appearance will suffice to illustrate the general features of the three New Zealand epiphytic species here dealt with.

L. Billardieri var. gracile grows abundantly throughout Westland on stems of the tree-fern Dicksonia squarrosa, and it has also been reported from various other parts of New Zealand in the same situation. The

Fig. 1.—L. Billardieri var. gracile. Complete mature prothallus in general view, showing basal cone intact, lateral branches, and plantlet. × 13. The small figure represents this prothallus at ¾ natural size.

prothalli and young plants occur more especially on the younger stems of the Dicksonia in between the bases of the stipites, which in this tree-fern run down the stem a considerable length before they begin to be overgrown by the mass of hard brittle aerial rootlets. It is in this more open part

of the tree-fern stem-surface that the prothalli and plantlets of both L. Billardieri var. gracile and of Tmesipteris occur. As the tree-fern grows in height the covering of aerial rootlets spreads up the stem, and so plantlets of increasing age have to be carefully dissected out from the mass of the brittle rootlets. The prothalli of this species of Lycopodium are often to be found adhering closely to the hard black surfaces of the stipites, and are there readily found by tearing away the humus and the debris of old tomentum which collects between the bases of the stipites. Their rather delicate, attenuated form is probably the result of this particular position of growth.

Fig. 6.—L. varium. Prothallus in general view, showing central body and branches, and a young plant. × 10.
Fig. 7.—L. varium. Prothallus in general view, showing basal end, and young embryo in the generative region. × 10.
Fig. 8.—L. varium. A branched “resting” process in general view. × 10.

The prothalli of L. Billardieri, on the other hand, occur for the most part in masses of humus on elevated positions in the forks of the forest trees and of their main branches. The central prothallial body is generally more bulky than that of L. Billiardieri var. gracile, but otherwise the prothallus is identical both in appearance and in structure. With regard to the tropical forms studied by him, Treub states that the prothalli of L. Hippuris are much larger and thicker than-those of L. Phlegmaria, while those of L. nummularifolium are exceedingly thin. These epiphytic prothalli have very much the appearance of a mass of root-ends, but a little

of epidermal cells. This lateral process was free of the fungus, except in certain isolated subepidermal cells which lay at the bases of the young rhizoids. From this latter fact it is apparent that, at any rate when the extension in length of a branch is rapid, infection may take place from without through the rhizoids, and that the distribution of fungus throughout the prothallus does not take place simply through its forward extension from the older regions. The original apex of the oldest conical region

Fig. 9.—L. Billardieri var. gracile. A young prothallus complete, in general view. × 60.
Fig. 10.—L. Billardieri var. gracile. Longitudinal section of the basal cone of the prothallus illustrated in fig. 1, showing remains of spore on the first-formed cell. × 170.
Fig. 11.—L. Billardieri var. gracile. Longitudinal section of central body of mature prothallus, showing general arrangement of tissues, and also foot of young plant. × 47.

was intact, and the first-formed cells were clearly to be traced. Most of the mature prothalli of these epiphytic species which I found were incomplete in their basal region, probably on account of the decaying-away of these oldest tissues through age; but I also found several mature prothalli both of L. Billardieri and of the variety gracile in which the original end was intact. This was the case with that shown in fig. 1. A highly

they are more equidimensional than are the adjacent cells (fig. 11). It is rather the centrally-placed, fungus-free cells which have become changed in form. They are always considerably elongated, and are somewhat narrower than the cells of the fungal zones (fig. 11). This is, of course, quite in accordance with the function of translocation which they are called upon to perform in these elongated prothalli. Certain of the epidermal cells grow out as rhizoids, a transverse wall separating the rhizoid proper from its parent cell. The width of the fungal zone in the main prothallus body varies in different parts, but, generally speaking, it diminishes towards the growing region. However, if lateral branches are borne well forward on the prothallus the fungus will be found well forward also. The fungus is always massed around the base of a branch, but in those portions of the main body which lie in between the branches

Fig. 14.—L. Billardieri. Transverse section of the prothallus shown in fig. 3, at the point c. × 60.
Fig. 15.—L. Billardieri. Transverse section of the prothallus shown in fig. 5, at the point a. Compare also fig. 2, c. × 60.
Fig. 16.—L. Billardieri. Transverse section of the prothallus shown in fig. 3, at the point a. × 135.

the fungal zone may be no more than one or two cells in width. In the basal cone-like region the cells, which are all equidimensional, are all infested with the fungus, there being here no centrally-placed conducting strand. At the actual base of the prothallus even the epidermal cells show the presence of the fungus (fig. 10). Fig. 12 is a transverse section of the basal cone, showing the first-formed lateral branch in longitudinal section. It will be seen that in the central region in this figure the fungus is inter-cellular as well as within the cell-cavities, and the cells here consequently appear roundish in shape. Farther forward still the centrally-placed cells are free from fungus and represent the beginning of the conducting-strand (fig. 13). Fig. 14 is a transverse section of the prothallus which is shown in fig. 3 taken at the point marked c. In this region of the main body the fungus has extended forward so that it underlies the old antheridia. The rhizoids are here ventrally borne, and the fungus is aggregated along the ventral side so that the prothallus-shows a dorsiventral structure. Fig. 15

is a transverse section of the prothallus shown in fig. 5 at aa, and represents the most forward position occupied by the fungus in this prothallus. It will be seen that the fungus in the lateral branch which is also included in this figure is not in connection with that in the main body. The same figure would also represent in transverse section the appearance of the prothallus shown in fig. 2 at cc.

