Pteropsida - Part I

Sporophyte with roots, stems and spirally arranged leaves (megaphylls) often markedly compound and described as ‘fronds’ (although some early members showed little distinction between stem and frond). Protostelic, solenostelic or dictyostelic, sometimes polycyclic (rarely polystelic). Some with limited secondary thickening. Sporangia thick- or thin-walled, homosporous or heterosporous, borne terminally on an axis or on the frond, where they may be marginal or superficial on the abaxial surface. Antherozoids multiflagellate.

Some botanists widen the definition of the Pteropsida to include, not only the megaphyllous pteridophytes, but also the gymnosperms and angiosperms, on the supposition that all three groups are related. While this may well be so, it seems preferable to retain the distinction between pteridophytes and seed-plants and to restrict the definition of the Pteropsida so as to exclude all but the ferns. Even so, the group is an enormous one, with over 9,000 species, and shows such a wide range of form and structure that it is almost impossible to name one character which is diagnostic of the group. The reader will have noticed that almost all of the characters listed at the head of this chapter are qualified in some way.

It will readily be appreciated that, in such a large group, the correct status of the various subdivisions is very largely a matter of personal preference. Accordingly, there are almost as many different ways of classifying the group as there are textbooks dealing with it, and this is particularly true of the fossil members of the group.

At this point, only the major subdivisions are presented, the details being deferred until each subgroup is dealt with.

List 8


This is a remarkable group of plants that first appeared in the Middle Devonian and survived until the end of the Palaeozoic. As the name suggests, they were probably the ancestors of modern ferns. They may be classified as follows:

List 9

The Cladoxylales are a particularly interesting group, whose correct phylogeny has long been a matter of controversy. On the one hand, they show a number of features in common with the Psilophytales and, indeed, Pseudosporochnus has only recently been transferred from that group. On the other hand, they show features in common with the Coenopteridales, whose later representatives had already begun to look fern-like before they became extinct. The group thus stands in an intermediate position which strongly suggests a genuine phylogenetic connection between the two groups.

Several species of Cladoxylon are known, of which the earliest is C. scoparium, and our knowledge of this is based on one specimen about 20 cm long from Middle Devonian rocks of Germany. According to the reconstruction of the plant by Kraiisel and Weyland (Fig. 17A), there was a main stem, about 1-5 cm in diameter, which branched rather irregularly. Some of the branches bore fan-shaped leaves (Fig. 17B) ranging in size from 5 mm to 18 mm long. Some leaves were much more deeply divided than others, but all showed a series of dichotomies. On some of the branches, the leaves were replaced by fertile appendages which were also fan-shaped, each segment terminating in a single sporangium (Fig. 17C).

The vascular system was highly complex and was polystelic; each of the separate steles was deeply flanged; both scalariform and pitted tracheids were present in the xylem. Such complex vascular structure is characteristic of all the species of Cladoxylon and some had the additional complication of secondary thickening. C. radiatum was similar to C. scoparium in that all the xylem was primary, and Fig. 17D illustrates the way in which several xylem flanges were involved in the origin of a branch trace system. It also illustrates the ‘islands of parenchyma’, as seen in transverse section, which are a common feature of the Zygopteridaceae too. Another feature, shared with the Coenopteridales, is the presence of ‘aphlebiae’ at the base of the lateral branch (or petiole?). These were similar in position to the stipules of many flowering plants and received separate vascular bundles (1).


Fig. 17E illustrates another type of stem structure, found in Cladoxylon taeniatum and several other species, in which each of the xylem strands has an outer region of radially arranged tracheids which are thought to have been formed from a cambium. The arrows in the figure indicate that three of the stem steles were involved in the origin of branch traces. Successive branches, petioles and pinnae of descending order had progressively simpler vascular structures, without secondary wood, and are described under separate form-generic names. Thus, Fig. 17F shows the Hierogramma type of stelar arrangement, in which there were six xylem regions, each with islands of parenchyma. Lateral branches from this had four xylem areas and are known as Syncardia (Fig. 17G). Clepsydropsis (Fig. 17H) was probably the next type of branch, or petiole, although there has been some disagreement among palaeobotanists about this. Its stele, as seen in transverse section, had the shape of an hour-glass (hence the generic name) and from it lateral pinna traces were given off alternately, along with a pair of aphlebia traces (1). It should be noted that similar clepsydroid steles are known from a number of plants belonging to the Coenopteridales.

Pseudosporochnus is represented in Middle Devonian rocks of Germany, Scotland, Scandinavia and North America, but our knowledge of its morphology is based chiefly on the German species P. Krejcii (Fig. 171). It had an erect stem with a swollen base and a bushy crown of branches which forked dichotomously and terminated in small sporangia. According to Kraiisel and Weyland there were no organs that could be called leaves and, for this reason, the plant was placed in the Psilophytales. However, Leclercq has recently discovered abundant remains of the plant and investigations by her and Banks show that there are, in fact, spirally arranged leaves with ‘several successive pairs of opposite divisions; each of the numerous divisions then divides by several successive dichotomies; the whole appendage is arranged in three dimensions’. Some of the leaves were fertile and their ultimate divisions terminated in a pair of sporangia. These workers were, furthermore, able to observe the vascular system of the plant and found that it had a very complex stellate form, as seen in transverse section. Their conclusion, published in a preliminary note, is that the genus has more in common with the Cladoxylaceae than with any other group of plants and that it should certainly be removed from the Psilophytales.


