Pteropsida - Part II


In the past, the name Filicales was applied in the broadest possible sense, so as to include all the ferns but, recently, its use has been restricted, and it is applied just to the homosporous leptosporangiate ferns (as in Engler’s Syllabus der Pflanzenfamilien). However, even when thus restricted, it is still by far the largest group of the pteridophytes, for it contains almost 300 genera and about 9,000 species. Details of their form and anatomy would occupy many volumes and can only briefly be summarized here, the following families, subfamilies and genera having been selected to illustrate the salient points (the classification is based on that of Holttum).

List 12

As might be expected in such a large group, there is a considerable range of form and growth habit, from tiny annuals to tall tree-ferns and from protostelic forms to those with highly dissected poly cyclic dictyosteles, yet all are alike in the early stages of development of the sporangium. This, together with its stalk, arises from a single cell. The first division of the initial cell (Fig. 22O) is into an apical cell (1) and a basal cell (2). Further divisions take place in each (Fig. 22P) and give rise to a primary sporogenous cell (shaded in Fig. 22Q) and a jacket cell (3). The former gives rise to a two-layered tapetum and to a number of spore mother cells, surrounded by a sporangium wall one cell thick. Further details of sporangium development differ according to species, for some have a long slender stalk, only one cell thick, while others have a short and relatively thick stalk; the majority have a vertical row of thick-walled cells, constituting the annulus, while some have an oblique row and others merely a group of thick-walled cells; some have a high spore output, while in most species it is thirty-two or sixty-four.

Most commonly the prothallus is either cordate or butterfly-shaped ranging in size from a few mm to I cm or more across. There is a midrib several cells thick, but the wings of the prothallus are only one cell thick. It is surface-living, green and photosynthetic, and there are rhizoids on the underside, among which antheridia and archegonia are borne; the archegonia are usually concentrated near the growing point, or ‘apical notch’. Departures from this typical form occur in certain families, e.g. some have filamentous prothalli, resembling an algal filament, while even subterranean prothalli are known, but this habit is extremely rare.

Stages in the development of the archegonium are illustrated in Figs. 22K-N, the only variations being in the number of tiers of neck cells at maturity. The structure of the antheridium is also fairly constant throughout the Filicales. Figs. 22B-H represent the various stages in the development of the commonest type. The way in which successive cross walls bulge upwards or downwards is peculiar and is responsible for the formation of the characteristic ring-shaped cells of which the mature antheridium wall is constructed. At maturity, the cap cell is pushed off (Fig. 22I) to release the antherozoids (usually thirty-two in number) (Fig. 22J). Some families have a slightly more massive antheridium, composed of a greater number of wall cells and containing more antherozoids; these are believed to be more primitive than the rest.

Development of gametangia and sporangia

The embryology of the leptosporangiate ferns is likewise very constant throughout. The first cross-wall is almost invariably longitudinal and the second transverse. Thus, the zygote is divided at a very early stage into four quadrants, two directed towards the apical notch of the gametophyte (called the inner and outer anterior quadrants) and two away from the notch (called the inner and outer posterior quadrants). The outer anterior quadrant ultimately gives rise to the first leaf, the inner anterior to the shoot apex, the outer posterior to the first root, and the inner posterior to the foot. This, at least, is the procedure described in classical studies, but more recently it has been stated that the fate of the four quadrants is not always so clearly defined.

Statements that certain characters are primitive and others advanced can be made with more certainty for the Filicales than for any other group in the plant kingdom, because of the large number of fossil representatives that are known. Some of the families had already become widespread by the Mesozoic, while others appeared as long ago as the Carboniferous. A comparison of these with the rest of the living Filicales makes it possible to draw up an extensive list of primitive characters for the group as a whole. The following list is based on that of Bower (as modified by Holttum) with additions by Stokey.

Rhizome — slender, creeping, dichotomous, with fronds in two ranks on its upper side, protostelic, covered with hairs.

Fronds — large, amply branched, dichotomous and of unlimited growth, the stipe (petiole) receiving a single leaf trace, the ultimate pinnules narrow and with a single vein; venation without anatomoses (i.e. ‘open’).

Sori — containing few sporangia, terminating a vein.

Sporangia — relatively large, with stout stalk, without a specialized annulus, developing and dehiscing simultaneously to liberate a large number of spores.

Spore germination — giving a plate rather than a filament of cells.

Gametophyte — relatively large, thalloid, with a thick midrib, slow to develop.

Antheridium — large, containing several hundred antherozoids; wall cells more than four in number.

Archegonium — with a relatively long neck.