The generative portion of the central prothallial body is always the most bulky, and shows no differentiation of its tissues whatever Immediately behind the growing apex of the prothalli shown in figs. 2, 3, and 5 young antheridia are being produced. Fig. 16 represents a transverse

Fig. 17.—L. Billardieri. Transverse section of the prothallus shown in fig. 3 at the point b. × 135.
Fig. 18.—L. Billardieri. Transverse section of the dorsal side of the prothallus shown in fig. 2 at the point b. × 135.
Fig. 19.—L. Billardieri. Oblique section through the prothallus shown in fig. 2 at the point a. × 135.

section of the prothallus shown in fig. 3 at a. Fig. 17 is a transverse section through an older antheridial zone of the same prothallus at b, while fig. 18 shows the grouping of the archegonia in amongst the paraphyses on the prothallus illustrated in fig. 2 at bb. The sexual organs occur in large numbers in distinct zones. The close grouping of the antheridia, for example, is shown in fig. 19, which is an oblique section taken through the point of the prothallus illustrated in fig. 2 at aa. The generative region in transverse section bears evidence of repeated cell-divisions, and the cell nuclei and contents are very prominent, as if there had been extensive translocation of food material to this region of the prothallus. A longitudinal section of the generative region shows that the extensive development of sexual organs and paraphyses along the dorsal surface brings it about that

cells, some of which have divided. Outside this again is another single layer of somewhat smaller cells which may or may not contain the fungus, bounded peripherally by the epidermal layer, whose external walls are cuticularized. These same layers can also be distinguished in fig. 22 in longitudinal section. Thinner branches show less cell-multiplication in the fungal zone. In longitudinal section it is apparent that the fungus-containing cells have divided transversely, so that they are more nearly equidimensional than those either of the central strand or of the epidermis. The fungus is present in the form of hyphal coils or clusters of oval dark-staining “spores.” The fungus also extends between the cells, so that the latter are frequently roundish in shape.

Sooner or later a branch will show a thickening of its tissues at the apex, and on this swollen region paraphyses and sexual organs will arise (figs. 6, 23, 24). Such a swollen region will be free from the fungus, but the hyphal coils are often aggregated very thickly in that portion of the branch which lies immediately behind it (fig. 24). It is always antheridia which are first formed in such branches, the archegonia occurring only on the main central region of a prothallus. However, a branch may thicken and develop to such an extent that it practically becomes a new prothallus. Probably not a few mature prothalli have arisen not from the germination of a spore, but by a branch having become detached from an older prothallus and having gone on growing in thickness. Such a branch will put forth lateral branches, and will eventually bear archegonia as well as antheridia on its main body. This will account for the fact that most of the mature prothalli found by me did not show the typical cone-like basal region. They are in fact prothalli which have arisen adventitiously. Fig. 27 shows in external view the terminal region of a stoutish branch of L. Billardieri var. gracile which has begun to develop antheridia and paraphyses. Some of the detached branches which are dissected out of the humus are imperfect at both extremities, a fact which indicates that they may persist in the humus and remain self-nourishing for a considerable time.

In dissecting out prothalli of these species of Lycopodium from the substratum in which they lie, one frequently comes across detached prothallial branches which have probably arisen by the decaying-away of the parent prothallus. Three such branches of L. Billardieri var. gracile are shown in figs. 28–30. In this particular species these branches are frequently very long and thin, and they may show the presence of paraphyses along extended portions of their length. Lateral branches frequently bear short club-shaped processes, sometimes in great numbers (fig. 30), which are quite dark with the fungal inhabitant. These processes are to be regarded as “resting” processes, and they may occur also on mature prothalli of the ordinary kind (figs. 1, 11). They are frequently met with detached in the humus, and are either of a simple nature or are branched (figs. 8, 25). The actual apex of the resting process is white and clear of fungus, but the rest is very dark, being thickly infested with it. Fig. 31 represents such a resting process in transverse section, from which it will be seen that the central cells contain the fungus as well as those more medianly situated, and that the fungus is intercellular as well as intracellular. The apex of one of the branches of the stout resting process shown in fig. 25 is given in longitudinal section in fig. 32. Immediately behind the apex the fungus is hyphal only and does not penetrate between the cells. Farther back, however, the hyphal coils have to a large extent been replaced by the “spores,” and the fungus is also intercellular. Instances may be met with

In the three New Zealand species dealt with above it is present at the actual base of the prothallus, having entered probably very early in its development. It occupies the whole of the tissues in the basal cone except at the meristematic apex, being found even in the epidermal cells at the basal point (figs. 1, 9). At first the fungus is in the form of hyphal coils which are in actual connection with the mycelium in the outside soil by means of hyphae which are to be found running through the rhizoids. As these rhizoids die off from the older parts of the prothallus, each being cut off at its base by a strongly thickened transverse wall, the fungus in these older regions becomes isolated from that in the soil, and probably ceases to function. The fungal coils soon disappear in many of the cells, their place being taken by clusters of darkly-staining oval “spores” (fig. 10). I do not know whether or not the fungus is at all used up by the growing prothallus, but I should say not, as in even mature prothalli the cells in the basal cone are still occupied either by the coils or by the “spores.”

A little forward from the basal cone the fungus has begun to penetrate between the cells as well as occupying their cavities, so that the cells appear roundish in section. It is especially present in this intercellular position at the centre of the prothallus, and produces its spores there also (fig. 12). A little higher up the prothallial body the fungus becomes more localized, avoiding the central core of cells which has begun to function in the translocation of food material. The cells of this fungal zone throughout the vegetative portion of the main prothallial body preserve their ordinary form and are in no way altered by the presence of the fungus. Their nuclei also appear large and healthy, showing that the fungus has exercised no harmful effect. I have not been able to distinguish in these prothalli the multinucleate vesicles which have been described as occurring in the fungal zone in the prothalli of the clavatum and complanatum types. All that are here apparent are the dense hyphal coils and the clusters of spores. As the prothallus grows, the fungus pushes forward, always occupying its particular zone, except that at length when the sexual organs and paraphyses are initiated it is confined to the ventral side of the prothallus body (fig. 14). In the forward, bulky, generative region the fungus is altogether absent. Its most forward position is shown in (fig. 15), and it will be observed that even here the coils have begun to be transformed into the spores.