This early group of ferns is a large one, consisting of many genera and species. It showed a wide range of growth habit, for some had creeping stems, others had erect trunks and yet others were epiphytes. As with members of the Cladoxylales, here too there is the problem of distinguishing leaf from stem and the term ‘Phyllophore’ is sometimes used for intermediate orders of branching.

Austroclepsis, occurring in Lower Carboniferous rocks of Australia, was first described as a species of Clepsydropsis, on account of its clepsydroid petioles. However, the mode of growth of the plant and the internal anatomy of the stem show that it was not a member of the Cladoxylales. It had a stout trunk, at least 30 cm in diameter and 3 m high, that must have looked superficially like modern tree ferns, but it differed fundamentally from these in that, within the mass of roots constituting the main bulk of the trunk, there were several stems instead of just one. These branched within the trunk and gave off numerous petioles in a 2/5 phyllotactic sequence and these, too, continued to run up within the trunk. Each of the many stems had a single stele, usually pentarch, in which there was a central stellate region of mixed pith surrounded by a zone of tracheids. The petioles had a rather narrow clepsydroid stele with two islands of parenchyma bounded by ‘peripheral loops’ of xylem, and it was from these peripheral loops that pinna traces were given off from alternate sides at distant intervals, each associated with aphlebiae.

Metaclepsydropsis duplex, from Lower Carboniferous rocks of Pettycur, Scotland, had a creeping dichotomous stem, from which erect ‘fronds’ arose at intervals. Its stele was circular in cross section or (just before a dichotomy) oval (Fig. 17J), with an inner region of mixed pith and an outer zone of large tracheids. The only protoxylem present was that associated with the origin of a leaf trace, there being no cauline protoxylem at all. The leaf trace was at first oval in cross section (Fig. 17K) but soon became clepsydroid (Fig. 17L). Pinnae were borne in alternate pairs, along with aphlebiae (1). In giving rise to a pair of pinna traces, the peripheral loop (2) became detached and then split into two (3). A new peripheral loop then quickly reformed.

Diplolabis Roemeri occurs in the same rocks and was very similar to Metaclepsydropsis. Its stem anatomy differed, however, in that the inner region of xylem consisted entirely of tracheids. The tracheids of the outer region were arranged in radial rows, but are nevertheless believed by some to have been primary in origin. The petiolar trace had a very narrow ‘waist’, with the result that it appeared X-shaped in cross section.

Dineuron ellipticum, also from Pettycur, on the other hand had no ‘waist’ at all in the petiolar trace, which was elliptical in cross section.

In all three of these Lower Carboniferous genera, the origin of pinna traces was the same, suggesting that pairs of lateral pinnae were arranged alternately along the petiole, or phyllophore. The frond was thus a highly compound one whose components formed a three-dimensional structure. However, it had not been realized just how complex they were until the important discovery, in 1951, of a mummified specimen of Rhacophyton zygopteroides. That this plant belonged to the Zygopteridaceae was established by examining its internal anatomy. It had a fairly stout stem bearing roots and spirally arranged fronds. The lowermost fronds were sterile and were bipinnate, the pinnules consisting of dichotomous branchlets, apparently without any flattening to form a lamina. The fertile fronds were much larger and more complex, and had pairs of pinnae arranged alternately, just as had been deduced for Metaclepsydropsis from a study of petrified material. Each of the paired pinnae was similar in its branching to a sterile frond. Whereas the lower pinna pairs had branched aphlebiae, these were replaced in the higher pinna pairs by profusely branched structures bearing numerous terminal sporangia. These were about 2 mm long and were without any specially thickened annulus.

At least eight species are known of the Upper Carboniferous genus Ankyropteris, which derives its name from the fact that the petiolar trace in some species, e.g. A. westfaliensis, was shaped like a double anchor (Fig. 17O). In some other species the petiolar trace was much less extreme, having less ‘waist’, but all were alike in that the islands of parenchyma were much extended tangentially and in that the peripheral loop remained closed throughout the origin of a pinna trace. Another peculiar feature was that the pinna trace was undivided, suggesting that the pinnae were single, instead of paired as in most other members of the group. A. Grayij from British coal measures, had a stem of considerable length which was over 2 cm in diameter. It was probably a climbing plant. The petioles were borne in a 2/5 phyllotactic spiral, corresponding with the five rays of the stellate stele (Fig. 17N). As in other members of the group, there were two distinct regions in the stele, but the inner region showed a clear distinction between a zone of tracheids and a central pith, while the outer region showed no evidence of radial arrangement at all and was clearly primary in origin.

The petiolar traces of Etapteris were peculiar in that the ‘peripheral loops’ remained open throughout (i.e. there were no loops at all). Two pinna traces became detached, fused and then separated again, before passing out into the paired pinnae (Fig. 17P). The Permian species, E. Lacattei, is interesting in having progressed further than other members of the group in the evolution of a photo synthetic lamina, for the ultimate pinnules were flattened (Fig. 17Q). In the fertile regions of the frond (Fig. i7R)the pinnules were replaced by groups of sporangia. These were club-shaped, slightly curved, and had a distinct broad annulus of thickened cells (Fig. 17S). Some Etapteris fronds were attached to trailing stems, while others belonged to tree-ferns with stout trunks. The nomenclature of the latter is, however, rather troublesome. The names Zygopteris and Botrychioxylon which have been used are probably synonymous. Z. primaria had a trunk about 20 cm in diameter, most of which consisted of a tangle of rootlets and leaf bases. In the centre was a single stem 1*5 cm across, with a five-rayed stele showing the usual two regions, but in this case the outer region looks very much as if it had been formed from a cambium, and many morphologists describe it as secondary wood. Botrychioxylon paradoxum had a very similar appearance, but in this stem the cells of the inner cortex were also regularly arranged in radial rows. It would seem, therefore, that the whole of the growing point of the stem must have been organized in a peculiarly regular manner and that great caution should be used in describing even the outer xylem as secondary.

