Art. XX.—Studies in the New Zealand Species of the Genus Lycopodium: Part III—The Plasticity of the Species.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 51, 1919, Page 161

Art. XX.—Studies in the New Zealand Species of the Genus Lycopodium: Part III—The Plasticity of the Species. By the Rev. J. E. Holloway, D.Sc. [Read before the Philosophical Institute of Canterbury, 18th December, 1918; received by Editor, 31st December, 1918; issued separately, 20th June, 1919.] Plates IX-XIV. Contents. Page Classification of the New Zealand Species 161 General 161 Species belonging to the Sections Selago and Phlegmaria 165    Habit and External Form of Plant, and Nature of Strobilus 165    Stem-anatomy 170    Prothallus, Sexual Organs, and the Young Plant 175 Species belonging to the Sections Inundata and Cernua 177    Section Inundata 179       External Form of Plant, and Nature of Strobilus 179       Stem-anatomy 179       The Young Plant 180    Section Cernua 180       External Form of Plant, and Nature of Strobilus 181       Stem-anatomy 182       The Prothallus 185       The Young Plant 189 Species belonging to the Clavata Section 190    External Form of Plant and Nature of Strobilus 191    Stem-anatomy 194    The Prothallus 198    The Young Plant 203       Heterophylly 203       The Plagiotropic Habit 204       The “Foot” 207 Summary 207 Postcript 215 Classification Of the New Zealand Species. Species belonging to the sections Selago and Phlegmaria— Lycopodium Selago Linn. L. Billardieri Spring. L. varium R. Br. L. Billardieri var. gracile T. Kirk. Species belonging to the sections Inundata and Cernua— Lycopodium Drummondii Spring. L. ramulosum T. Kirk. L. laterale R. Br. L. cernuum Linn. Species belonging to the section Clavata— Lycopodium volubile Forst. L. densum Labill. L. fastigiatum R. Br. L. scariosum Forst. General In the present paper I have gathered together all the facts which I have observed with regard to the general biology of the eleven species of Lycopodium which occur in New Zealand. The New Zealand biological region provides an excellent field for the observation of the remarkable plasticity of the members of this genus. L. Cockayne (10, pp. 2–3) has pointed out how peculiarly fitted New Zealand

is for ecological studies. In the particular paper cited he says, “Its vegetation is still in many places absolutely virgin; its climate varies from subtropical to subantarctic; some parts experience an annual rainfall of more than 500 cm. and other parts less than 30 cm.; the plant formations include mangrove swamp, rain-forest, heaths of various kinds, subglacial fell- and herb-fields, varied associations of rock and debris, subantarctic southern-beech forest, associations in and near hot springs, dunes, salt meadows, steppes, swamps, and moors—in fact, for an equal variety an ecologist would have to explore one of the larger continents in its entirety Further, the isolation of the region for a vast period of time far from any other land surface; the absence of grazing-animals, the moa (Dinornis) excepted; the diverse floral elements (Malayan, Australian, Subantarctic, &c.); the strong endemism; the numerous small islands where conditions are simpler than on the larger ones; and, finally, the presence of many areas whose vegetation has been changed within a very few years through the farming operations of the settler … all these attributes much enhance the importance of New Zealand for ecological research.” Some or other of the species of Lycopodium are found in abundance in practically every part of this very varied region, and their range of variability is remarkable. Three species belong to the subgenus Urostachya. Of these L. Selago Linn. occurs on the mountain-ranges of both Islands, but more particularly on those of the South Island, both in exposed and in shady situations. L. varium R. Br. a terrestrial species which, as will be shown, has affinities both with L. Selago and with L. Billardieri, is also to be found throughout New Zealand in open places in high country, and some noteworthy varieties occur on the subantarctic islands. L. Bil-lardieri Spring is a characteristic epiphyte, and occurs commonly throughout New Zealand in the mixed forest, and more particularly in the western botanical districts of both Islands where the rainfall is heavy. A small delicate variety of the last species, named L. Billardieri var. gracle T. Kirk, grows on tree-fern stems, and is especially abundant in the western districts of the South Island. The remaining eight species belong to the subgenus Rhopalostachya. L. Drummondii Spring, L. laterale R. Br., and L. ramulosum T. Kirk occur exclusively in swampy or peaty ground, the first-named in one restricted locality in the far north, and the other two more widely—L. laterale in the North Island, and L. ramulosum in the western botanical districts of the South Island and in Stewart Island. L. cernuum Linn. until lately has been known only from the northern half of the North Island, where it grows abundantly in Leptospermum scrub and in the neighbourhood of hot springs, but it has also recently been found in a very isolated locality at the northern extremity of the west coast of the South Island. These four species belong to the Inundata and Cernua sections of the genus. The four species L. densum Labill., L. volubile Forst., L. fastigiatum R. Br., and L. scariosum Forst. which belong to the Clavata section are found typically on open moors or amongst light scrub both at sea-level and at higher altitudes. Of these, L. densum is more common in the northern portion of the North Island, where it grows abundantly on clay lands, whereas L. fastigiatum and L. scariosum are to be found most abundantly in the mountain regions of the South Island. The two latter, with L. volubile, also grow luxuriantly on the west coast of the same Island in localities where the original vegetation has been disturbed by man. L. volubile occurs commonly throughout New Zealand in very varied

habitats, and more especially in heavy scrub or at the edge of the forest where it can climb. It will be seen that the eleven species mentioned cover all of Pritzel's five main sections of the genus. The study of these species as they occur in New Zealand under the manifold varieties of climate, soil, and altitude should present favourable data for the study of the plasticity of the genus as a whole. In many parts of New Zealand the vegetation is as yet quite virgin, but in many other localities the removal of the original forest by felling and burning, or the disturbance of the soil by alluvial gold-mining, road and railway cuttings, &c., has thrown open new ground for the various species of Lycopodium to occupy, and the rapid and luxuriant spreading of such species as L. cernuum, L. ramulosum, L. laterale, and also L. volubile, L. fastigiatum, and L. scariosum, over this new ground through sporegermination has provided excellent opportunities for the study of their prothalli and “seedling” forms. An examination of Baker's, and more especially of Pritzel's, classification of the species of the entire genus brings into view the wide question of the relation of elementary species to the species of taxonomy. The main sections of the genus are well defined, and are in accord with the main characters of both the gametophyte and the sporophyte generations in the life-history of the species. In each of these sections certain well-defined type species are to be recognized, and around each of these type species are grouped a number of variant forms, having each a more or less limited distribution, to which specific names have been given. Many of these latter species will be distinct and true-breeding forms, but probably there will be found to be instances where a form which bears specific rank and which is restricted to some particular country or other will prove to be taxonomically identical with some other form which may not have been distinguished by a specific name and which belongs to quite another biological region, the two forms having arisen quite independently either as epharmonic or non-epharmonic adaptations. In his Flora Antarctica Hooker says, “The importance of the question whether two perfectly similar plants from remote quarters of the globe are considered as belonging to one species has induced me to canvass very fully the claims of many supposed forms of Lycopodium to the title of distinct species. In all such cases my first object has been to determine whether the plant inhabits various intermediate countries. When … they are found to do so there need be little hesitation in referring them, after due examination, to one plant; in such instances the supposition of a double creation of the same species, or of one of them being a variety of some other really distinct plant, which plant wholly resembles another from other countries would be confessedly a gratuitous assumption. Where, however, no intermediate stations can be detected these suppositions become more plausible” (18, pp. 115–17). The genus Lycopodium has a remarkable distribution, some or other of the species occurring in every country, in practically all soils, and at every altitude. The persistence of such a very ancient Pteridophytic family to the present day, and its wide distribution, is probably due not to the different sections of the genus being representative of different parts of the old Carboniferous Lycopods, but rather to the extreme plasticity of the modern genus as a whole. This plasticity of both the gametophytic and sporophytic characters is clearly indicated by the study of the species which occur in New Zealand, and probably also of those in any other region. To quote L. Cockayne again (10, p. 13), “Nothing has been brought out

more clearly by ecological studies in New Zealand than the extreme plasticity of many species and structures, and their rapid response to a change of environment. This is so great in numerous instances that the idea of ‘normal’ loses its meaning.” With regard to the lycopodiums the experimental alteration of the environment will be a very difficult task to attempt, owing to the fact that the spores undoubtedly need very particular conditions for germination, and that the prothallus in many cases takes many years to grow. Therefore such a paper as the present, which contains no facts concerning the experimental cultivation of the prothalli and plants, cannot decide whether or not the varieties of the species described are actually true-breeding forms, although in some instances field observations can indicate with some certainty whether or not a particular variety is to be regarded as a hereditary polymorph. However, the details of field observations and of the general biology of the species will be of use, it is hoped, in indicating the degree of relationship of the species inter se, as also of the different sections of the genus to which they belong. The five main characters in the Lycopodium plant which lend themselves well to a study of its variability are (1) habit of growth and external form, (2) stem-anatomy, (3) nature of fertile region, (4) form and structure of prothallus, and (5) form of young plantlet. A natural classification of the genus will necessarily have to take account of all these characters, and must not be based upon any particular one. Professor F. O. Bower's (4) remarks on this subject may here be quoted: “The true basis for a natural system of classification is not one or two characters arbitrarily selected, but as many characters as possible. These characters will be found to vary as we pass from type to type, and the question will arise as to the relative ages of the extremes in these characters.” On the whole, the five main sections of the genus are each found to show very consistently certain characteristic types of structure in each of the five particulars just enumerated. This would seem to indicate not only that these sections are natural ones, but also that the characters themselves are closely interdependent. The stem-anatomy corresponds with the habit of the plant, as also does the character of the fertile region; and the type of prothallus, along with the form of the sexual organs and of the young plant, varies with the habitat. The genus as a whole being so widely distributed, and the species possessing such noteworthy powers of being able to spread themselves in new situations, it is not surprising to find that all the main characters of the plant are in a highly plastic condition, and are by no means so fixed in form as would be expected if the different sections of the genus had been separate from a very ancient geological age. There will thus be instances in which a species, while showing most of the typical characteristics of the section to which it belongs, possesses also in some one character or other a well-marked peculiarity which will be interpreted either as an extreme adaptation or as an instance of the retention of a phylogenetic character. The New Zealand species afford many instances of interesting variations, several individual species showing a remarkable range of variability in the form of such important organs as the prothallus and the strobilus. Besides the five main characters of the plant which are considered in this paper there are others, such as the form of the sporophyll, the development and form of the sporangium, the presence of a mucilage-canal in the leaf, and the development of bulbils on the stem, which will doubtless be of importance in indicating in conjunction with the main characters of the

plant the natural position in the genus to be assigned to any particular species. The facts brought forward in this paper concerning the varieties of the New Zealand species and their range of variability will thus have a double significance: they will provide new material for the study of the question as to whether or not epharmonic adaptations ever do become fixed and hereditary; and they will help to indicate the natural relationships which exist between the different sections of the genus, and between the individual species which are included within those sections. Species belonging to the Sections Selago and Phlegmaria. Habit and External Form of Plant, and Nature of Strobilus. It has long been recognized that as regards habit and external form these two sections are closely connected. The type species L. Selago and L. Phlegmaria are, of course, very distinct, the former being a short, upright, little-branched terrestrial form in which there is no differentiation between fertile and sterile leaves, the fertile zones alternating up and down the stems with sterile zones, and the latter being a much-elongated, pendulous, much-branched epiphyte in which there is a special sporophyll formation, the sporophylls being characteristically confined to the ends of the branches and the latter appearing as narrow whipcord-shaped strobili. But between these two extreme forms there are numerous species illustrating every grade of transition. The species which are grouped by both Baker and Pritzel under the heading Subselago possess the sporangia aggregated into quite easily recognizable terminal spikes which approach the Phlegmaria condition, although the transition from sterile to fertile leaves and from sterile to fertile regions of the stem in these species is very gradual. Also, included in both subsections Euselago and Subselago of Pritzel's Selago section are pendulous epiphytes and upright terrestrial forms, although on the whole there is seen to be a gradual transition in the section from the erect terrestrial form to the hanging epiphyte, this transition keeping pace with the transition in the fertile region from the Selago condition to the Phlegmaria condition. Again, in the Phlegmaria section are grouped forms some of which are robust and upright in growth, and others tender and pendulous; and of these the former possess short thick strobili in which the sporophylls show a more or less gradual transition from the sterile leaves, and the latter possess in some cases very distinct long whipcord-like strobili, while in others the fertile leaves may be not different from the sterile leaves. The New Zealand species which belong to these two sections present some very interesting transitions of this nature, which will be instanced. It will be at once apparent that a classification which is based upon one character alone—as, e.g., the nature of the fertile region—is bound to be unsatisfactory. The general habit of the Lycopodium plant and its external form must be considered along with the nature of the fertile region, for these three characters are closely interdependent. That these two sections together constitute a natural division of the genus seems to be suggested by the fact that they show certain common growth-features which are markedly absent from the remaining sections. In habit they are consistently orthotropic, so that their extent of growth is strictly limited. As a result of this orthotropic manner of growth it is found that throughout the two sections the roots are confined to the basal region of the plant, as many as six or seven roots appearing in the cortex of the stem in a transverse section

taken at its base. Again, branching of the stem is dichotomous and in no case monopodial. Now, the study of the distribution of the species of these two sections as given by Baker or by Pritzel shows clearly that the terrestrial forms occur, on the whole, in colder regions and the epiphytic forms in warmer regions. It would seem, then, that the evolution of the subgenus Urostachya has been determined largely by climatic conditions. That this is so is seen to be more probable still from the fact that the external form of individual species is in a highly plastic condition, varying greatly according to the habitat. Whether we are to regard the Selago form as having been derived from the Phlegmaria form, or vice versa, must at present remain an open question. It will be decided only by a comparative examination of all the other Lycopodium plant structures, for it involves the question whether the genus as a whole is to be read as a reduction or as a progression series. Especially must it be considered in the light of fossil evidence derived from later geological ages than the Carboniferous. In the meantime, considering that the more complex members of the genus are those which show the greatest adaptation to the environment, the onus of proof must lie upon those who would trace in the genus a general reduction in form rather than a progression. In the New Zealand biological region L. varium shows a remarkable variation in form according to locality.* See Postonpt, p. 215. In Plate IX, fig. 1, are illustrated three varieties. That named A is a plant which I have collected from the lower end of the Otira Gorge, where it grows abundantly in clumps on rocks and other exposed terrestrial positions. The Otira Gorge is situated on the western side of the Southern Alps, where the climate is exceedingly moist, at a height of about 1,500 ft. In Part I of these studies (16, pp. 254, 290) I have described the same form of this species as it occurs in enormous clumps on the floor of the excessively wet forest of Stewart Island. The plants are upright in habit, and the strobili characteristically curved. This habit is very similar to that of L. strictum Baker of the mountains of Madagascar, which is figured in Engler and Prantl (13, fig. 375). The plant of this particular form A is obviously akin to the typical form of L. Billardieri, differing from it in the rigid upright habit, the smaller size, and the short curved strobili, which are not quite so distinct as and are stouter than those of L. Billardieri. C on the same plate illustrates the same species as it occurs on the Dun Mountain, Nelson, in a drier, more exposed situation, and at a height of 4,000 ft. In this case the strobili in their tetragonous region are only from ½ in. to 1 in. in length, but the sporangia are continued still farther down the branches in the axils of gradually lengthening leaves. B shows two plants of this species as it occurs on the meadows of the Antipodes Islands.† I am indebted to Dr. L. Cockayne for specimens of this form. Spirit specimens of L. varium from Campbell Island were kindly supplied me by Dr. Charles Chilton from material in the Canterbury College Laboratory. For the herbarium material of the Macquarie Island species, and also for that of most of the species in my collection from localities outside the New Zealand biological region, I am greatly indebted to Mr. T. F. Cheeseman. In this particular variety the plants which occur in clumps stand only about 5 in. or 6 in. in height. They are very sparingly branched, and the fertile tips of the branches are very short and grade into the vegetative regions, as in the case of the form from the Dun Mountain. As well as these three forms of L. varium I have before me material of the same species as it occurs on Campbell Island, which was gathered by the Subantarctic Scientific Expedition of November, 1907. These plants are

only about 6 in. in height, and are very robust in form, with large leaves. None of the Campbell Island plants which I have seen show fertile regions, so that I suspect that they are not full-grown plants. They will, however, undoubtedly be identical with the plant from Campbell and Auckland Islands described by J. D. Hooker (18). He speaks of it as follows: “L. varium, in Lord Auckland's Group and Campbell's Island, is one of the finest of the genus; it grows nearly erect on the bare ground to a height of 1 ft. to 2 ft., branching upwards, copiously leafy, with large spreading leaves, bearing at the apices of the branches numerous pendulous or drooping tetragonous spikes 2 in. to 4 in. long. The stems of this species are often nearly the thickness of a swan's quill, with spreading leaves as broad as the middle finger. I have nowhere seen handsomer specimens of it than this island presents, and more constant ones, for it is confined to the woods, and does not ascend the hills, neither varying in the narrow belt it inhabits nor seeking other localities where it would be exposed to the influence of exciting causes.” There is also an interesting form from the hilltops of Macquarie Island gathered by Mr. H. Hamilton, and identified by Mr. T. F. Cheeseman as L. varium. This is shown in the present paper in Plate IX, fig. 2, B. In a letter to the writer with regard to this plant Mr. Cheeseman says, “I have for the present referred the Macquarie Island plants to L. varium because a few of the specimens have the branches narrowed towards their apices, with smaller leaves, thus approaching the spicate character of a true varium. It also differs from typical Selago in the larger, broader, and more coriaceous leaves. Still there can be no doubt that it comes very close indeed to Selago.” This plant stands from 4 in. to 6 in. in height, and is copiously supplied with bulbils, the latter developing in profusion while still attached to the stem. The presence of bulbils would seem to be an argument in favour of relating this plant to L. Selago. From these varieties it will be seen that L. varium stands midway between L. Selago and L. Billardieri, as Hooker pointed out. The external form of the plant approaches that of L. Selago in those localities in which the plant is exposed to a more rigorous climate, and on the other hand approximates to L. Billardieri in those varieties which occur in less exposed and shady situations. The character of the fertile region also varies along with the same change in the environment, the strobilus being less differentiated from the lower sterile portion of the stem in those varieties which are found in exposed positions. It will be convenient next to consider the forms of L. Selago as they occur in New Zealand. In Plate IX, fig. 2, A, are shown two plants of L. Selago which were collected by me in damp beech forest at Lake Rotoiti, Nelson. This is a drawn-out, largely unbranched, rather straggling form, green in colour, with comparatively large spreading leaves, and is very similar in appearance to that which is figured in the frontispiece of Professor F. O. Bower's The Origin of a Land Flora. The sporangia are very apparent in numerous fertile zones over almost the entire length of the stems, but the number of bulbils borne on the stems is quite small. C (Plate IX, fig. 2) is this species as it occurs in open tussock country on the hills around Cass, in western Canterbury, at an altitude of 1,500 ft. and over. This is a very common form in all such situations. It is much shorter and more rigid than the shade form described above, and is more profusely forked. The leaves are short, ascending, and densely crowded, giving a cylindric appearance to the whole stem. In the lower parts of the stem, however, they are larger

