live symbiotically with Algae as Lichens. Though obviously all these Lichen-fungi do not constitute a natural group or alliance, we propose treating them for convenience together. The general characters of Lichens and their mode of life have been already indicated in the chapter commencing at vol. i. p. 243; consequently little but an enumeration of the groups of Lichens and their methods of reproduction is required now. Briefly, a Lichen consists of a Fungus and an Alga upon which the Fungus lives parasitically. But it is something more than a mere parasitic Fungus on a green plant. The mycelium involves the Alga in the most complete manner (c/. fig. 392), but it doesn't kill it like an ordinary parasite. It lives upon the organic food which the Alga is able to manufacture in virtue of its chlorophyll, but without obvious injury to the Alga. Indeed, the algal cells often attain to a larger size and greater brilliance of colour than when growing freely. On the other hand, the Alga is nowhere in contact with the substratum (being inclosed in the substance of the Fungus), so that water 692 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. and dissolved salts are absorbed and supplied it by the Fungus. It is also protected by the Fungus, and able to exist in places where it could not live alone. Thus, for the Lichen, we speak of symbiosis, a living together. 'Tis true the Fungus is the predominant partner in this association, but it is not a parasite in the common acceptation of that term. Further, as predominant partner it is the Fungus which determines the form of growth and takes the initiative, the Alga following after. But even to this rule an exception has been found, and very likely others exist. For where two organisms live together, as in the Lichen, it may well happen that conditions may exist under which the Fungus can only control the Alga with difficulty, and that the Alga, attempting as it were to escape, compels the Lichen- fungus to follow it, not to lead. This indeed seems to be the case in one of the forms of that most remarkable of all Lichens, Cora pavonia, to be referred to below. The conception of the Lichen as a dual organism, compounded of Fungus and Alga, is of relatively modern origin. Its establishment is due to the researches of Schwendener, which date back some thirty years, and to those of Bomet, which shortly followed them. Since those days the continued study of Lichens has tended only to secure for the " Schwendenerian theory" (as it was formerly termed) a more wide and universal recognition. Previous to the Schwendenerian epoch the Alga was regarded as a definite portion of the Lichen-thallus, its cells as arising from the hyphse of the Fungus; indeed the Algae were termed " lichen-gonidia ". And for many years was the new view opposed by the majority of professed Lichenologists; but into this old controversy we have not space to enter here. It is sufficient to say that the Algse of Lichens are referable to known genera and species of free-living Algse, and that they have been determined for a number of cases. The Alga freed from the Lichen-fungus pursues its normal mode of life, and can then be identified; this is not always possible so long as it remains within the Lichen, owing to the change which the Fungus calls forth in it. It is a noteworthy fact that hitherto no Alga has been found so completely adapted to lichenism that it could not attain to its normal development outside the Lichen-thallus. On the other hand. Lichens have been raised from the spores of the Lichen-fungus allowed to germinate on free-growing Algse. In this way a number of Lichens have been synthesized; and it has been shown that one and the same species of Alga could serve for several Lichens. Finally, the spores of Lichen-fungi have been grown on nutrient solutions, and have attained to advanced stages of development. In nature, however, with one exception (Cora, see below), it is not certainly known that any Lichen-fungus can grow independent of its Alga as substratum. We must regard the Lichen-fungi as being members of various Fungus-families which have become so specialized to a peculiar form of nutrition that under ordinary circumstances they do not develop upon anything except their Algse. The Fungus- forms which occur in Lichens are vastly more numerous than are the Algse ; indeed the latter are drawn from relatively few families — from the Chroococcace^ and and Nostoccacese of the Blue-green Algse, and from the Protococcoidese, Confervoidese, THALLOPHYTA. 693 and Coleochseteae of the class Gamophycese of Green Algse. Thus the same Alga serves for many different Lichens. Classifying Lichens according to the characters of the fungal constituent, we find members of the following families: Discomycetes, Pyrenomycetes, Hymeno- mycetes, and Gasteromycetes. We may therefore speak of 4 families of Lichens: Biscolichenes, PyrenoUchenes, HymenoUchen'es, and Gasterolichenes. Discolichenes. — ^AU produce asci in apothecia after the manner of Discomycetes (c/. p. 682). The apothecia arise in numbers on the surface of the Lichen-thallus (c/. figs. 393 ^ 394, and 387 ^), and the spores are formed in the asci in the usual way. In many genera the spores are multicellular. Li not a few cases there are arrangements for the simultaneous distribution of the Alga-cells along with the Fig. 392.— Lichens. 1 Hyplise of Stereocaulon ramulosum enveloping filaments of the blue-green Alga Soytonema; x650. 2 Cladonia/urcata with Protoaoccus ; x950. ^ Coccocarpiamolybdcea, section oithaMna; x650. (After Bornet.) ascospores, so that it shall not be left to chance for the spores to alight upon a suit- able Alga. In these cases the algal cells bud actively below the layer of asci, and some are pushed up between the asci at the time when the latter reach maturity. Many Lichens possess also conidial fructifications, known as pycnidia (c/. p. 678). These are flask-like excavations, into which tiny conidia are budded by the lining cells. These receptacles were formerly termed spermogonia and the conidia sper- matia; but the evidence that they have any such function as the terminology sug- gests is of the most slender description. On the other hand, these conidia have been caused to germinate and produce characteristic Lichen-thalluses in a number of cases. Pycnidia may be easily seen in the Iceland Moss (Getraria islandica); they occur one in each of the tiny teeth on the margins of the ribbon-like thallus. Very common is vegetative propagation by means of brood-bodies known as soredia. These arise as little buds below the surface, and consist of an algal cell or two and a weft of fungal hyphse. Being formed in quantities together they burst through to the surface as a dust-like powder and constitute the " soredia-heaps ". They are distributed by the wind or washed away by rain. Both constituents of the Lichen 694 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. being present, they readily grow into fresh Lichens. In the majority of Lichens the Algae are arranged in a definite layer or stratum (c/. figs. 392 » and 387 ^); rarely are they distributed through the whole thickness of the thallus, as in the Gelatinous Lichens (Gollema, fig. 393 ^). Apart from these Gelatinous Lichens, the Discolichenes may be divided, according to their form, into Crustaceous, Foliose, and Fruticose Lichens. Crustaceous Lichens include such as are everywhere attached to their substratum (stones, bark of trees, &c.) and cannot be removed without injury. As examples may be mentioned the Lecidias, many of the Lecanoras, &c. The Foliose Lichens are not thus firmly attached; the thallus forms a shield- or ribbon-like expansion, readily separable from its substratum, as, for instance, the well-known orange Lichen Physcia parietina, the green-hued Peltigera canina (PI. XV.), with Fig. 393.— Gelatinous Lichens. I Epheie Kermri ; x 450. ^ Collema pulposum ; nat. size, s Section of thallus of CoUema pulposum, with Nostoe as Alga ; x 450. projecting umber apothecia, common in moist places amongst moss and stones, Umbilicaria, and many others. In the Fruticose Lichens the plant is only attached at one place, and has a shrub-like, branched appearance. Here are included many elegant and well-known forms, including Usnea barhata, the Old Man's Beard Lichen (PI. XV.), which hangs in tufts and festoons from the branches of trees, and sometimes produces large disc-like apothecia the size of sixpenny pieces. Another of these epiphytic forms is Ramalina reticulata, a Calif ornian form which forms beautiful gray-green, ribbon-like nets. Nor must Getraria islandica, the Iceland Moss (PI. XV.), and the Cladonias be omitted. The last-named are glaucous and generally erect-growing and branched. They include Gladonia pyxidata, the familiar Cup-moss (PI. XV.), G. rangiferina, the Reindeer-Moss (PL XV.), and G. coccifera, growing on heaths with its scarlet, and G. macilenta (PI. XV.), with orange apothecia. Pyrenolichenes. — In these the asci are contained in perithecia as in the Pyreno- mycetes (cf. p. 678). Otherwise they are in general agreement with the last group, and include crustaceous, foliose, and fruticose forms. There are only a PLATE XV. Cladonia rangiferina f'LA TK .W: FRONDOSE AND FRUTICOSE LICHENS. Mliteil rroiu lh(; uiigiiials by tlie BIBLIOGRAPHISCHES INSTITUT, Leipzig- THALLOPHYTA. 695 few genera of Pyrenolichenes, including Verrucaria, Ephebe, Endocarpon, and SphcBTophorus. Basidiolichenes.— These occur only in tropical countries, and a number of genera were formerly distinguished, including Cora, Bictyonema, and Laudatea. Cora pavonia, the best-known form, consists of a greenish-yellow, fan-like, concentrically - striated thallus which produces its basidia on the under surface and contains Ghroo- cocctts-ceUs as its Alga; Bictyonema, on the other hand, consists of thin plates of rather felty consistency, in which the radiating character of the strands is very apparent; these delicate plates, blue-green in colour, stand out from the tree-branch to which they are attached. Laudatea, though resembling Bictyonema, is a crus- taceous form. Both the latter forms have /Sc^/towema-filaments as Algse. Quite Fig. 394, — Lecmwra esculenta. recently it has been shown by A. MoUer, a naturalist who resided several years in Brazil, that all these supposed distinct Lichens are different growth-forms of one and the same Lichen. In addition to finding each of these forms in connection with the other — so that there is no doubt of their continuity — he found attached to the Cora-form the Fungus growing free from all trace of the Alga; this Fungus is one of the Telephorese (cf. p. 688), and when it is supplied with Chroococcus-cells grows into the Cora-form. This seems to be the only well-ascertained instance in which a Lichen-fungus has been found growing wild independent of an Alga. The BictyoncTna- and Laudatea-iorma consist of the same Fungus growing upon Scyto- nemxi instead of Ghroococcus. In the Laudatea-iorm the Alga seems to get the upper hand and to determine the growth of the thallus. Gora and its various growth-forms is certainly the most interesting, as it is also the most beautiful of all Lichens of which we have any knowledge. Gasterolichenes. — A Lichen from the Gasteromycetes has also been recognized; it is a little, shortly-stalked, puff-ball-like form resembling a Lycoperdon {cf. p. 690). The Alga (a Palmella) is restricted to the peripheral portions of the Lichen, which is named Ennericella variecolor. 696 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM, Phylum 3.— ARCHEGONIAT^. Plants showiDg a well-marked alternation of generations, that is of a sexual (oophyte) and an asexual generation (sporophyte). The latter reproduces the former by means of spores, the former the latter by means of egg-cells contained in archegonia. Fertilization is by means of spermatozoids which swim in water. Class I.— BEYOPHYTA. The sexual generation is ia the greater number of cases a leafy shoot, arising from a branched, filamentous protonema. The asexual generation, which arises from the archegonium, is a stalked or sessile capsule containing spores; it is destitute of leaves, and never becomes independent of the oophyte. Alliance XX. — Hepaticse, Liverworts. Oophyte nearly always dorsi- ventral; either thalloid or leafy. Protonema incon- spicuous. Sporophyte a stalked or sessile capsule containing spores and usually elaters. A calyptra is not present. Families: Ricciacece, Marchantiacew, Antherocerotacece, Jungermanniaceoe. The oophyte generation in the first three families is a branched, ribbon-like, or lobed thallus (cf. fig. 196 ^ p. 23), showing a distinction between its upper and lower surfaces, i.e. is dorsi-ventral. In the Jungermanniacese it is likewise thalloid in some forms, but in the great majority the oophyte has the form of a leafy shoot. The oophyte communicates with the substratum by means of rhizoids. The complete oophyte is preceded by an inconspicuous filamentous growth, the protonema, which arises directly from the spore, but this stage is not so well marked as in the Mosses. The sexual organs are borne usually in groups either in' little depressions or upon special outgrowths of the thallus, or, in the leafy forms, are collected together into little "flowers" at the tips of the shoots or in the axils of the leaves. The female organs or archegonia are flask-shaped bodies with long necks, as in mosses. The egg-cell is contained in the enlarged basal portion, and on fertilization develops into the sporophyte or spore-capsule. The antheridia are delicate, stalked, oval or club- shaped bodies in which the 2-ciliate spermatozoids are developed. They resemble those of Ghara, shown in fig. 374 '' (p. 660). The spore-capsule develops within the archegonium ; its lower portion (in all but the simplest cases) forming a " foot " or sucker which remains imbedded in the substance of the oophyte, its upper portion forming the capsule proper. In many forms, particularly in the Jungermanniacese, a stalk is developed between foot and capsule, so that the latter is ultimately hoisted up. The capsule does not burst through the archegonial wall until the spores are ripe, nor is a portion of the archegonium raised up as a cap on the capsule as happens in the Mosses. The capsule ultimately opens by splitting into valves. In nearly all cases elaters are present with the spores. They are long thread-like cells with spiral thickening of the wall, and as they dry become very hygroscopic and BRYOPHYTA. 697 contractile. They assist in the dispersal of the dust-like spores in many cases, though the details of their co-operation in this respect is not always the same. Before the ripening of the spores the elaters play an important part in the nutrition of the spores; they are sterile cells or filaments intermixed with the spores to which they conduct food-substances during their development. Attention has been already drawn to the brood-bodies or thallidia of Liverworts (c/. p. 24). RicciacecB. — These are very simple little forms occurring in wet places or floating in water (Riccia natans) like a Duckweed. The thallus is lobed or it may be ribbon-like and branched. The sexual organs are sunk in little chambers on the upper surface of the thallus; antheridia and archegonia may occur on the same or on different individuals. The fertilized egg-cell is here entirely converted into a spore-capsule, i.e. a sheath inclosing spores. No elaters are present in this family, Hg. 395i 1 Vertical section through an air-chamber ol the Liverwort Marchantia polymorpha showing the stomate-like pore and the assimilating filaments. " Transverse section of a leaf of a Moss, Bwrbula alaides, showing the projecting plates of cells. 1x300; 2x380. nor is there a sterile base or foot. The genus Riccia includes 107 species; there are 110 species in the whole family. MarchctntiacecB. — The oophyte is a lobed band-like thallus (cf. fig. 196^, p. 23) with marked dorsi-ventrality. On the upper surface are a number of chambers each opening to the exterior by a single stomate-like pore (Marchantia, fig. 395^). These chambers are really excavations of the upper surface which become closed in save for the pores in question. From the floor of each chamber sprout filaments of chlorophyll-containing cells (fig. 395 ^); these form the main assimilating tissue of the thallus. In Marchantia and Lunularia brood-bodies (or gemmae) are budded off in cup-like receptacles {cf. fig. 196 ^ p. 23). The sexual organs in this family are borne on special receptacles on the upper surface of the thallus. These may be either shield-like or stalked and umbrella-like (as in Marchantia, Fegatella, &c.). The male and female, organs respectively occur on separate receptacles and are variously arranged. There are special sheathing structures associated with the archegonia. The sporophyte generation or sporogonium has a sterile base or foot which remains embedded in the archegonium. The spore-capsule is joined to the foot by a narrow, isthmus-like neck. The capsule opens by splitting into teeth. Elaters are present. Considerable variety is shown in this family in the form of the receptacles of the sexual organs. 165 species have been distinguished. 698 THE STTBDIVISIONS OF THE VEGETABLE KINGDOM. AntherocerotaceoB.— The oophyte has the form of a little lobed disc, in depres- sions of the surface of which the antheridia and archegonia are sunk. The sporo- gonium is long and cylindrical, and is, invested in a circular sheath at its base. The spore-layer has the form of a hollow cylinder, leaving a sterile, central strand of tissue (the columella). Elaters are present amongst the spores, and serve as nutritive tissue for them rather than as instruments of dispersal. In this group Fig. 896.— Jungermanniaoese. 1 FruUania dilatafa growing on the bark of an Acer. 2 A small portion of this plant enlarged; it shows the stalked spore- capsule burst into four valves ; attached to the valves are the elaters. s A shoot of the same FruUania seen from the under side ; at the base of each leaf is a little pitcher containing a Rotifer. The little toothed scales lying on the stem are the amphigastria. * A single pitcher and its contained Rotifer. 6 The Rotifer (^Callidina synibiotica) removed from the pitcher, inat. size; 2x20; 8x26; 4x30; ^xlOO. — alone amongst the Liverworts — stomates occur upon the spore-capsule. It dehisces into two valves. There are 103 species. JungermanniaceoB. — Include both thalloid forms, in which the oophyte genera- tion has a general resemblance to that of a Marchantia, and creeping leafy forms (c/. fig. 396 ^) ; the latter are by far the more numerous. The sporogonium in both cases consists of a long-stalked capsule which splits into four valves (fig. 396^). Elaters are present, often attached to the capsule-wall. The JungermanniacesB grow for the most part on damp earth, stones, and bark of trees. The thalloid forms include the very common Pellia — like a Marchantia^ with BRYOPHYTA. 699 long-stalked capsules inserted on its thallus — Metzgeria, branched and ribbon-like, the curious aquatic Riella with its spiral membrane (see fig. 366, p. 611), and the rather complex Symphyogyna and Pallavicinia (chiefly tropical), with their creeping rhizomes and stalked fronds; these two forms, almost unique amongst the Liverworts, show a well-marked vascular system. The antheridia and archegonia are scattered over the surface of the thallus and not on stalked receptacles as in the Marchantiacese. The leafy forms ai-e very numerous and, as a rule, consist of branched axes which lie parallel to the substratum (e.g. FruUania dilatata, fig. 396 ^). There are three rows of leaves, two towards the upper surface and one towards the lower. These ventral leaves, which are termed amphigastria, are tiny and scale-like (see fig. 396 ^), and sometimes are only represented by hairs. The archegonia are usually borne in little groups at the tips of the main or lateral shoots, and are inclosed in involucres. The antheridia occur in various positions. The sporogonium is always a long-stalked capsule as in the thalloid forms. Not a few of the leafy Jungermanniacese produce curious little appendages or "auricles'' at the bases of their leaves, and these are often developed into little pitchers (fig. 396 '). In other cases the amphigastria bear pitcher-like appendages of the same kind. These structures seem to be receptacles for the holding of water by capillarity against times of drought. In some forms Rotifers inhabit these pitchers (e.g. FruUania dilatata, figs. 396 * and 396 ^), but there is no evidence to show that the Liverwort uses their bodies as food like ordinary pitcher-plants, or that any special relations exist between the Rotifers and the Liverwort. The pitchers are not gall-structures directly produced by the Rotifer — they develop equally well with or without them. They are probably formed by the plant simply for storing water, and are found by Rotifers and other small animals to be convenient abiding-places. There are more than 3500 species of Jungermanniacese Alliance XXI. — Musci, Mosses. The oophyte generation is the leafy moss-plant; it arises as a lateral bud from the simple and generally filamentous protonema. The sporogonium has a seta and spore-capsule, the latter usually possessing a central sterile mass of tissue, the columella. The upper portion of the archegonium is often raised as a calyptra by the elongating sporogonium. Families: Sphagnacece, Andreceacece, Archidiacece, Bryacece. Sphagnacece. — These are the Bog-mosses, and they include a single genus. Sphagnum. The form of the protonema here depends on whether the spore ger- minates in water or upon a solid substratum; in the former case it is branched and filamentous, in the latter it is a cellular expansion, not unlike a fern-prothallium. The leafy moss-plant arises by budding from the protonema, and is remarkable on account of its water-retaining properties. The character of the leaves has been already described and figured at vol. i. p. 219. The sexual organs arise on special 700 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. shoots at the apex of the plant (fig. 397 ''); these shoots are respectively male and female, and may occur on the same or on different plants. The antheridia are in the axils of the leaves of the male shoots, the archegonia in groups at the tips of the female shoots. The fertilized egg-cell develops into the sporogonium, the lower portion of which forms a large foot, whilst the upper part swells up into the spore- Fig. 397 « Polyirichum commune; the spore-capsule to the left is concealed by the cap (calyptra), that to the right is exposed. 2 The same moss in an earlier stage of development. ^ Spore-capsule of Polytrichum commune with its lid. * The same after the falling away of the lid. 6 Brywm ccespiticium, 6 Spore-capsule of same with its calyptra. ? The same after removal of the calyptra. ^ xhe same after removal of the lid, showing the teeth (peristome). ^ A portion of the peristome, enlarged. - 10 Antheridia, archegonia, and paraphyses of Bryurn cxspiticium. n EylocomAum splendens. 12 Spore- capsule of same, is Andrecea rupestris with hurst spore-capsule. 1* Sphagnum cymbifoUum ; the spore-capsules are still intact in the left-hand specimen, is A single capsule of the same. 1, 2 6, 11, 14, natural size; s, «, «. f, s, 12, is, is x6; 9, 10 X 150. capsule. The spore-layer in the latter has the form of a hemispherical shell. Ulti- mately the archegonium is burst irregularly by the enlarging sporogonium, and the .spores are set free by the removal of a circular lid at the summit. In SphagnvMi a true seta is not developed, the region between foot and capsule remaining quite short. The same result, however, is achieved by a considerable BRYOPHYTA. 701 elongation of the axis of the female shoot taking place in the region immediately below the group of archegonia. The capsule is thus hoisted up on a long stalk, though this stalk is no part of the sporogonium (c/. fig. 397 ^^). The remains of the Bog-mosses form an important constituent of peat. Andreceacece. — A small family, including the single genus, Andrecea. They are amongst the first settlers upon new and inhospitable rock-surfaces, and play an important part as soil-formers (c/. vol. i. p. 266). In them the mode of bursting of the spore-capsule is altogether peculiar amongst Mosses. Four longitudinal sHta Fig. 398.— Mosses. 1 A germinating spore. 2 A Moas-protonema. s Protonema giving rise to a bud from which will arise a leafy moss-shoot. 4 Longitudinal section of the tip of a male shoot of a Moss ; small, club-shaped antheridia are present between the scales; 5 Tip of a female shoot with archegonia ; two of them containing sporogoniums have enlarged, and in the left-hand one of these two the upper part of the archegonium (calyptra) has been torn from the basal portion. « Leafy female shoot bearing a fully developed sporogonium; the calyptra is still in position. ^,^,^ x 350-^00; 4x15; ^xsO; 6x5. arise in its wall, and the four valves remain attached to one another at the apex (c/. fig. 3971^). BryacecE. — This family includes the vast majority of the Mosses. The germinat- ing spore produces a simple, branching, filamentous protonema (figs. 398 ^ and 398 ^) on the surface of the ground, certain of its branches developing as colourless rhizoids and penetrating the substratum. From the protonema the ordinary leafy Moss-plant arises as a lateral bud (cf. fig. 398^). The curious properties of the protonema of the Luminous Moss {Schistostega osmundaoea) have been already described {cf. vol. i. p. 385, and PI. I. fig. p). The leafy shoots become rooted by the development of rhizoids from their lower extremities, and bear their leaves, as a rule, in three rows, though a slight twisting of the stem often disguises this fact. 702 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. The Luminous Moss just mentioned is an exception; in it the leaves are arranged in two rows (c/. fig. 399 ^). The leaves of Mosses are generally simple, and (unlike the Jungermanniaceae) provided with midribs. In many of the Polytriehese, and in Barhula aloides, &c. (of. fig. 395 ^), the upper surface of the leaf bears longitudinal ridges of thin- walled chlorophyll-containing cells, thus adding to its assimilating and transpiring surface. The Moss-plant can propagate freely by means of brood-bodies and gemmae. These sometimes take the form of modified leaves, sometimes of little stalked bodies on the leaves; occasionally they are collected together into little recep- tacles at the tips of the shoots, as in Tetraphis (cf. fig. 196, p. 23, where this and other ■cases are illustrated). The antheridia and archegonia are collected into little recep- tacles or "flowers'' placed either at the tips of the shoots (in the acrocarpous Mosses, cf. figs. 397 '^ and 398 ®), or laterally in the leaf -axils (in the pleurocarpous Mosses, cf. fig. 397 ^^). Occasionally both antheridia and archegonia are present together in the same "flower" (cf. fig. 397 ^°), but more frequently they are in separate receptacles (cf. figs. 398 * and 398 '). Mingled with them are sterile scales, the paraphyses. The structural details of the sexual organs and the mode of fertilization in Mosses has already been described (cf. pp. 64-66). After fertilization the egg-cell within the archegonium divides and enlarges, and gradually fashions itself into the sporogonium, the asexual generation of the Moss. For a time the archegonium stretches with the growing embryo, but sooner or later it is ruptured (cf. fig. 398 ^), and its upper portion raised aloft on the sporogonium as the calyptra. Sometimes the calyptra forms a closely-fitting cap, entirely investing the capsule as in Polytrichum (fig. 397 ^), or it may be a little hood split down one side as in Bryum (figs. 397^ and 398 «). After the raising of the calyptra by the elongation of the stalk or seta of the sporogonium the apex swells and develops into the capsule. Though in almost all cases the sporogonium consists of a capsule borne on a long smooth stalk (the seta), which is embedded below in the tissues of the female shoot of the Moss-plant, a very considerable amount of variety is met with in the structural details of the capsule itself. The seta may pass gradually into the capsule as in Bryvmi (fig. 397 «), or there may be a bulb-like enlargement (apophysis) at the base of the ■capsule as in Polytrichum (fig. 897^), or this enlargement may attain considerable dimensions, exceeding the spore-producing part of the capsule, as in Splachnum (fig. 399). This apophysis is of importance as an assimilating and transpiring organ, and it is the only portion of the whole Moss which bears stomata. Within the capsule is the spore-layer. This has the form of a hollow cylinder surrounding a central sterile tissue, the columella. External to the spore-layer, and between it and the wall of the capsule, is a lacuna generally traversed by chlorophyll-containing filaments of cells. Above the spore-layer the columella expands into a mass of tissue, which forms the lid of the capsule (operculum, cf figs. 397 ^ and 397 ')• At the periphery of the lid, where it abuts upon the wall of capsule, a ring of cells becomes marked out (the annulus); later, by the rupture of this ring the lid comes away, and the mouth of the capsule isguarded only by a set of teeth, the peristome {cf figs. 397 ^ 397 ^ 399 ^ and 399 «). By the time that the lid is ready to come BRYOPHTTA. 703 away the spores are ripe, and the columella, &c, have dried up and collapsed. The teeth of the peristome are exceedingly varied in pattern in the various genera of Mosses; typically there are two series of them forming an outer and an inner peri- stome (c/. fig. 397 »), but they are differently thickened, fused, and divided, and one or even both series (e.g. Gymnostomwm), may be absent. They are very hygro- scopic, and their function will be alluded to later on when we are discussing the Kg. 399.— Spore-capsules ol Mosses. 1 Splachnujn lutev/m. 2 An unripe capsule of the same. 3 A ripe and open capsule of the same. * Splacknum vasculomm. 5 Xongitudinal section of a ripe capsule of this moss, showing the large apophysis helow containing lacunse, and traversed in the middle by the columella; above is the capsule proper with persistent columella, spore-sac, and peristome. 8 Spktchnum ampullaceum. ' ' An unripe capsule. 8 A ripe capsule of the same. 9 and,io Schistostega osmundacea. 11 A ripe capsule of the same, i, * «, J" natural size ; 2, s x 2 ; ', b, 9 x 10 ; " x 16 ; « x 100. distribution of spores. In the Polytrichese the peristome is not quite the same as in other Mosses. In this group the teeth are very numerous and quite short, and from their apices a membrane (the epiphragm) remains stretched after the fall of the lid (c/. fig. 397 *). The spores here tumble out between the teeth. The position of the sporogonium is of course determined by that of the female "flowers"; where these are terminal the sporogonium will be terminal (acrocarpous), similarly where lateral (pleurocarpous). The number of genera of Bryacese is so 704 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. numerous that it is hardly possible to mention even the most notable here. Fontin- alis is of interest in being aquatic; Splachnum (of. fig. 399) in having a very large apophysis and being saprophytic on animal excreta (cf. vol. i. p. 118). Bimhaumia aphylla is an exceedingly simple form and vegetates in the protonema-stage. Leafy shoots are only foi'med in cormection with sexual reproduction, and even then they are very rudimentary. This plant has been thought to be a primitive type of Moss. Fossil Mosses are met with in Tertiary and more recent deposits. Class II.— PTEEIDOPHYTA, Vascular Cryptogams. As in the Bryophyta, so here, a well-marked alternation of generations is ex- hibited in the life-history. Whilst in the Liverworts and Mosses the oophyte is the dominant stage ("the plant") here the sporophyte constitutes "the plant". The oophyte is a mere prothallium of simple nature, the sporophyte is a complex struc- ture with root, stem, and leaves, and a well-marked vascular system. It becomes free from the prothaUium at an early stage in development. The Pteridophyte contain the following alliances: Filices, Hydropterides, Equisetales, Lyoopodiales. The plant or sporophyte generation attains to a wide diversity of form in the Pteridophytes; thus, amongst the Ferns the stem is often shott and bears a rosette of fronds, or is elongated and rhizome-like with leaves at intervals; in the Equiset- ales it is erect and jointed, and the leaves are reduced to toothed sheaths; and in many Lycopodiales the stem is procumbent, much -branched, and covered with simple scale-like leaves over the entire surface. Upon the leaves are borne the sporangia which contain the spores. The sporangia may be either scattered over ordinary leaves or on special leaves coUeeted into cones. There is one feature con- nected with the spores that must be described here. Though in the Ferns and ia many other Pteridophytes all the spores are of one kind and each gives rise to a prothallium bearing both archegonia and antheridia, there are Pteridophytes in the alliances Hydropterides and Lycopodiales in which two sorts of spores are produced. The latter are known as heterosporous, the former as homosporous. Where the plants are heterosporous the spores are of two sizes, and the larger ones (macrospores) are contained in fewer numbers in the sporangia than are the smaller ones (micro- spores). On germination the macrospore gives rise to a female prothallium only, the microspore to a male prothallium; i.e. growths which bear respectively arche- gonia and antheridia. The male prothallium is a very simple structure, and its part is played so soon as it has liberated its spermatozoids. The female prothallium having to nourish the young sporophyte for a while, until such time as it can live independently, is larger, and is usually well-provided with food-material. Contrasting the Pteridophytes and Bryophytes, the Fern-plant corresponds to the sporogonium of the latter and the prothallium to the Moss-plant or Liverwort thaUus. In the former the sporophyte, in the latter the oophyte generation is the more complex. But that a Fern-plant has been elaborated out of a Moss-sporo- gonium, or that the Fern-prothallium is a reduced or degraded Moss-plant, is exceed- PTERIDOPHTTA. 705 ingly improbable. It is more likely that the two groups have had a common origin, and have then developed along entirely different lines Alliance XXII. — Filices, Ferns. Families: Hymenophyllacete, Polypodiacew, Cyatheacece, Gleicheniacece, Schizceacece, MarattiacecB, Osmundacece, Ophioglossacece. With us for the most part Ferns have short underground stems bearing a rosette of leaves as in the Male Fern (Aspidium Filix-mas) and Hart's-Tongue Fern {Scolo- pendrium vulgare), or there may be an elongated horizontal underground rhizome, as the Bracken Fern (Pteris aquilina) and Polypody (Polypodium vulgare), the leaves being produced at intervals. In the tropics and sub-tropical regions, on the other hand. Tree Ferns are common. They belong mostly to the genera Gyathea, Alsophila, and Bicksonia. In these a considerable eaudex is developed (c/. fig. 347, p. 473, and vol. i. p. 714), which is often enveloped in. a perfect plexus of aerial roots. Many Ferns are epiphytic, especially in tropical forests (e.g. Platycerium alcicorne, fig. 349, p. 475); with us Polypodiutn vulgare is often met with envelop- ing the trunks and boughs of large trees. The tropical genus Lygodium is charac- terized by its climbing habit, the long rachis of its compound leaf twining like the stem of a twining plant. The view has been held, and is alluded to on p. 12, that the fronds of Ferns are not really leaves but modified stem-structures, the scaly structures that occur on the stem and fronds being regarded as the true leaves. It is sufficient to say that this view is not very widely held amongst Botanists, and that here the term leaf is used as synonymous with frond. Broadly speaking, Ferns love moist and shady habitats ; they grow especially in woods and forests, and on humid rocks beside streams, &c. Generally their leaves are thin and delicate, and ill-adapted to withstand prolonged desiccation. Borne on the leaves of Ferns are the sporangia, tiny capsules in which the spores are developed. The form of the sporangium and the arrangement of these bodies varies in the diflferent families of Ferns — indeed the sporangia afford characters which are used for the grouping and classification of Ferns. In the commonest Ferns (belonging to the predominant family Polypodiacese) the sporangium resembles two watch-glasses placed together, the rim being occupied by a series of large, thick- walled cells (the annulus), and the whole mounted on a little stalk (c/. fig. 400 ^*). In other families the stalk may be absent, the annulus incomplete, oblique, trans- verse or altogether wanting, &c., as will be pointed out in treating the several families. The sporangia are aggregated into clusters, the sori, and these are in many cases protected by little outgrowths of the leaf -surface (indusia) or under the infolded margins of the leaf. The form and arrangement of the sori and indusia provide the characters according to which the large family Polypodiacese is sub- divided. HymenophyllacecB.—Th.e Filmy and Bristle Ferns. There is generally a rhizome which bears delicate fronds at intervals (cf. fig. 400^); the lamina of the Vol. II. 95 706 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. leaf is often only one cell thick, and stomata occur only in the genus Loxsoma (New Zealand). The other genera, Hymenophyllum (the Filmy Fern) and Trichomanes, are both met with in Great Britain. The former is not uncommon on the rocks beside waterfalls, but the latter {Trichomanes radicans, the Killarney Fern) seems to be almost restricted to the south of Ireland. In this family the sporangia occur at the margins of the fronds on the excurrent veins (see fig. 400^). They are sessile, and the annulus is transverse, i.e. at right angles to the axis of the sporan- gium. The sorus is surrounded by an enclosure formed from the leaf -margin; this investment is cup-shaped in Trichomanes and bivalved in HyTnenophyllum. Often in the former genus the axis on which the sporangia is inserted projects considerably from the cup — hence the name Bristle Fern. In this family the prothallium is unlike that of other ferns, being frequently filamentous and branched; the filaments often bear local expansions, upon which the archegonia are inserted. There are about 200 species of Hymenophyllaceee. PolypodiacecB. — By far the largest family of Ferns; indeed this family includes more than three times as many species as all the rest of the Pteridophytes together. Almost all our familiar European Ferns belong to it. The character which they all have in common is a stalked sporangium (fig. 400"),. with vertical annulus. The distribution and form of the sori are exceedingly various, The PolypodiaceES have been separated into the following tribes: — Pteridece, Aspidiece, Aspleniece, Daval- liece, PolypodiecB, Grammitidece, Acrostichece. In the Pteridece the sori occur at the margin of the leaf; in the Bracken Fern (Pteris aquilina) the frond is much branched, and the sori are everywhere continuous on the pinnule-margin; they are covered in by an indusium derived from the margin; in the Maiden -hair Fern {Adiantum Oapillis-Veneris) the tip of the pinnule is folded back over the sorus. In the Aspidiece the sori are scattered, circular, and covered in by a circular or kidney-shaped indusium. Aspidium Filix-mas (the Male Fern) belongs to this tribe. The sorus is much elongated and linear in the Aspleniece, and the indusium is inserted on one side of it (e.g. Asplenium Ruta-muraria, figs. 401 « and 401'). The Lady Fern (Athyrium Filix-foemina), Hard Fern (Blechnum), Hart's-tongue (Scolopendrium), &c., are members of this tribe. In the Davallieoe, which include the large tropical genus Davallia, the sorus is near the margin, and inclosed in a pocket-like indusium. In the Polypodiece the sori are circular and scattered over the under surface of the frond. There is no indusium (see fig. 400 =). The Oram- mitidece resemble the last-named in the absence of an indusium. The sori usually follow the veins, frequently forming very elegant reticulations on the under surface of the leaf, as in the tropical genus Hemionitis. The Gold and Silver Ferns (Gym- nogramm^) belong to this tribe. In the AcrostichecB the whole under surface is covered with sporangia, and there is no indusium. Examples are, Rhipidopteris (fig. 400*), Platycerivm (fig. 349, p. 475), and Acrostichum. Nearly 3000 species of Polypodiaceae are known. Cyatheacece.— This family includes the' Tree-ferns (fig. 347, p. 473). The annulus of the sporangium is slightly oblique; it is only indifferently represented in PTERIDOPHYTA. 707 Fig. 400.— Various Ferns. 1 Nephrolepis Duffl. 2 Tnehomanes LyallU. * Sorus of the same fern, with cup-ahaped investment seen in longitudinal section. * Shipidopteris peltata. ^ Polypodiwm serpens. ' VoTtlou ot ironi ot Oleichenia alpina. i Schizosajlstnlosa. 8 BotryahiUTn lanceolatwm. 9 Under side of a fragment of the frond of Gleichenia alpina ; above the sporangia are concealed by a tuft of scales, below they are exposed, i" and " Fertile pinnule of Cyathea elegam. 12 Longitudinal section of a sorus of Cyathea. is Sporangium of Cyathea. " Sporangium of Polypodium. 16 Sporangium of Schizcea. 16 Under side of the prothallium of Asplenium. 1, 2, 4^ 6, 6_ 7, 8 natural size ; s, s. 10, n, 12, is, w, is, is x 6-20. 708 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. fig. 400 ^^ In Gyathea (figs. 400 !"■ ^^^ ^^) the indusium is cup-like, and closed until the spores are ripe. In BicJcsonia the sorus is marginal, with bivalved indu- sium; in Alsophila the sori are scattered, and the indusium absent or rudimentary; in Hemitelia the indusium is scale-like, and situated on one side of the sorus. Of CyatheacesB about 200 species are known. Fig. 401.— Life-history of a Tern. 1 A Fern-prothallium seen from the under side. 2 An arehegonium in longitudinal section, s An antheridium * Escape of spei™ato.oids from antheridium 5 Young sporophyte with first leaf arising from the prothallium. s complete sporo- vnr/nJh! r"" ^"'"-r™""- ^"^^^ ^^'^"^ °' """"'^ °' ^™^' *°™s "°^=" ^^n ^na l^td-al indusia. s A young prothalhum arismg from a spore ; the spore is below. ' natural size ; ' X 8 ; =, a, 4 x 860 ; « x 6 ; ' x 3 ; s x 240. Gleiche7iiacece.—M.oaily tropical forms. The sporangia have a transverse annu- lus, and are collected into little sori of 3 or 4 sporangia, often very closely packed (of. figs. 400 «, and 400 '). The frond usually forks repeatedly. There is only one genus, Gleichenia, which has some 40 species. Schizceacece.—The members of this family are also for the most part tropical. PTEEIDOPHYTA. 709 The sporangium is sessile, and the annulus is situated at the apex like a cap (c/. iig. 400 15). In Schizcea (iig. 400 ') the fertile pinnules bear two rows of sporangia partly sunk in little pockets; in the climbing fern Lygodium the leaflets bear little fertile spikes at the margin, and the sporangia are sunk completely in little pockets, one row on either side of the spike. In Aneimia the frond divides into two portions— a green vegetative portion, and several fertile branches whose ultimate ramifications are beset with naked sporangia. In habit Aneimia is not unlike a BotrycMum (cf. fig. 400 ^). Of Schizseaceae there are some 70 species. Marattiacece. — Tropical Ferns, many of them attaining considerable dimensions. The fronds care distinguished by possessing a pair of stipules at their base. The sporangia are more bulky than in the families hitherto enumerated, and in Angio- pteris are arranged in rows very close together, whilst in Marattia, Kaulfussia, fee, all the sporangia of each sorus are joined together into little button or bean-like bodies. There is no distinct annulus, though a little cap of cells possibly represents one. There are 25 existing species, but this family was much more abundant, than it now is, in palaeozoic times; their remains are abundant in the Coal Measures. Osviundacece. — Here also the sporangium is destitute of annulus, and possesses a little cap of cells in place of it. In OsTnunda regalis, the Royal Fern, the upper pinnules of the frond alone produce sporangia, but in such quantity that their whole surface is covered with them ; thus the tips stand out in marked contrast to the rest of the frond (hence the name " Flowering Fern"). The other genus of the family, Todea, resembles a Filmy Fern in the delicate texture of its leaves. There are only 11 species altogether. OphioglossaceoB. — A small family including the Adder's-tongue {Ophioglossum) and Moonwort (Botrychium). The frond here divides into a sterile and a fertile portion, the latter seeming to arise from the base of the former. In the Adder's- tongue the sterile portion is unlobed, and the fertile portion spicate, the sporangia being sunk in its substance. In the Moonwort (cf. fig. 400^) both parts are branched, the fertile portion resembling a panicle. The prothallium in this family is a little subterranean tuberous body. The origin from it of the sporophyte generation has not followed in any instance. There are twelve species of Ophio- glossacese. Alliance XXIII. — Hydropterides, Ehizocarps. This alliance is nearly associated with the Filices and more particularly with the earlier rather than with the last-mentioned families of that alliance. All the genera are more or less aquatic in habit; but their distinctive feature is the fact that they are heterosporous, i.e. that some sporangia contain macrospores (one in each sporangium) the others microspores. The sporangia are collected into sori, which are inclosed by metamorphosed leaf -segments into little fruit-like bodies. Families: Salviniacece, Marsiliacece. All the members of the alliance agree in their aquatic habit and in being hetero- sporous. The macrosporangia are larger than the microsporangia, and contain one 710 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. big oval macrospore; the microspores are produced in numbers in the microspor- angia. The macrospore, in addition to a hard wall, possesses an external gelatinous layer with stratified structure. On germination, the macrospore develops its (female) prothallium at one end, and, on the bursting of the spore- wall at the apex, this prothallium and the archegonia developed on its surface, are exposed. Of the contents of the spore, only a portion forijis the prothallium, the rest remaining as a reserve mass. The microspores, which are usually embedded in mucilage, undergo several divisions (forming antheridia), and liberate spermatozoids, which fertilize Fig. 402.— Hydropteridea. ' Salvinia natans, showing the floating and submerged leaves. 2 a portion of the same seen from the side, and showing the aporocarps at the base of the submerged leaves, s a section through two sporocarps of Salvinia natai}£ ; that to the left contains macroaporangia only, that to the right microsporangia. * Filularia globulifera ; one or two sporocarps are shown at the base of the needle-lilie leaves, s Marsilia guacLHfoUa, showing sporocarps. ^ y. i; the rest, natural size. (After Luerssen.) the archegonia. From one of the archegonia the young sporophyte arises, and gradually develops into the adult form. SalviniaceoB. — Include two genera, Salvinia and Azolla, both of which are floating plants. The former occurs in Southern Europe, the latter, though hardy in Europe, is a native of America, Australia, &c. Salvinia (figs. 402 ^ and 402 ^) has a stem which lies horizontally on the water, and develops both floating and submerged leaves. The latter divide into numerous filaments, which hang down like tassels in the water (cf. fig. 402^). They are the absorptive organs of the plant, and play the part of roots. True roots, however, are wanting, even from the embryo. The sporangia are borne on these submerged leaves close to the point of attachment to the horizontal stem, generally in three groups or sori, each of which is inclosed in a cup-like upgrowth arising from the submerged leaf (cf. fig. PTERIDOPHYTA. . 711 402 3). The arrangement of the sori is not unlike those in Triehomanes (cf. fig. 400^), except that the cup is closed round the sorus. In each sorus occur only macrosporangia or microsporangia; but in each group of three sori usually one is different from the other two. Each sorus with its investment constitutes a sporo- carp. The other genus, Azolla, resembles a floating, leafy Jungermannia (cf. fig. 396 \ p. 698); it is closely set with tiny leaves, and numerous true roots hang down into the water. The macrospores are provided with a floating apparatus and hooks; and the microspores which escape from their sporangia in packets have long barbed appendages, which become attached to the hooks of the macrospores. Thus the spermatozoids escape in the immediate neighbourhood of the female prothallia. There are 9 species of Salviniacese. Fossil residues occur in the tertiary formations. MarsiliaceoB. — Containing the two genera, Pilularia and Marsilia. Both grow in marshy or inundated ground, and spread their rhizomes horizontally, attaching them by means of roots. The leaves in Pilularia (fig. 402 *) are needle-like, and each bears at the base a very short branch which develops into a sporocarp. In Marsilia (fig. 402 ^) the leaves resemble those of Oxalis; near the base they give ofi" a branch which may bear several bean-like sporocarps. The sporocarps in both these plants do not — as in the Salviniacese — consist of mere sori with an investment ; but each is a leaf -segment in which a number of cavities develop (four in Pilularia, many in Marsilia), cavities which ultimately are quite cut off from the exterior, though they arise at first as pittings of the surface. In these cavities groups of sporangia arise — ^both macro- and microsporangia in each chamber. The sporocarp in this family is, therefore, in nature a leaf -lobe containing numbers of sporangial cavities, and of much greater complexity than in the Salviniacese. The sporocarps ultimately dehisce, the spores develop their prothallia, and fertilization takes place. There are 32 species of Marsilia and 3 of Pilularia. P. Globulifera alone is British. Fossils are found in tertiary formations. Alliance XXIV. — Equisetales, Horsetails. Possess jointed stems and small leaves inserted in whorls. The sporangia are produced on special leaves arranged in cones. All living examples are homosporous, but palaeozoic forms include heterosporous genera. Families: EquisetaceoB, Galamarice. The Equisetacem alone are represented by living plants, and include the solitary genus Eqwisetv/m, with about 40 species. The habit of growth of the Equisetums is exceedingly characteristic. There is a branching underground rhizome from which erect aerial shoots are produced each year. From the nodes of the underground stems numerous fine roots arise (fig. 403 ^). The whole of the aerial shoot is green and assimilating, and the leaves are represented by funnel-shaped sheaths bearing teeth inserted at the nodes. The inter- nodes are ribbed and the whole structure harsh to the touch, and often brittle owing Y12 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. to the large amount of silica contained in the epidermal membranes. The early spring shoots of many species are unbranched and terminate in spore-beanng cones (e.g. E. arvense, fig. 403 2), whilst later on other branching shoots arise which are sterile (fig. 403 1). In other cases the fertile shoots are also branched (fig. 403 0- Fig. 403. — Equisetaceas. 