The fungus is also present in the lateral branches, being confined there also to a particular zone, avoiding the centrally-placed conducting-cells and the epidermis. Here, too, it is intercellular as well as intracellular, and the clusters of spores are a well-marked feature (figs. 20–22). The distribution of the fungus in some of the long thin branches is discontinuous (fig. 1), from which it would appear that fresh infection can take place from without through the rhizoids (see also fig. 9). In those stouter branches which have begun to form sexual organs and paraphyses the fungus is only present along the opposite side of the branch (figs. 28–30), the branches thus being bilateral. Sometimes it is to be noticed that when a branch is beginning to swell at the apex preparatory to the formation of antheridia the fungus is thickly aggregated immediately behind this point, there occupying the whole of the tissues of the branch (fig. 24). The “resting” club-shaped branches are always very densely infected, there being no long narrow central conducting-cells (figs. 1, 25, 26, 30). The actual apex of the resting process is of course, clear of the fungus, but the latter reaches right up behind the apex, being found there as dense

be the normal rule in L. Selago—at any rate, in its surface-growing forms— and the latter in the prothallus of the Phlegmaria section. It is significant to note that both these methods of growth are represented in those New Zealand species which belong to the Cernua section, and which are described later in this paper, although in these species they are not so clearly differentiated as in the former. The prothalli of the Phlegmaria section represent the extreme attained within the genus by the elongated, cylindrical type of growth, and the prothalli of the clavatum and complanatum sections represent, on the other hand, the extreme attained by the continued cone-like manner of growth. The prothalli of L. Selago, as Lang has pointed out (10, p. 305), show that these two main types are not fundamentally dissimilar, for they both here occur in the same species. In saying this, however, it must clearly be remembered that, according to Bruchmann's account, the elongated, cylindrical forms of the L. Selago prothallus do not altogether correspond to the elongated prothalli of the Phlegmaria section, for growth in them is not truly apical, but is rather a one-sided marginal extension from the cone. The elongated prothallus of L. ramulosum, as will be shown, is really more comparable to the prothallus of L. Phlegmaria, & c., although, even here, the comparison is not an exact one. The elongated bilateral extensions from the basal cone in L. Selago are better compared with the flattened extension from the margin which Bruchmann described and figured in L. annotinum (1), being of the same nature but much more pronounced.

The more compact surface-growing form of the L. Selago prothallus is nearer to a self-nourishing, chlorophyllous type of prothallus, which must certainly be regarded as more primitive than a wholly saprophytic one. The very young prothallus, as has been said, begins with the cone form, and next has to set apart a central conducting-core which reaches up behind the actual growing apex. The apex, of course, will be more or less conical in shape, and at this stage is certainly not broad. If the prothallus does not proceed to elongate rapidly, the stimulation to increase in size will be largely confined to increasing its girth, and this will take place not only by a broadening of the apical mersitem, but also by divisions taking place in the body-cells generally. At the same time the fungus-free central core will extend in width as the apex broadens and the prothallus slowly extends in length, until it assumes above the form of a cushion of tissue. If chloropyhll is developed in it, this upper cushion will contain a general distribution of starch. In proceeding to explain how the original apical meristem could become transposed to a marginal ring it seems not unnatural to suppose that as the upper fungus-free tissue becomes more bulky, and so loses the tendency to a localized concentration of food material, the stimulation exerted by the supply of food is felt more in those parts of the prothallus where the fungus is youngest and therefore in its fullest functioning power. This will naturally be along the upper margin of the enveloping fungal zone which extends as a ring around the prothallus. Here, then, the meristematic activity of the prothallus will be localized, As the prothallus still continues slowly to increase in size the fungus will push forward bit by bit into the more newly formed cells, the prothallus still retaining the original cone form.

In the surface-growing prothallus of L. Selago this is what happens, and the sexual stage is fairly early initiated, the antheridia and later the archegonia arising from the meristem. A comparative study of the prothalli of the different species of Lycopodium shows that the cone type of

general form of the prothallus. There is sometimes a single basal tubercle of rounded form (7, figs. 17–20); or this basal fungal region may be extended laterally for a distance up one side of the shaft, thus giving to the prothallus a somewhat lop-sided appearance (7, fig. 21, and figs. 42–44 in the present paper). Again, there may be a second fungal region quite unconnected with the first and situated higher up the shaft (figs. 38, 39, 41). Lastly, there may be two distinct fungal regions situated side by side, the two constituting the basal portion of the prothallus and giving it a markedly thickset appearance (fig. 40). I have sectioned several prothalli which showed this latter form. Rhizoids are borne on both fungal swellings in those cases in which two are present. The lobes which are borne at the crown of the prothallus are flattened extensions (7, figs. 17–21), being generally not more than one to three cells thick in transverse section (figs. 40–42). In one or two instances I noticed from my serial sections that there was a group of lobes situated laterally on the shaft of the prothallus. In fig. 41, this is seen to be associated with the uppermost fungal region, as if the formation of this second fungal zone had initiated a meristematic activity in the adjoining regions of the shaft. The thickest part of the prothallus is always that situated at the base of the crown of lobes. Here is the meristem (figs. 42, 45) from which the lobes have arisen, and by whose continued activity the shaft of the prothallus can be extended in length and its upper region in width. Here, too, the sexual organs are to be found and the young plantlets (figs. 40, 44, 45). I have never found the chlorophyll extending much more than half-way down the shaft. Sometimes the lobes are bright green, or they may show a very slight tinge of colour or be quite colourless, this variation possibly depending simply upon the age of the prothallus.