and spreading, and in those regions the appearance of the stem corresponds more with the form A. In form C, however, the leaves are generally more or less reddish in colour, the particular plant figured being a bright golden red. The number of bulbils present is much greater than in the case of the forest form, but is not nearly so great as in the case of the Macquarie Island plant (Plate IX, fig. 2, B). There is a tendency for the sporangia to be confined to fertile zones in the upper half only of the stems where the leaves are of the short form. The tips of the branches where the sporangia are full assume quite markedly a special strobilar appearance, as will be seen by an examination of C, but this is only a pseudo-strobilar formation. However, it is instructive to compare the tendency, as seen in this form of the species, for the sporangia to be confined to the upper half of the stem with what Cheeseman says (as quoted above) with regard to the Macquarie Island plant. On the summit of Browning Pass, on the Southern Alps, at an altitude of 5,000 ft., I collected specimens of L. Selago which showed two very distinct forms. These are figured in Plate X, fig. 1, A, B. B corresponds very closely with the tussock-country form just described; but A, while it is similar to the former in its much-branched nature and short form, is yet very distinct. It is more flaccid in growth, and the leaves are large, green, and spreading, as in the forest form of the species. The upper parts only of the stem are fertile, and there is a very scanty development of bulbils. Now, the form figured A grew among grass in a small cavity in the ground partly sheltered by rocks, whereas that marked B grew in a more exposed position on the surface of the ground, but only a couple of feet distant from the other. This species must be in a very plastic state to be sensitive to such a small change in the environment. To pass now to L. Billardieri. When growing as an epiphyte it presents a very constant form. The plants occur in clumps and are pendulous, being in extreme cases as much as 4 ft. or 5 ft. in length. They are abundantly branched in all regions of the stems, so that the whole clump is quite bulky in appearance. In the lower parts of the stem the leaves are large and spreading, but there is a progressive diminution in their size in the ultimate branches until the strobih are reached. The latter, however, are quite distinct from the adjacent sterile regions of the branches, by reason of the fact that the sporophylls are at once broadly ovate in shape, with a well-marked keel, and are closely imbricating, and are also consistently arranged in four orthostichies. The phyllotaxy of the strobili, joined with the presence of the keel on the sporophylls, gives the strobilus a very distinct tetragonous form. In Plate X, fig. 2, A, is shown a much-forked fertile branch, alongside of which is the upper sterile part of the stem to which it was immediately attached. C in the same figure is the fertile portion of a branch of L. Phlegmaria Linn. from Fiji, from which it can be seen that the differentiation between fertile and sterile regions is more distinct than in the case of L. Billardieri. A specimen of L. Phlegmaria from the New Hebrides Islands which I have also in my collection shows the same very sharp distinction between sterile and fertile regions, although in this case the sterile leaves are tenderer and the strobili are markedly finer and more thread-like than in the Fiji form. Also in both these tropical forms of L. Phlegmaria the sporophyll keel is much less developed than in L. Billiardieri, so that the strobilus is cylindric rather than tetragonous in form. In the extreme north of New Zealand I have frequently found L. Billardieri growing terrestrially amongst Leptospermum scrub on the kauri-gum lands. These plants are from 2 ft. to 3 ft. in height, and the

Fig. 1.—A, L. varium Complete fertile plant from Otire Gorge, Westland. B, L. varium var. polaris Two fertile plants from Antipodes Island. C, L. varium: Uppei portion of fertile plant from Dun Mountain, Nelson. Fig. 2.—A, L. Selago Two fertile plants of mesophytic variety from Lake Rotoiti, Nelson. B, L. ? varuim: Macquarie Island variety. C, L. Selago One complete fertile plant of xerophytic variety from Cass, Canterbury.

Fig. 1.—L. Selago: Variety from Browning Pass, Southern Alps, showing two forms. Fig. 2.—A, L. Billardiei Strobili (typical form) B, L. Billardieri var. gracile. Two complete fertile plants. C, L Phlegmaria: Strobili, from Figi.

strobili are shorter than in the epiphytic form and are curved over. On the whole, I should judge that in form they come nearer to the typical L. Billardieri than to L. varium—even to that form of the latter which I have above described from Otira Gorge and from Stewart Island. Possibly we are to explain the existence of this terrestrial variety of L. Billardieri by the fact that the original epiphytic plants were forced to accommodate themselves to a terrestrial habit through the destruction of the old kauri forest which took place probably a century or so ago. At any rate, they represent a variety not far removed from L. Billardieri, and thus make almost perfect the chain of forms which can be traced from L. Billardieri, and, indeed, from the type species L. Phlegmaria, through L. varium to L. Selago. Perhaps the most interesting variety of L. Billardieri is the form which is known as L. Billardieri var.gracile. This I have found growing very commonly on the trunks of the tree-fern Dicksonia squarrosa in the coastal bush in the Western Botanical District of the South Island. This was first described by T. Kirk (23, pp. 376–77), and is there illustrated. In the present paper two plants are shown in Plate X, fig. 2, B. This variety is a very graceful, slender, flaccid plant, and always quite distinct from L. Billardieri. It is seldom more than one foot in length, and is sparingly branched. The fertile leaves show a wide range of variability in form. In some specimens they are in no wise different from the sterile leaves, although it must be noticed, as we saw in L. Selago, that the leaves are always largest towards the base of the stem. In others the fertile leaves are more bract-like, and the fertile regions then approach somewhat nearer to the typical L. Billardieri form. The two plants figured show well this varying character of the fertile regions, in the case of the plant on the right the various forms of the sporophylls occurring in intermixed zones on the same branch. The whole of the upper region of the stem is fertile, there being no intermixture of fertile and sterile zones as in typical L. Selago, but the fertile region generally extends to half-way or more down the plant. In the figure the limit of the fertile region on the two plants is indicated by crosses. Sir Joseph Hooker states both in his Flora Tasmaniae (20, pp. 155–56) and in his Flora Novae-Zelandiae (19, pp. 52–53) that L. varium, which in its ordinary state is very distinct, passes into L. Selago on the one hand, and, on the other, when it inhabits warmer latitudes, grows dependent from trees, is much branched, more slender and flaccid, and becomes L. Billardieri. In his Flora Antarctica (18) he expands more fully this view of a chain of forms uniting L. Selago through L. varium with L. Billardieri, drawing his illustrations largely from the varieties of L. varium as they occur in Tasmania. In the same place he also says that “the variations from it [i.e., L. varium] to Phlegmaria are not obscure, the variations of that plant being excessive.” In his paper cited above T. Kirk discusses the relation of the New Zealand forms L. varium, L. Billardieri, and L. Billardieri var. gracile to one another, and concludes, “I am compelled to consider L. Billardieri as merely one of the varieties of L. varium.” He proposes the following arrangement of the principal forms: L. varium Br.— (a) varium (the ordinary New Zealand form), (b) polaris (Campbell and Auckland Island form), (c) Billardieri, and (d) gracile. This arrangement certainly has the advantage of emphasizing the natural steps in the evolution of the forms concerned, although such a form as L. Billardieri deserves specific rank. It places the form gracile also in the right position

as a variation from varium rather than from Billardieri. The variety gracile is always quite distinct from L. Billardieri, and it would be a more startling reversion from that species, with its very distinct strobili, than it would be from L. varium, which, as has been emphasized above, is not so far removed from the Selago condition, unless, indeed, we are to take the view that the whole genus is to be read as a reduction series. It may be mentioned here that W. Colenso described the form now known as L. Billardieri var: gracile under the name L. novae-zealandicum (11, p. 275), noting that it grows as an epiphyte on fern-trees. He describes it as “a small species of the Selago section, apparently pretty closely allied to L. taxifolium Sw.,” and he speaks of L. varium as being its nearest New Zealand congener. Stem-anatomy. I will now pass on to consider the stem-anatomy of the mature plant of the New Zealand species of the sections Selago and Phlegmaria. Here also a high degree of plasticity will be apparent, although the entire subgenus Urostachya, so far as it has been investigated, can be seen to be characterized by a consistent type of both stelar and cortical anatomy which is markedly distinct from the type of stem-anatomy of any of the other sections of the genus. It will be necessary to make first some general statements. Owing to the plant being orthotropic in growth, of very limited size, and also more or less branched, it will be clear that in comparing the anatomy of these species care must be taken to section the different stems in corresponding parts of the plant. At the extreme base of the stem the stele has not attained its full development, and is also more disturbed by the giving-off of the roots than it is above. The typical form of the stele is best seen a little higher up the stem, though still in its lower half. Again, as is well known, the general configuration of the xylem and phloem groups and plates is by no means stable, even in a short length of the stem, so that it therefore becomes necessary to compare a number of sections of the stem of any individual plant in order to arrive at a fair idea as to its particular stelar anatomy. I hope to develop in this paper the conclusion, which I arrived at in a previous paper (16, p. 302), that, notwithstanding the fact that the genus Lycopodium includes several distinct types of stelar structure and also manifold variations of those types, yet the lines upon which the Lycopodium stele has evolved can be recognized, having been determined for each natural division of the genus by a combination of the inherited constitution with the acquired harbit of growth of that division. Each of these natural divisions shows a very distinct type of stelar and of cortical structure, so that this character becomes of value in checking the conclusions arrived at from the study of the other characters as to which section a particular species should be referred to. The sections Selago and Phlegmaria possess a stelar structure of the stellate or radial type, although at the extreme base of a plant this is sometimes obscured. The protoxylem groups are very massive and the protophloem very small, and except in the largest forms there is comparatively little wide-sized metaxylem formed. The phloem also shows very little differentiation of conspicuous sieve tubes in the smaller species. Fig. 1 is a drawing of the stele of the Macquarie Island plant shown in Plate IX, fig. 2, B, which has been identified by Cheeseman as L. varium. The massive xylem bands radiate from a common centre, which is occupied by large-sized metaxylem. The phloem is much less in

quantity than the xylem, and the cells composing it are very uniform in nature. They show abundant contents. The actual protoxylem elements are, as is the case throughout the genus, situated at the extreme periphery of the xylem, but the protophloem is not always so easily recognized. At the base of the stem there is a very marked disposition of the xylem and phloem in parallel plates. There are three parallel plates of phloem, and between these two plates of xylem, while outside each of the two flanking phloem plates there is a single massive xylem group. Whether or not there is a constant relation subsisting between this parallel arrangement and the forking of the stem in its lower region or the giving-off of roots I was not able to determine. There is a very narrow pericycle, two cells only in width, while immediately outside it there is a single layer of cells showing slightly cutinized and red-staining walls. This latter will be the innermost layer of the small-celled inner cortex, which is here only from two to three cells in width. The broad middle cortex is composed of large-sized and thin-walled cells forming a loose spongy tissue with air-spaces, while there is externally a broad zone of smaller, somewhat sclerenchymatous cells which practically alone gives to the stem its rigidity. Fig. 1.—Lycopoium varium. Transverse section of stele in middle region of stem of the Macquarie Island variety. × 137. Fig. 2.—L. varium. Transverse section of stele in middle region of stem of the Antipodes Island variety. × 137. Fig. 3.—L. varium. Transverse section of stele in middle region of stem of the Campbell Island variety. × 137. Fig. 4.—L. varium. Transverse section of stele towards lower region of stem of the Otira Gorge variety. × 90.

The stele of the Antipodes Island form of L. varium, which is shown in Plate IX, fig. 1, B, is illustrated in fig. 2. This is practically identical in all particulars with that of the Macquarie Island plant, although in fig. 2 it will be seen that the phloem and not the xylem radiates from a common centre; also, there is a close correspondence in the nature of the three cortical zones. There is a decided tendency at the base of the stem towards the formation of one or more plates of xylem and phloem extending right across the stele, and lying parallel with each other. I noticed in my sections that this was most marked at a point where the stele was about to fork; but this point needs further investigation. In fig. 3 is shown the stele of the Campbell Island form of L. varium. It will be remembered that this plant is described by Hooker as the largest form of L. varium seen by him. Probably in the full-grown stem the stele will be even larger than shown in this figure, but it will be clear from a comparison with figs. 1 and 2 that with the growth in size of the stele the simple radial form becomes more complex through the connecting across of some of the phloem bands or by the isolation of groups of phloem or of xylem into islands. This is the particular form of L. varium whose stele is figured in Part I of the present series of papers (16, fig. 87). The protophloem is clearly to be distinguished, and there is a large amount of phloem. The cells of the latter tissue are uniform in size, but those which lie immediately adjacent to the xylem bands show abundant contents, while those placed centrally in the phloem groups and bands are empty. There is a very distinct pericycle two to three cells in width, the cells of which show abundant contents. Thus along with the luxuriant growth of this variety of L. varium there go both a change in the configuration of the stele and also a tendency towards a greater differentiation of the tissues composing it. The larger girth of the stems of this particular variety is due to the presence of a very wide soft middle cortex. The outermost, usually thick-walled cortical zone is here very poorly developed, as we would expect from the habitat of the plant. The largest variety of L. varium which I examined was that which occurs at Otira Gorge and Stewart Island (fig. 4). In accordance with the size of the plant the stele is even larger than that of the Campbell Island form, and there is a greater disposition towards the formation of phloem and xylem islands. The typical radiate or stellate disposition of the vascular tissues is thus broken up, but that the stem-anatomy is essentially identical with that of the smaller forms of L. varium is quite clear both from an examination of it as it occurs in the smaller branches and also from an examination of the histology of the several tissues of the stele and cortex. In some of the larger phloem plates in the Otira Gorge plant there is a central row of cells that remain empty and are in clear contrast with the rest of the phloem, which flanks the central row, and which shows abundant contents. This is more marked still in the case of the Stewart Island plant, where the centrally-placed phloem cells are sometimes comparatively large in size. This feature is quite consonant with the robust growth of this variety from the two localities named, and especially with that from the latter place. The outer cortical zone in this variety is strongly sclerenchymatous. In both cases a considerable number of roots appear in transverse section in the middle cortex at the base of the stem, as many as six being noticed in some sections. The stem-anatomy of L. Selago in all the varieties in which it occurs in New Zealand corresponds closely with that described above for the smaller

forms of L. varium. At the base of the stem the xylem and phloem commonly extend across to form parallel plates. This is shown in fig. 5, which represents a section of the stele of the xerophytic variety collected on Browning Pass and illustrated in Plate B, fig. 1, B. The same tendency was observed in the basal region of the stem of the shade variety from Lake Rotoiti. Higher up the stem in this species the arrangement is distinctly radial. In the case of the shade variety collected from a woody gully in the neighborhood of Cass the centre of the stele in the lower part of the stem was occupied by an island of phloem, as is shown in fig. 6. Fig. 85 in Part I of these Studies illustrates the stele of the same material. Higher up the same stem it was found that sometimes the xylem and at others the phloem bands are joined up at the centre, there being thus a tendency towards the temporary formation of one or two more or less parallel plates of vascular tissue. Both xylem and phloem bands and groups are typically massive. The cortex is differentiated into three zones, the outer being Fig. 5.—Lycopodium Selago. Transverse section of stele near base of stem of the xerophytic Browning Pass variety. × 137. Fig. 6.—L. Selago. Transverse section of stele towards base of stem of the mesophytic Cass variety. x 137. Fig. 7.—L. Billardieri. Transverse section of stele towards lower unbranched region of stem of the typical epiphytic form. x 90. Fig. 8.—L. Billardieri var. gracile. Transverse section of stele towards lower region of stem. x 137.

more or less sclerenchymatous, the middle delicate and spongy, and the very narrow inner zone small-celled. The walls of the inner layer of the latter are usually slightly cutinized. One or more roots are to be seen in transverse section at the base of the stem of all the varieties mentioned. On the other hand, the stelar anatomy of the epiphytic L. Billardieri corresponds very closely with that of the largest varieties of L. varium described, albeit on a still larger scale. A transverse section of the lower unbranched portion of the stem of this species is shown in fig. 7. It will suffice to draw attention briefly to the fact that in this case we have an extreme illustration of the tendency to the grouping of the xylem and the phloem in temporary islands and curving bands which is to be observed in all the large stems of the species of the Phlegmaria section. There is an extensive development of large metaxylem, and most of the phloem groups and bands show large centrally-placed empty sieve tubes. The inner layer of the middle cells of the cortex have obviously cutinized walls, there is a spongy middle zone of thin-walled parenchymatous cells with copious airspaces, and the epidermal cortical zone is strongly sclerenchymatous. At the extreme base of a large stem as many as six roots are to be seen in transverse section in the middle cortex. The root stele as seen in the stem cortex is of the typical form so consistent right through both Selago and Phlegmaria sections—viz., a single large crescentic group of xylem with the bay between the two horns occupied by a single phloem group. The root originates from the stem stele by the giving-off of two groups of xylem from two adjacent plates and of a group of phloem from the intermediate phloem plate, the two xylem groups almost immediately joining to form the single crescentic group. The stele of the delicate plant named L. Billardieri var. gracile corresponds exactly with that of the smaller forms of L. varium. This is shown in fig. 8. The configuration of the stele is variable, at one time the plates of tissue being strictly radial, and at another a continuous plate either of xylem or phloem extending right across the stele. In the lower part of the stem one or two roots are to be seen penetrating the middle cortex. From the study of the chain of forms which occur in New Zealand connecting L. Selago through L. varium with L. Billardieri it will be seen, first, that there is a definite type of stelar anatomy characteristic of the whole series, and, secondly, that a gradual change takes place in the vascular arrangement from a strictly radial form in the smaller-growing species to a form in larger species in which the radial or stellate arrangement is broken up by cross connections, which result in the isolation of some of the xylem and phloem into islands. Moreover, there is a well-marked and gradually increasing tendency in larger and larger forms to the differentiation of the phloem into sieve tubes and phloem parenchyma. This is quite in accord with Jones's description (21) in his study of the anatomy of the stems of twenty species of Lycopodium. In the case of L. serratum Thunb. and L. reflexum Lam., both of which species belong to the L. Selago cycle of affinity, Jones's figures of the stem stele show that along with the fact that in general habit the two species named grow more robustly than L. Selago the arrangement of the vascular tissues is no longer simply radial, but the general configuration of the more abundant xylem and phloem is broken up by cross connections, with the consequent formation of phloem islands. His figures of the stele of the strongly growing epiphytic species L. squarrosum Forst., L. Dalhousieanum Spring, and L. Phlegmaria Linn are closely similar to that of the large New Zealand epiphytic L. Billardieri.