Some species of Tubicaulis were tree ferns, while others were epiphytes. They are characterized by a frond form which approached closely to that of a present-day fern, for the fronds coming off in spiral sequence from the stem were pinnate and the pinnae were arranged in one plane.

Stauropteris burntislandica

Stauropteris is represented by two species, S. burntislandica from the Lower Carboniferous and S. oldhamia from the Upper Carboniferous. Although the method of branching of the frond was similar to that of many of the Zygopteridales, differences in the vascular system are sufficient to warrant the creation of a separate family. The most important of these is the absence of islands of parenchyma in the xylem of the petiolar traces. It is believed that, so far, only portions of fronds have been found and that the stems have yet to be discovered. Fig. 18A shows how the frond of S. bumtislandica was constructed, pairs of pinnae arising alternately along the petiole, each associated with aphlebiae. Then each pinna gave rise to secondary pinnae in the same way and this pattern was repeated at all levels of branching within the frond. The vascular system of the petiole of S. oldhamia (Fig. 18B) consisted of four regions of xylem either contiguous or separate from each other, each with a mesarch protoxylem. The smaller branches, however, tended to have a single tetrarch strand.

Perhaps the most interesting feature of all about Stauropteris bumtislandica is the fact that it was heterosporous. Its megasporangia (Fig. 18C), when found isolated, are called Bensonites fusiformis. They were strangely fleshy at the base and most commonly contained two functional megaspores along with two very small and, presumably, abortive ones, although examples have been found with four, six or eight megaspores. It is believed that the whole structure was shed from the parent plant without prior dehiscence. The microsporangia of S. oldhamia (Fig. 18D) were spherical, were typically eusporangiate in having a thick wall and had a terminal stomium where dehiscence took place, but there was no annulus of thick-walled cells.

In the past, the Botryopteridaceae were often described as much simpler in their organization than the rest of the Coenopteridales, but recent investigations on both sides of the Atlantic have demonstrated that this is far from the case. Botryopteris antiqua, from the Lower Carboniferous of Scotland, is the earliest known species and was also the simplest in its internal anatomy. It had trailing dorsiventral axes up to 2 mm in diameter, which gave rise to erect, or semi-erect, radial stems bearing petioles, in spiral succession, and roots. The petioles then underwent branching, of up to five successive orders, to produce a multipinnate branch system. There was no flattening of the pinnules to form a lamina anywhere in the frond of this species and the distinction between stem and petiole is purely arbitrary. The three types of stele are illustrated in Fig. 18E, where ‘2’ indicates the one belonging to the trailing dorsiventral axis. It was a solid rod of tracheids with multiseriate pits or with scalariform or reticulate pits. The single protoxylem group was lateral and almost, but not quite, exarch. The radial stems were about the same diameter, but the stele was circular in cross section with the smallest tracheids (protoxylem ?) in the centre (3). The petioles were somewhat smaller, up to 1*4 mm in diameter, and had an oval stele with a lateral protoxylem (4). As the branches of the frond divided and subdivided, the stele became smaller and smaller until the ultimate pinnules had only a few tracheids or even only one. The sporangia were globose, up to 0-25 mm across, and had a multicellular annulus that occupied almost half the surface area.

A comparison of this early species with those of the Upper Carboniferous and the Permian shows that there was a trend in the evolution of the petiole trace towards a greater degree of dorsiventrality, together with an increase in the number of protoxylems. Thus, Botryopteris ramosa (Upper Carboniferous) had a shallow gutter-shaped stele with three protoxylems (Fig. 18F), whereas B. forensis (Permian) had a stele shaped like the Greek letter in transverse section, with up to fifteen protoxylems (Fig. 18G). Some of these later species, furthermore, are known to have had laminate pinnules (Fig. 181).

The complexity of the branching of the later species of Botryopteris is illustrated by the reconstruction of the stelar system of B. trisecta (Fig. 18H). Its erect stem had a cylindrical protostele and bore leaves in a spiral sequence. The petioles had an oval vascular strand and branched into three. The two lateral branches then trisected again but, whereas the median traces in each case were a> shaped, the lateral ones were cylindrical, like the stem stele. The whole frond was arranged in three dimensions, except for the ultimate pinnules which were disposed in one plane.

Associated with this plant were found some remarkable spherical masses, containing thousands of sporangia, which are believed to represent the fertile parts of the frond, although in the meantime they are described under a separate specific name, Botryopteris globosa. The whole mass was up to 5 cm across and had, running through it, a system of branches with co shaped steles, but how it was connected to the parent plant is not known. Each sporangium was pear-shaped (Fig. 18J) and the distal half consisted entirely of thick-walled cells, except for a stomium of thin-walled cells over the apex. In most species of Botryopteris, the sporangium wall is described as only one cell thick, suggesting that, in this respect at least, they were leptosporangiate. It is apparently true of some of the sporangia of B. globosa, but not of all, for some clearly had a second layer of thin-walled cells on the inside. This may well have shrivelled after the spores had been shed, so becoming invisible when petrified. Thus, although approaching the leptosporangiate condition, B. globosa had certainly not yet achieved it, and the same is probably true of all the species.

