Lycopsida

Sporophyte with roots, stems and spirally arranged leaves (microphylls). Protostelic (solid or medullated) sometimes polystelic (rarely poly cyclic). Some with secondary thickening. Sporangium thick-walled, homosporous or heterosporous, borne either on a sporophyll or associated with one. Antherozoids biflagellate or multiflagellate.

List 4

Protolepidodendrales

Until 1953, when Aldanophyton was described from Cambrian deposits in Eastern Siberia, Baragwanathia was believed to be the earliest representative of the Lycopsida, for it occurs along with Yarravia in Silurian rocks of Australia. It had fleshy dichotomizing aerial axes, thickly clothed with leaves, and must have had a most remarkable appearance, for the diameter of the axes ranged upwards from 1 cm to 6*5 cm (Fig. 8A). In the centre was a slender fluted rod of annular tracheids from which leaf traces passed out through the cortex into the leaves. The leaves were about 1 mm broad and up to 4 cm long and, in fertile shoots, they were associated with reniform sporangia arranged in zones. The preservation of the specimens is not good enough to show whether the sporangia were borne on the leaves or merely among them, but that they were indeed sporangia is established by the extraction of cutinized spores from them. Not much is known of the growth habit of the plant, but there are suggestions that the aerial branches arose from a creeping rhizome. Drepanophycus (=Ar thro stigma) and Protolepidodendron both occurred in Lower and Middle Devonian times: the former in Germany, Canada and Norway; the latter in Scotland and Germany. Of the two, Drepanophycus (Fig. 8B) was the more robust. Its aerial axes were up to 5 cm thick and forked occasionally in a dichotomous manner. It is believed that they arose from horizontal branching rhizomes. The aerial axes were covered with spine-like outgrowths (up to 2 cm long) in a manner reminiscent of Psilophyton, but with the difference that these outgrowths had a vascular strand and could therefore properly be called leaves. Some of them bore a single sporangium either on the adaxial surface (Fig. 8C) or in their axils, but these ‘sporophylls’ were scattered at random over the axes instead of being gathered together into a fertile zone.

Baragwanathia

Protolepidodendron (Fig. 8D) had dichotomous creeping axes from which arose aerial axes up to 30 cm high and less than 1 cm in diameter. All parts of the plant were clothed (sometimes densely) with leaves having cushion-like bases and, in most species, bifurcated apices. Stems from which the leaves had fallen showed a characteristic pattern of leaf-bases (Fig. 8E) arranged in a spiral manner. All the leaves were provided with a single vascular strand and some of them bore oval sporangia on their adaxial surfaces (Fig. 8F) but, as in Drepanophycus, such sporophylls were not aggregated into special fertile regions. Details of the stem anatomy of Protolepidodendron are difficult to make out, but there appears to have been a solid three-angled protostele in the centre, with some suggestion of a mesarch protoxylem.

Whether Aldanophyton was really a member of the Lycopsida cannot be determined with certainty, for no fertile portions of the plant have been described. It had stems up to 13 mm in diameter, clothed with narrow leaves up to 9 mm long and, although the preservation of the specimens leaves much to be desired, one published photograph looks not unlike Fig. 8E. Whatever its true affinities, this plant is an important discovery, for it seems fairly certain that it was a land plant and it therefore pushes further back into antiquity the origin of land plants by some 200 million years.

Lycopodiales

This group contains two genera of living plants, Lycopodium (‘Club-mosses’) and Phylloglossum, and one fossil genus, Lycopodites. Of the 200 species of Lycopodium, the majority are tropical in distribution, but some occur in arctic and alpine regions. Phylloglossum, by contrast, is monotypic and the single species, P. Drummondii, is restricted to New Zealand, Tasmania and the south-eastern corner of Australia. Not only do the various species of Lycopodium occur in widely different climatic regions; they also occupy widely different habitats, for some are erect bog-plants, others are creeping or scrambling, while yet others are pendulous epiphytes, and this wide range of growth form is paralleled by an extremely wide range of anatomical structure. Indeed, some taxonomists have suggested that the genus should be split into at least four new genera, so different are the various species from one another. Whatever their status, the following sections and subsections of the genus are recognized by most botanists.

List 5

Members of the Urostachya never have creeping axes, but have erect or pendulous dichotomous aerial axes, according to whether they are terrestrial or epiphytic. Their roots emerge only at the base of the axes, for although they have their origin in more distal regions, they remain within the cortex (many being visible in any one transverse section of the stem). Perhaps the most important character, phylogenetically, is the lack of specialization of the sporophylls which, as a result, resemble the sterile leaves more or less closely. Another characteristic is that vegetative reproduction may frequently take place by means of bulbils. These are small lateral leafy stem-structures which occur in place of a leaf and which, on becoming detached, may develop into complete new plants. The members of the Rhopalostachya, by contrast, never reproduce by means of bulbils. They are all terrestrial and, although the first formed horizontal axes may be dichotomous, those formed later have the appearance of being monopodial, by reason of their unequal dichotomy, as also do the erect branch-systems. Roots may emerge from the leafy branches, particularly in the creeping parts of the plant.

Lycopodium selago

Of the two sections, the Urostachya (and in particular those belonging to the Selago subsection) are usually regarded as the more primitive. The British species Lycopodium selago is illustrated in Fig. 9A. Its sporophylls (Fig. 9B) are very similar indeed to the sterile leaves (Fig. 9C) and occur at intervals up the stem, fertile zones alternating with sterile. L. squarrosum shows a slight advance on this, in that the sporophylls are aggregated in the terminal regions of the axes, yet they can hardly be said to constitute a strobilus, for the sporophylls do not differ from the sterile leaves to any marked extent. All the species of the Phlegmaria subsection are epiphytic. L. phlegmaria itself is illustrated in Fig. 9K. The pendulous dichotomous branches terminate in branched strobili in which the sporophylls are smaller and more closely packed than the sterile leaves but, nevertheless, afford relatively little protection for the sporangia.

The Inundata subsection of the Rhopalostachya is represented by the British species Lycopodium inundatum (Fig. 9D). Here, the strobilus is only slightly different in appearance from the vegetative shoot, for the sporophylls (Fig. 9E) are only slightly modified for protecting the sporangia (cf. sterile leaf, Fig. 9F). Within the Clavata subsection are three more British species, L. annotinum, L. clavatum and L. alpinum, of which the first two are illustrated (Figs. 9G and 9H). In this group, the sporophylls are aggregated into very distinct strobili and are very different from the sterile leaves, for they are provided with an abaxial flange (Fig. 9I) which extends between and around the adjacent sporangia belonging to the sporophylls below (cf. sterile leaf, Fig. 9J). Whereas the strobili of L. annotinum terminate normal leafy branches, those of L. clavatum are borne on specially modified erect branches, whose leaves are much smaller and more closely appressed. There are, thus, two different kinds of sterile leaf in this species. The Cernua subsection includes a number of species with very different growth habits. L. cernuum has a creeping axis, from which arise at intervals erect branch-systems resembling tiny fir trees in being apparently monopodial (for this reason sometimes called ‘ground pines'). In this species, all the sterile leaves are alike, but in L. volubile (Fig. 9L) there are three or four kinds of sterile leaves. It is a plant with a scrambling habit and its main axes are clothed with long needle-shaped leaves arranged spirally, while the lateral branches are dorsiventral and superficially frond-like. On these branches there are four rows of leaves, two lateral rows of broad falcate leaves (Fig. 9M), an upper row of medium sized needle-like leaves and a row of minute hair-like leaves along the under side. This species, therefore, like several others in this section is highly ‘heterophyllous’. The lateral branches in the more distal regions of the plant are fertile and terminate in long narrow strobili, which are frequently branched. As in the Clavata subsection, the closely appressed sporophylls have, on their dorsal (abaxial) side, either a bulge or a flange which provides some protection for the sporangia below.