Thus it may be assumed that the general type of stelar structure which has been shown in the present paper to hold throughout the whole chain of New Zealand forms holds also for the whole of the subgenus Urostachya, and that all modifications of this type as seen in the different species of the subgenus are merely the result of the particular size of each species and of its habit of growth, which, as has been shown above, varies with the environment. In his paper cited above Jones arrived at rather different conclusions. He would classify all the species' he examined, in so far as their anatomy is concerned, into two groups, “the second containing a number of epiphytic forms, and the first plants with horizontally-growing stems, while species with erect stems may be included in either group” (21, p. 31–32). He would link the group of L. Selago with the clavatum type. My own conclusions, derived from a thorough study of the New Zealand species, are that, using the sections of Pritzel's classification, which are certainly more natural than those of Baker, there are three main strains of stelar anatomy to be traced in the genus Lycopodium, the first of which is characteristic of the sections Selago and Phlegmaria, the second characteristic of the sections Inundata and Cernua, and the third of the section Clavata. This, it seems to me, is in accord with Jones's own figures and descriptions of the stelar anatomy of the species examined by him, and also goes hand in hand with the other main characteristics of both sporophyte and gametophyte. Prothallus, Sexual Organs, and Young Plant. In my previous paper already cited (16, p. 264) I have given a short general description of the external features of the prothallus of L. Billardieri. Since this was written I have found in the neighbourhood of Hokitika, Westland, an additional large number of specimens of the prothallus of this species, also a very large number of prothalli of L. Billardieri var. gracile from two or three localities in the same neighbourhood, and about a dozen specimens of the prothallus of the Otira Gorge variety of L. varium. Anticipating here what I hope to describe more fully in a further paper with regard to these prothalli, I may say that they are all closely alike. The majority of the specimens found of all three species show, on the whole, longer processes than those figured in the above-named paper, figs. 1–6. There is in each complete prothallus a more or less bulky central region on which the sexual organs and paraphyses are formed, and from this arise the long vegetative processes, which are richly covered with long rhizoids. The ends of some of these processes sometimes also become more bulky and bear antheridia. There is, however, a wide range of variability in the general form of the prothallus, brought about by the greater or lesser bulkiness of the central region, and also by the varying development of the processes both as regards numbers and length. In all cases the first stages in the development of the prothallus from the spore result in the immediate formation of a cone-shaped tissue-body, which by elongating and thickening becomes the central portion of the mature prothallus, and on this the processes arise adventitiously. Thus the first-formed region of the prothallus of these three species conforms to the fundamental structure plan of the Lycopodium prothallus, although the mature prothallus is variable in form. In the case of all the three New Zealand species mentioned I dissected out numerous detached prothallial processes which were evidently developing independently; and also not a few of the mature prothalli bore evidence of the fact that they had originated not directly from the spore. but from such detached processes.

Treub (30) has described the prothalli of the four following tropical epiphytic species: L. Phlegmaria Linn., L. carinatuim Desv., L. nummularifolium Blume, and L. Hippuris Desv. Of these four species, L. Phlegmaria and L. nummularifollis belong, according to Pritzel's classification, to the Phlegmaria section, L. Hippuris to the Euselago subsection, and L. carinatum to the Subselago subsection of the Selago section. They thus cover between them all the main divisions of the subgenus Urostachya. According to Treub's conclusions, the prothalli of all these four species are alike, belonging to that form of the Lycopodium prothallus known as the Phlegmaria type. That of L. carinatum is exactly like that of L. Phlegmaria. The prothallus of L. Hippuris, however, is much larger and thicker, and that of L. nummularifolium very much thinner and more delicate, than that of the type species. Miss Edgerley (12, pp. 104–9) has described the prothallus of the New Zealand species L. Billardieri, and her results are in close accord with my own. It corresponds very closely with that of L. Phlegmaria both as regards external form and internal structure. There are, however, two characteristic differences to be noted. In the case of L. Phlegmaria the paraphysis is composed of a linear row of as many as a dozen cells, and it may also branch, whereas in L. Billardieri it is much simpler in form, is always unbranched, and is generally only three cells in length. Also, although I have examined a very large number of prothalli of the three New Zealand species mentioned above, I have never observed any vegetative buds of either of the kinds described by Treub as occurring freely in the prothallus of L. Phlegmaria. But in the New Zealand species short clubshaped richly-stored resting vegetative processes commonly occur. The prothallus of L. Selago has been described by Bruchmann (5), and presents some exceedingly interesting modifications of form. Bruchmann says (5, p. 85), “This variety of form of the prothalli seems to be dependent mostly on the soil in which they are produced. The elongated cylindrical forms are found especially in firm soil, in which they strove towards its surface mostly in a vertical direction. In loose soil, especially near the surface, I came across more thickset and flat forms of prothalli.” Moreover, besides the subterranean forms of prothallus of this species, he found some which grew wholly or partly at the surface of the earth, and which in their upper part showed a thoroughly green colouring. Such prothalli lived as semi-saprophytes, and by their manner of life formed, as Bruchmann himself says, “an interesting transition between the assimilating and the merely saprophytic forms” of the Lycopodium prothallus. Every complete prothallus, whether of the elongated or thickset form, shows at the original end a tiny, usually bent point, which is the region immediately developed from the spore, and which develops above into the cone-shaped tissue-body. The prothallus is abundantly supplied with long rhizoids; and many-celled paraphyses, similar to those described by Treub in L. Phlegmaria, are present, along with the sexual organs. Spessard (25) has recently given a short account of the prothallus of L. lucidulum as it occurs in America. It is a cylindrical elongated body, bearing paraphyses at its uppermost end. In discussing the different types of Lycopodium prothallus Lang (24, pp. 305–6) says with regard to that of L. Selago, “The two forms of prothallus found in L. Selago give the clue to the more specialized saprophytic types, which in the deeper-growing subterranean species retain the radial symmetry while becoming modified in shape. On the other hand, the

type of prothallus growing in rotting wood has lost the radial symmetry, and consists of cylindrical but more or less clearly dorsiventral branches.” Curiously enough, Bruchmann himself does not read any signs of the transition existing between the different types of prothallus from the significant variability of that of L. Selago. He gives to the prothallus of this species the rank of a new type, and concludes also at the end of his paper (5, p. 108), “From the above facts it follows that the above-treated-of Lycopodium groups, characterized especially by their generative generation, do not stand to each other in near relationship—that is to say, not in such as one would expect with plant species that are found in the same genus. This knowledge leads to a separation of the lycopodiums into groups, or, better still, into genera.” However, the facts known concerning the gametophyte generation of the species which comprise the subgenus Urostachya undoubtedly point to the fact that the two sections Selago and Phlegmaria are closely related; and, seeing that the species L. Selago is probably to be regarded as the primitive type of the subgenus, if not, indeed, of the whole genus, it follows that the Phlegmaria type of prothallus, which has been found to occur in all the main divisions of the subgenus where the species have an epiphytic habit, has arisen as a modification of the Selago type in accordance with that habit of growth. One variation in the prothallus of the New Zealand species, however, is not in accord with the view that the series Selago-varium-Billardieri-Phlegmaria represents a linear series—namely, the paraphysis in the prothallus of L. varium and L. Billardieri differs markedly in size from that of L. Selago, whereas the L. Phlegmaria paraphysis is similar to the latter. This isolated fact would indicate that evolution in the form of the prothallus in the subgenus Urostachya has proceeded along several parallel lines, and that there are to be traced, as Lang suggests (24, p. 313), “instances of independent adaptation to similar conditions.” I am not prepared in this paper to enumerate the variations in the form of the sexual organs to be observed from a study of the different types of prothallus as they occur in New Zealand, or to discuss the question of their modification from the ancestral type in accordance with the form and habit of the prothallus and sporophyte. Such a study has been suggested to me by Professor Charles Chamberlain, of Chicago, in a letter in which he points out the great interest that would come from seeing “what the resultant between the force of heredity and the influence of environment might have on the antheridium and archegonium and young embryo,” and I hope to be able at some future time to carry out the suggestion. The “seedling” plants of the New Zealand species which belong to these two sections conform to the type described by Treub in L. Phlegmaria and by Bruchmann in L. Selago. I have noticed no variations in the New Zealand species. Bower (3, pp. 346–47) has expressed the view that this type of embryo is the least modified in the genus. However, we can see in the variation in length of the hypocotyl according to the depth at which the prothallus grows another indication of the great plasticity of the Lycopodium plant, which is to be noticed in almost every organ. Species Belonging to the Sections Inundata and Cernua. In these sections there is less variability to be observed in the habit and external form in the New Zealand species than in the last two sections. However, in the gametophyte generation and the young plant there are certain interesting modifications.

The sections Inundata and Cernua comprise comparatively few species. The species which belong to the former are fairly widely distributed, but are confined to boggy and marshy habitats. I venture to propose that the composition of the Cernua section as arranged by Baker, and by Pritzel following him, should be altered in accordance with our increased knowledge of the main characters of the species which have hitherto been included in it, certain of these species, such as L. densum, L. volubile, and possibly L. obscurum and L. casuarinoides, being removed to the section Clavata, while L. ramulosum, and possibly also L. diffusum, should be removed from the latter section and placed near to L. cernuum. This suggestion has been previously made by me (16, p. 301). The reason for this is that the cernuum type of prothallus, a protocorm stage in the embryo plant, and the “mixed” type of stelar structure undoubtedly go hand in hand, and that although the species which show these characters do not altogether show a close similarity in general habit of growth and in external form, yet the presence of the first-named characters should be given chief position in determining the limits of the Inundata and Cernua sections. Now, Baker and Pritzel in their classifications of the genus have both described the species L. cernuum and L. densum as upright, tree-like forms. This is quite misleading. It is only the lateral branches in these two species which are erect. There is a trailing main stem which in the case of the former species is snake-like in growth, spreading above-ground for as much as 12 ft. to 15 ft. in a series of loops, and rooted to the ground at each node, and in the case of the latter species is a subterranean rhizome which attains an extreme length of 8 ft. to 10 ft. The species L. laterale also, which is placed near L. cernuum by Pritzel, is described by him as nearly or quite erect. This also is an incorrect description, for whereas the aerial branches are more or less erect, the mam stem is a short much-branched underground rhizome which ramifies extensively in the soil. One rather suspects that the supposed upright habit of growth of these species has been one of the reasons which led the two systematists into grouping together such really widely different species as L. cernuum and L. densum. For the same reason, the great point that is made by Jones (21, p. 32), that “the fact must not be overlooked that the banded structure is well marked in the stem of the erect-growing species of L. obscurum,” likewise loses its significance in view of the fact that in this species also the main stem is subterranean and creeping, and that it undoubtedly belongs to the Clavata section. The type of prothallus and young plant and also the stelar anatomy of L. densum and L. volubile indicate that these species, as well possibly as the two species L. obscurum and L. casuarinoides, should be included in the section Clavata. This, then, would leave in the Cernua section only the type species L. cernuum with all its varieties, and the species L. laterale with its congeners L. ramulosum and L. diffusum. The fact that one or two species in the Inundata section—e.g., L. contextum Mart. and L. cruentum Spring—show the Selago habit in the fertile region is open to various interpretations. Either they have preserved unaltered a phylogenetic character; or perhaps it is an indication that the whole genus is really to be read as a reduction series, and that in these particular species, as also in the case of the New Zealand plant L. Billardieri var. gracile, mentioned above, actual transitions in the process are to be seen. However, isolated instances such as these are not to be interpreted apart from the main evolutionary tendencies to be observed in the genus as a whole.

Section Inundata. There is only one species in New Zealand which belongs to the Inundata section, this being L. Drummondii. I have gathered this species on the sphagnum peat-bog at the outlet of Lake Tongonge, Kaitaia, North Auckland, the only locality where it has been found in New Zealand. It will be convenient to consider this species by itself. External Form of Plant, and Nature of Strobilus. In Plate XIV, at C, are shown two small specimens of this plant. It may attain a length of 7 in. to 8 in., and possess one to three spikes. The variation in form from the normal which I have observed in this species is that occasionally the cone, which corresponds in form to that of L. carolinianum figured by Pritzel in Engler and Prantl (13, fig. 378), may be interrupted. In some plants the lower half of the cone clearly belongs to the past season's growth, being dark in colour, with empty outstanding sporophylls, and there is a well-defined demarcation between this and the farther prolongation of the cone in the succeeding season. Again, in other instances there is a sterile zone as much as ½ in. in length separating the two portions of the cone. In Plate XIV, C, the left-hand specimen shows an interrupted cone, the point of demarcation between the two seasons' growth being indicated by a cross. Also, in some specimens the erect fertile branch which generally appears as a very distinct peduncle to the cone, with leaves in scattered whorls, shows instead throughout a greater or lesser portion of its lower region crowded leaves such as are borne on the trailing stems. Thus, although in this species the fertile region is differentiated as a definite club-shaped cone, and even the erect branch on which it is borne is also differentiated as a peduncle, yet variations occur which indicate that this character is in a state of plasticity. Stem-anatomy. The stelar anatomy in the main stem is shown in fig. 9. It will be well for me to describe this in some detail here, as it is typical for all the New Zealand species which belong to the sections Inundata and Cernua. The general appearance of metaxylem and protoxylem, as well as the configuration of these tissues, is in striking contrast to that of the same tissues in the stem of the Phlegmaria and Selago sections. In L. Drummondii the metaxylem is somewhat feebly lignified, and does not show the presence of small-sized elements flanking the larger ones. Also the degree of coherence of the metaxylem elements one to another into plates or groups is much slighter, so that there is a very characteristic mixing of the xylem with the phloem. The protoxylem elements are much fewer in number in this species than in those of the Selago and Phlegmaria sections, where the protoxylem groups are exceedingly massive, but they are so extended Fig. 9.—Lycopodium Drummondii. Transverse section of stele of creeping stem. × 137.

peripherally as to form an interrupted cylinder around the stele. They are also only feebly lignified. This nature of the xylem and protoxylem in L. Drummondii gives a very characteristic appearance to the stele as a whole, and one which is in marked contrast to its appearance in the other divisions of the genus. Moreover, it is found, as will be shown below, not only in the Inundata, but also in the Cernua section. In the ultimate branches, and in the fertile branch of L. Drummondii, the configuration of the vascular tissues is more definite, as would be expected where the vascular elements are much fewer in number, and the xylem might also be described as radial. The protoxylem groups, however, are just as much extended peripherally as in the larger parts of the plant, and the individual elements of the metaxylem are just as feebly lignified, so that it is apparent that right throughout the plant a consistent type of stelar anatomy and of vascular histology is to be found. Jones (21) describes the stelar anatomy of the strobilus of L. inundatum as radial, and notes that in the branches a striking tetrarch structure is obtained. In his figure of this he shows both xylem and protoxylem as being very broad at the periphery. This is very similar to what I have just described for L. Drummondii. The cortex in L. Drummondii shows an inner narrow more or less sclerenchymatous zone which is more strongly developed in the erect fertile branches, whilst the main cortical tissue consists of large spongy parenchyma with abundant air-spaces, grading off externally into a very ill-defined epidermis. I have been unable to find a mucilage-cavity in the leaf such as has been described in L. inundatum. The Young Plant. I was not successful in my search for the prothallus of this species, but along with the adult plants I discovered a large number of plantlets of all sizes which seemed to have originated vegetatively. All the specimens which I have preserved are of small size and are unbranched. They very early adopt the plagiotropic habit, and at once begin to develop adventitious roots. Such specimens as are complete show that at the base the plantlet is prolonged directly into its first root, and that the vascular strand is continuous throughout this root and the shoot. In one or two instances I could clearly trace a broken prominence where the shoot passes into the root, at which spot presumably the plantlet had been attached to the parent body. One specimen consisted of a small old piece of rhizome with its roots, and attached to the rhizome was one of these plantlets showing both shoot and root. From this specimen, and from indications on many of the others, I therefore judge that the plantlets are propagative shoots which arise on the older rhizomes at the season of the year when the latter are beginning to die off. They give evidence of no structure which can be likened to a protocorm. This is noteworthy, for we should rather expect to find that the embryogeny of this species would show a protocorm stage, and that even vegetatively produced plantlets would also possess this structure, as do those of L. ramulosum, which I have described elsewhere (17). Section Cernua. Belonging to the section Cernua are the New Zealand species L. laterale L. cernuum, and probably also L. ramulosum.

A, L. cernuum: Fertile branch with one loop of the trailing stem (typical). B, L. cer nuum: Portion of fertile branch from New Hebrides Islands.

Fig. 1.—A, L. laterale: Portion of underground rhizome bearing fertile aerial branches, slender variety. B, L. ramulosum: Complete fertile plant showing rhizomes and aerial branches. Fig. 2.—A, L. volubile. Portion of lateral branch showing vegetative and fertile regions. B, L. scariosum: Portion of creeping stem with two fertile erect branches.

External Form of Plant and Nature of Strobilus. Of these L. cernuum is illustrated in Plate XI, A.This figure shows one loop of the main plagiotropic stem with two nodes on which roots are borne, and also a single erect fertile branch arising from the upper side of the main stem. The figure which Pritzel gives in Engler and Prantl (13, fig. 379, A) of the erect fertile branch of this species shows roots at its base. This is obviously wrongly figured. The only variation to be noted here in the external form of the plant is that the New Zealand variety is a very robust-growing plant, while certain tropical forms of the same species are much more delicate in all parts of the plant. This is well seen from a comparison of the two forms marked A and B in Plate XI, A being the New Zealand variety and B a variety from the New Hebrides. Specimens from Fiji and from the West Indies which I have in my possession correspond very closely with B, and Mr. Cheeseman informs me that this is also the form of the species as it occurs in Melanesia and north Australia. An examination of Pritzel's classification of the section Cernua shows that several of the species included in it are varieties of the typical form L. cernuum. Two erect-growing aerial shoots of L. laterale, together with a small portion of the underground rhizome from which they arise, are shown on Plate XII, fig. 1, A. This illustrates well the form assumed by this species when it grows amidst tangled vegetation composed of Leptospermum, Gleichenia, &c. The aerial shoots in such a habitat are very slender, and attain an extreme height of 2 ft. to 3 ft. It is obvious that such shoots depend upon the surrounding vegetation to support them in an erect position, and that therefore they are abnormal in form. The specimen illustrated shows that the extreme height of these shoots represents the growth of two seasons, the point at which the second season's growth commenced being indicated by a cross. When growing on open, boggy hillsides, as it does commonly on the kauri-gum lands of the Auckland Province, the aerial branches of this species assume a much shorter, stouter, and less-branched form, and are reddish in colour. The rhizome, with its vascular tissues, also in this case is stouter. In this species the cones are typically lateral and sessile, but they are also sometimes borne on short, leaf-covered peduncles, and are then to be regarded as terminal This variation in the position of the cones in L. laterale provides a transition to the next species, in which one of the chief distinguishing characters is the terminal position of the cones. The species L. ramulosum is described by Pritzel as occurring in the New Zealand mountains; but its habitat is, on the contrary, wet ground at low altitudes, especially that of sphagnum bogs.* See Postcript, p. 216. It was first described by T. Kirk (22) from Westland, and later still from Stewart Island. I have studied this species in both these districts, and have some interesting variations to record. A single plant is shown in Plate XII, fig. 1, B. Cheeseman figures it in his Illustrations of the New Zealand Flora (9), and a plant, incomplete in its lower regions, is also figured by Kirk (22, pl. 19, fig. B). In his first description of it Kirk says, “Not infrequently two spikes are produced from the apex of a branch, and rarely the fertile branch is overtopped by a luxuriant ‘usurping shoot’ so that the spike appears to be lateral, showing its close affinity with L. laterale, which is still further strengthened by the fact that in that species the spikes are not invariably sessile, but occasionally are developed on very short leafy peduncles.” The most obvious point of difference between the two species is the usually prostrate and densely matted habit of L. ramulosum, and, as will be seen