1 Summer sterile shoot o! Equisetuw, arvense. 2 Vernal, spore-bearing shoot ol Equisetwm arvense. s Fertile cone of the same. * A single sporangiferous scale (sporangiophore) of the same. 6 and « Spores with "elaters" expanded and coiled. ? Equisetmn sylvaticmn with cone. ^ Prothallium of a Horse-tail with antheridia. ^, ^ ' natural size ; * X 3 ; *X6; 5, 6x26; « X30. The branches arise from the main axis in whorls at the base of the leaf-sheaths, and in most cases perforate the latter as they develop (fig. 403'). They repeat the structure of the main axis, save that they are smaller and have fewer teeth to their leaf-sheaths. E. vruwovmum, common in this country in damp places, attains a height of two metres, and is the largest British representative of the group, but PTEKIDOPHYTA. 713 E. giganteum (Tropical America) is stated to reach as much as ten metres. The Dutch Rush {E. hiemale) is largely used for polishing owing to the quantities of silica it contains. Characteristic of the stems of Equisetums is the large central air-space, which is only interrupted by diaphragms at the joints. Other spaces also occur associated with the vascular bundles and in the cortex, alternating with the bundles. The spore-bearing cones (figs. 403 2. s. 7) consist of stalked, shield-like leaves borne on the club-shaped termination of the axis. The scales bear numerous sporangia on their under surface (fig. 403 *), and in these are the curious and characteristic spores. The wall of the spore is three-layered, and the outmost layer splits away from the one below it, forming four arms attached to the spore at one point (fig. 403=). These arms, termed elaters (not to be confused with the elaters of Liverworts, cf. p. 696) are extremely hygroscopic, and though at first coiled around the spores (fig. 403^) become extended as the spores dry, and as their humidity fiuctuates contract and expand again. In this way the spores become entangled with one another and are distributed in groups, arm-in-arm. The importance of this circum- stance appears to be as follows: — The spores, though all of one sort (i.e. homosporous), give rise to dioecious prothallia as a rule (cf. fig. 403^, representing a male pro- thallium); consequently it is of advantage for promoting fertilization that a number of prothallia should arise in the same neighbourhood. This result is achieved by a linking of the spores. The prothallia are richly lobed, but not unhke those of Ferns. The GalamarioB are found as common fossil remains in the carboniferous forma- tions. They include casts of the medullary cavity, impressions of the surface, and actual portions of the stems and cones in a petrified state. Many members of this family attained gigantic proportions, and their stems underwent a well-marked secondary increase in thickness. An examination of the cones shows that these former Equisetales possessed both micro- and macrospores. Alliance XXV. — Lycopodiales, Club-mosses. Forms usually with elongated, branching stems and small leaves distributed over them. The sporangia are borne on the upper surface of the leaf or in the leaf -axil; the fertile leaves are in many cases aggregated into cones. Both homosporous and heterosporous families occur. Families: Lycopodiacece, Psilotacew, Selaginellacece, Lepidodendracece, Sigilla- riacece, Isoetacece. Whilst in the Filices and Equisetales several or many sporangia are present on the fertile leaves, in this alliance there is only one, and this is situated on the upper surface or in the leaf -axil. The sporangia in this group differ from those in many of the Filices (e.g. Polypodiacese) in being more massive and in having origin not from single epidermal cells, but from a row or group. Their form also is in many cases peculiar. The Lycopodiacese and Psilotaeese are homosporous, the other families heterosporous. In the former the prothallia generally resemble those of Ferns, in 714 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. the latter their condition parallels that of the Hydropterides. Lepidodendracese and SigillariacesB are represented by fossil forms only. LycopodiacecB. — The Club-mosses proper include some 100 species, distributed over various parts of the globe. The habit of a typical Lycopodium is indicated in the accompanying figure of L. annotinuTn, with its branching stem closely set with simple, scale -like leaves and terminal cones. The species common in mountain regions in this country are L. clavatum, L. alpinum, and L. Selago ; L. annotinum Fig. iOL— Lycopodium annotinum. (fig. 404) is also met with. Of these L. Selago alone is devoid of cones, its sporangia occurring on the ordinary leaves. The sporangium is generally large and kidney- shaped, and is attached to the base of the upper side of the leaf (fig. 405*); its concavity is directed towards the axil of the leaf. Till recently the life-history of Lycopodium was unknown, as difficulty was experienced in causing the spores to germinate. It was first observed in certain tropical species, L. oernuum and others. In this species the prothallium has the form of a tiny tuberous body, with a lobed fringe on which the antheridia and archegonia are developed. The history of development of the sporophyte from the egg has been followed and is of some mterest. In L. cernuum the young plant consists of a tubercle bearing a tuft of PTERIDOPHYTA. 715 leaves above. Gradually the stem elongates and the adult form is assumed. The special interest attaching to this stage is that it is characteristic of the mature Phylloglossum referred to below. It has been suggested that the last-named genus is a primitive form w^hich retains as adult character what is but embryonic in Lycopodium. The genus Phylloglossum (found in parts of Australia and New Zealand) possesses, in addition to its tubercle and tuft of leaves, a stalk which terminates in a cone of sporangium-bearing leaves. There is only a single species. Psilotacece. — Includes two genera, Psilotum and Tmesipteris. Psilotum is tropical ; it has delicate, angular, forking stems, and its leaves are reduced to tiny scales. It is rootless and grows epiphytically. Its sporangia are three-chambered and are borne on reduced leaves. Vegetative bulbils are frequently met with, especially on those shoots which grow upon the substratum. Tmesipteris is also an epiphyte (New Zealand and Australia). It has conspicuous, pointed leaves and long, trailing stems. The ordinary leaves are simple, but the fertile ones fork like a V, and the sporangium (which is two-chambered) is inserted on the upper surface a^ the junction of the V. The prothallial stage is not known in either of these genera. SelaginellaceoB. — A family of some 300 to 400 species, which are in large part tropical, and all belong to the genus Selaginella. The shoots are forked and are dorsiventrally flattened. The leaves are borne in four rows — two rows of smaller overlapping leaves right and left of the median dorsal line, and two rather larger ones along the edges of the stem (c/. fig. Ill-', vol. i. p. 421). A very common species in the alpine regions of Europe is Selaginella helvetica, whilst >Si. selaginoides ( = 8. spinosa) is British. The last-named species, unlike the majority of Selaginellas, is not flattened, and its leaves are distributed around the stem as in a Lycopodium. A characteristic feature is the presence of a little tongue inserted in the median line of the upper surface of the leaf near its point of insertion; this is known as the ligule. The roots in most cases arise, not directly from the stem, but from special branches termed rhizophores. Selaginella is heterosporous. The sporangia are spherical and arise in the axils of the fertile leaves, which are collected into cones. The macrosporangia contain four macrospores, and the microsporangia numerous microspores. Both kinds of sporangia occur usually in one cone, the former below; or they may be in rows along the sides of the cones; or, finally, the two sorts of sporangia may be on different cones. The product of germination of a microspore consists of a single, simple anther- idium, containing spermatozoids, which are provided with two flagella attached to the pointed end. The macrospore produces a small, green female prothallium at one end (as in the Hydropterideae, p. 710), whilst the rest of the spore, which here divides into large cells, serves as a reserve of food-material. The green portion bears the archegonia, and is exposed. After fertilization, an embryo arises, and gradually develops into the Selaginella--p\a,nt. The embryogeny presents various features of interest. In particular may be mentioned the production of a suspensor from that portion of the embryo which is towards the neck of the archegonium. 716 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. By the elongation of this suspensor the embryo proper is brought down into the aforementioned food-reserve, where it continues its development. This process is quite similar to the corresponding stage in Flowering Plants, where the suspensor is almost universally found. LepidodendracecB.— This family, represented only by fossils from the Devonian and Carboniferous formations, consisted of large-growing Lycopod-like forms, with huge stems clad with linear leaves. They exhibit a secondary growth in thickness (wanting in recent Lycopods), and both micro- and macrospores were produced in the cones. Casts of Lepidodendron-stems bear characteristic rhomboidal areas corresponding to the leaf bases, a,nd upon these the actual leaf-scars may be seen. Fig. 405.— Lycopodiales. 1 IsoBtes lacustris. 2 Expanded fcase of leaf showing tlie sporangium immersed in its socket and partly hidden by the velum and the ligule above, s Longitudinal section of base of leaf showing the strands crossing the sporangium and the inser- tion of the ligule. * Leaf from the cone of Lycopodium davatum showing the kidney-shaped sporangium, s A single 3pore of L. davatum. 6 Prothallium of L. annotinum with young plant attached, i natural size ; 2, »_ 4^ 6 x 10 ; s x 100. (After Luerssen.) SigillariacecB. — Another family which flourished in carboniferous times. Like the LepidodendraeesB, their stems are gigantic, thickened and scarred; they were also heterosporous. The marks on the stems are not rhomboidal, but shield-like, and they stand in vertical rows. The curious branching remains named Stigmaria constitute the root (or rhizome) of Sigillaria. Isoetacece. — Is a small family of aquatic mode of life, containing the single genus Isoetes, which is represented by some 50 species in various portions of the globe. Unlike the other members of the alliance Lycopodiales, Isoetes possesses an ab- breviated stem, bearing a tuft of lance-like leaves. The common British species Isoetes lacustris may serve as type of the genus (fig. 405^). It grows in quantity in the mud at the bottom of upland tarns and lakes in the northern parts of the country, and is attached by delicate roots which repeatedly fork. The very short PHANBROGAMIA. 717 stem is furrowed on two sides, and from these furrows arise the roots. On rare occasions it is branched. From this stem arise numerous pointed leaves, which are slightly expanded below. Each leaf bears a sporangium, immersed in a socket on its upper surface (fig. 405 \ and partly covering the sporangium a membrane, the velum. Immediately above the sporangium is the ligule, a little tongue-like emergence of unknown function (figs. 405 ^ and 405 ^). Macrosporangia, containing several large macrospores, are generally found on the outer leaves, and micro- sporangia, with very numerous microspores, on the inner ones. The sporangia are crossed by strands of cells (c/. fig. 405 »), termed trabeculse, but these do not partition them into chambers. The germination of the spores presents certain resemblances to the same event in Selaginella, but it cannot be followed out in detail here. An interesting feature in the structure of Isoetes is the existence of a cambium- like zone in the stem just outside the central bundle-cylinder. This adds new tissue, both towards the inside and outside, but most abundantly towards the outside. This latter secondary cortex is parenchymatous, but in time it becomes corky. To its formation is due the curious form of the stem. Though many species of Isoetes live below water, others are terrestrial or semi-aquatic in habit. /. lacustris, as it grows at the bottom of a mountain tarn, is very similar in general appearance to two flowering- plants which affect the same situation, viz.. Lobelia Dortmanni and Littorella lacustris; a closer inspection, however, will readily distinguish it. Phylum IV.— PHANEROGAMIA, Flowering Plants The general characters of Flowering Plants have been so fully dealt with in previous sections of this work that little more is needful here beyond a bare outline of the classification of their divisions and alliances. The Phanerogamia are characterized by the production of seeds. The macro- sporangia of heterosporous Archegoniatse are here represented by ovules, the macro- spores by embryo-sacs, and the microspores by pollen-grains. The macrospore (embryo-sac) remains inside its sporangium (ovule), and here produces the reduced female prothallium (endosperm), which never has an independent existence. An egg-cell is formed within the embryo-sac, and this is fertilized by the pollen-tube which has arisen from a pollen-grain lodged upon a suitable receptive surface in the vicinity of the ovule. Ultimately, after the embryo has attained a certain differentiation, the whole macrosporangium, with contained embryo and food- material, comes away, and is known as the seed. The oophyte or prothallial generation is thus suppressed as an independent stage in the life-cycle. The sporophyte, on the other hand, attains to a markedly more complex development than in the groups already treated. Fertilization of archegonia on free-growing prothallia by swimming spermatozoids is here replaced by a direct penetration of pollen-tubes to the ovules. To the " flower " also new duties are 718 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. allotted. In the Pteridophytes, fertile " cones '' are frequently met with. They are assemblages of leaves bearing sporangia, and with the shedding of the spores accomplish their function. But in the Phanerogams it is not so. The stamens, having shed their pollen-grains (microspores), truly are done with. But the carpels which bear the ovules persist in situ until the ripening of the seed. And in the vast majority of Phanerogams, structures accessory to the stamens and carpels have become associated with the flower. These, forming the perianth, promote the transfer of pollen by attracting insects to the flowers in the innumerable ways already fully indicated in this volume. A minority of species depend on wind, and are destitute of attractive perianths. It is worthy of passing notice that wind- pollinated plants, though relatively few in species, are well represented in number of individuals in the various Floras of the globe. It is sufficient to instance the Conifers, Grasses and Sedges, Palms, Amentacese, and Urticacese. The phylum Phanerogamia is divided into two sub-phyla, Gymnospermse and Angiospermse, which difier technically in that in the former the ovules are exposed on scales and receive the pollen-grains direct into the micropyle, whilst in the latter the ovules are borne in closed chambers, the ovaries, and the pollen is received on a special organ, the stigma Sub-phylum A.— GYMNOSPEEMiE. The pollen is received direct upon the nucellus of the ovule, whence the pollen- tube penetrates to the egg-cell. The embryo-sac (macrospore) is filled with the endosperm (prothallium) which bears archegonia sunk in its substance at . that end which is directed towards the micropyle. In almost all cases the archegonia possess ■ neck- and canal-cells in addition to the egg. The phenomena accompanying fertilization and seed-production in the Gymno- spermse having been described at pp. 418 and 437, brief statements of the general external characters alone are given below. The Gymnospermse are divided into 3 Classes: Cycadales, Coniferse, Gnetales. Class I.— CYCADALES, Cycads. Alliance XXVI. Family: CycadacecB. In habit the Cycads generally resemble the Tree-Ferns and Palms. They possess for the most part unbranched columnar stems terminating in a crown of large pin- nate leaves. The surface of the stem is scarred with the bases of the fallen leaves, and recalls in appearance that of the fossil Lepidodendrons (c/. p. 716). In height Cycads do not exceed about 12 metres, and usually they do not attain even these dimensions. The flowers take the form of cones of closely aggregated scales which vary in number from 30 to 600. The cones are respectively male and female, and GYMNOSPERM^. 719 720 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. are produced in close proximity to the apex of the plant, and sometimes (as in the female cone of Cycas) are actually terminal in position. The scales (stamens) of the male cones are usually boat- or shield-shaped, and bear very numerous pollen- sacs on their under surfaces, often aggregated into little groups or sori. The female cones are generally more massive than the male, and their scales (carpels) are shield- like or scale-like in form. In Cycas revoluta (of. fig. 208 '', p. 74) the carpels more nearly resemble the foliage -leaves, and the ovules are borne in the position of leaflets. In other genera the number of ovules is restricted to two. Pollination is accomplished by the wind, and the ovule as it ripens develops a hard shell enveloped in a fleshy coat. The seed thus exhibits the superficial characters of a drupe (of. p. 428). Cycads are restricted to the warmer regions of the globe, and are especially characteristic of the Floras of Australia (Macrozaniia, Bowenia, Cycas) and Central America (Za')nia, Ceratozamia, Dioon). There are nearly 80 species, belonging to 9 genera, and they have mostly a somewhat restricted distribution. Fossil remains, from the Cretaceous onwards, are abundant, and show that in former times the Cycads formed a much more important constituent of the vegetation than at the present day. A living collection of these interesting plants is cultivated in the Palm-house in Kew Gardens; it is exceedingly rich in forms, all the genera, and a large number of the species being represented. Class II.— CONIFERS. Alliance XXVII. Families: Araucariacece, Abietinece, Taxodiece, Cwpressinece, TaxacecB. The Coniferee, which include the various Pines, Firs, Junipers, Cypresses, and Yews, have characteristically branched stems. The leaves are usually linear and needle-like or scaly, rarely possessing an expanded lamina. The flowers are unisexual, and occasionally the sexes are on difierent individuals. In by far the larger number of Conifers the flowers are cone-like, i.e. aggregates of scales set upon a central axis and bearing respectively ovules and pollen-sacs. The stems of Conifers thicken up in the manner characteristic of Dicotyledons, but the secondary wood is composed entirely of tracheides (fibre-shaped elements), with peculiar bordered pits (c/. vol. i. figs. 10^- ^ '■ p. 45); vessels are absent from it. In a great number of forms resin-ducts are present. The families above given fall into two groups. The first of these includes the Araucariacese, Abietinese, Taxodie*, and Cupressinese, and is characterized by the female flowers being cone-like. In the Taxaeete, on the other hand, the female flowers are rarely in cones. Araucariacece.— This group is familiar to everyone in the widely cultivated Chili Pine or " Monkey-puzzle " {Araucaria imbricata). The cones are made up of scales spirally arranged, and the ovules are solitary on the scales of the female GYMNOSPERMiE. 721 cones. The only other genus is Agathis (Dammara). These two genera include 14 species, distributed in the Southern Hemisphere only. Abietinece.— This family includes the majority of familiar Conifers of the Northern Hemisphere. They are distinguished by the fact that the scales of the female cones are divided into an upper ovule-bearing scale (the ovuliferous scale) and a lower subtending bract scale. The ovules are borne in pairs on the former, Fig. 407.— Female Coue and Scales in AbietineEe. I Cone of the Silver Fir {Abies pectinata). 2 Bract scale and ovuliferous scale of the same seen from the outside (the bract scale is pointed). » Ovuliferous scale of same seen from above, showing the two winged seeds, and the bract scale behind. * Longitudinal section of bract and ovuliferous scales, showing a seed in situ upon the latter. 5 A winged seed of the same. ^ Longitudinal section of the seed. ' Ovuliferous scale of the Scotch Pine (Pinus sylvestris) seen from above; it bears two ovules, s single ovuliferous scale of Larch (Larix europcea) showing two ovules on its surface and bract scale - (with bristle) below it. 9 Longitudinal section of the ovuliferous scale of the Larch, i nat. size; the other figs, enlarged. and on ripening into seeds are provided with membraneous wings in most cases. The relations of the parts of the scales and of the ovules are fully illustrated in the accompanying fig. 407. The pollen-grains also are characteristic,- being in nearly all cases provided with two sac-like appendages which promote transit by wind (c/. fig. 217 ^ p. 98). Included here are the Pine (Pinus), Cedar (Gedrus), Larch (Larix), and the VOL. n. 96 722 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. various kinds of Firs {AUes and Picea). All have needle leaves, but variously arranged. In Pinus the needles are borne in tufts of two (figs. 408 * and 408 ^) or three or five. The tufts are really short branches which arise in the axils of relatively Fig. 408.— Mountain Pine (Pinus Pumilio). 1 A single polliniferous scale (stamen) seen from above, 2 Three polliniferous scales, one above the other, seen from the side. The pollen falling from each anther alights on the upper surface of the stamen next below, s Two spikes of polliuiferou& scales. *^ranch with apical group of stamhial flowers from which pollen is being discharged, sj'emale flower. ^,2 xlO: 3x8; *x2; 4 natural size. inconspicuous scales (of. fig. 408 ^) and though these branches are produced plenti- fully, permanent long branches arise only at the yearly limits of growth. The GYMNOSPERMA 723 K". 409.— The Scotch Piae (Pintts sylvestris). 724 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Fig. 410.— The AroUa Pine (Piims Cembra). GYMNOSPEEMiE. 725. Pines may conveniently be divided into 2- and 3-leaved forms on the one hand, and 5-leaved on the other. The former agree in possessing cones of the type shown in fig. 337 2, p. 443, in which the ends of the ovuliferous scales visible at the surface of the cone are pyramidal, whilst the 5-leaved forms are without these terminal pyramids. Among the 2-leaved Pines (20 species) are included the Scotch Pine (P. sylvestris, fig. 409), the Stone Pine (P. Pinea), the Idountain Pine (P. humilis or montana, Plate X., and fig. 135, vol. i. p. 549), the Corsican and Aleppo or Shore Pines (P. Laricio and P. halepensis), and many others. Of the 16 species of 3- leaved Pines the majority are North American and Oriental. The Himalayan P. longifolia has needles nearly half a metre long, and the Californian P. Coulteri cones attaining a like length. The 5-leaved forms include about 35 species, of which the beautiful Arolla Pine (P. Gembra, fig. 410) is a European, and the Weymouth Pine (P. Strobus) a North American example. The genus Cedrus is represented by 3 forms, the Cedar of Lebanon (C. Lihani),. common in Asia Minor, the Deodar ((7. Deodara) of the Himalayas, and the Atlas Cedar (C. atlantica). Some authorities regard these as but varieties of one species. The Larch (Larix, cf. fig. 337 ^, p. 443, and fig. 354, p. 483) bears its needles, which are deciduous, in tufts. The Firs which comprise the genera Abies, Picea, and Tsuga, differ from the Pines, &c., in that the leaves are borne directly upon the elongating branches. Abies, typified by the Silver Fir (Abies pectinata, cf. fig. 177, vol. i. p. 717), has erect cones (cf. fig. 407^, p. 721), with conspicuous bract and ovuliferous scales; on ripening, the scales disarticulate from the axis of the cone. Picea, typified by the Spruce Fir (Abies excelsa, cf. fig. 105, vol. i. p. 415), has pendulous cones, with persistent scales, and, as a rule, the bract scale remains relatively short. Tsuga Bouglasii, the Douglas Fir, and the Hemlock Firs are in certain respects intermediate between the Silver and Spruce Firs. In all there are 120 species of Abietinese distributed over the cooler parts of the Northern Hemisphere. Taxodiece. — Are characterized by bearing more than two ovules on the scales of their cones. They include the two big Sequoia-STpecies of North America, S, gigantea, the Wellingtonia or Californian Mammoth Tree, and S. sempervi/rens, the Eed-wood, both of which attain to huge dimensions. Taxodium distichum, the so-called Deciduous Cypress, interesting on account of the curious knee-like roots which it produces above ground, Cryptomeria japonica, and Sciadopithys verti- cillata, the Umbrella Pine of Japan, all cultivated in this country as ornamental trees, are members of this group, which comprehends some 12 species in all. GwpressinecB. — Are characterized by the fact that their cones have their scales in whorls, not spirals (cf figs. 336 « and 336 ^ on p. 442, and figs. 337 s- *- s, p. 443). They include about 80 species, amongst them the Cypress (Gupressus), Arbor Vitse (Thuja), Juniperus, and other ornamental genera. TaxacecB. — Have generally few scales in their female cones, and sometimes, as in Taxus, the Yew (cf fig. 234, p. 145, and figs. 336 "■• 2. s. 4 b, p 442), the ovule is 726 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. terminal on a little shoot of its own. The seeds in this group are frequently embedded in a fleshy investment (often arillar in nature). Besides the Yews, there are included several peculiar Australasian forms, and the Maidenhair Tree, OingJco hiloha {cf. fig. 337 ^ p. 443). There are about 70 species of Taxacese. Class III.— GNETALES. Alliance XXVIII. Family: Gnetacece. This family which includes three very dissimilar genera, Ephedra, Gnetwm, and Welwitschia, is by no means easy to define. In habit they are all of them quite unlike the members of the two preceding classes, but yet they fall under the sub- phylum Gymnospermee in view of the fact that the pollen-grain has direct access to the nucellus of the ovule and from the resemblance (rather remote) which their ovules and embryogeny presents to other Gymnosperms. They agree amongst them- selves in possessing flowers with perianths, vessels in their wood, and in the absence of resin-ducts from their tissues. Ephedra is a good example of a switch-plant, having jointed assimilating shoots with little scale-like leaves at the nodes, as in Gasuarina or Equisetum. The flowers which are borne in little clusters, are small and unisexual. The male flower consists of a central columnar stamen bearing 2-8 anthers and inclosed in a 2-leaved perianth. The female flower has an ovule with one integument and a little perianth. As the seed ripens the bracts around the flower become red and fleshy. There are some 20 species scattered over the warmer regions of the globe, including the Himalayas, Mediterranean, and Mexico. Gnetum occurs as a Kane or erect tree, and has expanded leaves like a Dicoty- ledon, in decussating pairs. The flowers occur in clustered, catkin-like spikes, on which they are arranged in whorls. The male flowers are very like those of Ephedra, the female have a central ovule with 2 integuments inclosed in a flask- shaped perianth. On ripening, the perianth becomes fleshy, and the outer integu- ment of the ovule hardens to a stone. There are 16 species, distributed in the tropics. Welwitschia mirabilis is a plant altogether unique. Discovered some thirty- five years ago by the botanical explorer Welwitsch, it has formed the subject of a classical monograph by Hooker. It occurs in the desert regions of West Tropical Africa (Angola, Damaraland, &c.). The stem is dwarf and top-shaped {cf. fig. 411), and may attain more than a metre in diameter. The summit of the plant never reaches far above the surface, and it bears two huge leathery leaves which sprawl on the sand on either hand. Actually 4 leaves are produced, the 2 cotyledons, which fall away whilst the plant is still quite young, and an additional pair placed at right angles to the cotyledons and persisting throughout the life of the plant. These 2 leaves grow continually at the base whilst their apical regions become G-yMNOSPERM^. 727 728 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. tattered and broken. Welwitschia, once established, does not increase much in length but continually in thickness. From the edge of its disc, in the leaf-axils, arise branches which bear cones (cf. fig. 411). The male cones are small, and in the axils of their scales occur flowers consisting of 6 stamens united together into a sheath and surrounding a central structure which resembles an ovary, and contains a single ovule, which is, however, so far as is known, always abortive. This ovary- like structure is provided with a trumpet-like " stigma ", and the occurrence of this remarkable structure in the male flowers points to the fact that the ancestors of this plant possessed hermaphrodite flowers. The fertile female flowers occur in the axils of the scales of other much larger cones, which become bright red in colour. Each flower consists of a perianth containing an ovule with 2 integuments, but although the inner of these integuments is very long, there is no stigma-like structure as in the male flower, and the pollen-grain reaches the nucellus. The developmental history of the ovule and embryo is exceedingly peculiar — as it is also in Gnetum — but we cannot enter into these matters here. Sub-phylum B.— ANGIOSPERMJl. Ovules contained in closed ovaries. Pollen received on a specialized portion of the carpel known as the stigma, and fertilization achieved by means of pollen-tubes which penetrate hence to the ovule. Angiosperms fall naturally into two classes, Monocotyledones and Dicotyledones. Class I.— MONOCOTYLEDONES. Includes Flowering Plants whose flowers typically have their parts arranged in whorls of three, embryos with one cotyledon, vascular bundles scattered through the stem and not thickened by a cambium, leaves usually parallel-veined. The Monocotyledones may be divided into 6 alliances :—Liliiflor88, Scitaminese, Gynandrse, Fluviales, Spadiciflorse, and Glumiflora. Alliance XXIX. — LiliiflorsB. Families: Juncacece, LiliacecB, AmaryllidacecB, Iridacece, Bioscoreaceoe, Brome- liacece, Gommelynacece, Pontederiacece. In this alliance the flowers are actinomorphic, and their parts arranged in whorls of three, i.e. two whorls constituting the perianth, two (or one) whorls of stamens, and a whorl of three carpels united together. This condition may be briefly repre- sented by the following formula:— P 3 + 3, A 3 + 3, G <3). in which P, A, and G stand for perianth, androecium, and gynaeceum respectively. The bracket inclosing the number of carpels indicates that they are united (syncarpous). The ovary is 3-celled, and may be either superior or inferior; the seeds contain endosperm. The LiliiflorsB are for the most part herbs with perennial imderground bulbs, corms, and rhizomes. In relatively few cases is a permanent above-ground system ANGIOSPERM^, MONOCOTYLEDONES. 729 Fig. 412.— Liliiflorffi. lutea. ' Oalanthits nivalis, s Leucojv/m vemum. * Colchicum auty/mnale, in flower and in fruit, s Section of capsule of Colchicum. . ANGIOSPERMiE, MONOCOTYLEDONES. 733 J^^;S^. -^JW- ^ (Step, -fety^ Kg. a6.—JEch.mea pcmimUata (after Baillon). 734 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM, Amaryllidaceae. — Resemble Liliaceae, but have inferior ovaries, and in many cases a corona to the perianth. They include the Snowdrop (Galanthus nivalis, fig. 412 2), Snowflake (Leucojum vernum, fig. 412^), Olivia, Amaryllis, Crinum, Narcissus (with well-marked corona, fig. 248, p. 177); also, Agave and Fourcroya {of. vol. i. p. 657). Iridacece. — Resemble Amaryllidaceae, except that they have three stamens only, with extrorse dehiscence. The stigmas are commonly very conspicuously developed. They include the Iris (figs. 412^ and 265), Crocus (fig. 223, p. 113), Gladiolus, the flowers of the last-named being slightly zygomorphic, and many others. BromeliacecB. — Possess distinct calyx and corolla. Ovary superior or inferior; fruit a berry or capsule. The fam.ily is tropical American, and very many of its members are epiphytes, showing interesting adaptations to this particular mode of life. The accompanying fig. 416 of the Peruvian JEchmea paniculata gives a good idea of their mode of growth, with rosette of tough, leathery, sword-shaped leaves and dense, terminal infiorescence. Not infrequently the bracts which accompany the flowers are very brightly coloured. Two of the chief tribes of this family are Tillandsiece, with capsular fruits, hairy seeds, and entire leaves; and BromelieoB, with baccate fruits and toothed leaves (c/. fig. 416). The former includes the rather aberrant Tillandsia usneoides (cf. vol. i. p. 614), a widely-distributed American epiphyte which covers trees much in the same way as does the Lichen Usnea harbata in temperate zones. To the Bromeliese belong numerous forms, including the .^chmea figured here, and the Pine-apple (Ananassa sativa), the various portions of the fruiting-spikes of which become entirely succulent and con- fluent, forming the collective fruits referred to on p. 436. The family includes about 400, and the whole alliance over 4000 species. Alliance XXX. — Scitamineoe. Families: MusacecB, Zingiberacece, Cannacew, Marantacece. This alliance includes tropical plants with rhizomes and lai^e conspicuous leaves. In the flowers there is more or less reduction of the andrcBcium, often com- bined with a production of petaloid staminodes. The ovary is inferior, and usually 3-celled, and the seeds, which are often inclosed in arils, have perisperm. The flowers are zygomorphic, or destitute of any sort of symmetry. As a whole this alliance is one of the most remarkable amongst the Monocotyledons. Musacece.—The flowers of this family agree most nearly with those of typical Monocotyledons. Of the six stamens one only is absent or developed as a stami- node. The flowers are zygomorphic. They include Musa sapientum (the Banana) and M. paradisiaca (the Plantain), widely cultivated for their fruits; Strelitzia, a remarkable South African genus, and Bavenala Madagascariensis, the Traveller's Tree, so named from the water which accumulates in the excavated sheaths of the leaf -stalks. This plant attains a height of 10 metres, and has a remarkable appear- ance (cf. fig. 417) owing to the fact that its huge leaves (amongst the largest in the ANGIOSPERM^, MONOCOTYLEDONES. 735 Fig 417 —Ihe Traveller s Tree (iictyenaia Vadarja^caiiemih) Aftei i ]ia-vMu„ ) J belleij 736 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. vegetable kingdom) are borne in one plane. The fruits are large capsules, and the seeds are inclosed in arils with blue metallic lustre. Zingiberacece.— One stamen, the posterior of the inner whorl, alone is fertile. The two others of this whorl are joined together to form a petaloid structure — the labellum. Here are included the Ginger (Zingiber), Alpinia (fig. 283 \ p. 289), and Hedychium, cultivated for the beauty of its flowers; also Globba and Mantisia (the Opera Girl) with flowers altogether remarkable; M. saltatoria presents some resem- blance to a ballet-dancer, hence the name. Cannacece.^Flowevs altogether asymmetrical. The fertile stamen possesses only a half-anther and several petaloid staminodes are present forming the most conspicuous portion of the flower. Ganna indica, the Indian-shot, and other species are much cultivated, and have given rise to numerous brilliantly-coloured hybrids. MarantaoeoB. — Have asymmetrical flowers and a fertile half -stamen only. The other stamens are modified into pecuUar staminodes, and the family is nearly allied to Cannacese. Arrowroot is obtained from the rhizomes of species of Maranta. The Scitaminese includes about 600 species. Alliance XXXI. — Gynandrae. Families: Orchidacece, Burmanniacece. Flowers with petaloid perianth and inferior ovary. The stamens reduced to one, two, or three, and generally united with the gynseceum to form a column (hence the name Gynandree). The fruits are capsules, and the seeds exceedingly small and numerous. OrchidacecB. — After the Gompositae, the largest family of flowering plants, numbering some eight thousand or more species. Its members are chiefly tropical epiphytes, and their mode of life has received frequent mention in vol. i. Very characteristic of the Orchid flower is the median petal, developed as a labellum (c/. fig. 2581 p. 227, and fig. 268 ^ p. 265). The family may be divided into two divi- sions according to the number of stamens present. Division 1. Diandroe. — Having usually two polliniferous stamens and a large and conspicuous staminode (which corresponds to the single fertile stamen of the Monandrse). The Lady's Slipper (Gypripedium, cf. figs. 267 ^ and 267 ^ p. 249, and description on p. 253) is the best-known representative of this division. A little group of plants, the Apostasiacese, is also included here. To it belongs Neuwiedia, with 3 fertile stamens (i.e. the two found in Gypripedium and the one which is represented by the staminode). Division 2. Monandrce. — Includes the greater portion of the family, with one stamen only united with the gynseceum into the column and producing pollen in masses (pollen-masses). The single stamen is inserted above the stigmatic surface, from which it is separated by the rostellum, and is on that side of it which is away from the labellum, the usual alighting place for insects. The structure of the monandrous Orchid flower having been fully described and figured (pp. 253-257), ANGIOSPEEMyE, MONOCOTYLEDONES. 73r repetition is not needful here. The Monandrse may be divided into four tribes, the Ophrydece, Neottiece, Vandece, and EpidendrecB. The Ophrydece include most of the British and European Orchids, which are not Fig 418 — Angrcecum ebumeum epiphytic on a tree trunk (Madagascar) epiphytes but terrestrial, with swollen tuberous roots, including Orchis, Ophrys,. Gymnadenia, Habenaria, and the South African Lisa. The NeottieoB also include some European forms, Cephalanthera, Listera ovata (the Twayblade), &c., and a series of coloui'less forms of saprophytic habit, which are destitute of foliage, including Epvpogium aphyllum (fig. 257 i"' p. 226), Neottia, VOL. II. 97 738 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Nidus-avis (the Bird's Nest Orchid), and Corallorhiza (the Ooral-root). To this tribe belongs the climbing Vanilla-orchid (Vanilla planifolia) the capsules of which yield the spice. To the Vandece and Epidendrece belong the numerous tropical epiphytes of which many genera are widely cultivated in greenhouses, such as Lceliaj Epiden- drum, Gattleya, Masdevallia, Phajus, Gatasetum, (figs. 275 1- ^' '> *■ ^' ^' ''• p. 269), Stan- hopea (Plate XIIL), Phalcenopsis (fig. 258 \ p. 227), Odontoglossum, (Plate XIII.), Dendrohiwm (fig. 275 ^), Oncidium (Plate XIIL), Angrcecum (fig. 418), and many ■others besides (vol. i. Plate III.). To give any idea of the enormous varieties of floral structure met with among the Orchids would require a whole volume. In this family of all others do we find adaptations to insect visits carried out on a gigantic scale, and in not a few •cases mechanisms of singular beauty and delicacy. For 'full details Darwin's Fertilization of Orchids should be consulted. BurmanniaceoB. — A small family of some 60 species, largely represented in Borneo and New Guinea, is of interest, as it seems to connect the highly specialized Orchidaceae with more typical Monocotyledons like Amaryllidacese. They have •curious flowers, with three or six stamens, and several of them are saprophytic. Alliance XXXII. — Fluviales. Families: Potamogetacew, Haiadacece, Aponogetacece, Juncaginacece, AlismaceoB, ButomacecB, Hydrocharitacece. This alliance includes a large number of aquatic forms, some with narrow, some with broad leaves. The gynaeceum is superior, except in the Hydrocharitacese. The stamens and carpels show a tendency to an increased number as compared with typical Monocotyledons. On the other hand, many forms with reduced flowers occur. Endosperm is generally absent. Potamogetacece.—lnclnde chiefly submerged forms, some of which raise their inflorescences above the water-level, and are wind-pollinated {Potamogeton, fig. 419), whilst the others, including the Sea-grass (Zostera marina), Zannichellia, &c., are pollinated below the water. Potamogeton (Pondweed) is a large genus of some 50 species, met with in fresh and brackish water; Zostera- grows on sandy shores between tide-levels, often forming extensive belts. The embryos in this family are peculiar. They consist of a much -thickened hypocotyl with a relatively small cotyledon inserted upon it. They are termed macropodous. There are about 74 species of Potamogetacese. Aponogetacece.— Contekins two interesting genera, Aponogeton and Ouvirandra. The plant is submerged, and raises a spicate inflorescence (often forked) above the water. The flowers are imbedded in the spike, and consist of some 6 stamens and 3 free carpels, and a small number of perianth-segments. Aponogeton distachus is often cultivated in this country on account of its beautiful white flower-spikes. Ouvirandra fenestralis is the Lattice-leaf plant of Madagascar. ANGIOSPERMjE, monocotyledones. 739 There are 15 species in all. Alismacece. — Possess a 6-leaved perianth, and stamens with tendency to increase by division; carpels numerous. Here are included Alisma Plantago, the Water Plantain, and Sagittaria sagittifolia, the Arrowhead. There are about 50 species. Butomacece. — Includes Butomus umhellatus, the Flowering Rush, interesting, from the fact that it bears ovules all over the internal surface of its carpels. Fig. 419. — Curled Pond weed (Patamogeton crispvs). HydrocharitacecB. — Is distinguished from the foregoing families in that it includes submerged forms with inferior ovaries. The flowers are frequently uni- sexual, and in one form (Halophila) are pollinated under water, as in so many of the Potamogetacese. To this family belong Vallisneria (see fig. 155, vol. i. p. 667 and fig. 227, p. 132), Elodea, the American Water-weed (alluded to on p. 457), Lagarosipon and. Unalus (p. 133), Stratiotes aloides, the Water-soldier (vol. i. p. 552) and Hydrocharis Morsus-rance, the Frog-bit, with expanded floating leaves. Contains about 60 species. 740 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Alliance XXXIII. — Spadiciflorse. Families: Palmaceoe, Aroideas. Monocotyledons with small and usually unisexual flowers crowded on spikes or spadices, and inclosed in one or more conspicuous sheaths, the spathes. The ovaries are superior. This alliance may be contrasted with the group Compositse of Dicotyledons in which also the individual flowers are merged in dense crowded inflorescences which superficially resemble single flowers (cf. Arum and Chrysanthemum). PalmacecB. — Include plants with cylindrical, woody stems and tough fan-shaped or feather-like leaves of large dimensions having a plaited vernation. The flowers are borne in branched, fleshy spikes often inclosed in large sheathing leaves; they are hermaphrodite or unisexual and actinomorphic; the parts are arranged in threes, and are inconspicuous. The gynaeeeum consists of three carpels, each containing one seed. Stamens six, pollen dust-like. Fruits are berries, drupes, and nuts, and contain three, or by suppression, one seed. The endosperm is copious, and generally hard and stony. The majority of Palms possess upright, columnar caudices sur- mounted by a huge tuft of crowded leaves {cf. vol. i. p. 289, and PI. VIII.). In several species the caudex attains a height of 30 metres, and in one (Ceroxylon andicola) 57 metres. The Climbing Palms have slender branched stems, and by the aid of the hooks on their leaves mount to the summit of trees and stretch like lianes from crown to crown {cf. vol. i. pp. 363, 675, and 676). The stems of these Palms reach a length of 150-200 metres, and yield the rotang cane. The opposite figure shows the interior of a forest penetrated by Climbing Palms and two natives rolling the stems into a coil. Old Palm-stems are either smooth and show the scars of the fallen leaves, or they still bear the disintegrated fragments of former foliage- leaves. Others again are armed with spiny girdles and scales. The leaves are folded in bud and undivided, and as they unfold they split along the creases, and the blade is divided pinnately or like a fan; we may distinguish between the feather-leaved and fan-leaved Palms. Often in young Palms the leaf splits at the apex into two pointed lobes only, as in Areca disticha, represented the foreground of fig. 420. The dimensions of Palm-leaves and the gigantic inflorescence of the Talipot Palm {Gorypha umbraculifera) have already been alluded to {cf. vol. i. pp. 287 and 745). In Oreodoxa regia the sheathing base of the leaf attains a length of 2 metres and a half. The fruits of many species (e.g. Ghamcerops excelsa) are borne in grape-like bunches; in others they attain to great size and weight. The Double Cocoa-nut, the fruit of Lodoicea Sechellarum, is prominent in this respect {cf. p. 452). Most Palms are eminently tropical in their distribution. Some genera are met with throughout the tropics, others (e.g. Mauritia, Oreodoxa, and Iriartea) are confined to the New World; others again, as Borassus {B. flabelliformis, PI. VIII.). Raphia, Garyota, and Galamus to the Old. Ghamcerops humilis, alone of the ANGIOSPERMiE, MONOCOTYLEDONES. 741 Fig. 420.— Primeval forest in Ceylon with Climbing Palms (Calamus) and Areca disticha in foreground to the right. (Drawn from nature by v. Bansonnet.) 742 THE SUBDIVISIONS OF THE VEGETABLE KINGDOftT. Palms is indigenous to Europe; Ceroxylon andicola is found in the Andes growing at a height of 270 metres. Fossil remains are found in the formations of the secondary and tertiary formations. The number of living species is about 1100. ^ Aroidece.—Tevenmal xAants with tubers, rhizomes, and climbing stems which Fig. 421.— Avoids. 1 Ai-um mamlatum. 2 Spadix of .4. moinilatom, tlie spathe removed, a Fruiting spike of same. * Inflorescence of Colocasid antiguorum. 5 Basal portion of tliis inflorescence witli part of spatlie removed. ^ Ovary of Coloeasia antiquorum. 1 The same in longitudinal section, s Columnar androecium of same. ^Ariopsis peltata. ^, *, » nat. size; \ ^ *reduced; «, 7, 8 enlarged. genera,lly bear large foliage-leaves. ' The flowers are borne on unbranched, fleshy spadices ..which are inclosed in large, expanded spathes (c/. figs. 421 1'*-^'^); they are unisexual or hermaphrodite. The parts are inserted in whorls of 2 or 3; the perianth-members being inconspicuous and often absent. The androecium is very various. In Coloeasia antiquorum (fig. 421 *) it consists of a whorl of stamens ANGIOSPERM^, MONOCOTYLEDONES. 743 Fig 422.— Raphidophora decursiva climbing in a primeval forest of the tropical Himalayas (from a photograph) 744 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Fig 423 —Climbing Aroids (Philodendron pertvmm and P. Imhe) with cord-like aerial roots. |^iJJi!Sn?P ANGIOSPEKMiE, MONOCOTTLEDONES. 745 united into a single abbreviated column. Endosperm is present in the seeds. The tribe PistecB includes floating plants with leaves arranged in rosettes and propagat- ing vegetatively by means of stolons. The Arece, of which the Arum (figs. 421 1- ^- ^) may be taken as a type, have subterranean tuberous stems, from which arise the leaves and spadices. Numerous representatives of other tribes, including Ariopsis, Caladium, Bracontium, and Amorphophallus have tubers. Amorphophallus titanum, the giant of this family, has a tuber 50 centimetres in diameter, and produces umbrella-like leaves on stalks 2-5 metres long and with segments in proportion. The inflorescence is a huge spadix some 2 metres high, encircled by a sheath of beautiful mottled green with purple lining and frilled edge. When this plant flowered at Kew in 1890 (for the first and as yet only time in captivity) it was one of the sensations of a London season. It is a native of Sumatra. A few species of Montrichardia (if. linifera) and Philodendron (P. bipinnatifidum) have erect cylindrical stems, whilst the Snake-root (Calla palustris) . and Sweet Flag (Acorus Calamus) have creeping rhizomes. Many tropical Aroids belonging to the tribes Monsterese and Pothoidese climb up the stems of trees, fastening themselves by their aerial roots, and pass from crown to crown like lianes. The Himalayan Baphidophora decursiva (fig. 422) is an* example of this type of growth. Many of these climbing Aroids send down pendent aerial roots into the humid air of the forest (fig. 423), and these not unfrequently reach the ground, take root, and become stretched taut. The majority of Aroids are tropical, less than 10 per cent of the species being met with in temperate regions. Acorus Calamus, Arv/m maculatum, and Calla palustris reach the furthest north. The curious Ariopsis peltata (fig. 421 *) occurs in the Himalayas to a height of 1600 metres. There are about 900 species of living Aroids. The Lemnacece is a little family of reduced forms allied to Aroidese. The flowers are unisexual, and consist of a stamen and a carpel respectively. They are floating, flattened forms, and include Lemna (the Duckweed), and Wolffua, which is destitute of roots. Associated with the Spadicifloreae are the Pandanaceae, which include Pan- danus utilis, the Screw Pine (c/. vol. i; fig. 186, p. 758); Oyclanthacese, climb- ing and palm-like; Sparganiacese and Typhaceae, marsh plants, which include SpargaTWiim, the Bur-reed, and Typka, the Bulrush. AUiance XXXIV.