In the more advanced ferns, the dermal appendages are usually scales instead of hairs and, as the stem assumes an erect position, the leaves tend to form a crown at the apex. With increasing size, the stelar anatomy becomes more complex, the leaf-gaps overlap, and a dictyostele results. True vessels are known to occur in at least two genera. The fronds become reduced in size and may have a simple broad lamina with an entire margin and with anastomosing veins, while the stipe receives a number of leaf traces. In the most advanced ferns, the fronds are frequently ‘jointed’ at the base, i.e. they are shed by means of an absciss layer, a habit which may well be associated with life outside the tropics, in regions where seasonal changes in climate may be severe. Evolution of the sorus appears to have taken place in stages, the first of which involved a regular gradate sequence of development of the sporangia. The next resulted in a mixed arrangement of old and young sporangia within the sorus. Still more highly advanced is the condition desscribed as ‘acrostichoid,’ where the individuality of the sorus is lost and the sporangia form a ‘felt’ that covers the dorsal surface of the lamina, irrespective of the position of vein endings.

Circular phylogenetic classification of the Filicales

The various stages in soral evolution are often held to be the most important indicators of relative advancement and, on this basis, many pteridologists subdivide the Filicales into Simplices, Gradatae and Mixtae. It is important to realize, of course, that these subdivisions represent levels of evolution and not taxonomic groups. However, it is debatable whether one character should be weighted to this extent, for it is almost universally agreed among taxonomists that the maximum possible number of characters should be used in the assessment of phylogenetic status. If all the primitive characters listed above are taken into account, it is possible to calculate roughly an average ‘advancement index’ for each family or subfamily, ranging from o per cent (the most primitive) to 100 per cent (the most advanced). This has been done for the families and subfamilies selected for detailed treatment, and they have been arranged (Fig. 23) on a circular scheme, according to their advancement index. The most primitive families are near the centre and the most advanced are near the outside. The broken lines, enclosing ‘areas of affinity’, indicate which groups are most closely related to each other (in the main, the views expressed here accord most closely with those of Holttum). Such a scheme may be thought of as a view, looking down from above, of the ‘tree of evolution’ of the Filicales, and while it may not be acceptable to all taxonomists, it does avoid the error, which is common to most phylogenetic classifications, of suggesting that one modern family has evolved from another modern family.

The two most primitive families are the Schizaeaceae and the Gleicheniaceae, and they are also the oldest, being represented in Carboniferous deposits by Senftenbergia and Oligocarpia respectively. Both are represented in the Mesozoic, too (viz. Klukia and Gleichenites).


The Schizaeaceae are represented today by four genera and about 160 species, most of which are tropical or subtropical in distribution. In all of them, the sporangia are borne singly instead of in sori (‘monosporangial sori’) and they show the most primitive type of dehiscence mechanism known in the Filicales. In all, the annulus consists merely of a terminal group of thick-walled cells (Figs. 24A-D) and dehiscence is longitudinal. The stalk of the sporangium is short and thick and the spore output from each is 128 or 256. The sporangia arise simultaneously, on the margin of the frond, and are unprotected, except by the inrolling of the margin, or marginal flaps, of the pinnule. Lygodium is one of the few modern genera of ferns to have fronds of unlimited growth, forming twining structures 30 m or more in length. Unlimited growth is a feature which, in most plants, is taken to distinguish stems from leaves. When it occurs in fronds, as in this case, it is, therefore, taken as evidence that they have evolved from stem structures (or are still in the process of doing so). Further evidence that the frond of Lygodium is very primitive is provided by the structure of the leaf trace, which shows only slight departures from radial symmetry. The other three genera have leaf traces which are clearly dorsiventral and ‘gutter-shaped’. Their stem structures, too, are more advanced for, whereas Lygodium has a creeping protostelic rhizome, Schizaea has an oblique rhizome with a medullated protostele, Anemia has a creeping or oblique rhizome which is either solenostelic or dictyostelic, while Mohria is dictyostelic. It is interesting to note, also, that Mohria is the most advanced in its dermal appendages, for they are glandular scales, whereas those of the other three genera are hairs. In Anemia, only the two lowermost pinnae are fertile.

Sporangia of Filicales

The prothalli are flatthalloid structures, except in Schizaea, where they are filamentous, with occasional mycorrhizal cells and with the gametangia at the tips of short lateral filaments. Bower remarked that these filamentous prothalli are ‘the simplest prothalli known among the Pteridophyta. They suggest a primitive state, and provoke comparison with green Algae’. However, their simplicity is now regarded as the result of evolutionary specialization, instead of representing a primitive state.

Of the four genera, Lygodium has the most complex antheridial wall and the highest output of antherozoids (156).