below, the prothallus of the latter is peculiar. Cheeseman (9, p. 250) states that L. ramulosum “is more closely allied to the Australian L. diffusum than to any other species, principally differing, as Mr. Baker has remarked, in its entirely terminal spikes, whereas in L. diffusum they are frequently lateral.” It would seem, then, natural to place these three species very close together in the same section, and not to dissociate them, as both Baker and Pritzel in different ways have done. When growing on sphagnum bogs throughout Westland, where it is associated with Gleichenia alpina and Cladium teretifolium, the plants are closely intermatted, and their rhizomes penetrate the peaty soil in all directions. The aerial branches, by reason of the close nature of the vegetation which covers the ground, are more or less erect. When it is growing amongst thicker vegetation on a hillside I have often observed that L. ramulosum occurs in dense, mossy cushions, the intermatted aerial branches in these cases being particularly well developed and drawn out. On wet soil on which there is a very scanty mossy covering individual plants of this species assume a habit in which the branches are closely pressed to the surface, the cones alone standing erect, almost pedicelled. In this case the rosette form of the plant, produced by the continued dichotomies of the branches, is very striking. The main stems show a considerable range of variability in their form. They may be, in drier situations, above-ground and green, and covered with ordinary vegetative leaves, but on the bogs they function as subterranean rhizomes, white in colour, with scattered scale leaves. Again, the more deeply penetrating rhizomes in the latter situation are often quite naked and brown in colour. L. Cockayne (10, p. 17) has made the interesting observation that the cones of L ramulosum “are absent or scantily produced in shade plants, but extremely abundant in those growing in bright light.” He has suggested that this is a point which would lend itself well to experimental investigation. It is apparent, then, that the habit of L. ramulosum is very variable, although in its typical form on sphagnum bogs it is quite characteristic. The external form of L laterale also is in a plastic state, and the position of the cones in neither species is quite fixed. Thus in respect of these characters they are more or less closely allied, and also their affinity with the Australian L. diffusum seems to be clear. Stem-anatomy. In describing the characteristic nature of the vascular tissues of L. Drummondii I indicated that this mixed type of stele with the very extended protoxylems is that also of those New Zealand species of Lycopodium which belong to the section Cernua. It is very distinct from the first type described in this paper, which, as I have tried to show, holds, in spite of the many modifications in the external form of the plant, throughout the subgenus Urostachya, and also from the third type, which is characteristic of the Clavata section. It therefore should be given prominence to, along with the character of the prothallus and of the embryo plant in the sections Inundata and Cernua, as a character to be taken into account in a natural classification of the genus. In fig. 10 is shown the vascular cylinder of the main rhizome of L. cernuum. It will be seen both from the amount of vascular tissue and from the scale of the magnification that the stele of this species is an exceptionally large one. For that reason it shows even more clearly than do the other species the nature of the mixed type of structure. Both metaxylem and phloem

are much broken up into curving bands and isolated groups, and the protoxylem is markedly extended around the periphery of the stele. The protoxylem and metaxylem elements are poorly lignified. The metaxylem elements are large and are all of one size, there being no small flanking tracheides present. In fig. 10 it is apparent at several places that the metaxylem elements are separating from each other. In accordance with the larger size of the phloem groups there is a distinct differentiation between centrally placed large phloem elements and flanking phloem parenchyma, the latter in the region immediately behind the apex of the rhizome showing abundant contents. In the nature of its cortical tissues this species differs considerably from L. Drummondii. There is an outer zone of sclerenchyma which includes the epidermis, while the rest of the cortex is thinwalled but not of a spongy nature. Immediately surrounding the vascular cylinder the cortex is small-celled, as will be seen in fig. 10. The pericycle in the mature stem is very indistinct, but in the region immediately behind the rhizome apex it is seen to consist of a single layer of cells adjoining the protoxylem, showing abundant cell-contents. The cells of the innermost cortical layer show quite distinctly thickened angles: this will be the endodermis. Fig. 10.—Lycopodium cernuum. Transverse section of stele of main trailing stem. × 60. Fig. 11.—Lycopodium laterale. Transverse section of stele of rhizome of strongly growing variety. × 90. Fig. 11 is that of the stele of the more strongly growing variety of L. laterale. This is somewhat larger than that of the tall slender form, but otherwise does not differ from it. Behind the stem-apex the peripherally extended nature of the protoxylems is very marked, but in the mature stem it is not quite so clear, owing to the fact that the smallest protoxylem elements are very feebly lignified, and also become crushed, and so are not always easily recognized. The vascular tissues are quite clearly of the mixed type. The cortical tissues are very similar to those of L. Drummondii,

there being an inner slightly sclerenchymatous zone, which is more strongly developed in the aerial branches, merging outwards into a thin-walled spongy tissue which is continued right up to the epidermis. In dissecting out the rhizomes of this species from the soil and cleaning them it was noticeable how very easily the white spongy outer cortex could be stripped away from the central core. In the ultimate branches I found a tetrarch condition of the stele. The four extended protoxylems and the four compact groups of phloem which alternate with them are very distinct, but the few large metaxylem elements, even in these small branches, are continually changing their disposition. Mention has been made of the two kinds of underground stem of the species L. ramulosum. Those which penetrate the soil most deeply are brown in colour and smooth, being devoid of scale leaves. They are thicker than the white scaly rhizomes, and their vascular cylinder is proportionately larger. Fig. 12 shows the stele of the larger kind, while the figure given in Part I of these Studies (16, fig. 94) is that of the smaller form The cortical tissues differ considerably from those of L. laterale. There is an inner, thin-walled zone, which in the large more deeply growing rhizomes is relatively wide and contains much starch, a median sclerenchymatous zone, and an outer, thinwalled region in which there are abundant air-spaces. The vascular tissues correspond in nature and arrangement to those of the other species in this section of the genus. Practically surrounding the main vascular tissues there is a ring of flattened and distorted cells. In sections of the mature rhizome it is difficult to determine the exact nature of this ring, but from a study of sections taken immediately behind the apex of the rhizome it is clear that the original protoxylem has contributed to the greater part of it. It is evident, then, that the vascular anatomy of the four New Zealand species described above as belonging to the sections Inundata and Cernua belongs to a common type which is best described as “mixed,” this being seen best in L. cernuum. In all these species, although the habit is plagiotropic, branching takes place all around the stem, so that no directive tendency is present towards a dorsiventral disposition of the vascular tissues as in the species of the Clavata section. The mixed nature of the vascular tissues, however, is not merely the result of the all-round branching; it is present from the early seedling stages of the plant, and goes along with certain characteristic features in the histology of the xylem and protoxylem, and it is best regarded as part of the inherited constitution of this division of the genus. The only variation in the nature of the vascular tissues is that in L. cernuum the phloem shows considerable differentiation, this arising probably simply from the large size of the Fig. 12.—Lycopodium ramulosum. Transverse section of stele of deeply penetrating rhizome. × 137.

plant and of its vascular tissues. The stem-anatomy of only one other species belonging to these two sections has been examined—viz., that of L. inundatum, described and figured by Jones (21). His figure of the tetrarch structure and broad protoxylems in a small branch of this species corresponds very closely with what I have observed in the corresponding parts of the plant of all four New Zealand species. The cortex in the New Zealand species varies very markedly. The position of the sclerenchyma zone in L. cernuum is outermost, in L. ramulosum it is median, and in L. Drummondii and L. laterale it is innermost. There is a distinctly spongy region of the cortex in the three latter species which is quite in accord with the nature of their habitat. The external position of the sclerenchyma in the stems of L. cernuum is perhaps to be explained by the fact that these are above-ground, and that this species seeks drier situations than do the other three. It is not quite so clear, however, what is the particular physiological reason for the median position of the sclerenchyma in the cortex of L. ramulosum unless it be because in this species the innermost cortex is used as a storage zone. The Prothallus. The prothallus of these two sections, so far as it is known, belongs to the L. cernuum type, and this is one of the main arguments which determines that the two sections together form one of the natural divisions of the genus. However, the prothalli of all the species that have been examined show striking variations from the typical form, and in the case of the three New Zealand species, L. cernuum, L. laterale, and L. ramulosum, there is a wide range of variability in the form of the prothallus in each individual species. In fact, in these species there is a far greater degree of plasticity shown by the gametophyte generation than by the mature sporophyte. This is to be expected, for whereas on the one hand the species are limited in their distribution, and hence the sporophyte is under a comparatively constant set of external conditions, on the other hand the prothallus, being very delicate in nature and combining the chlorophyll condition with the saprophytic, must necessarily show considerable variation from the typical form in accordance with the varying depth at which the spores germinate. In these two sections of the genus, in contradistinction to what obtains in the other sections, the prothallus lasts for only one season. The spores germinate only in consistently damp, loose soil, more especially that which possesses a thin covering of short moss. Being partly dependent upon the presence of light, they cannot germinate at any great depth in the soil, but yet amongst the moss and other delicate debris of vegetable matter there is, of course, a considerable range of variation in the conditions under which the development of the prothallus must take place. The prothallus of L. inundatum has been described by Goebel (14). This corresponds fairly closely with the L. cernuum type. Treub (30) has described the prothallus of L. salakense, a species which is apparently a variety of L. cernuum. This he says belongs to the cernuum type, but it is not so closely similar to it as is the prothallus of L. inundatum. Several filaments may develop from the primary tubercle, one of these afterwards developing into the main prothallus-body. There are no foliaceous lobes produced on the crown. Treub has given a very full account of the prothallus of L. cernuum. I have not had access to his original papers, and hence am not aware whether or not he has described in the case of any of his specimens the

variations in structure which I have found in the prothallus of this same species as it occurs in New Zealand. In a previous paper (16, p. 266) I noted that the length of the shaft varies greatly in the case of different individuals, it being comparatively long in some and in others almost absent, and stated that this arose simply from the fact of the variation in depth at which the spores germinate. With regard to the internal structure of the prothallus, I must anticipate here what I hope to describe more fully in a future paper. The lower primary tubercle not infrequently consists of two distinct swellings, and not one only. The extreme basal end of the prothallus constitutes the original tubercle, but the second swelling is separated from this by a slight constriction and occupies a position higher up one side. In some prothalli the development of this second laterally placed swollen region gives a somewhat lopsided appearance to the prothallus. This will be seen in fig. 21 of the paper cited above (16), and also in fig. 13, which is that of a similar prothallus of L. laterale. In some of the short prothalli the two swellings appear side by side, together forming the basal part of the prothallus, the crown of lobes arising from both together: this, of course, serves to impart a thickset appearance to the prothallus as a whole. Contrary to what botanical writers, quoting from Treub's original papers, have stated with regard to the internal structure of the primary tubercle in L. cernuum, I find that a well-marked differentiation of fungusinhabited tissues is there to be seen, though, of course, to a much less extent than in the prothalli of the clavatum or complanatum types. The outer peripheral layer of cells is for the most part only one cell in thickness, and this contains the spherical coils of fungal hyphae. Those cells of the interior of the tubercle which immediately abut on the peripheral layer are very distinctly elongated at right angles to the latter. They are narrow, and in transverse section are roundish in outline, and in this layer the fungus occupies a position in the cell-walls. In fact, this layer may be compared to the palisade cells in the vegetative part of the clavatum or complanatum prothallus. In some cases I noticed that the cells of this interior tissue contained abundant small bodies which I took to be starch-grains. Apart from those instances in which there is a second swollen fungal region occupying a lateral position, I have not observed the fungus penetrating into the region of the shaft. The two fungal tissues occupy the whole of the first-formed tubercle, but in the case of the second swelling they are, as I have said, laterally placed. If now prothalli of this species occurred in which the fungal regions extended in this way farther up the main body of the prothallus, not only in one localized position but uniformly all around it, leaving a central core of undifferentiated cells, we would have a condition approaching that of the prothallus of L. complanatum in its basal region, as described by Bruchmann. I have not observed such cases in the prothallus of this species, but what has just been described becomes of some significance in this connection when compared with the further variations in the prothallus of L. ramulosum to be described below. In the case of the complanatum and clavatum types of prothallus the main characteristic is that the fungal tissues have become so important a part of the prothallus-body that their development is not left to the mere spreading of the fungus to additional regions of the prothallus, or to the successive infection of those regions from without through the rhizoids, as is obviously the case in L. cernuum and L. laterale, but proceeds uniformly from the meristem, the vegetative part of the prothallus consisting of one large swelling and not several small distinct ones. On

re-examining my serial sections for the purpose of this paper I found that in one or two instances the second fungal swelling was associated with the development of assimilating lobes arising laterally on the prothallus-shaft and on the opposite side of the swelling. The prothallus of L. cernuum is thus in a somewhat plastic condition, and the variations in form which are to be observed make it easier to institute comparisons between this type of prothallus and those of the other sections of the genus. In my previous description of the prothallus of L. laterale (16, p. 265) I noted that, as in L. cernuum, the shaft is variable in length, the shorter prothalli having thus a somewhat solid, compact appearance, and that the leafy expansions on the crown of the prothallus of L. laterale are less lobelike and more filamentous than in the case of the other species. Another character noted was that these assimilating outgrowths occur also normally in a lateral position on the shaft of the prothallus. The presence of a long, narrow, club-shaped process attached to the tubercle of two of the prothalli was also mentioned, and I was inclined to regard this as the actual first-formed portion of the prothallus and not as a branch arising from the “primary” tubercle. As in the prothallus of L. cernuum, the fungal zone may extend for a certain distance up one side of the shaft, giving rise to a second swelling distinct from the lower one. In none of my preparations did I find that the interior cells of the swollen areas which adjoined the fungus-infected peripheral cells were so clearly differentiated as a distinct fungal tissue as they are in the largest prothalli of L. cernuum. The fungus extends between these cells, in one particular instance as far as into the lower region of the shaft. These cells have without doubt a definite arrangement, with their long axes at right angles to the peripheral layer, but in longitudinal and transverse section they appear little differentiated in form from the other cells of the shaft. Thus the fungus zone in this species is of the same nature as that in L. cernuum, but it shows a somewhat less degree of differentiation of its tissues. I found one very young prothallus of this species entangled closely amongst the rhizoids of an older prothallus. It consisted of a filament nine cells long, the older half of this prothallus being but one cell wide, and the younger half two cells wide. At the base of the prothallus was to be seen the original spore, and four or five rhizoids were borne on the older cells. There was no suggestion of a primary tubercle, nor was the presence of a fungus to be seen, but all the cells showed numerous chloroplasts. This young prothallus may be compared with those of the same early stage in L. ramulosum; described below. The prothallus of L. ramulosum is very variable in form (16, pp. 269–71), much more so than that of the other two species. The long-drawn-out forms have several swollen fungus-infected areas, bearing rhizoids, some of the largest prothalli showing as many as five such areas. Each of these fungus swellings has usually a group of assimilating lobes associated with it, and at the base of the lobes antheridia or archegonia are developed. Not infrequently the form of the prothallus is strangely altered owing to the fact that a somewhat massive body of tissue is formed at a place where a group of lobes is associated with a fungus area. Generally the last-formed uppermost part of the prothallus is the bulkiest, but two or even three such masses of tissue are often to be seen in these long-drawn-out forms of prothallus. In every case in which a young plant was attached to a prothallus it had developed from the uppermost of these bulky regions. One or two prothalli were found which had branched at the point of a swollen fungal region, the two branches being equally developed. As well

as the long-drawn-out forms there are short, comparatively massive individuals. These correspond to such a group of lobes with its associated fungal region as has just been described, the development of the massive body of tissue having proceeded to an unusual extent. Sometimes it is seen that in these massive prothalli there has been an original first-formed, long-drawn-out portion which has withered away, but always the subsequent growth of the prothallus is confined to increasing the massiveness of the main tissues and does not result in any further extension in length. In many cases, especially at the actual crown of the prothallus, the group of lobes has withered, there being a consequent browning of the upper surface at this point. In some instances, instead of lobes being present, there are only feebly developed warty excrescences. Again, in other instances the lobes are fairly thick in form. Serial sections of the prothalli show that the fungus swellings are all identical in nature and correspond very closely with what is found in the prothallus of L. laterale. The peripheral cells contain the spherical masses of hyphae, while those in the interior of the swelling show the fungus only within the cell-walls. These latter cells have not assumed such a definite tissue-form in accordance with their particular function as is sometimes to be seen in L. cernuum. The mitial stage in the formation of a fungus area is sometimes to be seen, one or two epidermal cells showing the presence of the fungus, but no swelling having yet taken place. In not a few prothalli, both of the massive and of the long-drawn-out form, the actual first-formed region nearest the original spore was to be seen. Moreover, several very young prothalli were found. There is first a longer or shorter filamentous stage, in which the young prothallus is only one cell wide, after which it proceeds gradually to increase in width as it lengthens. The prothallus is from the first quite green in colour, and infection by the fungus seems to take place subsequently to these early stages of growth. However, in one or two prothalli of the massive form I noticed that the original but very short filament passed immediately into the bulky fungus-containing tissue, the complete prothallus having very much the form of an inverted cone with the spore filament at its apex. The resemblance of this to the clavatum type of prothallus is obvious. The prothalli of the three New Zealand species which belong to the section Cernua are thus seen to be in a condition of great plasticity. The main reasons for the modifications of the typical cernuum form seem to be the varying depth at which the spores germinate, together with the extent of infection by the fungus element, this resulting in all three species, and especially in L. ramulosum, in a form of prothallus more or less elongated, which develops rhizoids at intervals along its length, a form which may be well compared with that acquired by the humus-growing, wholly saprophytic prothalli characteristic of the Phlegmaria type. The various forms of the prothalli of L. Selago and L. ramulosum show how it is possible for both the much-branched Phlegmaria type and the compact, more massive clavatum or complanatum type to arise as modifications of the original form. The prothalli even of L. cernuum and L. laterale give mdications of this This, taken in conjunction with the fact of the great plasticity of the species as seen in all their main characters, and the occurrence of numerous transitional forms both between one species and another and also between one type and another, indicates that the main natural divisions of the genus correspond with its biological distribution, and that the different sections into which the species are classified by systematists bear a certain definite relationship to one another. Only a comparative examination of the chief

characters of both gametophyte and sporophyte generation can show which particular characters are more fixed than others, and so indicate what is the nature of this interrelationship between the sections of the genus. The Young Plant. In all the species belonging to the two sections Inundata and Cernua whose embryogeny is known the embryo plant passes through a protocorm stage. Goebel has described this in L. inundatum, and Treub in L. cernuum. In two previous papers (15, 16) I have described the protocorm of L. laterale, and in the second of these papers I have given a fairly complete account of it not only in this species, but also in L. cernuum and L. ramulosum. Chamberlain (7) has also published an account of the protocorm of L. laterale. In Part II of these Studies (17) I have demonstrated the presence of a protocorm in adventitiously developed plants of L. cernuum and L. ramulosum, and have shown that the large protocormous rhizomes of the latter species can branch, give rise to two or even three stems, bud off bulbils, or by partial decay break up into a number of distinct portions, each of which can develop into a young plant. Goebel also has described adventitious protocorais in L. inundatum. There will be no need for me to repeat here the description of this organ as it occurs in the three New Zealand species, beyond pointing out the facts relating to its variability in form. In the case of L. cernuum I have shown (16, pp. 283–84) that usually the growth in size of the protocorm ceases after the first three or four protophylls have been formed, but that in some instances a certain amount of lateral growth takes place, as many as seven protophylls being formed along the top of the extended protocorm before a stem-axis is initiated. Since that account was written I have examined another colony of young plants of L. cernuum, several of which showed very large protocorms quite comparable in size to the largest of those of L. laterale found by me. It is thus apparent that whereas under normal conditions this peculiar structure does not assume a large size in L. cernuum, yet under certain conditions, occasioned probably by a dry season, it may function for a much longer period and assume a much larger size. In both L. laterale and L. ramulosum the protocorm is also very variable in size, although it is much larger than it usually is in L. cernuum. In the paper quoted above (16, pp. 277–83) I have set myself to show that the large size of this organ in L. laterale, and also in L. ramulosum, is merely a special adaptation suited to carry the young plant over the dry season. In all these New Zealand species which possess the delicate short-lived type of prothallus the young plants are formed during the spring and early summer. During the dry months of midsummer probably the majority of them die, except those in exceptionally favourable situations, or those whose protocormous rhizome has become sufficiently tuberous to be able to withstand a period of drought. Then, as the wet season comes round again, growth is resumed and a stem-axis is initiated. According to this view the protocorm is a plastic organ whose form and importance is mainly dependent upon external conditions. This also suggests that the protocorm in general, as it occurs throughout the sections Inundata and Cernua, is a physiological specialization. This is the view taken by Bower (3, pp. 225, 355). Goebel also (14) was not able to see any phylogenetic significance in this organ. Thus there are certain consistently present characters in these two sections which indicate that they together form a natural division of the genus