The apical region of the stem in Lycopodium differs markedly from species to species, for it is almost fiat in L. selago, yet extremely convex in L. complanatum. In the past, opinions have differed as to whether growth takes place from an apical cell, but it now appears that this is not the case and that any semblance of an apical cell is an illusion caused by studying an apex just at the critical moment when one of the surface cells is undergoing an oblique division. All species are now held to grow bv means of an ‘apical meristem’, i.e. a group of cells undergoing periclinal and anticlinal divisions.

The sporelings of all species are alike in their stelar anatomy, for the xylem is in the form of a single rod with radiating flanges. In transverse section these flanges appear as radiating arms, commonly four in number. As the plant grows, the later-formed axes of most species become more complex, the xylem splitting up into separate plates or into irregular strands. However, some species retain a simple stellate arrangement throughout their life, as in L. serration (Fig. 11F) where there are commonly five or six radiating arms of xylem. It is interesting that this species belongs to the Selago subsection which on other srounds is regarded as the most primitive, for some botanists, applying the doctrine of recapitulation, have held that the embryonic structure of a plant indicates what the ancestral adult condition was like. The stele of L. selago is similar to that of L. serratum, and the number of radiating arms of xylem may be as low as four. This is supposed to represent the ancestral condition and the Selago subsection is regarded as primitive in its stelar anatomy, as well as in its lack of a well defined strobilus. Alternating with the xylem arms are regions of phloem, separated from them by parenchyma, and the whole is surrounded by parenchymatous ‘pericycle’, outside which is an endodermis. The xylem strand of L. selago sometimes shows a slight advance on this arrangement, in that it may be separated into several areas with a variable number of radiating arms. L. clavatum has a number of horizontal plates of xylem, alternating with plates of phloem. An even greater number of such plates is found in L. volubile (Fig. nC). To some extent this trend appears to be bound up with an increasing dorsiventrality of the shoot, which reaches its culmination in the heterophyllous L. volubile. L. annotinum lends some support to this idea, for its horizontal axes are like those of L. clavatum, whereas its vertical axes are more like those of L. selago. However, exceptions are numerous and it may well be that no valid generalization of this kind can be made.

Quite a different kind of complexity is illustrated by Lycopodium squarrosum (Fig. nE), also placed in the Selago subsection. A transverse section of the stem of this species shows not only radiating arms of xylem, but also islands, within the xylem, lined with parenchyma and containing apparently isolated strands of phloem. Actually, however, the whole structure is an anastomosing one, so that no regions of phloem, or of xylem, are really isolated. This process of elaboration has gone even further in L. cernuum, where the appearance is of a sponge of xylem with phloem and parenchyma filling the holes (Fig. nD).

Throughout the genus, the stele is exarch, the protoxylem elements being clearly recognizable by their ‘indirectly attached annular thickenings’ (i.e. occasional interconnections occur between adjacent rings), while the metaxylem tracheids are either scalariform or have circular bordered pits. The phloem consists of sieve cells which are elongated and pointed, with sieve areas scattered over the side walls. The endodermis is clearly recognizable, in young stems only, when casparian strips may be seen. In older axes, however, the walls become heavily lignified along with the cells of the inner cortex and their identity becomes obscured. This lignified region extends through most of the cortex in some species, whose stems are consequently hard and wiry, while in other species, e.g. L. squarrosum, the stem may be thick and fleshy. Stomata are present in the epidermis of the stem and in the leaves where, in some species, they are on both surfaces (‘amphistomatic’) and, in others, only on the under side (‘hypostomatic’). The leaves of some species are arranged in a whorled or a decussate manner, but in most are spirally arranged. However, in these, the phyllotactic fractions are said to be unlike those of other vascular plants in forming part of the series 2/7, 2/9, 2/11 etc. (whereas the normal phyllotactic fractions, 1/2, 1/3, 2/5, 3/8, 5/13 etc., are dervied from the Fibonacci series). Each leaf receives a single trace, which has its origin in one of the protoxylems of the stem stele and continues into the leaf as a single unbranched vein composed entirely of spirally thickened tracheids. It is of interest that, in L. selago, the bulbils also receive this kind of vascular bundle, for this supports the view that, at this level of evolution, there is no clear morphological distinction between the categories ‘leaf’ and ‘stem’. This is further supported by the fact that leaf primordia may be transformed by suitable surgical treatment into regenerative buds.

The so-called ‘roots’, too, show varying degrees of similarity to stems. All, except the first root of the sporeling, are adventitious and endogenous in origin, arising in the pericycle, and they are peculiar in not bearing endogenous laterals. Instead, they branch dichotomously (very regularly in some species). They are provided with a root cap and their root-hairs are paired (a most peculiar arrangement). The majority are diarch with a crescent-shaped xylem area, but in some species the stele is very similar to that of the stem, as in Lycopodium clavatum, where the xylem takes the form of parallel plates.

Lycopodium clavatum

Variations from species to species in the shape of the sporophylls have already been described. In addition, there is considerable variation in the manner in which the sporangium is borne in relation to the sporophyll. In some, e.g. Lycopodium selago and L. inundatum, the sporangium is in the angle between the sporophyll and the cone axis, i.e. it is axillary. In others, e.g. L. cernuum and L. clavatum (Fig. 10C), the sporangium is borne on the adaxial surface of the sporophyll and may be described as ‘epiphyllous’. The sporangial initials arise at a very early stage in the ontogeny of the strobilus, normally on the ventral side of the sporophyll, but in some species actually on the axis, whence they are carried by subsequent growth changes into the axil. The first sign of sporangial initials is the occurrence of periclinal divisions in a transverse row of cells (three to twelve in number) (Fig. 10A). The innermost daughter cells provide the archesporial cells by further division and also contribute to the stalk of the sporangium, while the outermost cells (the jacket initials) give rise to the wall of the sporangium (Fig. 10B). This is three cells thick just before maturity, but then the innermost of the layers breaks down to form a tapetal fluid. Like the sterile leaves, the sporophyll has a single vein, which passes straight out into the lamina, leaving the sporangium without any direct vascular supply. The mature sporangium is kidney-shaped and dehisces along a transverse line of thin-walled cells, so liberating the very numerous and minute spores into the air.