— Glumiflorse. Families: Graminece and Gyperaceoe. This alliance, which includes some 6000 species, consists exclusively of Grasses and Sedges, forms with insignificant flowers destitute of coloured perianths and pollinated by wind. GraminecB. — ^Annual and perennial plants with upright, jointed haulms, and in the case of perennials, provided with creeping rhizomes. The leaves consist of an 746 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. undivided, linear, parallel-veined lamina (vol. i. fig. 150^), and a sheathing basal portion. At the junction of blade and sheath is inserted a little scale — the ligule. Flowers hermaphrodite and unisexual, arranged in spikelets (c/. fig. 231, p. 139). Perianth absent, its place being taken in many cases by 2 tiny scales, the lodi- Gules, sometimes regarded as reduced perianth-leaves. Each flower is inclosed in a sheathing scale known as a palea, whilst outside this and subtending the flower is a bract-like structure, the flowering glume; this is often awned. Every flower is thus inclosed in a palea and flowering glume, whilst the whole spikelet is inclosed in a Uttle 2-leaved involucre consisting of 2 outer glumes. The ovary bears 2 feathery stigmas, and contains a single ovule. The stamens are generally 3 in number, though variations are met with. The pollen is dust-like. Pollination has been fully described on pp. 140-142. The fruit or grain is indehiscent, and is known as a earyopsis. The seed contains a floury endosperm, and an embryo placed at one side (cf. vol. i. flgs. 141 ^' *■ ^, p. 599). The intemodes of the haulm are in Zea, Andropogon, Panicum, &c., filled with pith; in the majority of Grasses they are hollowed. The Bamboos and numerous other tropical Grasses have upright perennial stems, and form an arborescent vegetation {cf. vol. i. p. 713). Bamboos attain a height of 25 metres and a diameter of nearly half a metre. But the majority of Grasses produce new haulms each year from their subterranean rhizomes, and these die down at the end of the season. The female flowers of the Maize {Zea Mais) are borne on thick spadices (cobs) inclosed in sheathing bracts. Grasses are widely distributed over the globe, the tropics being richer in species than the temperate regions, but poorer in individuals. Grasses are found extending into arctic and alpine regions to the extreme limits of phanerogamic vegetation; thus in the Alps Poa laxa has been found at an elevation of 3000 metres. The Bamboos are tropical and sub-tropical; in the Steppes certain Grasses are very pre- dominant, e.g. the genera Stifa and Festuca {cf. Plate VI. vol. i. p. 616). In moist, temperate climates. Grasses form a continuous carpet, the basis of meadow land. In marshy places and by river banks reed-like Grasses occur in great quantities (e.g. Phragmites communis). Gramine^ number about 3500 species. Gyperacece. — Annual and perennial plants with upright, haulm-like stems, jointed below and with long upmost segment. The leaves much resemble those of Graminese, but the ligule is wanting. Flowers hermaphrodite and unisexual, aggre- gated into spikelets, inclosed in bract-like scales. Perianth absent, or represented by scales, bristles, or hairs. The ovary is 2- or 3-carpellary. Stamens in one or two whorls of 3 each; pollen dust-like, pollination by wind. The seed contains endosperm. In the Scirpeae the leaf-blades are frequently obsolete, and assimila- tion is carried on by the stems. Scirpus lacustris reaches a height of 1, Papyrus antiquorum (or Cyperus Papyrus, fig. 424) of 3 metres and a diameter of 10 centi- metres. The pith of the larger flowering stems of this plant cut into thin strips, united together by narrowly overlapping margins, and then crossed under pressure by a similar arrangement of strips at right angles, constituted the papyrus of ANGIOSPEEMjE, monocotyledones. 74T 748 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. antiquity; it grows in the Upper Nile, Syria, Palestine, &c. The stem of Papyrus bears at the summit an umbel-like tuft of filamentous branches, upon which the inflorescences arise. Cyperacese grow for the most part on damp moors, and by the banks of streams and lakes, and in mountain regions. Many of them are social forms, noteworthy in this respect being Garex stricta, which forms hummocks in marshy places, standing up above the water, often thousands together. Several Sedges, e.g. Garex sempervirens and C. firma, contribute largely to the turfy carpet of alpine slopes (of. Plate XII.). The family is distributed over the whole world. Garex, Sriophorum, and Soirpus are found especially in cooler and northern zones; Gyperus and Papyrus in warmer regions. About 2500 species are known. Class II.— DICOTYLBDONES. Flowering Plants whose flowers typically have their parts arranged in whorls of four or five, embryos with two cotyledons, vascular bundles arranged in a ring and undergoing a secondary increase in thickness, leaves more complex than in Monocotyledones and usually reticulately veined. The Dicotyledones may be divided into three Sub-classes: Monochla/mydecB, Monopetalce, and Polypetalce. The Monochlamydese have a simple perianth, or in some cases the perianth may be wanting. The Sub-class is an artificial one, as it includes forms whose ancestors probably possessed a double perianth and others which are primitively simple. The Monopetalse and Polypetalse possess both calyx and corolla; in the former the parts of the corolla are united together, in the latter free. Sub-class I. — M0N0CHLAMTDE.E. Alliance XXXV. — Centrospermae. Families: Piperacece, Polygonacece, GynocrambacecB, Urticacece, Ghenopodiacece, NyctaginacecB, Amaranthacece, Paronychiacece, Garyophyllacece. Annual or perennial .herbs, shrubs, and trees. Venation of the leaf -blades palmate or pinnate. Flowers solitary or in cymes; the cymes arranged in fascicles, glomerules, or spikes. Flowers actinomorphic, hermaphrodite, pseudo-hermaphrodite, monoecious, and dioecious. Floral-leaves in one or two whorls; all sepaloid, all petaloid, or (in a few cases) the outer whorl sepaloid and the inner whorl petaloid. Where a corolla is developed the petals are free. In the case of dioecious fiowers there is no difference between the male and the female flowers in respect of the development of floral-leaves. The ovary is superior; 1-5-carpellary, unilocular. The ovules are borne in the centre of the ovary on a stalk which rises from the bottom of the ovary, and is sometimes long, sometimes short. Stamens 1-30, arranged in one or two whorls, the outer ones inserted in front of the sepals or sepaloid perianth segments. Fruit an achene, capsule, or berry. The seed contains an abundant farinaceous or mucilaginous endosperm. Cotyledons not thickened. ANGIOSPERM^, DICOTYLEDONES. 749' The Centrospermee are extremely rich in inorganic salts, and in the case o£ many of the species soda is extracted from the ash obtained by burning the plants. The Piperace^ contain aromatic and pungent substances; the Urticaceee secrete enzymes in their stmgmg-hairs (see vol. i. p. 441). The leaves are lobed in Urticaceae and Ohenopodiacese, in the rest they are undivided and have entire margins. In several Cheno- podiacesB the cauline leaves are squamiform, and assimilation is then effected by the green cortex of the branches, which are trans- formed into phylloclades. The Piperacese are distinguished by a peculiar distribution of the bundles in the foliage-leaves. The lateral strands do not branch' off from the midrib in the usual way, but are appressed to it and can be traced to the base of' the lamina. The Urticace^ also, particularly the genus Parie- taria, exhibit a peculiar disposi- tion of the bundles (see vol. i. p. 629). The Chenopodiace^ are destitute of stipules, the Paronychiaceae have large mem- branous stipules which protect the foliage -leaves, the Poly- gonacese are distinguished by curious sheathing stipules. In the Caryophyllaceas and some Paronychiaceae the floral enve- lopes are differentiated into calyx and corolla; in Nyctagin- acese, Amaranthacese, and most Polygonacese there is a petaloid perianth, whilst in Chenopodi- acese and Urticaceae there is a sepaloid perianth. The perianth in Nyctaginaceae resembles a corolla most strongly when the bracts are connate and form a sepaloid envelope or involucre, as is the case, for instance, in the Marvel of Peru (Mirabilis Jalapa, see fig. 425). The lowest portion of the perianth in Nyctaginaceae continues to grow after the flower has faded and forms a leathery or woody investment to the fruit (see fig. 425 ^). . In several Chenopodiaceae and; Fig. 425.— Nyctaginacese, Mirabilis Jalapa, 1 Flowering brancli. 2 Fruit inclosed in the persistent base of tlie peri- antii. 3 Longitudinal section tlirougli tlie same; the true fruit is seen within. (After Baillon.) 750 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Urticaoese also the perianth persists as a similar investment (e.g. Morus). In Amaranthacese the pedicels are furnished with bracts which resemble the perianth- segments and, like them, envelop the fruit. In those Centrospermse which are destitute of floral-leaves (achlamydeous), e.g. the Piperacese, the floral envelopes are replaced by bracts. The Caryophyllacese have distinct calyx and corolla, and are often placed in the sub-class Polypetalee; they have, however, decided affinities with' the Centrospermse. The androecium is composed of one whorl in Urticaceae and Chenopodiaceae, and of two whorls in most of the other families. The filaments are inflexed in the bud in Urticacese, but spring up when the perianth opens and so eject the pollen from the anthers (see p. 306). Most of the Centrospermse have dust-like pollen, but in Caryophyllacese and Nyctaginacese the pollen is adhesive. In Urticacese and in some Caryophyllacese the embryo is erect, in the rest it is horse-shoe shaped or spirally curved (see fig. 426^). The copious farinaceous endosperm of some Polygonacese and Chenopodiacese (Polygonum Fagopyrum, P. TataricuTn, Ghenopodium Quinoa) is used for flour. The Centrospermse are -distributed in every quarter of the globe. The Piperacese, Urticacese, Polygonacese, Amaranthacese, and Nyctaginacese are developed in the greatest variety in the tropics. Most Centrospermse, however, are found in the temperate zones. The Mediterranean Flora is especially rich in Caryophyllacese, whilst Equatorial America abounds .particularly in Amaranthaceee and Nyctaginaeeas. The Polygonacese grow chiefly on the banks of streams; the Chenopodiacese are very prevalent by the sea-shore and on salt steppes, especially in Central Asia. Several Caryophyllacese flourish also on the confines of perpetual snow. Silene acaulis (see Plate XII.) is ■one of the most remote outposts of the Phanerogamia and has been met with in Franz Joseph's Land at 81° north latitude, and in the Central Alps at a height of 3160 metres above the sea-level. Fossil remains of Urticacese and Piperacese have been recognized in the deposits of the Mesozoic and Tertiary periods. The number ■of species now living amounts to about 4200. Alliance XXXVI.— Proteales. Family: Proteacece. Perennial herbs, with underground stems which project but little above the ■earth, or herbs and small trees with entire or variously lobed and incised stifi' foliage-leaves without stipules. Flowers in capitula or spikes (see fig. 426 ^); actino- morphic or zygomorphic, hermaphrodite, pseudo- hermaphrodite, monoecious and dioecious. Perianth 4-partite, petaloid; the four segments are connate at the base, and, in the bud, have their free ends closed together like valves (see fig. 426 2). Ovary superior, free, unilocular. Number of ovules one to many. Placentation parietal. Each ovule has a double integument; the micropyle is directed towards the base of the ovary. The latter is surrounded by tissues which secrete honey. The number of stamens is equal to that of the perianth segments; the short filament is adnate to the perianth-segment behind it (see fig. 426 ^). Fruit a drupe, a nut, ANGIOSPERMiE, DICOTYLEDONES. 751 a capsule or a follicle (see figs. 426 * and 426 «, and fig. 324, p. 429). The seed contains an embryo furnished with two large, thick, fleshy cotyledons, but no endosperm. The Proteales are for the most part much-branched shrubs. The arboreal Kg. 426.— Proteales. I Banksia ericifolia. 2 Single flower otBanksia littoralis with the spoon-shaped perianth-segments still closed. 3 Longitudinal section through the same flower ; the style is in the form of a barbed hook, and the stigma rests between the anthers ; the filaments are adnate to the concave surfaces of the spoon-shaped pSrianth-segments. * Fruiting spike of Banksia erici/oHa. 5 Fruit of Xylomelum pyri/orme. 2 and ' magnified; the rest nat. size. (After Baillon.) species Knightia excelsa, a native of New Zealand, attains a height of 30 metres. The foliage-leaves are sometimes glabrous and sometimes clothed with scales, and they possess peculiar stomata (see vol. i. p. 296). The genus Rakea exhibits in 752 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. some of its species pinnate and bipinnate leaves, in others cylindrical and needle- shaped leaves. The flowers, which are crowded together, are sometimes surrounded by an involucre of many scales reminding one of the involucre of Compositse. The ovary is often borne on a special stalk. The style and stigma are very various. In many species, as, for instance, in Banksia ericoides and B. littoralis (fig. 426^), the style is hooked and breaks through the perianth-tube in consequence of the excessive longitudinal growth of its lower region, but the capitate extremity remains between the anthers, which are adherent to the spoon-shaped extremities of the perianth. Only when the perianth segments disunite and open back does the top of the style become free. The pollen is often deposited upon the end of the style with- out, however, immediately reaching the stigmatic surface, and in many species there are special hairs or brushes for collecting the pollen, whilst pockets and recesses for its temporary reception also occur. The Proteales flourish chiefly in regions where a short rainy season alternates a long rainless period. Australia and the south-west district of the Cape are richest in species; the alliance is represented by many fewer species in the tropical region of South America, in Chili, in New Caledonia, in New Zealand, in the tropical parts of Eastern Asia, in Madagascar, and in the mountains of tropical Africa. Fossil remains of Proteacese occur in strata of the Tertiary Period. The number of existing species is about 1000. Alliance XXXVII. — Daphnales. Families: Mceagnacece, ThymelacecB, Lav/racecB. Annual and perennial herbs, shrubs, and trees with green foliage, or leafless parasites. Stipules absent. Flowers in fascicles (see fig. 427 ^); actinomorphic, her- maphrodite, pseudo-hermaphrodite, or dioecious. Perianth of one or two whorls of 2, 3 or 5 leaves each, sepaloid or petaloid. Gynseceum 1-3 carpellary. Ovary unilocular. Style single, free, at the bottom of a cup-shaped receptacle from the margin of which spring the perianth-leaves (see fig. 427 ^). Ovule solitary. An- droecium 1-4 whorls with 2-4 stamens in each inserted on the inner margin of the cup-shaped receptacle. Fruit a one-seeded berry, drupe, or nut. The seed contains no endosperm. The embryo is furnished with large, fleshy cotyledons. The Cassythse, belonging to the family Lauracese, are parasites poorly suppKed with chlorophyll, with thin twining stems and squamiform leaves. Most of the Daphnales, however, develop woody stems with leafy branches. The leaves of Elaeagnacese are clothed with scaly covering-hairs (see vol. i. p. 322, fig. 78^). The foliage-leaves of most Lauracese exhibit a curious distribution of the strands in the laminae (see vol. i. p. 631, fig. 149*, and accompanying fig. 427 1). The majority of Laurace^ contain ethereal oils and aromatic substances. Especially to be mentioned in this connection are the Bay-Laurel (Laurus nobilis), the Cinnamon-tree (Cinna- Tnomum Zeylanicum), and the Camphor-tree {Camphora officinarum). In the Sea Buckthorn (Hvppophae, see p. 109, fig. 220), the perianth is 2- and the andrcecium 4-membered; in Ulceagnus the perianth and andrcecium are each composed of two ANGIOSPERM^, DICOTYLEDONES. 753 2-membered whorls; in Daphne the perianths has two 2-membered whorls, and the andrcecium two 4-membered whorls (see fig. 427 '); in Laurus the perianth consists of two and the andrcecium of four 3-membered whorls; in the genus Gnidium there are two kinds of floral-leaves, the lower ones sepaloid the upper petaloid in colour, and these are spoken of as calyx and corolla. The same arrangement is found in several Lauraceee. The anthers of Elseagnacese and Thymelaceae dehisce by longitudinal slits, those of Lauracese by valves (see fig. 427 ^). In the ElEsagnacese the cup- shaped receptacle persists as an envelope around the fruit, and becoming succulent Kg. 427.— Daphnales. I Camjphora ojicinarum (Family Lauraceas), flowering branch. 2 Longitudinal section through the flower of Cmntmifynvum Zeytanicum (Family Lauraceee). 3 Flower of Daphne Mezerewm (Family Thymelacese) cut open and rolled back • reduced: 2 and s magnified. (Partly after Baillon.) outside and strong within, the result is a false drupe. In some of the Lauracese also, as, for instance, in Nectandra, the receptacle continues to grow with the fruit, and forms a cup-shaped envelope resembling the so-called cupule in the fruit of the Oak. In Thymelacese and Lauracese the ovule is pendulous (see fig. 427^), in Elseagnacese it is erect. The Daphnales are scattered over all parts of the earth. The Thymelacese are best represented in countries where the climate is temperate; the Cape and Australia are particularly rich in species of that family. Daphne striata attains its highest elevation in the Central Alps at 2500 metres. There is a striking concentration of several species of the genus Daphne on a strictly limited area in the mountainous parts of Southern Europe. One of these species is the plant known in Carniola under the name of the Konigsblume (Daphne Blagayana), VOL. II. 98 764 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. This name of King's Flower was popularly accorded to the plant because, in 1838, Xing Augustus of Saxony travelled to Carniola on purpose to see this rare species flowering in its restricted habitat. The Lauracese are principally tropical and sub- tropical plants; Eastern Asia, the Sunda Islands, and Brazil are especially rich in species of this family. The Lauracese reach their northernmost boundary below 50° in Eastern Asia, below 46° in Europe, and below 45° in North America. In the Southern Hemisphere the Lauracese range as far as 43° S. lat. Fossil remains of the Daphnales, especially of Lauracese, are found in the strata of the Mesozoic and Tertiary Periods. The number of existing species hitherto discovered is about 1400. Alliance XXXVIII.— Santalales. Families: Santalacece, Viscacece, Loranthacece, Olacacece, GrtibbiacecB. Herbs, shrubs, and trees, of which most are parasitic on the roots and stems of other green-leaved Phanerogams, although they are themselves capable of assimila- tion owing to the presence of chlorophyll in the foliage-leaves. The leaves have entire margins; there are no stipules. Flowers actinomorphic, solitary or in cymes, which are combined into spikes, racemes, umbels, and capitula; hermaphrodite, pseudo-hermaphrodite, monoecious, and dioecious. Perianth composed of 2- or 3- membered whorls; either sepaloid or petaloid. Gynseceum 2-3 carpellary; ovary sunk in the discoid or cup-shaped receptacle, inferior or semi-inferior, unilocular. Style single. Ovules 1-5, without integument. Stamens as many or double as many as the perianth-segments; in the former case they are inserted in front of those segments. Fruit usually a berry or drupe. Seed-coat either single or absent ; the embryo either partially or entirely surrounded by fleshy endosperm. For a description of the sinkers and haustoria of the parasitic Santalacese see vol i. p. 177 and pp. 205-213. Several of the Loranthacese have thin twining stems which put out roots, i.e. sinkers, from their nodes. Such of the Loranthacese as are destitute of green foliage-leaves have thickened and flat expanded branches. In certain Santalacese several bracts are united so as to form a cup-shaped involucre. In Grubbiacese and Olacacess the lower portion of the ovary is septate, at least in the first stage of development. In the Santalacese and some Olacacese 1-5 pendulous ovules are borne upon a cellular structure which is either adnate to the internal wall of the ovary or else rises freely in the cavity; in the Loranthacese they completely fill the ovary, and are united with the carpels into a solid mass. In Grubbiacese the stamens of the outer whorl alternate with the leaves of the perianth, and there are double as many stamens as perianth-segments. The stamens of the Mistletoe {Viscum album, see p. 87, fig. 214^^) are adnate to the perianth-leaves behind them, and their anthers have 6-20 loculi, each of which liberates pollen through a pore. The Santalales are widely distributed. The majority of the species are tropical and sub-tropical. The Olacacese only occur in the tropical parts of South America and Africa, and the Grubbiacese only at the Cape, whilst the Santalacese are chiefly natives of Africa and Australia. The ANGIOSPERMiE, DICOTYLEDONES. 755 Mistletoe (Viseum album, found in Scandinavia as far north as 59° 30') and several species of the genus Thesium reach furthest north. Thesium alpinum attains its highest limit in the Alps at 2400 metres. Fossil remains occur in the strata of the Tertiary Period. The number of existing species is about 750. Alliance XXXIX.— Rafflesiales. Families: RafflesiacecB, Apodanthacece, and Gyhnacece. Plants destitute of chlorophyll, parasitic on the roots of green-leaved woody plants. Flowering axis greatly thickened, fleshy, Flowers solitary or in racemes, hermaphrodite or pseudo-hermaphrodite. Perianth 4-6 partite. Ovary inferior. The cavity of the ovary is divided irregularly into chambers which are filled with strands and ridges bearing the ovules. Above the ovary rises a columnar style with a discoid thickening at the top, and upon the under margin of this disc the stigmatic tissue is situated. The stamens are inserted underneath the stigmatic tissue in a circle. The fruit is fleshy, baccate, and crowned by the persistent column. The seeds have hard coats. The embryo consists of few cells, has no cotyledons, and is surrounded by an oily endosperm. For a description of the suction-organs see vol. i. pp. 199-204, and for the size of the flowers see vol. ii. p. 185. The Eafflesiales live in the tropical and sub-tropical regions of both the Old and the New World; two species of the genus Gytinus (see vol. i. p. 201) belong to the Mediterranean flora. No fossil remains are known. The number of extant species hitherto identified is 29. Alliance XL. — Asarales. Families: Aristolochiacece, Asaracece. Perennial plants, some with subterranean tuberous or creeping rhizomes, some with twining liane-like stems (see vol. i. fig. 95 ^, p. 364). Foliage-leaves broad, with entire margins, sometimes lobed. Venation apical (see vol. i. p. 633). Flowers hermaphrodite, solitary, or in cymose inflorescences, especially in axillary fascicles. Perianth of 3 petaloid leaves, united at the base. Gynseceum 4-6 carpellary; ovary inferior or semi-inferior. Styles united into a column bearing a radiating stigma. AndrcEcium composed of 2-12 whorls of 3 stamens each. Ovules numerous in the loculi of the ovary. Fruit a capsule (see p. 431, fig. 325^). The seed contains an abundant endosperm, and a very small embryo with two cotyledons. The perianth in Asaracese is actinomorphic (see p. 279, fig. 279 ^^-i^), whilst in Aristolochiacese it is zygomorphic or else unsymmetrical, and the tube of the peri- anth is variously curved and inflated (see p. 166, fig. 242, and p. 226, fig. 257 ^' ''• ^' '). These flowers are very striking, on account not only of their form, but also of their dark-brown colour; moreover, in many cases they attain to an extraordinary size. Mention has already been made of Aristolochia gigas (see p. 185), and recently a Birthwort (Aristolochia Goldeana) has been found in West Africa which 756 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. has a perianth 66 cm. long and 28 cm. broad. In the Asaracese there are some- times three small teeth alternating with the three perianth-segments, and these are looked upon as reduced inner perianth-segments. The stamens of Aristolo- chiacesB are adnate to the stylar column (see p. 292, fig. 284 1^). The Aristolochiacese are distributed in all parts of the world. The majority of the species are found in tropical and sub-tropical regions. The genus Asarum reaches furthest north. The northern limit of the Asarabacca {Asarum Uuropceum) and its highest elevation are the same as those of the Beech. Fossil remains are found in the strata of the Mesozoic and Tertiary Periods. The number of extant species hitherto identified is about 200. Alliance XLI. — Euphorbiales. Family: Ewphorhiacece. Annual and perennial herbs, shrubs, and trees. Flowers in racemos^or umbel- late cymes; actinomorphic, hermaphrodite, pseudo-hermaphrodite, monoecious, and dioecious. Floral-leaves difierentiated into calyx and corolla. Calyx and corolla 3-12-merous. The corolla is often suppressed, and sometimes the calyx also is wanting. In these cases the floral-leaves are replaced by bracts and involucral leaves. The gynseceum is superior, and is composed of 3-20 carpels, which are arranged in whorls round a central column. The carpels are joined together to form a multilocular pistil. In the inner angle of each loculus are 1-2 pendulous ovules. The stamens vary in number from 1 to over 100. At the base of the flower are some peculiar glands, which are looked upon as outgrowths of the receptacle. They are either in the form of separate cellular structures, arranged in a whorl, or else are coherent in the form of a cup. In the cases where these structures do not occur they are replaced by similar glands, which are seated upon the margin of the cup-shaped involucre. The fruit is a schizocarp or drupe; sometimes it is baccate. The embryo is imbedded in an abundant fleshy endosperm. It is difficult to describe the Euphorbiales in few words on account of their extraordinary variety. Some of them contain watery juices; the majority are full of latex. Several of the laticiferous species are poisonous. Many have green foliage-leaves; whilst some are destitute of foliage-leaves, and assimilation is then effected by means of the green cortical tissue of switch-like or cactiform branches and phylloclades. In many genera, especially in Euphorbia, the inflorescences have the appearance of being single flowers. A large number of male flowers are assembled together within a cup-like involucre, the free edge of which is furnished with glands as though with petals. Each of these flowers consists, however, merely of a bract and a stamen, and in the midst of them is a female flower, borne at the end of a long stalk, and resembling a stalked ovary. In many species of the genera Groton and Poinsettia the inconspicuous flowers are surrounded by bright-coloured bracts and involucral leaves. In the majority of instances three carpels are devel- oped, which are remarkable for their rotundity. They are laterally coherent, and usually separate when mature, and become detached from the central column. The ANGIOSPERM^, DICOTYLEDONES. 757 Euphorbiales are distributed in every quarter of the globe. The majority are found in the tropics, and several arboreal species form entire woods in those regions. Some grow in marshy lowlands, whilst others inhabit steppes and the rocky declivities of mountains. Ewphorhia capitulata grows on the mountains of the Balkan Peninsula. Euphorbia Austriaca stretches as far as the alpine region of the Eastern Alps. Mercurialis perennis attains in the Alps an elevation cor- responding to the upper limit of the Beech-forests. Several annual species of Euphorbia are encountered as weeds in cultivated ground, as far as the limits of the arctic region. Fossil remains have not been definitely ascerta,ined to exist. The number of extant species hitherto identified is about 4000. Alliance XLII. — Podostemales. Family: Podostemacece. Perennial herbs with creeping roots which are fastened to the substratum. The shoots spring laterally from these roots, and, are clothed by small scales arranged in two or three rows; these leaves are either entire or pinnately lobed, and they are sheathed at the base. Not infrequently the shoots are transformed into phylloclades, and sometimes shoots and roots are fused together into a thalloid structure. In these cases the assimilation of carbon is effected by the phylloclades as well as by the green branches of the thalloid tissue clinging to the substratum. The branches in question are ribbon-shaped or filiform, and are submerged. The flowers occur singly at the ends of the shoots, or else are sunk in the margins of the phylloclades in rows, and together form a sort of flat club. They are actinomorphic and zygomor- phic, hermaphrodite, monoecious, and dioecious. The floral-leaves are small, greenish, squamous, free, or connate, and are arranged in a 3-5-partite whorl. When the floral-leaves are suppressed, they are replaced by sheathing involucral leaves. The gynseceum is composed of 1-3 carpels; the ovary is superior, and either unilocular or else divided by delicate partition- walls into three chambers. The ovules spring from cushions of tissue which project from an axial column in the ovary. The number of stamens varies greatly, the flowers being either monandrous, diandrous, or polyandrous. In the last case the stamens are arranged in several whorls. The anthers dehisce longitudinally. The fruit is a capsule. The seeds are very small, and do not contain any endosperm. The embryo has two thick cotyledons. The Podostemacese are foimd in running water, especially in waterfalls, clinging to rocks, loose stones, and stumps of trees which have been stripped of their bark. Almost all of them inhabit the tropics, the only exception being one species in South Africa and one in North America. No fossil remains have been found. The number of existing species hitherto described amounts to 175. 758 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. Alliance XLIII. — Viridiflorse. Families: Leitneriacece, Gannabinacece, Dorsteniacece, Artocarpacece, Ficacece, GonocephalacecB, Moracece, Ulmacece Annual or perennial herbs, shrubs, and trees. The laminae of the foliage-leaves veined with pinnate or radiating bundles. Flowers in glomerate, fasciculate, or spicate cymes; actinomorphic, hermaphrodite, pseudo-hermaphrodite, moncecious, and dioecious. Perianth composed of 2-8 inconspicuous greenish segments. Gynseceum superior, 1-2 carpellary and unilocular. Ovule solitary, pendulous. Stamens as many or double as many as the perianth-segments; all or those of the outer whorl are Fig. 428.— ' Living bndge formed of the aerial roots of the India-rubber and other kinds ol Figs in Sikkim-Hinialaja. (After Hooker.) inserted opposite the segments of the perianth. The pollen in dusty. The fruit is a one-seeded achene or a drupe. The cotyledons of the embryo are not thickened. The UlmaceEe and Cannabinacese contain watery juices; the plants of the other famiUes produce milky juice in greater or smaller quantity. The milky juice of the Indian Fious elastica is used in the manufacture of caoutchouc, that of the South American Cow-tree (Galactodendron utile) as an article of diet. The milky juice of the Upas-tree (Antiaris toxicaria), native to Java, contains poisonous substances. In the Hop {Humulus Lupulus) and in the Hemp {Gannahis sativa) bitter and ANGIOSPERMiE, DICOTYLEDONES. 75& Fig. 429,— Amentales. ' Birch (Betula alba) shoot with male and female catkins (the former at the apex). 2 Ripe female catkin of same. 8 Winged nut of same. * Subtending scale of fruit of same. « Shoot of Hornbeam (Carpitms Betidus) with male and female catkins (latter to right). « Scale of female catkin with flowers of same. ' Scale from male catkin with stamens, s Scale of female catkin with ripe fruit. », *, ", 1 enlarged ; the rest nat. size. 760 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. aromatic substances (lupulin and hashish) are produced in special cells and groups of cells. The curious tabular and columnar roots of the Ficacese have been described in detail and illustrated in vol. i. pp. 755-757. Here is represented a Ficus (fig. 428) with aerial roots, which Hooker saw used by the natives in the Himalaya as bridges. " The property of the fig-roots, which inosculate and form natural grafts, is taken advantage of in bridging streams, and in constructing what are caUed 'living bridges' of the most picturesque forms." The axis of the inflorescence is thickened in many cases, notably in the Dorsteniaceee, Artocarpacese, Ficacese, Conocephalacese, and 7 Fig. 430. — The Oak {Quercus sessUiJtora). 1 Cluster of female flowers. ' Single female flower, s Longitudinal section of a female flower showing the ovary with ovules, small perianth and the young cup or cupule. * Three male flowers. 5 Cluster of nuts (acorns) with their cupules. ^ The seed, 7 Longitudinal section of seed, s Transverse section of seed, i, ^, s, ^ enlarged ; rest nat. size. Moracese; sometimes it is discoid, and sometimes hollowed out into the shape of an urn (see p, 157), It also enters into the structure of certain fleshy succulent collec- tive fruits, which afibrd such an important article of food in hot countries, viz. the figs of Ficus, and the Bread-fruit of Artocarpus incisa. Sometimes the perianth also takes part in the formation of the fruit, as, for instance, in the Mulberry (Morus), where it is converted into a fleshy envelope inclosing the fruit. On the other hand, in several other Viridiflorse the perianth is almost completely suppressed. In DorsteniacesB and Moracese the stamen-filaments are inflexed in the bud, and spring up after the perianth has opened, scattering the pollen-dust in the air (see fig. 229, p. 137). The filaments are straight in the flowers of the other families. The embryo is curved in most cases. There is either a very small quantity of endosperm or none at all. The Viridiflorse are found in all parts of the world. The Ulmacese es;tend ANGIOSPERM^, DICOTYLEDONES. 761 Fig. 431.— The Beech (Fajrus syUaUca). M >^ ...«.^-=-^ 762 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. beyond 66° north latitude and 36° south latitude. Representatives of Viridiflorse are most abundant in the tropics. Fossil remains occur in the deposits of the Mesozoic and Tertiary Periods. The presence of leaves of a Bread-fruit tree (AHo- carpus Dicksoni) in Greenland chalk, and of quantities of remains of Ulmacese (Geltis, Zelkova, Ulmus) in the Miocene strata, are points of special interest. The number of species ascertained to exist at the present day is about 1000. Alliance XLIV. — Amentales. Families: Betulacece, Corylaceoe, Cwpuliferoe, Juglandacece, MyricacecB, Casuarinece, SalioacecB. Are all large shrubs and forest trees, forming a very conspicuous feature in the landscape in temperate climates. The flowers are unisexual and arranged in catkins or heads. Pollination is by wind, and, in the majority of cases, a perianth is wanting. In the Cupuliferse, in which it is present, it is inconspicuous and con- solidated with the ovary. The group is an exceedingly interesting one, owing to the recent discovery of several instances of chalazogamic fertilization within its limits; of this, full details were given at p. 413. In this alliance are included such familiar trees as the Birch {Betula, figs. 429 1-^, 3,4)^ Alder {Alnus, p. 135), Hazel {Gorylus, p. 147), Hornbeam {Garpinus, p. 433 and figs. 429 ^■^■^'^), Oak (Quercus, p. 298 and fig. 430), Beech (Fagus, fig. 431), Chestnut {Castanea, p. 445), Walnut (Juglans), Sweet-gale (Myrica), Casuarina, and the Willow (Salix, pp. 299 and 424) and Poplar (Populus). Between 500 and 600 species have been distinguished. Members of this alliance date far back amongst the secondary rocks, and they are supposed by many to represent a primitive group of Angiosperms. Alliance XLV. — Balanophorales. Families: Hydnoracece, SarcophytaceoB, Cynomoriacece, Balanophoracece, Scybaliacece. Plants destitute of chlorophyll, parasitic on the roots of green-leaved woody plants, with tuberous, cylindrical, or angular stems with flowering lateral branches. The flowering axes are thickened and fleshy; the flowers hermaphrodite or pseudo- hermaphrodite, monoecious or dioecious. Perianth of 2-8 segments, sometimes transformed into a cup on the top of the ovary, or absent and replaced by scaly bracteoles and hairs. Gynseceum 1-3 carpellary; ovary inferior, unilocular, styles either absent or 1-2 in number, filiform, and terminating in small stigmas. Stamens 1-60 inserted below the limb of the perianth. Fruit a kind of berry, nut, or drupe. The embryo is very small, has no cotyledons, and is imbedded at the top of a fleshy, oily endosperm. For a description of the suckers and scale-leaves, see vol. i. pp. 186-198. The flowers of Hydnoracese are solitary; those of Balanophoracese, Cynomoriacese, and Scybaliacese are crowded in large numbers on unbranched clavately-thickened axes. ANGIOSPERM^, DICOTYLEDONES. 76S and those of Sarcophytaceas on branched clavate axes. In Hydnoracese the cavity of the ovary is occupied by numerous ridges which project from the walls and bear the ovules; in Sarcophytaceae, Scybaliacese, Cynomoriacese, and Balanophoraceae the placentation is parietal, and the number of the seeds is 3 in Sarcophytaceaj, 2 in Scybaliaceffi, 1-3 in Cynomoriacese, and 1 in Balanophoracese. In Hydnoracese and Sarcophytacese there is no style, and the free upper extremities of the masses of tissue which bear the ovules act as stigmas. The Cynomoriacese and Balanophoracese have one, and the ScybaUacese two, fihform styles with small papillose stigmas. In Hydnoracese the stamens are inserted between the lobes of the perianth, and form a fleshy ring; in the other families they stand in front of the segments of the perianth; in the Balanophoracese the filaments are connate. Most of the Balanophoracese live in the tropical parts of Asia and America; a few species inhabit South Africa and New Holland. Cynomorium coccineum, the only species of the Cynomoriacese, grows in the Mediterranean area and in Western Asia (see vol. i. p. 197, fig. 42). Fossil remains are not known. The number of extant species hitherto discovered is about 45. Sub-Class II.— MONOPETALtE. Alliance XLVI. — Caprifoliales. Families: Ruhiaceoe, Caprifoliacece Annual and perennial herbs, shrubs, and trees. The foliage-leaves are opposite, stipules are present at their bases (see fig. 432^). The flowers are in cymes, actinomorphic and zygomorphic, hermaphrodite and pseudo-hermaphrodite. The floral-leaves are differentiated into calyx and corolla. The calyx is composed of one 2-6-sepalous whorl. The calyx-tube clothes the inferior ovary, whilst the limb consists of small green teeth. The corolla is a whorl of 3-6 connate petals (see fig. 432^). The gynseceum is composed of 2-5 connate carpels,; ovary inferior, 2-5 locular. The placentas are axile. The andrcecium is a whorl of ,3-6 stamens, adnate to the corolla-tube. The pollen is either adhesive or powdery (see p. 265).. The fruit is a berry, drupe, schizocarp, or capsule. The seed contains endosperm. Most of the Eubiacese are herbaceous, whilst the species of the other families are mostly shrubby and arboreal plants. In the roots of several Rubiacese (e.g. Rubia tvnctorum, and Oalium boreale) there is a red colouring matter (madder- red); the Coffeacese and Cinchonacese contain alkaloids (caffeine, quinine, &c.); the sweet-scented Woodruff (Asperula odorata), the herb used to make the German May- wine, is famous for the kumarin it contains. No laticiferous tubes or latex, however, are contained in the tissues of any species belonging to this alliance. The foliage-leaves are always opposite and in pairs, which are at right angles to one another; the venation of the laminae is pinnate. In the Stellatse section of Rubiacese the stipules are of the same size, colour, and form as the laminse of the opposite leaves to which they belong, and are inserted between them. The conse- quence is that at each node there is a whorl of leaf -structures arranged in the form 764 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. of a star. In the Cinchonaceae and Coffeaceae, the stipules are squamiform, and sometimes lacerated (see fig. 432 ^). In the CaprifoliaiceEe they are either very small and in the form of stalked glands, or else they are adnate to the base of the petiole, and have the appearance of being narrow sessile segments of the leaf. The cjrmose inflorescences may be contracted into glomerules and fascicles, in which case each is surrounded by an envelope of bracts, as, for instance, in the Ipecacuanha plant {Gephaelis Ipecacuanha; see fig. 432 ^), or they may form pyramidal panicles, as in the Cinchona, or, lastly, they may be flat cymes, as in the Elders (Sambucus nigra Fig. 432.— Caprifoliales: Cephaelis Ipecaeuanna CPamily Rubiaceaj). 1 Entire Plant. 2 Inflorescence, s Single flower, i reduced; » and s magnaed. (Alter Baillon.) and 8. Ehulus). In the Caprifoliacese, especially in the genera Linncea and Lonicera, two-flowered cymes also occur, and in several species of the genus Lonicera, the ovaries of the two flowers in each cyme are connate. The flowers of several Caprifoliacess (Linncea, Lonicera, &c.) are zygomorphic, whereas the other members have actinomorphic flowers. In the Stellatee the fruit is a schizocarp which breaks up into two mericarps; in Cinchonacese, it is a capsule which dehisces from the base upwards (see p. 431, fig. 325 1"). In Coff'eacese, Sambucaceee, and the genus LinncBa, the fruit is a drupe, and in Gardeniese and some of the Caprifoliacese, a berry. The baccate fruits of several species of the genus Lonicera {L. alpigena L. coerulea, &c.), coalesce to form a collective fruit. Each chamber in the fruit of ANGIOSPERM^, DICOTYLBDONES 765 Rubiacese, Coffeaceae, Sambucacese, and of the genus Linnoea, contains one seed, m most Caprifoliacese, as in the genus Lonicera, several, and in the CinchonaceEe, many seeds. The seeds of Cinchonacese are winged (see p. 423, fig. 318 ''). The Caprifoliales are distributed over all parts of the earth. The Cofieaeese and Cinchonacefe are chiefly tropical plants, whilst the Rubiacese, Sambucacese, and Caprifoliacese belong principally to the North Temperate Zone. The Cinchona is found wild only in the Cordilleras in South America (from 10° north lat. to 22° south lat.). Tropical Africa is supposed to be the original home of the Cofiee-tree {Goffea Arabica). Linncea borealis, a plant named after the Swedish botanist, Linnaeus, is scattered over the Alps, in the low-lying part of Germany adjoining the Baltic, and in Scandinavia. Several species of the genus Galium, of the family Rubiacese, belong to the flora of the extreme North and of high mountains. Foesil remains have been preserved in the deposits of the Mesozoic and Tertiary Periods. The number of extant species discovered up to the present time is about 4800. Alliance XLVII. — Asterales. Families: ValeriaTiaceoB, Bipsacece, Calyceracew, Brunvniaceoe, CompositcB. Annual and perennial herbs, shrubs, and trees. Foliage-leaves extremely various in form but always destitute of stipules. Inflorescence a cyme or a capitulum. Flowers actinomorphic and zygomorphic, hermaphrodite, pseudo- hermaphrodite, monoecious, and dioecious. Floral-leaves differentiated into calyx and corolla. Calyx of 2-5 sepals; the limb, which crowns the inferior ovary, is in the form of a pappus, bristles, scales, teeth, callosities, or membranous borders, and is destitute of chlorophyll. The corolla is composed of 4-5 connate petals. The gynseceum consists of 2-3 connate carpels. The ovary is inferior and contains only one devel- oped loculus with a single ovule in it (see p. 73, fig. 207 ^). The andrceciuni consists of 1-5 stamens. The filiform filaments are adnate at the base to the corolla-tube. The fruit is a unilocular, one- seeded achene. The plants belonging to this alliance exhibit for the most part herbaceous growth, but some Compositse are shrubby (e.g. Baccharis), and some arboreal (e.g. Vanillos- mopsis, Lychnophora). Several Valerianacese and Compositse, e.g. the Dahlia and Jerusalem Artichoke (Dahlia variabilis and Helianthus tuberosus), are distin- guished by underground tuberous structures. The inflorescence in Valerianacese is a much-branched cyme (see p. 305, fig. 289 ^). In Dipsacese also the arrangement of the flowers is cymose, but the cymes are usually grouped together in capitula (see p. 121, fig. 225^). In some genera, such as Morina, they are arranged in opposite fascicles in the same manner as in Labiatse. The flowers of Compositse are situated at the extremity of a thickened axis which is conical, hemispherical, or flat, and compressed, as the case may be; they are spirally arranged and are grouped to- gether in capitula (see p. 242). In many cases they spring from the axils of scales ("palese"), or else their place of origin is surrounded by bristles. Not infrequently they spring from little depressions, and then the axis is seen to be pitted when the 766 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. flowers have fallen off The number of flowers in a capitulum varies greatly. In many species several hundreds of flowers (florets) are crowded together, in Adeno- styles and Eupatorium (see p. 320, fig. 294 ^) there are only a few flowers in each capitulum, and in Echinops it is limited to a single one. The capitulum is sur- rounded by an involucre of bracts crowded together. The form of these involucral leaves exhibits extreme variety. In Thistles their apices are transformed into prickles, in the species of the genus Xeranthemum, Helichrysum, &c., they are like paper or parchment, dry, and distinguished by white, yellow, violet, and red colours. They preserve these characteristics unchanged even when dried, and can therefore be made up into bouquets and wreaths which do not fade. These composite flowers, which are known as " immortelles ", are everywhere used as symbols of immortality and as memorial tokens. The Cape is exceptionally rich in Everlasting Flowers; among the species found there is Helichrysum eximium. The Edelweiss {Gnaphalium Leontopodium, see vol. i. p. 315, fig. 76) may also be looked upon as an immortelle, although here the bracts are not themselves dry and membranous, but are only covered with a dry, white felt of hairs. In many species the capitula are themselves grouped in capitula or glomerules. One of the most striking instances of this is afforded by the species of the genus Haastia, which are shown on p. 188. In the genus Echinops a large number of one-flowered capitula are grouped together in spherical heads, usually of a steel-blue colour. The capitula often look like single flowers, and in former times they were looked upon by botanists as compound flowers {flores compositw), whence the name of Compositse. In many species, e.g. the Sun-flower (Helianthus annuus), the capitula attain to a diameter of 40 centi- metres. In the actinomorphic corollas a tube and a bell-shaped 5-partite limb may be distinguished (see p. 360, figs. 302 1- ^' ^). The zygomorphic flowers are either two- lipped, the upper lip being composed of one or two petals, and the under lip of four or three petals, or else ligulate, in which case the tube is greatly abbreviated and the free end of the ligule usually exhibits five segments or teeth (see p. 121, fig. 222*, and p. 236, fig. 261 ^). In Valerianacese the corolla is usually produced on one side into a kind of sac, which in the genus Valeriana is short and blunt (see p. 289, fig. 283 '), and in the genus Centranthus is in the form of a long, slender, pointed spur (see p. 240, fig. 263 ^ and p. 306, figs. 289 ^' ^). In the capitula of Compositse the flowers with tubular, ligulate, and bilabiate corollas respectively are grouped to- gether in a great variety of ways. It is not common for all the flowers of a capitu- lum to have tubular flowers, but that is sometimes the case (e.g. Eupatorium, p. 320, fig. 294^); much more frequently all the fiowers in a capitulum have ligulate corollas (e.g. Hieracium, p. 112, fig. 222 ^), and in the majority of instances the flowers in the middle of the capitulum are furnished with tubular corollas, and those near the periphery with ligulate or bilabiate corollas (see p. 360, fig. 302 ^). The distribution of the sexes has been dealt with on pp. 295-297, and pp. 318-321. In the Compositse the anthers of the five stamens are united into a tube. The anthers are not connate in the other families. In Dipsacese the androecium consists usually of four stamens, and in "Valerianacese usually of three stamens (see. p. 289, ANGIOSPEBM^, DICOTYLEDONES. 767 fig. 2838); the genera Morina and Fedia have two stamens in each flower, whilst the genus Centranthus (Red Valerian) has only one (see p. 240, fig. 263 2). For a description of the pollen of Compositfe see p. 99. The gynaceum in Valerianacese is composed of three carpels, and the ovary is originally 3-locular, but two of the carpels are abortive, and only the third loculus is completely developed. In the other families the ovary is unilocular from the first. The ovule and the seed result- ing from its development is pendulous (see p. 178, fig. 249, and p. 240, fig. 263 2) in Dipsaceffi and Valerianacese, basal (see p. 73, fig. 207 s) in Compositffi. In most cases the calyx remains adnate to the mature fruit and assumes the form of a crown of hairs or bristles, which is termed a " pappus " (see p. 432), or else constitutes a membranous limb. In a later chapter we shall deal with the significance of these structures. In the Dipsaceee the fruit is surrounded by a saccate involucre called an involucel. The alliance is distributed over all parts of the earth; its members flourish both in the tropics and in the arctic regions, and are met with on the sea- shore and by the side of glaciers, in bogs and on arid ground, in shady woods and on sandy steppes. The greatest number are natives of the North Temperate Zone. In the Himalayas several Composites occur at an elevation of 4500 metres. Fossil remains have been found in small quantities in the deposits of the Mesozoic and Tertiary Periods. The number of extant species identified up to the present time is about 10,700. Alliance XLVIII. — Campanales. Families: Campanulacece, Lobeliacece, Stylidiacece, Goodeniacece. Annual and perennial herbs with entire exstipulate foliage-leaves arranged spirally. Flowers in capitula or racemes, or else solitary; actinomorphic or zygo- morphic, hermaphrodite or pseudo-hermaphrodite. Floral-leaves diflerentiated into calyx and corolla. Calyx of one whorl of 3-8 sepals, corolla of one whorl of 3-8 petals. The calyx-tube clothes the inferior ovary, and the calyx-limb is in the form of 3-8 comparatively large, green segments which crown the top of the ovary. The petals are joined. The gynseceum is composed of 2-5 connate carpels; the ovary is inferior and 2-5 locular. The ovules are numerous, and are borne on axile placentas. The andrcEcium consists of one whorl of 3-8 stamens, which are attached to the bases of the petals. The filaments are free; in the young flower the anthers are in close contact, forming a tube surrounding the style (see p. 360, figs. 302 1"' ^^). Sometimes they are connate, and in that case the tube persists even when the flower begins to fade. The pollen is adhesive. The fruit is a capsule (see fig. 340 \ p. 448). All the Campanales have laticiferous tubes running through them, and in several species the leaves and stems are copiously supplied with latex. The flowers are actinomorphic in Campanulaceee, zygomorphic in the other families. In the Stylidiacese, only two of the stamens develop pollen capable of effecting fertilization, whilst three stamens are abortive; in the other families all the stamens produce 768 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. pollen, which ripens effectually. The Campanales are distributed over all quarters of the globe. The Campanulacese are mostly natives of the North Temperate Zone, the Lobeliacese of the South Temperate Zone and the Tropics. Some Campanulaceffi are also found amongst the flora of the Arctic regions, and of high mountains. The Stylidiace£e and Goodeniacese are confined to Australia. No fossil remains have been discovered. The number of identified species now living is about 1300. Alliance XLIX. — Ericales. FamiHes: Biapensiaceoe, Pyrolacece, Monotropacece, Lennoacece, Arbutacece, RhodoracecB, Ericacece, Epacridacece, Empetracece, Ebenacece, Sapotacece. Perennial herbs, shrubs, and trees. Flowers actinomorphic and zygomorphic, hermaphrodite, pseudo-hermaphrodite, monoecious, and dioecious. The floral-leaves arranged in two 3-6 partite whorls. The lower whorl constitutes a calyx, the upper a corolla. The petals are free in Pyrolaceae and Monotropacese; in the other families they are coherent, at any rate at the base. The gynseceum is composed of 3-10 carpels; the ovary is superior, and 3-10 celled. The ovules arise from an axile placenta. The androecium is composed of one or more whorls of 4-5 stamens each. In many cases some of the stamens are metamorphosed into gland-like structures. The fruit is a capsule, berry or drupe. The embryo is erect, and imbedded in the fleshy endosperm. The Diapensiacese and Pyrolacese (see figs. 277 '■ *- p. 273), are perennial herba- ceous, or suffruticose plants, which grow in black humus, and have stifij evergreen foliage-leaves; the Monotropaceae and Lennoaceae are parasites or saprophytes (see vol. i. p. 252), and are destitute of chlorophyll. The Arbutacese, Rhodoracess, Ericacese, Epacridacese, and Empetracese are, for the most part, dwarf shrubs; only a few are trees. Erica arborea, one of the Mediterranean Flora, when able to grow without hindrance, attains the height of 8 metres. The branches of Ericaceae, Epa- cridacese, Empetracese, and of Loiseleuriaov Azalea procumbens, one of the Rhodoracese, are thickly covered with stiff', rolled leaves (see vol. i. p. 303 ^). The species of the genus Rhododendron have flat foliage-leaves (see Plate X.), as have also the Arbutacese (see fig. 433^). The leaves of Arctostaphylos alpina, Azalea Pontica, and A. mollis are green in the summer only, whilst most of the Arbutacese and Rhodoracese have flat, evergreen foliage. The Ebenacese sind Sapotacese exhibit, in a large proportion of their species, arboreal growth and leathery foliage-leaves. All the Ericales are distinguished for their solid timber. Some of the Rhodoracese have zygomorphic flowers; the rest of the Ericales have actinomorphic flowers. In Loiseleuria, and in the Empetracese and Epacridacese, the androecium is composed of one whorl; in the others it consists of two or more whorls. The anthers of Arbutacese and of many Ericaceae possess two peculiar horn-like appendages (see figs. 433 3. and 433*, and figs. 216 9.10.11. p. 91). In Epacridacese the anthers are unilocular, and dehisce longitudinally; in Ericaceae, Arbutacese, Rhodoraceas, and Pyrolacese they usually open by terminal chinks or pores (see p. 91, figs. 216 8' *> i"- H' i^). ANGIOSPERM^, DICOTYLEDONES, 769 The pollen is dusty in Ericaceae, but adhesive in most other cases. The pollen-cells are united in fours in EricacesB, Rhodoracese, and Pyrolacese, and in the Rhodoracese these groups are connected by tough threads (see figs. 219 2- ^'^' p. 101). The fruit is capsular in Diapensiacese, Pyrolaceae, Monotropacess, Rhodoracese, Ericaceae, and Epacridacese, and baccate in Arbutacese, Empetracese, Sapotaceae, and Ebenacese. In the Lennoacese the fruit resolves itself into 10-28 one-seeded portions. The Ericales are distributed over the whole world; the Ebenacese and Sapotacese live chiefly in the tropics; the Lennoacese are confined to the southern half of North America, and the Epacridacese to Australia. The species of Ericaceae are most abundant at the Cape. Most of the species of the genus Rhododendron inhabit Fig. 433.