clearly marked off from the remainder. These characters are—a plagiotropic habit of growth with all-round branching of the trailing stem; a mixed type of stelar anatomy with exceedingly broad protoxylem groups; a delicate short-lived surface-growing prothallus; and a protocorm stage in the embryogeny. It may safely be assumed that a type of prothallus which possesses chlorophyll and is largely self-nourishing is less modified from the original type than are those which are wholly saprophytic in mode of nutrition. This assumption is made by most writers on the subject. It is not, however, necessary to assume that the delicate nature of the cernuum type of prothallus is also a primitive character. Such a type of prothallus demands a damp habitat. Another consequence of the delicate nature of the prothallus is that it becomes necessary for the young plant to quickly gain independence and establish itself, and hence probably arose the protocorm as a physiological specialization. Such damp situations as are suitable for the development of this type of prothallus are liable to seasonal periods of drought, and hence the protocorm is further developed as a resting tuber. The mixed type of stelar anatomy is initiated by the fact that in the very young plant the leaf-trace system precedes the formation of the cauline cylinder, and thus the stelar tissues show from the first a loosely aggregated character. Moreover, in the older stems the branching of the stele in all four directions tends to continually disturb the tissues. However, this is not sufficient to account for all the peculiar features in this type of stele, which would seem rather to be an expression of the inherited constitution of this division of the genus. The plagiotropic habit and unlimited growth of the plant are a modification suited to assist it to spread over wide areas, and are able to counterbalance as means of vegetative propagation the uncertainties attached to the sexual reproduction of these species. Thus all these characters are largely dependent upon the environment, being under its direct influence, and with respect to them the different species show a high degree of plasticity. Species belonging to the Clavata Section. I have already enumerated certain reasons for removing the species L. densum, L. volubile, and possibly also L. obscurum and L. casuarinoides, from Pritzel's Cernua section, and placing them instead in the Clavata section, and for removing the species L. diffusum and L. ramulosum from Pritzel's Clavata section and placing them in the Cernua section. Such a rearrangement of these species would seem to be in accord with the facts known concerning their main gametophytic and sporophytic characters. This section is marked by a number of very definite characters. The habit of growth of the species is strongly plagiotropic—indeed, so much so that several show a bilateral structure in the leaf-arrangement, with heterophylly, and, at least in the New Zealand species, the branching is always in the plane of the ground, and not all round the stem as in the species of the last section. The fertile regions are upright club-shaped cones borne on distinct pedicels, or numerous short cones at the tips of the densely-branched aerial branches, except in the case of L. volubile, where the scrambling habit of the plant has probably been accountable for a noteworthy modification. The stelar anatomy shows a characteristic dorsiventral disposition of parallel plates of xylem and phloem. The prothallus is large, compact, and deeply buried, and the young plant possesses a strongly developed foot. These characters clearly mark off the Clavata section from the other sections of the genus. It is, of course, obvious

that these same characters are under the direct influence of the environment, even more so than in any of the other sections, and therefore we may expect to find not only variations in form and structure in the species themselves, but also indications that the section itself includes more than one strain. This seems to follow, for example, from the fact that there are two types of prothallus, and that, in the case of two of the New Zealand species, along with these types of prothallus there go two distinct types both of heterophylly and of strobilus-form. External Form of Plant, and Nature of Strobilus. L. volubile has a scrambling and climbing habit, and its form and structure show various modifications in accordance with this. The following quotation is from an unpublished work by L. Cockayne on the Vegetation of New Zealand: “L. volubile is an interesting example of the transition of a creeping ground-plant through a winding liane, by way of a scrambling plant. This plant, as a creeper, in many places extends its slender, woody, stiff stems far and wide, rooting in the soil and raising them unsupported for 60 cm. or so into the air. Thus a prop may be gained, and, this happening, the method of climbing depends upon the nature of such. If the support be twiggy the liane merely scrambles through the branches, its lateral branchlets at about a right angle to the axis, the sporophyll-bearing branches and the hook leaves of the stem all functioning as climbing-organs, the last-named also aided by the flexuous stems. When the support, however, is smooth and with few projections the stem of the Lycopodium twines strongly, gripping the support tightly.” In localities where this species is growing in luxuriant masses upon and amongst low vegetation, one frequently sees the growing shoots standing erect and unsupported to a height of as much as 3 ft. to 4 ft, the flexible ends of the shoots being curved similarly to those of a hop These shoots are thus able to lift themselves up in a vertical direction and so reach the lowest branches of small trees. Sometimes two or even three neighbouring shoots in such situations will twine tightly round each other, and so lift themselves up into the air to an even greater height. Although somewhat slender, the stems and branches are exceedingly rigid and strong, owing to the fact that the whole of the cortical tissue of the stem is strongly selerenchymatous. When growing on the ground the characteristic foliage of L. volubile is luxuriantly developed, such specimens being commonly used for decorative purposes. When climbing, however, this foliage is generally very poorly developed, and frequently is entirely wanting, the branches being present simply as short peg-like projections scantily clothed with the acicular scale leaves. The hooks are present on all rapidly elongating stems and branches that are bare of the usual dorsiventral foliage. The hook is formed by a short downward prolongation of the base of the ordinary acicular scale leaf. They point backwards from the elongating region of the stem, and, when the stem is scrambling or twining, serve to take the weight of its lower parts. I have elsewhere (15, p. 366) described the characteristic heterophylly of this species. Not infrequently in sheltered places throughout Westland I have come across mature plants in which the dorsal and ventral scale leaves, which are normally very reduced in size, are longer and more spreading, the laterally spreading large leaves at the same time tending to the acicular form. This is of the nature of a reversion foliage, and such plants have an exceedingly beautiful, feathery appearance. When the plants

are scrambling over scrubby vegetation the adventitious roots are sometimes of great length. They may be as much as 3 ft. to 5 ft. in length, and a very characteristic feature during the wet season is the thick envelope of mucilage which covers from 3 in. to 12 in. of the growing root-tip before it reaches the ground. This mucilage envelope also occurs to a much less extent on the exposed root-tips of such other species as L. cernuum, L. densum, L. scariosum, and L. fastigiatum. The spikes are of a very characteristic form, which may be regarded as a direct adaptation in accordance with the scrambling habit of the plants. They are produced only when the plant can raise its lateral branches to a considerable height, and then the ultimate twigs on the entire terminal portion of such a branch become developed as long cylindrical pendulous strobili. Plate XII, fig. 2, A, shows a portion of such a fertile branch, at the point where the ordinary foliage passes into the fertile region. The terminal portion of the branch has been broken off at the point marked with a cross. Frequently on such branches fertile regions are intermixed with sterile, and vice versa, as also will be clearly seen in the figure A close comparison may be instituted between the fertile region of this species and that of the pendulous epiphytic species of the Phlegmaria section, the obvious deduction being that in both cases the form of the strobilus is simply the result of the pendulous habit, and that the Lycopodium strobilus must be in a very plastic condition for this habit to appear in the Clavata section. The main stem of L. densum is subterranean, but sometimes—as, for example, at the edge of a bank—it emerges from the ground and continues its growth, becoming green in colour. The aerial branches arise right and left of the main rhizomes, and at once bend sharply upwards, emerging from the ground with a stiff, erect, dendroid habit. There are three very distinct forms of aerial branches, which take their character from the nature of their foliage. These are shown in Plate XIII, at A, B, and C, that lettered B showing its own particular form of fohage in the upper portion of the branch and the foliage of form A in the lower portion. These three varieties of foliage are almost invariably quite distinct from one another, and do not grade into one another, being confined for the most part to separate branches. Where more than one variety does occur on the one branch, as in the specimen figured, the two forms nevertheless keep quite distinct. Now, all three varieties of foliage are commonly to be met with in practically every locality where L. densum occurs, growing side by side, so that it is obvious that the different forms are not by way of being adaptations to the environment. It would seem, then, that they are hereditary polymorphs, true-breeding races which possibly have arisen by mutation, and which hybridize. These three forms are by no means plastic in their nature, although individual twigs may very occasionally be observed which seem to show gradations in their foliage. When growing amongst tangled vegetation, more especially in hollows, the aerial branches may attain a great height. They are then very scantily branched; indeed, I have seen aerial stems as much as 9 ft. in height, some of which have been quite unbranched. The cones are very numerous, are short, and are borne solitarily at the tips of the ultimate branchlets very much after the same manner and appearance as in L. cernuum. Instances are to be met with sometimes in which there has been a vegetative prolongation of the cone, as at the points marked with a cross in Plate XIII, D. In this particular figure the lower of the two fertile regions thus indicated may well have been differentiated after the whole twig had been formed.

A, B, C, L densum: Three varieties of fohage, D, L. densum' Abnormality in the fertile region.

A, L. fastigiatum Fertile aerial branch of mesophytic variety, from lowlands, Westland. B, L. fastigiatum Portion of rhizome with two sterile branches, xerophytic variety from Browning Pass. C, L. Drummondii: Two incomplete plants showing portion of creeping stem with fertile branch.

L. fastigiatum varies very much in the form of its aerial branches in accordance with the environment. Plate XIV, A and B, represent two extreme forms, the latter from Browning Pass at a height of 5,000 ft. and the former from Westland at practically sea-level. The form lettered B represents a portion of a plant as it grows commonly on the hard dense cushions of Phyllachne clavigera. The main rhizome is deeply buried in the rotting substratum of the cushion, and the lateral branches are also rooted and buried, but at a lesser depth. Arising from the lateral branches are the short tufted foliage shoots, which appear as little compact rosettes firmly resting upon the surface of the cushion. The cones are numerous, and stand erect on very short pedicels to a height of ¼ in. to ½ in. The plant figured is sterile. This is an extreme xerophytic form, and is in marked contrast to that lettered A, which represents the species as it occurs typically around the alluvial goldfields in the very wet climate of the lowlands of Westland. In these localities the aerial branches attain a height of 1 ft. to 3 ft., and are often of a striking red or golden colour, handsomely branched, with a terminal bunch of long, club-like cones. On the Canterbury mountains, where this species occurs abundantly, I have often observed that when the plants are growing in a tussock-clump or amidst other sheltering vegetation the aerial branches are tall and slender and open-leaved, whereas those, perhaps of the same plant, growing in the open in the full light of the sun only a foot or two away are short, sturdy, and compact, with much thicker stems and closely imbricating leaves. This species occurs also commonly in subalpine Nothofagus cliffortioides forest, and there its habit is characteristically mesophytic. There is a certain range of variability in the form of the strobilus and its pedicel, even in individuals from the same locality. On some branches the tips of practically all of the ultimate twigs are fertile and there is no pedicel, there being a gradual transition from the vegetative leaves of the twig to the sporophylls. On other branches, again, the main axes of the branch are continued on as long scantily-leaved pedicels, bearing one or more long, club-shaped cones, which are often branched. The former condition is, generally speaking, characteristic of the xerophytic type of branch, and the latter of the mesophytic, although this distinction is not always kept. Among the mesophytic form of plants some very extreme examples of pedicel-formation are sometimes to be met with. As in the case of the last-named species, the wet climate of the lowlands of Westland encourages a luxuriant growth of L. scariosum. Plate XII, fig. 2, B, shows a portion of the rhizome of this species with two small erect lateral branches bearing cones. The development of heterophylly in this species, as in the case of L. volubile, has been elsewhere described by the writer (15, p. 366). In L. volubile the acicular leaves are finer in form than in the other species, and there are about twice as many orthostichies. Heterophylly in L. volubile arises through two orthostichies on each lateral face of the branch approximating to one by being flattened in the plane of the ground, the leaves of these orthostichies assuming the larger form. At the same time the leaves of the two or three orthostichies placed ventrally and also dorsally become much reduced. All stages in the development of this dimorphism can be seen in young plants, and also in the reversion foliage which has already been noted as occurring in this species. In the case of L. scariosum it is the leaves of the two dorsally placed orthostichies which become flattened in the plane of the ground and assume the larger form, there being two to four orthostichies of much reduced scale-like leaves

on the ventral side. The stages in the development of this form of distichous arrangement can also be seen in the young plants, but not so well as in L. volubile, since the heterophylly is developed rapidly and very early. When growing in dry situations—as, for example, at considerable altitudes—the lateral branches of the mature plant of L. scariosum are more or less flattened in the plane of the ground, but when amongst other vegetation they tend to assume a more upright position. This species occurs in wide, luxuriant spreads around the abandoned gold-mining claims throughout north Westland, and there the lateral branches are characteristically erect-growing, although the heterophylly is still strongly marked. Frequently, as a result of the upright habit of growth, the twigs of these branches show an all-round spread, and even the flattened falcate leaves are more outstanding, but inspection shows that both twigs and leaves arise in the characteristic dorsiventral manner. The long club-shaped cones are borne in great numbers, being raised up on pedicels 3 in. to 6 in. in height, which are formed by the continued growth of the main twigs of the branch. The appearance of these pedicels previous to the formation of the cones is very striking, standing stiffly up as they do in massed numbers above the dense foliage. The gradual transition from heterophylly to the spiral arrangement of ordinary acicular leaves can be well seen at the base of these pedicels. When growing amongst ferns and saplings at the edge of the forest, the lateral branches of this species sometimes become very long-drawn-out and adopt a semi-scrambling habit. Owing to the rapid growth the scale leaves are somewhat scattered along these branches, and in those regions in which the characteristic heterophylly has appeared its development has taken place irregularly. Instances are not infrequently met with of internodes showing, contrary to the usual rule, one, two, or even three dorsally placed orthostichies of scale leaves, the type of heterophylly then approximating to that of L. volubile. Stem-anatomy. As in the case of the two other natural divisions of the genus, the Clavata section shows a very characteristic and consistent type of stelar anatomy. This has been described by Jones (21) in the six species L. clavatum, L. annotinum, L. complanatum, L. chamaecyparissus, L. alpinum, and L. obscurum; by Boodle (2) in L. volubile; and by myself (15, 16) in the four New Zealand species which belong to this section. The central cylinder consists throughout of parallel plates of alternating xylem and phloem disposed dorsiventrally, surrounded by a pericycle. Consequent on the comparatively large size of the plant in the New Zealand species, the number of plates of vascular tissue is also large. The phloem is markedly differentiated into centrally-placed empty sieve tubes of large size and flanking phloem parenchyma with abundant contents. I have previously described (15) in L. volubile and L. scariosum the development of the dorsiventral structure characteristic of the mature stele from the radial structure characteristic of the young plant, showing that it is initiated in the developing plant at the point of branching, which in this section of the genus always takes place in the plane of the ground, and tends thereafter to be preserved in between the branchings. Assuming that this is the main cause of the dorsiventral structure, it is quite possible that it will be found in species which are not closely related, and thus this character taken by itself will not be a sure sign of natural affinity. Owing to the great plasticity of the genus Lycopodium as seen in all its main

characters, the natural classification of the species must take all these characters into consideration, weighing one against another, and can only become an accomplished thing when the full life-history of many more of the more important species is known. This is, of course, true in a greater or less degree of every genus and family of plants, but it is especially true of the present one. The stelar and cortical anatomy is very consistent in structure as it occurs in the case of each of the four New Zealand species, but certain small distinguishing features are worth recording. The text-figures illustrating the vascular structure of the four species show in each case the ends of two bands of both xylem and phloem with the protoxylem and protophloem groups, and also the adjacent pericyclic and cortical regions. As in the case of the other New Zealand species, semi-diagrammatic drawings of the complete stele have been given in another paper (16). In the stem of L. volubile (fig. 13) a very characteristic feature is the exceedingly large size of the sieve tubes. The chief metaxylem elements are also large, and are noticeably flanked on both sides of each band by small-sized tracheides. The pericycle is narrower than in the other species. Practically the whole of the cortex is strongly scleren-chymatous. Three zones can be recognized, however—an inner narrow, exceedingly strongly thickened zone, in which the cells have been flattened tangentially by mutual pressure and the cell-cavities almost obliterated; a broad median zone of circular, thick-walled cells, which passes into a narrow epidermal zone of thin-walled cells showing air-spaces and bounded externally by a strongly cuticularized epidermis. The whole stem and the vascular cylinder is much smaller in size than in the case of the other three species. The main features thus given of the stem-anatomy of L. volubile—viz., the comparatively thin stem, the woody cortex designed rather to impart a wiry quality to the stem than to act as a starch-storing tissue, and also the large open vascular elements—can be put into connection with the scrambling habit of the plant. In the large adventitious roots the configuration of the stele is strikingly stellate (16, fig. 93). There is no modification of the vascular tissues in the ultimate branchlets by the heterophyllous habit. The rhizome of L. densum is from two to three times as thick as that of L. volubile, and the vascular cylinder is very large. The cortex functions as a storing tissue. There is a somewhat narrow, strongly sclerenchy-matous, mner cortical zone, while the rest of the cortex consists of much less thickened circular cells containing abundant starch. In the aerial stems the whole of the cortex is strongly sclerenchymatous. The same rhizomes which show the presence of starch in the cortex show also the cells of the phloem parenchyma to be practically empty, but immediately behind the Fig. 13. — Lycopodium volubile. Transverse section of portion of stele of main stem. x 137.