In some species, the spores germinate without delay, while still on the surface of the ground, but in others there may be a delay of many years, by which time they may have become deeply buried. Surface living prothalli are green and photosynthetic, but subterranean ones are, of necessity, colourless and are dependent on a mycorrhizal association for their successful development. Indeed, a mycorrhizal association appears to occur in all species growing under natural conditions, whatever their habit. As a generalization, it may be said that those species inhabiting damp tropical regions germinate rapidly and have green prothalli, whereas those of cooler regions tend to germinate slowly and produce subterranean prothalli. Lycopodium selago is interesting in this respect, for it shows variability. Fig. 10E illustrates a surface living prothallus with photosynthetic upper regions, in addition to the fungal hyphae in the lower parts (and rhizoids). Fig. 10F, on the other hand, is of a subterranean prothallus, with fungal hyphae in the lower regions but covered all over with rhizoids. Archegonia and antheridia are restricted to the upper parts in both cases. L. cernuum provides an example of a surface-living prothallus (Fig. 10G). It is roughly cylindrical and the upper regions bear numerous green photosynthetic lobes, among which are borne the gametangia. In L. clavatum (Fig. 10H) and L. annotinum the prothallus is colourless and subterranean; it is an inverted cone with an irregular fluted margin, growing by means of a marginal meristem which remains active for many years, and the gametangia are developed over the central part of the upper surface. Epiphytic species, e.g. L. phlegmaria, also have colourless prothalli, but they are very slender, they branch and they exhibit pronounced apical growth.

Archegonia and antheridia each arise from a single superficial cell in which a periclinal division occurs. The subsequent cell divisions in the antheridial initials are similar to those described for Tmesipteris (Fig. 7), but the mature antheridium differs in being sunken into the tissues of the prothallus. The archegonium differs from that of Tmesipteris in having several neck canal cells, which vary in number according to whether the prothallus is subterranean or surface living. In the latter species, the neck is very short e.g. Lycopodium cernuum (Fig. iol), and there may be just a single canal cell, apart from the ventral canal cell. At the other extreme, the number of canal cells may be as high as fourteen in L. complanatum (in the Clavata subsection), while L. selago is intermediate, with about seven. Various stages in the development of L. clavatum are illustrated in Figs. 10J-N. At maturity all the canal cells break down and part of the neck may also wither (Fig. 10O). The antherozoids are pear-shaped and swim by means of two flagella at the anterior end, attracted chemotactically by citric acid diffusing from the archegonium.

The orientation of the embryo in Lycopodium is endoscopic and this is determined at the first division of the zygote, with the laying down of a cross wall in a plane at right angles to the axis of the archegonium (Fig. 10U). The outermost cell, called the ‘suspensor’, undergoes no further divisions, but the innermost cell gives rise to two tiers of four cells, called the ‘hypobasal’ and ‘epibasal’ regions respectively (Fig. 10W). It is from the epibasal (innermost) tier that the young plant is ultimately derived, by further divisions. The hypobasal region remains small in some species, and in others it swells up into a structure commonly called a ‘foot’. L. clavatum is an example of the latter and various stages are illustrated in Figs. 10U-Z. In Fig. 10X, the three regions of the embryo are clearly demarcated (the suspensor cell, ‘s’; the middle hypobasal region, already beginning to swell into a foot, ‘f’; the epibasal region with a stem apex, ‘x’, becoming organized), and the axis of the embryo has bent through a right angle. This bending of the axis proceeds further in Fig. 10Y and is completed in Fig. 10Z, where, by turning through two right angles, the stem apex is pointing vertically upwards. The first root, ‘r’, is seen to be a lateral organ, not forming part of the axis of the embryo, as indeed is the case in all pteridophytes: not until the level of the seed plants does the root (radicle) form part of the embryonic spindle. L. selago (Fig. 10P) is similar, except that the hypobasal region does not swell up into a large foot.

Lycopodium cernuum (Figs. 10Q-T) is an example of a very different kind of embryology. As in L. selago, the hypobasal region remains relatively small, but the organizing of a stem apex is considerably delayed. The epibasal portion breaks through the prothallial tissue and swells out into a tuberous ‘protocorm’, ‘t’. Roughly spherical at the start, it is provided with rhizoidal hairs and mycorrhizal fungus. On its upper surface a cylindrical green leaf (‘protophyll’), T, appears and then, as the protocorm slowly grows, further protophylls appear in an irregular manner. This stage may persist for a long time and secondary protocorms %’ may be formed as shown in Fig. 10T. Finally, however, a stem apex ‘x’ becomes organized and a normal shoot grows out. This type of development has led, in the past, to much speculation as to its phylogenetic significance, for the protocorm was held by some to represent an atavistic survival of an ancestral condition. However, Wardlaw has offered an alternative explanation, based on the metabolism of the prothallus and young sporophyte in the various species of Lycopodium. He suggests that an abnormally high carbon/nitrogen ratio may delay the organization of a stem apex and may lead, also, to a swelling of the tissues, such being expected where mycorrhizal nutrition is supplemented by photosynthesis. On this basis, the protocorm might well be regarded as a derivative and retrograde development, rather than as a sign of primitiveness.

When all facts are considered, it is Lycopodium selago which is usually regarded as the most primitive species, in lacking an organized strobilus, in having a relatively simple vascular structure and in showing variability in behaviour in its prothallus, but such conclusions can only be speculative in the absence of clear fossil evidence. While there are fossil remains, known as Lycopodites, they contribute little to these discussions. No petrified specimens have been found and some of the mummified remains are now known to be those of conifers. Some had well organized strobili; others did not. Lycopodites stockii, from the Lower Carboniferous of Scotland, appears to have been heterophyllous, with its leaves in whorls, and to have had a terminal cone as well as scattered sporophylls among the sterile leaves. Clearly, therefore, this species was very different from the modern L. selago and, in some respects, was nearer to some members of the Phlegmaria subsection.

The sporophyte of Phylloglossum Drummondii, illustrated in Fig. 10D, is never more than about 4 cm high and appears above ground only during the winter months, when it develops a few cylindrical leaves like the protophylls of Ly cop odium cernuum. The most robust specimens develop, in addition, a single erect stem terminating in a tiny strobilus. During the hot summer months, when the ground is baked hard, all the aerial parts wither and the plant survives this unfavourable season as a tuber. Each year a new tuber is formed (sometimes two or even three) from the apex of a lateral stem-like structure, which grows out and downwards. This parallel with the behaviour of the protocorm of L. cernuum (Fig. 10T) has led to the suggestion that Phylloglossum exhibits ‘neoteny’, in being able to produce sporangia while still in an embryonic stage of development. Whatever the truth of this, it would certainly seem that some of its peculiarities are adaptations which enable it to survive adverse environmental conditions as a geophyte. From the morphogenetic point of view, it is possible to see the tuberization as a response to a high carbon/nitrogen ratio, since the prothallus is both photosynthetic and mycorrhizal. Perhaps all three ‘causes’ may apply, for they are not incompatible with each other and merely represent different ‘grades of causality’.

Chromosome counts for Phylloglossum show a haploid number n = about 255, with many unpaired chromosomes at meiosis, suggesting a high degree of hybridization in its ancestry. Such a high number is believed by some to be characteristic of primitive plants, and in this connection it is interesting to find that Lycopodium selago has a haploid number n=i30, whereas species in other sections of the genus have lower numbers (L. clavatum and L. annotinum n = 34). But such a belief is justified only as a generalization. High chromosome numbers may well point to ancient origins in the majority of cases, but not in all, for polyploidy could have occurred at any stage in the evolution of an organism. Whenever it did occur, further evolution would be retarded because of the masking of subsequent mutations. Thus, if it occurred long ago, the ancient condition would have become ‘fixed’, as may have happened in L. selago; whereas, if it had happened recently, it would be possible for an advanced morphological condition to be associated with a high chromosome number, as in Phylloglossum perhaps.