— Ericales: Arbutus CTnedo of tlie lamlly Arbutacese. 1 Flovfering branch. 2 Three flowers magnified. « Longitudinal section through a flower. * Flower from which the corolla has been remoyed. 6 papillose berry. 2, s and * magnified. (After Baillon.) the mountains of Central Asia, e.g. the Himalayas. The genus Kalmia belongs to the mountains of North America. The Diapensiaceae live in the arctic regions, as also do several Ericaceae. Loiseleuria or Azalea proGumbens is widely distributed in the arctic regions, and also occurs in exactly the same form on the mountains of Central and Southern Europe; in the Central Alps it attains its maximum elevation of 2700 metres above the sea-level. Most of the Ericales grow sociably on rocky declivities in mountainous districts, and on sandy soil in plains. Many only flourish on moorland, or when rooted in a deep layer of humus, and these play an important part in the formation of peat. Fossil remains are found in the deposits of the Mesozoic, Tertiary, and Diluvial periods. The number of extant species known is about 2300. Vol. ii. 99 770 THE SUBDIVISIONS OF TJHE VEGETABLE . KINGDOM. Alliance L. — Vaociniales. Families: Vaociniacece, Oxycoccacece. Woody plants, presenting all gradations in form, from that of delicate dwarf shrubs lying upon the ground to that of stately trees. The foliage-leaves arranged spirally, exstipulate. Flowers in racemes and fascicles, or solitary; actinomorphic, hermaphrodite. Floral-leaves differentiated into calyx and corolla. The calyx is composed of a whorl of 4-6 sepals. The calyx-tube clothes the inferior ovary; the calyx-limb is in the form of short, green teeth, crowning the top of the ovary. The corolla consists of a whorl of 4-6 petals; the petals are united or free. The gynseceum is composed of 4-6 connate carpels. The ovary is inferior and 4-6- locular. The placentas are axile. A honey-secreting tissue is situated on the top of the ovary. The androecium consists of two whorls with 4-6 stamens in each. The stamens surround the nectary, and are free from one another and from the corolla. The members of the outer whorl are opposite the petals. The fruit is a berry or a drupe. The seed contains a fleshy endosperm. The Vaociniales have no laticiferous tubes or latex. In Vacciniacese the petals are united, and the anthers are furnished with horn-shaped appendages, in Oxy- coccacese the petals are free, and the anthers have no horns. The Vaociniales are distributed in aU quarters of the globe, and in all latitudes. The species which belong to the Temperate Zones grow in peat-bogs and in the humus of woods and heaths, the species native to the mountains of tropical regions are, in some cases, epiphytic on the bark of old trees. Many are of social habit, and cover extensive tracts of ground. This is the case, for instance, with the various species of the genus Vaccinium: the Cow-berry (Vaccinium Vitis-Idcea), the Bilberry {Vac- cinium Myrtillus), and Vaccinium uliginosum. These species are also found within the area of the Arctic Flora. Vacciniv/m uliginosum ranges furthest to the North, and in Greenland forms with the dwarf Birch (Betula nana) and dwarf Willows, a low undergrowth which reaches to 73° N. Lat. They clothe the. moun- tain sides in the Central Alps as far as 2400 metres above the sea-level. Fossil remains have been found in the deposits of the Mesozoic, Tertiary, and Diluvial Periods. The number of extant species hitherto recognized amounts to about 360. Alliance LI. — Primulales. Families: Primulacece, Plumhaginacece, Myrsincuseoe. Annual and perennial herbs, shrubs, and small trees with alternate, opposite, and verticillate foliage-leaves. Fowers solitary, or in spikes and racemes; actinomorphic, hermaphrodite or pseudo-hermaphrodite. The floral-leaves are arranged in two whorls of 4-8 segments each. The lower whorl constitutes a calyx, the upper a corolla. The petals are coherent. The pistil is superior, 5-earpellary, unilocular. The ovules are supported in the middle of the ovary on a column of varying length ANGIOSPERM^, DICOTYLEDONES. 771 rising from the bottom of the ovary. The stamens, five in number, are inserted in front of the petals, and are adnate to them (epipetalous). The fruit is a unilocular capsule or drupe. The seeds contain an endosperm, in which the embryo is embedded. The ovary is surmounted by a single style in Primulacese and Myrsinacese, by five styles in Plumbaginaceae. The capsules of Primulaceae are many-seeded, those of Plumbaginaceae are one-seeded. In the genus Glaux only one floral envelope is developed. It has the appearance of a perianth, and resembles that of Polygo- naeese. It is interpreted as being a petaloid calyx. The fact that in Glaux the stamens occupy the same position in relation to the sepals as the petals do in other cases warrants our supposing that what is usually designated as the corolla in Primulaceae is only a whorl of stamens with connate petaloid filaments. The PrimulacesB are distributed mainly in the temperate zone of the Northern Hemi- sphere. Most of the species of the genera Primula, Soldanella, and Avdrosace are alpine plants. The Alps and the Himalayas are particularly rich in these species. Androsace glacialis (see fig. 221 «) occurs in the Alps in the neighbourhood of glaciers at a height of 3160 metres above the sea-level Primula pubescens, a plant obtained by Clusius in 1582 from the Gschnitzthal in Tyrol, was the original species from which Auriculas were derived during the fashion for their cultivation which prevailed in the seventeenth century. The Plumbaginacesa are represented by large numbers of species on the shores of the Mediterranean and in the saline steppes of the East The MyrainaceaB grow exclusively in the tropics. Fossil remains of Myrsinacese are known amongst the deposits of the Tertiary period. The number a£ species now existing is about IIQO. ADiance LEL — ^Tubiflois. Families: G&iitiaiutcem, AsclepieidacecB, Apoeynacece, Loganiaeeae, Gcmvolvulacece, Pt^Wifmiaeeai, Myd/FopkyllacecB, Boragmac&B, MolanacecB, Solanacece, ScropkulariaceoE, LentibulariaeecB, Bignoniacece, Acanthacece, Gesneracece, OrobanehaeecB, Globulariaceos, PlantaginacecB, Myoporacece, Verbenacece, LabiatoB, Oleacece, Jasminacece. Annual or perennial herbs, shrubs, and trees. Flowers actinomorphic and zygomorphic, hermaphrodite and pseudo-hermaphrodite. Floral -leaves in two 4-6-partite whorls; the lower whorl in the form of a calyx, the upper in the form of a corolla. Petals united. Gynseceum 2- or more celled, ovary superior. The ovules are developed either on the turned-in margins of the carpels or on an axile placenta. The androecium is composed of a whorl of 2-5 stamens. The fruit is either a succulent berry, a capsule with various modes of dehiscence, or a drupe. The Solanacese, Scrophulariacese, Loganiaeeae, and Asclepiadaceae contain poison- ous alkaloids, the Gentianacese contain bitter substances, and the Labiatae contain etherial oils and aromatic substances. The majority of Tubiflorae possess green foliage-leaves. Some Scrophulariaceae, e.g. the species of the genus Rehmannia, are 772 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. in the form of switch-shrubs, and several Asclepiadacese, e.g. the species o^ tl^^ gf^^« BUvdia, have cactiform steins. In these the assimilation of carbon is effected by the green cortical tissue. The Orobanchace^ are parasites destitute of chlorophyU (see vol i p 183). Amongst Convolvulaceae, and more especially amongst Scrophu- lariace^, there are many species which live as parasites and saprophytes, and are partially deficient in chlorophyll (see vol. i. pp. 171-183). An account has already Fig 434 — AcantliacedtJ Acanthv^ nioUis on the coast of Dalniatia. iSfe^ been given of the way in which the Lentihulariacese, e.g. the species belonging to the genera Utricularia and Pinguicula, derive a portion of their food from the bodies of insects which are caught by them (see vol. i. pp. 120, 140). In Gentianaceae, Oleacese, Apocynacese, Asclepiadacese, Convolvulacese, and many Boraginacese and SolanacesB the corolla is actinomorphic. The Labiatse, Scrophulariacese, Verbenacese, Acanthacese, Lentihulariacese, and some genera of Boraginacese and Solanacese bear distinctly zygomorphic flowers. In the Ash genus (Fraxinus), which belongs to the family of Oleacese, the corolla is often entirely suppressed. Most Labiatse have four ANGIOSPERMiE, DICOTYLEDONES. 773 didynamous stamens, but some of them, e.g. those of the genus Salvia (see fig. 271, p. 262), have two stamens, as have also the species of the genus Veronica (see fig. 257, p. 226) of the family Scrophulariaceae, and the majority of the Jasminaceae and Oleacese (see fig. 283 ^). Most of the Tubiflorse possess five stamens. The curious modification of the andrcecium of Asclepiadaceae has been fully described on p. 257, Fig. 435.— JBanunculaoece, ' Helleborus niger (reduced J). 2 Mvosurus minimug, complete plant with flowers and flower-buds (nat. size). * A single flower of Myoffurus (magnifled). et seq. In .the ApocynaceBe the two opposite carpels are separate at the base and connate at the upper end only. The fruit of Labiatse and Boraginacese resolves itself when it is ripe into four one-seeded nutlets. The seeds of Apocynacese and Asclepiadaceae are furnished with a plume of hairs. In most of th'e Tubiflorse the base of the pistil is partially or completely surrounded by swollen tissue which 774 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. secretes honey. The Tubiflorse are distributed in every quarter of the globe. Several families, such as the Loganiaeese and Bignoniacese, are confined to tropical and sub- tropical regions. The Acanthacese also chiefly inhabit the warmer parts of the earth. The genus Acanthus grows particularly in the region of the Mediterranean Flora. The leaves of several species of Acanthus, e.g. Acanthus spinosissimus (see vol, i, fig. 116, p. 437) and Acanthus mollis (see fig, 434), frequently served the Greek and Roman sculptors as patterns for their ornaments. The genus Stapelia, of the family Asclepiadacese, is confined to the Cape; the Labiatse are most abundantly represented in the Mediterranean Flora; the Gentianaceae and Scrophulariaeese inhabit mountainous regions of the Old and the New World in large numbers of different forms, and several species of the genera Gentiana, Veronica, Euphrasia, and Pedicularis thrive best in proximity to glaciers both in mountain districts and in the arctic regions. Fossil remains occur in the strata of the Tertiary period. The number of species now living which have been identified up to the present time is about 16,500, Class III.— POLYPETAL.^. Alliance LIII, — Ranales. Families: Ranunculacece, DilUniaceoB, GalycanthacecB, Ma^noliacece, Anonacece, MenispermaceoB, Berberidacece, Lardizabala^ece, NymphcBacece. Stamens rarely definite. Carpels, free or immersed in the receptacle, very rarely connate. Embryo minute, embedded in a fleshy endosperm. In the Ranunculaeeae the petals are not infrequently modified into honey-glands, and the sepals petalline. The carpels are free from one another, and sometimes indefinite and spirally ar- ranged, sometimes definite and whorled. In Galycanthus, the parts of the flowers are inserted in a continuous spiral upon a hollow receptacle, and pass gradually the one into the other. In Berberidaceae, the anthers open by means of valves. The Nymphseaceas include marsh and water plants (e.g. Nymphcm, Nu/pha/r, Ndumhium, cf. fig, 436, Victoria Regia, Plate XL). In several of these "the carpels are united together into a large ovary with shield-like stigmatic disc. In Nelumbium (cf. fig. 334, p. 440), the carpels are borne in distinct sockets. The fruit in the alliance is very varied, and includes achenes, follicles, berries. Fossil remains occur in the Tertiary Strata, Total number of living species about 3000 Alliance LIV, — Parietales, Families: Sarraceniaceos, Papaveracece, Fumariacece, Cruciferce, Capparidacece, Moringacece, Resedacm, Cistinece, Violacece, Bixaceoe. Annual and perennial herbs, shrubs, and trees, Flowers solitary or in spikes, umbels, racemes, and racemose cymes; actinomorphic and zygomorphic, herma- phrodite and pseudo-hermaphrodite. Floral-leaves diflTerentiated into calyx and corolla; the calyx composed of a 2-5-partite whorl, the corolla of two 2-partite ANGIOSPERM^, DICOTYLEDONES. 775. whorls or one 5-partite whorl. Petals free. Gyii£eceum composed of 2, 3, or several carpels joined together to form a unilocular (or spuriously bilocular), free, superior ovary. Ovules attached to the interior walls of the carpels on ridges, or springing directly from the walls of the ovary (see fig. 437 1). The androecium is composed of either one whorl or many whorls of 2-5 stamens; the stamens are free, and generally of equal length, and are not joined together or to the corolla (see fig. 243, p. 168). The fruit, in most cases, is a many-seeded capsule (see fig. 437 ^); in the genus Fumaria it is a small one-seeded drupe (see figs. 322 ^ and 322 \ p. 427). Sc^V3^V^t^ ' Fig. 436. — J^elumbium specwsum, the Indian Lotus, growing in a marsh, near Pekin (from a photograph). }f^*^^'^f!^^^=^ In the Capparidacese, it is borne on a long stalk. The Resedacese are interesting, in that in many of them the ovary is open from the beginning, the stigmatic tissue being formed by the swollen lips. The Cruciferse form a large and important family of over 200 genera. For systematic purposes they are divided into the following tribes: — Pleurorhizese, Notorhizese, Orthoploceae, Spirolobese, and Diploco- lobese. Annual or perennial herbs and suffrutices with the foliage-leaves in spirals, venation pinnate. Flowers in racemes, hermaphrodite, pseudo - hermaphrodite, actinomorphic and zygomorphic. Floral-leaves difierentiated into calyx and corolla, each of which is composed of two 2-merous whorls. Petals free. Ovary free, superior. The carpels spring from below the end of the conical receptacle, and are of two kinds: the two lower carpels bear no ovules, but form valves, whilst the two 776 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. upper are transformed into ribs and form a framework to which the valves are applied. The two superior carpels are separated by a thin membrane, and bear the ovules in two rows (see p. 75). The androecium is composed of two short and four , long stamens (see p. 292, fig. 284®). The pollen is adhesive. The fruit is a siliq[iaa (see p. 75 and p. 431, fig. 325 ^'-i^). The seeds have no endosperm. The embryo is curved. The cotyledons and also the foliage-leaves and . roots of most Cruciferse contain pungent and oily substances, particularly Oil of Mustard, as is well known in the cases of the Mustard-plant, Water-cress, Garden- cress. Radish, and Horse-radish. The Qruciferse are distributed over the Northern Hemi- sphere. They occur in greatest variety in the steppe-dis- tricts of the Old World. The Mediterranean, Arctic, and Alpine floras also include large numbers of these plants. Amongst those Phanerogams which survive at the very confines of vegetation in the Arctic regions, and on moun- tains, are to be reckoned several Cruciferae. No fossil remains are known. The Sarraceniacese are marsh- and water-plants, and their leaves are adapted to the capture of insects (see vol. i. p. 143 et seq.), whilst tlie rest of the Parietales grow chiefly on rocky or sandy ground. The Parietales are distributed over the warm and temperate parts of the Old and the New World; the Cistacese belong especially to the flora of the Mediterranean. The only known instance of fossil remains is the fruit of a Poppy which was found in a deposit of the Tertiary Period. The number of extant species hitherto identified is about 3000. Fig. 437.— Parietales. 1 Bj'xa Orellana (Bixaoeje). lon- gitudinal section through a flower-bud which Is about to open. 2 Argemone Mexicana (Papaveracese). Longitudinal section through the ovary. (Magnified.) Alliance LV. — Malvales. Families: Malvacece, Sterculiacece, Tiliacece. Flowers actinomorphic, parts in whorls of 5. Sepals free or connate, often valvate in bud. Stamens various, often united. Carpels 3 to indefinite, united. Placentation axile; seeds with endosperm. The Malvaceae often possess an epi-calyx; fruit usually splitting into 1-seeded mericarps. They include the Mallows (Malva), Cotton-plants (Gossypium), Hollyhock (Althcea), &c. The Tiliaceee are represented by the Linden (Tilia), and the Sterculiaeeae include Theobroma Cacao, from which chocolate is derived. There are nearly 200 species in this alliance. ANGIOSPERM^, DICOTYLEDONES. 777 Alliance LVI. — DisciflorsB. Families: Unace(B, Erythroxylacece, Oxalidacece, Humiriaceoe, Malpighiacece, Zygo- phyllacece, Geraniacece, Balsaminece, Tropceolacece, Rutacece, AuriantiacecB, I)ios7nace(B, Zanthoxylacece, SimarubacecB, Ochnacece, Burseracece, MeliacecB, IlicinecB, Celastrinece, Bhamnacece, AmpelideoB, Sapindaceoe, Aeervneoe, HippocastanecB, Sabiacece, Terebinthacece. Annual and perennial herbs, shrubs, and trees, with simple and compound foliage-leaves. Flowers actinomorphic and zygomorphic, hermaphrodite, pseudo- hermaphrodite, monoecious and dicecious; arranged in varying types of inflores- cence. Florai-leaves in tw6 4-5-merous whorls; the lower whorl a calyx, the upper a corolla. The gynseceum^is composed of a whorl of carpels borne on a swollen disc. Ovary superior. Each carpel has a separate loculus. In Aurantiacese and Ampelideae the carpels are (completely united so as to form a single pistil; in Rutacese and Zygophyllacese they are united at the base and form a lobed ovary, whilst in Zanthoxylaceae, Ochnacese, and Simarubaceae, they are quite separate (see fig. 438 ^). In Terebinthacese only one carpel is developed, but there are usually traces of suppressed carpels close to it. The ovules are in the inner angles of the loculi; in Aurantiacese, Rutacese, and ZygophyUacese their number exceeds two in each loculus, in the other families it is only 1-2. The stamens are arranged in 1-2 whorls, and number 4-5 in each whorl; they spring from the edge or from the surface of the floral receptacle, which is swollen and forms a ring round the ovary; their place of origin is always lower than the base of the ovary (see figs. 438 ^- ^' ''). The pollen is adhesive. The fruit contains either few seeds or a single compara- tively large seed. The Disciflorae are in most instances woody plants, containing etherial oils and aromatic, resinous substances like turpentine. Amongst the Malpighiacese, Celas- trinese, and Ampelideas are many lianes. The foliage-leaves are undivided in Erythroxylacese and Celastrinese (see fig. 438^), lobed in most Aceracese and Ampelidese, and variously segmented and compounded in the other families (see fig. 438?). The petals are usually small, and of a greenish-yellow colour. The filaments in Melianthacese and Aurantiacese are connate all together, or in groups. The fruits are extremely various. In Staphyleacese and Diosmacese they are follicles; in Celastracese and Rutacese, capsules (see fig. 325 ^ p. 431); in Zygo- phyllaceae, Aceracese, and Malpighiacese, schizocarps; in the Tree of Heaven, of the family Simarubacese, winged achenes (samaras, see fig. 323^, p. 428); and in Ampelidese and Aurantiacese, berries. The Disciflorse are distributed over the whole earth. The majority belong to the tropics, and several, e.g. Burseracese, Ochnacese, and Malpighiacese are exclusively tropical. The Diosmacese are conflned to South Africa, the Rutacese to the districts of the Mediterranean and the Black Sea. Comparatively few species occur in the Northern Temperate Zone, or in corresponding situations on mountains. The Mountain Maple covers about the 118 THE SUBDIVISIONS Ot' TH!E VEGETABLE klNGDOM. - Fig. 43S.— piscifioiw. ' Euonymus EuropcBua (Family Celastrinese), flowering branch, » tongitudinal Beotioli through a flower. > Qvassia amam- (Family Simaruhacedj), flowering branch.. *0yn8Bce^m' and 'floral receptacle. -^Qeh^ia (Family Ochnaceis), receptacle, gynasceum, and one stamen. ° Frnit of same. 1 Longitudinal section through the flower of the Tree of Heiiven {AUanUxa, Family Simarttbacese). (Partly afte^BaiUtn.) ■ ' .', .' ' ''■■':'.. ' '<% ANGIOSPEEM^, DICOTTLEDONES. 779 same ground as the Beech, and. in the Central Alps, even extends beyond the upper hmit oi the Beech. Fossil remains are found in the Mesozoic and Tertiary strata. Ihe number of known species now living is about 9000. Alliance LVII— Craterantha, Families: Legumim>s<^, Rosacec^, Saxifragacece, Escalloniacm, Cephalotacece, FrancoacecB, GrassulmecB, Hydrangeacm, Eibesaceoe, PhUadelphacm, Styracacece, Hama-melidacecB, RhamnacecB. Annual and perennial herbs, shrubs, and trees. Flowers abundant; actino- morphic and zygomorphic; hermaphrodite, pseudo-hermaphrodite, moncBcious, and dioecious. Floral - leaves in two 4-5-merous whorls, the lower whorl a calyx, the upper a corolla. Both whorls spring from the pitcher-shaped, bowl- shaped, or flat hypanthium, the petals always from the edge; the sepals, in part, also from the base of the hypan- thium. In the last case the tube of the calyx is adnate to the external surface of the hypanthium. The gynseceum is in the middle of the hypan- thium, and consists either of a single carpel with a uni- locular ovary (see figs. 438* 1' ^' *), or of several separate unilocu- lar carpels (see fig. 438* \ and p. 74, fig. 208 2), or of 2-many united carpels inclosing a multilocular oVary which may be adnate to the surrounding hypanthium at the base only, or from the base to the middle, or from the base to the top (see p. 74, fig. 208*'^'*). The ovules are situated on the ventral suture, and therefore in the inner angles of the loculi. The stamens spring from the edge of the hypanthium (see fig. 438*), and are in 1-2 whorls of 3-5 members each. The fruit is very various (pod, follicle, drupe, nut, berry, &c.), and the diversity in this rfespect afibrds the best means of distinguishing the numerous families belonging to this alliance. The hypanthium also varies considerably, and Pig. 438*.— Crateranthse. Longitudinal sectlona through the flowers of: i Cadia varia (Family Xeguminosse, division Csesalpinese). 2 Agrimonia Eupaiorium (Family Kosaceee, division Agrimoniaceee). ^ Chrysobalanua (Family Itosacese, division ChrysobalanacesQ). * Anthyllia Vutneraria (Family Legumiuosse. division Fapilionaceee). (After BaiUon.) 780 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. the families above enumerated may be placed in several groups according to the form of that structure. In the first group the hypanthium is short as compared with the floral-leaves, and dries up or detaches itself when the fruit is ripe (Papilionaceaj, Csesalpinese, Mimosese, Amygdalese, Rhamnacese); the pistil is simple. In the second group the hypanthium is flat, and bears the floral-leaves and stamens on its margin, and several separate unilocular carpels arranged in spirals in the middle of its surface; it does not fall off when the fruit ripens. In the third group the hypanthium undergoes further growth when the fruit ripens, and is converted into an envelope surrounding the nut-fruits, which have developed from the separate carpels in the interior of the pitcher-shaped cavity of the hypanthium (Agrimonese, RoseEe; see fig. 438 *^ and fig. 208 1' 2. p. 74). In the fourth group the multilocular pistil is adnate to the hypanthium which surrounds it entirely, and is converted into a fleshy pericarp (Pomacese; see fig. 208*'^'^> p. 74). In the fifth group only the lower half of the 2-carpellary gynseceum is adnate to the hyp- anthium, so that its upper half projects above the hypanthium, which is clothed by the calyx-tube (some of the Saxifragacese). In the sixth and last group the short hypanthium is only adnate at the base to the multicarpellary, actinomorphic gynseceum (Crassulacese, Styracacese, &c.). These groups are by no means sharply defined, and the links connecting them are again described as special families. It is also worthy of note that honey-secreting tissues in the flowers exhibit extreme variety of form and position ; sometimes they constitute a fleshy lining to the inner, i.e. the upper-surface of the hypanthium (several Dryadese), sometimes a swelling round the base of the ovary (several Saxifragacese), sometimes an annular ridge, or a group of separate wart-like glands, which are seated on the edge of the extremely short hypanthium, and are looked upon as metamorphosed stamens (Crassulacese). Amongst Csesalpinese, Pomacese, and Hamamelidacese are many species of arboreal growth, and amongst Mimosese, Amygdalese, Eosese, Spirseese, Rhamnacese, and' Hydrangeacese are large numbers of shrubs and under-shrubs. The majority of the herbaceous plants of this alliance occur in the families of Papilionacese, Dryadese, Agrimoneffi, and Saxifragacese. The Cassalpinese include several climbing lianes, the Papilionacese afford numbers of instances of switch-shrubs, and the Mimosese exhibit many shrubs with phyllodes. Amongst the Saxifragacese and Crassulacese many species with thick leaves (see vol. i. p. 327) occur. Cephalotus is insec- tivorous (see vol. i. p. 131). Compound pinnate or digitate foliage-leaves occur especially in Rubeas, Dryadese, Rosese, Papilionacese, Csesalpinese, and Mimosese (see vol. i. p. 533), whilst entire foliage-leaves are found particularly in Amyg- dalese, Styracacese, Crassulacese, Philadelphacese, and Rhamnacese. The flowers of PapiKonacese and Csesalpinese, and of some of the Saxifragacese and Chrysobalanese, are zygomorphic; those of the other families are actinomorphic. In some Mimosese, Crassulacese, and Styracacese the petals are connate at the base. Small, incon- spicuous, greenish petals are exhibited by some Agrimonese, Dryadese, Saxifragacese, Crassulacese, and by many Hamamelidacese and Rhamnacese; but most of the species of the alliance Crateranthro have brightly-coloured petals. Dusty pollen haa ANGIOSPERMiE, DICOTYLEDONES. 781 only been observed in a very few species (e.g. Poterium). In some Dryadace^ and Chrysobalanese the style springs in a curious manner from the base of the ovary (see fig. 438*8). The fruit is a pod (legume) in PapiMonaceee, Ceesalpinege, and Mimosese, and these three sub-families are hence often classed together by botanists under the name of Leguminosss. The fruit of Amygdalese, Chrysobalanese and RubesB is a 1 -stoned drupe, that of Rhamnacese a 3-stoned drupe. The Agrimoneae and Dryadeee are distinguished by small nut-like fruits, and the Spirseese, Saxi- fragacese, and CrassulacesB have follicles which dehisce at the upper part of the ventral suture. In most of the families above enumerated the seeds contain no endosperm; on the other hand, the thick cotyledons are crammed with reserve materials, and several of these seeds are used as important articles of human food (e.g. beans, peas, lentils, &c.). The Crateranthse are distributed in all quarters of the globe and in all latitudes. Csesalpinese and Chrysobalanesa belong chiefly to the tropics, whilst Dryadese and Saxifragacese live principally in the arctic regions and on high mountains. The Papilionacese are found most abundantly in the area of the Mediterranean flora and in the steppes in the south-west of Asia. More than 800 species of the genus Astragalus alone are known to exist in the last-mentioned districts. The Mimosese, especially the species of the genus Acacia, are represented in Africa and Australia by many characteristic forms. Rosese and Rubese, e.g. the genera Bosa and Rubus, occur in an astonishing variety of species in Central Europe, whilst the Spirseeae and Amygdalese are in like abundance in the west of Asia. Crassulacese are most abundant at the Cape and in Mexico, but they are also represented by a great number of species of the genus Sempervivum in the mountainous parts of Southern Europe. Rhodiola rosea, which belongs to this family, occurs in the arctic flora, and Sedum repens is found in the Alps at a height of 3000 metres above the sea. Of the Saxifragacese, Saxifraga oppositifolia reaches the furthest north, it having been met with at the northernmost spot hitherto visited in Franz Joseph's Land, at 81° N. Lat. In the Central Alps this Saxifrage is found at an elevation of 3160 metres. Fossil remains of Rosacese, Leguminosse, and Rhamnacese have been identi- fied in the deposits of the Tertiary Period. The number of extant species hitherto discovered amounts to about 10,000. Alliance LVIII. — Myrtales. Families: Myrtacece, Oranatacece, Onagraceoe. Annual and perennial herbs, shrubs, and trees, with entire foliage-leaves. Vena- tion consisting of a main axial strand, with sinuous lateral strands branching pinnately from it. Flowers hermaphrodite, actinomorphic or zygomorphic. The floral- leaves spring from an annular or tubular hypanthium, which is fused with the inferior ovary; they are differentiated into calyx and corolla, each of which consists of a 2-6-merous whorl. The stamens are in 1, 2, or more whorls of 2-6 members 782 THE SUBDIVISIONS OF THE VEGETABLE KINODOM. each, and spring from the fleshy annular or tubular hypanthium which rises up above the ovary. The fruit is baccate, drupaceous, or capsular. The Granataceaj and Onagracese contain watery juices, whilst the species, belong- ing to the other families are more or less rich in aromatic substances and etherial oils. Several are used as spices. Allspice is derived from Pimenta officinalis, and cloves are prepared from the flower-buds of Eugenia caryophyllata. The Circseese are small and delicate herbs, the Chamselaucese are dwarf shrubs; whilst the genus Eucalyptus of the family Myrtacese includes several species which are amongst the highest trees in the world (see vol. i. p. 723). The flower in Girccea exhibits two sepals, two petals, and two stamens in order (see fig. 261 ^ p. 236). In (Enothera and Epilobium the calyx and the corolla are composed of one 4-partite whorl each, and the andrcecium of 2 such whorls (see fig. 281, p. 282, and fig. 300, p. 354). In Eucalyptus. Myrtus, and many other genera the number of stamens amounts to over Fig. 439.— Myrtales. 1 Melaleuca. Longitudinal section tlirough the flower. 2 Plower-bud ol Eucalypbm globulus; the connate sepals are detached fromithe receptacle as a lid when the bud opens. * Fruit ol Enealyptua globulus. (Alter Baillon.) 100; in Melaleuca (see fig. 284*, p. 292) the stamens are coherent in bunches. Where the petals are suppressed, the filiform filaments are white or bright red and yellow. In the Fuchsias the sepals are coloured like petals; in some species of the genus Eucalyptus the sepals, which are joined together so as to form a lid, become detached from the floral receptacle before the flower opens. This remarkable phenomenon is shown in flg. 439^. The hypanthium which invests the ovary exhibits every possible transition in form, from a shallow saucer to a long tube (see fig. 266, p. 247). The ovary is in several families divided by septa, composed of tissue pertaining to the receptacle, which extend from the central column to the wall of the ovary. In Granatacese (Punica Granatum) the ovary is in addition divided by a plate of tissue into an upper and a lower story, and placentas are developed in the chambers of both stories. The fruits of several Lecythidacese, e.g. those of Lecythis, open with a lid; those of Eucalyptus are in the form of hard woody capsules, which dehisce at the top by means of valves, pores, or slits (see fig. 439'). The fruits of Gi/rccea are indehiscent; those of Epilobium, &c., are capsular, and open by valves. Many Myrtacese and the Fuchsias have baccate fruits. The fruit of the Pomegranate (Punica Qranatum) resembles an apple. ' The seeds of Bertholletia excelsa of the family Lecythidacese are known as Brazil nuts, and have a coat as hard as stone; ANGIOSPERM^, DieOTYLEDONES. 783 those of the Pomegranate have a fleshy coat, and those of EpiloUum are furnished with hairs. The Myrtales are distributed over all quarters of the globe. The ChamselaucesB and Leptospermacese are natives of Australia and the islands of the Pacific Ocean. Several species of this family help to form Australian bush, whilst others, especially species of the genus Eucalyptus, constitute entire forests (see Elate XVI.). The Lecythidacese grow chiefly in South America. The Myrtaceee are most abundant in America, and occur also in Asia and Africa. Europe only possesses one species, Myrtus communis, which belongs to the Mediterranean Flora. The Fuchsias are indigenous to Central and South America. The Epilobiums and Fig, 440, — ^MelastomacesB. Melastffma Malabathricum. (After BailloD.) Circsaas live_ principally in the North .Temperate Zone, and some species of the genus Epilohiu/m occur in the arctic regions and on mountain heights. Fossil remains of Myrtacese and Granatacese have been foimd in the strata of the Tertiary Period. The number of species ascertained to exist at the present day is about 2500. Alliance LIX. — Melastomales. Family: Melastomacece. Perennial herbs, shrubs, and trees, with opposite or whorled foliage-leaves. Leaves entire, with 3-11 curved ribs connected by transverse anastomoses (see fig. 440). Flowers hermaphrodite or pseudo-hermaphrodite; slightly zygomorphic. Floral-leaves difierentiated into calyx and corolla. The bowl-shaped or tubular 784 THE SUBDIVISIONS OP THE VEGETABLE KINGDOM. receptacle is covered externally by the tube of the 4-6-sepalous calyx, is surmounted by the segments of the calyx-limb, and bears upon it the 4-6 petals which alternate with those segments. The gynisceum is composed of 3-8 connate carpels. The 3-8-locular ovary is adnate to the hollowed receptacle at the base only, or from the base to the middle, or from the base to the top. An axis rises up in the middle of the ovary and bears the placentas, which project into the separate loculi. The androeeium is composed of 1-2 whorls of 4-5 stamens each. At the base of each anther is a spur-like appendage; dehiscence is apical by 1 or 2 small holes (see fig. 216 ", p. 91). The pollen is powdery. The fruit is a berry or a capsule which dehisces by valves. The seeds contain no endosperm. The Melastomales belong chiefly to tropical America. Fossil remains have not been identified with certainty. The number of species ascertained to exist at the present time is about 2000. Alliance LX. — Lythrales. Families: Lythracece, Gwpheacece, and LagerstrcemiacecB. Annual or perennial herbs, shrubs, and trees with opposite or whorled foliage- leaves. Laminae entire, with pinnate venation. Flowers hermaphrodite, actino- morphic or zygomorphic, with calyx and corolla. The cup-shaped or tubular receptacle is covered externally by the tube of the 3-16-sepalous calyx, the segments of which project beyond it and alternate with the 3-16 petals which are borne upon the receptacle. The gynseceum is composed of 2-6 connate carpels. The 2-6-locular ovary is free, and is situated at the bottom of the hollow receptacle. An axial column rises up in the middle of the ovary and bears the placentas, which project into the separate loculi. The androeeium is composed of 1-2 whorls of 3-16 stamens each. The anthers have no appendages, and dehisce by longitudinal slits. The pollen is adhesive. The fruit is a capsule coated by the cup-shaped receptacle. The seeds contain no endosperm. The Lythrales are distributed in all quarters of the globe. They exhibit greatest variety in tropical America. In the North Temperate Zone they are represented by the genera Lythrum, Peplis, and Bidiplis. No fossil remains are known. The number of identified species now living is about 400. Alliance LXI. — Hygrobiae. Families: HippuridaceoB, Callitrichacece, Myriophyllacece, Ounneracece, TrapacecB. Herbs and under-shrubs living in water or in wet places. Flowers hermaphro- dite, pseudo- hermaphrodite, monoecious, and dioecious; actinomorphic. Floral- leaves inconspicuous, in 1-2 whorls of 2-4 leaves each. Gynseceum ©f one carpel or 2-4 connate carpels. The under half or the whole of the 1-4-locular ovary is adnate to the sepals, which cohere so as to form a cup. Each loculus contains one ovule in its inner angle. The androeeium is composed of 1-8 stamens. The fruit is ANGIOSPEEM^, DICOTYLEDONES. 785 a schizocarp (Callitriche; see p. 427. figs. 822= and 322*) or a drupe covered with a thin coat of pulp; it becomes detached from the receptacle. In the Water Chestnut (Trapa natans; see vol. i. p. 607, fig. 144) the two whorls of two sepals each which are adnate to the ovary become a part of the fruit, and their apices project in the form of four stiff points. The Hygrobise are distributed in every quarter of the globe, but belong especially to the North Temperate Zone. The Gunneracese inhabit the Southern Hemisphere. Fossil remains of a plant resembling Myrio- phyllum have been found in strata of the Tertiary Period. The number of extant species known is about 100. Alliance LXII. — Passiflorales. Families: Passifloracece, Loasacece, Datiscacece, Samydacece, Turneracece, Papayacece. Annual or perennial herbs, shrubs, and trees, with palmately-lobed foliage-leaves. Venation palmate (radiating). Flowers hermaphrodite or pseudo-hermaphrodite, and dioecious; actinomorphie. The floral-leaves spring from a cup-shaped hypan- thium in one or two 4-5-merous whorls. The gynseceum is composed of 3 connate carpels. The unilocular ovary is free, and is raised upon a more or less elongated stalk from the bottom of the receptacle, or else it is sessile and adnate to the cup- shaped receptacle either half-way from the base or from base to top. The ovules are borne upon three placentas which project in the form of cushions from the internal wall of the ovary. The androecium is composed of 4-5 stamens which spring from the edge of the cup-shaped hypanthium. The fruit is a berry or a capsule opening by valves. The seeds contain a fleshy endosperm, in which is imbedded a straight embryo. The DatiscacesB have a sepaloid perianth. In the Loasacese and Passifloracese the floral-leaves are in two whorls, both of which are petaloid. In the Passifloracese a many-membered corona is inserted between the androecium and the petals. The Passiflorales belong chiefly to tropical America. Fossil remains have not been identified with certainty. The number of extant species known is about 700. Alliance LXIII. — Pepones. Families: CucurbitacecB and Begoniacece. Annual and perennial herbs and under-shrubs (suffrutices). Venation of the foliage-leaves radiating (palmate). Flowers solitary or in cymes; actinomorphie; pseudo-hermaphrodite, monoecious and dioecious. The uppermost part of the , receptacle, which is deeply hollowed, is developed as a hypanthium, and from it spring the floral-leaves in 1-2 whorls of 2-5 segments each. When two whorls are present they are either both petaloid in colour or the under whorl is a calyx and the upper a corolla. The petals are either free or partially coherent. The ovary is inferior. The ovules are borne on thick pads which are split in two longitudinally. Vol, II. 100 786 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. and project into the middle of the ovary. The androecium is composed of 5 or many stamens which spring from the hypanthium, and are joined at the base to the corolla. The fruit is baccate or capsular. The seeds contain no endosperm. The Cueurbitacese have symmetrical foliage-leaves — no stipules, but often tendrils (see vol. i. p. 696, fig. 165); the Begoniacese have oblique, unsymmetrical laminse, large lacerated stipules, and no tendrils (see vol. i. p. 420, fig. 110^). The whorls of fioral-leaves are 5-merous in Cueurbitacese ; in Begoniacese the floral- leaves of the female flowers are arranged in two whorls of 3-5 segments each, and those of the male flowers in two whorls of 2-5 segments each. Three winged ridges project from the inferior ovary in Begoniacese. The stalks of the ovules of Cueurbitacese fill the cavity of the ovary so completely that only small interstices are left between them. In many Cueurbitacese these stalks are converted when the seeds ripen into a succulent mass {e.g. in cucumbers, melons, and gourds). In Begoniacese also they project from the walls into the middle of the ovary, and the latter looks in consequence as if it were divided into loculi. The androecium ■exhibits great diversity. In some Cueurbitacese the five stamens are free ; in others they are partially united, and in a third group they are all completely fused together into a column. In Begoniacese, also, the stamens are connate and form a •column. In many cases the anthers are sinuous, and in the genus (Cyclanthera) there is a continuous anther all round the column. The Pepones are mainly tropical plants. The Begoniacese grow especially in the tropical forests of America, where they are not infrequently epiphytic. There is still some doubt as to the place of origin of Melons, Gourds, and Cucumbers. The alliance is represented by the Squirting Cucumber (Materium) in the South of Europe, and by Bryony {Bryonia) in Central and Northern Europe. No fossil remains have been dis- covered. The number of extant species hitherto identified is about 1100. Alliance LXIV. — Cactales. Families: Opuntiacece and Cactaceoe. Perennial plants, whose stems are much swollen or flattened. Flowers solitary; actinomorphic or zygomorphic; hermaphrodite. The ovary consists of a hollow, cup-shaped floral receptacle, from the inner wall of which spring filiform stalks bearing the ovules (see fig. 209 !■ ^, p. 77). The external surface of the receptacle is clothed by floral -leaves arranged in a spiral; the lower leaves are small and inconspicuous, the upper petaloid and brightly coloured (see figs. 441 1.2,3,4,6) inside the upper tubular prolongation of the receptacle are crowds of stamens arranged spirally. The pollen is adhesive. The fruit is unilocular and succulent (see fig. 441 ^). The seeds contain no endosperm. The genus PeresJcia alone exhibits thick, green foliage-leaves; in the other genera the leaves are replaced by small caducous scales or thorns, and assimilation is effected by means of the green cortical tissue of the swollen stems (see vol. i. pp. 327 and 440). The species of Rhipsalis and Phyllocactus, which are epiphytic ANGIOSPERM^, DICOTYLEDONES. 787 on the branches of old trees, have much -branched and segmented phylloclades, which often hang in curves; the segments of the stem in Opuntias are laterally compressed, and more or less discoid (see vol. i. Plate IV.). The cylindrical stems of the Queen of Night (Gereus nycticalus, vol. i. Plate VII., in foreground) are prismatic, and climb up rocks and the bark of trees by means of clinging roots. Other species of Gereus, such as Gereus giganteus, which grows to a height of 20 metres, possess erect columnar stems (Plate VII., left hand, middle distance). Another set of Cereus-species, including the many forms of Mamillaria, Melocadus, EcMnocactus, and Echinopsis, are spherical or truncate; they are covered either with papillae, each of which is crowned by a bunch of prickles (see fig. 441 ^), or Fig. 441.— CactacesB. 1 Flower. 2 Fruit of Cereus giganteus. s MamUlaria pectinata. 4 Cereus dasyacanthus. 5 ^ckinocactits hfyrizontalis. All the figures reduced. with tubercles, which coalesce into crests and ribs (see figs. 441 * and 441 *). The Cactuses are natives of the New World, and inhabit regions where a short, rainy season is followed by a prolonged period of drought. The largest number of species is found in Mexico. Some species also occur at high elevations in the Andes. No fossil remains are known. The number of species ascertained to exist at the present time is about 1300 Alliance LXV. — Ficoidales. Families: Portulacece, MoUuginacecB, Ficoidacece, Mesembryanthemacece. Annual and perennial herbs and under-shrubs, with entire fleshy foliage-leaves. Flowers solitary or in fascicles and glomerules; actinomorphic, hermaphrodite. The floral -leaves free or else connate at the base; in one, two, or several 2-5- merous whorls. Either all the floral-leaves, or only those of the lowest whorl, are sepaloid; in the latter case, the upper whorl or whorls are petaloid. The 788 THE SUBDIVISIONS OF THE VEGETABLE KINGDOM. gynEeceum is composed of 3, 5, 8 or more connate carpels. The ovary is adnata to the bowl-shaped receptacle at the base only or as regards the lower half of its surface, or from the base to the top; it is crowned by a radiating stigma of 3-10 rays, and is uni- or multi-locular. A column rises up in the middle of the ovary, and bears the ovules. In the case of multilocular ovaries, the ovules are borne on ridges and strands which project from this column into the loculi. The androecium is composed of one or several whorls of 3-10 stamens each. The anthers have no spurs, and they dehisce by longitudinal slits. The pollen is adhesive. The fruit is a capsule or an achene. The seeds contain a mealy albumen and a curved embryo. The Ficoidales chiefly inhabit dry localities. Only a few species (e.g. Montia fontana) live in water and on marshy soil. They are distributed all over the globe. Most of the Portulacese belong to South America and the Cape. The Mesembryan- themacese are developed in extraordinary variety in South Africa. There are over 300 species of the genus Mesembryanthemum alone at the Cape. No fossil remains are known. The number of extant species hitherto identified is about 500. Alliance LXVI.— Umbellales. Families: CornaceoB, AraliacecB, and UmbelUferce. Annual or perennial herbs, shrubs, and trees which flower profusely. Flowers in capitula, umbels, and cymes. Floral-leaves diiFerentiated into calyx and corolla. The calyx 4-5 sepalous, with its tube clothing the inferior ovary and the limb re- presented by 4-5 small teeth, which surround the top of the ovary. The corolla is likewise 4-5 petalous, the petals free and alternating with the sepals. The gynae- ceum is composed of a whorl of connate carpels; ovary inferior, 2-5 locular. Every loculus corresponds to a carpel, and contains a single ovule, which is suspended near the upper end of the loculus (see fig. 442 *). On the top of the ovary is a glandular disc, which secretes honey (see figs. 442 ^' *■ '). The androecium consists of a whorl of 4-5 stamens. The stamens are quite separate, and stand in a circle round the honey- secreting disc. The fruit in Umbelliferse is a schizocarp (see p. 427, fig. 322 ^•^■'^ and fig. 442^), in Cornacese and Araliaceee a berry or drupe. The seed contains an abundant endosperm, in which the embryo is imbedded. The Comaeese are for the most part woody plants, with entire, opposite foliage- leaves, possessing a venation of arched strands (see p. 231, fig. 260 and vol. i. p. 630). The Araliacese, of which the Ivy (Hedera Helix, see vol. i. p. 703, fig. 167) may be taken as a type, are woody plants with climbing roots, or shrubs and herbs with radiately-veined foliage, and the Umbelliferse, which are very rich in aromatic substances, oils, and resins, are for the most part herbs whose stems in many species reach a length of 3-4 metres, as, for instance, in Ferula communis and Euryangium Sumbul. The foliage-leaves of Umbelliferae are usually much divided (see fig. 442^), those of Eydrocotyle vulgaris, a plant which lives in swamps, are peltate (see ANGIOSPERM^, DICOTYLEDONES. 789 fig. 442^). The calyx, corolla, and androecium are 4-merous in Cornaceee, 5-merous in Umbelliferse and AraliacesB (see figs. 442 ^ and 442 «, and p. 289, fig. 283*). The Umbellales belong chiefly to the North Temperate Zone, but the Araliacese are Fig. 442.— Umbellales. * Heracleum Sphondylium (Family Umbelliferse), iiowering plant. 2 Single flower, s Fruit. * Longitudinal section through the iiower of Eryngium mariti'mwm (Family Umbelliferse). 5 Hydroeotyle vulgaris (Family Umbelliferse), entire plant, fi Cornus mas (Family Cornacese), inflorescence, f Longitudinal section through a flower. ^ Fruit. 1 reduced ; 2, s, 4^ s^ e^ 7, and « magnified. (After Baillon.) also represented by a number o£ species in the Tropics. Several of the Umbelliferse are natives of the arctic area of vegetation and of alpine regions. Gaya simplex occurs in the Central Alps as high as 2600 metres above the sea-level. Fossil remains, belonging chiefly to the families of Araliacese and Cornacese, have been 790 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. found in the deposits of the Mesozoic and Tertiary Periods. The number of extant species identified hitherto amounts to about 1800. 4.— THE DISTEIBUTION OF SPECIES. Distribution of Species by Offshoots.— Piatribution of Species by Fruits and Seeds.