List 10


It was customary in the past to describe the Carboniferous as the Age of Ferns. This was because of the abundance of large fern-like fronds in the coal-measures, but it is now known that many of them really belonged to gymnosperms, for they have been found in association with seeds. Indeed, it is now suspected that most of them were gymnospermous. However, there can be no certainty about sterile fronds and these must, therefore, be placed in a number of form genera defined on the basis of the overall shape of the frond and on the shape and venation of the pinnules. Pecopteris is one of these and a large number of species are known. Some of them were certainly gymnosperms, but others were equally certainly ferns, for they bore sori of thick-walled sporangia. The frond, sometimes as much as 3 m long, was many times pinnate and the pinnules were attached along their entire base, each with a single midrib. The lateral veins were somewhat sparse and branched dichotomously once or twice (Figs. 19D and 19E) or remained unbranched.

Asterotheca is the name given to pecopterid fronds bearing sessile sori made up of four or five sporangia fused at the base into a synangium, but with the distal part free (Fig. 19F). The sori were commonly arranged in two series along the pinna, as illustrated in Fig. 19E, each associated with a veinlet in the lamina. Scolecopteris was similar, except that the sorus was elevated on a short pedicel, or receptacle (Fig. 19A). In Acitheca, the sporangia were elongated and pointed and were arranged round a central plug-like receptacle (Fig. 19C). Eoangiopteris is regarded as a more advanced type of sorus since it was linear instead of radial. Each had a cushion-like receptacle, on which were five to eight sporangia (Fig. 19B). In all these genera, the sporangium wall was very massive, many cells thick, and the number of spores liberated from each was very high, e.g. up to 2,000 in Asterotheca parallela.

Fronds of these various types are often found in association with stout trunks that bore a superficial resemblance to those of modern tree-ferns, some of them as much as 15 m high, but organic connection between fertile frond and trunk has not yet been demonstrated conclusively. The evidence, nevertheless, suggests that Asterotheca fronds were borne in a crown at the summit of trunks known as Psaronius. Many species, belonging to a number of subgenera of Psaronius, are known and most of them had remarkably complex stelar anatomy. Some of them were as much as 75 cm across, but most of this width was occupied by a thick mantle of roots, for the single stem in the centre was only a few cm in diameter. The stele of most species was a poly cyclic dictyo stele which, in the more complex types, contained as many as eleven interconnecting coaxial cylinders (or, rather, inverted cones fitting inside one another). Each was dissected into a number of mesarch meristeles completely surrounded by phloem, and the leaf traces at any particular level arose from the outermost system, while the inner systems were concerned with the origin of leaf traces at higher levels. The earliest examples, however, were simpler than this in their internal anatomy, e.g. P. Renaultii from the Lower Coal Measures had an endarch solenostele; and there is evidence that even the complex Permian species had a relatively simple structure near the base of the trunk, as would be expected by analogy with present-day ferns. Although the trunks were widest at the base, this was not because the stem within was wider but because there were a greater number of rootlets in the mantle; the stem was actually smaller towards the base. Some species had the leaves arranged distichously, some in three or four vertical rows, while others had them arranged spirally, as in most modern representatives of the group.

The Marattiales are represented at the present day by about 200 species, placed in six (or seven) genera, most of which are confined to the tropics. Angiopteris (100 species) is a genus of the Old World, extending from Polynesia to Madagascar, while Danaea (thirty-two species) is confined to the New World. Marattia (sixty species) is pan-tropical and extends as far south as New Zealand. Christensenia (= Kaulfussia) is monotypic and is confined to the Indo-Malayan region. Most species have massive erect axes, but they never attain the dimensions of the fossil Psaronius. The largest, although reaching a diameter of 1 m, seldom exceed this in height. Christensenia and some species of Danaea, however, have creeping horizontal axes. The fronds of some species are larger than in any other living ferns and may be as much as 6 m long, with petioles 6 cm in diameter. They may be as much as five times pinnately compound or, in some species, only once pinnate, like a Cycad leaf, while a few species have a simple broad lamina. Christensenia is peculiar in having a palmately compound frond, as the specific name, C. aesculifolia, implies. It is also peculiar in having reticulate venation, for all the other genera have open dichotomous venation. All show circinate vernation, i.e. the young frond is coiled like a crozier and gradually uncoils as it grows. This is a feature which they share with Leptosporangiate ferns but which is absent from the Ophioglossales. With the exception of Danaea trichomanoides, all the living members of the group have very leathery pinnules in whose ontogeny several rows of marginal initials are active (instead of a single row of marginal initial cells, as is more usual in leaves of other plants). In many species there are swellings, or pulvini, at the base of the pinnae and pinnules, which play a part in the geotrophic responses of the leaf, and in all species there are thick fleshy stipular flanges at the base of the petiole. Fig. 19V illustrates the appearance of the growing point of Christensenia, showing how the stipules (1) are joined by a commissure (2), and how they are folded over the primordium. After the frond has died and has been shed, the stipules and the leaf base remain attached to the axis and contribute much to its overall diameter.