By contrast with the Protolepidodendrales and the Lycopodiales, which are homosporous, the three remaining orders of the Lycopsida (Lepidodendrales, Selaginellales and Isoetales) are heterosporous. Another feature that they share is the possession of a ligule, on the basis of which they are sometimes grouped together as the Ligulatae. The ligule is a minute tongue-like membranous process, attached by a sunken ‘glossopodium’ to the adaxial surface of the leaves and the sporophylls. A study of living heterosporous lycopods shows that it reaches its maximum development while the associated primordium of the leaf or the sporophyll is still quite small. The mucilaginous nature of the cells and the lack of a cuticle have led to the suggestion that the ligule may keep the growing point of young leaves and young sporangia moist, but the fact is that no-one knows its true function. It may even be a vestigial organ whose function has been lost.

Lepidodendrales

The Lepidodendrales, over 200 species of which are known, first appeared in Lower Carboniferous times and reached their greatest development in the Upper Carboniferous swamp forests, in which members of the Lepidodendraceae, Bothrodendraceae and Sigillariaceae were co-dominant with the Calamitales and formed forests of trees 40 m or more in height. The fourth family, Pleuromeiaceae, is represented by a much smaller plant, Pleuromeia, from Triassic rocks, and approached more nearly to the modern Isoetales. The Carboniferous genera had stout trunks, some with a crown of branches, others hardly branching at all, but all possessed the same type of underground organs, known collectively as Stigmarian axes. Some species of Lepidodendron, (e.g. L. obovatum, Fig. 12A) showed very regular dichotomies in its crown of branches, but others approximated to a monopodial arrangement because of successive unequal dichotomies. While the trunks and branches of all species of Lepidodendron and Lepidophloios were protostelic and exarch, there was nevertheless considerable variation in stelar anatomy, from species to species, and from place to place within one individual. Some species had solid protosteles, others medullated protosteles; some had abundant secondary wood produced by a vascular cambium, some had little and others had none at all; in some, the stele of the trunk had secondary wood, while that of the branches lacked it altogether. Thus, Lepidodendron pettycurense and L, Rhodumnense (both Lower Carboniferous species) had solid protosteles, the former having secondary wood in addition, but the latter being without it. Lepidodendron selaginoides (= L. vasculare), from the Coal Measures, provides an interesting case of partial medullation, for the central region of the axis consisted of a mixture of parenchyma and tracheids, round which was a solid ring of tracheids. The secondary wood of this species was often excentric in its development, as illustrated in Fig. 11B.

Various Lycopod steles

Lepidophloios Wuenschianus, from the Lower Carboniferous of Arran, is known in considerable detail, for examples have been found in which portions of the stele from various levels had fallen into the rotted base of the trunk before petrifaction occurred. This has made it possible to discover something about the growth processes taking place in the young aerial stem. The primary wood near the base was solid and only 55 mm across, half way up the trunk it was medullated, while near the top (Fig. 11 A) it was 15 mm across and had a hollow space in the centre of the medulla. It is concluded that, as the stem grew, its apical meristem grew more massive and laid down a much broader procambial cylinder. Meantime, the cambium in the lower regions had laid down more secondary wood than higher up, with the result that the total diameter of the wood (primary and secondary together) was about the same throughout the length of the trunk (about 7 cm). In proportion to the overall diameter of the trunk (40 cm), however, this quantity of wood is surprisingly small, when compared with that of a dicotyledonous tree, where most of the bulk is made up of wood. The difference probably lies in the fact that the wood of modern trees is concerned with two functions, conduction and mechanical support, whereas the wood of Lepidodendrales was concerned only with conduction. Mechanical support was provided mainly by the thick woody periderm which was laid down round the periphery of the trunk.

The metaxylem was composed of large tracheids with scalariform thickenings, while the protoxylem elements were much smaller and frequently had spiral thickenings. The secondary wood consisted of radial rows of scalariform tracheids and small wood-rays, through which leaf-traces passed on their way out from the protoxylem areas. In most specimens the phloem and even some of the cortex had decayed before petrifaction occurred, but what is known of the phloem suggests that it was small in amount and very similar to that of modern lycopods.

The primary cortex was relatively thin-walled and, within it, a number of different regions are recognizable. Of these, the most interesting is the so-called ‘secretory tissue’, made up of wide thin-walled cells, whose horizontal walls became absorbed in the formation of longitudinal ducts. Each leaf-trace, as it passed through this region, acquired a strand of similar tissue which ran parallel with it before splitting into two ‘parichnos strands’ on entering the base of the leaf. It is believed that the secretory tissue was in some way connected with the aeration of the underground organs, providing an air path from the stomata of the leaves, through the mesophyll to the parichnos strand and so to the secretory zone, which was continuous with a similar region in the cortex of the Stigmarian axes.

Lepidodendron

The leaves, known as Lepidophyllum, were borne in a spiral with an angle of divergence corresponding to some very high Fibonacci fraction such as 55/144, 89/233, etc. They were linear, up to 20 cm long, triangular in cross-section and with stomata in two longitudinal grooves on the adaxial side. The vascular strand remained unbranched as it ran the length of the leaf. The leaves were shed from the trunk and larger branches by means of an absciss layer, and the shape of the remaining leaf base and scar provides important details for distinguishing the various genera and species. Fig. 12B shows the appearance of the trunk of a Lepidodendron where, characteristically, the leaf bases were elongated vertically. In some species, the leaf bases became separated slightly as the trunk increased in diameter, but in others they remained contiguous, even on the largest axes. No doubt this was brought about to some extent by an increase in the size of the leaf base, much as a leaf scar becomes enlarged on the bark of many angiospermous trees, but such increase must have been relatively slight, for otherwise the leaf bases would have become much broader in proportion to their height. Evidently, therefore, the largest leaf bases must have been large from the start, from which it follows that the axes bearing them must also have been large, even when young. Details of a typical Lepidodendron leaf base are illustrated in Fig. 12C. Within the area of the leaf scar (2) are to be seen three smaller scars, representing the leaf-trace (3) and the two parichnos strands (4). Above this lies the ligule pit (1) and, in some species, below it are two depressions that were once thought to be associated with the parichnos system, but are now known to be caused by shrinkage of thin-walled cells within the leaf cushion.

Lepidophloios is distinguished by its leaf bases being extended horizontally, instead of vertically. Otherwise, the anatomy of the trunks is indistinguishable from that of Lepidodendron. Indeed, it has been suggested that the differences do not warrant a separation into two genera. However, there were differences; in the way the cones were borne. In Lepidodendron, they were nearly always terminal, whereas in Lepidophloios, they were borne some distance behind the branch tip in a cauliflorous manner.

The cones of both genera are known as Lepidostrobus and they consisted of a central axis around which sporophylls were arranged in a compact spiral, their apices overlapping so as to protect the sporangia. Further protection was afforded by a dorsal projection, or ‘heel’, as illustrated in the idealized longitudinal section, Fig. 12D. The cones varied in length from 5 cm to over 40 cm and must have looked like those of modern conifers. Some cones contained only megasporangia, others only microsporangia, while others were hermaphrodite. In the latter, the megasporophylls were at the base and the microsporophylls towards the apex, as illustrated in Fig. 12D. This is the reverse of the arrangement in gymnosperms and angiosperms, where the microsporangial organs lie below the megasporangial whenever they happen to be associated in a hermaphrodite ‘flower’. The sporangia of Lepidostrobus were elongated and attached throughout their length to the ‘stalk’ of the sporophyll, which was relatively narrow, compared with the expanded apex of the sporophyll (Fig. 12E). The sporangium wall was only one cell thick at maturity and dehisced along its upper margin. Megaspores and microspores must have been produced in enormous numbers, for they are extremely abundant in all coal-measure deposits. Some megaspores have been found with cellular contents, representing the female prothallus, retained within the megaspore wall (‘endosporic’) as in Selaginella today, and occasionally archegonia can be recognized.