rhizome-apex they have abundant contents. The sieve tubes are very much smaller than in L. volubile. There is a much scantier development of small-sized flanking xylem elements, but the main metaxylem tracheides are large (fig. 14). The dorsiventral structure of the main rhizome stele passes over into the lower regions of the aerial stems, but this is more or less quickly replaced there by the stellate or radial form. The largest aerial branches not infrequently possess towards their base a larger stele than do the main rhizomes. At the base of one such stem I found as many as twenty-one protoxylem groups, whereas the number in a full-grown rhizome is generally about seventeen or eighteen. In this connection reference may be made to Jones's remark with regard to the vascular structure in the stem of L. obscurum (21, p. 32). He says, “In any hypothesis which endeavours to explain the dorsiventral structure in the creeping stems the fact must not be overlooked that the banded structure is well marked in the stem of the erect-growing species of L. obscurum.” There is no doubt that L. obscurum, as also L. densum, must be removed from Pritzel's Cernua section and placed in the Clavata section. L. densum is not an erect-growing plant, as stated by Pritzel, but is plagiotropic, and it is practically certain, judging from the sporeling plants which I have found, that the prothallus belongs to one or other of the deep-growing, massive types. Mr. Cheeseman has informed me that in both Bretton's and Asa Gray's manuals of the flora of the United States of America L. obscurum is described as possessing subterranean creeping “rootstocks.” Moreover, Spessard (26) has found the prothallus of this latter species to be of the L. clavatum type. Even if the material of L. obscurum examined by Jones was not that of the main creeping stem but of the lateral branch, the fact of its dorsiventral structure would not be significant, since, as I have shown in L. densum, this frequently persists for some distance up a branch after having been carried over from the main stem at the point of forking. The main subterranean rhizome of L. fastigiatum is much shorter than that of the other three species, and is more slender than that of L. densum or L. scariosum. Two characteristic features of its vascular tissues are the small size of the xylem elements and the sieve tubes, and also the wide double-zoned pericycle. The nature of the pericycle in this species will be seen from fig. 15. It consists of two distinct zones, the inner of which is from four to five cells wide, and shows abundant contents, and stains a reddish-brown with safranin, and an outer zone three to four cells wide, also showing abundant contents, but staining purple with the haematoxyn Evidently this wide pericycle has been developed as a storage tissue, and the stored substance in both zones appears to be starch, although the reason Fig. 14. — Lycopodium densum. Transverse section of portion of stele of main rhizome. x 137.

for the different staining-qualities of these two zones is not clear. The cortex consists of an inner, narrow, very strongly sclerenchymatous region of exactly the same nature as in the rhizome of L. densum, and a broad median zone of much less thickened cells which are abundantly stored with starch. The outer region of the cortex in this species is peculiar. There is a layer of small-celled compact parenchyma, three to four cells in width, and bounded externally by a strongly cuticularized epidermis which is wholly without stomata, and this is separated from the median rather thick-walled cortical region by a layer of enormously large thin-walled empty parenchyma cells, two or three cells in width. It is rather difficult to say what is the significance of this peculiar wide-celled zone. Possibly it functions as a water-storing tissue. If the rhizome had been epigaeous it might have been interpreted as an aeration tissue. The outermost cortical zone easily tears away in this region, and in dried herbarium specimens of the Browning Pass form I have also noticed the same. Fig. 16.—Lycopodium fastigiatum. Transverse section of portion of stele of main rhizome. × 137. Fig. 16.—Lycopodium scariosum. Transverse section of portion of stele of main stem. × 137. The main stems of L. scariosum are stout and epigaeous, and not very widely spreading. In point of actual size the stele is not much larger than that of L. volubile, but it contains twice the amount of vascular tissue, owing to the fact that the xylem and phloem elements are all very much smaller in size than in the latter species. This will be at once apparent from fig. 16, which should be compared with that illustrating the stele of L. volubile. In L. scariosum there is also an entire absence of the small tracheides flanking the large metaxylem elements, which present so characteristic a feature in the stele of L. volubile, and to a smaller extent also in that of the other two species. The number of protoxylem groups in the full-grown stem of L. volubile is from ten to sixteen, but in that of L. scariosum it is from eighteen to twenty-seven. The stout size of the rhizome is due to the very large cortex, which consists of an almost uniform tissue of not very strongly thickened sclerenchyma. There is no inner strongly sclerenchymatous zone such as exists in the other three species. I did not observe starch in the cortical tissues, although this may possibly be present at certain seasons

of the year. The outer zone of the cortex is noteworthy, and may be compared with that in the rhizome of L. fastigiatum. It consists of a rather loose tissue of small roundish thin-walled parenchyma, three to five cells in width, showing air-spaces, and bounded externally by an epidermis which is only slightly cuticularized. There are abundant stomata present. In some places this aerating tissue seems to be in process of breaking down, while there are quite lengthy stretches in which there is seen to be the same development of enormously large irregular-shaped cells separating the outer from the main cortical zones, as also occurs in L. fastigiatum. It is possible that in L. scariosum this is an adaptation for aeration purposes, and that in L. fastigiatum is still persists although the rhizome has become subterranean. The subterranean habit has probably arisen as a modification from the epigaeous trailing stem, and there is no doubt that it is not an absolutely fixed character, as is seen from the case of L. densum. However, a simpler explanation is that in both L. scariosum and L. fastigiatum the peculiar large-celled tissue functions as a water-reservoir. As in the other three species, the aerial branches show a stellate configuration, and there is no modification of the stele in the strobili or in the heterophyllous regions of the plant. There is thus seen to be a very characteristic form of stele in all the species which belong to the section Clavata. This type of stele gives evidence of a greater degree of modification than do the other two main types which occur in the genus. The habit and external form of these species also present the greatest degrees of modification, and it is obvious that the stelar anatomy is in direct correspondence with the habit. The four New Zealand species show some small but interesting peculiarities which can all be best explained by reference to the environmental factors. The stem-anatomy of this section of the genus is thus seen to be in a plastic condition, and some very interesting experiments could be conducted (though perhaps with difficulty) to determine whether or not these modifications are fixed characters. Whether or not the dorsiventral type of stele covers but a single cycle of natural affinity in the section can only be determined by a full study of the other main characters of the species concerned. The Prothallus. The prothallus throughout this section is large, compact, long-lived, and more or less deeply buried, and in its wholly saprophytic habit is clearly the most modified of all the Lycopodium types. There are two main types of this form of prothallus—viz., the clavatum type and the complanatum type; and of the seven species of this section whose prothalli are known five conform to the clavatum type and two to the complanatum type. Bruchmann (5, 6) has described the prothalli of L. clavatum, L. annotinum, and L. complanatum, and Lang (24) that of L. clavatum. Spessard (25, 26) has recently given a preliminary account of the prothalli of L. clavatum, L. complanatum, L. annotinum, and L. obscurum as they occur in America. The prothalli of the New Zealand species L. volubile and L. scariosum have been described by Miss Edgerley (12), by Chamberlain (7), and by myself (15, 16). I have also described (16) the prothallus of the other New Zealand species, L. fastigiatum. These types of deeply buried prothalli, living under consistent conditions, do not show in any individual species such a range of variability as do those forms of prothallus which can reach the surface. In fact, the main variations from the normal which I have met with are related to the actual

bulk of the prothallus or to the fact that, as in L. volubile, it occasionally grows at the surface of the ground and develops chlorophyll. The different species, however, show certain characters, more especially in the structure of the fungal zones, in which they differ from the other species of the same type. These might be viewed as variations from the two main stocks represented by L. clavatum and L. complanatum, or they may be taken as indicating that the adaptation of the prothallus to the subterranean, wholly saprophytic mode of life has occurred independently in a number of species. A comparison of Bruchmann's account of the prothallus of L. clavatum with that of Lang shows that, although in the main the two descriptions are very similar, there are yet certain particulara in which they differ. These refer more particularly to the main fungal and store tissues. It will be worth while to notice these details, in view of the fact that, as will be seen below, some of the distinguishing details in the two New Zealand species are remarkably characteristic. Some of the differences manifest between Bruchmann's and Lang's accounts and figures are to be explained by the fact that the longitudinal sections which the latter figures and describes are not median. Lang himself in a footnote mentions that the first-formed conical projection at the base of the prothallus was not at first noticed by him and hence does not appear in his figures. On account of this, the layer of flat, elongated cells belonging to the central core, which Bruchmann describes as lying next to the starch-containing store tissue where the latter abuts on to the centrally placed core of large-celled empty tissue, does not appear in Lang's figures and is not mentioned by him, for in the sections on which he bases his description it was, of course, cut transversely. Also, Bruchmann's figures show a very much thicker subepidermal cortical fungal zone than do Lang's—at least, towards the upper half of the prothallus. Again, Bruch-mann speaks of the starch-containing store cells as polyhedral in shape, but Lang says that they are not infrequently elongated in the same direction as those of the palisade layer. Bruchmann figures the palisade layer as a very clearly marked one in which the single row of palisade cells is very much elongated in a direction at right angles to the surface, but in Lang's figures the cells are much shorter and are sometimes in two rows. Now, all these differences can be accounted for by the fact that Lang's sections, not being exactly median, have cut the various cell-layers somewhat obliquely. These differences in the two accounts, therefore, must not be taken to show that the structure of the prothallus of L. clavatum is in a variable condition. A more important difference between the two accounts lies in the fact that Lang describes the presence of the fungus in the starch-containing store tissue as penetrating not within the cell-cavities but in the cell-walls, whereas Bruchmann says quite definitely that it stops short with the palisade layer, and in his figures he shows it to be altogether absent from the store tissue. Is it possible that the fungus enters this tissue in rather old prothalli only, and that this must be held to account for the difference? Lang's description in this particular agrees with my own observations on the prothallus of L. volubile and L. fastigiatum. There is no need for me here to enter into any of the other details of the prothallial tissues. In the prothallus of L. volubile the most characteristic feature is the enormous development of the palisade layer. This layer constitutes generally by far the greater bulk of the whole prothallus. The cortical fungus layer corresponds with that of L. clavatum except that in the New Zealand species it is only from two to four cells in width. In a median

longitudinal section the elongated cells of the palisade layer are seen to be from four to six deep. They overlap one another, and, contrary to what Lang has described in L. clavatum, they show the presence of dense fungal coils in all the cells, as well as the mycelium which inhabits either the cell-wall itself or the spaces between the cells. At its interior boundary the cells of the palisade layer are shorter in form, and where they abut on to the upper region of empty parenchyma cells the cells of the latter tissue are much flattened in form. Miss Edgerley, in her account of this prothallus, (12, p. 96) says that “the whole of the central part of the prothallium is occupied by large thin-walled parenchymatous storage cells, in which starch is stored in great abundance. The cells of this tissue bordering on the palisade layer are often smaller and more densely filled with starch-granules.” I found no starch in any of the older or the young prothalli which I sectioned. Possibly this difference is to be accounted for by a difference in the season of the year in which Miss Edgerley's and my own prothalli were collected. The upper fungus-free region of the prothallus, which bears the sexual organs, is in full-grown specimens very irregularly developed into large rounded protuberances, which overlap the original saucer-shaped upper surface. The fungal zones extend into such of these protuberances as lie towards the rim of the upper surface. In such cases as these in some longitudinal tangential sections the prothallial epidermis and the cortical fungal zone extend almost entirely around the section, the whole of the interior being occupied by palisade tissue, while a very small upper corner of the section consists of the upper non-fungal tissue. The youngest prothalli sectioned consisted almost entirely of fungal tissues, more especially of the palisade layer, with a very feeble development of the upper non-fungal zone. There is no store tissue containing starch either in the young or mature prothalli such as Bruchmann figures in L. clavatum—in fact, the only starch which I found present consisted of a few scattered grains is some of the palisade cells. One very large prothallus, which grew at the surface of the ground, and whose upper region was vivid green in colour, showed a considerable divergence from the normal structure of the prothallus as just described. The palisade layer was four to five cells in thickness, but by far the major bulk of the prothallus consisted of the upper non-fungus-containing parenchyma. Moreover, this contained a large accumulation of starch both in the layer immediately underlying the small-celled, generative region and also around the meristem. The starch was also thickly accumulated around the “foot” of a developing plantlet. In the palisade cells there was an abundant accumulation of oil-globules and starch, and the epithelial cells of the “foot” of the young plant were densely crowded with contents. I have several times come across these green, surface-growing prothalli. The nature of the fungal zones in the prothallus of L. fastigiatum is identical with that of L. volubile. The most striking feature in it is the enormous development of the palisade layer. In some of the largest prothalli sectioned lobing of the upper surface had taken place and the fungal zones had extended into these lobes, giving to the sections a very irregular appearance. Also, in the large prothalli the lip sometimes greatly grows over inwards, with the result that the fungal zones become somewhat superimposed upon one another, as has been described above in L. volubile. The youngest prothalli found by me were of two forms. The first were of the usual saucer shape, showing the first-formed conical projection on the lower surface. Those of the second form were dichotomously forked, as I

have previously described (16, pp. 271–73), showing the two equally developed cylindrical branches inclined to one another at an acute angle, the point of the angle being the first-formed part of the prothallus. The usual fungal zones extend up the outside of the two branches, while their opposed faces, which are in point of fact the upper surfaces of the branches, consist of the empty large-celled parenchyma. There is a very marked groove at the extreme end of each branch, where the meristem is situated. This lies somewhat towards the inner side, so that the antheridia, which are present in large numbers, extend partly down the inner side of the arms. I have not up to the present been able to trace the next stages in the growth of the prothallus from this branched form, but enough has been said to show that it constitutes a very interesting modification of the usual compact saucer-shaped form, and possibly provides a transition between it and the branched epiphytic type in which the meristem is confined to the ends of the branches. One other variation from the normal must be described. I sectioned a fairly young prothallus of the saucer form, which showed absolutely no fungal infection or differentiation of its tissues into zones. The narrow outer layer of cells around the outer portion of the prothallus had collapsed, but I could see no traces of fungus in the collapsed cells. The rest of the prothallus consisted of quite uniform large cells of the usual parenchyma form. There was an abundant development of antheridia around the entire lip. I cannot now tell whether the prothallus was situated at the surface of the ground or not. It is, of course, possible that the whole of the fungal zones had been eaten away uniformly by an insect while the prothallus was still growing. The fact that the inner boundary of the palisade layer in this type of prothallus is always exceedingly well marked and even lends itself to such an explanation. I considered this peculiar instance worthy of record, not in order to include it here as a striking example of the plasticity of the L. fastigiatum prothallus, but so that it might be compared with any other instance of a fungusless Lycopodium prothallus, should such ever be found. The other type of prothallus in this section is that of L. complanatum and of the New Zealand species L. scariosum. The following details of the structure of the former will be given, so that a comparison may be instituted with the New Zealand species. The prothallus is more or less carrot-shaped in form, with its lower region tapering even when fully grown, so that the zones of tissue in the vegetative region are much narrower than they are in the prothallus of the clavatum type. However, as a result of its greater length these tissues about equal in amount those of the latter type. The central core of cells is narrow, and the cells themselves are poor in contents and are elongated in the direction of the prothallus axis. The palisade tissue consists of very narrow, much-elongated cells, which show their greatest length in the thickest part of the prothallus. This zone, which is only one cell deep, serves as a store for reserve substances, there being no store tissue corresponding to that which is found in L. clavatum. In the lower region of the prothallus the palisade cells are empty, but higher up they are full. The fungal hyphae are found only in the cell-walls of this zone, and they sometimes form between the cells large oval spheromes. There is the usual fungal cortical tissue. This prothallus sometimes produces one or more secondary prothalli as shoots on the primary prothallus. With regard to the prothallus of L. scariosum, Miss Edgerley says that the palisade layer is several cells wide; but I find that, whereas in sections

which are not median the palisade tissue presents this appearance, in sections where it is cut exactly longitudinally it is found to be only one cell thick. These cells are exceedingly long and narrow, and occasionally divide transversely. Again, Miss Edgerley describes the cell-layer of the central core which abuts on to the palisade tissue as containing abundant starch and acting as a storage tissue. I was not able to find this in any of my sections, but, as in the case of the prothallus of L. volubile, this may be due to the fact that our specimens were collected at different seasons. On the whole, the main feature in which the prothallus of L. scariosum differs from that of L. complanatum is in the large, irregular size attained by the mature individual. Young or even half-grown prothalli sometimes show the tapering carrot form, but sooner or later this becomes modified owing to the enormous development of the centrally placed core-cells, which evidently must function as a store tissue. However, even very young prothalli sometimes show a rounded lower vegetative region instead of the tapering one, the first-formed conical region being in these cases quite blunt. This difference, then, is to be referred to the manner of development of the prothallus from the beginning. Whereas in L. complanatum the increase in girth is very gradual, in L. scariosum it is generally rapid from the actual point upwards. This may be occasioned possibly by the early development of the storage function of the central core of cells, or it may be due to a deeper constitutional difference between the two types. In this connection, however, it is significant to note that the type of heterophylly in both species is alike. In Part I of these Studies (16, figs. 49–52) I have figured several very irregularly grown large prothalli. As Miss Edgerley notes, the most striking feature in the large prothallus of this species is the relatively small proportion of the bulk of the vegetative region which is occupied by the fungal tissues, a feature in which it is strikingly different from that of L. volubile and L. fastigiatum. It will thus be seen from a comparison of the prothalli of the species mentioned that as regards their fungal zones the two types are not very dissimilar. There are two main points of difference to be noticed: first, the increase of girth of the clavatum type from the original point upwards is more rapid, and hence the height of the prothallus is less than in the case of the complanatum type; and, secondly, the extent of the upper generative tissue in L. clavatum is greater than in the latter species, these two features being closely interdependent. The fact that in L. clavatum the upper generative region is more or less saucer-shaped at its surface, with a distinct rim, and is often accompanied in large specimens by a development of warty protuberances, whereas in L. complanatum it takes the form of a compact semicircular crown, would seem to be simply the physical result of the two types of growth. Lang suggests that “the flattened and still more the trough-like form which these older prothalli present may be an adaptation to facilitate fertilization” (24, p. 296). There is no doubt that this particular shape does serve this purpose, but I would be inclined to explain the difference in form between the two types in rather a different way. I have found that the prothalli of L. scariosum always occur at a greater depth than those of L. volubile and L. fastigiatum. In localities where the prothallus of L. scariosum and L. fastigiatum were growing together I always unearthed the latter from the layer of humus which immediately underlay the carpet of moss, &c., whereas the former had to be dug out of the deeper-lying clay. Not infrequently, too, I have found the prothalli both of L. volubile and of L. fastigiatum amongst the thick

moss itself, where they show no adhering earthy matter at all, or even, as in the case of the former species, growing on the surface of the soil with the upper region of the prothallus exposed to the light, being then in their upper region a vivid green in colour. In Part I of these Studies (16, p. 262) I noted that the prothalli of L. scariosum often lie at a depth of 8 cm. to 10 cm. I have not been able to find in Bruchmann's papers any reference to the comparative depth at which he found the prothalli of the three European species. I would suggest that the two different types are an expression of the fact that the species of prothalli which conform to them grow respectively in a shallower and in a deeper stratum of soil, the deepergrowing prothallus having more thoroughly departed from the self-nourishing chorophyll habit. The fact that there are two such distinct types of prothallus in the Clavata section, and that along with these there go two distinct types of heterophylly, may perhaps show that more than one distinct strain of evolution is included in this section. The Young Plant. Heterophylly. — The development of heterophylly in the seedling of L. volubile and L. scariosum opens up the question as to what is the nature of the stimulus which calls forth this character. That it is an extreme instance of adaptation is beyond doubt. Heterophylly is present in the genus Selaginella, but it would be difficult to take the view that in the Lycopodium genus it is a phylogenetic character. Thus modern systematists have abolished Baker's subgenus Diphasium, in which he groups together the heterophyllous species, and have distributed these species according to their more natural affinities. Heterophylly is an adaptation which has appeared also in the section Inundata. What has been the particular external stimulus to call forth this character ? It is to be noted that heterophylly is not found in any species whose main stem is subterranean, and whose lateral branches accordingly emerge from the ground in a stiffly erect and tree-like form. L. densum and L. fastigiatum possess this latter habit. Their erect, tree-like aerial branches are obviously stimulated by the light in no one direction more than in any other. Certainly the aerial branches of L. scariosum frequently grow erect in the lowlands of Westland, as described above, and still possess the characteristic heterophylly, but this is simply on account of the luxuriant massed development of the plants in these localities. And, again, the species L. Jussiaei Desv of South America, which is joined by both Pritzel and Baker with L. scariosum, shows more or less erect-growing branches along with the heterophyllous habit. But in both these species the flattened nature of the branching is preserved even in the more erect-growing branches, the case being very different with the dendroid, fastigiately branched shoots of L. densum and L. fastigiatum. Heterophylly, then, is an adaptation which goes hand in hand with the more or less flattened habit of the branches in which they are somewhat spread out in the plane of the surface over which they are growing. The particular stimulus to the development of the large leaves and to their dorsiventral arrangement is probably that simply of the light, which, falling more or less at right angles upon the branches, stimulates the leaves, which are tiny and needle-shaped and catch very little sunlight, to place themselves in such a position and to acquire such a form that they can intercept the maximum amount. In both L. volubile and L. scariosum the large-shaped leaves are not naturally borne in the lateral position but adapt themselves to it. In the case of L. volubile two neighbouring