The young parts of Marattia and Angiopteris are covered with short simple hairs, while those of Christensenia and Danaea bear peltate scales. Bower suggested that the nature of the dermal appendages in ferns can be a useful indicator of primitiveness or advancement, hairs being more primitive than scales; on this basis, therefore, Christensenia is relatively advanced and this conclusion is supported by its possessing reticulate venation. A comparison of fossil and recent members of the group suggests that there has been progressive reduction in height, from the tree-like Carboniferous forms, through an intermediate stumpy erect axis, to an oblique or horizontal creeping rhizome; and on this basis, too, Christensenia along with Danaea is to be regarded as relatively advanced.

The stem grows by means of a bulky type of meristem, not referable to a single initial cell and is characterized by the absence of sclerenchyma. Mucilage canals and tannin cells are abundant throughout and give the tissues a very sappy texture. The vascular anatomy of the stem is the most complex of all living pteridophytes and is surpassed in complexity only by fossil members of the group, such as Psaronius. A transverse section of the stem of Angiopteris (Fig. 19O) reveals a number of concentric rings of meristeles which, in a dissection (Fig. 19S), are seen to be part of a series of complex and irregular meshworks lying one within the other, yet interconnected by ‘reparatory strands’. The whole system may be described as a highly dissected polycyclic dictyostele, but can best be visualized as a series of inverted cones of lace stacked inside each other. Although each meristele in the sporeling is surrounded by an endodermis, in the adult state the endodermis is completely lacking.

The earliest protoxylem elements to lignify are ‘annular-reticulate’, i.e. adjacent rings of lignin are interconnected by a network of strands, whereas later ones are reticulate. The metaxylem elements are scalariform and, in Angiopteris, the orientation of the elongated bordered pits is sometimes longitudinal, instead of transverse. This peculiar arrangement has been called ‘ob-scalariform’ and occurs elsewhere in the Ophioglossaceae and a few leptosporangiate ferns (Dennstaedtia and Blechnum).

Each leaf, in a mature plant, receives a number of traces which arise from the outermost system of meristeles (the cut ends of the leaf traces are represented in black in Fig. 19S), but the root traces may arise from the innermost regions of the stele, threading their way through successive cones on their way to the cortex (cross-hatched in Fig. 19O). In those species with erect axes, the roots may emerge from the cortex some distance above the ground, so forming proproots. They are polyarch, with as many as nineteen exarch protoxylems and, while the aerial portions are medullated, as soon as the roots penetrate the soil the xylem extends right to the centre. Those of young plants usually contain a mycorrhizal fungus within the cortex (an oomycete known as Stigeosporium marattiacearun).

In all genera, the sori are borne in a ‘superficial’ manner, i. e. on the dorsal surface of the lamina, and beneath a vein or a veinlet. Christensenia has circular sori irregularly distributed between the main veins (Fig. 19P)but, in all other genera, the sorus is more or less elongated beneath a lateral vein (Figs. 19Q and 19R). In Angiopteris the sporangia are free from each other (Figs. 19G and 19H), but in Marattia, Danaea and Christensenia they are fused into a synangium (Figs. 19I-N). Danaea is peculiar in having fleshy flanges of tissue (3) projecting between the adjacent synangia (or, according to some, in having the synangia sunken into a very fleshy pinnule).

The first stage in the development of a sporangium is a periclinal division of a single epidermal cell, of which the inner half gives rise ultimately to the archesporial tissue, while the outer half gives rise to part of the sporangium wall, the rest of the wall being produced by the activity of adjacent cells. At maturity, the sporangium wall is many cells thick and there is a tapetum formed from the innermost wall cells. The occurrence of numerous stomata in the sporangium wall is an interesting feature rarely found elsewhere and presumably associated with its massive structure. Very large numbers of spores are produced from each sporangium (e.g. 1,440 in Angiopteris, 2,500 in Marattia and over 7,000 in Christensenia) and, since all the sporangia within a sorus mature and dehisce simultaneously, prodigious numbers of spores are shed.

In those species with free sporangia, e.g. Angiopteris, there is a crude kind of annulus of thickened cells, whose contractions pull the sides of the sporangium apart along a line of dehiscence on the inner face (Fig. 19H). Those with synangia have no such device; instead, a thin part of the sporangium wall dries and shrinks to form a pore through which the spores can fall (Figs. 19K-N). The whole sorus in Marattia is very woody and, when ripe, splits into two halves which are slowly pulled apart, so as to expose the pores in each sporangium (Fig. 19J).

Germination of the spores is rapid, occurring within a few days of being shed, and they develop directly into a massive dark green thalloid prothallus, which is mycorrhizal and is capable of living for several years. An old prothallus may be several centimetres long and may resemble closely a large thalloid liverwort (Fig. 19T). The prothallus is monoecious but, while the antheridia occur on both the upper and lower surface, the archegonia are confined to the lower surface, where they occur on the central cushion along with rhizoids. Both types of gametangia are sunken beneath the surface of the prothallus and the antheridium is large and massive. The archegonium (Fig. 19U) has a large ventral canal cell (except in Danaea) and a neck canal cell with two nuclei. The antherozoids are coiled and multiflagellate, as in other ferns.