The number of megaspores produced within each megasporangium varied considerably from species to species, and in some was restricted to one. In Lepidocarpon (Figs. 12F and 12G) the megaspore was retained in the sporangium, which, in turn, was enveloped by two flanges from the stalk of the sporophyll. The whole structure was shed like a seed from the parent plant and has been regarded by some botanists as actually being a stage in the evolution of a seed. It would be much safer, however, to regard Lepidocarpon as merely analogous to a seed, for the sporophyll flanges are quite unlike the integuments of true seeds, except perhaps in function. It is not known whether the microspores germinated within the slit-like ‘micropyle’ while the megasporophyll was still on the tree, or whether it did so after it had fallen to the ground.

Sigillaria (Fig. 12H) is characterized by the arrangement of its leaf bases in vertical rows (Fig. 12I). It branched much less than Lepidodendron or Lepidophloios and it bore its cones in a cauliflorous manner. Furthermore, the leaves were much longer, up to 1 m, grass-like and, in some species, had two veins, possibly formed by the forking of a single leaf-trace. Species from the Upper Carboniferous were similar to Lepidodendron and Lepidophloios in their internal anatomy, having a medullated protostele with a continuous zone of primary wood. Some of the Permian species, e.g. S. Brardi, however, showed a further reduction of the primary wood, which was in the form of separate circummedullary strands. This is most interesting, for it represents the culmination of a trend which was also taking place, at the same time, among several groups of early gymnosperms, from the solid protostele, through medullated protosteles (first with mixed pith and then with pure pith) to a pith surrounded by separate strands of primary wood.

From a distance, Bothrodendron must have looked very similar to Lepidophloios, for it had a stout trunk with a crown of branches covered with small lanceolate leaves and its cones (Bothrodendrostrobus) were borne in a cauliflorous manner. It differed, however, in the external appearance of the trunk, for it had circular leaf scars that were almost flush with the surface.

The underground organs of all the genera of Lepidodendrales so far described were so similar that they are all placed in the form genus Stigmaria, and many are placed in a single artificial species, S. ficoides. The base of the trunk bifurcated once and then immediately again, to produce four horizontal axes, each of which continued to branch dichotomously many times in a horizontal plane. These Stigmarian axes were most remarkable structures in many respects. Thus, even at their growing points, perhaps 10 m from the parent trunk, they were frequently as thick as 4 cm. They bore lateral appendages, commonly called ‘rootlets’, in a spiral arrangement. These were up to i cm in diameter and were completely without root hairs. Internally they show a remarkable resemblance to the rootlets of the modern Isoetes in having had a tiny stele separated from the outer cortex by a large space, except for a narrow flange of tissue (Fig. 12J). In origin, they were endogenous, although only just so. The axes on which they were borne were peculiar in being completely without metaxylem. In the centre was either pith or a pith-cavity, round which were protoxylem regions directly in contact with a zone of secondary wood. This consisted of scalariform tracheids interspersed with small wood-rays, but there were also very broad rays (through which the rootlet traces passed) which divided the wood into very characteristic wedge-shaped blocks.

The true nature of Stigmarian axes has long been a problem to morphologists, for although doubtless they performed the functions attributed to roots in higher plants (absorption and anchorage), yet they were different in so many respects from true roots and, at the same time, were so different from the aerial axes that they appear to have belonged to a category of plant organization that was quite unique. Even the nature of the ‘rootlets’ is open to question, for specimens of Stigmarian axes are known which bore leaf-like appendages instead of rootlets. Once more one is forced to the conclusion that the categories root, stem and leaf have no clear distinction at the lower levels of evolution.

Pleuromeia (Fig. 13A) was a much smaller plant than the other members of the Lepidodendrales, for its erect unbranched stems were little more than 1 m high and 10 cm in diameter. The lower parts of the stem were covered with spirally arranged leaf scars, while the upper parts bore narrow pointed ligulate leaves about 10 cm long, attached by a broad base. The plant was heterosporous and dioecious, and the sporangia were borne in a terminal cone made up of numerous spirally arranged circular or reniform sporophylls. Although early descriptions described the sporangia as on the abaxial side of the sporophyll, most morphologists believe this to be an error and it is usually accepted that, as in all other lycopods, they were on the adaxial side. Verification of this must await the discovery of better preserved specimens, for no petrified material has yet been found. For this reason, also, little is known of the internal anatomy of the plant.

Below the ground, Pleuromeia was strikingly different from the other members of the Lepidodendrales, for, instead of having spreading rhizomorphs of the Stigmaria type, it terminated in four (or sometimes more) blunt lobes. From these were produced numerous slender forking rootlets, very similar anatomically to those of Stigmaria and also to those of Isoetes. Indeed, Pleuromeiais commonly regarded as a link connecting the Isoetales with the Carboniferous members of the Lepidodendrales.

Isoetales

Apart from the fossil genus Nathorstiana, the Isoetales contain only the two living genera Isoetes and Stylites.

The genus Isoetes is world-wide in distribution, some seventy species being known, of which three occur in the British flora and are commonly called ‘Quillworts’. I. lacustris and I. echinospora grow submerged in lakes or tarns, while I. hystrix favours somewhat drier habitats. Most of the plant is below the level of the soil, with only the distal parts of the sporophylls visible. These are linear structures from 8 to 20 cm long in I. lacustris, but up to 70 cm in some species growing in N. America and in Brazil. They constitute the only photo synthetic parts of the plant and, as in many aquatic plants, they contain abundant air spaces (lacunae). The expanded bases of the sporophylls are without chlorophyll and overlap one another to form a bulb-like structure which surmounts a peculiar organ, usually referred to as a ‘corm’. The true morphology of the corm has long been the subject of controversy, for it is obscured by a remarkable process of secondary growth, involving an anomalous cambium. This produces small quantities of vascular tissue from its inner surface and large quantities of secondary cortex towards the outside. This secondary cortex dies each year, along with the sporophylls and roots attached to it, and it becomes sloughed off when the new year’s growth of secondary cortex is produced. Vertical growth of the corm is extremely slow, with the result that the body of the plant is usually wider than it is high.

Fig. 13F is a diagrammatic representation of a vertical section through an old plant of Isoetes. To the right and left are the shrivelled remains of the previous year’s growth, the several sporophyll-traces and root-traces being visible within it. All the rest represents the present year’s growth surrounding the perennial central regions. Occupying the centre is a solid protostele, the lower part of which is extended into two upwardly curving arms, so that the overall shape resembles an anchor (Fig. 13G). This is made up of mixed parenchyma and peculiar isodiametric tracheids with helical thickenings. Towards the outside the tracheids are arranged in radial rows but, nevertheless, they are of primary origin (indeed, some workers hold that the whole of the primary wood is protoxylem). Surrounding this primary wood is a narrow zone of phloem (not shown in Fig. 13F), and outside this is the tissue produced centripetally by the anomalous cambium. This commonly consists of a mixture of xylem, phloem and parenchyma and is described by the non-committal term ‘prismatic tissue’. The cambium, represented in Fig. 13F as a broken line, cuts through the sporophyll-traces and root-traces of previous years, leaving their truncated stumps still in contact with the primary wood.