orthostichies on either side approximate to one, and in L. scariosum two dorsal orthostichies bend over one to either side. In both species the ventrally borne leaves, which catch no direct sunlight, become mere scales, and in L. volubile they become also very scattered and few in number. In sheltered, shady places plants of L. volubile are found which show the reversion foliage. Here no direct sunlight falls on the branches, the light being diffuse, and the stimulus to a dorsiventral disposition of the leaves is almost completely lacking. The equal all-round-the-stem stimulus exerted by the diffuse light results in the tendency to a more all-round development of the leaves; or would it be more correct to say that the inherited constitution of the plant has the opportunity to assert itself over the acquired character? Which is the correct way to express it depends, of course, on whether or not the heterophyllous character which originally came about as an epharmonic adaptation has altered the hereditable constitution of the species. Some botanists, of course, are quite ready to believe that such a thing is possible, and that this takes place in nature more readily than is generally imagined, while others would hold that an epharmonic adaptation must inevitably revert as soon as the controlling stimulus is removed. In the young plants of L. volubile and L. scariosum the characteristic heterophylly appears while they are still erect in growth. In. L. scariosum it appears almost from the very first, there being but a few scattered scale leaves formed before the characteristic dimorphism is in evidence, and there being practically no transition stages. In L. volubile the dimorphism of the leaves appears first when the plant has attained, compared with L. scariosum, a considerable size (16, figs. 97, 98). It appears first in some particular branchlet or other, and develops its characteristic appearance in gradual stages, so that a plantlet possessing six to ten branchlets will show probably every stage in the development. Since the heterophylly appears in the plantlet before it has adopted the plagiotropic habit, it would seem that, after all, this character has actually become fixed, and that this has become so to a greater degree in L. scariosum than in L. volubile, for it appears there much earlier in the ontogeny. I have not actually observed whether or not the young plants have their dorsiventral branches turned at right angles to the direction of the light, but I should judge that, if such were the case, heterophylly ought to begin in all the branches at the same time. However, the plantlets grow generally amongst thick moss and other vegetation where they get no direct sunlight at all. The Plagiotropic Habit.—The sporelings of L. volubile, L. fastigiatum, and L. scariosum maintain an erect growth for a much longer period than do those of the species which belong to the Inundata and Cernua sections. The upright stems which arise from the protocormous rhizomes in L. cernuum,L. ramulosum, and L. laterale, and the stems of the vegetatively produced plantlets of L. Drummondii, almost immediately bend over and flatten themselves in the plane of the ground, and a strong adventitious root emerges at right angles from the stem and binds the latter to the ground. Whether or not it is the early development of the first adventitious root which compels the young plant to so soon adopt the plagiotropic habit, or whether rather it is the strong plagiotropic habit which determines that even the first adventitious root shall emerge at right angles from the stem at the point at which it is given off from the vascular cylinder, and that it shall not penetrate down the tissues of the cortex as it does in L. Selago, I am not able to say. The fact remains, however, that the species which belong to the Inundata and Cernua sections characteristically differ from those of

the Selago and Phlegmaria sections in this respect, for in the Selago section, although roots do emerge at different points on the lower more or less sprawling adult stem, yet in the young plants they emerge at the base, and throughout the life of the plants they may be seen in transverse sections traversing longitudinally the cortical tissues of the stem. In the young plants of L. volubile, L. fastigiatum, L. densum, and L. scariosum, which belong to the Clavata section, there is a very marked erect stage in the ontogeny. Here the prothallus is subterranean, whereas in the Inundata and Cernua sections it is surface-growing, and thus necessarily the stem is erect before it can even develop its foliage. However, I have often observed that the naked subterranean stems on the deeply growing prothalli of L. scariosum may be straggling and bent; but this will probably be the result merely of the presence of stones, &c., in the soil around which the stem has to find its way. Thus in these plants the stem-axis emerges perpendicularly out of the soil. However, contrary to what takes place in the young plant of the Inundata and Cernua sections, the stem continues to grow erect. A young plantlet of L. densum found by me was no less than 4½ in. in height, being in every respect truly erect. Besides the first-formed root there was present a second, the first “adventitious” root, which was borne on the underground portion of the stem just above the “foot.” The young plants of L. scariosum frequently grow erect and branch to a height of 1½ in. to 2 in., and those of L. volubile and L. fastigiatum to an even greater height. It will thus be seen that in the Clavata section there is in the ontogeny a strongly marked erect stage which precedes the adoption of the plagiotropic habit. The adventitious roots I found in no case to travel down through the cortical tissues of the stem, but they emerge immediately from it at right angles. Miss Wigglesworth (31), however, records an instance in a plantlet of L. complanatum in which she found one rootlet which had travelled for some distance down the cortex of the stem instead of pushing its way directly to the periphery. This must certainly be regarded as an abnormality. In the erect plantlets of all the New Zealand species the adventitious roots commence usually to arise on the subterranean portion of the stem just above the foot, and are also frequently to be seen projecting out at right angles from the aerial region of the stem. There can be no doubt that when these roots do reach the ground they help the plant to bend over and adopt the trailing habit of growth. I have not infrequently seen well-grown erect plantlets of L. volubile, which had branched several times, on which one or more strong, naked adventitious roots an inch or two in length, which had not yet reached the soil, were present immediately behind the tips of the branches, the terminal portion of one or other of the branches having begun to increase in stoutness in anticipation of its greater extension in length. In the Selago section some of the species are stiffly erect, but in others the lower part of the stem is somewhat recumbent. L. Selago itself shows in the different forms in which it occurs in New Zealand both these habits of growth. L. varium also in its smallest forms is more typically erect than in the larger forms. Whatever may have been the extent of growth of the ancestral stock of the modern genus, herbaceous or more tree-like, the erect species that now exist do not possess the capacity of extensive growth. It would seem to be more probable that the modern species have sprung from an herbaceous stock than that from the whole genus the character of secondary growth, whether of vascular or of cortical tissues, has been completely lost without the slightest trace having been left behind.

In the Selago and Phlegmaria sections, owing to the fact that the primitive dichotomous nature of branching is the characteristic one, the plant is strictly limited in size. In the Inundata and Cernua sections both dichotomous and monopodial branching is to be found. The branching of the young plant of L. cernuum, L. laterale, and L. ramulosum is always at first dichotomous. In L. ramulosum this is more or less maintained throughout the life-history, the rosette form which arises as the result of the repeated dichotomies being very characteristic. This species shows an interesting transition from the subaerial to the subterranean habit of growth, the branches in some localities growing down into the soil and functioning as rhizomes. In L. laterale this has taken place to a greater extent, so that there the branches of the plant have become permanently subterranean. However, in this species also the growth of the plant is limited, for sooner or later these rhizomes emerge from the ground as the aerial shoots. The branching of the mature plant of L. Drummondii, however, is monopodial, and the plant itself is wholly above-ground, so that the way is open for it to spread extensively; but it never attains a greater length than 6 in. to 8 in. All the members of the Inundata and Cernua sections are very limited in their extent of growth, except L. cernuum, the replacing in them of the dichotomous by the monopodial habit of growth resulting in very little increase in the size. In L. cernuum, however, the case is different. Here too the branching is dichotomous in the young plant, but the adoption of the monopodial habit leads to the unlimited growth both of the main stem and of some of its lateral branches. This species alone of those which belong to these two sections spreads itself over dry localities, the others being all confined to a marshy habitat. It would seem best, then, to regard this character in L. cernuum as quite a recent adaptation, this species, at any rate with respect to its habit and external form, being by no means typical for the two sections. The trailing habit of L. cernuum is not so characteristically plagiotropic as is that of the species comprising the Clavata section. It provides in this respect a transition, for its main stem is more ascending than creeping. The apex is continually striving to ascend, and it is only through being weighed down to the ground through its own increase in length that the stem then puts forth a bunch of adventitious roots, the series of such loops in which the plant grows showing that it consistently maintains the ascending habit. Amongst the Leptospermum scrub in the gum lands of North Auckland, where this species grows most luxuriantly I have sometimes seen the stems scrambling over the vegetation. However, this is by no means so marked a habit as in L. volubile, in which latter species striking modifications have taken place as a result of this habit. In the young plants of the species which comprise the Clavata section the monopodial habit of branching is present before ever the young plant becomes plagiotropic; in fact, in most cases there is no dichotomous branching at all (16, figs. 97–101). The terminal region of the main stem suddenly assumes a stouter form and puts forth an adventitious root. This most often takes place after the plant has begun to bend over. If we are to argue from the facts of the ontogeny we must certainly conclude that the plagiotropic habit of growth is a recent adaptation, and that the power of unlimited growth is the most recent feature of this habit. This is the conclusion also to which the facts relating to the Inundata and Cernua, sections seem to lead us. There is one feature which probably is a still further adaptation, and that is the subterranean

habit of growth of the main stem. In L. volubile and L. scariosum this has not taken place, but in L. fastigiatum and L. densum the young trailing stem early buries itself in the soil. This also is the manner in which the subterranean habit comes about in the developing plant of certain of the species of the Inundata and Cernua sections. The “Foot.”—In all the species which belong to the section Clavata the young plant possesses a very large “foot” by which it is attached to the prothallial tissues. This foot evidently functions as an absorbingorgan for a considerable time, for the epithelial cells show the presence of abundant contents even after the sporeling has attained a considerable size. This I have already indicated earlier in this paper in my description of a large surface-growing prothallus of L. volubile. Miss Wigglesworth (31) suggests that it may even function after the disappearance of the prothallus itself. Lang (24) figures the same epithelial layer in the case of the young plant of L. clavatum. Both Miss Wigglesworth and I have found that a strand of vascular tissue containing both xylem and phloem passes off from the main stele of the plant into the foot. The former has demonstrated this in the case of L. complanatum, illustrating it with a series of figures, but she states that she did not find this condition in the young plants of L. clavatum. In a previous paper (16, p. 285–86) I have described the presence of this foot strand in the young plants of L. volubile, stating that whereas in the smaller plantlets the strand consists of small thin-walled cells with abundant cell-contents, in older plantlets a few tracheides are also present. I also stated in the same place that in the exceptionally large “foot” of L. scariosum only in the sections nearest to the main stele was any small-celled tissue to be seen, and this contained no tracheides, and also that in the case of L. fastigiatum the strand was not developed to the same extent as in L. volubile. I concluded that “the development of vascular tissue in the foot of the young plant varies in extent in different individuals of the same species, and possibly this is dependent simply upon the size to which the parent prothalli may grow.” It is evident, therefore, that in this organ also, as in the form and structure of the prothallus, the species composing the Clavata section are in a condition of plasticity. Summary. In this paper I have described the variations which I have observed to occur in the main characters of the New Zealand species of Lycopodium, and my aim has been to show that these characters must be regarded as being more or less in a condition of great plasticity. Also, at the risk of being tedious, I have noted, in order to institute comparisons, some of the facts relating to the plasticity of other species of the genus. These variations in form and structure can, of course, be viewed either as fixed characters or as characters which are maintained only so long as the controlling external conditions are present. This is a question which can only be settled by experimental cultivation of the plants and prothall concerned. The modern genus shows certain main types of form and structure in accordance with which Baker, and more recently Pritzel, has classified the various species. That the latter's classification is a natural one seems to follow from the fact that these types do not relate to one character only, but that all the main characters of the plant, both gametophytic and sporophytic, are more or less consistently interdependent. The different types of the prothallus have been regarded by some botanists as almost unrelated to

one another, representing quite distinct Lycopodiaceous stocks. This was the conclusion arrived at by Treub from his study of the prothalli and young plants of several species belonging to the Selago, Phlegmaria, and Cernua sections, and by Bruchmann also from his study of the prothalli of several European species. Lang, however, pointed out (24) that in spite of the great differences existing between these types there was a common fundamental structure to be traced in all, and that the various modifications of this structure were all obviously in accord with the particular mode of life peculiar to each prothallial type. He stated his belief that it was the change from the self-nourishing chlorophyll condition to a saprophytic condition of life which has determined the lines upon which the Lycopodium prothallus has evolved, and he pointed to the variations which are known to occur in the prothallus of L. Selago as illustrating clearly how it has been possible for the more modified types of prothallus to arise. Further, taking into account the fact that the mode of life of the different types of prothalli, as exemplified in the twelve species whose prothallus was then known, was in close accord with the habit of the sporophyte generation, he suggested that it was possible that the genetic affinities of the species of Lycopodium will be found to coincide exactly with the biological divisions of the genus. The facts which I have brought forward in this paper relating to the variations of the main types which occur in the New Zealand species, and also to the great range of variability which the individual species show under the manifold external conditions under which they are found in this biological region, seem to be thoroughly in accord with Lang's suggestion. Not only do the prothalli of these species provide transitions between the main types in the same sense as does that of L. Selago, but so also do the other main characters; in fact, these variations show that the whole genus is in a state of great plasticity, and that the various types of habit and external form of the sporophyte, of the nature of the fertile region, of the vascular anatomy, and of the form of prothallus and young plant, can be best explained only by viewing them together as adaptations which have proceeded more or less hand in hand. At the same time, the eleven New Zealand species bring to light no new main types either of prothallus, of young plant, or of stelar anatomy. Our knowledge of these species serves to emphasize the fact that there are three main cycles of affinity to be distinguished in the modern genus—namely, the Selago-Phlegmaria group, the Inundata-Cernua group, and the Clavata group. In the different parts of this paper I have tried to institute a distinction between characters which are recent and adaptive and those which are phylogenetic, and in the light of these facts have expanded more fully the conclusions with regard to the interrelationships of these main groups which I reached in a former paper (16, p. 302). Stated briefly, these conclusions were that the Selago section must be held to comprise the most primitive and least modified members of the modern genus, and that the Phlegmaria and Clavata sections have been independently derived from it, the former being less modified than the latter. The Inundata and Cernua sections I suggested should best be placed in a group apart, as having been derived from ancestors common to themselves and to the Selago section but independently of the latter. I will now proceed to sum up the results of the present paper in terms of this thesis. The statement that the Selago section comprises the most primitive members of the genus by no means suggests that with respect to all its

characters it is primitive. The Inundata-Cernua group is generally held to possess the most primitive form of prothallus, being one which is least modified of all from the original self-nourishing chlorophyllous condition. Nor does this belief in the primitive character of the Selago section rest upon a basis of palaeontological fact, very little evidence of this kind being forthcoming, but it is founded upon the broad fact that in a genus which is in an exceedingly plastic condition the characters of this section show, on the whole, the least degree of adaptation to external conditions. The species of the Selago section are typically orthotropic in habit, and dichotomously branched. L. Selago itself shows very little differentiation between fertile and sterile regions of the stem, although in some forms the beginnings of this may be seen. Seeing that the strobilar habit had been thoroughly adopted by the Carboniferous Lycopodiales, it might seem natural to infer that the modern members of the Selago section were in process of losing it by reduction. However, from another point of view it would be unnatural to infer this, for the type species, L. Selago, is the most widely spread and highly variable of all the species of Lycopodium, and this would certainly not be the case if it occupied a position as the most recent member of a long reduction series. The most natural view would seem to be that L. Selago, with its congeners, has originated from a stock which did not possess the strobilar habit. The stelar anatomy in this section is typically radial, a type of stele which is undoubtedly as primitive as the circular protostele. The prothallus of L. Selago is in a much more plastic condition than are those of the Phlegmaria or Clavata sections. The latter are greatly modified in accordance with their respective modes of life, having practically abandoned the chlorophyll condition, and have acquired a definite and probably fixed form in relation to the saprophytic condition. The prothallus of L. Selago is very variable in form according as it occurs at or beneath the surface of the soil, and shows a transition stage between the chlorophyll and the completely saprophytic habit. Bower indicates the significance of this when he says, “A plant which shows such plasticity is clearly not far removed from the self-nourishing condition of the prothallus which was probably the primitive condition for them all” (3, p. 345). Lastly, the L. Selago type of embryo is the simplest and least modified of all the types in the genus, and may well be regarded as primitive. These, then, are the evidences of primitive simplicity which point to the Selago section as comprising the most primitive members of the genus, and from this as a premise the argument as to the relative position of the other sections can be built up. A chain of forms links up the two typical forms L. Selago and L. Billardieri with respect to the external form of the plant and the differentiation of the fertile region, L. varium being an important connecting-link. The stelar anatomy is identical throughout the whole chain of forms, what modifications there are being dependent simply upon the size of the plant and occurring alike in both sections. In all the epiphytic species in which it is known, whether of the Selago or of the Phlegmaria section, the prothallus is of the branched Phlegmaria type. This is so in the rock-epiphyte L. varium and the tree-epiphytes L. Billardieri and L. Billardieri var. gracile. This form is regarded as bearing reasonable comparison with the form which is adopted by the prothallus of L. Selago when growing well below the surface. The form of the sporeling plant of the three New Zealand species is identical with that of L. Selago, the first leaves being large, and not scale-like, and similar to the mature foliage. The typical