This family is represented by about 130 species belonging, mostly, to the one genus, Gleichenia (some taxonomists prefer to split the genus into four). A number of rather different types of leaf morphology occur, two of which are illustrated in Figs. 25C and 25D, but in all of them the growth of the main rachis is arrested, until a pair of primary laterals has formed. In some species, these are of limited growth (Fig. 25C), but in others they too may terminate in dormant buds, so producing a variety of patterns, some looking superficially like a series of regular dichotomies (although, in fact, they are pseudo-dichotomies, because of the dormant apical bud in each angle). In others, there is a zig-zag arrangement of branches (Fig. 25D). As in the Schizaeaceae, therefore, the fronds are of indefinite growth, and some attain a length of 7 m or more. They arise from a creeping dichotomous rhizome which in most species is protostelic. A few, however, achieve a solenostelic condition, e.g. G. pectinata, a relatively advanced condition which is associated with a larger number of sporangia in the sorus than is usual in the genus. Yet, in this species, the dermal appendages are hairs, whereas scales are commonly present in others. Divergent facts such as these serve to emphasize the point that the evolution of different characters does not necessarily keep step, the result being that most organisms show a combination of advanced and primitive characters. This is why it is unwise to focus attention unduly on one character, when attempting to assess the relative advancement of taxonomic groups.

Leaf form

The sporangia, in strong contrast to those of the Schizaeaceae, are borne superficially on the adaxial side of the frond. They develop simultaneously and are arranged in sori containing, often, only a single ring of sporangia, seated either at a vein ending or, more usually, over the middle of a vein. There is no indusium at all covering the sorus, whose only protection is a covering of hairs or scales. Each sporangium is pear-shaped (Figs. 24E-G), has a stout stalk, and dehisces by means of an apical slit. Dehiscence is brought about by the contraction of the thickened cells of the annulus, which runs obliquely round the sporangium wall. Large numbers of sporangia are liberated from each, ranging from 128 to more than 1,000.

The gametophyte is primitive, in that it is large, massive and slow growing. When old, it becomes much fluted and develops an endophytic mycorrhizal association. The antheridia are larger than in any other leptosporangiate fern and resemble those of the Osmundales. Those of G. laevigata are as much as 100 u in diameter and contain several hundred antherozoids.


This group is commonly referred to as ‘the filmy ferns’, because of their delicate fronds, the lamina of which is usually only one cell thick. There are some 300 species of Hymenophyllum, of which two occur in the British Isles, and 350 of Trichomanes, of which there is one in the British Isles. Because of their delicate nature, almost all of them are confined to moist habitats, and most of them are restricted to the tropics, where they commonly grow as epiphytes. The British species H. tunbrigense may be seen growing on rocks constantly wetted by the spray from waterfalls.

Most filmy ferns have a thin wiry creeping, protostelic, rhizome, from which the fronds arise in two rows. In one species, the stele of the rhizome is reduced to a single tracheid, while in another there is said to be no xylem at all. Some species are completely without roots. The leaf trace is a single strand, which at the base of the stipe shows marked similarity to the stem stele but, higher up the stipe, broadens out into a gutter-shaped strand. The frond is usually much branched, each narrow segment having a single vein, but various degrees of ‘webbing’ occur and, in one species, Cardiomanes reniforme (= Trichomanes reniforme), there is a single expanded lamina. Nevertheless, the venation is open in all species.

The sori are marginal, and most species are strictly gradate. The vein leading to the sorus continues into a columnar receptacle which, in Trichomanes, can grow by means of an intercalary basal meristem until it forms a slender bristle. The receptacle of Hymenophyllum has more limited powers of growth or may lack them altogether. In such species, the sporangia are produced simultaneously, but, where the receptacle can grow, new sporangia arise in basipetal sequence. Surrounding the sorus is a cup-shaped indusium in Trichomanes (Fig. 25F, where the broken line indicates where the indusium was cut away to show the base of the receptacle) and a two-lipped indusium in Hymenophyllum.

The sporangium has a relatively thin stalk and an oblique annulus, which brings about dehiscence along a lateral line (Figs. 24J-L), by a process of slow opening, followed by rapid closure as a gas phase suddenly appears in the cells of the annulus. This mechanism is found throughout the more highly evolved members of the Filicales, and results in the forcible ejection of the spores. The spore output varies from 128 or 256 in Hymenophyllum to as low as thirty-two in some species of Trichomanes.

The prothallus of Hymenophyllum is a strap-shaped thallus, often only one cell thick, but, by contrast, the few species of Trichomanes whose prothalli have been studied have a filamentous structure which, like that of Schizaea, is mycorrhizal.