— Limits of Distribution. — Plant-Communities and Floras. THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. When the dreaded Dry-rot spreads unhindered over the surfaces of wooden beams, in a dark, damp cellar, its mycelium presents quite a strange appearance. Grouped in a circle around a dark centre of dead, disintegrating, and crumbling wood are a number of white spots, joined by indistinct lines to a centre. But this was not always the case. That which now forms the dead and crumbling centre was formerly the seat of the first development of the mycelium, then composed of a coherent network of mycelial threads and appearing to the naked eye as a single rounded white spot. The mycelial threads then crept out like rays all round the periphery, and as the white spot increased in diameter its centre became proportion- ately dark. The mycelial threads forsook their first settlement; they died oflp, and the wood they had destroyed then appeared merely as the dark centre of a white ring. In consequence of its continual widening the ring at length becomes segre- gated into stars, and is gradually transformed into a wreath of isolated mycelia, or, in other words, a group of separate but distinct mycelial spots arranged in a circle arises from a single mycelium in consequence of its radiating method of growth. The mycelium of Gasteromycetes, of many Fungi allied to Morels, and especially of many Agarics growing in the forest mould or in meadow humus, also exhibit under favourable conditions this ring and wreath formation. Although it is not possible to see the subterranean growth directly, its results are readily recognizable, since the receptacles rise above the ground from the separate portions of mycelium and indicate their distribution; these receptacles occur in regular circles, and when their colour contrasts with the surroundings they are especially conspicuous. Rings of this kind are shown in fig. 443 formed by the Ascomycete Spathularia flavida. The subterranean mycelium of this Fungus exercises no injurious influence on , plants in the immediate neighbourhood — at any rate, the mosses, grasses, and weeds which compose the carpet of the meadow round about show no sign of weak- ness, but are equally fresh and luxuriant within and without the rings. But it is not so in meadows where Agarics of the genus Marasmius and others have settled. The meadow-plants whose roots and root-stocks have been penetrated by their mycelia die off", and the places can be easily recognized by the withering and discoloration of their green aerial parts. On first looking at these spots one might easily suppose that the foundations of old circular walls were lying close under the turf which had RING-LIKE ARRANGEMENT. 791 792 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. in consequence dried up above the stones; that this is not so may be readily proved by digging, and this shows us at the same time that the humus and roots in these places are quite riddled and wrapped round by the mycelium of the Fungi named. The brown and grey ring- and crescent-shaped stripes show up most clearly on meadows because a parallel stripe of a specially vivid green is usually seen close to them. The reason of this is that, after two years, new plants develop in all those places which were formerly occupied by the mycelium, and have been abandoned by it in its centrifugal growth. Strangely enough, these plants are not the same species which were killed there in the previous year, but are herbs and grasses which find a suitable nourishing soil in the places which have lain waste for 1-2 years. The roots and root-stocks of the plants killed by the mycelium have mean- while decomposed, together with the remains of the mycelium and the receptacles of the Fungi. The soil is thus manured, and plants which usually settle on fallow ground grow there in abundance. They raise luxuriant stems and leaves, and thus is produced on the inner side of each bare stripe a parallel one of a bright green colour. This phenomenon has long been noticed by country people, particularly in regions where pastures are an important feature. It is associated with the influence of ghosts, witches, and elves; hence the name of fairy rings. In Upper Austria these bare dry spots are regarded as the rendezvous of the witches, and Walpurgis Night (1st May) is supposed to be the time when they are produced. In the Tyrol and other primitive countries the most varied superstitions are held to account for these curious stripes and patches. Fairy rings are sometimes formed by plants with underground rhizomes and runners, although not so frequently as by the subterranean mycelia of the Fungi named. Some Composites (for example Petasites niveus and P. officinalis, Arnica Ghamissonis, Achillea Millefolium), Labiates (Betonica grandiflo'ra, Mentha alpigena). Irises (Iris arenaria and I. Pallasii), Grasses, Sedges, and Rushes {Hierochloa borealis, Sesleria coerulea, Carex Schreberi, Juncus trifidus), under suitable conditions of soil form ring-shaped and garland-like colonies independently of Fungi. The mode of growth in these plants is like that of the Dry-rot. Young plants grow up with closely crowded shoots; these then spread out on all sides, and the connecting links die off simultaneously. In this way the original settlement is left a bare patch with dried remains surrounded by a circle of distinct and vigorous offshoots. Though shoots are very numerous they still stand close together even after they have severed connection, and if their annual growth is but slight it is some time before an actual ring is formed. It is in this case, however, the more striking, so that even a casual passer-by cannot fail to notice it. This happens principally in the above-named Grass-like plants, and among them especially in Sesleria carulea, which has attained a certain celebrity in Sweden as the ring- forming plant. It is there popularly termed elf dansar, and legend has it that the elves are especially fond of holding their nightly dances on places where rings of this Grass have been formed. RING-LIKE ARRANGEMENT. 793 Of course good rings are only produced by the plants named if the foremost buds produced by the subterranean internodes, i.e. those which form the terminations of the radiating stock or rhizome, undergo further development, while the intervening ones perish. This may not be the case under certain conditions, particularly if the growth of the terminal buds is retarded or stopped. For this reason fairy rings are formed much less frequently on stony, uneven ground than on flat homogeneous soils; and the best lands for this kind of fairy rings are pastures stretching over a moun- tain plateau, or the even floor of a valley. If specimens of the plants here described are planted on smooth ground, in good soil in a garden, in places where there is no obstacle to their spreading, they will form the rings and wreaths in question within a few years. But in spite of this, very few people are ever able to witness this interesting spectacle in gardens, because gardeners will not leave the rings alone, regarding the bare patch in the centre as unsightly and that the existence of a ring is a slur upon their craftsmanship. I remember noticing this many years ago in the Botanic Gardens at Innsbruck. The perennial plants were cultivated in certain beds close together, and to each species was allotted a limited amount of space. When the spring came round the gardener dug up the periphery of the circle, and planted it in the centre, to catch the escaping plants, as he put it. In the spots where Mentha alpigena had stood the previous year only a few withered stumps were to be seen, and not a single living shoot could be found. But shoots with their tops above the ground could be seen in a circle in the neighbouring beds, and also in the paths between the beds all round the space set apart for this species of Mint. These shoots were ruthlessly dug up and planted again in the forsaken spot. Every year or every second year this capturing of the fugitives was repeated, not only in the case of the Mint, but in many other instances, as, for example, Achillea asplenifolia and A. tomentosa, Betonica grandifiora, and Lysimachia thyrsiflora. Amongst aerial-sprouting plants which form rings and wreaths may be numbered the majority of Moulds, Lichens, and Mosses. The Mould, Penicillium glaucv/m, which settles on the fruit rind of oranges, apples, and pears, at first makes its appearance as a mere point, but later as a circular spot, and finally as a distinct ring surrounding a brown and rotten centre. The most striking of the ring-forming Lichens are those which stand out from their substratum on account of their colour. Most noticeable in this respect are the white Parmelia conspersa, which contrasts with the dark slate rock, and the saffron-yellow species Amphiloma callopisma and Oasparrinia elegans. The gelatinous Lichens, dark olive-green normally, but black when dried, especially Collema muUifidum and G. pulposum, often form such regular wreaths on a light background of lime- stone that they look as if they had been drawn with compasses, and the tiny yellowish-red Physcia cirrochroa has a particularly elegant appearance when it has radiated out from the hundreds of spots where it established itself on the flat surfaces of a steep calcareous rock. One might almost think that the small orange wreaths had been painted in with a brush. They also remind one of the fleecy 794 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. clouds in the evening sky, whose edges are reddened with the rays of the setting sun; and if I am not mistaken, this Lichen has obtained its name on account of this resemblance. The chief Liverworts and Mosses which form rings and wreaths when they grow on the flat surfaces of steep rock-faces and on the bark of old tree-trunks, are Frullania dilatata, Radula complanata, Amblystegimn serpens, Anomodon viti- culosus, and Hypnum Halleri. When they first settle they are scarcely noticeable on account of their minuteness, but they spread very rapidly, their firmly adherent stems forking and radiating out in all directions, the whole plant at a little distance now forming a greenish-yellow spot of circular outline. While growth proceeds in this way round the periphery of the Moss-plant, covering the rock or bark like a carpet by the multiplication of its outer forked branches, the older parts near the original place of settlement become dry, disintegrate, and are blown away like dust by the wind, the naked rock or bare bark thus again coming into view. In this way 5, 10, or 20 new Moss-plants are derived from the original one, and stand in a circle round the bare centre. This circle widens from year to year, until at last it is interrupted by gaps, and then 20 or more specimens of the Moss are seen adhering to the substratum arranged in a circle more than a span from the original settling place. In order that the ring or wreath arrangement of the offshoots above described should obtain, it is necessary that the original plant should dry up and decompose, and that the shoots which radiate from it should also die off behind in proportion as their growing points travel away from the centre of the settlement, and, finally, that no new ring-forming species should establish itself, or spread on the dead centre for a considerable time. These conditions are only comparatively rarely fulfilled, and this is the reason that ring and wreath formations are relatively so scarce. It happens much more frequently that the plant forming the starting-point of a colony, after it has sent out creeping threads of cells, runners, shoots, and the like in all directions, does not itself perish, but remains living and active in the centre of its separated shoots, even sending out new shoots year after year. In the same way the separated shoots repeat the parent-method of growth, i.e. they send out shoots in all directions like the mother-plant, though perhaps less regularly, and thus of necessity some of the young shoots come back to the bare centre and settle down where the mother-plant originally stood. The following phenomenon may also be observed: A plant gives off annually a pair of horizontal shoots on one side only, let us say on the south; their buds in the course of time become independent plants, and each again sends out a few horizontal shoots towards the south. In a few years' time these offshoots give rise to 20-30 plants, which are more or less distant from the starting-point, according to the length of the shoots. In all these cases the offshoots are not arranged in a ring or wreath round an empty centre, but in lines or clusters. Like the ring- and wreath-forming colonies, the offshoots, forming lines and clusters, may be underground or aerial. The receptacles of many Fungi emerge in a LINEAR AND CLUSTERED ARRANGEMENT. 795 clearly lineal arrangement from the mycelial threads running below the surface of the ground and in dead, rotten tree-trunks. Some Mosses form colonies in very regular lines from their rhizoids and horizontal underground protonemas. The most conspicuous, however, is the line formation produced by roots which run horizontally below the soil. The Aspen (Populus tremula), the Sea Buckthorn {Hippophae rhamnoides), Lycium barbarum, the Raspberry (Rubus Idceus), the Dwarf Elder (Sambucus Ebulus), Asclepias Gornuti, various species of Linaria and Ewphorbia, and numerous other plants (c/. p. 27) produce special horizontal underground roots, which give off buds towards the upper side. The shoots arising from these buds form separate independent plants after the root which formed them has died away. Obviously the plants follow the direction of the roots, and are arranged in rows. Even for years afterwards the line-like arrangement of the individuals in such colonies can still be recognized. When the bud-forming roots are of considerable length, the terminal offshoots are sometimes situated at some distance from the mother-plant. I saw single offshoots from the root of an Aspen push up through the ground 30 paces from the woody parent stem. Stems of Asclepias Gornuti spring up from the thick horizontal roots deep under the ground, at intervals of about 40 cm., and in them also can the linear arrangement be sometimes very clearly seen. When the individual offshoots in their turn give rise to horizontal roots, the line-formation is lost sight of more and more, and a scattered group spread over a wide area is the result. Sometimes the older portions of the colony die off completely, and as the individuals in one direction disappear, those in the other grow more luxuriantly. One might almost suppose the whole group to have taken a few steps forwards. This phenomenon can be seen particularly well in Raspberry bushes. On suitable soil a group of Raspberries will move about 2 paces every year, and therefore, after 10 years, they may have moved about 20 paces. If Raspberry bushes are planted near an inclosed piece of ground along a fence or hedge, it may happen that ten years later not a single one can be seen in the original place, while on the other side of the fence, in the neighbouring piece of ground, quite an assemblage of Raspberry plants has come into existence. The clustered or linear colonies which spring from underground tubers have the following very simple history. After a tuber has been fully formed on the under- ground shoot of a plant the slender bridge-like connections which have hitherto served for the conduction of food break down by the decay and decomposition of their tissues. The new tubers thus separated from the mother-plant send out stems from their buds, after the necessary period of rest these push up above the ground and also give rise to new subterranean shoots with tuberous swellings. These fresh tubers, after they have become disconnected, again form the starting-points for tuber-forming plants. This goes on until after a few years the soil all round the place where the first tuber had been is crowded with hundreds of separate tubers, and corresponding to these above the ground is a group of hundreds of separate leafy stalks. It depends of course on the number and length of the underground tuber-forming shoots whether the group is crowded or scattered. In the Artichoke 796 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. {Helianthus tuberosus) the tuber-forming shoots are short; the colony is therefore crowded, and only spreads slowly over a larger area. The Alpine Enchanter's Nightshade (Circoia alpina; see fig. 444^) forms elongated tubers at the end of shoots 6 cm. in length, about 5 of them round the mother-plant, whilst each of the new plants arising from these tubers repeats this formation in the same colony. Since 6 cm. is a considerable length compared with the size of the Enchanter's Nightshade, the group is scattered and in a few years extends over a considerable area. The tuber formation of Thladiantha dubia, a gourd-like plant gi'owing in Fig. 444.-Plant9 with tubers and bulbs whose mode of growth leads to the formation of colonies arranged in lines and clusters, i Musmri racemosum. a Thladiantha dubia. s Circcea alpina. Eastern Asia, is especially luxuriant; its shoot-formation is shown in fig. 444 \ In this plant a whole series of tubers which are chained together by thin threads 4-8 cm. long is formed on each shoot. Usually they form series of 6-10, and such a chain is about 50 cm. long. As a new plant grows from each tuber and again produces chains of underground tubers, the Thladiantha in a few years may occupy an area of 10 sq. metres, and will form a cluster which is both crowded and rapidly mcreasing in circumference. A further excellent example of the same thing is Glaux maritima which often spreads in the most diagrammatic manner in bare sandy places near the sea-shore. Many tuber-forming plants producing clustered colonies inhabit marshes, espe- cially such as are liable to great alteration in the level of the water and are exposed in years of drought to the danger of temporarily drying up. Many Pond- LINEAR AND OLUSTEEED ARRANGEMENT. 797 weeds (Potamogeton), e.g. Fotamogeton pectimatus, form a large number of small tubercles on their shoots which creep horizontally through the mud; Scirpus tuberosus, like the Alpine Enchanter's Nightshade, forms underground shoots 10-15 cm. in length, each terminating in a tuber as large as a chestnut, and since the plants to which these tubers give rise themselves repeat this formation of off- shoots, the diameter of the clustered colony increases about 20-30 cm. every year. The Arrow-head {Sagittaria sagittifolia) also develops peculiar tubers. In the autumn, offshoots whose scale-leaves terminate in a sharp point not unlike those of the Couch-grass spring from the knotty stems hidden in the mud. The leaf which envelops the swollen end of the offshoot has a stiff point and plays the part of an earth-borer or rather of a mud-borer, since it makes a path for the offshoot which may elongate as much as 25 cm. The swollen end of the offshoot, which is about the size of a hazel-nut, bears a small bud with greenish, closely-folded leaves, and this, together with its tuberous support, remains alive during the winter, while the plant to which the offshoots owed their origin perishes. In the following spring each of the small buds grows up into a new plant at the expense of the reserve- materials stored in the tuber, and now instead of the old dead plant we have a small group of young independent plants rising from the mud. The colonies of offshoots arranged in lines and clusters, which are developed from underground rhizomes and shoots, elongate horizontally, and form buds laterally and at their growing point, and in the same proportion as they fork and divide in front they die off behind, so that the individual sprouts become separated. To this category belong several species of Dentaria, Anemone, Couch-grass (Agropyrum), Mint (Mentha), Yarrow (Achillea), Willow-herb (Epilohium), Butterbur (Petasites), and the Woodruff (Asperula odorata). The length of the underground shoots which form the buds in these plants is very varied, as will be clearly seen from the table we insert here. Centimetres. Anemone rwnuncvZoides Monarda fistulosa . . Achillea Millefolivm . Equisetum arvense . Asperula taurma . Oxalis cornieulata. . Betonica gramdiflora . Tamaeel/mn Balsamita Aster salignus . . . Dentaria glamd/ulosa . Carex arenaria . . . Jimcus arotieus . . ^5-10 10-15 ^ 15-20 Centimetres. Epimedium alpimum . SHene alpestris . Mentha viridis . Asperula odorata Mentha piperita Huhia ti/nctorv/m, Senecio Fuohsii Mercurialis perennis . . Mentha crispa . . Agropyrum repens . . JEgopodiwm Podagraria Convolvulus arvensis . . Saponaria o^ci/nalis . PotentiUa hifwrca . . ■■K 20 h20-25 y 25-30 |30-35 35-45 Centimetres. Hieroehloa borealis . Urtica dioica . . Oarex pilosa Glaux maritima . Arnica Chamissonis Daphne Philippi . Senecio fluviatilis . . Tussilago Farfara Solidago canadensis . Petasites niveua . . Mentha alpigena Na/rdosmia fra^rans . Epilohium angustifolium Petasites officinalis 35-45 45-55 55-60 |-60-75 [75-85 ^85-100 100-150 These numbers do not represent the lengths of single internodes, but those oi the whole annual underground shoots which may consist of many internodes. For example, the year's underground shoot of the Umbellifer jEgopodium Poda- graria has 8 internodes, of which the proximal one is the longest and the distal the shortest. 798 THE DISTKIBUTION OF SPECIES BY OFFSHOOTS. The rapidity of the extension and the dimension of the colonies which are pro- duced from rhizomes and runners depends upon the length of the annual shoot, and upon whether or not the soil is favourable to the spreading of the offshoots, just as it does in the linear or clustered colonies arising from underground roots and tubers. In wood-clearings and on the banks of rivers many of these plants develop in a surprisingly short time, as, for example, Galamagrostis Epigeios, Epilobium angusti- folium, the North American Golden Rod and Rudbeckias (Solidago Canadensis and Eudbeckia laciniata), and these also have the property of suppressing and destroying all other vegetation in places where they have taken possession. This fact is turned to practical account by farmers who use certain Grasses which form linear and clustered offshoots to bind together loose soil, especially river sand. But there are also plants in this category which are veritable plagues to the farmer, the establishment and propagation of which he opposes by every means in his power. Examples of these dreaded plants, which, when they establish themselves in the fields and garden-beds, hinder the development of other plants, are furnished by the Gout- weed {^gopodium Podagraria), the Stinging Nettle (Urtica dioica), and the Fuller's Thistle {Girsium arvense). Wherever these have settled on cultivated ground and penetrated the soil with their offshoots there is nothing for it but to dig up the whole ground and to carefully remove all the shoots. Unfortunately even this laborious task is not always rewarded with the desired result, for in spite of the utmost care it may happen that small fragments remain, and these form the nucleus of a new colony of weeds. In a short time a new group appears above the soil which has been cleared with so much care, and a fresh digging and still more careful clearing of the ground is necessary. These clustered colonies have a characteristic appearance when foliage-leaves spring from their underground stem-structures, the large blades being borne on almost equally long erect stalks, as is the case, for example, in the Butterbur (Petasites officinalis) and numerous tropical Aroids. Wide tracts are then to be seen covered exclusively with their large luxuriant foliage- leaves, all other vegetation being suppressed. The formation of offshoots and the production of clustered colonies also occurs to a remarkable extent in the common Reed (Phragmites communis). Once settled on suitable soil it will cover the widest areas in uninterrupted and unhindered march, suppressing and destroying all other plants. On the lower Danube there are many lowlands so thickly set with Reeds that in several hours' journey only a few small inconspicuous plants will be seen beside the Reed haulms. This Reed is also interesting from the fact that its offshoots can arise just as well under water as under the ground, and it may serve, in some respects, as the type of a group of plants which, by reason of their amphibious nature, play an important part in the transformation of submerged into dry land. On the other hand, the variety of the protonemal threads, runners, shoots, and creeping stems which spread above-ground from the offshoots of these colonies is almost inexhaustible. And this is readily intelligible. The processes which are connected with their formation are much more varied in plant-members which grow CLUSTERED ARRANGEMENT. 799 in the light and in open air than in those which develop under the water or the soil, or, perhaps we should rather say, that above the ground the greater fluctuations in light, moisture, and temperature bring about corresponding modifications in the vital processes. Moreover, the substratum presents every imaginable gradation from shifting quicksand to heavy clay, from steep rock-faces in one place to the bark of old tree-trunks in another, all these having by no means the same effect on the formation of offshoots. One of the most noteworthy processes occurring above the ground leading to the formation of clustered offshoots is that exhibited by Moss-protonemas. By protonema is meant a web of threads which spreads some- times as a loose, open network, sometimes as a thick felt, over rock, clay, sand, earth, humus, bark, and decayed wood, the individual cells becoming the starting-points \-'-;<>X"'Ji rig. 446.— A section tliiaiigh soil permeated by the protonemal threads of the Moss Pottia intermedia. (Magnified.) l>t«^js of new Moss stems. This protonema may be compared to a web of Strawberry runners which has spread over the ground in a wood-clearing. Just as small plants spring up from the thread-like runners in this case, so Moss-plants are produced from the protonemal threads, and by the dying away of the latter become isolated. In many Mosses the end comes with the formation of this clustered arrangement, as, for example, in the tiny Mosses classed together as Pottiacese, of which one species, Pottia intermedia, is shown in fig. 445. This plant has the following remarkable history. During the period when most other plants are engaged in active nutrition and reproduction it remains with its rhizoids and part of the protonemal threads imbedded in the ground. Numerous scattered spores also remain resting in the ground until at length the time for aerial development arrives. Strangely enough, however, this is not until late in the autumn, when the leafy trees have discarded their foliage and autumnal mists drift through their bare branches. Then on the surface of the bare, cold, damp earth appear green threads which at first look like algal filaments, and on these small buds are formed (see fig. 445). In the course of a few weeks Moss-plants grow up from these buds 800 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. which become independent by the gradual withering and decomposition of the connecting pieces of thread. They form spherical spore-capsules, and with the scattering of the spores they wither and die. These plants are only a few milli- metres high, but they are clustered together in such thousands that they form a velvety carpet over the soil, their emerald-green colour being the more striking as the last remains of the neighbouring vegetation have assumed the dull hues of decay. The Luminous Moss (Schistostega osmundacea) growing in the holes and clefts of slate mountains (already described in vol. i.), the protonema of which is depicted in Plate I., also forms loose colonies of separate Moss-plants from the green threads which creep over the clayey soil in the hollows. These plants die off after they have ripened their fruits. Of course the development is in this case not so rapid and does not occur in the late autumn as in Pottia intermedia. A peculiar formation of offshoots may be noticed in epiphytes which climb over the bark of old trees and possess only short ribbon-like roots adhering to the damp bark, but none which grow down into the ground. Their stems and leaves invest the substratum like a carpet, as, for example, in several tropical Aroids of the genus Pothos, and in Marcgravia. The growing stem forks, and later on by the dying away of older portions behind the fork the two branches are separated and isolated. Each in its further growth may go a different way, one climbing up this and the other up that branch of the tree-trunk which serves as support; and, since this process is repeated, several independent plants of Marcgravia and Pothos may be found on the crown of the tree, all of which are to be regarded as natural offshoots. The same thing occurs in numerous Ferns, which grow on the bark of trees and in the humus-filled clefts of rocks, and in all those plants whose creeping aerial stems grow and branch at one end while they die off to a corresponding extent behind, as in many creeping species of clover, for example. As the annual increase in the stem of these plants is but small, the separated individuals move very slowly from one another, and several years elapse before the offshoots have formed a group which extends over an area of a square half -metre. The result is obtained comparatively much quicker when the offshoots are formed by runners and shoots. In one section of these plants, of which the Saxifraga flagella/ris (fig. 446), a plant widely spread through the Arctic region and in the high- mountain districts of the Himalayas, Altai, and Caucasus, may be taken as a type, only a single bud is developed at the end of a slender thread-like shoot. This takes root where it touches the ground, and grows up into a rosette. Not until the nourishment of the rosette by the rootlets which have been sent into the ground is assured does the long thread, terminated by the bud, die off, the connection with the mother-plant being thus severed, whilst the rosette now forms an independent plant. Since the shoots are usually numerous and radiate outwards the mother- plant in course of time becomes surrounded with an actual garland of rosette-shaped offshoots, and in a few years a fairly large area is covered with hundreds of larger and smaller rosettes, which, however, no longer show the circular arrangement, because the shoots of neighbouring rosettes often cross, and consequently the circles intersect. OFFSHOOTS ON EUNNERS. 801 Everyone knows the long runners of the Strawberry plant (Fragaria vesca). Here buds arise at the intermediate nodes as well as at the tip of the runner, and these develop into new plants after the thread-like connecting portions have perished. Suppose a Strawberry stock sends out three runners during the summer; each takes root at 5 nodes, and from each node a bud, i.e. an offshoot, develops, so that the following year the mother-stock is surrounded by fifteen daughter-plants. It should be noted that the length of the internodes in each runner is unequal. For example. Fig. 446. -Formation of a clustered colony by means of aerial runners in Saxifraga MieUaris. in one which had extended over the ground in the shade of the wood, the first mter- node was 37, the second 34, the third 31, the fourth 30, and the fifth and last 22 cm • thus the offshoots were the closer together the greater their distance from the mother-plant. Next summer fifteen new offshoots were again formed from each of the original fifteen, arranged in exactly the same way, and in the forest- glade where two years previously there had been only a single Strawberry plant occupying a space of 50 sq. cm., there would now be 200 plants distributed over a space of about 3600 sq. cm. The lesser Spearwort (Ranv^nculus reptans), the Ground Ivy {Glechonm hederacea), and the creeping Cinquefoil iPot.ntiUa reptans) display^quite as VOL. II. 802 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. considerable an increase and distribution as Strawberries. The accompanying table gives the length of runners and shoots of some well-known species in which the formation and rapid distribution of blFshoots is particularly noticeable on suitable substrata. Centimetres. Centimetres. Centimetres Saxifraga Aizoon . i Lycopodivm annotimum . 30-40 Vinca herbacea . . . . . 70 „ euneifolia . . 6 Saxifraga sarmentom . 40 Fragaria Indica . . . 85 „ Geum . 8 Ranunculus Flammmla . ) PotentiUa anserina 110 „ fiagellaris . ■ 10 Geum reptans .... h50 Glechoma hederacea . . 126 Sempervivurn stenopetalvm . 12 Glyceria fluitans . . J PotentiUa reptans . . . . 130 Viola odorata . . 13 Idthospermum pwrpureo- Rubus saxatilis . . . 140 Arabia proawrrens . 16 coervZeum . ... 56 Fragaria vesca . . . . 150 Androsaoe sarmentosa 18 Eanimcvlus reptans . . 60 Vinca major . . 200 Ajuga reptans .... 20 Tiarella cordifolia 65 Rubus Radula . . . . 300 Hieraeimn Jlagellare . . 30 Vimea Libanotica . . . . 66 „ bifrons . . . 650 In those cases in which plants change their position by the development of offshoots in any direction, whilst they die off in tbe opposite one, progress is always restricted. The offshoots penetrate only by slow degrees in the surrounding soil, and many years elapse before a space of 100 metres is traversed in this way. The change of position is much more rapid when the offshoots become detached from their place of origin and are carried to a new spot by special mechanisms of transit, by currents of water, the wind, or finally by the help of men or animals. In this way it may happen that single detached cells, cell-groups, buds, and shoots may be carried vastly further than 100 metres in a few minutes, through long valleys, over steep precipices, or even over high mountain ridges. This rapid distribution is not indeed so certain in its result as the slower mode of progression. It may easily happen that the wind or water current lands the detached offshoot on some spot where further development is impossible, where it must inevitably perish. Appa- rently, however, this disadvantage is compensated for by the immense quantity of such detached offshoots. Again, there are plants which form two kinds of offshoots, those which propagate slowly but surely, which are few in number, and others, developed in large numbers, which are distributed rapidly but less certainly. Only a very small proportion of plants develop offshoots which after they become detached reach a new locality spontaneously, by means of special organs of motility. This class of brood-body is always aquatic and of very small size, and its development can only be followed under the microscope. The best-known examples are Fungi, belonging to the Saprolegniacese and Chytridiacese, the dark green Vaucherias, and other species of Algae. The Saprolegnias are sapro- phytes growing in and on the bodies of animals which have died in the water — not only fish, crustaceans, and insect larvae, but also birds. They form delicate, thread-like, tubular hyphse, which ramify repeatedly, and part of which penetrate into the corpse like a root-plexus, while the rest rise up above the body in the form of white or grey felt, which floats in the water. Single tubular erect hyphse assume a knob or club-shaped form, and their protoplasm divides up into numerous portions. Ultimately the club-skaped tube opens at the apex, and the little proto- DISPERSAL BY WATER. 803 plasmic bodies (swarmspores) escape (c/. fig. 192, p. 17). What happens next differs according to the species. In the genus Saprolegnia the individual swarmspores have two cilia, by means of which they immediately swim away (see figs. 192 ^ and 192'); in Achlya, on the other hand, the swa-rmspores group themselves into a round ball in front of the opening of the tube as they escape (cf. fig. 192 1- 2. s. *), and at first possess no cilia. They surround themselves there with a delicate capsule, which apparently consists of cellulose, but they do not remain long in this condition. A few hours after, they leave the capsule and assume a bean-shaped form, being now provided with cilia which enable them to swim about in the water. They only swim about for a comparatively short time. When they have settled on some spot they lose their cilia, surround themselves with a cell-wall, and become the starting- point of a new plant; therefore they must certainly be regarded as ofishoots. The Chytridiacese have a similar offshoot formation. These too are devoid of chlorophyll, but they are true parasites, not saprophytes like the Saprolegniacese. They prefer green water-plants for their hosts, penetrating into their cells, killing and destroying the protoplasm, and then develop thick tubes which project beyond the host-plant, and in which the protoplasm becomes divided up into numerous spherical portions. The tubes open at their apex sometimes by the raising of an actual lid (see fig. 192^), sometimes by the dissolution of a limited portion of the cell-wall, so that a hole results from which the isolated protoplasts are expelled. On its escape each of these offshoots is spherical or egg-shaped in form, and possesses a single long cilium. This cilium serves as a swimming organ which in many species actually causes a hopping and springing movement. In order to avoid repetition, we may refer to the description of the swarmspore-formation given at vol. i. p. 29, in the case of the Vaucherias and Sphaerellas. On the whole, as we have already stated, the formation of offshoots which swim about independently in the water and seek out new spots suitable for settlement is restricted to a very small section of water-plants. Offshoots which, after their detachment from their place of origin, are carried passively by water currents with- out exercising any directive influence, and are stranded at some distant spot, are of much more frequent occurrence. Of these water-plants we might mention in the first place the filamentous green Algse which cover with slimy masses the surface of slowly moving water or stones at the bottom of rapid streams. In many of these plants several times during the year do the dividing membranes between the individual cells break down into mucilage so that the cells become free and are ■carried away by the flowing water. Each of these cells may again give rise to a new thread by repeated division. We cannot easily conceive a more simple method of propagation and distribution than this. The offshoot-formation in the Florideae is hardly less simple. Whether the whole plant is composed of rows or of open networks of cells, four protoplasmic balls, the so-called tetraspores, are formed in various situations on the plant; these are liberated into the surrounding water and carried away by the current. They adhere to some firm spot under water and there grow up into new plants. In most instances the protoplasm of the cell in 804 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. which the offshoot formation occurs is divided into four, more rarely two or eight clumps are formed by the splitting and dividing of the protoplasm, and most rarely of all is the whole undivided protoplasm of a cell transformed into a single offshoot. The small group of water-plants known as the Hydrodictyaceee display a peculiar offshoot-formation. In the elegant Water-net {Hydrodictyon utriculatum, cf. p. 640), whose cylindrical cells form a closed net with hexagonal meshes, the cells each originate new plants as Water-nets in miniature. The protoplasm in one of rig. 447.— Frogbit (Hydrocharis Morsut-n The winter buds in process of detachment from the ends of the submerged stolons. the ceUs which is preparing for offshoot-formation divides into many thousand parts which quiver in a remarkable way and pass between one another, and are said to undergo the so-called swarming motion. This lasts about half an hour, then the swarming portions, whose rod-like form can be recognized in spite of their minute- ness, come to rest, arrange themselves into nets with hexagonal meshes (see figs. 370 3.*. 5, p. 640), and now each cell contains a tiny Water-net. The outer layer of the cell-wall in which this grouping has taken place is partially dissolved. The little net, at first still inclosed in a pellicle of protoplasm, slips out and swims freely m the water as an offshoot. In 3-4 weeks it has attained the size of the Water-net, from one of whose cells it emerged, and in each of its own cylindrical cells the same process may be repeated. A similar process is observed in the small water-plant DISPERSAL BY WATER. 805 called Pediastrum, which is closely related to the Water-net, and of which one species is shown in figs. 370 ^■''•s. p. 640. The distribution of bud or sprout-like offshoots is seen especially in the Duck- weeds, Alismacese, Potamogetons, Utriculariaceffi, Droseraceee, and Primulacese. Most of the Duckweeds (e.g. Lemna polyrrhiza and L. arrhiza), which float during the summer on the surface of still water, towards the autumn form organs on their flattened stems which become detached from the summer plants, sink to the bottom Fig. 448. — Frogbit (Hydroeharis Morsus-ranoB). i Winter buds rising to the surface in spring. 2 Young floating plants which have developed from such buds, s Older floating plants. of the lake, and stay there during the winter. Each of these organs is pocket- shaped, and in the hollow the next year's shoot is already laid down — of course, as a minute structure whose semicircular free end scarcely projects above the closely- adjoining edges of the pocket. These detached winter buds sink because their cells, even those of the epidermis, develop large starch gi-ains which are crowded together, and literally fill up the lumens of the cells. There are no air-spaces like those which cause the summer-plants to float on the surface of the water; the stomata as yet are closed, and the whole body, hermetically sealed from the outer world, now has a specific gravity which causes it to sink down to the bottom of the water, where it is protected against the frost. There it remains in a resting condition during the winter. At the beginning of the warmer season the bud wakes from its winter 806 THE DISTEIBUTION OF SPECIES BY OFFSHOOTS. sleep, the starch-grains are used up in the building of the young stem, and the growing buds of these species of Lemna again rise to the surface of the water, because the cells which had served as storehouses for the starch become empty, and because air-spaces are formed in the new tissues. The same change of position during the year is also observed in the offshoots of the Frogbit (Hydroeharis Morsus-ranoe), which is common in still waters through the whole of Europe and a great part of Asia. Although this plant is abundantly provided with roots, it never fixes itself by them to the slimy bottom of the pool in which it lives. Throughout the summer it remains floating on the surface, spreading its foliage like the Water-lilies on the top, while its roots hang below in the upper layers of water. Its propagation in the summer is very rapid by the formation of offshoots. These arise in the axils of foliage-leaves from the very short, erect, floating stem, and are rather long, resembling thick threads, which keep close to the surface of the water, and grow in a horizontal direction. Each shoot terminates in a bud, and this quickly opens, sending up green foliage to float on the water, and a bunch of roots below. In a very short time the plant thus formed resembles the parent which gave it origin, and is itself able to develop new shoots. Thus it comes about that in a few weeks the surface of the water is covered with innumerable floating plants of Frogbit, every 10-20 being connected together by horizontal strands. The pretty flowers now rise above the surface from the stronger plants. The flowering is, however, of short duration, and is seldom successful, i.e. fruits with fertile seeds are rarely produced. As soon as the blossoming is over and autumn approaches, new shoots ending in buds appear. They are shorter than those of the spring, and they sink lower on account of the greater weight of the buds they carry. The buds, too, have a rather different form. They are flrm, and wrapped in closely-fitting scale-leaves, and they almost attain the size of a small date-seed. As soon as the bud is provided with the requisite amount of starch and other reserve food-substances, it becomes detached from its filamentous support, and sinks down (see fig. 447) till it rests on the mud at the bottom of the pond. The plants floating above, which gave rise to them, die off completely and decay. It is high time indeed to quit the field above, for the surface of the water is soon covered with a sheet of ice, which renders all vital activity impossible for months. When spring again arrives, and the ice vanishes from the pools and ponds, new life rises up from the mud below. The buds of the Frogbit which have passed the winter there become spongy, the cell-cavities fill with air, and the whole structure rises to the surface (see fig. 448 1). Arrived there the scale-leaves rapidly separate, green leaves expand their blades on the sunlit surface, roots hang down into the water, and before long, shoots are again developed as already described (fig. 448 2). Obviously deviations of position and sometimes considerable changes of place are brought about by the sinking and rising of the buds in the water. It is observed, too, that the Frogbit is very variable in regard to its position, and that sometimes a place whose surface was one year covered with numberless plants will in the year following present no trace of them, while new colonies will have developed at a distance. DISPERSAL BY WATER. 80T The Bladderworts (Utricularia), Aldrovandias (Aldrovandia vesiculosa, see vol. i. p. 151), and the Water Violet (Hottonia palustris), which desert the cold upper strata of water as winter sets in, and sink down to the relatively warm depths below, develop special wandering buds for this purpose; these are not enveloped in scale-leaves like those of the Frogbit; they are in reahty merely much abbreviated shoots whose leaves are so crowded and folded so closely together, that the whole shoot looks like a rounded green ball. These balls at first remain con- nected to the piece of the floating stem which gave them origin. This attachment is lost towards the end of autumn, and the little buds sink down to the bottom of the pond and necessarily get distributed in various directions. Next summer, when the balls leave their winter quarters and are again carried to the upper strata of water, they expand into foliage-bearing plants. It has been already stated (vol. i. pp. 76 and 668) that the Water Soldier (Stratiotes aloides), which is closely- related to the Frogbit, undergoes similar changes during the year, and we need here only draw attention to the fact that it sinks down to its winter quarters at the bottom of the pond as an open rosette, and not in the form of buds, and rises again the ensuing spring when the weather is more favourable. The Pondweeds Potamogeton crispus, obtusifoUus, pusillus, and trichoides behave differently from the marsh and water plants hitherto described. Here, as autumn approaches, buds are developed which become detached from the old decaying stems (fig. 136, vol i. p. 551), and sink down to the bottom of the pond; but in the following summer they remain sticking in the mud at the spot where they fell, and do not rise again to the surface. They send out roots and develop much-branched leafy stalks, and these rapidly grow up to the surface of the water. These Pondweeds, firmly rooted to the bottom of the pond, multiply not only by these free-swimming offshoots, but also by stolons which creep far and wide through the mud; but of course the plants are distributed to much greater distances by the sprouts or buds which are developed in the autumn on the upper intemodes, and which then become detached and float in the water, than would be possible by the creeping stolons alone. A very remarkable distribution of offshoots is to be observed in the marine Cymodocea Antarctica, which is very common on the coast of Australia, south of the Tropics. This plant has an erect stem, thickly covered with dull-green foliage- leaves, arranged in two rows. The lower leaves fall off prematurely, and the bare scarred stem then carries only a bunch of ribbon-shaped leaves at its summit. Towards the close of the winter the end of the stem above these leaves is seen to become peculiarly modified. Its internodes become much contracted, and at the lowest node is developed a scale-leaf with four lobes, which surrounds the leaves developed from the upper nodes, like a cup. Buds arise in the axils of one or two of these leaves, while the leaves themselves die and decay. The parenchyma of the four-lobed, cup-shaped scale-leaf also decays, and only its stiff" veins remain, so that instead of the cup, there are now only comb-like scales. After this alteration has taken place, the tissue of the stem below the pectinate scales breaks across. 808 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. and the whole shoot-apex, separated from the lower part of the stem, which has long been in a leafless condition, is carried away by the currents of the water. How far and for how long the shoot is carried about depends upon the local con- dition of the sea-shore. Sooner or later its career of wanderer is arrested by the trailing comb-like scales assuming the r61e of anchors. As soon as the anchor is fast, some 2-4 roots develop from the lower intemodes of the shoot; these pass between the teeth of the comb-like anchor, and grow down into the muddy sub- stratum, thus fixing the offshoot. All this happens at the end of winter. During the following summer, the shoot, which is about 8 cm. long, and is anchored and rooted in the mud, again grows up into a stem about a metre high, and next winter its top again falls off just in the same way. It has already been mentioned (p. 457) that this strange sea-plant very rarely blossoms or fructifies— a circumstance which confirms the supposition that the boundless colonies of it round the coasts of Tasmania owe their origin to the offshoots distributed by the sea currents. The distribution of offshofcs by sea-water is a much simpler affair in the Sea- wracks, Ulvas, and Florideae than in these other water-plants. When the sea is stirred to its depths by violent storms and the spring-tide is higher than usual the retreating waves leave any quantity of fragments of these plants behind them. These have been torn by the raging water from the firmly -fastened ribbons, nets, and threads below, and are then carried away by the billows. The water drives them into clefts of the rocky coast or imbeds them in the sand and mud of the shore, and, wonderful to relate, not a few of them flourish again, granted of course that they are not speedily removed by subsequent tides and that the circumstances are otherwise favourable. Much the same kind of thing happens on the banks of rivers and streams. A portion of the plant- fragments brought by floods and stranded on the mud of calm inlets undergoes decomposition. A larger portion remains fresh and living, sending out roots and vigorous shoots. In the bed of the Danube, in addition to the abundant creeping shoots of the Reeds {Phraginites) and various Sedges, Bul- rushes {Scirpus, Typha), broken twigs of Salix fragilis, bits of roots of the Sea Buckthorn (Hippophae rhatmioides), fragments of the rhizomes of (Enanthe Phel- landrium and Acorus Ccdamus, leafy twigs and stolons of various species of Pondweed, Water-Milfoil, and Water Eanunculus (Potamogeton, Myriophyllwm, RanuTiculus aquatilis) are all distributed in this way. Sometimes these growths settle in places where formerly no specimen of the kind had been seen for miles, and the fact may be easily confirmed that the distribution of their offshoots is actually brought about by flowing water in a very short time to great distances and in great abundance. The distribution of offshoots in little brooks which flow down between Reeds and Rushes with a moderate fall, and scarcely ever overflow their banks takes place more quietly. A rapid flow occurs only in the middle of the channel^: but near the bank, and especially in the small inlets, the water is almost as still and calm as in a closed-in lake. Here in these quiet spots are also to be found floating DISPERSAL BY WATER. 