L. laterale.—The prothallus of this species corresponds very closely with that of L. cernuum. I have not found that there is such a marked variation in the length of the shaft as is to be found in the latter species, the majority of the prothalli of L. laterale being more of the intermediate length (7, figs. 13–16, and pl. 17, fig. 3). The longest prothallus I have found is that shown in fig. 46. One prothallus of a very large size is shown in fig. 48. It was exceedingly massive in form, and of irregular growth. Its greatest length was in a horizontal direction, and it possessed a crown of irregular short lobes all along the top. It was almost colourless, though perhaps this was the result of its age. It bore a young plant which was of a bright-green colour, the foot of the plant also being bright green and showing very conspicuously through the prothallial tissues. Chamberlain (3, p. 57) states that in the prothalli of this species examined by him he found no sharply differentiated primary tubercle like that described by Treub in L. cernuum. From his figures it would appear that this was because his prothalli are rather of the short, thickset form. The basal tubercle is always more distinct when the shaft is well developed. I have found that, as in L. cernuum, the basal, fungal region may extend laterally for a short distance up one side of the shaft, thus giving a lop-sided appearance to the prothallus as a whole (7, fig. 13). Also, as in the other species, there is sometimes a second fungal region higher up the shaft and quite distinct from the basal tubercle (fig. 47). The leafy expansions on the crown of the prothallus of L. laterale are more filamentous and less lobe-like than those of L. cernuum (7, figs. 13–16, and pl. 17, fig. 3, and figs. 46 and 47 in the present paper). This appears also from Chamberlain's figures. A large proportion of the prothalli examined by me showed

a lateral group of lobes borne on the shaft. I previously thought that this was a distinguishing feature between the two species L. cernuum and L. laterale, but have since found it also in the prothalli of the former. In one of the figures given in Part I of this series of papers (7, fig. 16) I showed that a long, thin, club-shaped process is sometimes to be found attached to the basal tubercle of the prothallus of L. laterale, and I there expressed the opinion that this was the first-formed part of the prothallus, as it so often is also in the prothallus of L. ramulosum. My serial sections of this prothallus unfortunately do not make this satisfactorily clear, but none the less I feel confident that this is the right interpretation. The base of this first - formed process (7, fig. 16) shows the presence of the fungus, and it may be compared with the corresponding parts of L. ramulosum illustrated in the present paper in figs. 52–54. As in L. cernuum, the prothallus of L. laterale is normally short-lived, having decayed away by the time that its plantlet has developed two or three protophylls. Probably the large size of the prothallus in fig. 48 was due to the fact that for a considerable time no embryo was formed on it. Moreover, in this case, on account of its large size, the prothallus would persist attached to the young plant for a much longer time than usual.

Fig. 50.—L. ramulosum. Complete mature prothallus of elongated form in general view, showing several swollen fungal regions. × 30.

L. ramulosum.—The prothalli of this species vary very remarkably, both in form and structure. I have discovered these prothalli in several different localities and in considerable numbers, and so am able to give a fairly complete account of their typical form and of the variations of it. The prothalli, as in the other species belonging to the cernuum type, are more or less green in their upper regions, being situated at the surface of the ground. Some of the prothalli of this species are very similar to those of L. cernuum, while others show on the one Hand the elongated and on the other hand the compact, massive habit which I have described as occurring both in L. cernuum and in L. laterale, but here to a much more marked extent. Figs. 50–55 illustrate prothalli of the elongated form

and figs. 56–60 prothalli of the compact, massive build. Some of these figures were given in Part I of the present series of papers, but are here reproduced in greater detail. Photos of three elongated prothalli of this species in general view are also given in Plates XII, XIII, and XIV, and

Fig. 51.—L. ramulosum. Complete mature prothallus of elongated form in general view, showing several fungal regions and a young plantlet. × 30.

a photo of a more massive prothallus with its young plantlet is given in longitudinal section in Plate XV. The prothallus shown in Plate XII is the same as that given in fig. 50.

In the elongated forms the fungal regions may be swollen so as to present the form of rounded tubercles, or, again, they may be scarcely swollen at all. Each fungal region invariably bears a group of rhizoids. In the massive forms the prothallus may be compact and show a certain

similarity in general appearance to the surface-growing forms of the L. Selago prothallus, or they may be so irregularly extended in length that their build is difficult to distinguish. In the elongated forms, along with the development of a number of fungal areas, there goes a corresponding development of several distinct generative regions bearing sexual organs and filamentous or lobe-like processes. Frequently, in this latter form of prothallus, the first-formed generative regions are brown and withered, suggesting that the prothallus had resumed growth on one or more occasions at the close of a dry period or on account of there being no embryo developed. I have always found that a young plant when present was borne on the last-formed region of the prothallus, so that growth in length is apparently not continued after the development of a young plant. The prothalli sometimes persist attached to the developing plant for a much longer time than in L. cernuum. I have found several instances of a healthy prothallus still attached to a plantlet which showed as many as seven or eight protophylls along an extended protocorm. The lateral processes are sometimes filamentous, and at others short and lobe-like. Owing to the fact that the prothallus of L. ramulosum is so variable in form, it is impossible to give many details in a general description of it. I will now proceed to enter more into detail in connection with the particular individuals figured in this paper.

The prothallus shown in fig. 50 and in Plate XII possessed five fungal areas, quite distinct from one another, each being swollen and rounded in form and bearing a group of rhizoids. There are two generative regions showing the presence of archegonia and bearing filamentous processes. The basal fungal swelling bore no evidence of any original filamentous stage, and was probably formed immediately on germination of the spore. As will be described later, a first-formed filamentous stage is frequently to be observed in the prothalli of this species. The uppermost fungal region is, of course, the youngest. It occurs at the base of the meristematic apex of the prothallus, and it is evident that both this growing apex and its fungal region have just begun to form. Both from the figure and from the photo it will be seen that here, curiously enough, the fungus lies internally to the epidermal cells. There is no special apical cell or group of cells, the whole apical region of the prothallus being meristematic. The succession of five swollen fungal regions shows very clearly that it is the fungus which causes the swelling in the tissue of the prothallus, and that, at any rate in this prothallus, the primary tubercle is in no wise different from any of the later-formed tubercles.