The first recorded occurrence of a fossil member of the Dicksoniaceae is of Coniopteris, from Jurassic rocks of Yorkshire. Like modern members of the group, it had highly compound fronds with marginal sori, protected by two flaps (the upper and lower indusia). In the modern genus Cibotium, the fronds are borne on stout creeping stems or on low massive trunks, while some species of Dicksonia are tall tree-ferns (e.g. D. antarctica), with a crown of leaves at the summit of a tall trunk. All are characterized by a profuse hairy covering over the stem and the base of the stipe, the hairs being as much as 2 cm long in Cibotium barometz.

The stems are solenostelic or (in species with erect axes) dictyostelic, and the stele is deeply convoluted around a large central pith region. There is a single gutter-shaped strand entering the base of the stipe, but this soon breaks up into numerous small bundles.

The sporangia are truly marginal in origin and arise in strictly gradate sequence within a purse-like box, formed by the two indusia (Fig. 25G). They are long-stalked and have an oblique annulus (Figs. 24M and 24N) which, in some species, is very nearly vertical. The typical spore output per sporangium is sixty-four.


This is a most interesting family, containing the two genera Phanerosorus, from Sarawak and New Guinea, and Matonia, from Malaya, Borneo and New Guinea. In spite of its rarity at the present day, the family had many fossil representatives in the Triassic. So characteristic is the method of branching of the frond (Fig. 25B) that there can be little doubt that the fossil Matonidium is correctly placed in this family. After an initial dichotomy, each half of the frond undergoes a regular series of unequal catadromic dichotomies (i.e. each takes the main growing point further from the median plane). Each pinna is pinnatifid and there are anastomoses in the veinlets, particularly in the neighbourhood of the sori. Phanerosorus (Fig. 25A) has a frond of indefinite growth which is long and slender and bears dormant buds at the tips of some of its branches.

The stem of Matonia is creeping and hairy, and has a very characteristic polycyclic stelar structure, with two co-axial cylinders surrounding a central solid stele. From these, a single gutter-shaped leaf trace is formed, both cylinders playing a part in its origin.

The sori are superficial and consist of a small number of sporangia arranged in a ring round the receptacle, which continues into the stalk of an umbrella-shaped indusium (Fig. 25E represents a vertical section through a young sorus). There is an oblique convoluted annulus round the sporangium, dehiscence being lateral, although there is no special stomium of thin-walled cells (Figs. 24H and 24I). The spore output is sixty-four.


This family is represented at the present day by some eight species of the single genus Dipteris, restricted to the Indo-Malayan region, but in Triassic times there were at least three genera, Clathropteris, Dictyophyllum and Camptopteris. Again, the architecture of the leaf is quite characteristic, and there can be little doubt as to the correct taxonomic placing of these fossil forms. After an initial dichotomy, the frond shows successive unequal dichotomies in an anadromic direction (i.e. towards the median plane). This pattern is represented in present-day species, in the venation of the two halves of the frond. However, while the primary veins are dichotomous, the smaller ones form a reticulum of a highly advanced type, with blind-ending veinlets, as in the leaves of many flowering plants.

The fronds arise at distant intervals along a creeping hairy rhizome, whose vascular structure is a simple solenostele. While some species have only a single leaf trace, others have two entering the base of the stipe.

The sorus is superficial, completely without an indusium, and the sporangia are interspersed with glandular hairs. In Dipteris Lobbiana the sporangia arise simultaneously, but in D. conjugata they are mixed. Thus, the single genus cuts right across the division of the ferns into Simplices, Gradatae and Mixtae.

The sporangia have relatively thin stalks (only four cells thick) the annulus is oblique (Fig. 24Q-S), and dehiscence is lateral. The spore output is sixty-four.


This is the family to which most of the tree-ferns of the world belong. Indeed, at one time, all tree-ferns were placed in it, but Dicksonia must clearly be removed on account of its marginal sori, for the Cyatheaceae, as now constituted, have superficial sori. The earliest known fossil representative of the growp is Alsophilites from the Jurassic. Bower recognized three living genera within the family: Alsophila with about 300 species, Hemitelia with about 100, and Cyathea with about 300.

Although the largest may attain a height of 25 m, some species are comparatively low-growing. Much of the diameter of the trunk is composed of matted adventitious roots and persistent leaf bases, while the stem within is relatively small. Nevertheless, its stelar anatomy is highly complex for, in addition to a convoluted dictyostele, there are abundant medullary strands, and sometimes cortical strands too. Broad chaffy scales form a dense covering over the stem apex and the base of the frond.

The stipe receives a number of separate leaf-traces from the lower margin of the associated leaf gap. While the fronds of most species are several times pinnate, those of Cyathea sinuata are simple. The venation is open in the majority of species, except for very occasional vein fusions.