809 plants brought by birds; their roots are either not fixed to the ground but sway about in the water, or they may be altogether absent; examples are, Riccia fluttans and R natans, Lemna and Wolffia, and in tropical regions AzolUi and P^stia. All these multiply very rapidly. While they continually branch at one end, forming spreading lobes and sprouts, they die away on the other, the result being of course a separation into several pieces, i.e. into offshoots. These fragments spread themselves like a green mosaic over the surface of the water. As the off- shoots increase in numbers a certain number of them will extend beyond the calm mlet by the bank into the flowing water in mid-stream. Here they are hurried away by the current, and it often happens that they travel some distance before they are again stranded in some calm spot near the bank to form again the starting- point of a fresh aggregate of offshoots. Rain-water also plays an important part in the distribution of offshoots. Those of the widely spread Liverwort, Marchantia polymorpha, so frequently met with on damp earth, are especially noticeable in this respect. Their development is represented in fig. 196, p. 23. On the surface of the dark-green leaf-Kke thallus of this Liverwort cups arise, at the base of which papillae give origin to plate-like brood-bodies (gemmae, cf. figs. 196 ^ and 196 % Other papillae behave differently, and undergo only slight enlargement. The heads of these latter then swell up forming a gelatinous mass, and as this swells up it raises the green gemmse higher and higher out of the bottom of the cup (fig. 1962). At last they get close to the edge and are washed out of it by the rain. The offshoots of other Liverworts are also chiefly distributed by rain-water, as for instance the gemmae which arise in the crescent-shaped pockets of Lunularia, and in the flask-shaped cavities of Blasia pusilla. The pairs of cells which arise on the upper surface of Aneura multifida, the single cells which become detached from the edge of the fronds of so many Liverworts, the multicellular offshoots which are given off by Badula com- planata so common on the bark of trees, the round cell-plates growing on the edge of the leaf -like thallus of Metzgeria pubescens, and finally the ball- and disc-shaped groups of cells which develop on the surface of the leaves of numerous Mosses (e.g. on various species of the genera Leucobryum, Orim,'n%ia, Zygodon, Ortho- trichum, Barhula, Calymperes). In many of these cases the small offshoots are detached as well as distributed by the action of rain-water, but in others the loosen- ing occurs before the rain begins, and in Blasia and Aneura, as well as in Marchantia, the offshoots are first separated by mucilaginous membranes, and are thus raised up from their attachment. Not until afterwards are they washed out and distributed by the falling rain. These small offshoots can of course also be carried away from their place of origin by strong gusts of wind. Even breathing strongly on them is sufficient to detach the uppermost gemmaa of Marchantia, but in dry air and in dry soil they rapidly shrivel up and perish. The distribution by currents of air is therefore not attended by success, but the offshoots of the Liverworts and Mosses washed out by showers of rain immediately begin to grow, and quickly attain to further development. This mode of distribution plays an 810 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. important part in the covering of tree-trunks with Mosses and Liverworts. A small patch of Badula, Metzgeria or similar plant having once taken hold, when a downpour of rain beats upon the trunk quantities of tiny ball- and disc-like offshoots float away to be caught again by projecting irregularities of the surface; indeed the rapid covering of old trunks with green carpets and mantles of Liverworts and Mosses is for the most part effected by rain-water. It is comparatively seldom that bud- and sprout-shaped offshoots are distri- buted by rain. But there is one very interesting example of this, viz. the widely- spread Lesser Celandine (Banunculus Ficaria), a single plant of which is shown in fig. 343 ^ p. 460. In the axils of the foliage-leaves of this plant are developed off- shoots which have the form of small tubers, and which are not unlike the youngest stages of small potato-tubers (fig. 343^). When the leaves and stalk of the Lesser Celandine begin to turn yellow and wither in the early summer, the tubers break away from the stem and fall to the ground. There they usually escape observation, since they are hidden by the yellowing foliage; but should there come a heavy storm of rain the withered leaves are pressed down on to the soil by the force of the rain-drops, and the scattered tubers become visible. Sometimes the impact of the falling rain-drops hastens the detachment of the tubers from the mother-plant. When the rain is so heavy that the water flows away in the form of small rivulets, the loose tubers are washed off in abundance. A sudden downpour of rain in a region abundantly overgrown with Lesser Celandine is sufficient to float away numbers of the tubers, and heap them up on the borders of irrigation channels when the rain disperses. In such places the quantity of tubers which have floated together is often so large that one can hardly gather them in one's hands. In this way arose the idea that the tubers had fallen from heaven with the rain, and the myth of a rain of potatoes. The small tubers which arise in the axils of the leaves of Oagea bulbifera (of. fig. 343 \ p. 460), a plant growing on the steppes of Southern Russia, are distri- buted by rain-water just like those of the Lesser Celandine. This brings us to the question of the much-discussed manna-rain in steppes and deserts, which in reality is nothing but the distribution of the offshoots of a Lichen, viz. the Manna-lichen. This Lichen, which was termed Lichen esculentus by the older Botanists, but in recent times has been referred to the genera Urceohxria, Lecanora, Ghlorangium, and SpJmrotkallia, and which apparently consists of three species, viz. Lecanora esculenta, L. desertorum, and L. Jussufii, is spread over an enormous region in south-west Asia, and extends as far as the south-east of Europe and the north of Africa. This Lichen is met with in the neighbourhood of Constantinople, in the Crimea and Caucasus, in Persia (whence the illustration at page 695), also in Kurdistan, Arabia, and the Anatolian high land from Bulgar Dagh in the Taurus (where it is very often met with at a height of 2700 metres above the sea), and finally in the Sahara and the deserts of Algeria. It first forms thick wrinkled and warted crusts on the stones, preferably on small fragments of limestone lying about; the outer colour of the crust is a grayish yellow, while on breaking it appears as DISPERSAL BY WATER. 811 white as a crushed grain of corn. As they get older the crusts become rent, and separate either partially or wholly from their substratum. When they first become loosened the edges of the detached portion become somewhat rolled back. The roll- ing then continues, and ultimately the loosened piece forms an elliptical or spherical warted body with a very much contracted central cavity. Small stones are some- times imprisoned in this way within the cavity of the sphere, in which case the weight of the loose Lichen is correspondingly increased, but as a rule the hole is filled with air, and when dried the pieces weigh very little. Ten loose pieces of Manna-lichen, each as large as a hazel-nut, weighed 3-36 grams, and the weight of a single piece therefore was on an average only -34 grams. It is easy to see that the loose portions will be rolled about by the wind, and that a storm will sometimes sweep them up from the ground and carry them hither and thither through the air. This method of distribution appears to be the prevailing one in regions where the supply of water is not abundant in the rainy season, and where violent storms rage from time to time. That this is so is confirmed by the circumstance that the Manna- lichen after the storms lies chiefly piled up behind the low bushes and undergrowth,. i.e. just where the force of the storm has been to some extent broken, and where the shifting sand has been heaped up into little hillocks. Where a period of heavy rains succeeds the long dry summer, however, and where such a quantity of water falls on the parched land that it cannot all be absorbed, some of the rain collects into small rivulets. These carry away with them everything which is movable and capable of floating. The turbid rivulets flow down over the inclined soil to the lowest parts of the country and there unite into larger streams, or if it can find no outlet the water remains for some time in the hollows as small pools and puddles, and deposits there the mud and vegetable debris it has carried with it. The latter is more especially the case on the steppe soil overstrewn with small stones where between the slight elevations there is a labyrinth of shallow channels and winding depressions resembling ploughed land. In such regions the Manna-lichen is chiefly washed into the depressions by the rain-water, and in some years in such quantity that they form heaps a span high, and a single man can in a day collect 4-6 kilograms (about 12,000-20,000 pieces, varying in size from a pea to a hazel-nut). This is especially the case in the steppe region and in the high lands of South-west Asia, where the Manna-lichen is used as a substitute for corn in years of famine — being ground in the same way and baked into a species of bread. That the rain- water is the agent which transports the Lichen in these regions is beyond doubt,, because the pieces heaped up in the hollows are not in the least rubbed on their outer surfaces as would certainly be the case if they had been rolled and dragged even for only a short distance over stony ground. It is also remarkable that all the great so-called rains of manna, of which news has come from the East to Europe, especially those of the years 1824, 1828, 1841, 1846, 1863, and 1864, occurred at the beginning of the year between January and March, i.e. at the time of the heaviest rains. When we remember that the inhabitants of the district actually thought that the manna had fallen from heaven, and quite overlooked the 812 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. fact that this vegetable structure grew and developed (although only in isolated patches and principally as crusts on stones) in the immediate neighbourhood of the spots where they collected it, we need not be surprised at the conclusion of our own peasants who thought the tubers of the Lesser Celandine had fallen with the rain from heaven. It should be mentioned that the manna sent to the Israelites on their journey out of Egypt to the Holy Land is identical with the Lichen described here and figured on page 695, and the older view that the manna of the desert was the sap of a Tamarisk {Tamarix gallica inannifera) exuded under the influence of a parasite is without any foundation. Spores take the first place among the reproductive bodies which are distributed by wind. Many Ascomycetes develop some of their spores by abstriction from the free ends of special hyphss. These rise up into the air from the substratum, which is permeated or covered by the mycelium. In this way the separated but loosely- adhering spores can be carried away by the slightest atmospheric movement. In the Moulds known as Aspergillus and Penicillium, whole series of spores are cut ofi" from the end of each hypha (see figs. 193 *> ^' ^- ^, p. 18), and as they are crowded closely together a single breeze carries oiF innumerable quantities of spores. By breathing only lightly on the small forest-like colonies of supports the spores are whirled as dust into the air, and as they are extremely light they not only remain a long time suspended in it, but even in perfectly still air are carried sometimes up, sometimes down, by the currents due to slight diflferences of temperature, again being carried horizontally or whirled along until at last they settle, and become the starting-point of a new Mould formation. The spores abstricted from the ends of the so-called sterigmata in the Hymenomycetes (see figs. 389 and 390 ') may also be detached and carried away by wind, but apparently most of the spores in these Fungi separate spontaneously in calm air and fall to the ground, covering it with a delicate layer of dust, to be afterwards carried away from this resting-place by breezes. The spores of Ustilaginese and those in the Eecidia of Uredinese (see p. 686) are at first covered with delicate membranes and sometimes inclosed in special receptacles. As soon as they are mature they form a powdery mass, which bursts through the covering membrane, and the now exposed spores are blown away as dust by the wind. If they have developed in deep receptacles shaking is necessary before they can be blown away. The spores then fall from the mouth of the receptacle into the currents of air. In many Myxomycetes and Gasteromycetes (see fig. 367 \ p. 618, and fig. 391 ^ p. 690) delicate twisted threads called the capillitium are developed simultaneously with the spores. The web of threads with the spores between them is inclosed in a membrane (see fig. 449 1), When this membrane bursts at maturity and the receptacle is thrown open only the spores in the immediate neighbourhood of the opening can be blown away by the wind, the deeper ones being held back by the capillitium. The lower layers of the capillitium are then raised by tha action of dry winds, and thus quantities of new spores are continually carried from below up to the opening. In this way it happens that the spores of these plants are only DISPERSAL BY WIND. 813 distributed in small detachments, and only at a suitable time, i.e. when a dry wind IS blowing. A similar contrivance is exhibited by the Muscineffi in the Mar- chantiaceae, Anthocerotace^, and JungermanniacejB. Peculiar filamentous, very hygroscopic cells with spiral bands of thickening on the cell-wall, are found with the spores in the receptacles of these plants (see p. 696). These have been called elaters, because it was thought that their movements caused the ejection of the spores. Their significance, however, rather lies in the fact that they serve to hold the spores together after the opening of the receptacle, and only expose them by degrees to the wind. They also help to burst open the receptacles, but that hardly concerns us just now. Only three of the most striking of the varied contrivances for spore distribution by wind in Mosses (which are destitute of elaters) will be here described. First, those which are observed in the Andreseacese (see figs. 450 ^ and 450 2). Here the capsule opens by four longi- tudinal clefts which, however, do not extend quite to the free end, and the four pieces into which the wall is thus divided may be com- pared to the staves of a barrel joined together at the top. In damp weather they become approximated, so that the clefts are closed (fig. 450^). In dry weather the valves become arched, the clefts widen, and the spores may be blown out from the interior of the capsule by the dry wind (fig. 450 ^). The distribution of the spores is effected quite difierently in the Polytrichums, one species of which is illustrated in figs. 450 ^'*'^'^'^'^. After the roof (operculum) which formerly surmounted the cap- sule has fallen ofi" a delicate whitish membrane comes into view, which is held fast by the points of numerous sharp teeth, and is stretched like the skin of a drum over the mouth of the capsule with its annulus. If rain and dew moisten the Moss the teeth are seen to be much bent inwards, the membrane lying upon the annulus, and completely closing the receptacle (fig. 450 ^ and 450 ''). But in dry air, especially when a dry wind is blowing, the teeth turn rather outwards, raising the membrane above the annulus, and thus small holes are left between the teeth through which the spores can escape (figs. 450 ® and 450 ^). The same dry wind which causes the alteration in the position of the teeth now shakes the spores out of the capsule, which is borne on an elastic seta. Griimmia ovata, one of the Bryacese (see figs. 450^ and 450^"), may be taken as the type of a third contrivance for ex- posing the matured spores to the wind in dry weather, retaining them in the receptacle when it is damp and protecting them there from the injurious efieets of moisture. The circular mouth of the pipe-bowl-shaped receptacle is furnished with Fig. iid.—Trichm davata. 1 The membrane of the sporangium has burst, and the capillitium has bulged out raising up the spores embedded between its threads and expos- ing them to the wind ; x 20. a Threads of the capillitium with the spores lying between them; x250. 814 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. teeth, each of which terminates in a free point. The tissue of these teeth is hygro- scopic, and their direction and position alter to a surprising extent according to the ■degree of humidity of the air. In damp weather the teeth are so close together that they completely shut the capsule (fig. 450 »), but in diy weather they bend outwards (fig. 450 "), and the spores are shaken out of the capsule and scattered by the wind. We shall have to describe presently how the sporangia of most Ferns dehisce Fig. 450. — Dispersal of spores by wind. 1 Spore-capsule of an Andrecea in damp weather. 2 The same in dry weatlier. s Spore-capsule of a Polytrichum in damp weather. * In dry weather, s The spore-capsule of a Polytrichum, the edge of the mouth beset with teeth and covered with a membrane, in damp weather, e In dry weather. ? A part of the peristome more highly magnified, in damp weather. 8 In dry weather, a Spore-capsule of a Grimmia in damp weather. 10 In dry weather. 11 Racemose sporangia of a Botryehium in damp weather. 12 in dry weather, is a. single sporangium of this Botrychium enlai'ged, side and front view, in damp weather. 1* In dry weather. ^, * " and 12 nat. size ; the others enlarged. suddenly so as to scatter the spores. In such Ferns the sporangia are developed on the under side of the frond, and this position protects them excellently against any injury which might befall them from rain or dew. But there are some Ferns whose sporangia are exposed to both rain and dew, and whose spores are not suddenly scattered by the bursting of the sporangia. Among others, the Moonwort {Botryehium) may be mentioned. Its branched spike of sporangia is represented in figs 450" and 450 ^^ The elliptical sporangia of the Moonwort open by a trans- verse slit, but the two valves thus formed only separate in dry weather (figs. 450 ^^ DISPERSAL BY WIND. 815 and 450") when the spores may be shaken out and blown away. As soon as the sporangia are moistened the two valves immediately shut together (figs. 450" and 45015) and obviously the spores can no longer be shaken out. A similar opening and shutting of sporangia according to dryness or moisture may also be observed in the Lycopodiaceae (see fig. 405 *, p. 716). A similar phenomenon may also be observed in the sporangia of the Horse-tails (see fig. 403 *, p. 712). Here not only the sporangia but the spores themselves present a very different appearance according as they are dry or damp. The wall of these spores consists of several layers, of which the outermost sphts up spirally to form two arms which remain joined to the spore at one spot. In dry weather the two spiral bands, which are arranged in the form of a cross, unroll (see fig. 451 ^) and constitute four appendages which afford enough purchase to the wind to enable the comparatively large and heavy spores to be carried away. If the spores fall on to Fig. 451.— Spores of the Horse-tail i: ^- «), which grows commonly on rotten tree-trunks in Pine forests in mountainous regions, might also be mentioned. It develops multicellular disc-shaped gemmae at the top of certain erect shoots. The discs are supported on delicate filamentous stalks and are embedded in a cup of closely crowded leaflets (figs. 196 ^' ^' '- ^). After the supporting threads have withered and the small multicellular discs have become detached, a slight shaking by the wind is sufiicient to make the gemmae fall out and to scatter them. The .same breeze which shook the stem now whirls the tiny green discs far over the forest ground and transports them to other places of attachment where they continue their development. In some Mosses whose little leaves are very brittle when dry, for example, in Gampylopus (see fig. 196^^), the leaves themselves serve as offshoots. How the detachment of these leaflets is brought about is to some extent an enigma; appar- ently they separate and are thrown off spontaneously, not unlike the foliage-leaves which fall from the branches of trees in autumn. This is immaterial to the question under consideration here, however. This much is certain, that in the remotest mountain ravines, and on inaccessible ledges in precipitous places where the dis- turbance of passing animals is quite impossible, the turf -forming crowded stems of Gampylopus always carry detached and partially split leaflets which adhere loosely to the support. When after a few dry days a storm rages through the ravines, these loose leaflets are torn away, and do not again come to rest until they are far distant from the spot from which they were taken. The offshoots of Mosses composed of groups of cells, and the last-mentioned detached leaves which function as offshoots, do not grow up immediately into new Moss-plants, but first of all develop protonema- like cell-filaments, and it is from these that the young Moss-plants originate. It also happens that whole Moss-plants with elongated axis, numerous leaves, DISPERSAL BY WIND. 817 and abundant rhizoids are distributed by wind. This is observed in Mosses of very different genera (e.g. Leucodon sciuroides, Thuidium abietinum, Hypnum rugosum, Myurella julacea, Conomitrium Julianum, Anoectangium 8endtnerianum). The development of this form of offshoot is shown in figs. 156^ and 156", p. 23, in Leucodon sciuroides, which is common on the bark of old trees. In the angles made by the leaves with the axis of old shoots, buds first arise which grow into miniature Moss-shoots. These tiny shoots then become loosened at their base, and push up towards the top of the leaves supporting them. This happens more especially in rainy weather. When it is dry their leaflets lie close to the axis, but when saturated with moisture they stand out and bend backwards, and thus raise themselves out of the deep niche in which they have hitherto been concealed. Many of these loosened shoots are without doubt carried away by rain-water, and so transported some little distance, but most of them are whirled off by the wind, and carried far away over mountain and valley. Bud-shaped offshoots, which become detached from the aerial portions of plants, and whose distribution is effected by wind, are comparatively rare. A remarkable instance is furnished by the Club-moss Lycopodium Selago (see fig. 343 ^ p. 460). This plant, which is found in mountainous districts in the Northern Hemisphere of the Old and New Worlds, forms buds in the axils of its stiff, dark-green leaves, especially near the top of the shoot, which might, at first sight, be mistaken for small winged fruits. These buds are so provided with little leaves as to offer a good purchase to the wind, and by this means they are transported (cf. fig. 343^). The North American Lycopodiwm, lucidulwm, L. refiexum, L. Haleakala, L. ser- ratum, L. erubescens, behave in just the same way as Lycopodiwm Selago, and it is not improbable that many other allied species exhibit this kind of offshoot. Most detached bud-like offshoots, which develop in the axils of foliage-leaves and bracts on larger plants, e.g. on the bulbiferous Coral-wort (Dentaria bulbifera; see p. 461), can hardly be said to be distributed by wind. They are spherical or ovate, and not flattened like those of the Club-moss, and they are too cumbrous for transport on the wings of the wind. And yet the wind plays an important part in the distribution in such cases. The bulbils are borne on fairly stiff shoots, and the nature of their attachment is very fragile. Thus, as the shoot rebounds after the blast, many of the bulbils become detached, and are jerked away to a considerable distance. There are three types of offshoots which are jerked from the plant in the above- mentioned manner. First, those which have the form of closed buds or small bulbs, and which consist of a very much abbreviated stem or bulb-axis, and a few much- thickened scale-leaves filled with reserve materials. These are found in the bulbiferous Coral-wort, which grows commonly in Central European Beech forests, and has been selected as typical; on the bulbiferous Saxifrage (Saxifraga bulbifera), widely distributed in meadows in Eastern Europe; on several Lilies (e.g. Lilium bulbiferum, tigrinum, and lancifolium); and on the Persian Gagea (Gagea Persica), in the axils of the upper foliage-leaves; on Foucroya gigantea, growing on the VOL. II. "== 818 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. Antilles, and above the disc-shaped bracts on the top of the stalk on a considerable number of species of Allium (e.g. Allium Moly, vineale, oleraceum, carinatum, arenarium, Scorodoprasum, sativum). A second type, growing in the axils of bracts on the upper part of the stem, is shown by Polygonum bulbiferum and viviparum, natives of the far north, and of the Alpine regions of Europe and the Himalayas (see figs. 452 1' ^' ^- *' ^' *''''■ ^). These are not bulb-like structures, but small tubers or corms with a minute terminal bud projecting like a little horn, and the tissue of the tuber is abundantly filled with starch and other reserve materials (see figs. 452^ and 4521"). The third type is observed in species of the genus Globba, belonging to the Scitamineee, more especially in the East Indian Qlobba bulbifera and in Globba coccinea, which grows in Borneo. These rare plants develop offshoots in the axils of bracts on the uppermost part of the rigid stem. They consist of a small bud, from whose minute axis a thick, fleshy root filled with reserve materials grows down, so that in reahty the chief part of the offshoot consists of a root-structure. When the closed bulb-like offshoots, tubers, or buds with thickened roots have been thrown from the wind-swayed stem they remain unaltered in the spot where they have found a resting-place through the whole winter, or the whole dry period of summer. At length, when the most suitable time of year arrives, little absorbent roots make their appearance (see fig. 452 °) at the expense of the stored- up reserve materials, and these fix the offshoots in the soil and convey fluid nourishment to them. The axis of the offshoot elongates and grows into a stem, putting out leaves and forming a new independent plant. The entire sprouts, which are detached from aerial stems and become offshoots, can obviously not be transported very far by wind. They are much too heavy, and offer no suitable hold to the wind, which can only influence them by shaking the stem on which they are supported, or by rolling them along after they have fallen to the ground. In the former case the sprout-like offshoots are jerked off, and the action of the wind is therefore only indirect. Some plants bear side by side on the same stem tubers with undeveloped buds, and also some whose buds have begun to grow into sprouts, and have developed green foliage-leaves. These- form a connect- ing-link between the groups just described and those we are now about to consider. One of them is the already mentioned viviparous Polygonum, (Polygonum viviparum, fig. 452), in which it often happens that all possible stages of development occur close together on a single spike. In Grasses especially it is often the case that the offshoots when ready to be detached have the form of developed, leafy sprouts. In the Grasses of the Arctic flora belonging to the genera Poa, Festuca, and Aira, the formation of leafy sprouts which become offshoots is so usual that in places the plants bearing offshoots are more common than those bearing flowers in their panicles. On our high mountains also there grows a grass {Poa alpina, cf. fig. 342 ^ p. 455), in which the panicles as often bear offshoots as flowers. On the plains of Hungary flourishes a species of Meadow-grass {Poa bulbosa), in which the same thing happens to such an extent that in the many thousand plants which cover the ground all the panicles d^yelop DISPERSAL BY WIND. 819 offshoots exclusively. The detachment takes place in different ways in these "Viviparous" Grasses. Usually the sprouts loosen from the erect panicles of the Fig. 452. — Polygonum vivipanmi. 1 Entire plant ; one spike bears flowers only, the other carries tubers on the lower half and flowers above. 2 a whole plant whose spike bears tubers only. On some of the tubers small foliage-leaves have already developed, s-b Fallen tubers in successive stages of development ; nat. size. * A fallen tuber magnified, lo The same in longitudinal section. swaying haulm, and are scattered by the wind; but sometimes the separation does not occur until the stem is bent down to the ground with the weight of the crowded offshoots in the panicle. In this case the offshoots strike root where the panicle 820 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. touches the soil, and the result is that closely-crowded groups of new plants grow up round it. The same thing may be observed in CMorophytum comosivm, a native of the Cape often cultivated as a basket plant by gardeners under the name of Cordyline vivipara. In this plant leafy shoots are very regularly developed in the floral region instead of flowers, and as these increase in size and become heavier, the long, comparatively slender and very supple stem which bears them sinks down so that the sprouts are suspended on a green thread. If the ground below is suitable the pendent shoots which have meanwhile developed roots may settle there. If they do not come into contact with any suitable soil they remain a long time swaying in the air, growing and themselves forming long, thin stalks in their turn in whose floral region fresh, leafy sprouts with roots arise, and years after three or four generations of shoots connected together by a slender green stalk may be seen hanging down for the length of a metre. At length one or other of the swaying and wind-tossed sprouts strikes firm ground and takes root, separating itself from the old plant, or it falls like the fruit from a tree and rolls down below until it finds a place of settlement possibly at a considerable distance from the old plant. Among the Rushes also there are many species which develop pendent sprouts. In one species which is very widely spread over Northern Europe, viz. Juncus swpinus, it is much more usual to find sprout-like oflshoots in the floral region than flowers. In many of the Saxifrages of the far North, viz. in Saxifraga stellaris, S. nivalis, and S. cernua, very reduced shoots with small rosettes of foliage-leaves are formed on the terminal branches of the floral axis, or bulb-like buds arise in the axils of the bracts on the upper part of the stalk which, like those of the viviparous Polygonum, send out green foliage-leaves before they fall or become loosened (see figs. 342 1-2. 3, 4, 6, 6, 7, p_ 455-) Sedum villosum, which grows on moors, develops short, leafy sprouts with thread-like axis in the axils of the stem-leaves. As soon as the stem begins to wither these sprouts loosen and are carried to a short distance by gusts of wind. They send out delicate roots as soon as they find a resting-place and new plants are established. A very peculiar mode of detachment and distribution of sprout-like oflshoots is found in many species of House -leek (Sempervivum). The Sempervivum sobolifervm illustrated here may be taken as an example. The thick, fleshy leaves of this plant are arranged, as in all House-leeks, like rosettes on abbreviated axes, and the new rosettes are always laid down as minute buds in the axils of the rosette-leaves. From these buds proceed thread-like runners, furnished with smaU adherent scales, ending in a reduced shoot. The crowded leaves of this reduced shoot enlarge, forming a small rosette, the leaves being folded so closely together that the whole structure has a spherical form. For some time the round rosette is nourished by means of the fllamentous runner from the old plant, but afterwards the runner withers and dries up and the rosette breaks away from it. It is now quite separated from the parent plant (see fig. 453). A gentle breeze is sufficient to roll along the small detached balls; and as the House-leeks in question choose DISPERSAL BY WIND. 821 as their habitat narrow ledges in rocky places, it is inevitable that some of the separated rosettes should fall over the steep wall, and should not come to rest till they have travelled a considerable distance from the mother-plant. Roots are soon developed from the base of the detached rosettes, by which they become fixed to the substratum. Usually a parent plant produces 2-3 rosettes, but frequently as many as six, and the neighbourhood of the terraces overgrown with the species of House -leek figured, and with other allied species {Senvpervivum arenarium. Fig 453 — Sempervivum soboliferum. On the lower step of the rock he five hall-shaped offshoots which have become detached from the upper rocky platform and have rolled down. The butterfly and snail are introduced into the picture to show the true proportions of the offshoots. S. Neil/reichii, S. hirtum) often looks as if it had been sown with the ball-like rosettes, which have rolled down. SeduTn dasyphyllum (see fig. 454^), which grows in rocky crevices and in the niches of old stone walls, develops offshoots partly in the floral and partly in the foliage region. In the floral region the offshoots originate by the metamorphosis of floral-leaves into foliage. Instead of flowers there are small rosettes (fig. 454^) of thick, ovate, green scales, like those which take the place of flowers in Saxifraga nivalis and 8. cernua (cf. p. 455). These rosettes in the autumn break away from the flower-stalks, and behave just like those of Sempervivum. In the foliage region the offshoots arise in three ways. In the axils of the uppermost leaves there is formed a bud which is hardly perceptible to the naked eye. It is embedded in the 822 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. shallow depression on the upper side of the thick leaf, and possesses 2-3 leaflets about 5 mm. in diameter (fig. 454 ^). In the axils of the lower foliage-leaves short sprouts are formed, whose axes are furnished with fairly large crowded rosette- shaped leaves (fig. 454^). In the axils of the lowest arise rudimentary sprouts, with an elongated thread-like axis bearing 8-14 thickly-crowded leaflets at its end {fig. 454*). As soon as the stem carrying the flowers begins to wither, the foliage- leaves and the buds or sprouts in their axils loosen from it and fall to the ground. The succulent, very turgid, almost hemispherical leaves are comparatively heavy, and if the spot where they first fall is sloping they do not lie still, but roll down until they are caught by some pro- jecting ledge, or a mossy cushion, or arrive on level ground. Since they carry with them the buds and sprouts formed in their axils, they to a certain extent function as a means of transport. As soon as the offshoots come to rest, they develop rootlets at their base at the expense of the reserve materials of the de- tached succulent leaf. Rootlets are often formed even while the leaves are still adhering to the decaying stem. It is worthy of note that the aqueous tissue of the fallen leaves also plays a part in the establishment of these ofishoots. If the spot where they have come to rest is exceptionally dry, as is usually the case in places where Sedunru dasyphyllum grows, the supporting leaf may for a long time provide the water necessary for the mainten- ance of the ofishoot, and so protect it from perishing. The formation of sprout-like offshoots is very remarkable in the Kleinias, natives of the Cape, which belong to the Compositae. Some species of this genus, viz. Kleinia neriifoUa and K. articulata, remind one very much in their appear- ance of certain Cacti. The fleshy, much-thickened cylindrical branches are connected with one another by thin strands, and the whole plant looks as if it had been constricted at intervals by ligatures. The strands joining the heavy cylindrical branches break at the slightest pressure, and the upper shoots especially may be broken off even by a violent gust of wind. The result of the fracture at the constricted plaxjes, however, is that the branches fall to the ground. If the plant grows on a slope, the fallen cylindrical shoots roll down until they are stopped by a projecting stone or some other obstacle. When they come to rest they develop numerous roots Fig. 454.— The formation of oilshoots in Sedum dasyphyllum. » Entire plant; nat. size. 2, s, and *, Offshoots which have developed at different levels on the stem in the axils of the leaves. 6 off- shoots from the floral region. DISPERSAL BY WIND. 823 where they touch the ground, and at the same time send up new lateral branches from the opposite side, as shown in fig. 455. It should be mentioned that in K. articulata the roots often begin to develop before the branches have broken and fallen off, always appearing on the side of the shoot which is turned towards the soil. This also is shown in the figure. Fig. 455. — The formation of offshoots in Kleinia articulata. In all these instances the ofishoots are detached by the force of the wind. Another method by means of which the same end is attained depends on the hygroscopic properties of the tissues concerned, and on the alternate swelling and contraction from this cause. Several Fungi of the group Peronosporese, among others the unwelcome Potato-disease Fungus, Phytophthora infestans, multiply by spores formed on delicate hyphal threads, which are protruded from the stomata of the 824 THE DISTfilBUTION OF SPECIES BY OFFSHOOTS. host plant. These hyphal threads bifurcate, and the end of each branch swells up into a spore. The supporting hyphal branch then grows out again below each spore, elongates, and extends upwards, and pushes the spore on one side. The result of this oft-repeated process is a structure which resembles a small much- branched tree, with egg-shaped fruits hanging from the boughs. The hyphal branches, on which the spores are set like fruits, are cylindrical, stiff, and turgid in damp air, but in dry air, especially when they are ripening, they become ribbon-like and spirally twisted so as to resemble cotton-cells. They are extremely hygroscopic, and the slightest change in the humidity of the surrounding air is enough to increase or diminish their spiral torsion. Even mere breathing on them produces an alteration in the twisting, and if a rapid and marked alteration occurs in the hygroscopic condition of the environment, the branches with their hanging spores are whirled hither and thither, and the spores, which are only attached but slightly, are scattered in all directiojis. This cannot of course be seen except under unusually favourable circumstances, on account of the minuteness of the spoi'es. The shedding of the spores can be observed with the naked eye in the Mould Pilobolus cristaZlinus, one of the Mucorineee, shown in figs. 456 ^ and 456 ^. The mycelium of this Mould consists of a colourless, much-branched tube, and grows on the excrement of horses and other mammals. Enlargements arise on the mycelium, and from each is produced a sporangial mechanism composed of two parts, a colour- less, barrel-shaped, stalk-cell and a dark head. The latter contains a colourless jelly, which swells up in water, together with numerous spores, and is to be regarded as a sporangium. Its wall is covered with calcium oxalate, so that its elasticity is completely lost and it becomes brittle. The cell-wall of the barrel-like swollen stalk, however, remains soft and elastic. At the junction of the dark sporangium with its colourless stalk a circular layer of separation is formed. When the turgidity of the sporophore increases in consequence of the absorption of water from the mycelium the tension at last becomes so great that it causes a rupture round the circular line mentioned. At the same moment, however, the elastic wall of the part of the sporophore immediately below contracts, and the fluid contents are pushed out with great force. The push is transmitted to the dark sporangium above the split, and both the fluid contents of the club-shaped support and the entire sporangium are thrown off (see fig. 456 ^). The force of the explosion is so considerable that the dark mass is raised about a metre in height. The whole process, which, as we have said, may be seen with the naked eye, usually occupies 18-20 hours. The development of the mechanism begins at mid-day; during the night the spores are formed in the vesicle, and the next morning the explosion occurs as soon as daylight appears. A no less interesting spectacle is afforded by the scattering of the unicellular off- shoots, i.e. conidia, in species of the genera Empusa and Untomophthora. These live on the dead bodies of caterpillars, flies, aphides, and other insects, the commonest and best known being Empusa musccB, which lives on the common house-fly. When a conidium of this Empusa falls on the body of the fly it puts out a tube which pene- DISPEESAL BY EXPULSIVE MECHANISMS. 825 trates into the body-cavity, and there it divides up repeatedly, forming numerous cells throughout the body. The infected fly, sickening under the injurious influence of the Fungus and almost at the point of death, seeks for some quiet spot in which to die. It frequently chooses for its last resting-place a window pane, in which case it is possible to thoroughly investigate the further development of the Fungus, After the death of the fly the round cells of the Empusa, hitherto hidden in the body-cavity, grow out into long tubes which pierce the skin of the fly's corpse and appear as short club-shaped structures on the surface. A single egg-shaped conidium Fig. 456. — Distribution of spores by expulsive mechanisms. i Pilobolus cHstallinus before the sporaQgium breaks away. 2 xhe same at the moment when the sporangium is thrown off. 3 Sporangium of Nepkrodium Filix-mas closed. * and ^ The same in the act of splitting and scattering the spores. 6 Sphcerobolus stellatus at the moment when the balls filled with spores are thrown off. '^ Peziza aurantia. 8 Longi- tudinal section through this Peziza. The spores are escaping from two of the asci. All the figures magnified is then cut off" from each club-like end of the tube, and this is thrown off" in exactly the same way as the sporangium of Pilobolus (cf. fig! 383 '', p. 672). Here, too, a place is formed for the splitting, and here again the mucilaginous contents are thrown off simultaneously with the conidium by the sudden contraction of the club- shaped end of the tube, and the conidium is thus always surrounded by a gelatinous adhesive mass (fig. 383 *). The distance of the projection may be as much as 2-3 cm., which, considering the extraordinary minuteness of the conidia, is proof of great power. The dead fly then appears to be surrounded by a veritable halo of detached conidia which are firmly attached to the substratum (fig. 383 ^). This is to be accounted for by the fact that, as( already stated, a part of the sticky mucilaginous contents of the club-shaped end of the tube are thrown out with the conidia. This serves as an adhesive material, and causes the conidia to adhere particularly firmly to glass window panes. If a living fly which happens to be near is struck by the projected 826 THE DISTEIBUTION OF SPECIES BY OFFSHOOTS. conidia, they stick to it so firmly that it cannot succeed in getting rid of them or freeing itself in spite of all its attempts. Each adhering conidium then again sends a tube into the body-cavity of the fly, and the development is repeated in the way just described. The same thing happens in Entomophthora radicans, which lives on the caterpillar of the Common White butterfly (Pieris Brassicce). It is represented in figs. 383 ^' ^' ^' *■ ^ p. 672). Tufts of delicate thread-like hyphse come out of the body of the caterpillar for the purpose of forming conidia (fig. 383 ^). These gradually form a thick web round the dying caterpillar, and at a cursory glance one might think it had woven its covering and changed into a chrysalis (fig. 383 ^). The tubes, looking like fine threads, unlike those of Empusa, are here much-branched, and actual tufts of hyphee arise from whose ultimate somewhat swollen ends the long, sticky conidia are abstricted and scattered (figs. 383 * and 383 ®). The scattering of the spores from the asci of Ascomycetes takes place in a characteristic manner. They are developed in groups of 2, 4, 8, 16, or 32 in the tubular asci, and numerous thread-like hyphal ends, the so-called paraphyses, occur between the asci (see fig. 456 ^). In addition to the spores the asci contain proto- plasm and cell-sap, and are considerably distended by the large amount of the latter. As the dilatation increases the asci burst, and their cell- wall, which is at a high tension, exercises a powerful pressure on the cell-contents, which are extruded with great force. The place where the rupture of the wall of the ascus occurs is determined beforehand, so that the extrusion of the cell-contents and spores always takes place in the same way. In many species the top part of the ascus-wall is raised like a lid, in others a transverse splitting occurs, and in others again the spores are ejected through a small circular hole. A slight shake or a dry breeze is quite enough to cause the ejection, and in Spatularia flavida, for example (figured on p. 791), or in Peziza aurantia (see fig. 456^), it is easy to observe how small clouds of extruded spores rise from the surface of the fructification as soon as these Fungi are brought from a damp place into a dry atmosphere, or when a dry wind blows over them. In some species of Ascobolus, minute black or waxen yellow Fungi living on the excrement of animals, the spores are not only ejected, but the turgidity of the tissue surrounding the tubes is so great that the whole tubular layer is extruded with the spores. Some Gasteromycetes have special contrivances for scattering the spores. In species of the genus Geaster (see figs. 391 * and 391 ', p. 690) the threads of the capil- litium and the spores imbedded between them develop within a tough, leathery, bladder-like envelope which separates into two layers when the spores are ripe. The inner layer has the form of a bladder, and opens only at a spot at the apex. The outer layer, on the other hand, splits into 4-12 radiating lobes. The position of the lobes alters remarkably according to the hygroscopic condition of the atmo- sphere. In damp weather they fold together over the vesicle, but in dry weather, especially in sunshine and when a dry wind is blowing, they bend back so forcibly that some of the spores are shaken from the mouth of the vesicle. Travellers in Central America tell us of the gigantic Puff"-balls which literally explode on being DISPERSAL BY ANIMALS. 827 shaken, sending such quantities of reddish spores into the air that it is impossible to breathe in their vicinity. In Europe a minute Puff-ball, Sphoerobolus stellatus (fig. 456 % grows on decaying stems, leaves, &c. The wall of the fruit divides, as in Geaster, into two distinct layers: one remains closed and assumes the form of a ball, but the outer one when the spores are ripe divides by radiating clefts into several lobes. These bend back rapidly on drying, and as the central portion round which the lobes are placed becomes strongly arched upwards, at the same time the ball containing the spores is shot out with considerable force. The dissemination of spores in some of the Ferns is illustrated in figs. 466 ^■*'^. Sporangia are developed on the under surface of the frond, where they are arranged in various ways. Those of the Nephrodium Filix-mas, which is here selected as a type, consist of a stalk and a flattened bi-convex vesicle. Eound the latter runs a ring of darker-coloured cells, whose side-walls are much thickened, while their outer walls remain thin and delicate. When the sporangium is ripe its bursting is brought about by the contraction of the cells of the ring. With regard to the distribution of offshoots by animals we may distinguish two classes, those in which the offshoots are first conveyed to the animals by special disseminating mechanisms, so that two methods of distribution are combined, and, secondly, those in which animals alone effect the transport of the offshoots from one place to another. We have already spoken repeatedly of the former class. Of the latter the distribution of spores by food-seeking animals is the first to be considered. The Pyrenomycetous Fungus known as the Ergot of Eye {Claviceps purpurea) is a well-known instance. The thick web of hyphal threads which invests the ovaries of the Eye is penetrated by labyrinthine passages, whose walls are formed by the ends of hyphal threads arranged in rows and tufts (see fig. 386 ^, p. 680). Spherical spores are abjointed from these somewhat club-shaped ends. Simultaneously with this abjunction the outer layer of the cell-wall of both spores and hyphae forms a sugary fluid by the absorption of water and subsequent breaking down. This fiUs the winding passages, and the innumerable abjointed spores are im- bedded in it. The sweet-tasting fluid gradually collects into drops on the exterior, and even comes into view on the spikes of Eye between the glumes which surround the infected ovaries. This is the "honey dew" by which the presence of the parasitic Claviceps in the interior of the spike is recognized, and which is viewed with some apprehension by the farmer. Insects, especially wasps and flies, eagerly seek out these springs of sweet fluid and suck and lick up the juice, which is crowded with numberless spores. It is therefore inevitable that small quantities of spores should stick and remain hanging to portions of their bodies, and when they fly to the spikes of other Eye-plants the spores are easily rubbed off, and in a very short time may again grow up into a mycelium involving the ovaries there. A similar phenomenon may be observed in the Phalloidess, belonging to the Gasteromycetes, of which the best known species, the Stink-horn Fungus (Phallus impudicus), may be taken as an example. The cap, borne on a white cylindrical and spongy stalk, is bell-shaped and covered with a greenish-black viscous fluid in 828 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. which numerous spores are imbedded. This fluid gives off a far-reaching carrion smell which allures many insects, especially carrion-flies. The fact that the fluid contains sugar which serves as nourishment for the insects also contributes to the allurement. A fly which alights on the cap of the Stink-horn cannot leave it without spores adhering to its whole body. Some of them may, perhaps, fall off while it is flying away, but the majority will not be brushed off until it again alights and cleanses itself from the uncomfortable appendages {of. also p. 691). It is well known that the fleshy fructifications of Hymenomycetes provide food for numerous insect-larvse. Frequently as soon as the receptacles appear above the soil the flesh of the stalk and cap are riddled by passages in which live the larvae of various gnats and beetles. These leave their dwellings before the decay and decomposition of the Fungi set in, and enter the chrysalis stage in the ground. In this way numerous spores which have adhered to the animals are carried away and disseminated. The spores of various Fungi, especially of the Hymenomycetes and Truffles, are without doubt distributed by animals which eat the fleshy spore- bearing portions. The spores pass unharmed through the alimentary canal and then germinate in the deposited excrement. Earthworms and swine in particular seem to take part in this distribution. The dissemination of detached bud- and shoot-like offshoots is comparatively seldom effected by animals. Of the cases known the following are the most note- worthy. First, where the offshoots are taken up as food by animals, but are again got rid of in an undigested condition, and grow up into new plants in the place where they have been deposited. This has certainly been observed in Poly- gonum viviparum, which grows commonly in the far North and on the high mountains of Central Europe (see fig. 452). The bulbils of these plants are a dainty morsel to ptarmigan, and are eagerly sought for by them. The ptarmigan seizes the lower half of the spike of the Polygonum with its beak, and by a quick movement of its neck passes the bill the whole length of the spike, and so puts dozens of bulbils at a time into its crop. Numerous observations have shown that the bulbils of Polygonum viviparum and cranberries are the commonest food found in the crops of ptarmigan shot on the Alps, and I also always found these bulbils in great quantity in the crops of Norwegian .ptarmigan. The portion which passes from the crop into the muscular gizzard is of course crushed and digested, but it has often been noticed that part of the food so greedily swallowed by the ptarmigan is thrown up again, and this is particularly the case with the bulbils when they have been taken in excess. When thus extruded, they have the power of further development; far from being destroyed, they grow up very rapidly into new plants, and as the places where the superfluous food is thrown out are always at some distance from the spot where the ptarmigan obtained the bulbils, this process is really a mode of distributing the Polygonum viviparum. The second method of distributing detached offshoots by animals is effected by means of barbed bristles and hairs, such as are represented in fig. 457, in the Mamillarias (Mamillaria placostigma and gracilis) of the high mountains of Mexico. DISPERSAL BY ANIMALS. 