The prothallus shown in fig. 51 illustrates some interesting points in comparison with that just described. There are six distinct fungal areas, and the four lowest are closely approximated and are scarcely, if at all, swollen. There is a well-developed filament at the base of the prothallus, and the prothallus as a whole increases quite gradually in girth from this upwards. Even the first-formed fungal region is practically unswollen. I should judge that the extension in length of the prothallus in these early stages had been comparatively rapid, there being no cessation in growth in length consequent on the formation of a first generative region or of the storing of food material. This state of things can bear close comparison with what takes place normally in the cylindrical much-elongated epiphytic prothalli, in which growth is continuous and rapid and there is an entire absence of swellings in the fungal regions. The adoption, or rather the preservation, of the definite apical meristem in the epiphytic type is probably only the result of this habit of growth. The prothallus

of L. ramulosum shown in fig. 51 may best be described in its entirety as a gradually widening filament. The young plantlet borne at its upper extremity was peculiarly diminutive in size. There were no antheridia

Figs. 52–54.—L. ramulosum. Complete mature prothalli of elongated form in general view. a is the basal fungal region; a′ and a″, secondary fungal regions; b′ and b″ are generative regions bearing lobe-like outgrowths. In fig. 52 there is a lateral outgrowth (b) associated with the basal fungal region. The shading on b′ and b″ represents browning due to withering. × 20.
Fig 55.—L. ramulosum. Complete mature prothallus of elongated form in general view, showing branching above the second fungal region. × 20.

either on this prothallus or on that in fig. 50. The photo of the prothallus given in Plate XIII also illustrates the nature of the elongated manner of growth with the formation of several fungal regions.

The four prothalli outlined in figs. 52–55 show at their base a distinct fungal tubercle, that in fig. 53 bearing evidence of an original filament. The prothallus in Plate XIV shows a lower fungal region, which was, however, not the first-formed portion. The shaft is long, and bears a lateral process and also a terminal broad generative region which is associated with a second fungal area. In figs. 52 and 54 there is a fungal swelling located behind the growing end, it being evident also that in both these prothalli the apex is at the point of renewing its growth. A large number of prothalli of this species which I found were of the peculiar step-like form shown in figs. 52–54. Here the generative regions are sometimes of considerable extent, and, relatively to the long axis of the prothallus, lie horizontally. The position of these elongated prothalli in the soil is always vertical. The older generative regions are frequently brown and withered, either as the result of a dry spell in the weather or because they have ceased to be supplied with food and have begun to decay. These generative regions are frequently fairly bulky. It is probable that the irregularly massive forms of prothallus described next have acquired their form through an exceptionally bulky development of one or more of these successive generative regions, the intermediate shaft-like stages of growth having been suppressed. Sometimes these massive prothalli show the remains of the lowest shaft and first-formed tubercle still attached to them. Those prothalli of L. laterale which bear club-like processes attached to the primary tubercle, as described in a previous paper (7, fig. 16), probably come under this category. Fig. 55 shows a prothallus which has branched in its middle region, the two branches being borne on a particularly large fungal swelling. These two branches are clearly not to be regarded as processes of the kind usually borne on the generative regions. One of them showed a fungal region with rhizoids near its apex. Both are browned at the tip. It is certainly interesting to find that along with its power of cylindrical elongation the prothallus of this species is able also to branch, these variations suggesting that the Lycopodium prothallus is very plastic in character and that the several main types are by no means so fixed as has been supposed.

I pass now to the description of the massive and compact prothalli. Some of these show a very regular build (figs. 56, 57, 60, and Plate XV), being comparable with the surface-growing forms of L. Selago (see 1, pl. 6, fig. 37). They do not, however, show the same internal structure as the latter. The prothallus illustrated in fig. 56 possesses one fungal region only, which occupies the whole of the base. This leads up into a short, massive shaft and a wide, upper, generative region. The whole of the uppermost region is meristematic, there being no localized marginal, ring-like meristem as is the case in the compact prothallus of L. Selago. At the base of this prothallus there is a filament consisting of one long cell bent in the middle (fig. 56a) which does not show the presence of fungus. It leads up into a swollen region which can be described as the basal tubercle, this latter passing gradually again into a higher and more swollen region. The prothalli shown in figs. 57 and 60 and Plate XV are also of this same form, there being but one large basal fungal region. That in Plate XV is in longitudinal section, but the section does not pass medianly through either the base or the shaft, nor does it show the full width of the prothallus. This massive form of the prothallus of L. ramulosum is similar to that sometimes adopted by the prothallus of L. cernuum and L. laterale, howbeit in a more marked degree.

seen only in certain standard books of reference, show that a tubercle is at once initiated, which is apparently at first devoid of fungus. This tubercle was called by Treub the “primary tubercle,” and he supposed that it was a characteristic feature of the Lycopodium prothallus. He was also successful in experimentally cultivating the prothalli of L. salakense. This also begins with a globular body, the “primary tubercle,” from which several thin filamentous branches arise, one of the latter eventually thickening and producing the sexual organs. This prothallus remains altogether free from a symbiotic fungus. Treub also germinated the spores of L. curvatum Sw., from which he obtained “primary tubercles,” but he was unable to induce them to develop farther. All the mature prothalli of L. cernuum which I have found show the rounded basal tubercle, although in some instances this is somewhat pointed below (see 7, figs. 17–21, and figs. 38–44 in the present paper).