The three genera recognized by Bower are distinguished by the character of the indusium but, otherwise, the sori are very similar in their gradate development. In Alsophila there is no indusium at all, in Hemitelia there is a large scale at one side of the receptacle, and in Cyathea (Fig. 25H) it extends all round the receptacle to form a cup which completely covers the globose sorus when young, but which becomes torn as the sporangia develop and push through it. Holttum, however, regards this distinction between the three genera as artificial, and prefers to merge them into the one genus Cyathea. Furthermore, he has recently changed his opinion as to the affinities of the family, for he now draws attention to the close similarity between the scalelike indusium of some species and the lower indusium of Dicksonia.

The sporangium is relatively small, with a four-rowed stalk, an oblique annulus (Figs. 24O and 24P), and a fairly well marked lateral stomium. The spore output ranges from sixty-four to sixteen, and even eight in some species.

We now come to the large assemblage of ferns whose sori show the mixed condition and which Bower grouped together in the one big artificial family, the Polypodiaceae. Some, he believed, had affinities with the Dicksoniaceae, some with the Cyatheaecae and some with the Osmundaceae, yet all had achieved the same advanced type of sporangial structure, with a thin stalk, a vertical incomplete annulus, and lateral dehiscence. Figs. 24T and 24U are two views of the sporangium of Adiantum, which demonstrate the small number of cells constituting the capsule, and the way in which the stalk is composed of just one row of cells, in the most highly evolved types.

In 1949 Holttum suggested a more nearly natural classification of these ferns, by creating a new family, the Dennstaedtiaceae, within which he grouped a number of subfamilies which, he believes, have affinities with the Dicksoniaceae. In this new scheme of classification, the Polypodiaceae constitute a very restricted family, having affinities with the Matoniaceae, the Dipteridaceae being absorbed into it. Within the Dennstaedtiaceae, so many evolutionary processes have taken place that the group is hard to define; indeed, it would almost seem that the subfamilies warrant elevation to family status.


This is the most primitive of the subfamilies of the Dennstaedtiaceae, for some species still retain the gradate arrangement of sporangia in the sorus. Most have creeping rhizomes with solenosteles. The sorus of Dennstaedtia (Fig. 25I) is very similar indeed to that of Dicksonia in having two indusia. In Microlepia, however, the upper indusium is greatly expanded (Fig. 25J), so that, in spite of its marginal origin, the sorus appears to be superficial at maturity. This represents an early stage in the evolutionary process which Bower called the ‘Phyletic Slide’, whereby the sorus ultimately has a superficial origin despite is marginal ancestry.


Davallia, likewise, has a superficial sorus at maturity, covered by a funnel-shaped indusium, but which, nevertheless, is marginal in origin. The stem is creeping, with a peculiar type of dissected solenostele, and is clothed with scales.


Nephrolepis has upright, dictyostelic stems with long runners, by means of which vegetative reproduction occurs, for the tips of the runners are capable of rooting and turning into normal erect stems. Within the genus, there is a wide range of soral form. N. davallioides (Fig. 25N) is very similar to Microlepia, in that the upper indusium is scarcely larger than the lower. In N. acuta, the sorus is superficial, not only at maturity, but also in origin. N. dicksonioides shows a different evolutionary trend, in that adjacent sori are sometimes fused, and this trend has proceeded so far in N. acutifolia that the margin of the pinna has a sorus running continuously along it, between two linear indusia.


It is generally accepted that the sorus of Pteridium evolved in a similar way to that of Nephrolepis acutifolia, for it, too, is continuous along the margin of the pinnule (Fig. 25Q) and is protected by two indusia. The upper indusium (1) is relatively thick, but the lower one (2) is thin and papery. Pteridium is one of the most successful ferns in its ability to compete with flowering plants and this may, to some extent, be due to the great depth at which its rhizomes spread beneath the surface of the soil. Its stele is a dicyclic perforated solenostele. Pteris also has a continuous sorus near the margin of the lamina (i.e. the sorus is superficial in origin) and the margin becomes inrolled to protect it (Fig. 25O). In some species, e.g. Pteris cretica (Fig. 25P), the soral region is somewhat expanded and indicates the way in which the acrostichoid condition might have evolved. Indeed, Bower suggested a close relationship between Pteris and Acrostichum. Chromosome counts support this view, but they also suggest that Pteris is wrongly classed with Pteridium. The haploid numbers are, for Pteridium fifty-two, for different species of Pteris twenty-nine and 120 and for Acrostichum thirty. The fact that in most of the Adiantaceae n = 29 or 30 suggests a possible affinity between Pteris (and Acrostichum) and this family. It should be noted that Lygodium, too, has a haploid number n = 29 or 30. Holttum believes that another acrostichoid genus Stenochlaena is closely related to Acrostichum, but a chromosome number lying somewhere between seventy and eighty casts some doubt on this. Bower placed it near Blechnum but, for the time being, Stenochlaena should perhaps remain unplaced.