829 Here some of the spherical, closely-crowded lateral shoots growing from the old plant loosen spontaneously and fall to the ground; others again remain in situ but adhere very slightly, so that a passing contact or a gentle touch is enough to complete the separation from the old plant. Now bristles are formed at the top of each papilla of these Mamilkrias, some of which end in barbs, so that the spherical shoots resemble burs. They adhere just like burs to the hairy paws or fur of grazing animals, which carry them away unconsciously. Afterwards Fig. 457.— Distribution of detached sprout like oflshoots by means of animals, i Mamilldria placostigma, 2 Mamillaria gracilis. the animals when resting seek to rid themselves of the inconvenient appendages, brushing them off and leaving them behind on the ground. Here they may strike root and grow up into new plants. The third method of distribution of sprout-like oifshoots by animals is seen in aquatic plants, which fasten either entirely or in fragments to passing water-birds. Certain species, which very rarely blossom or form fruit, but nevertheless occur in innumerable widely distant spots and often appear unexpectedly in newly-formed ponds, in artificial lakes, and in other waters, are for the most part distributed by water-birds. Some of these water-plants, e.g. the Frogbit and Bladderwort (Hydro- charis and Utricularia), develop peculiar slimy coverings round their buds, which 830 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. cause them to stick to the feathers of birds which come in contact with them as they swim by. Others, such as the small Duckweeds (Lemna minor, gibba, folyrrhiza), hang by their long, somewhat twisted, floating roots, and many filamentous Algse, Aldrovandia, the delicate Riccias (Riocia natans and fiuitans), the Ivy-leaved Duckweed (Lemna trisulca), &c., become attached in their entirety to the coot and duck swimming in the ponds and lakes. These fly away with them, but as soon as they again enter other water the adhering plants fall oflT or are cleaned off by the birds, and in this way they are distributed quite fresh and living over great distances. We might also mention in this connection the peculiar distri- bution of Ulvas, Floridese, and Sea- wracks by means of crabs, which was described at vol. i. p. 77. We will only allude in passing to the fact that many economic plants are propa- gated and distributed by ofishoots to a very great extent by man. Bananas whose fruits contain no fertile seeds, Potatoes, Artichokes, and many other tuberous and bulbous plants are continually multiplied by the help of slips, tubers, bulbs, &c. The intentional artificial propagation by offshoots has of course no apparent influence on the development of a natural method of distribution in such species. Although planted and cultivated in large quantities they do not become naturalized; and if it were not for the artificial maintenance and propagation by offshoots they would soon vanish again from such places, leaving no trace behind. This is, how- ever, not the case with the unintentional distribution of offshoots of certain plants by man. The keels and bottoms of ships journeying over wide seas become, like the stakes and buoys of the harbour and the sea-walls and rocks of the shore, quite overgrown with sea-weeds. If these are removed by chance or intentionally from their substratum they do not necessarily perish. They may remain alive in the sea- water, and under favourable conditions may attach themselves to some other firm spot of ground. In this way plants may be transmitted from one coast to another over very wide distances. Another unintentional distribution of plant-offshoots by man occurs on cultivated ground in vineyards, fields, and gardens. By ploughing, digging, and throwing up the soil the bulbous or tuberous offshoots embedded in the ground undergo a change of position. The offshoots of certain plants may in this way be distributed so uniformly over a whole field by spade and ploughshare in the course of a year that it almost looks as if they had been purposely planted there. It is curious on journeying through the vine-planted districts of Northern Italy to see one of two adjacent vineyards abundantly covered with wild Tulips, while not one is visible in the other. In Central Europe the same thing happens with the Gageas (Gagea arvensis, 0. stenopetala) growing in the fields, and with the tuber- forming Earth-nut pea (Lathyrus tuberosus). One field looks as if it had been sown with Gageas and yet its neighbour is completely devoid of them. On the Giinsel- hohe in the Lower Austrian Erlaf thai I once saw a rectangular ploughed field over- grown from one end to the other with plants of the Bulbiferous Lily {Lilium bul- biferum), while only isolated specimens of this plant could be seen in the adjoining fields. There is no doubt that here the bulbils thrown on to the ground from the DISPERSAL BY ROOT-SHORTENING. 831 leaf -axils of a few plants had been distributed equally by ploughing over the whole field, although this distribution had certainly not been intended by the ploughman, It would of course be a mistake to explain the uniform distribution of bulbous, plants over a large stretch of land exclusively by the ploughing and overturning of clods of soil full of bulb-like oifshoots. In many instances the distribution of such oflshoots is also produced by the pulling action of the roots. This process is so remarkable that we must describe it somewhat in detail. The multiplication of subterranean bulbs is known to take place by the formation of buds in the axils of the scale-leaves, and these, in the course of a few months, themselves grow up into small bulbs. When mature, they may form the termination of a slender shoot which, of course, never attains any considerable length, but which in many cases is thread- like, as shown in Muscari racemosum (fig. 444 ^). The small bulbs are pushed by this thread-like shoot out of the region of the protecting scale-leaf near the old bulb, and there they develop long root-fibres in abundance. In other instances the shoots remain extremely short, and the small bulbs are not pushed out, but the protective scale-leaf, in whose axil they originated, decomposes, and they send out their roots through the decomposing tissue into the surrounding soil. In both cases they become detached at the end of the vegetative period in which they originated; they are then no longer connected with the old bulb, but are quite independent. Many species form only one bud in the axil of a bulb-scale, others a whole series which all grow up into bulbs; in the latter case the old bulb in the autumn is. surrounded by a whole family of small young bulbs. There is a species of Garlic called AlUuTU pater-faTnilias whose old bulb gives rise to about a hundred young- ones in a year. It is impossible for so many to develop properly when closely crowded together round the plant from which they sprang; mutual pressure would be unavoidable in their further growth, and if next year each of these bulbs should in its turn form new oflTshoots, and again become the centre of young bulbs, it would become imperatively necessary to make room, and to thin and separate the dense crowd. Since all the bulbs are placed with their apices pointing upwards- they cannot be moved apart by the elongation of their stems; the mutual pressure of neighbouring bulbs as they enlarge would certainly cause a trifling displacement,, but this would not prove an efficient remedy. The remarkable pull of the roots, which was described in vol. i. p. 768, now comes into play. Only a few of the roots arising from the base of a young bulb strike downwards; by far the greater number grow out at a right angle to the axis of the bulb in a direction parallel with the surface of the soil (see fig. 444 ^). When these very long roots have stopped grow- ing they contract, and thus draw the young bulb to which they belong away from the old one. The young bulbs now form a wide open wreath round the old one (which has meanwhile disintegrated), and thus obtain sufficient room for further development. This happens not only in the Muscari described, but also in Ornithogalmn nutans, Tulipa sylvestris, and indeed in quite a number of bulbous, plants. Since this process is repeated annually a fairly wide area of soil may in the course of years be covered with the bulbs in spite of the slight distance through, S32 THE DISTRIBUTION OF SPECIES BY OFFSHOOTS. which they have been shifted under the ground. Some soil containing bulbs of Tulipa sylvestris was once put in a garden in Vienna in the middle of a grass plot shaded by Maple-trees. As the grass was mowed every year before the flowers opened there was no formation of seed, and the Tulips could only multiply by offshoots. After about 20 years the lawn was covered with Tuhp-leaves, which arose from subterranean bulbs occupying an area 10 paces in diameter. Thus in the time mentioned the bulbs had spread for about 5 paces in all directions, in consequence of the pull of the contracting roots. It is more than probable that the offshoots of many perennial plants with erect stem and napiform or tuberous roots, e.g. the blue-flowered species of the Monkshood (Aconitum Napellus, A. JSfeubergense, A. variegatum) undergo a change of position by the pull of their horizontal root- flbres; and that the clustered arrangement of these plants is the result of the root-pull. A review of the very varied modes of origin and distribution of offshoots leads to the conclusion that they may be formed on all parts of the plant, that the form of the offshoot is constant for each species, or, in other words, that the form of the individual parts of the offshoot in succeeding generations is repeated as exactly as the flowers and fruit, but that one and the same species may frequently form two or even three kinds of offshoots. The Fungus Glaviceps purpurea develops spores which are distributed by honey-sucking insects, also the sclerotia known as " ergot ", which are scattered from the dry spikes by the swaying movement of the stem, and thirdly, filamentous spores, which are extruded from asci, and distributed by wind. The Liverwort Blasia pusilla, develops thallidia in special flask- shaped receptacles on the surface of the thallus, and spores in the sporogonia. The form of the offshoot is always adapted to the season and to the distributive agents available where they are formed. In one case it is more suitable that the offshoots should be distributed slowly, and step by step, in another quickly and by bounds. In the spring it may be more advantageous if they are distributed by wind, by animals in the summer, and by self -scattering mechanisms in the autumn. Steppe- plants must develop different offshoots from those formed by plants living on the damp, shady floor of the forest. It is just as obvious that offshoots, which creep along, above, or under the ground without leaving the soil, must be equipped quite differently from those which are detached from their place of origin, and either roll along or are carried by wind, or have to travel long distances as the appendages of wandering animals. In the former, it is all-important that they should be able to overcome possible obstacles in the soil; in the latter, that they should not perish during their journey for lack of food and water. When separated from the soil they are greatly exposed to the danger of drying up, and even when they have settled somewhere, the supply of water they require for the formation of organs of attachment and absorption is by no means assured. Settlers of this kind must either be so .organized that they can sustain a long-continued drought without injury, like the offshoots of the Mosses and the soredia of Lichens, or they must themselves bring with them the necessary water supply, and care must be taken THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. 833 that this supply is not lost prematurely by evaporation. As a matter of fact, such detached offshoots, e.g. those of Sempervivum, Sedum, Kleinia, or Mamillaria, are not only provided with a special aqueous tissue, but also with a cuticle which is very effective in preventing excessive transpiration. All offshoots, when liberated from their place of origin, are also provided with the necessary reserves, i.e. con- structive materials, so that immediately after settling, they can send out absorbent roots and green leaves of their own initiative, obtain a firm footing in their new locality, and extract nourishment from it. When the offshoots are distributed by water-currents, they require neither an aqueous tissue nor protection against drying up, and it may be due to this fact that detached offshoots are relatively more fre- quent in aquatic than in land plants and lithophytes. THE DISPEESION OF SPECIES BY MEANS OP FEUITS AND SEEDS. On the heights of the Kahlenberg, at Vienna, at the edge of the wood, grows an under-shrub which bears the name of Borycnium herbaceum. It is one of the Papilionacese, and develops spherical one-seeded fruits, which ripen in October. I once collected from this plant several twigs laden with fruit, for the purpose of a comparative investigation on which I was engaged, and brought them home and laid them on my writing-table. Next day as I sat reading near the table, one of the seeds of the JDorycnivmi was suddenly jerked with great violence into my face. Shortly afterwards I saw a second, third, fourth, and ultimately about fifty seeds let fly from the small clusters of fruit, and each time I heard a peeuKar sound which accompanied the bursting open of the fruits and ejection of the seeds. The rays of sunshine from the window had evidently heated and dried the fruits, and occasioned this surprising phenomenon. The incident reminded me of the following passage in Goethe's Travels in Italy: — "I had brought home several seed-capsules of Acanthus m,ollis, and put them away in an open box, when one night I heard a crackling noise, and immediately afterwards a sound like the impact of small bodies against the walls and ceiling. I could not understand it at first, but found afterwards that my pods had burst and scattered their seeds all over the place. The dryness of the room ha-d caused the fruits to ripen in a few days to the requisite degree of elasticity." The fruits of Dorycnium and Acanthus may be taken as types of a large group designated by the name of Sling-fruits. It is found that when these fruits are ripe, the tissue around the seeds becomes highly tense. The first result of the tension is that the tissue is rent at particular spots, and this rupture is followed by a sudden contraction of the segments, which double back and roll up, at the same time expelling the seeds resting upon them. Sometimes the roUed parts of the fruits, and, more rarely, the entire fruits themselves, are jerked off simultaneously with the seeds. There is the greatest variety in this respect, but all the con- trivances for expelling seeds resemble one another in the fact that through their agency the seeds reach places beyond the range of the mother-plants. VOL. II. 103 834 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. In one class of expulsive fruits the high degree of tension which finally results in the disruption and rolling up of particular tissues is caused by a swelling up of the cell-membranes or by the turgidity of the cells. One of the most curious instances is that of the Squirting Cucumber {Ecballium, Materium), which is shown in fig. 458 1. This plant belongs to the Cucurbitacese and its fruit resembles a small fleshy cucumber beset with bristles and borne by a hooked stalk. The end of the stalk projects into the interior of the fruit like a stopper. When the seeds are quite .^SiSajn.. Kg. 468.— Sling-fruits. f**^^^*""^ • Ectallium Elaterium ; branch bearing flowers and fruits. ' A fruit detached from its stalk and with its seeds squirting out. 8 Oxalis Aoetosella; entire plant with one unripe fruit on a hooked stalk, and one ripe fruit on an erect stalk ejecting its seeds ; nat. size. * Vniipe trait ot Oxalis AcetoseUa; x6. '''Ri'gettmtoiOxalis Acetosella ejeatlngttieseedLs; x6. ripe the tissue surrounding them is transformed into a mucilaginous mass. Also the tissue in the neighbourhood of the conical stopper just referred to breaks down at the same time, and thus the connection between the stalk and the fruit is loosened. In the wall of the fruit there is a layer of cells which is under great tension, and endeavours to stretch itself out. As long as the fruit is unripe such expansion is prevented by the tense tissue close to the stalk, but with the ripening of the fruit this obstacle is removed. The fruit then severs itself from the conical end of the stalk and at the same moment the expansion of the strained layer of tissue takes place. The consequence is that the interior of the fruit is subjected to great pressure, and the seeds, together with the surrounding mucilage, are squirted out SLING- FRUITS. 835 with considerable force through the hole which was previously closed by the end of the stalk (see fig. 458 ^). The DorsteniacesB behave in a manner no less remarkable. As in the case of Figs so also in these plants, numbers of small flowers are seated upon an enlarged receptacle, which remains fleshy and succulent after the small one-seeded fruits have developed from the flowers. The lower portion of each fruit has thick walls, and is embedded in the receptacle like a hair-follicle in the human skin, whilst the delicate- Fig. 469.— Sling-fr«its. 1 Orobus vermis. 2 and 8 Oeranium palnstre. ' Viola elatior. 5 Cardamine impatiens. « Impatiens Nolitangere. ^ and 8 Acanthus mollis. ^ and 10 Ricinus communis. coated portion projects above the receptacle in the form of a papilla. When the seed is quite ripe the turgidity of the outer cellular layer of the thick wall of the fruit increases, the thin-walled top is torn, the thick walls suddenly close, and the seed hitherto enveloped by them is violently ejected. A special case of the expulsion of seeds as from a sling is also found in Oxali- .daeese, of which the common Wood-sorrel (Oxalis Acetosella, see figs. 458 ^' *> ^) may be taken as an example. In this case it is the seed -coat that possesses a special tumescent tissue adapted to the expulsion of the seeds. One of the deeper layers of the seed-coat is composed of tense cells and is itself in a highly strained condition, 836 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. whilst the outer layers of cells of the seed-coat are not in a state of tension. When the seed is quite ripe the cell-membranes in the strained layer of tissue swell up, the outer layer of the seed-coat, being no longer able to withstand the pressure to which it is subjected, is rent asunder and the edges of the slit thus formed roll suddenly back. A violent jerk is given to the inclosed seed, in consequence of which it flies out through the fissure in the capsule immediately in front of it (fig. 458 *). The fact of the ejection of the seeds of Balsamacese also has long been known. The fruit of Impatiens Nolitangere, one of the members of that family, is an oblong capsule composed of five carpels (see fig. 459 ^). The walls of this capsule are con- structed of three layers of cells. The layer lying immediately beneath the epidermis consists of large and highly turgid cells, and is called the turgescent layer. It is in a state of great tension, and when the seeds are ripe and the union between the five carpels gives way along the lines of union, a relaxation of the tension takes place, the loosened tissue of those lines is torn, the five carpels roll up, and their rapid movements of involution result in the expulsion of the seeds contained in the fruit. Gydanthera explodens and Thladiantha dubia, plants belonging to the Cucur- bitacese, as also several Crucifers of the genera Dentaria and Gardamine, in particular the species shown in fig. 459 ® (Gardamine vmpatiens), exhibit similar phenomena, except that in these cases the carpels do not roll inwards but outwards. In the instances hitherto dealt with the cause of the expulsion is the turgidity of cells or the swelling up of cell-membranes with a concomitant maintenance of a state of extreme tension in a particular layer of tissue situated in the wall of the fruit. In the next class of cases the phenomenon depends on the desiccation and consequent contraction of a special layer of the fruit-wall which leads to a rupture and subsequently to a bending over and rolling up of particular parts of the fruit. This change is accomplished with great rapidity and has the effect of hurling away the seeds or the separate parts of the fruit or even the entire fruit itself. We wiU only mention some of the best known instances of this kind. The fruit of the Marsh Crane's-bill (Geranium palustre; see fig. 459^) has a 5-angled column rising up in the centre of a circle of five carpels. The carpels are hemispherically inflated at the base, and terminate above in long bristles or beaks. Each contains a single seed. When the seeds are ripe, the tissue composing the beaks undergoes desiccation, which, however, is not of uniform intensity throughout. The outer layer, consisting of several plates of succulent cells, dries up more quickly than the inner layer, which is composed of thick- walled cells. The result is that the beak lifts itself away from the axial column, and curls up externally like a watch- spring. No resistance to this movement is afforded by the delicate dried tissue which has hitherto served to hold the carpels together, and as the cavity of each carpel is open along the inner surface, and the seed lies in it simply as though it were resting in the hollow of a hand, the rapid drawing up of the beak has the effect of ejecting it in a wide curve away from the carpel (see fig. 459^). In the Marsh Crane's-bill, as also in the other large-flowered species of the genus Geranium, the tops of the beaks continue attached to the axis, and the latter, together with the five SLING-FRUITS. 837 empty and rolled-up carpels, resembles a chandelier in shape (shown to right of fig. 4593). Those Violets which have aerial stems, such as Viola elatior (see fig. 459*), develop capsular fruits, each of which resolves itself into three valves when it bursts open. The valves are boat-shaped, and the marginal parts which form the sides of the boats are thin, whilst the keels are very thick and swollen. Inside each boat, near and parallel to the line of the keel, are two rows of seeds. The valves them- selves have an exceedingly complex structure. A cross section through one of them shows a layer of thin-walled parenchymatous cells, a layer of elongated curvilinear cells, and a layer of broad, greatly thickened cells. The unequal desiccation of these layers is the cause of the curving up of the lateral walls of the valves, which at last approach so near to one another as to exercise considerable pressure on the seeds in the middle. The result of this pressure is that the smooth seeds are shot out with about the same force as is imparted to a cherry-stone when it is flicked to a distance by the finger and thumb. The seeds are ejected in regular succession. The foremost seed of the first carpel goes first, and the seeds at the opposite extremity are dis- charged last. It is not till the first carpel is quite empty that the second begins to part with its seeds, and the third only comes into play when the second is finished. The drawing together of the two sides of the valve always begins at the free extremity of the valve, and lasts until all the seeds have been ejected. In many Mimoseae, Csesalpinese, Papilionacese, Sterculiacese, and Acanthacese the seeds are expelled by means of a spiral torsion of the valves of the fruit at the moment that the legume or capsule opens. The wall of the fruit of these plants includes a soft succulent layer of thin- walled parenchymatous cells, and a hard layer of strongly-thickened elongated cells, which run obliquely from one edge to the other in each valve. The rupture of the fruit, and the spiral torsion of its valves at the moment of their separation, depend upon these diagonal cells of the hard layer. Each one of these cells winds itself into a spiral as it dries, and consequently the entire layer undergoes a corresponding torsion. The tissues composed of thin- walled cells, which are in connection with the hard layer, ofier no resistance to the move- ment, and the rotation is therefore so sudden and violent that the seeds contained in the pod are projected to a distance. If the fruit is short, the valvular torsion is confined to |-1 twists; if long, the spiral includes 2 or even 3 complete coils, and the valves of the empty fruit are curled up like ringlets (e.g. Lotus corniculatus, see p. 431, fig. 325^, and Orohus vernus, see fig. 459^). The force of projection varies according to the thickness of the hard layer. In Gastanosperinumfi australe, where the pod- valves attain to a thickness of 5 millimetres, the sudden torsion causes the expulsion of spherical seeds, measuring 3'5 centimetres in diameter, and weighing 16 grams. In these cases the valves of the fruit persist upon the fruit-stalks after the ejection of the seeds, and herein lies the essential difference between them and those expulsive fruits of which the carpels break away from the stalks with the seeds. To this class of expulsive fruits belong also several Papilionacese, such as the Dorycnium mentioned at the beginning of this section, and besides them the genus 838 THE DISPERSION OF SPECIES BY MEANS OF FETJITS AND SEEDS. Kitaihelia of the family Malvaceae, Alstrcemeria amongst the Liliacese, several Acanthaceas, including the Acanthus mollis (see figs. 459 ' and 459 ^), which Goethe has made familiar to us, the wonderful parasite, Lathrcea clandestina, and many Euphorbiacese (e.g. Euphorbia, Hura, Hyosnanthe, Mercurialis, Micinus, see figs. 459^ and 459^°). In all these plants the fruit- valves are comparatively short, and the spiral torsion is therefore less clearly manifested. The impulse given to the seeds by the twisting of the valves is supplemented by various other contrivances which cannot here be described, and, as a matter of fact, the range of projection in this group of sling-fruits is wide as compared with that of others. A peculiar form of sling-fruit is found in several of the Diosmaceae, Eutacese, and Zygophyllaceae. In these plants a complete separation of the hard from the soft layer takes place. When the seeds are ripe the external soft layer dries, splits along the ventral suture, and contracts strongly. In consequence of this contraction the hard internal layer, which is in the form of a case inclosing the seeds, is forced out of the slit. As soon as the hard case is thus set at liberty its two lateral walls part asunder, assume the shape of the screw of a steamer, and eject the seeds to a distance. Similar processes occur in the genus Gollomia of the family Polemoniacese, but in this instance it is not the soft outer layer of the valves but the calyx, which, on drying, exercises pressure on the inclosed case, and the latter, which is extruded, is not the hard layer only but the entire dry capsule. The liberation of the case is here materially assisted by the circumstance that the three valves of the capsule disunite at a time when they are still surrounded by the calyx, and hence exert a counter-pressure upon the calyx. When once the capsule is freed from the grasp of the calyx, its valves diverge still more widely from one another and eject their seeds. In EschschoUzia also the entire fruit is jerked off" the receptacle, but here the phenomenon depends on the fact that the two valves of the siliquose fruit attain to a high degree of tension on desiccation and tend to curve outwards. When the tension has reached a sufficient pitch to sever the connection between the fruit- valves and the receptacle, the whole fruit is shot away from the stalk in a curve. In the Stork's-bill (Erodium, see vol. i. p. 619, figs. 147^ and 147*), and in several Umbellifers (e.g. Scandix), the entire fruit is not thrown off, but the constituent parts of the fruit with their tightly inclosed seeds are jerked away from the central axis. This cursory survey is sufficient to give an idea of the great variety existing amongst the sling type of fruit. Of course the fruits in question are always placed in such a position as to render a free flight of the seeds possible. In every case where the fruits before ejecting their seeds or being themselves jerked away are for any reason hidden under foliage-leaves, or are borne by stalks which bend downwards, as in the Wood-sorrel and the Violet (see figs. 458 ^ and 458 *), the stalks straighten out just before expulsion takes place and lift the fruits up above the leaves. In most instances the angle of projection is 45°, and, as is well known, the greatest range of flight is thus attained. The ejected seeds are spherical, oval, bean-shaped, or lenticular. In the last case they are expelled in such a manner as SLING-FRUlTS. 83& to cut edgewise through the air, and it is the invariable rule for seeds to be so ejected as to encounter as slight a resistance from the air as possible whatever their shape may be. Contrivances for determining the direction in which the expelled body is to move are rare. A first indication of some such adaptation occurs in the Wood-sorrel (see fig. 458 ^) and in Ricinus (see fig. 459 "), where the seeds are thrust through an opening of definite shape. In the Acanthacese {Justuda, Acanthus, &c.), the path of projection is determined by the circumstance of the seeds resting before their expulsion against rigid curved bars springing from the partition-wall which runs through the fruit (see fig. 459 ^). The act of expulsion is usually accompanied by a characteristic noise like that of the bursting of a bladder, and the sound amounts to a regular detonation in the case of the dehiscence of the fruits of Hwra crepitans. The range of projection is least when the seeds are small and light, and greatest when they are large and heavy, as is shown by the following table: — Name of Plant. Shape of Seed. Longest Diameter of Seed in Millimetres. Shortest Diameter ol Seed in Millimetres. Weight of Seed in Grams. Range of Projection in Metres. Cardamine impatiens Viola canina Borycn ium decumbens Oeranium columbinum Geranium palustre I/upinus digitatus Acanthus mollis Hura crepitans ellipsoidal oval spherical spherical cylindrical cubical bean-shaped lenticular lenticular 1-5 1-6 1-5 2-0 3-0 7-0 14-0 20-0 30-0 0-7 1-0 1-5 2-0 1-5 7-0 10-0 17-0 18-0 0-005 0-008 0-003 0-004 0-005 0-08 0-4 0-7 2-5 0-9 1-0 1-0 1-5 2-5 7-0 9-5 14-0 15-0 It will be noticed that as a means of distribution the agency of expulsive fruits is confined to a very restricted range. As compared with the distances to which seeds are conveyed by other means, such as the wind, the range of projection of the most powerful contrivances for expulsion, viz. 15-0 metres, is inapprecialbly small. This may account for the facts, firstly, that expulsive fruits are produced by comparatively few plants; and secondly, that such plants as do possess them are for the most part denizens of localities that are sheltered from the wind, where, therefore, the conditions are not favourable to dispersion by that agency. Cardor- rnvne impatiens, Bentaria, Impatiens, Lathrcea clandestina, Mercurialis perennia, Orobus vernus, OxalisAcetosella, Viola canina, and V. sylvatica all inhabit retired and shaded woodlands, whilst others, as, for instance, Geranium palustre and Lathyrus sylvestris, climb over bushes and hedges on the borders of woods. Mention must also be made of the fact that in many cases a second mode of dispersing fruits and seeds acts conjointly with that of expulsion, as is indicated by the name of Impatiens Nolitangere, i.e. "Touch me not". Those sling-fruits, for instance, in which the high degree of tension is due to the swelling up and turgidity of particular layers of cells, are so constructed that the slightest touch on the outside causes a relaxation of the tension and the ejection of the seeds in the direction of the object that has touched the fruit. The animals which frequent the shady woods 840 THE DISPEESION OF SPECIES BY MEANS OF FRUITS AND SEEDS. where Impatiens, CardaTnine, JDentaria, Oxalis, &c., grow, brush against the fruits of those plants in the course of their wanderings, and at once receive a charge of seeds, some of which are sure to be left sticking to the creature's fur or feathers. It has long been known that when animals pass over places that are overgrown by Materium (see fig. 458^) and brush against its fruits, which hang down from hook-shaped stalks, they are bespattered with the mucilaginous mass in which the expelled seeds are embedded, and that as soon as they reach a place of rest they endeavour to get rid of the unpleasant encumbrance. Several contrivances for the distribution of fruits and seeds remain to be described which, so far as regards their results, exhibit the greatest resemblance to the above sling-fruits, although the causes which determine the phenomenon in their case are utterly different. In the last-named the forcible expulsion is due to cellular turgescence, or to movements brought about by the drying up of hygro- scopic cell-layers; in the cases now to be described the result depends solely on the elasticity of stems and fruit-stalks. The stems and stalks in question are strongly resilient, and are strained and curved by a force acting from without. The moment the force ceases to act, their quality of resilience causes them to return to their former position, and in doing so they jerk the fruits and seeds borne by them to a distance. Of these contrivances, which are called balistic means of dispersion of fruits and seeds owing to their analogy to catapults or balistas, we will here deal with five forms. The simplest occurs in the Compositse, whose fruit-capitula are borne upon erect, comparatively long, elastic, flexible stems. The small fruits of the capitulum are already detached from their short pedicels by the time they are ripe and are deposited upon the central disc of the receptacle, which is surrounded by involucral scales, or at the bottom of the basket-shaped fruit-eapitulum into which the floral-capitulum develops. They are so deeply bedded in this situation that it is not possible for them to fall out unless subjected to some external impetus. But the erect resilient stem which bears the capitulum has only to be bent to' one side by a gust of wind or by the touch of an animal for the fruits lying on the fruit-capitulum (which is flat or excavated as the case may be) to be shot off" by the recoil which ensues. In many of the Compositss the involucral scales which form the enveloping basket bend towards one another at the top so as to constitute a roof; they are, however, elastic and flexible and very smooth on the inner surface, so that the fruits when ejected easily slip by them, and yet are to a certain extent guided in the course they take by the tips of the scales. In other Composites, of which the genus Telekia is an example, the floral receptacle is thickly clothed with so-called palese, and the fruits to be ejected, which, it may be noted incidentally, have no pappus, are embedded amongst these palese. The palese are erect and stiff; and are edged with small, upturned teeth; the slightest shock sends the fruits a little higher up amongst the scales, and they cannot then return to their former position, as the stiff- marginal teeth bar the way. The fruits thus seem to make their way up the scales, step by step, as though they were ladders. If, when they have nearly reached the top, there comes a gust of wind which sets the peduncles of the capitula CATAPULT FRUITS. 841 rocking to and fro, the fruits are thrown out from between the elastic tips of the scales and describe an open curve before they reach the ground. A third group of Composites, which may be represented by Oentaurea Pseudophrygia and G. steno- lepis, exhibits the following arrangement: The receptacle is destitute of paleffi, but the involucral scales form a sort of basket at the bottom of which are the fruits. In damp weather the tips of the bract-scales close tightly together, and the short bristles of the pappus crowning each fruit are applied closely to one another. In warm, dry weather especially, under the influence of a dry wind and sunshine, the scales part asunder and the basket stands wide open. At the same time the hairs 1 and 2 Salvia verticillata. ^, 11 Polygonum Virginicum. Fig. 460.— Catapult fruits. , 6 and fi Teucrium Evgancev/m. ? and s TeucHum fiavum. 9 Monarda fistulosa. lo and 1, 3 and i» nat. size ; the others magnified. of the pappus bristle up, and in so doing raise the fruits to the open mouth of the basket. If the peduncle supporting the capitulum is now set in motion, the fruits are tossed out like shuttle-cocks. The bristly pappus-hairs are not in this case organs of flight ; they are short and stiff", and, besides raising the fruits, serve also to determine the direction of their fall. Balistic apparatus very similar to that just described is also found in several Iridacece, Liliacece, CaryophyllaceoB, Prmiu- lacecB, and Scrophulariacece, only in them the erect, resilient stem does not bear a fruit-capitulum but a capsule, and the ejected particles are not fruits but seeds. The seeds are comparatively large and heavy, and are destitute of membranous or hairy appendages. In all these cases the capsule is situated with its orifice upwards and only opens in dry weather. As its cavity is very deep, no ejection of the seeds ensues except when the resilient stalk which carries it sways somewhat violently to and fro. 842 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. The manner in which the fruits of Labiatse are thrown oif is particularly remarkable. The fruits in question are spherical, oval, or ellipsoidal nutlets, and when ripe are still hidden at the bottom of the persistent calyx. The calyx is either bell-shaped or tubular, and faces laterally; the pedicel supporting it is resilient, and usually bowed (see figs. iGO^-^'^'^). If one presses upon the stiff points of the calyx with some hard object such as a piece of wood (fig. 460*) the pedicel is subjected to a strain, and as soon as the pressure ceases it springs back to its former position, and the nutlets are shot out with great force (fig. 460®). The path of projection is in this case determined by the two inferior calyx-teeth, which curve upwards like sledge-rujiners (see fig. 460^). In many instances, as, for example, in Teucrium flavum, T. EugcmoBum, and Monarda fistvZosa (see figs. 460 ^' *• ^> ®' ''■ ^' ^), there is yet another contrivance for ensuring the proper direction of flight. This consists in the presence of stiff though pliable convergent hairs in the calyx-tube, and their function may be compared to that of the grooves in a rifle. Again, in Scutellaria the lobes of the calyx-limb, which is in the form of a tilting helmet with the visor down, determine the path of the seeds after expulsion. The result thus artificially attained by bending down the stalks of the fruiting calyces and letting them fly up again is brought about in nature by gusts of wind, by drops of rain, and most frequently of all by animals brushing against the calyces. In the last-mentioned event one or other of the ejected nutlets may stick to the animal's coat and be carried to a much greater distance than would otherwise be the case. This kind of apparatus rarely occurs in plants other than Labiatse. The nearest analogy is found in the ejection of seeds from the fruits of several species of the Chickweed genus, e.g. Cerastiuw, macrocarpum, (see p. 448, fig. 340 *), where the fruits are curved like the letter S, are borne on stiff stalks and hold the ends that open upwards. One of the most curious forms of mechanism of the catapult variety occurs in the North American Polygonum Virginicum (see figs. 460 1" and 460 "). In this plant the fruits are on short stalks, and are arranged in spikes on long switch-Kke stems. The fruit-stalks are remarkable for the fact that the cells of the cortical parenchyma, which is greatly developed, have their walls strongly lignified, though only slightly thickened. It is also noteworthy that between the stalk and the fruit there is a layer of separation which looks like a joint to the naked eye. The style is transformed into a decurved beak, which is seated upon the fruit, and terminates in two little divergent hooks. When one of these fruits is pushed by a passing animal it is at once detached at the separation-layer and springs away to a distance. The pressure applied to the fruit is apparently transmitted to the short stalk, and gives rise to a condition of tension in the tissue of the stalk analogous to that of a watch-spring. As soon as the pressure ceases the tension relaxes, and the fruit is cast away with great force. For a long time it remained a mystery how these fruits were thrown off in the absence of any animals to give the initial impulse. A few years ago, however, I succeeded in observing the manner in which the long fruiting switches are swayed backwards and forwards by a boisterous wind, CREEPING MECHANISMS. 843 and how they brush against one another and against the branches of neighbour- ing shrubs as they swing, and thus receive the stimulus necessary to cause them to throw oif the fruits. The contact of animals is, however, a more advantageous means of dispersion, inasmuch as the fruits may be left hanging to their coats by the hard styles and the range of distribution be greatly increased thereby. When there is no assistance from animals, and the cast-off fruits simply fall to the ground, the range of projection is not more than 2-3 metres, which is a comparatively small distance from the spot where the fruits were ripened. The limitation of the range of dispersion is still more marked in the case of fruits which creep or hop along the ground than in those where the action is that of a sling or of a catapult. The fruits in question have stiff and very hygroscopic bristles projecting on one side from their external coats, and these bristles continually change their position according to the varying state of the environment in respect Fig. 461.— Creeping and hopping fruits. 1 jSgilops ventricosa. 2 ^gilops ovata. " Crupiim vulgaris. * Trifolimn stellatum of moisture, and by so doing propel the fruit or seeds, as the case may be, in a definite direction. The awns which project from the glumes of Grasses (e.g. Elymus crinitus, Secale fragile, and various species of u^gilo-ps; see figs. 461 ^ and 461 ^), the strong bristles in which the bract-scales of the flowers in Restiaceae terminate {e.g. the South African plant, Hypodiscus aristatus), the calyx-bristles and stiff pappus-hairs in Scabiouses and Composites (e.g. Grwpina vulgaris, see fig. 461 % and the divergent calyx-teeth in Papilionacese (e.g. Trifolium stellatum, see fig. 461 *) constitute structures whereof the different parts alternately approach and recede from one another and so cause a movement resembling that of creeping. In all these cases the hygroscopic structures are furnished with small teeth. Sometimes the teeth are on both sides, sometimes on one side, and sometimes only at the tip (see figs. 462 1-^-^-*). The teeth render retrogression impossible, and to that extent determine the direction in which the fruit moves. In AveTia elatior, Avena pratensis, and several other Grasses the awns which project from the base of the enveloping glumes are bent elbow-wise. The part below the bend is spirally twisted, and as the tissue is extraordinarily hygroscopic, the spiral relaxes or contracts according to the amount of moisture in the air. This spiral motion causes the part of the awn which is above the bend to move like the hand of a watch, but now to one side, now to the other. Of course this movement can only 844 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. take place provided the part of the awn which undergoes it is not fixed down anywhere by an obstacle. If one of the lever-arms of the awn encounters a fixed object on the ground the spiral motion of the lower arm sometimes has the effect of forcing the tip of the awn over the obstacle, so that the whole structure shoots obliquely upwards. This phenomenon is especially conspicuous in Avena sterilis. In this case two glumes furnished with strong bent awns are to be seen on the fruiting spikelet after it has fallen. An alteration in the environment in respect of moisture causes the two awns to twist in opposite directions, so as to cross one another. After pressing one upon the other, they ultimately slip apart with a sudden jerk, which causes the whole fruit to spring up. This movement is much more like hopping or jumping than creeping. ^ The distance traversed by creeping, hopping, and bounding fruits is seldom greater than a few decimetres. The movements generally land the fruits almost Fig. 462.— Fruits wliicii creep or hop along the ground. 1 Awn of j^gilops ventricosa. 2 Xvfn of j^gilops (yvata. 3 Bristles of the pappus of Cmpina vulgaris. * Calyx-tooth of Tr^oHum stellatum; all the figures magnified. immediately in some cul de sac, where they remain, or else the awns gets entangled with the above-ground stems and leaves, and in that case the result of the move- ments is to imbed the seeds, which are concealed in the fruit-scales, in the earth (see vol. i. p. 617). In such cases no doubt the most important function of the movements in question is to fix the plants in the soil, but on the other hand it cannot be denied that a limited form of dispersion may be and is as a fact achieved by these movements. The dispersion of fruits through the agency of water takes place in all plants which undergo fertilization under water and detach their fruits when they are quite ripe. To this class of fruits belong the Fungi of the family Saprolegni- acece, and most of those Cryptogams which are known as Algse. Such facts as have been ascertained by botanists concerning the distribution of the fruits of these plants in the water have been recorded in previous pages (see pp. 49 and 64). This method of dispersion is of less importance in the case of Phanerogams, which are fertilized and ripen their fruits in the medium of the air. At first sight one might '■ The hopping movements of the fruits of the Mexican plant named Sehastiana Pavoniana, and of those of Tamarix Gallim, which belongs to the Mediterranean floral area, are not due to alterations in the tension gf particular parts of the fruit-coat, but are caused by inseot-larvse which live inside the fruits. In the case of the "Mexican Jumping Bean", the larvae are those of the small lepidopterous insect named Carpocapsa saltitaiis, and in Tamarix Gallim those of the beetle JVanodes Tamarisci. « DISPERSAL BY WATER. 845 suppose that rain-water running off the plants and then trickling along the ground would be a very effectual means of dispersing fruits and seeds, but closer observa- tion convinces one that distribution is comparatively seldom effected in this manner, and that wherever such dispersion does occur it is invariably supplemented by some other means of dissemination. The best-known instances are afforded by two plants which, on account of their extraordinary properties, were brought to Europe from the East by pilgrims and crusaders in the Middle Ages. They were called " Eoses of Jericho", and all sorts of marvellous tales were told concerning them. One of these plants is Anastatica Hierochuntica, a Crueifer which grows on the Steppe- lands of Egypt, Arabia, and Syria, and which has the peculiarity that its branches Fig. 463. — Fruits wiiicli open upon being wetted with w iL e 1 Anastatica Sierochunfica, dry. 2 The same when wetted. 3 Fruit of Mesembryanthemum CaTutolleanwm, dry. * The same when wetted. 