Pleuromeia, Stylites and Isoetes

What little vertical growth there is takes place by means of apical meristems at the top and bottom of the corm. The lower of these is extended as a line beneath the anchorshaped primary xylem and is buried deeply in a groove. Roots arise endogenously along the sides of this groove in a very regular sequence and are carried round on to the undersides of the newly formed cortex. The stem apex is also deeply sunken between the ‘shoulders’ of the corm and is said to contain a group of apical initial cells. Sporophylls arise in spiral sequence (with a phyllotaxy of 3/8, 5/13 or 8/21 in mature plants) and, as new secondary cortex is formed, they are carried up on to the shoulders.

Stages in the development of the young sporophylls are illustrated in Figs. 13J-L. At a very early stage, when the primordium is only a few cells high, one conspicuous cell on its adaxial surface undergoes a periclinal division to produce a ligule primordium (i). This soon gives rise to a membranous ligule a few mm long which, for a time, is much larger than the young sporophyll. Next, a velum initial appears, from which is developed the velum (2) — a flange of tissue which partly hides the sporangium in the mature sporophyll, except for an oval opening called the ‘foramen’. The sporangium (3) arises as the result of periclinal divisions in a group of superficial cells near the base of the sporophyll, on its adaxial side. The inner daughter cells are potentially sporogenous, while the outer (peripheral) cells give rise to the sporangium wall, three or four cells thick. Isoetes is peculiar among living plants in that some of the potentially sporogenous cells become organized into trabeculae of sterile tissue which cross the sporangium in an irregular manner. They subsequently become surrounded by a tapetal layer which is continuous with the one derived from the innermost layer of the sporangium wall.

The general appearance of the base of a mature sporophyll is indicated in Figs. 13I and 13H, representing a longitudinal section and an adaxial surface view respectively. The sporangia of the Isoetales are larger than those of any other living plant and have a very high spore content indeed. The sporophylls formed earliest in the year and which, therefore, lie outermost on the apex of the corm are megasporangial and contain several hundred megaspores. Those formed later are microsporangial and are estimated to contain up to a million microspores each. Finally, a few sporophylls with abortive sporangia are produced late in the season.

There is no special dehiscence mechanism and the spores are released only when the sporophylls die and decay, as they become sloughed off at the end of the season. The first cell division within the microspore is an unequal one which cuts off a small ‘prothallial cell’. The other cell is called the ‘antheridial cell’ since, by successive divisions (Figs. 13M and 13N), it gives rise to a jacket of four cells surrounding a central cell from which four antherozoids are formed (Fig. i3o). These are spiral and multiflagellate (Fig. 13P) and are released by the cracking of the microspore wall. As already mentioned, this mode of development, where the prothallus is retained within the spore wall, is described as ‘endosporic’.

The female prothallus likewise is endosporic. Within the megaspore, free nuclear divisions take place for some time, i.e. nuclei continue to divide without any cross-walls being laid down between them. Then, when about fifty such nuclei have become distributed round the periphery of the cytoplasm, cross-walls are slowly formed, starting in the region immediately beneath the tri-radiate scar, but gradually spreading throughout. Meanwhile, the megaspore wall ruptures at the tri-radiate scar and an archegonium is formed in the cap of cellular tissue which is thereby exposed. Stages in the development of the archegonium are illustrated in Figs. 13Q-S. If fertilization does not occur immediately, further archegonia may develop among the rhizoids that cover the apex of the gametophyte.

Stages in the development of the young sporophyte are illustrated in Figs. 13 T-Y, in which the megaspore is supposed to be lying on its side, as is commonly the case. The first division of the zygote is in a plane at right angles to the axis of the archegonium, or slightly oblique to it. That part of the embryo formed from the outermost half, designated ‘the foot’, is indicated in the figures by oblique shading. As growth proceeds, the orientation of the embryo changes so that the first leaf and the stem apex are directed upwards, while the first root is directed obliquely downwards. It is of interest that there is no quadrant specifically destined to produce a stem apex, and that it appears relatively late in a position somewhere between the first leaf and the first root. In some species, there are no clearly defined quadrants at all.

Despite the absence of a suspensor, the embryology of Isoetes may be described as endoscopic, since it is from the inner half of the dividing zygote that the shoot is ultimately formed.

For some time, the young embryo continues to be enclosed within a sheath of prothallial tissue which grows out round it, but ultimately the various organs break through and the first root penetrates the soil.

A chromosome count of n = 10 has been obtained in one species of Isoetes, and of n = 54 - 56 in several others.

Isoetes is clearly a remarkable genus, not only in its peculiar method of secondary thickening, but also in the fact that all its leaves are, at least potentially, sporophylls. For this reason, some morphologists regard the upper half of the corm as representing a cone axis. The lower half they regard as a highly reduced rhizomorph, homologous with Stigmarian axes, and this is supported, not only by the regular arrangement of the roots on the corm, but also by the extraordinary similarity of the roots to Stigmarian rootlets internally. If this view is correct, then the stem, as such, must have become completely suppressed, along with its leaves.

Stylites was unknown until 1954, when it was first discovered, forming large tussocks round the margins of a lake at an altitude of 4,750 m in the Peruvian Andes. Since then, it has been examined in great detail by Rauh and Falk, who claim that there are two species. Stylites is no less remarkable than Isoetes, for it likewise exhibits a kind of anomalous secondary thickening, though less active, and all its leaves are potential sporophylls. It differs from Isoetes, however, in having limited powers of vertical growth and in being able to branch, both dichotomously and adventitiously, so as to form the characteristic tussocks. Two plants are illustrated in longitudinal section, one young and unbranched (Fig. 13C), the other older and branching (Fig. 13D). Perhaps the most remarkable feature is the way in which the roots are borne up one side only and receive their vascular supply from a rod of primary wood which is quite distinct from that supplying the sporophylls; the two run side by side within the axis (Fig. 13E). The nature of the axis is, therefore, even harder to interpret than in Isoetes. Rauh and Falk draw a comparison with the Cretaceous Nathorstiana (Fig. 13B) in which the roots arise from a number of vertical ridges round the base of the stem. This in turn may be compared with Pleuromeia and ultimately, therefore, with the Lepidodendrales.