I have dissected out a number of young prothalli of the species L. ramulosum along with the more mature forms, and some of these are shown in figs. 61–65. I judge these to belong to this species since they are of the typical Lycopodium form and no other species of Lycopodium were present in the two localities from which they came. No other kind of prothallus was ever found by me in the turves from which I dissected out those of L. ramulosum, nor was any species of fern present in the near neighbourhood, except Gleichenia dicarpa. Finally, the individuals which are shown in figs. 61–65 form a series which leads on to the more mature prothalli which undoubtedly belong to L. ramulosum. The youngest prothallus was that shown in fig. 61. In it there was no basal swelling, nor was there any indication of the presence of a fungus. The lower half consisted of a single linear row of cells, whilst in the upper half there was a gradual increase in the number of cells and in the girth of the filament towards the apex. This upper half was not a flat expansion of cells as in the usual fern prothallus, but was radial in build. The cells at the apex were smaller than those farther back, and evidently functioned as the meristem. The whole filament contained chlorophyll corpuscles. The original spore-case was still attached to the basal cell. In fig. 62 is shown a prothallus in which the first-formed filamentous stage was very short, passing almost immediately and suddenly into a globular mass of cells which could be called the “primary tubercle.” This prothallus also was quite free of fungus, being green throughout. The apical region consisted of small-celled tissue, and a young sexual organ was developing near by. A group of rhizoids was borne on the basal tubercle. In the formation of a fungusless primary tubercle this young prothallus corresponds very closely to those early stages in L. cernuum, L. salakense, and L. curvatum described by Treub. With regard to the mature prothallus of L. ramulosum, the conclusion I arrived at was that any swelling in the fungal regions was primarily due to the localized presence of the fungus. From these other three species, however, it is apparent that there may be at first a primary tubercle quite apart from the presence of the fungus, and this also appears in such a young prothallus of L. ramulosum as that shown in fig. 62. However, this is not always the case, as in the particular prothallus shown in fig. 61. The three young prothalli shown in figs. 63–65 all possessed a primary tubercle, which was infested by fungus.

Are we to consider that a primary tubercle is a fundamental feature in the structure plan of the Lycopodium prothallus, or is it to be regarded as an added feature? Possibly we are to regard the radially-built filament

as being the original typical condition, this filament being drawn out, or short and bulky, according to whether growth has taken place rapidly or slowly. The basal tubercle of the mature prothallus when present would, according to this view, be the result largely of secondary cell-divisions

Fig. 61.—L. ramulosum. Very young prothallus, in general view, with spore still attached. × 75.
Fig. 62.—L. ramulosum. Young prothallus, in general view, showing basal tubercle with no fungus. × 45.
Figs. 63, 64.—L. ramulosum. Young prothalli, in general view, showing basal tubercle with fungus. × 30.
Fig. 65.—L. ramulosum. Young prothallus, in general view, showing first-formed lobes and the initiation of a second fungal region. × 45.

which had taken place in the first-formed basal cells owing to the stimulus excited by the storing of food material. Prothalli cultivated under artificial conditions seem generally to develop slowly, and this may explain why it is that none of Treub's prothalli showed a first-formed filamentous

is sometimes a second fungal region quite distinct from the first, which may be situated either higher up the shaft or at the base alongside the primary tubercle. The characteristic difference in structure between the basal tubercle and an upper fungal area is clearly shown in fig. 39. The details in the basal structure of the other form are illustrated in fig. 40, the right-hand tubercle here being cut not quite medianly.

The shaft is built up of large cells whose long axes are directed towards the growing apex. In some prothalli this shaft is long and thin (figs. 38, 42), but in others it is short and thick (figs. 40, 44). An instance in which a lateral group of lobes is borne on the shaft is given in fig. 41, these lobes

Figs. 66, 67.—L. cernuum. Transverse sections of the basal tubercle of prothallus, showing distribution of fungus. × 42.
Fig. 68.—L. cernuum. Longitudinal section of portion of fungal area shown in fig. 41. × 135.
Fig. 69.—L. ramulosum. Median longitudinal section of basal tubercle of massive prothallus shown in Plate XV. × 42.
Fig. 70.—L. ramulosum. a and b, longitudinal section of an elongated prothallus with three fungal regions, the basal tubercle being shown at b. × 48. c, longitudinal section of apical head of the prothallus shown in fig. 70a. × 75.
Fig. 71.—L. ramulosum. Longitudinal section of two adjoining fungal regions in a large elongated prothallus. × 42.

here being in conjunction with a second fungal zone. At the top of the shaft is the crown of lobes, these being in transverse, section from two to four cells in thickness (figs. 40–42, 44). In amongst the bases of these lobes are situated the sexual organs (figs. 40, 44). Some of the prothalli bear organs of one kind only, that shown in fig. 40 bearing numerous antheridia but no archegonia. Others, again, bear both, as in the case of the massive prothallus shown in fig. 44.

L. laterale.—The structure of this prothallus corresponds more or less closely with that of L. cernuum. In the fungal regions, whether at the base of the prothallus or higher up the shaft, the hyphal coils are confined to the epidermal cells, and the hyphae penetrate the adjoining tissues in an

first by transverse walls arising in the terminal cell (fig. 61), but later it would seem from fig. 49 that oblique walls in the terminal cell cut off segments alternately on one side and on the other, so that the filament becomes two cells in width. All indications that I have noticed point to the fact that this first-formed part of the filament does not subsequently thicken. It is to be found in this thin form still attached at the base of older prothalli (figs. 51, 53, 56), so that it has nothing to do with the “shaft” of the prothallus. Presumably the tubercle is formed by the activity of a single triangular apical cell as in Treub's material, though, of course, my material is not young enough to show this. The young prothalli of L. ramulosum given in figs. 63–65 have grown on from the primary tubercle not in a filament one cell wide, but in a shaft of the same width as the tubercle. This shaft becomes progressively wider as it grows on. If a tubercle shows subsequently a distinct rounded form it must have been due to secondary cellmultiplication having taken place in it owing to the presence of the fungus. In the prothalli of L. ramulosum shown in figs. 62, 63, and 64 the whole of the small-celled terminal portion is clearly meristematic, so that here, possibly after a period of rest, the whole of the forward end of the tubercle has become meristematic. In the prothallus shown in fig. 61 the original filament gradually thickens as it grows on, apparently by longitudinal walls appearing in the various cells, and the whole of the apex has taken on the function of growth.

In L. cernuum the apex of the shaft grows on and eventually gives rise to a crown of lobes. Some of Treub's figures seem to indicate that the lobes are initiated when the prothallus is not more than of middle age. The meristematic zone is then largely localized at the base of the lobes, and by its continued activity the lobes are extended in length, and also the length of the shaft is added to. At the same time the shaft can increase slowly in width by the division of its cells in a longitudinal direction. Under certain conditions, perhaps inducing quickness of growth, the shaft becomes somewhat elongated, and the lobes also are well developed (fig. 42). The opposite conditions bring about a thickening of the shaft unaccompanied by any great increase in its length (fig. 44). Thus in the prothallus practically the whole of the main body is more or less meristematic, and the localization of this function at the growing apex is not so definite as in the prothalli of the epiphytic or of the terrestrial subterranean types. Meristematic activity at the base of the lobes becomes most pronounced when the sexual organs are being formed, so that they are embedded in a small-celled tissue (figs. 40, 45).