6 Fruit of MesembryantJtemum annuum, dry. 6 The same after being wetted. curve inwards when the fruits are ripe in such a manner as to form a trellis round the closed, pear-shaped siliquas, which are very numerous and are situated at the extremities of the ramifications, and to protect them from being touched (see fig. 463 ^). The shape of the structure in this condition is something like that of an unopened rose, and it remains unaltered so long as it keeps dry. When moistened the branches at once open back and stretch straight out (see fig. 463 ^). The fruits also open at the same time, and the seeds are then liable to be washed out of the fruit-valves by falling rain. When growing wild Anastatica remains closed during the long drought which follows the maturation of the fruits, and it is not till the winter rains set in that the tangle of branches opens and the seeds are washed out of the fruits. The second "Rose of Jericho", Asieriscus pygmceus, is a small plant of the Composite family, and ranges from the northern portion of the Sahara to Palestine, being met with in especial abundance in the neighbourhood of Jericho. In this case the branches do not close together when the fruits arrive at maturity, but the involucral leaves, which are arranged in a rosette, close up over the capitula 846 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. of fruit and do not open until the winter rains set in and cause the "rose" to unfold, whereupon the fruits are washed away. Similar phenomena in connection with the rainfall are exhibited by the fruits and seeds of the so-called Ice-plants (Mesembryanthemum) which occur in a great variety of forms at the Cape. The capsular fruits of these plants remain closed in dry weather; but the moment they are moistened the valves covering the ventral sutures of the fruit-loculi open back, dehiscence takes place along the ventral sutures, and the seeds, hitherto retained in a double shroud, are washed out of the loculi by the rain (see figs. 463 ''*'^'*). Amongst plants belonging to the flora of Europe, the Yellow Stone-crop {Sedum acre) responds to the influence of rain in a manner which reminds one strongly of the Ice-plants above referred to. The carpels are arranged radially, and are furnished at the base with wing-like borders, whilst the central part of the external surface of the fruit is in the form of a shallow basin, In dry weather the five fruit-loculi are closed; but the moment a drop of rain falls upon the concave centre they open wide, and the next few drops wash out the seeds, which are of small size, and convey them to the ground. As the rain trickles into the tiniest crevices in rocks and walls, the seeds are carried into holes in vertical or even overhanging clifis where it would be scarcely possible for them to be deposited by any other means of dispersion. In the case of Veronica Cyriiba- laria, which grows on walls in the south of Europe, the fruits likewise remain closed so long as the weather is dry and only open when they are thoroughly soaked. The seeds are then carried, like those of the Stone-crop, into the holes and crevices of vertical walls by means of the infiltration of rain. Similarly in the cases of Veronica agrestis and Veronica serpyllifolia, species of Speedwell which grow profusely on cultivated ground, the seeds are washed out of the gaping capsules by rain and conveyed to spots where they find conditions favourable to germination. It is worthy of note that the capsules of Veronica Anagallis, V. Beccabunga, and V. scutellata, species which grow on banks and in running water, also do not open until they are thoroughly wetted by rain. The explanation of this curious fact must be as follows. If the wind were to act as the means of dispersion there would be a risk of the seeds being deposited on dry places where they would be doomed to perish. On the other hand, the rain carries the seeds on to the wet soil of the marsh or into the shallow water of the brook or pond, as the case may be, where the plant in question finds favourable conditions. I must again repeat that actual contrivances with a view to seeds being washed out of open fruits by rain are comparatively rare. This, of course, does not exclude the possibility of fruits or seeds unprovided with such contrivances being dispersed by rain, or by the little tributaries of rivulets, which result from showers of rain if once they are transferred by any means into the channels in question. The rills of water which run swiftly down to join larger streams after a violent fall of rain collect not only sand and earth, but also any seeds that may have been deposited on the ground by the wind, and they subsequently set them down -syith the mud at the edge of the stream. Those fruits and seeds also which fall by chance into running DISPERSAL BY WATER. 847 water, during transportation by aerial currents, may be floated along, and finally deposited by the stream. Numbers of fruits and seeds of the most various kinds of plants are invariably found to have been deposited on the banks of sand by the sides of mountain-torrents, and on the margins of rivers and rivulets after the water has subsided from a state of flood. Many of them, it is true, have no chance of developing, but perish, either because the conditions are unfavourable, or because they have lost their capacity for germination in the transit; others do, however, germinate, and some even thrive luxuriantly. But such seeds can only be said to have been accidentally dispersed by running water, and must not be considered as instances of adaptation to that method of dispersal. The same statement applies generally to the chance deposition of the fruits or seeds of land-plants in the sea. They may be carried away to a great distance by ocean-currents, may float about for months, and finally be stranded on some remote coast. Experiments have frequently been made with a view to ascertain which fruits and seeds retain their power of germination notwithstanding prolonged immersion in salt water. As a result of these experiments it has been established that the seeds of Asparagus officinalis, Hibiscus speciosus, and several other plants do not lose their capacity for germination after immersion in sea- water for a period exceeding a year in duration, a fact which is in itself of great interest. But such results are without significance in relation to the dispersion of fruits and seeds, unless it be also ascertained that the fruits and seeds in question keep afloat upon the surface of the water. For most fruits and seeds sink at once, and sooner or later undergo decomposition at the bottom of the sea. The number of fruits or seeds capable of keeping afloat on the surface for any length of time is extremely small. Of the fruits which are found floating on the sea we may mention first the hard-coated fruits of the group of Palms named Lepidocarynse. They have a smooth, scaly, completely closed envelope which is impermeable to water, and looks very like a coat-of-mail, and, owing to the fact that this envelope is not in im- mediate contact with the fruit, but is separated from it by a layer of air, the fruits are able to float on the surface of the water. The large fruit of the Cocoa-nut Palm also is rendered buoyant by a substantial layer of fibres, which incloses a quantity of air, and is itself coated by a layer with fatty contents which prevents the infiltration of water. If fruits of this kind fall into the sea and are cast up by the waves, the seedlings inclosed in them may develop and become denizens of the shores to which they have drifted, provided the conditions, in respect of climate and soil, are such as to permit it. As a matter of fact, fruits cast up by the sea on to remote islands in the Tropics have been known to develop without any human interference. The phenomena connected with the dispersion of fruits and seeds in still water are altogether peculiar. Currents arising from the slope of the ground do not occur in such water, whilst currents set in motion by the varying temperatures of different layers of water, for the most part, ascend and descend merely, and can occasion very little horizontal displacement of fruits and seeds. The wind is the only agency in 848 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. these circumstances that can supply the propelling force necessary to drive such fruits and seeds as can keep afloat from one shore to another. Special mention must be made of three groups of fruits and seeds belonging to this category. These are, firstly, dry fruits which are rendered buoyant by air-inclosing envelopes, as, for instance, in the case of the marsh-plants known as Sedges {Carex ampullacea, Q. vesicaria, &c.), where the fruit is surrounded by an inflated utricle; secondly, the fruits of Water-Plantains, Flowering-rushes, &c. (Alisma, Bwtomus, Sagittaria, Sparganium, &e.), which are furnished with a thick air-filled cortical parenchyma; and, thirdly, the seeds of some "Water-lilies. In the case of the white Water-liKes 464. — Dispersion of fruits and seeds by the wind. ^ Anthyllis Vulneraria; two fruiting -calyces are falling from the plant. 2 Longitudinal section through a fruiting calyx belonging to the same plant; the pod is visible in the interior, s Trifolium tomentomTn; one head of inflated fruiting calyces is detached, and another is still attached to the stalk. ^ Longitudinal section through a fruiting calyx belonging to the same. ^ sfedicago scutellata. « Ostrya ca/rpinifolia ; branch with two fruit-spikes. ? Longitudinal section through the saccate cupule which envelops the nut in this plant (NymphoBO), each seed is enveloped in a coat (arillus), which loosely clothes the outer integument (testa) of the seed, so as to leave a layer of air between the two. In the species of the genus Nuphar there is no arillus, but the carpels separate when the fruit is ripe into two layers, of which the outer one is green and succulent, whilst the inner one is white and charged with air, and incloses a large number of seeds. In all these cases the seeds are enabled to float by their envelopes, and are driven along on the surface of the water by the wind. In a similar manner the wind causes certain detached as well as aggregated fruits to roll along upon level ground. This phenomenon is observed particularly in regions where a long period of drought follows the short summer season of development; and accordingly the plants concerned are especially abundant in the vicinity of the Mediterranean Sea and in Steppe-lands. Several Umbellifers indigenous to the high DISPERSAL BY WIND. 84& Steppes of the East produce smooth, ellipsoidal fruits about the size of a hazel-nut and so light that if one of them is laid on a person's open hand when his eyes are shut he does not perceive its presence. The extraordinarily small weight of these fruits is due to the fact that their structure includes a layer resembling the pith of the Elder. A fruit of Oachrys alpina measures 13 mm. in length and 10 mm. in thickness and weighs 0'07 grm.; another Oachrys fruit from Shiraz is 15 mm. long and 10 mm. thick and weighs only 0-06 grm. When fruits of this kind fall they are rolled along over the Steppe by the wind and only come to rest when they are caught in some crack in the parched clay soil or get lodged in a hole in a rock. A few Papilionacese also produce rolling fruits of the kind. One of the groups of species belonging to the Medick genus, of which Medicago scutellata (see fig. 464 ^) may be taken as a type, has pods which are spirally curled into round balls and which, when their seeds are ripe, detach themselves from their stalks and are rolled Fig. 465. — Dispersion of fruits and seeds by the wind. Plantago Cretica. a Uttle way along the ground every time there comes a gust of wind. The same thing happens in the case of BlumenbacMa Hieronymi, a native of South America, belonging to the family Loasaceae. Although the spherical fruit of this plant has a diameter of 2-5 cm. it only weighs 0'34 grm. when thoroughly dried. As soon as the seeds are ripe the fruit-stalk withers and the round fruits, which are then left lying loose upon the ground, are rolled away by the gentlest breeze. If their career is stopped anywhere, and they get wetted by rain, the openings which are already formed in them become enlarged and a quantity of wrinkled seeds fall out. Paronychia Kapella (see fig. 468 % a plant of wide distribution in the floral area of the Black Sea, where it grows on dry rocky soil, brings small fruits to maturity in the height of summer, each of which is surrounded by silvery white membranous bracts. "When the season for the dispersion of these fruits arrives the entire tuft of fruits, which is in the form of a spherical glomerule, becomes detached from the branch on which it grows and lies lightly on the ground, where the least puff of wind imparts to it a swift rolling motion. Sometimes if the ground is uneven the rolling is converted into a hopping and springing motion, and occasionally such masses of fruit are raised by powerful gusts of wind and carried considerable distances through the air. In several species of Clover, such as Trifolium globosum, T. subterramum, and T. nidifieum (see fig. 4681") there are only a few perfectly VOL. II. lo* T 850 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. developed flowers in the cluster growing at the end of the flower-stalk, whilst a number of abortive flowers are crowded together in a tuft in the middle of the inflorescence. At the season when the legumes are formed from the fertile flowers the calyx-teeth of the abortive flowers increase in sdze and assume the shape of long hairy bristles, which bend over outwards and form a loose globular inclosure round the head of leguminous fruits. These balls afterwards become detached from the stalk and are rolled away by the wind. Even entire plants are in some cases uprooted or have their stems severed from the roots at the base in the fruiting season, and are then rolled along like balls by the wind. The most remarkable instance is that of Plantago Cretica, which is shown in fig. 465. This is an annual plant possessing an abbreviated main axis from which springs a tuft of stiff", erect flowering stems. When the fruits begin to ripen the stems curve down in coils to the periphery c5f the plant, and by so doing give a strong pull to the abbreviated axis and to the simple tap-root, which is inserted in the earth in a vertical position. The soil on which Plantago Cretica grows being completely dried up in summer is full of cracks, and the pull imparted in the manner described is in consequence sufficient to uproot the plant. The plants now in the fruiting stage have the form of flattened balls and are very light, so that the entire structure is rolled along by gusts of wind. Plantago Cretica is also a type of the so-called "Steppe-witches" and " wind- witches ", which are a source of so much wonder to travellers in the regions of Steppes. On the high table-land of Persia there is a plant named Gundelia Toumefortii which grows in loose, round, prickly sods, and has a tap-root deeply sunk in the earth. When the fruits are ripe the neck of the root rots away and the round sod then rests simply with its stiff lower branches in contact with the ground. Whenever the slightest wind begins to blow innumerable quantities of these sods are set in motion, and are thus dispersed over the plateau. The herbaceous plants of the Steppes of Southern Russia which exhibit the phenomenon of a decay of the bases of the stems in the fruiting season and a consequent liberation of the dry aerial portion of the plant belong to families of the most various kinds. The most common are Alhagi caTnelorum, Gentaurea diffusa, Phlomis herha-venti, Rapistrwrn perenne, and Salsola Kali. It often happens that a number of these dry, branching herbs get hooked and entangled together as they roll along, until at length they form a ball as big as a cartload of hay. Such balls have also been seen lifted up by whirlwinds and driven bounding over the plain. It is not surprising that this marvellous phenomenon has appealed to the imagination of the inhabitants of the Steppes, and has even become a subject for witch-lore whence have arisen the names Wind- witch and Steppe-witch. It only happens in a small proportion of these cases of rolling fruits, wind- witches, and the like, that the seeds are strewn out as they are bowled along; when this does occur it is usually occasioned by some unevenness in the ground which gives a sudden jog to the rolling body. In the majority of cases the seeds do not escape until the fruits are brought to rest by encountering some insurmountable obstacle, the reason being that the seed-vessels only open when they become wet. DISPERSAL BY WIND. 851 This brings us back to the fruits of Mesemhryanthemum and Anastatica, which were described on pp. 845, 846. Sometimes these also play the part of rolling fruits. The capsules of Mesembrianthemum detach themselves from their stalks, the plants of Anastatica become partially uprooted, and lie during the dry season of the year loose upon the earth. A puff of wind blows them into hollows in the ground or cracks in rocks, where they are held prisoners. The seed-cases, however, still remain closed. At last the winter rains set in, whereupon the capsules open, the seeds are washed out, and after a short time they germinate on the saturated ground, to which the rain has conveyed them. Innumerable are the cases of wind-dispersed fruits and seeds which remain floating in the air for a period of more or less duration after severance from the mother-plant, and which have their fall retarded by special contrivances for the purpose. The conformation of fruits and seeds of this category must be such that the air offers great resistance to their fall, and it is important that they should possess as small a weight as possible in relation to their size. It is well known that the spores of Fungi often remain for a long time floating in the air as constituent particles of the dust. Some seeds, too, are so extraordinarily light that they also look simply like dust and are able to remain for a comparatively long period sus- pended in the air. Amongst such dust-like seeds those of Orchids must be men- tioned first. A single seed of Goodyera repens, for instance, weighs only 0-000002 grm. Several other plants, particularly parasites and saprophytes which live in deep beds of humus, possess extremely light seeds, as is shown by the annexed table: — Name of Plant. Weight of Seed in grams. liTame of Plant. Weight of Seed in grams. Stanhopea oculata Monotropa glabra 0-000003 0-000003 0-000004 0-000006 0-000008 0-00001 Semperifivum acuminatum Parnassia palustris Sedwm maximv/m Lepigonv/m Tnarginatwm 0-00002 0-00003 0-00004 0-00007 0-00008 0-0001 uinbilicus srectus CryftiThddQThxo, conopsBfZ .. Spiraea Aruncus Veronica aphylla OTobctnche ioTicLHthcL To enable these seeds to float in the air for as long a time as possible they are more or less flattened, and their centre of gravity is so placed that they always present the broad side to the direction of descent. The same form of adaptation ■occurs in seeds which are shaped like leaflets, scales, or delicate discs. A compressed seed is usually surrounded by an attenuated margin, a membranous border, or a radiating fringe of extremely fine processes, as in FunJcia, Lilium, Tulipa, Fritil- laria, Rhinanthus, Veronica, Lepigonum, Cinchona, Bignonia, Dioscorea, and Heliosperma (see p. 423, figs. 318 *- *- \ and figs. 466 ^- *• *). In some cases the entire pericarp is modified in this manner, as in Hymenocarpus, Mattia, Peltaria, Ptelea, and C/Zmus (see fig. 467_*, and p. 143, fig. 232 ^). Amongst Umbelliferse, Mimosete, Papilionacese, and Cruciferae cases also occur in which the mericarps, the segments of silieulas and lomenta, or the seed-studded valves of ordinary pods and siliquas. 852 THE DISPERSION OF SPECIES BY MEANS OF FEUITS AND SEEDS. according to the particular plant considered, are in the form of scales and leaflets which become detached separately. Instances of this kind are afforded by Artedia sqvMmata, Megacarpcea laoiniata, Mimosa hispidula, JEschynomene glabrata, and Lunaria rediviva (see figs. 467 1- ^' ", and p. 445, fig. 339 S and fig. 466 ^). With these forms may be classed also such fruits and seeds as are furnished with wing-shaped appendages. The wings are either produced from the seed-coat, as in Pines and Firs (see p. 441, fig. 335 % or else arise from the carpels. A single wing, which stands out to one side, is developed in the case of the pods of some tropical Leguminosae (e.g. Secwrida virgata and Gentrolobium robustum; see p. 445, fig. 339 ^), and in the separate parts of the double fruit of the Maple and of the Fig. 466.— Dispersion of fruits and seeds by the wind, 1 SilicLUOse fruit of Liinaria rediviva ; tlie two valves of the fruit have become detached ; seeds are fastened to the inside of each valve. 2 Opened capsule of a BigTwnia from which winged seeds are being carried off by the wind, s Capsule of Heliosperma gimdrifidum after dehiscence; the seeds are being shaken out by the wind. ^A seed of Seliosperma quadrijidum magnified, 6 Capsule of a Dioscorea after dehiscence, the winged seeds being blown away by the wind. Banisterias, belonging to the Malpighiacese (e.g. Acer Monspessulanum- and Banisteria Sinemariensis; see figs. 467 "^ and 467 "). The achenes of Birches and of the Tree of Heaven (e.g. Betula verrucosa and Ailanthus glandulosa; see figs.. 467^ and 467^^) bear two laterally placed wings in each case. The mericarps of many Umbelliferse (e.g. Opoponax Gretica and Laserpitiwm latifolium; see figs. 467^ and 467^^) have wings projecting from the back; the fruits of some Polygonums (e.g. Polygonum dumetorum and P. Sieboldi; see fig. 467^) are furnished with three wings, and those of Triopteris bifurca, one of the Mal- pighiacese, with four wings, of which two are large and two small (fig. 467 ^). In other cases some of the floral -leaves are transformed into wings for the fruit, as, for instance, in Dryobalanops, of the family Dipterocarpese, in which five sepals are in the form of long wings (see fig. 468 % and in Oyrocarpus, of the family Combretacese, in which two of the 4-7 unequal segments of the calyx are similarly adapted (see fig. 467 ^). It is of common occurrence for the fruits to becomes winged DISPERSAL BY WIND. 853 in consequence of the continuous growth after the flower has faded and the ultimate desiccation of persistent bracts, as is seen in the Hop (Humulus Lwpulus), the Eig 467 —Dispersion of fruits and seeds by the wind. ■'TST-.r-r.i^sriTS,. :s==i. *.sfr£» ■.r^.^z. iilridiasguamata. i^ Betula verrucosa. ^> Laserpitium laU/ohum. Oriental Hornbeam (Oarp^r^us Orientalis), and the Lime (Tilia i'rUer^dia^ (see figs. 468 1 and 468 ^). In many cases, as, for instance, in the Tree of Heaven iA^lan- 854 THE DISPEESION OF SPECIES BY MEANS OF FRUITS AND SEEDS. thus), the two wings exhibit a slight spiral twist resembling a propeller; this occasions a peculiar gyratory motion of the fruit as it sails along in the air. Wherever there is only a single wing which projects from one side, the centre of gravity has an eccentric position, and the fruits and seeds of this class spin quickly Fig. 468.— Dispersion of fruits and seeds by the wind. 1 Carpinus Orientalis. 2 TUia intermedia, s Armei'ia tdpina. * Melica altissima. 5 Bryobalanops. * Paronychia Kapella. '^ Briza maxima. ^ Scabiosa graminifolia. ^ Sumulus Lupulus. ^f> Tri/oliumnidificum. as they fall freely through the air. The motion in question is particularly well marked in the half -fruits of the Sycamore and the seeds of Pines. The same object as is attained in the above cases by the development of alate processes is brought about in other plants by the transformation of dry bracts or floral-leaves into light, loose, saccate, or inflated envelopes round the fruits or seeds. DISPERSAL BY WIND. 855 When quite dry, these envelopes are extremely thin and delicate, and sometimes their weight is still further reduced by a portion of the tissue being torn during desiccation, in which case the whole assumes a sieve-like or latticed appearance. The small fruit within the envelope defines the position of the centre of gravity, and consequently determines also the attitude of the structure as a whole that best adapts it to dispersion by the wind. In several Papilionacese, as in Callipeltis cucullata and the yellow -flowered species of Clover (e.g. Trifolium agrarium and T. badium; see figs. 469 L 2. s, 4. 6)_ the dried petals of the corolla are fashioned into an envelope which incloses the small 1-seeded legume, and in several species of Lady's Fingers (e.g. Anthyllis tetraphylla and A. Vulneraria; see figs. 464^ Kg. 469.— Dispersion of fruits and seeds by the wind. Trifoliwrn badiuTn. — i Inflorescence. 2 Same with fruit ripened, s Flower. * Fruit enveloped in the dried petals. 6 Longitudinal section through the fruit in its envelope of petals. — Vertieordia oculata. « Fruit. 1 Longitudinal section through the fruit, s Five " feathers " from the fruit, s, *, s, and s magnified. and 464 2), and some species of Clover of the tribe Vesicastrum (e.g. Trifolium fragifenom and T. tomentosum; see figs. 464' and 464*), the inflated calyx plays the same part. In many Labiates also (e.g. Galaminta, Salvia, Thymus), the calyx is converted into a dry, saccate envelope, which is severed from its stalk by any external stimulus, and then serves as a means of dispersing the ripe nutlets con- tained in it. In the Hop-hornbeam (Ostrya, see figs. 464 ^ and 464 ''), the small nut is enveloped in the sac-like bract; and in many Grasses, as, for instance, Briza maxima and Melica altissima (see figs. 468 * and 468 ^j, the dry glumes constitute a covering to the small fruit which adapts it to dispersion by the wind. One of the commonest devices for keeping fruits and seeds suspended in the air is of the nature of a parachute. This form of mechanism occurs in the shape either of tufts of hairs or of membranous borders. In Willow-herbs (Epilobium; see fig. 472 ^), Asclepiadaceas (e.g. Cynanchum, see fig. 471 ^), and several Bromeliacesa 866 THE DISPERSION OP SPECIES BY MEANS OF FRUITS AND SEEDS. (e.g. nilandsia; see fig. 475 ^) only one pole of the seed is furnished with a tuft of hairs, whilst in Adenium (see fig. 471 ^), belonging to the family Apocynacese, both poles are so provided. In Valerianacese (e.g. Valeriana; see fig. 4713) ^nd i^ Composite (e.g. Seriecio and Taraxacvm; see figs. 471 1.^'^) the tuft of hairs which acts as a parachute springs from the upper extremity of the achene. Sometimes the parachute and the body it keeps in suspension are connected by a slender stalk (e.g. in TiUandsia and Taraxacum); but usually the former is directly sessile on one extremity of the seed or indehiscent fruit as the case may be. In VeHicordia (see figs. 469 «• ''■ ^), of the family Myrtacese, a strange and beautiful parachute is formed by five petals which are in the form of little fans, each composed of ten Fig. 470. — Dispersiori of fniits and seeds by the ■wind. Bombax. s Anemone sylmstris. » Qossypium Barbadense. feather-like lobes, and in some Labiatse, as, for instance, Micromeria nervosa (see fig. 471 '), the radiating, hair-studded segments of the fruiting calyx constitute a similar apparatus. On the other hand, in several other Labiates (e.g. BaUota acetabulosa), in many Plumbaginacese (e.g. Armeria; see fig. 468 *), and in several Dipsacese (e.g. Scabiosa; see fig. 468^) the parachute is developed from the dehcate, dry membranous calyx or from the epicalyx. Nor must reference to the Cape Silver Tree (Leuoadendron argenteum, one of the Proteacese) be omitted. The fruits here are produced in large cones not unlike those of the Stone Pine {Pinus Pvnea) in form and dimensions. Each bract of the ripe cone subtends a fruit consisting of a nut with persistent wiry style and stigma. The 4-lobed perianth also persists as a membranous parachute, its originally free apices haAring become connate above the nut and around the style. Ultimately the original attachment of the perianth below the ovary becomes dissolved, and as the nut falls out of the cone DISPERSAL BY WIND. 857 the style (with nut suspended below) slides out of the hole, around which the perianth-lobes are connate, until its further progress is arrested by the button-like Fig. 471.— Dispersion of fruits and seeds by the wind. 1 Senecio vulgaris. 2 Adeiiium Eonghel. » Valeriana tripteris. * TypJia Schuttlewortkii. 6 JSriophoruTn angustifolium. « Cynamchumfuscaturn. ' Micromeria nervosa. » and » Taraxacum oficitiale. lo Salix JHyrsinites. stigma. The perianth here forms a beautiful parachute, with the nut hanging freely below at the end of a string, like an enterprising balloon-gymnast. From the fruits and seeds equipped with parachutes we pass to those which are embedded in masses of wool or in envelopes of silky hairs, and are thereby enabled 858 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. to remain poised in the air. The hairs arise either from the surface of the seed-coat (testa), as in the Cotton trees (Bombax and Gossypium; see figs. 470 ^ and 470 ^), or else they spring from the base of the seed, as in Poplars and Willows (Populus and Salix; see p. 423, figs. 318 « and 318*; p. 424, fig. 319 and fig. 471 1"). In the Bulrush (Typha; see fig. 471 *) they take their rise from the pedicels of the fruits, and in several Ranunculaceae (e.g. Anemone sylvestris; see fig. 470 2) from the achenes themselves. In other cases they arise from the floral-leaves, as, for instance, in the Cotton-grass (Eriophorum; see fig. 471 =), where the structure which repre- Fig. 472. — Dispersion of fruits and seeds by tlie wind. 1 Melica Balansce. 2 Caiamagrostis Ejoigeios ; nat. size, s The same magnified. * Geurn mohtanum. ' * ^schynanVius specioBus. 6 ^pilobium coUinv/m. ' Cl&vmtis Flammula. sents the perianth is transformed into delicate hairs, and in Trifolium plumosum, where the fruiting calyx is wrapped in wool. In many Grasses the glumes are beset with extremely fine hairs (e.g. Melica and Galamagrostis; see figs. 472 ^\^- ^), in Micropus, of the Compositse, long hairs project from the scales of the involucre and envelop the entire capitulum in a flocculent mass, and in the Venetian Sumach or Wig-plant (Rhus Gotinus) the stalks of abortive flowers are covered with a woolly down, which serves for the dispersion of the fruits,, whose stalks are usually free from wool. Lastly, we have the cases where the fruits or seeds are kept suspended in the air for a more or less prolonged period by means of special hairy tails. Either the seeds are tailed at both ends, as in jEschynanthus (see fig. 472 % one of the Gesneracese, in which the tiny seeds are furnished with two long hairs, one at each end, or else the style lengthens after the flower has faded and becomes converted into a spirally-curved tail, which remains attached to one side of the DISPERSAL BY WIND. 859 achene, and acts like a parachute, as may be seen in Oeum, Atragene, Pulsatilla, aaid Olematis (see figs. 472 * and 472 0- In some Grasses, such as Stipa (see vol. i. p. 619, fig. 1471), an awn is developed in the form of a long feather, which soars above the tightly-closed glumes inclosing the fruit. Several of the fruits and seeds above described are directly exposed to the wind. Owing to the fact that the desiccation of the envelopes and stalks of the fruits at the time of ripening of the seeds renders certain layers of tissue brittle, a moderate wind is sufficient to cause the fall of such fruits, and the same gust that brings about their severance from the plant drives the fruit along in a horizontal direction. The fruit does not fall to the ground until the wind drops, or until its pro- gress is arrested by some obstacle. Many other fruits and seeds detach themselves spontaneously from the mother-plant when they are ripe, but are not directly exposed in consequence to the full shock of the wind. In these we find many contrivances for the purpose of ensuring that the parts to be dispersed shall be brought out from their shelter, and given over to the wind at the proper time. In some tropical Orchids which are epiphytic on the bark of old trees (viz. Aerides, Angroecimi, Sarcanthus, Saccolabium, &c.), the capsular fruits contain, in addition to the small seeds, hair-like cells, with spirally-marked and obliquely-pitted walls (see fig. 473). Vanda teres (see fig. 475 ^) may be taken as a type of this group. The hair-like cells in question are woven together into a sort of felt. They are extremely hygroscopic, and twist and turn about in a curious manner if the slightest change of condition in respect of moisture occurs. When the valves of the capsules move apart under the influence of a dry wind, an active movement is simultaneously initiated in the matted hairs. The felt becomes to a certain extent puffed up, and consequently it squeezes out between the valves of the capsule, and drags the seeds, which are imbedded amongst the hairs, from the interior to the surface of the capsule, where they are liable to be blown away by ,the least breath of wind. This happens, as was said, when a dry wind is blowing. In wet weather the capsules close up, and conceal both hairs and seed once more in their interior. Similar phenomena may be observed in the fruit-capitula of some Com- posites whose fruits are spontaneously detached from the receptacle on ripening. In damp weather the loose achenes lie hidden in the involucral cup, as though at the bottom of a basket, and the hairy pappuses appended to the achenes are clubbed together. When the atmosphere is dry, the involucre, which is composed of hygro- scopic scales, opens, and the pappuses of the fruits within spring apart, and so act as rig. 473.— Seeds of the Orchid Vanda teres, which are moved from the interior to the surface of the capsule by hygroscopic hair-like cells, and are thus exposed to the wind; xlOO. 860 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. levers. The fruits are speedily raised by this means above the edge of the open involucre to a sufficient height to expose them to the wind. In some other Com- posites, such as the Dandelion {Taraxacum), the fruits do not detach themselves spontaneously from the floral receptable when they are ripe. The segments of the involucre close together in wet weather, as do likewise the hairs or plumes of the pappus. In dry weather the involucre opens, whilst the feathers of the pappus diverge so as to assume the form of a parachute, and in that condition offer a com- paratively large surface for the wind to act upon. A moderate gust of wind is now able to lift the fruits, with their expanded parachutes, off the receptacle and carry them away (see fig. 4)71*). If no breath of wind stirs, they remain upon the receptacle; the damp atmosphere of evening causes both parachutes and involucre Fig. 474.— Dispersion of fruits and seeds by the wind. Fruits of a Tliistle (Cirsium nemorale) floating in the air and becoming detached from their parachutes and dropping to the ground whenever they encounter an obstacle in the course of their flight. to close up again, and the process of dispersion is suspended until next day, when the air is dry once more and the sun shining. In Andropogon Ischoemv/m, Avena pratensis, and many other Grasses, the flowering glume has an awn composed of spirally-marked and highly hygiroscopic cells, and bent like an elbow, and this awn undergoes a marked spiral torsion, accompanied by a slight downward flexure whenever the air is dry. The distal arm of the awn is liable to get pressed against objects in the course of these movements, and it then acts as a lever in raising the fruits above the outer glumes. They are then easily blown away by a puff of dry wind. In several Scabiouses, also, the breaking up of the fruit-capitulum, and the raising of the fruits with a view to their dispersion by the wind, are occasioned by a bristling movement on the part of the hygroscopic setse of the calyx. Each fruit- let in the Valerian is surmounted by a pappus of delicate feathery hairs. When the air is damp these feathers are folded together; when it is dry they become unfurled (see fig. 471 8). In this condition of divergence, they present an ample surface to the wind, and the slightest gust detaches the fruits and blows them away. A similar phenomenon occurs in Lryas, and in some other plants; but we cannot now enter into the details of these cases. DISPERSAL BY WIND. 861 In the case of Willow-herbs (Epilobium) and of some Pines (Pinus nigricans, P. sylvestris, &c.) the fruit-valves and fruit-scales which cover the seeds only open back under the influence of the sun's warmth, and when a dry wind is blowing, and the same wind which thus operates on the valves and scales also carries off the seeds the moment they are exposed, they being furnished with wings or tufts of hair with a view to aerial dispersion. The reader must be referred to p. 447 for a description of the manifold effects of a dry wind on the fruits and seeds in question. First, the dry capsules open; secondly, the seeds, hitherto lying in the interior of the fruits, where they are protected against moisture, are shaken out by the swaying to and fro of the elastic fruit-stalks; and thirdly, these seeds are caught up and scattered by the wind. The distance to which fruits and seeds Which are adapted to aerial transit by means of wings, hairy tails, parachutes, inflated envelopes or woolly coverings, as the case may be, are conveyed by the wind depends on the degree of perfection of their mechanism, on the condition of the air in respect of moisture, and on the strength of the current of air by which they are transported. When the weather is calm and sunny, innumerable of the lighter fruits and seeds are carried up to a great height by the ascending currents which are generated in the atmosphere; but they usually descend again after sunset at a little distance from the spot where they were taken up. Such excursions do not conduce so much to a dispersion of plants over large areas as to their deposition on shelves and in crevices of steep walls of rock, where seeds would not otherwise easily acquire a footing. Currents moving in a horizontal direction may, it is true, convey their freight of fruits and seeds over extensive tracts of country, but very exaggerated notions are usually entertained concerning the distances thus attained. Amongst the numerous species of fruits and seeds blown by storms of wind to the tops of the Alps and left upon the snowfields above the glaciers, not a single one derived from distant parts (i.e. from another district) has been found after careful examination of the deposited matter; and from this we may infer that, even on mountains, fruits and seeds are scarcely conveyed any further by a raging wind than when they are blown from one side of a valley to the other. In many plants the wings or parachutes, as the case may be, only remain attached to the seeds or fruits for the period of their journey thi-ough the air. If the winged seed of a Pine gets stranded anywhere the membranous wing drops off, and the seed is then no longer capable of flight. This phenomenon is even more marked in the fruits of Thistles (e.g. Garduus and Girsium; see fig. 474). The achenes, which are comparatively large, are supported by parachutes and float quietly in the air, but the moment one of them strikes against any obstacle the fruit severs itself from the parachute and falls to the ground. There can be little doubt that to this mode of dispersion must be attributed the common occurrence of Thistles at the foot of walls and in hedgerows, inasmuch as the floating fruits are carried against such structures with especial frequency. In other cases the fruit or seed maintains permanently a firm connection with the parachute, and the latter 862 THE DISPEESION OF SPECIES BY MEANS OF FEUITS AND SEEDS. serves to fasten it to some place where the conditions requisite for germination are present. For instance, when the seeds of Tillandsia (see 475 ^) come into contact with the boughs of old trees, as they are blown along in a horizontal direction, they fasten on to the bark where they are able to germinate immediately. Thus the pappus to which the seed owes its buoyancy serves subsequently to anchor it to a substratum favourable to its development. The modes of dispersion of fruits and seeds through the agency of animals are Fig. 475. — Dispersion of fruits and seeds by the wind. ^ Capsule of Vanda teres, from which the seeds have been transferred to the air by means of hygroscopic hairs, and are being blown away. ^ Open capsule of a TiUandsia ; the seeds are being lifted out by the wind by means of their parachutes. If a-seed is blown against the bark of a tree it is anchored there by the hairs of the parachute. almost as varied as the different methods of dissemination by the wind. In many cases such dispersion is brought about by the animals using the fruits and seeds in question for food; the undigested parts are excreted, and any embryos which may have survived the passage through the alimentary canal subsequently germinate. As the fact of this mode of dispersion has been a matter of dispute amongst botanists, and could only be established by experiment, i determined to feed various animals with selected fruits and seeds, and to ascertain first of all whether the embryos preserve their vitality after passing through an animal's intestinal canal. Fruits and seeds belonging to 250 different species of plants were used for the purpose, and DISPERSAL BY ANIMALS. 863 the following birds were fed with them: blackbird, song-thrush, rock-thrush, robin, jackdaw, raven, nutcracker, siskin, goldfinch, serin-finch, titmouse, bullfinch, cross- bill, pigeon, fowl, turkey, and duck; and also the following mammals: marmot, horse, ox, and pig. After each meal the fseces were examined, to ascertain what seeds they contained, and were then laid on a separate bed of earth, and at the same time fruits and seeds of the same plants which had not been used for food were planted in an adjoining bed. It would be out of place to set forth here all the precautions which it was necessary to take in conducting these laborious researches, and I shall confine myself to a statement of the most important results obtained from 520 separate experiments. As regards the mammals subjected to experiment very few words will suffice. A.lmost all the fruits and seeds administered to them, whether they took them voluntarily or unawares mixed with their ordinary food, were destroyed either at once or upon being chewed with the cud. It is true that a few millet-seeds germin- ated from the ox-dung, and must therefore have escaped being crushed during rumination, and that one or two solitary specimens of lentil-seeds and oat-fruits similarly passed uninjured through a horse, whilst Cornus alba, Hippophae rham- noides, Ligustrum vulgare, Malva crispa, Rhaphanus sativus, and Robinia Psewdacacia all germinated after passing through a pig; but the number of the seedlings so obtained was scarcely appreciable as compared with the number of fertile seeds swallowed in the animals' food, and the fruits and seeds of about 60 other species of plants completely lost all power of germination during their passage through the intestines. The birds resolve themselves into three groups in relation to the matter in question. The first group includes those which grind up even the hardest fruits and seeds in their muscular and hard-coated " gastric mills " which are in addition usually filled with small stones and sand. Amongst these, some strip the fruits and seeds when they first lay hold of them, and thereby condemn them to destruction. To this group the following birds of those employed in the experi- ments belong, viz. the turkey, the hen, the pigeon, the cross-bill, the bullfinch, the goldfinch, the siskin, the serin-finch, the nutcracker, the titmouse, and the duck. No seed capable of germination was found under ordinary conditions in the excre- ments of these birds; only when on a few occasions food was forcibly administered to the hen and to ducks, so that their crops must have been overloaded, were a few seeds found to have escaped pulverization, and to still possess the power of develop- ment. The seeds in question belonged to Arenaria serpyllifolia, Papaver Rhceas, Sisymbrium Sophia, Ribes rubrum, Ligustrum vulgare, Fragaria Indica, and other species. Eavens and jackdaws form a second group, in that the stones of the drupes and hard-coated seeds of the berries which they ate passed uninjured through the intestine, whilst soft-coated seeds and fruits were all destroyed. It is worth mentioning in particular that after these birds had been fed with cherries their excrements contained cherry-stones 15 mm. in diamater, every one of which was able to germinate. Of the birds selected for experiment, the blackbird, the song- thrush, the rock-thrush, and the robin belonged to a third group. Of these the 864 THE DISPERSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. blackbird was the least fastidious about its food. It even swallowed the fruits of the Yew without afterwards relieving its crop of the stony seeds, and it never rejected a single fruit that was mixed with its food. The song-thrush refused all dry fruits of 5 mm. diamater or more, even when they were mixed with the finely- chopped meat with which the bird was fed. They also avoided certaia strong- smelling fruits, such as that of the Yarrow. On the other hand, the aromatic fruits of Umbelliferse (e.g. Buplev/rum rotundifoUum and Garum Carvi) were eaten with great avidity. The seeds of the Tobacco-plant, Henbane, and Foxglove mixed with the food were not rejected and caused no ill effects, and no more did the berries of the Deadly Nightshade, which were devoured greedily. On the other hand, a song- thrush sickened after eating berries of Phytolacca. When fleshy fruits with seeds of diameter exceeding 5 mm., such as those of Berberis, Ligustrum, Opuntia, and Viburnum were introduced into the crop, the pulp passed thence into the gizzard, but all the seeds were thrown up. Many seeds, as, for example, those of Lychnis flos-Jovis, were carefully removed from the rest of the food with which they had been mixed. The seeds of fleshy fruits which were greedily devoured were thrown out of the crop if the stones which they inclosed measured as much as 3 mm. The interval of time between ingestion and evacuation was surprisingly short in the birds of the third group. A thrush fed with Ribes petrceum at 8 o'clock in the morning excreted numbers of the seeds after the lapse of three quarters of an hour, and seeds of SaTnhucus nigra were found to have passed through the alimentary canal in half an hour. The majority of seeds took from 1\ to 3 hours to perform the journey. Curiously enough, the small smooth fruits of Myosotis sylvatica and Panicwm diffusum were retained for the longest period. Of the fruits and seeds which passed through the intestine of one or other of these birds, 75 per cent germinated in the case of the blackbird, 85 per cent in the case of the thrush, 88 per cent in the case of the rock -thrush, and 80 per cent in the case of the robin. The germination of fruits and seeds that had undergone ingestion and excretion was usually (i.e. in from 74 to 79 per cent of the cases) tardy as compared with that of similar fruits and seeds which had not been treated in this way but were only germi- nated for the purpose of comparison. Only in the case of a few berries (e.g. Berberis, Ribes, Lonicerd) was the, period of germination hastened. The seeds of such plants as grow on richly-manured soil (e.g. Ama/rcmthus, Polygonum, Urtica) after passing uninjured through a bird's intestine produced stronger seedlings than did those which were cultivated without such preliminaries. From these experiments it is evident that the dispersion of edible fruits through the agency of thrushes and blackbirds is not, as was formerly supposed, an exceptional phenomenon obtaining in the Mistletoe only, but one that may take place in the case of many other plants, and other observations prove that, as a matter of fact, it does take place. Plants possessing fleshy fruits are undoubtedly often disseminated in this manner. The occmrence of such plants as epiphytes upon trees, and also their unexpected appearance on the tops of high rocks and old walls thus receives a natural explanation. DISPERSAL BY ANIMALS. 865 The phenomenon in question also enables us to interpret the meaning of the changes undergone by fleshy fruits at the season when their dispersion becomes, desirable, inasmuch as they serve the purpose of attracting animals, and the same consideration applies to the contrivances whereby animals are discouraged from taking the fruits before they are ripe. Mention has already been made of these latter contrivances on p. 4.44; and as regards the attraction of animals with a view to the dispersion of ripe fruits the following particulars are of especial interest: Fruits and seeds that are still unripe are hidden amongst the leaves of the mother- plant, have a green colour resembling that of the foliage, and are destitute of scent. On ripening the fruits are exposed, the coats of the fruits acquire a conspicuous coloration, and frequently emit a strong scent. In the cases where the seeds alone are dispersed and the pericarps are left behind, as, for instance, in Pceonia Russi, Euonymus verrucosus and Magnolia grandiflora, the capsules or follicles burst- open, and the seeds are of a bright red or yellow colour, sometimes flecked with steel- blue and black, which renders them visible from afar. In the above-named species of Euonymus and Magnolia they emerge from the pericarps and hang at the ends of threads which renders them even more conspicuous. The particular colour assumed by fruits and seeds at the time of maturity varies according to that of the foliage by which they are surrounded. The difierent tones of red stand out best from a green environment; therefore, for plants with evergreen foliage (e.g. Ardisia, Oaulteria, Ilex, Taxus, Arbutus Unedo, Arctostaphylos uva-v/rsi, Vacciniu^n, Vitis-Idcea) a red coloration is the most advantageous. Also in the case of plants- with foliage which, although not evergreen, does not acquire an autumnal tint at the season when the fruits are ripe, e.g. the Strawberry, the Easpberry, the Currant,, the Wild Cherry, and the Ked-berried Elder (Sambuous Ebulus) the red hue of the fruits is of great value. On the other hand, red fruits would stand out but little against a background of foliage that had already donned the red or yellow tints of autumn by the time they ripened, and accordingly the fruits of Ampelopsis hederacea, Cornus sanguinea, Prunus Padus, Arctostaphylos alpina, VacciniuTn Myrtillus and V. uliginosumn, &c., are, as a fact, blue or black. Sometimes the. fruits are black and the fruit-stalks red, as in Sambucus nigra, or the fruits are only coloured on the side exposed to view, as in the Apple and the Pear. The fruits of the Quince and the Pine-apple are set off by their yellow colour from the blue-green foliage. White berries, such as those of Cornus alba and Symphoricarpus, occur principally in plants which cast their leaves before the fruit is ripe. Standing out against the brown or gray background formed by the leafless branches and the fallen leaves of late autumn these white fruits are clearly visible. The extent to which fruits are advertised by their scents is a matter of common experience, and we need only refer for illustration to the Strawberry, the Easpberry, the Quince, and the Pine-apple. Seeing that the seeds and stones containing seeds of the fleshy fruits eaten by thrushes and blackbirds only remain a short time in the crop and intestine of the bird, it is probable that the plants, in question are disseminated by this agency to VOL. II. los 866 THE mSPEKSION OF SPECIES BY MEANS OF FRUITS AND SEEDS. the distance of a few leagues at most, in the course of a single year, and that it takes many years to distribute them, step by step, as it were, over large areas. We may reasonably suppose that distribution is effected principally in the direction of those parts of the world towards which thrushes and blackbirds are in the habit of journeying