Holttum leaves this subfamily unplaced in his classification, while Bower thought that it shows some affinities with the Cyatheaceae and with the Blechnoideae. It contains two genera, Matteuccia (two species) and Onoclea (monotypic). Both are markedly dimorphic, with specially modified fertile fronds. The fertile pinnae are narrow and the margins are tightly inrolled so that protection of the sorus is derived more from them than from the indusium, which is thin and papery (Fig. 25K). Both are dictyostelic and covered with scales. Matteuccia has open venation and Onoclea reticulate.


The ferns in this subfamily have a short stout stem which is more or less erect, covered with scales and dictyostelic. The stipe receives numerous leaf traces, and the venation is open. The sori are superficial on the veins, or at vein endings, and are covered by an indusium which in Dryopteris is reniform (Fig. 25L) and in Polystichum is peltate (Fig. 25M). Of these the reniform type is probably the more primitive, for it is not far removed from the condition figured for Nephrolepis (Fig. 25N). From this type, it is easy to imagine the evolution of the radially symmetrical indusium of Polystichum, by the extension of the ‘shoulders’ round the point of attachment, followed by a ‘fusion’ to form a disc, with a central point of attachment.


Some species of Athyrium have indusia that are identical in shape with those of Dryopteris, but most have two types on the same frond, as does the British A. filix-femina (Fig. 25R). Here, there are some sori with reniform indusia and some in which the indusium is extended along the lateral veins. The vascular supply to the stipe of the frond consists of two leaf traces, which unite into a single gutter-shaped strand higher up.


All the members of this subfamily are acrostichoid. There has been much discussion as to their affinities, but they probably lie with the Davallioideae, for the stele of Elaphoglossum is very similar to that of Davallia, in having two large meristeles connected into a cylinder by a network of smaller bundles.


This subfamily, too, is believed by Holttum to have affinities with the Davallioideae. The sorus of Asplenium (Fig. 25W) is extended along the lateral veins and is protected by an indusium which is usually acroscopic (i.e. its free margin is directed towards the apex of the pinna). In this, it resembles most of the sori of Athyrium. However, the vascular supply to the stipe is different from that in Athyrium, for the two bundles which enter it fuse into a single four-armed strand, instead of into a gutter-shaped strand. The same is true of Phyllitis. That Asplenium and Athyrium are closely related seems fairly certain, since they have the same basic chromosome number, n = 36, and hybrids between them are known to occur. In Phyllitis the sori occur in pairs, facing each other, along the lateral veins (Fig. 25V), one acroscopic and the other basiscopic.


Blechnum punctulatum forms a possible intermediate between Phyllitis and the more typical species of Blechnum, for on one and the same frond both types of sorus may occur, some in pairs facing each other and some showing various degrees of fusion along a commissural vein. Woodwardia has a series of box-like sori, on either side of the midrib, whose indusia are like hinged lids. The typical Blechnum sorus is a continuous one, as if the adjacent sori of a Woodwardia had become fused together, with the indusium facing the midrib of the pinna (Fig. 25U). Each has beneath it a commissural vein, which is visible in Fig. 25T, where part of the two sori have been removed to expose it. The British species, here figured, shows a considerable reduction of the fertile lamina, and this reduction process has gone much further in other members of the subgenus Lomaria, where the lamina is almost completely lacking. Such species are markedly dimorphic, for the sterile fronds have a normal unreduced lamina. The genus shows a wide range of habit, for some species are creeping, some are climbing, while several have erect trunks, like small tree ferns.


This is a very diverse family, some members of which show marked similarities with Mohria (Schizaeaceae). Their sori are without indusia and occur along the veins or else form ‘fusion sori’ near the margin, much as in Pteris. Adiantum has the sporangia restricted to the under side of special reflexed marginal flaps of the lamina (Fig. 25S). The majority of the members of the family inhabit fairly dry regions and some are markedly xeromorphic, e.g. Cheilanthes and Pellaea. However, at the other extreme, Ceratopteris is a floating, or rooted, aquatic plant, now widespread in tropical countries, where it chokes up canals and slow moving rivers. Anogramma leptophylla is interesting, in having a subterranean perennial prothallus, from which arise delicate annual sporophytes.