The first division of the zygote is at right angles to the axis of the archegonium, and the embryo is endoscopic. Thus, since the archegonial neck is directed downwards, the embryo is orientated with its shoot uppermost and, as it grows upwards, it bursts its way through the tissues of the prothallus. A minute suspensor is present in Danaea (Fig. 19Y) and in some species of Angiopteris, but Marattia, Christensenia and most species of Angiopteris are completely without a suspensor. This lack of constancy is paralleled in the Ophioglossales and has led to speculation as to its phylogenetic implications. A suspensor is generally held to be a primitive character and its presence even if not universal in the Eusporangiatae, places them at a lower level of evolution than the remaining ferns, from which it is completely absent.

The epibasal hemisphere gives rise to the shoot apex (x) and the first leaf (l) (Fig. 19W), but there is no regular pattern of cell divisions and the hypobasal region gives rise to a poorly developed foot (f) and, somewhat later, to the first root (r) (Fig. 19X). Chromosome counts give a haploid number 11 = 40 in Angiopteris.


This group of plants, completely without any early fossil record, is represented by about eighty living species, belonging to three genera. Botrychium (thirty-five species) is cosmopolitan in distribution and Ophioglossum (forty-five species) is nearly so, but Helminthostachys (monotypic) is restricted to Indo-Malaysia and Polynesia. Two species are fairly common in the British Isles, Botrychium lunaria, ‘Moonwort’ (Fig. 20A) which grows in dry grassland and on rocky ledges, and Ophioglossum vulgatum, ‘Adder’s Tongue’ (Fig. 20G) in damp grassland, fens and dune-slacks, while a third species, O. lusitanicum, is restricted to grassy cliff tops in the Channel Islands and the Scilly Isles.

The stem, in most species, is very short and is erect, except in a few epiphytic species of Ophioglossum and in Helminthostachys, where it becomes a horizontal rhizome as the plant grows larger. Where the stem is erect, the leaves arise in a spiral sequence, but in temperate regions it is normal for only one leaf to be produced each year. In Helminthostachys, the leaves are borne in two ranks along the rhizome; they are large and ternately compound, but in the other two genera they are usually much smaller. Those of Botrychium are pinnately compound; those of Ophioglossum are simple or lobed and, unlike those of the other two genera, have a reticulate venation. At the base of the petiole there is a pair of thin stipules which enclose the apical bud; and the next leaf, when it begins to grow, has to break its way through the thin sheath covering it. Unlike all other living ferns their leaves are not circinately coiled when young.

In all three genera, the fertile fronds have two distinct parts, the fertile part being in the form of a spike which arises at the junction of the petiole with the sterile lamina, on its adaxial side. The fertile spike is pinnately compound in those genera with a compound lamina and simple in Ophioglossum, where the lamina is simple. Its morphological nature has been the subject of some considerable discussion in the past but is now generally thought to represent two basal pinnae which have become ontogenetically fused, face to face (i.e. it is believed that some early ancestor of the group had two fertile basal pinnae, whose primordia became fused during subsequent evolution). Today, the only evidence for the double nature of the spike lies in its vascular supply.

The roots are peculiar in being completely without root hairs, a feature which is possibly connected with their mycorrhizal habit.

Growth of the stem apex is from a single apical cell, and its products are characteristically soft and fleshy, for they are without sclerenchyma. The stem of the young sporeling is protostelic, but soon becomes medullated. Later on, the stem of Botrychium becomes solenoxylic, i.e. there are leaf gaps in the xylem, but not in the single external endodermis. Ultimately, the appearance of a sporadic internal endodermis may give rise to a rudimentary solenostele. Botrychium is the only genus of living ferns to show secondary cambial activity, and in some species it may give rise to a considerable thickness of secondary wood, composed of tracheids and wood-rays. Rhizomes of Helminthostachys pass through much the same stages of stelar organization, but the largest specimens go one stage further and achieve true solenostely, with an internal as well as an external endodermis. Ophioglossum varies considerably in its internal anatomy, according to species. Some possess an outer endodermis, but in most species it is absent, even in the young stages. The leaf gaps in the xylem overlap one another, giving rise to a network of meristeles, which form a rudimentary kind of dictyostele.

The xylem is endarch in Botrychium and Ophioglossum, but mesarch in Helminthostachys. The earliest formed protoxylem tracheids are very similar to those of the Marattiales; later ones are reticulate (some being obreticulate) but scalariform tracheids are absent. A pronounced feature of all three genera is the distinctly bordered circular pits in the metaxylem tracheids, but early accounts of the universal presence of a torus in the pit closing membrane appear to be incorrect. Bierhorst records them only in Botrychium dissectum and states that even in this species they are not a constant feature.

The sporangia in all three genera are ‘marginal’ in origin. In Botrychium, they are borne in two rows along the ultimate pinnules of the fertile spike (Fig. 20B) and each receives its own separate vascular supply from a vein running into the pinnule (Fig. 20C). In Helminthostachys, the axis of the fertile spike bears numerous ‘sporangiophores’ in several rows, each bearing several sporangia and a few tiny green lobes at the tip. The spike of Ophioglossum bears two rows of sporangia fused together, beyond which the axis projects as a sterile process (Fig. 20G). A number of vascular bundles run longitudinally up the middle, anastomosing occasionally and giving off lateral branches to the sporangia (Fig. 20I).

Early stages of development of the sporangium are similar to those in the Marattiales; a single initial cell undergoes a periclinal division, the inner half giving rise ultimately to the archesporial tissue, while the outer half goes to form part of the sporangium wall. Adjacent cells contribute further to the wall, which is very massive and several cells thick at maturity. A tapetum of several layers of cells is formed from the inner regions of the sporangium wall, which break down to form a continuous plasmodium in which the spores develop. As in Marattiales, there are stomata in the sporangium wall.