epiphytes are for the most part pendulous, although some, by reason of the strongly thickened cortex, are more rigid. L. Billardieri sometimes grows on the ground, and is then more or less erect, with shortened overhanging strobili. L. varium also provides a striking gradation in form from the stiffly erect to the pendulous habit. The entire genus is characterized by the fact that the roots are adventitious and arise behind the growing apex. In the sections Selago and Phlegmaria, in accordance with the habit of growth, the roots emerge at the base of the stem, penetrating down the cortical tissues in order to do this. In L. Selago the characteristic orthotropism is modified by the somewhat sprawling character of the lower region of the stem, and the roots emerge from the stem throughout this region. In L. varium and L. Billardieri the stem is more vertical, whether erect or pendulous, and the roots are borne only in a bunch at the base. The transition between the Selago and Phlegmaria conditions of the fertile region is strikingly exemplified in L. varium and L. Billardieri var. gracile, which may on the one plant show practically all stages between a wholly undifferentiated condition of the fertile region and a special sporophyll and special strobilar formation. At the base of mature xerophytic plants of L. Selago the leaves are of the larger form which is characteristic of the mesophytic variety, and in the embryo plant the first leaves are large and not scale-like. This would seem to indicate that the immediate ancestors of the species of the Selago type possessed leaves which were larger than the acicular or scale-like leaves of the common Lycopodiaceous form. The three New Zealand species which belong to the Phlegmaria section, as also many others, all possess the larger form of leaves. The fact that not a few epiphytic species have the acicular form of leaf may indicate that the large leaf is not merely a mesophytic character, but is an indication of the presence of more than one line of evolution in the Phlegmaria section. It is, of course, probable that the epiphytic habit has been adopted by species which are not immediately related, but that there have been parallel cases of adaptation to similar conditions, and it is possible that these species may have sprung not only from members of the modern Selago cycle of affinity other than L. Selago itself, but even from related forms now wiped out. However, in the New Zealand species we can trace a continuous chain from L. Selago to L. Billardieri through L. varium. Thus a comparison of the New Zealand members of the Phlegmaria section with L. Selago brings forward facts which are in close accord with the belief that the epiphytic species have all been derived from the Selago cycle of affinity, and that the evolution of the characteristic Phlegmaria plant-form, strobilus, and prothallus has been determined by the epiphytic habit. Moreover, the close similarity between the two sections in respect to their chief characters may be regarded as sufficient ground to justify the grouping of these two sections as one natural division of the genus. When we turn to those species which are classified in the Clavata section we find that the main characters of both gametophyte and sporophyte are in a less variable condition than are those of the Selago and Phlegmaria sections. They have become more fixed in form and structure, and are all obviously in direct harmony with the mode of life. The individual species do not, on the whole, show such a wide range of variability in the external form of the sporophyte or in the nature of the fertile region as do, for example, the species L. Selago, L. varium, and L. Billardieri, although certain characteristic “fixed” features are to be found in almost every species. However, in the life-history there is to be found striking evidence of the fact

that the mature form and structure of the plant is a more or less recent modification. In the young plant the habit is orthotropic for a lengthy stage. The stimulation of the cortical and stelar tissues to an increase in girth does not take place till after the stem has begun to assume the plagiotropic habit, and is then localized in its terminal region. This indicates that the character of unlimited growth from a broad apex has been an added feature in the phylogeny. The character of heterophylly makes a somewhat earlier appearance, though not so early in the sporeling of L. volubile as in that of L. scariosum. The stelar anatomy of the stem is primarily radial, as in the Selago and Phlegmaria sections, but this becomes modified to a characteristic dorsiventral arrangement by the restriction of the branching of the mature stem to one plane. The large size of the stele and the differentiation of the xylem and phloem into large conducting elements and flanking storage elements has come about simply through the extended growth of the plant, this being most marked in the scrambling stem of L. volubile. The subterranean habit of the main stems of L. densum and L. fastigiatum is an extreme modification and is only found in a few of the species. In some species—for example, L. densum—the cones are very numerous, every terminal branchlet in the upper portion of the erect aerial shoot being fertile. In this case the cones are quite short and are not always erect, and there is no suggestion of a pedicel. In other species—for example, L. scariosum and L. fastigiatum—there is a great restriction in the cone-formation, only the main branchlets being fertile. The cones are here long and club-shaped, and are raised on long pedicels through the continued growth of the axes of the branchlets. Moreover, the pedicels sometimes branch, and also sometimes even the cones themselves. This club-shaped condition of the fertile region must be considered the most modified form of strobilus in the plagiotropic species of Lycopodium, except that the pendulous strobili of L. volubile are an extreme modification. That the club-shaped cone is not a phylogenetic feature, but an adaptation only, seems to be indicated by the fact that in L. fastigiatum both the densum condition and the scariosum condition occur in a most marked degree. The prothallus shows two main types which are adapted to a subterranean mode of life. It is noteworthy that whereas in L. Selago the form of prothallus which grows deepest is the branched form, while the compact massive form occurs at the surface, and also that in L. Billardieri, &c., the subterranean habit has resulted in a much-branched structure, in the species of the Clavata section the subterranean habit goes along with two types of prothallus, both of which show the compact and massive form in a most marked degree. The branching which occurs in the young prothallus of L. fastigiatum may be significant in this respect, as indicating that this type has not altogether lost the disposition to branch. In fact, in all the New Zealand species the form of the prothallus is not invariable, L. volubile and L. fastigiatum showing a lobing of the upper surface, and a pseudobranching being found in some large individuals of L. scariosum (16). Bruchmann (5, 6) has shown also that the lip of the prothallus of L. annotinum may grow out into long frill-like protuberances, and that the prothallus of L. complanatum may bud. All these features, however, possibly have no phylogenetic significance at all, but are recent modifications. The prothalli of both the Phlegmaria and the Clavata sections have adapted themselves to a subterranean habit, and have departed from the chlorophyll condition, although the clavatum type has done this to a less extent than has the complanatum type. As regards the differentiation of the

fungal tissues the terrestrial subterranean forms of prothallus show greater modification than does the epiphytic form, but, on the other hand, as regards external form the former have probably kept nearer to the ancestral type than have the latter. I will return to this again in discussing the prothallus of the Inundata and Cernua sections. It is quite likely that the immediate ancestral stock of the species which comprise the Clavata section were erect-growing forms which belonged to the Selago cycle of affinity. The stelar anatomy is fundamentally similar throughout, the characteristic differences being readily explained. The adoption of the plagiotropic habit presents no great difficulties of explanation, nor does the difference in the form of the strobilus. The presence of paraphyses on the prothallus of the Selago and Phlegmaria sections, while serving to emphasize the fact that these two sections together constitute a natural division of the genus, certainly serves also to distinguish between the prothalli of the Selago and Clavata sections. If this particular character can be considered as being amenable, along with so many of the other characters of the prothallus, to changes in the mode of life, its absence from the clavatum and complanatum types of prothallus will have no phylogenetic significance. But there is no doubt that it is to such small and constant features as this, especially those connected with both the sexual and asexual reproductive processes, that we are often to look for the most reliable indications of affinity or other-wise. On the whole, it is possible to relate the Clavata section with the Selago section, although the degree of relationship is clearly not so close as in the case of the Selago and Phlegmaria sections. In other words, whereas many of the species of the Phlegmaria section may be without much doubt linked up with certain of the modern members of the Selago section, and even with L. Selago itself, the species of the Clavata section have possibly been derived from the Selago stock at an earlier date, and even from forms of that stock not now existing. Lastly, we must consider the evidence which is afforded by the variations in the New Zealand species of the Inundata and Cernua sections as to the natural position of this group. These two sections seem together to constitute a natural division of the genus which is just as clearly defined as is the Selago-Phlegmaria division. But there can be no doubt that the Inundata-Cernua group stands more or less apart from the rest of the genus, its chief distinguishing characters being the mixed type of stelar anatomy, the surface-growing chlorophyll-possessing prothallus, and the protocorm condition of the young sporeling. The species that are included in these two sections would seem to have departed from the primitive erect habit of growth less recently than have even the thoroughly plagiotropic species of the Clavata section, for they show practically no erect stage in the young plant. Moreover, the dichotomous type of branching has been in these sections very little replaced by the monopodial. The cones in L. laterale are laterally placed and are sessile, but they are not infrequently terminal, and the branching of the aerial shoots is dichotomous. L. Drummondn branches monopodially, and so especially does L. cernuum, more particularly in its aerial shoots, but in the young plants the branching is always dichotomous. The branching of the creeping stem in these sections is never confined to the plane of the ground, as it is in the Clavata section, but the aerial shoots arise dorsally. Thus the plagiotropic habit as seen in this group is different in nature from that in the Clavata section. In a genus in which the main character of the prothallus throughout is a greater or less adaptation to a saprophytic mode of life, the occurrence of a type in

which the saprophytic condition is still subordinate to the chlorophyll condition must be regarded as a primitive feature. Included in these two sections are several forms of strobilar formation. In L. contextum and L. cruentum it represents a near approach to the undifferentiated Selago condition, and this may be regarded, along with the dichotomous branching of the stem and the chlorophyllus nature of the prothallus, as an indication of the ancient character of the Inundata-Cernua group. The long, erect, club-like pedicelled strobili, and the short unpedicelled numerous strobili borne at the tips of all the branches, both occur in this group, as also in the Clavata section. Possibly these two types represent extreme modifications, and, seeing that they both occur in the one species, L. fastigiatum, it is probable that they do not indicate any phylogenetic differences between the species. The protocorm condition in the young plant has been found in all those species of the Inundata and Cernua sections whose embryogeny is known, but it has never been found in any other section of the genus. Treub considered that he had found an indication of this organ in the embryo plant of L. Phlegmaria, but this has never been established. The protocorm is thus a distinguishing character of the Inundata-Cernua division of the genus, and is always associated with the surface-growing chlorophyllous prothallus of the L cernuum type. The monotypic genus Phylloglossum is also characterized by the possession of a protocorm, and Thomas has shown (29) that its prothallus is of the L. cernuum type. Treub elaborated the theory that the Lycopodium protocorm is an exceedingly primitive organ, and Phylloglossum came to be regarded as the most primitive of vascular plants. However, Bower and Goebel have both doubted that it represents a primitive condition in the phylogeny of vascular plants, regarding it rather as a physiological adaptation. In my accounts of this organ as it occurs in L. laterale, L. remulosum, and L. cernuum (16) I have concluded on the one hand that it is a physiological development, but that on the other hand, occurring as it does throughout the whole Inundata-Cernua division, and being always associated with a type of prothallus which is regarded as little modified from the ancestral type, it must certainly be considered as primitive for this division. Further, the fact that Goebel has described adventitious protocorms in L. inundatum(14), and that I have also found them in L. ramulosum (17), would seem to indicate that this organ is not to be regarded on a par with the seedling “foot” as an adaptation which has altogether been governed by the nature of the prothallus with which it is associated, but that it is part of the inherited constitution of this division of the genus. The prothallus in this group is always chlorophyllous and surface growing, but it is remarkably variable in form, the different forms possibly resulting from the different depths at which the spores germinate. In L. cernuum and L. laterale the “shaft” may be somewhat elongated or almost suppressed, in the latter case the prothallus appearing more compact and massive in form. Very extreme forms are met with in L. ramulosum. Here the short form is relatively exceedingly massive and very much like a young prothallus of the clavatum type in external appearance, almost the whole of the bulk of the prothallus, except for the actual crown, consisting of opaque fungal tissues. The elongated form may show as many as five distinct fungal swellings along its length, each swelling being usually associated with a local group of assimilating lobes and sexual organs. This long-drawn-out form may also branch. The prothallus of L. salakense is also much elongated.

The question arises as to whether the ancestral type of prothallus was compact and massive or long-drawn-out and delicate in nature. Being situated at the surface of the ground, it would seem that the prothallus would be short rather than filamentous. This would be in accord with the fundamental structure-plan of the prothallus as seen in all the modern Lycopodiaceous types, in which the lower first-formed vegetative region has the form of an inverted cone, and appears especially likely from the fact that this form occurs not only in the Inundata-Cernua group but in L. Selago also when the prothallus is surface-growing. However, on the other hand, we are probably not to assume that the primary tubercle” is a primitive feature, for it seems to be abundantly clear in the three species L. cernuum, L. laterale, and especially L. ramulosum that these tubercular swellings are due simply to the localized presence of a fungus element which was, of course, not an original feature of the prothallus. In the prothalli of the Selago, Phlegmaria, clavatum, and complanatum types the increase in girth from the spore upwards is always rapid, there being apparently no filamentous stage at all; but this can be readily explained by the fact that infection by the fungus seems there to take place from the very beginning. Also in L. cernuum this is apparently the rule, so that the tubercle constitutes the basal portion of the prothallus; but in L. laterale and L, ramulosum there is generally a well-marked filamentous stage. The young prothalli of both these species are green from the first and remain free from fungus for a considerable period, assuming during this stage the form of an elongated flat filament of cells which shows no localized tubercular swelling of its tissues. Even some of the largest and most drawn-out forms of prothalli of L. ramulosum are seen to have their basal region of this form (16, figs. 32B and 32D), the increase in girth of the prothallus having taken place very gradually and the infection of it by the fungus being not localized in one much-swollen region but distributed over several distinct areas. Thus it is probable that the primitive, wholly selfnourishing type of prothallus was filamentous rather than massive, and that the erect growth and radial build came later as the result of the more rapid development of the tissues consequent on the infection of the lower region by the fungus element. The “mixed” type of stelar anatomy with much-extended protoxylem groups characteristic of the Inundata and Cernua sections is in striking contrast with the definitely radial type which is the primary stelar structure of the rest of the genus. This mixed type is initiated by the precedence of the leaf-trace system over the cauline cylinder in the young plant, but is marked throughout the whole life of the plant. This character, along with others that have been mentioned, gives to the Inundata-Cernua group a position definitely apart from the other sections of the genus. Thus, although some of the main characters of the species of these two sections are probably recent and adaptive, especially as regards the external form and habit of growth of the sporophyte, others are undoubtedly primitive. Also, the possession of a protocorm stage in the embryogeny, and the mixed type of stelar anatomy, brand these sections as together standing very much apart from the other sections, which, as we have seen, are probably somewhat closely interrelated. Being thus primitive in some respects and modified in oters, the Inundata and Cernua sections are best considered to be a natural division of the genus whose ancestors diverged from the ancestors of the Selago stock at a probably early period.

Literature Consulted. 1. Baker, J. G., Handbook of the Fern Allies, London, 1887. 2. Boodle, L. A., On the Structure of the Stem in Two Species of Lycopodium, Ann. of Bot., vol. 14, pp. 315–17, 1900. 3. Bower, F. O.,The Origin of a Land Flora, 1908. 4. " On the Natural Classification of Plants, as exemplified in the Filieales (Hooker Lecture), Jour. Linn. Soc., vol. 44, pp. 107–24, 1918. 5. Bruchmann, H., Uber die Prothallien und die Keimpflanzen mehrerer europdischer Lycopodien, Gotha, 1898. 6. ", Das Prothallium von Lycopodium complanatum L., Bot. Zeitsch., vol. 66, pp. 169–81,1908. 7. Chamberlain, C. J., Prothallia and Sporelings of Three New Zealand Species of Lycopodium, Bot. Gaz. vol. 63, No. 1, pp. 51–65, 1917. 8. Cheeseman, T. F., Manual of the New Zealand Flora, Wellington, 1906. 9. ", Illustrations of the New Zealand Flora, Wellington, 1914. 10. Cockayne, L., Observations concerning Evolution derived from Ecological Studies in New Zealand, Trans. N.Z. Inst., vol. 44, pp. 1–50, 1912. 11. Colenso, W., On a New Species of Lycopodium, Trans. N.Z. Inst., vol. 19, p. 275, 1887. 12. Edgerley, K. V., The Prothalli of Three New Zealand Lycopods, Trans. N.Z. Inst., vol. 47, pp. 94–111, 1915. 13. Engler, A, and Prantl, K., Pflanzenfamilien, I Teil, iv Abteilung, Lycopodiaceae (Pritzel), 1900. 14. Goebel, K., Organography, pt. ii, 1st ed., Eng. transl., 1905. 15. Holloway, J. E., A Comparative Study of the Anatomy of Six New Zealand Species of Lycopodium, Trans. N.Z. Inst. vol. 42, pp. 356–70,1910. 16. ", Studies in the New Zealand Species of the Genus Lycopodium, Part I, Trans. N.Z. Inst., vol. 48, pp. 253–303, 1916. 17. ", Studies in the New Zealand Species of the Genus Lycopodium, Part II, Methods of Vegetative Propagation, Trans. N.Z. Inst., vol. 49, pp. 80–93, 1917. 18. Hooker, J. D., Flora Antarctica, vol. 1, 1844. 19. ", Flora Novae-Zelandiae, vol. 2. 20. ", flora Tasmaniae, vol. 2. 21. Jones, C. E., The Morphology and Anatomy of the Stem of the Genus Lycopodium, Trans. Linn. Soc., ser. 2, Bot. vol. 7, pp. 15–35, 1905. 22. Kirk, T., Description of a New Species of Lycopodium, Trans. N.Z. Inst., vol. 2, pp. 456–57, 1879. 23. ", On Lycopodium varium R. Br. and Lycopodium Billardieri Spring, with a Description of a New Form, Trans. N.Z. Inst., vol. 17, pp. 376–77, 1884. 24. Lang, W. H., The Prothallus of Lycopodium clavatum, Ann. of Bot., vol. 13, pp. 279–317, 1899. 25. Spessard, E. A.. Prothallia of Lycopodium in America, Bot. Gaz., vol. 63, No. 1, pp. 66–76, 1917. 26. ", Prothallia of Lycopodium in America (a Correction), Bot. Gaz., April, 1918. 27. Sykes, M. G., Notes on the Morphology of the Sporangium-bearing Organs of the Lycopodiaceae, New Phytologist, vol. 7, pp. 41–60, 1908. 28. ", A Correction, New Phytologist, vol. 8, 1909. 29. Thomas, A. P. W., Preliminary Account of the Prothallium of Phylloglossum, Proc. Roy. Soc., vol. 69, pp. 285–91, 1901–2. 30. Treub, M., Some Words on the Life-history of Lycopods, Ann. of Bot., vol. 1, pp. 119–23, 1887. 31. Wigglesworth, G., The Young Sporophytes of Lycopodium complanatum and L. clavatum, Ann. of Bot., vol. 21, pp. 211–34, 1907. Postscript (May, 1919). Since this paper was written I have observed in the southern-beech forest on Mount Greenland, Westland, L. varium growing in extensive patches on the forest-floor in a similar manner to what I have described as occurring in Stewart Island. In one instance the clump of plants grew on the edge of a rock, and instead of the plants being rigidly upright in growth, with the characteristic short-curved strobili of the typical

L. varium, they were lax and pendulous and long-drawn-out, and the fertile regions of the stem, were as little differentiated as is the case in the plant L. Billardieri var. gracile. It was only the fact that L. varium was abundant over the whole locality and that L. Billardieri and the variety gracile were altogether absent which indicated that this particular clump of plants belonged to the former species, for in habit of growth they showed the Billardieri form and in the nature of the strobilus that of gracile. It would seem to follow from this that these three species are best to be distinguished from one another by the particular stations that they normally adopt, rather than by their habit of growth or nature of the fertile region, for the latter characters are very variable. It must be added, however, that the strobilus seems to be more fixed in form in L. Billardieri than in the other two species. This particular clump of L. varium was further noteworthy by reason of the fact that nearly all the pendulous fertile branches showed rejuvenation at their tips. These new shoots were precisely similar in appearance to young normally — developed plantlets, and presented a marked and sudden contrast to the fertile branches from which they arose. Moreover, close to the base of the new shoots one or more roots were borne, which, in the case of those branches which reached to the ground, had begun to ramify in the soil. It is possible that the large clumps of this ground-growing species owe their spread a great deal to this mode of vegetative propagation. I have found, also, L. ramulosum on the summit of Mount Greenland, at a height of about 3,000ft., associated in damp peaty places with Cladium teretifolium, very stunted Leptospermum scoparium, and cushions of Phyllachne clavigera. The Lycopodium was here densely matted, and of a very short, creeping, and much-branched form, the numerous cones being about ⅛ in. long, and borne on stiffly erect branchlets ¼ in. to ¼ in. high. In sheltered spots in this locality the plants were, on the contrary, of a more lax and drawn-out form, and bore very few or no cones. (See Cockayne, 10, p. 17.)