Within this family are placed a number of genera of ferns, all of which completely lack any kind of indusium. There are about 1,000 species in the family, almost all tropical in distribution (but note that Polypodium vulgare occurs in the British flora), and most are epiphytic. Many have highly complex anastomosing venation and some are acrostichoid, e.g. Platycerium. This genus is markedly dimorphic, with ‘nest leaves’ appressed to the tree trunk on which it is growing, while the fertile fronds are quite different in shape and give rise to the name ‘Stag’s horn fern’.

It will be clear, from this brief survey of the Filicales, that there is much scope for disagreement among pteridologists as to the relationships and detailed phylogeny of the group, and that much more research is necessary before final conclusions can be reached. The areas of affinity indicated in Fig. 23 must, therefore, be regarded as only tentative. On the evidence so far available, it would seem that the group might well be diphyletic, with two evolutionary starting points, one with marginal and the other with superficial sori. Furthermore, it seems clear that even among those with marginal origins there has been a trend towards the superficial condition. Should the Filicales prove to have been monophyletic, however, then it is most probable that the ancestral type had marginal rather than superficial sori, and that the early Superficiales underwent a ‘phyletic slide’ early in their evolution, while the Marginales are proceeding more slowly in the same direction.

‘Water Ferns’

There are two interesting groups of leptosporangiate ferns which, at one time, were classified together as the Hydropterideae. Features which they show in common are heterospory and a hydrophilous habit, but in other respects they are so different as to warrant a much wider separation, from each other and from the rest of the ferns. Accordingly, their taxonomic status has been elevated to the Marsileales and the Salviniales respectively.

List 13

All the members of the Marsileales have creeping rhizomes, bearing erect leaves at intervals, on alternate sides. The only member of the group represented in the British flora is Pilularia globulifera (‘Pillwort’). Like all species of Pilularia, its leaves are completely without any lamina (Fig. 26A). The leaves of the monotypic Brazilian genus Regnellidium have two reniform leaflets. Marsilea occurs in temperate and tropical regions, many of its sixty-five species occurring in Australia. Its leaves have four leaflets and somewhat resemble a ‘four-leaved clover’ (Fig. 26B). All have solenostelic rhizomes but, in Pilularia, the vascular structure is much reduced, and the internal endodermis may be missing.

The sporangia, in all three genera, are borne in hard beanlike sporocarps, attached either to the petiole, near its base, or in its axil, either stalked or sessile. The morphological nature of these sporocarps has been the subject of much discussion, but it can most conveniently be regarded as a tightly folded pinna (like a clenched fist) enclosing a number of elongated sori, each covered by a membranous indusium. Figs. 26C and 26D represent, very diagrammatically, the structure of the sporocarp of Marsilea as seen in vertical and horizontal sections, respectively (for clarity, the number of sori has been reduced to two rows of five). Each receptacle bears microsporangia laterally and megasporangia terminally, and receives vascular bundles from a number of strands running down in the wall of the sporocarp. Arching over the top, is a gelatinous structure sometimes called a ‘sporophore’ (cross-hatched in the figures) which swells up at maturity and drags the paired sori from the sporocarp, as it dehisces (Fig. 26E).

The sporocarp of Pilularia is similar in construction, except that there are only four sori.


The sporangia are typically leptosporangiate in origin, the sporangium wall is very thin, and there is a tapetum of two or three layers of cells. The microsporangia contain thirty-two or sixty-four microspores but, in the megasporangia, all but one of the potential spores degenerate. On dehiscence of the sporocarp, the delicate sporangium wall rapidly decays and the spores begin to germinate almost at once.

The male gametophyte (Fig. 26F) is extremely simple, as in most heterosporous plants, consisting of nine cells only. There is a single small prothallial cell, and six wall cells surround two spermatogenous cells (cross-hatched in the figure) that give rise to sixteen antherozoids each.

The first cross-wall in the germinating megaspore is excentrically placed and cuts off a small apical cell, whose further divisions give rise to a single archegonium (Fig. 26G) with a short neck of two tiers of cells, one neck canal cell, a ventral canal cell and a large egg cell.

The first division of the zygote is longitudinal, and the second transverse, giving four quadrants (Fig. 26H) of which the outer two develop into the first leaf (l) and the first root (r), while the inner two develop into the stem apex (x) and the foot (f). Meanwhile, the venter of the archegonium grows and keeps pace, for a time, with the enlarging embryo so as to form a sheath round it, from the underside of which a few rhizoids may be produced. The first few leaves in Marsilea are without a lamina and, therefore, closely resemble the leaves of Pilularia.