Selaginellales

This group contains two genera, one living (Selaginella) and one fossilized (Selaginellites). More than 700 species of Selaginella are known, of which some occur in temperate regions, but the vast majority are confined to the tropics and subtropics, where they grow in humid and poorly illuminated habitats, such as the floor of rain-forests. Some, however, are markedly xerophytic, inhabiting desert regions, and are sometimes called ‘resurrection plants’ because of their extraordinary powers of recovery after prolonged drought. Relatively few are epiphytes, unlike Lycopodium. Some form delicate green mossy cushions, others are vine-like, with stems growing to a height of several metres, while many have creeping axes, from which arise leafy branch systems that bear a striking superficial resemblance to a frond. Hieronymus divided the genus into the following sections and subsections:

List 6

The Homoeophyllum section is a small one, consisting of fewer than fifty species, all of which are isophyllous and have spirally arranged leaves. The only native British species, Selaginella selaginoides (= S. spinosa, = S. spinulosa) (Fig. 14A), is a typical example of this kind of organization and is placed in the subsection Cylindrostachya because the spiral arrangement extends also to the fertile regions. Species belonging to the Tetragonostachya subsection differ in that the sporophylls are arranged in four vertical rows, giving the cone a four-angled appearance. All the members of the Homoeophyllum section are monostelic, but S. selaginoides is peculiar in that the stele of the creeping region is endarch (Fig. 14I), whereas that of the later-formed axes is exarch (Fig. 14H), as in all other species. According to Bruchmann there is a limited amount of secondary thickening in the so-called hypocotyl region of this species; this is the only record of cambial activity in the whole genus.

The Heterophyllum section is characterized by a markedly dorsiventral symmetry and by anisophylly, for the leaves are arranged in four rows along the axis, two rows of small leaves attached to the upper side and two of larger ones attached laterally. The fertile regions, however, are isophyllous and the cones are four-angled, which makes them very clearly distinguishable from the vegetative regions (Fig. 14F). The section is divided, somewhat arbitrarily, on the number of megasporophylls in the cone and is further subdivided on the number of steles in the axis.

Most commonly the axis is monostelic and contains a ribbon-shaped stele, e.g. Selaginella flabellata (Fig. 14K), but some species have more complicated stelar arrangements. S. Kraussiana, now naturalized in parts of the British Isles, has a creeping habit and has two steles which run side by side (Fig. 14J), except at the nodes where they interconnect. S. Braunii (Fig. 14E) is one of the many species which have a creeping stem with erect frond-like branch systems: the creeping axis is bi-stelic, with one stele lying vertically above the other, while the erect axes are monostelic. S. Willdenowii is a climbing, or vine-like, species and may have three ribbon shaped steles (Fig. 14L) or even four. The most complex of all is S. Lyallii, where the creeping axis is dicyclic and the aerial axes are polystelic. The central stele of the creeping axis is a simple ribbon of metaxylem, without any protoxylem, surrounded by phloem, pericycle and endodermis. This is surrounded by a cylindrical stele which is amphiphloic (i.e. has phloem to the inside as well as to the outside of the xylem) and is bounded, both externally and internally, by endodermis. Both steles play a part in the origin of the many steles in the aerial axis, which number twelve or thirteen, four of them being main ones to which the leaf traces are connected, while the rest are accessory steles.

It is important to realize that, however complex the stem of a mature plant may be, the young sporeling Selaginella is invariably monostelic, there being a gradual transition along the axis until the adult condition is achieved. This observation has naturally, in the past, led to the supposition that species which are monostelic throughout are more primitive than the more complex species. However, much caution is necessary before accepting this view. In the first place, not all monosteles are directly comparable (e.g. S. selaginoides and S. Braunii). In the second place, it would seem that some members of the monostelic Homoeophyllum group are highly advanced in other respects. Thus, S. rupestris and S. oregana (in the Tetragonostachya subsection) are remarkable for the possession of vessels in their xylem. Among the tracheids are lignified elements whose transverse end-walls have dissolved, leaving a single large perforation plate so that, like drain pipes placed end to end, they provide long continuous tubes. While it is true that vessels are known in some other pteridophytes, they are not of this advanced type which occurs, elsewhere, only among the flowering plants. The metaxylem tracheids have scalariform thickenings, while the protoxylem elements are helically thickened and exhibit a feature which is found elsewhere only in Isoetes—viz. the helix may be wound in different directions in different parts of the cell.

Selaginella selaginoides

The phloem, composed of parenchyma and sieve cells, is very similar to that of Lycopodium and is separated from the xylem by a region of parenchyma one or two cells thick. To the outside of it is a region of pericycle, and then comes a trabecular zone which is characteristic of Selaginella. This zone differs markedly in detail from species to species, but is usually a space, crossed in an irregular fashion by tubular cells or by chains of parenchyma cells. Endodermal cells are recognizable also in this region because of their Casparian bands, but it sometimes happens that a single Casparian band may encircle a bunch of several tubular cells. Whatever the exact constitution of the zone, however, it is very delicate, with the result that the stele usually drops out of sections cut by hand. The outer regions of the stem are frequently made up of thick-walled cells and the epidermis is said to be completely without stomata.

In Selaginella selaginoides, roots arise in regular sequence from a swollen knot of tissue in the hypocotyl region, but in most creeping species they arise at intervals along the under side of the stem. They are simple monarch structures, which branch dichotomously in planes successively at right angles to each other, as they grow downwards into the soil. Root-caps and root-hairs are present, just as in the roots of other plants. A mycorrhizal association has been demonstrated in S. selaginoides. In species with aerial branches, the roots are associated with peculiar organs, usually referred to as ‘rhizophores’, and some morphologists describe the roots as borne on them, while others describe the rhizophores as changing into roots on reaching the soil. Of these, the second interpretation is probably the more accurate. Rhizophores are particularly well developed in climbing species, such as S. Willdenowii (Fig. 14G), where they grow out from ‘angle meristems’ which occur in pairs, one above and one below, at the junction of two branches. In some species, only one of these is active while the other remains dormant, as a small papilla. The active one grows into a smooth shiny forking structure without leaves. Its branches are without root-caps until they reach the soil, but then root-caps appear and all subsequent branches take on the appearance of typical roots. This is the normal behaviour, but the fate of the angle meristems appears to be under the influence of auxin concentrations, for damage to the adjacent branches may result in their giving rise to leafy shoots, instead of rhizophores. It is clear that the rhizophore fits neither into the category ‘stem’ nor into the category ‘root’, but exhibits some of the characters of each. It is not surprising, therefore, that in the days when botanists believed in the reality of these morphological categories, the rhizophores of Selaginella were the subject of much argument.

The stem apex shows an interesting range of organization, from species to species, for those with spirally arranged leaves tend to have a group of initial cells, while dorsiventral species usually have a single tetrahedral apical cell. Leaf primordia are formed very close to the stem apex and, in some species, appear to arise from a single cell. They give rise to typical microphylls, receiving a single vascular bundle which continues into the lamina as an unbranched vein. The ligule, which is present on every leaf and sporophyll, appears early in their ontogeny and develops from a row of cells arranged transversely across the adaxial surface near the base of the primordium. When fully grown it may be fan-shaped or lanceolate and has a swollen ‘glossopodium’ sunken into the tissue of the leaf. There is much variation, according to species, in the structure of the lamina of the leaf, for some species possess only spongy mesophyll, while others have a clearly defined palisade layer also. In some, the cells of the upper epidermis and, in others, some of the mesophyll cells contain only a single large chloroplast, a feature which is reminiscent of the liverwort Anthoceros. In other species, all the cells of the leaf contain several chloroplasts. There is much variation, also, in the occurrence of stomata, some species being amphistomatic and others hypostomatic.