Dehiscence of the sporangium is transverse in Botrychium and Ophioglossum (Figs. 20B and 20H), but longitudinal in Helminthostachys, and large numbers of spores are released (more than 2,000 in Botrychium and as many as 15,000 in Ophioglossum).


The prothallus in all three genera is mycorrhizal. Indeed, the presence of the appropriate fungus is essential for the growth of the prothallus beyond the first few cell divisions. In most cases the prothallus is deeply buried in the soil and lacks chlorophyll, but cases have been reported of superficial prothalli, in which some chlorophyll was present. Some have abundant rhizoids, but others are completely without them.

The prothallus of Botrychium virginianum (Fig. 20D) is a flattened tuberous body, up to 2 cm long. Antheridia appear first and are deeply sunken. Large numbers of antherozoids are liberated from each and escape by the rupturing of a single opercular cell. The archegonium has a projecting neck several cells long, a neck canal cell with two nuclei, and a ventral canal cell (Fig. 20E).

The prothallus of Ophioglossum vulgatum differs in being cylindrical, and may be as much as 6 cm long (Fig. 20J). Frequently, there is an enlarged bulbous base, in which the bulk of the mycorrhizal fungus is located. (In both Figs. 20D and 20J, the extent of the fungus is indicated by a broken line.) As in Botrychium, the antheridia are sunken and produce very large numbers of antherozoids. Unlike Botrychium, however, its archegonia are sunken too. In Fig. 20K, the archegonium is illustrated at a stage just before maturity, when there are visible two nuclei in the neck canal cell, but just before the basal cell has divided. Indeed, a ventral canal cell has rarely been seen, presumably because it disintegrates almost as soon as it is formed.

As in the Marattiales, the first division of the zygote is in a plane at right angles to the archegonial axis. In Helminthostachys, the outer (epibasal) hemisphere undergoes a second division, so as to produce a suspensor of two cells, while the hypobasal hemisphere gives rise to a foot, a root and, later, the stem apex. The embryo is thus endoscopic, but during its further development its axis becomes bent round through two right angles, so as to allow the stem to grow vertically upwards. The embryo of some species of Botrychium is likewise endoscopic and has a small suspensor (Fig. 20F), but in others including B. lunaria, there is no suspensor and the embryo is exoscopic; and this is true of all species of Ophioglossum (Fig. 20L). In all cases there is considerable delay in the formation of the stem apex, and in some species it may be several years before the first leaf appears above the ground, by which time many roots may have been formed. These long delays suggest that the mycorrhizal association is an important factor in relation to the nutrition, not only of the prothallus, but also of the young sporophyte.

Chromosome counts show a surprising range within the group, for Botrychium has a haploid number n = 45, Helminthostachys n = 46 or 47, while in Ophioglossum vulgatum n = 250 - 260 and in Ophioglossum reticulatum n = 631 + 10 fragments.

Despite these divergent chromosome numbers, there can be little doubt that the three genera of the Ophioglossales are fairly closely related, nor that they represent an ancient and primitive group of ferns despite the lack of fossil representatives. The reticulate venation of Ophioglossum, its consolidated fertile spike and its complete lack of a suspensor together suggest that it has reached a more advanced stage of evolution than either of the other two genera. As in the Marattiales, it seems that the upright stem is the basic condition, since even in Helminthostachys the young plant has an erect axis.

Regarding the relationships between the Ophioglossales and the Marattiales, it is not easy to decide which characters are significant. Of the many characters common to the two groups, most indicate merely that they have reached roughly the same stage of evolution, rather than that they are closely related. These may be briefly listed as 1. basically erect axis, 2. stipules at the base of the petiole, 3. absence of sclerenchyma, 4. sporadic endodermis, 5. massive sporangium wall, with stomata, the sporangia showing a tendency to fusion, 6. large spore output, 7. prothallus long lived, 8. massive antheridium, 9. suspensor present in some, absent in others. Characters which suggest that the two groups are only distantly related are the circinate vernation of the Marattiales and their superficial sori, contrasting with the absence of circinate vernation from the Ophioglossales and their marginal sporangia.

List 11

The modern representatives of the Osmundales occupy an isolated position among the ferns, intermediate in many respects between the Eusporangiatae and the Leptosporangiatae but not necessarily, therefore, linking the two groups phylogenetically, for they are an extremely ancient group with an almost complete fossil history extending as far back as the Permian. Those that have survived to the present day can truly be described as ‘living fossils’.

All have erect axes, bearing a crown of leaves; and the same is true of the fossil members, some of which had trunks 1 m or more in height. Among the earliest representatives, in the Permian, were several species of Zalesskya. These had a solid protostele in which there were two distinct regions of xylem (an inner region of short tracheids and an outer one of elongated tracheids forming an unbroken ring). The same was true of Thamnopteris Schlechtendalii, but T. Kidstonii had a slightly more advanced stelar anatomy, in that the central region was occupied by a mixed pith of tracheids and parenchyma. Osmundites Dunlopii from the Jurassic was similar to T. Kidstonii, but the contemporaneous O. Gibbeana showed some dissection of the xylem ring into about twenty separate strands. Nevertheless, the stele was still strictly a protostele, since there was a continuous zone of phloem (and, presumably, endodermis) round the outside. The term ‘dictyoxylic stele’ can conveniently be used to describe this arrangement. Poor preservation does not allow any statement to be made about the central pith regions of these two forms, but in the Lower Cretaceous O. Kolbei there was definitely a mixed pith. The Cretaceous species O. skidegatensis had a pith of pure parenchyma and showed a further advance in having some internal phloem, while O. Carnieri was the most advanced of all, in being truly dictyostelic. This is most interesting, for it is a condition not achieved by any modern representatives of the group. Most of these are no further advanced in stelar anatomy than the Jurassic Osmundites Gibbeana.