It is often claimed that this interesting group of ferns represents an evolutionary offshoot from an ancient schizaeaceous stock. Arguments for this are based on the leaf form, the type of hairs, the form of the sorus and the vestigial annulus round the apex of the sporangium in Pilularia, but the evidence is not very convincing and, in the absence of early fossil representatives, the group must be regarded in the meantime as an isolated one.

List 14

Whereas most of the members of the Marsileales are rooted in the soil, either in or near water, all the members of the Salviniales are actually floating. Azolla has pendulous roots, but Salvinia is completely without them. Like the Marsileales, their sporangia are borne in sporocarps. However, the morphological nature of the sporocarps is quite different, for each sporocarp represents a single sorus whose indusium forms the sporocarp wall.

The only member of the group to be represented in the British flora is Azolla filiculoides, described as recently naturalized from N. America. However, it was a native British plant in Interglacial times. It has an abundantly branching rhizome, with a minute medullated protostele, and with crowded overlapping leaves about i mm long (Fig. 27A). These have two lobes, within the upper of which is a cavity containing the blue-green alga Anabaena azollae.

Sporocarps arise on the first leaf of a lateral branch and are usually of two kinds—large ones containing many microsporangia and small ones containing a single megasporangium, although sporocarps with both types of sporangium are sometimes present. The early stages of development are similar in both types of sporocarp, for there is an elongated receptacle on which numerous sporangial initials arise. However, during development, the microsporangial initials abort in the one case (Fig. 27B) and the megasporangial initials abort in the other. In both types of sporangium there is an abundance of mucilaginous ‘periplasmodium’, which becomes organized into ‘massulae’. In the megasporangium there are four such massulae, in one of which the single megaspore is buried. Fig. 27C illustrates the dehiscence of a megasporangium, the apex of which is cast adrift as a cap over the four massulae. The megaspore then germinates to produce a cap of prothallial tissue, within which several archegonia develop (Fig. 27D).

When the micro sporangium dehisces a variable number of spherical frothy massulae are liberated, each with several microspores near the periphery. Each bears a large number of peculiar anchor-like ‘glochidia’ (Fig. 27E). These become entangled with the massulae surrounding a megaspore and, together, they sink to the bottom, where the microspores germinate, without being released from the massulae. The male gametophyte (Fig. 27F) has a single antheridium from which eight antherozoids are liberated.


The cleavage of the zygote is typical of the leptosporangiate ferns, and as soon as the first leaf appears the sporeling rises, carrying the massulae etc. once more to the surface.

There are about twelve species of Salvinia, several of which occur in Africa. Its horizontal floating stems, up to 10 cm long, have a much reduced vascular anatomy and bear leaves in whorls of three (Fig. 27G), two floating and one submerged. Whereas the floating leaves are entire and covered with peculiar unwettable hairs, the submerged leaves are finely divided into linear segments that bear a striking resemblance to roots (Fig. 27H). However, it is doubtful whether they perform the functions of roots. Growth is rapid and fragmentation occurs easily, with the result that ponds and lakes in tropical regions may rapidly become covered and canals choked.

The first few sporocarps to be formed in each cluster contain megasporangia, up to twenty-five in each, and the later ones microsporangia, in large numbers, on branched stalks (Fig. 27I). All except one of the potential megaspores in each megasporangium abort, and the functional megaspore becomes surrounded by a thick perispore (Fig. 27J), which later becomes cellular and comes to look, superficially, like the pollen chamber of a gymnosperm seed.

Within the microsporangium, the sixty-four microspores come to lie at the periphery of a single frothy massula (Fig. 27K). They remain within the sporangium throughout and, as they germinate, the male prothalli project all round. Each male prothallus contains two antheridia (Fig. 27L) producing a total of eight antherozoids.

The megaspore, too, remains throughout within the sporangium, after it has become detached. The female prothallus protrudes, as a cap of tissue from which extend backwards two narrow horizontal wings, or ‘stabilizers’. Several archegonia develop, in a row, across the upper side of the projecting cap, each with a short neck, a neck canal cell with two nuclei, and a ventral canal nucleus (Fig. 27M).

Fig 27N illustrates a young sporeling, still attached to the female prothallus within the megasporium, and shows the peculiar development of a ‘column’ (1), which separates the first leaf (2) and the stem (3) from the foot, which remains embedded in the prothallus. The early stages of segmentation of the zygote are peculiar and their morphological relationships are not fully established. At no stage is a root primordium distinguishable.

If the relationships of the Marsileaceae are obscure, those of the Salviniales are even more so. The gradate origin of the sporangia within the sporocarp, the intercalary growth of the receptacle in Azolla and the vestigial oblique annulus have led to the suggestion that the group has affinities with the Hymenophyllaceae. However, this hardly seems acceptable, in view of the many extraordinary features that mark them off from all other ferns.