In L. ramulosum the prothallus is capable of much more extensive growth, probably on account of its continued infection by the fungus. At a fairly early stage the apex gives rise to the lobe-like outgrowths, the initiation of which is shown in figs. 63 and 64. In the young prothallus in fig. 65 two outgrowths are well formed and a sexual organ is developing at their base. The lobes are browned and slightly withered, as if they had been touched by a dry spell in the weather. At the same time the shaft of the prothallus is proceeding to extend on past the lobes so that these latter will eventually come to occupy a lateral position. Immediately behind the developing shaft-apex three of the epidermal cells have become infected with the fungus and one epidermal cell has grown out as a rhizoid. It certainly seems as if this fresh infection by the fungus had been the cause of the renewed activity of the prothallus. The study of the prothallus given in fig. 50 suggests that the apical meristem has alternately

The development of an embryo seems to bring all further growth of the prothallus to an end. The prothallus of this species is generally small and delicate, and does not long persist attached to its young plant. It has in most cases decayed away by the time that the plant has developed two or three protophylls. However, I have several times found prothalli of L. ramulosum still persisting, in one case healthily green in colour, attached to plantlets of seven or eight protophylls.

The intraprothallial portion of the embryo plant is developed as an absorbing-organ, but only to a slight extent as compared with the same organ in the young plant of the Selago and Phlegmaria types. Still less is it comparable in importance with the large “foot” of the clavatum and complanatum types. This progressive importance of the foot in the different types, taken in the order in which they are mentioned, goes hand

Fig. 72.—L. cernuum. Longitudinal section of prothallus and young plant showing foot. × 30.
Fig. 73.—L. cernuum. Transverse section of suspensor and foot of young plant. × 42.
Fig. 74.—L. cernuum. Longitudinal section of foot and part of protocorm of young plant. × 42.
Fig. 75.—L. ramulosum. Oblique section of portion of basal fungal region showing three large “spore” capsules of undetermined nature. × 250.

in hand with the increase in the size of the prothalli and of the extent of the dependence of the young plant upon its parent prothallus. From the figures mentioned above it will be seen that the surface cells of the foot in the three species of the cernuum type are only slightly developed as an epithelial layer. They are best illustrated in fig. 73, which is a transverse section of the foot and suspensor of the young plant of L. cernuum, and fig. 74, which shows the foot in longitudinal section.

In Part II of these papers (8, p. 92) I stated that I had never observed the presence of fungal hyphae in any of the protocorms which I had examined. This statement I must correct. The fungus is certainly never present as

passes over into the elongated cylindrical main body which bears sexual organs on the dorsal side, so becoming bilateral. The prothalli of L. Selago grow upright in the soil. They either preserve the cone form throughout their life—being, however, much smaller than the prothalli of the clavatum or complanatum types—or they give rise to elongated bilateral extensions of portions of the upper margin of the cone. In the upright prothalli of the clavatum and complanatum types there is practically no disposition towards elongation, the cone form being preserved throughout the life of the comparatively massive prothallus, any extra meristematic activity showing itself merely in irregular lobing of the upper margin or in adventitious budding. The angle of divergence of the cone from its apex upwards is different in these two types, being greater in the former than in the latter, so that the mature forms of the two prothalli are somewhat different. The prothalli of the species which belong to the Cernua and Inundata sections are also upright in growth, and have a radial build, so that in transverse section the main body is circular in outline, but the cone form is not always present. Treub's description of the prothalli of L. cernuum suggested that a short cone-like stage always followed immediately upon the germination of the spore, and the” primary tubercle,” as it was called by him, was thus looked upon as a characteristic feature in this type of prothallus. In all the species of this type whose prothalli are known a primary tubercle is, indeed, often present, but not invariably so, as appears, for example, in the prothalli, both young and old, of L. ramulosum.

The cone form adopted by the Lycopodium prothallus must be considered along with the presence in its tissues of the symbiotic fungus. It is, of course, well known that this fungus is one of its most characteristic features, just as it is also of other subterranean pteridophytic prothalli. Treub states that the prothallus of L. salakense shows no fungus, and it is quite possible that other species belonging to either the Cernua or Inundata sections may prove to be without it. From what we know of the young prothallus of L. laterale and L. ramulosum it would seem that in the sections Cernua and Inundata the infection may be postponed considerably, for the fungus is frequently absent altogether from young prothalli of quite fair size, and also not infrequently mature prothalli show signs of the same initial fungusless region at their base.

In the prothallus of Tmesipteris the fungus is found even in the basal filament, when that is present; but whether the infection is initiated from the very first or subsequently spreads back into the extreme lower parts of the prothallus is not known. I have found many prothalli of Tmesipteris, both young and also mature, which showed a short filament of cells at their base, and also, on the other hand, complete prothalli in which no filament was present, in the latter case the remains of the original spore being sometimes clearly apparent on the cell at the actual apex of the basal cone. Sometimes the Tmesipteris prothallus takes the form of a steadily tapering cone, and at others in its lowest regions it increases in girth by a progressive series of gentle swellings. One cannot avoid the suggestion that the dominating factor in the Tmesipteris prothallus is the presence of the fungus, the nature of the swelling being due to the consistency or intermittency, as the case may be. of its activity, and the omission of a filamentous stage being due to its very early entry into the germinating spore. When we turn to the prothalli of the Lycopodiaceae the same conclusion seems to be forced upon us. Large mature prothalli of L. ramulosum may show no primary tubercle at all, and the cone form may be altogether absent, so that the general form of this prothallus can then