Of the living genera, Osmunda (fourteen species) is widespread in both hemispheres, Leptopteris (six species) is confined to Australasia and the South Sea Islands, while Todea is represented by the single species T. barbara, found in S. Africa and Australasia. (Some taxonomists include Leptopteris in the genus Todea.) Only one species, Osmunda regalis — the ‘Royal fern’ — is represented in the British flora. Its stems are massive and branch dichotomously to form large hummocks. Todea barbara may have a free-standing trunk I m or more high, and so also may Leptopteris hymenophylloides, while one species of Leptopteris from New Caledonia attains a height of 3 m.

A transverse section of the stem of a mature Todea (Fig. 21K) exhibits a typical dictyoxylic condition. The central medulla is surrounded by separate blocks of xylem, outside which there is phloem and a continuous endodermis. Occasionally, some internal phloem occurs, but no internal endodermis. Most species of Osmunda are similar, but O. cinnamomea sometimes has an internal, as well as an external, endodermis (Fig. 21B). The types of xylem element present are similar, in some respects, to those of the Marattiaceae, and the position of the protoxylem ranges from endarch in Todea to nearly exarch in Osmunda.

The leaves, in most species, are leathery in texture, but those of Leptopteris hymenophylloides are comparable with those of the Hymenophyllaceae (‘filmy ferns’) and have a thin pellucid lamina, only two or three cells thick, from which stomata are completely lacking. During their development the leaves of all species exhibit circinate vernation and are covered with hairs. The base of the petiole is broad and winged in a manner reminiscent of the Eusporangiatae and, after the frond has been shed, the leaf base is persistent, adding considerably to the diameter and the mechanical strength of the stem.


The fronds of Osmunda regalis are twice pinnate, those produced first in each season being sterile. These are followed by partially fertile fronds (Fig. 21 A), while the last to be produced are often completely fertile. The fertile pinnules are very reduced tassel-like structures, representing just the midrib. In the absence of a lamina, the sporangia cannot be ‘superficial’ and are usually described as ‘marginal’. In partially fertile fronds of O. regalis, the fertile pinnules occupy the distal regions, but in those of O. Claytoniana they occupy the middle regions. Todea barbara has once-pinnate fronds in which the fertile pinnules show scarcely any modification and the sporangia are superficial, being densely scattered over the under-surface of the lamina. They occupy the basal regions of partially fertile fronds (Fig. 21 FI). The fronds of Leptopteris hymenophylloides are large and many times pinnate, with the sporangia scattered sparsely along the veinlets of unmodified pinnules (Fig. 21F). In no case is there any tendency for the sporangia to become aggregated into sori, nor is there any sign of an indusium.

The sporangium is not strictly leptosporangiate, for several cells play a part in its initiation and, at maturity, it is relatively large and massive with a stout short stalk. There is some variation in the shape of the archesporial cell, as illustrated in Figs. 21I and 21J, for it may be tetrahedral, as in leptosporangiate ferns, or it may be cubical, as in the Eusporangiatae. The tapetum is formed from the outermost layers of the sporogenous tissue, unlike that of the Eusporangiatae, and there is also a layer of tabular cells, formed from the same regions, which becomes appressed to the inner side of the sporangium wall. For this reason, at maturity, the wall appears to be two cells thick. There is a primitive kind of annulus, formed by a group of thick-walled cells, on one side of the sporangium and a thin-walled stomium, along which dehiscence occurs, extends from it over the apex of the sporangium (Fig. 21G). Relatively large numbers of spores are released from each sporangium (e.g. about 128 in Leptopteris and more than 256 in Osmunda and Todea). The spores contain chlorophyll and must germinate rapidly if they are to do so at all.

The prothallus (Fig. 2TC) is large, fleshy and dark green, resembling a thalloid liverwort, up to 4 cm long. The antheridia (Fig. 21D) project from the surface, as in Leptosporangiatae, but are larger, have more wall cells and produce a greater number of antherozoids than do most of them. The archegonia (Fig. 21E) are borne along the sides of the midrib; they have projecting necks and differ from those of leptosporangiate ferns only in the number of neck cells (six tiers, instead of the usual four).

The embryology of the young sporophyte, too, shows some features which distinguish the Osmundales from the Leptosporangiatae. Not only is the first division of the zygote vertical, but so also is the second. It is the third division which is at right angles to the axis of the archegonium, instead of the second. Subsequent divisions are somewhat irregular and the embryo remains spherical for a relatively long time. Ultimately, however, a shoot apex, cotyledon, root and a large foot appear, but there is some irregularity in their derivation from the initial octants.

Despite the marginal position of the sporangia in Osmunda, as compared with their superficial position in the other two genera, the three genera are so similar in other respects that they are, without doubt, closely related, and this conclusion is supported by chromosome counts. The haploid number is n = 22 throughout.