Early stages in the development of the sporangia in Selaginella are very similar to those in Lycopodium, and there is a similar range of variation in the location of the primordium. Thus, in some species, it arises on the axis, while in others it arises on the adaxial surface of the leaf, between the ligule and the axis. However, at maturity the sporangium comes to lie in the axil of the sporophyll. The first division is periclinal and gives rise to outer jacket initials and inner archesporial cells. The jacket initials divide further to produce a two-layered sporangium wall and the archesporial cells produce a mass of potentially sporogenous tissue, surrounded by a tapetum. In microsporangia many cells of the sporogenous tissue undergo meiosis to form tetrads of microspores but, in the megasporangia of most species, all the sporogenous tissue disintegrates, except for one spore mother cell, from which four megaspores are formed. Some species, however, retain more than the one functional megaspore mother cell, so that up to twelve or even more megaspores may result. Yet other species are peculiar in that, out of the single tetrad of megaspores, one, two or three may be abortive, so that in the extreme condition the megasporangium may contain only one functional megaspore. S. rupestris usually has two megaspores in each sporangium and sometimes only one, while S. sulcata regularly has only one. S. rupestris is further remarkable in that the megaspores are not shed, but are retained within the dehisced megasporangium and fertilization takes place while it is still in situ. Thus, it happens that young sporophytes may be seen growing from the cone of the parent sporophyte. Few seed-plants have achieved this degree of vivipary, yet in Selaginella it occurs in a species belonging to the allegedly primitive Homoeophyllum group.

In those species whose cones contain both megasporophylls and microsporophylls, it is usual for the former to be near the base of the cone and the latter near the apex. This further emphasizes the point, already made, that the arrangement in lycopods is the inverse of that observed both in the gymnospermous Bennettitales and in hermaphrodite flowers of angiosperms.

Fig. 14D illustrates the appearance of a megasporophyll in longitudinal section, with the ligule (1) and the differential thickening in the sporangium wall, while Figs. 14B and 14C illustrate a dehiscing microsporangium and megasporangium respectively. Contractions of the thick-walled cells of the megasporangium cause the megaspores to be ejected for a distance of several centimetres, but dispersal of the microspores is mainly by wind currents.

Long before the spores are shed, nuclear divisions have started to take place, so that the prothallus is well advanced when dehiscence occurs. The stages in the formation of the male prothallus are very similar to those figured for Isoetes and, at the moment of liberation, the male prothallus commonly consists of thirteen cells (one small prothallial cell, eight jacket cells and four primary spermatogenous cells, of which the latter undergo further divisions to produce 128 or 256 biflagellate antherozoids — Fig. 14Y). Within the megaspore, a large vacuole appears, around which free nuclear divisions occur and then, subsequently, a cap of cellular tissue becomes organized beneath the tri-radiate scar. In some species, this cap is continuous with the rest of the prothallus, which later becomes cellular too, but in others a diaphragm of thickened cell walls is laid down, as illustrated in Figs. 14R and 14U (3). Rupturing of the megaspore allows the cap to become exposed and it frequently develops prominent lobes of tissue, covered with rhizoids, between which are numerous archegonia. It has been suggested that the rhizoids, as well as anchoring the megaspore, may serve to entangle microspores in close proximity to the archegonia.

The archegonia are similar to those of Isoetes, except that the neck is shorter and consists of two tiers of cells only.

The embryology of Selaginella is remarkable for the very great differences that occur between species. These are illustrated in Figs. 14M-X, all of which, for ease of comparison, are drawn as if the megaspore were lying on its side. The first cross wall is in a plane at right angles to the axis of the archegonium (Fig. 14M) and the fate of the outermost half in the different species is indicated by oblique shading. In S. Martensii (Figs. 14M-P) the outer half gives rise to a suspensor (s), while the inner half gives rise to all the rest of the embryo, with a shoot apex, (x), a root (r) and a swollen foot (f). The axis of the embryo, in this species, becomes bent through one right angle so as to bring the shoot apex into a vertical position. S. selaginoides (Fig. 14Q) is similar, except for the absence of a foot. S. Poulteri (Figs. 14R-T) is a species with a well developed diaphragm, through which the embryo is pushed by the elongating suspensor. A curvature through three right angles then brings the shoot into a vertical position. S. Kraussiana (Figs. 14U and 14V) likewise has a diaphragm, but in this species the venter of the archegonium gradually extends through it (a), so carrying the embryo with it into the centre of the prothallus. The outerhalf of the dividing zygote provides, not only the vestigial suspensor, but also the foot. The archegonium of S. Galeottii behaves in a similar way, but the embryo is different (Fig. 14W) in that the outer half provides the suspensor, the foot and also the root. S. denticulata (Fig. 14X) has the various parts of the embryo disposed as in S. Martensii (i.e. the root lies between the suspensor and the foot) but they are derived in a completely different way, for they all come from the outermost half of the dividing zygote.

Such extraordinary variations as these are very puzzling and have occupied the thoughts of many morphologists. Some have held that the presence of a well developed suspensor is a primitive character and that the reduction of this organ in some species is a sign of relative advancement. Its reduction seems to be correlated with the transference of its function to the venter of the archegonium, and this would certainly seem to be an advanced condition. As to the ‘foot’, all that can be said is that it has little reality as a separate organ, since it can apparently be formed from various regions of the embryo and may even be dispensed with altogether.

Selaginella is peculiar among pteridophytes for its low chromosome numbers, n = 9 being the commonest number, and its chromosomes are minute.

Selaginellites is the genus to which are assigned all fossil remains of herbaceous lycopods that are known to have been heterosporous. The recent examination of Selaginellites crassicinctus is of particular interest, for within its cones were found the megaspores Triletes triangularis, which have long been known as one of the commonest spores in coal measure deposits, but whose origin was hitherto unknown. This discovery suggests that Selaginellites was probably an important component of the flora of those times, contemporaneous with the tree-like Lepidodendrales. Whereas this species was similar to most Selaginella species in having four megaspores in each sporangium, others had sixteen or even thirty-two, which suggests that they had not progressed so far in the direction of heterospory. While there is no general agreement among botanists as to how the various groups of the Lycopsida are related to each other, it is generally supposed that the heterosporous forms must have evolved from some homosporous ancestor.

In this connection, it is perhaps significant that, among Selaginella species, the type regarded as the most primitive (S. selaginoides) approaches most nearly to the Lycopodium species which is regarded as the most primitive (L. selago). Both are erect and isophyllous, with spirally arranged leaves showing the least difference between fertile and sterile regions and both having simple protostelic vascular systems. The similarities extend even to the young embryo, as a comparison of Figs. 10P and 14Q will show. The lack of a well developed foot in each is interesting, and makes one wonder whether it might have been absent from their ancestors also.

The most important differences between these two plants, therefore, seem to be the heterospory of Selaginella and its possession of a ligule. If it be accepted that heterospory is derived from homospory, there remains only the ligule to be explained. This is, indeed, difficult. There is no obvious reason why, in lycopods, this structure should invariably be associated with heterospory. Selaginella is usually grouped with Isoetes and the Lepidodendrales on the basis of the possession of these two characters, yet on other grounds Selaginella stands apart from Isoetes. The multiflagellate antherozoids of the latter suggest very fundamental differences. On the basis of the number of flagella, Lycopodium and Selaginella should be grouped together. Unfortunately, of course, we have no knowledge of the antherozoids of the Lepidodendrales, but one’s guess would be that they were multiflagellate, like those of Isoetes and Stylites. One thing is fairly certain — that Selaginella is not a direct descendant of the Lepidodendrales. Apart from this, one’s views on the relationships of the Lycopsida must depend upon a decision as to whether the ligule is more significant phylogenetically than the number of flagella.