best be described as a radially-built filament. The immediate effect of the presence of the fungus also can be clearly seen in the prothalli of this species. It almost invariably causes a swelling of the tissues of the prothallus at the infection spot, this swelling being due to the greater or less development of the internal fungal tissue in which the hyphae are intercellular. Sometimes the fungus is restricted to one place in the prothallus, this being towards or at its base, so that there is present a basal swelling. But in all three New Zealand species which belong to the cernuum type, and more especially in L. ramulosum, more than one distinct fungal area is frequently formed. Another form sometimes adopted by the prothallus of L. ramulosum, and to be seen also in those of L. laterale and L. cernuum, is that in which the usual shaft is practically absent, the crown of lobes immediately surmounting the basal fungal region. The latter is then more bulky than usual, the whole mature prothallus bearing more truly the cone form than is normally the case in the prothalli of these sections. That this particular massive form is not a fixed character appears from the fact that only infrequently are these bulky prothalli of regular shape and growth. However, just as the long-drawn-out, more filamentous form may bear some comparison with the epiphytic, elongated type, so may the massive form be compared with the deep-growing clavatum and complanatum types, and the prothallus of L. ramulosum is thus seen to possess as great significance as that of L. Selago in bearing witness to the plasticity of the Lycopodium prothallus.

When we turn to the epiphytic type of prothallus we find that there also the form of the prothallus must be considered along with the nature of the fungus-distribution in its tissues. The whole of the basal cone constitutes, as it were a primary turbercle which contains the fungal element throughout its tissues, being comparable to the first-formed tubercle in L. cernuum, when that is there present. The elongated, cylindrical regions of the epiphytic prothalli, have completely passed out of the cone form, nor do they show any localized swellings induced by the presence of the fungus. The fungus is distributed throughout the branches and the vegetative region of the central body, and the absence of swelling is probably due to the rapid elongation of the prothallus in all its parts. Even the discontinuous infection which is to be seen often in the branches does not result in any swelling of the tissues as it does in L. ramulosum. Thus, although the elongated prothalli of L. ramulosum may be compared with the epiphytic type, yet they never attain to the characteristic form or structure of the latter. It will suffice here to allude quite briefly to the statement made by certain other writers, notably Lang, that this difference is probably merely the expression of the fact that in the Cernua and Inundata types fungal assimilation has not assumed the same degree of importance in the life of the prothallus as it has in the epiphytic types, for the prothallus is surface-growing and never gives up its chlorophyllous, self-nourishing habit. The epiphytic prothalli, however, live in soil of a different nature, and have adopted a wholly subterranean mode of life, in which they depend altogether upon the activities of the fungus for their nourishment. This has been the immediate cause of their extensive ramifications and their character of continuous cylindrical growth. In this respect, therefore, they are to be regarded as more specialized than the chlorophyllous type.

Although I have not in the present paper given any description of the prothalli of the species L. volubile, L. fastigiatum, and L. scariosum, which belong to the clavatum and complanatum types, yet it is appropriate here to compare briefly their form and structure with that of the other New

Zealand species already dealt with. In these two types the cone form is at once initiated and is continued, throughout the life of the prothallus. There is no extension in length comparable even to that of the Cernua and Inundata types, although, as I mentioned in a former paper (7, p. 273) and hope to describe more fully in a later one, the young prothallus of L fastigiatum branches after it has passed through the initial cone stage. It assumes the shape of the letter “Y,” the arms of the prothallus being true branches, as indicated by the fact that in each of them there is a central core of fungus-free cells completely surrounded by the fungal zone, the meristem being located near, but not actually at, the apex of each arm. Although the mature prothallus in this species has the saucer form, yet it frequently bears evidence of this early branching. The prothalli of these two types have departed altogether from the self-nourishing, chlorophyllous habit, and have become wholly dependent upon the symbiotic fungus, although the prothallus of L. volubile is able to develop chlorophyll if it happens to grow at the surface of the ground. We find that the external tissues in the whole of the vegetative portion of the prothallus have been given over to the fungal inhabitant, so that in form this vegetative region is one large regularly-shaped cone. The prothalli of the clavatum and complanatum types, owing to their terrestrial habit, occur in soil of a nature different from that of the humus-growing, epiphytic prothalli. There seems to be no need for the prothallus to elongate or branch in order to come in contact with a sufficient supply of food, although, judging from the branching by which the young prothallus of L. fastigiatum assumes the saucer form, it could do so if necessary. The fact that these types of prothalli are so much larger in size than the surface-growing form of L. Selago is obviously due to the much greater degree of dependence upon them of the young plant than in the latter species. They function largely as storehouses, the large quantities of starch and oil which they contain having been noted by all who have described them.

The compact, massive form of the prothallus in L. ramulosum, L. laterale, and L. cernuum must not be compared too closely with that of L. Selago or of the terrestrial subterranean types, for the position of the meristem in the latter is. altogether different from that in the former. Nor can the elongated form of L. ramulosum be compared exactly, with that of L. Billardieri. It would seem, however, that the difference between the marginal meristem of the Selago prothallus and the apical meristem of the epiphytic types is not so great. The prothalli of both these types in variably begin with the cone form, and the subsequent manner of growth of each, and the form which the meristem takes, can be explained quite naturally (as I have attempted to do earlier in this paper) as resulting simply from the position of the prothallus in the soil. That there is a somewhat close relationship between these two types appears also from the fact of their similarity in the embryo plant and in the presence of paraphyses. The clavatum and complanatum types of prothallus also seem to be allied to that of the Selago type, possessing the same kind of marginal meristem, although they lack paraphyses. Moreover, the embryo plant of these deeply subterranean forms differs from that of L. Selago only in the abnormal development of the “foot,” a character which has quite obviously resulted from their position in the soil.

The prothalli of the Cernua and Inundata sections seem to stand rather apart from the rest of the genus. The meristem is of a quite different nature, the fungal habit has not been adopted to the same extent, the filamentous manner of growth is largely present, and the form of the