Sexual Generation.

The spores of the four tree-ferns studied all germinated very quickly—in two or three weeks (fig. 52). The slits of dehiscence were generally very narrow, and the spore-case remained attached. The normal heart-shaped prothallium was rapidly attained, and was similar in form and development to that of the Polypodiaceæ. But the tree-ferns' prothallium exhibits excessive variability. The apical cell may arise (especially in Dicksonia) in the cell next to the spore (fig. 57), or a long filament be formed; or even after the apical cell is formed it may grow out into a filament (fig. 62). In well-nourished prothallia, after about seven segments have been cut off, by a vertical pericline in the apical cell a three-sided initial is cut out, and a small-celled meristem now comes to occupy the depression at the apex. Normal prothallia produce a few antheridia and then archegonia on the “cushion.” “Ameristic” prothallia, as usual in ferns, produce antheridia only.

The prothallia (of Dicksonia especially) produce adventitious “shoots” very readily if conditions are unfavourable. Filiform upright branches spring especially from the margins of male prothallia, and produce abundant antheridia. In a few cases one of these “shoots” formed an apical cell and formed a normal prothallium.

Antheridia.

All the forms examined were similar in the structure of the complex normal type of antheridium and in the variety of the reduction forms.

Normal Development.

(a.) Rudiment: lighter green, and more densely protoplasmic. (b.) Cap cell. (c.) Upper ring cell. (d.) Lower ring cell. (e.) Pedicil.

In Dicksonia an opercular cell was often cut out from the cap cell, and the ring cells were sometimes divided. In the reduced antheridia few walls are formed.

Absence of Pedicel.

The sperms take some time to mature, and during this time the wall is not easily permeable. The wall seems to be chemically altered for a time, so that the nearly mature sperms may not be injured if the prothallium is suddenly wetted.

The sperms are ejected rather flatly coiled, and as soon as the pellicle is softened in the water they spring out of it as if they were in a state of great tension. This movement is very jerky, especially at first. After half an hour they swim more regularly, and straighten out more as death approaches.

The “ring wall” in Cyathea is peculiar in that it is attached to the peripheral wall. Does this give us a suggestion as to how the ring wall originated from a form as in Osmunda?

Osmunda.

Cyathea.

Campbell (“Mosses and Ferns”) considers that the antheridia are intermediate between the Polypodiaceæ and the Hymenophyllaceæ.

Archegonia.

The archegonia, as Campbell states, are simply those of the Polypodiaceæ. It was found that the chief variations were in the basal cell and the ventral-canal cell. A single basal cell was nearly always present; there were rarely two (fig. 12), and rarely the cell seemed to be absent.

The ventral-canal cell was cut off from the apex of the central cell. Rarely it seemed to be due to the primary neck cell.

In young prothallia forming the first few archegonia the divisions of the segments at the apex do show some regularity. The basal segment may become the archegonium-rudiment (fig. 21a).

Archegonia may be formed at a distance behind the apex.

View of Cover Ell from Above.

The first wall is parallel to the surface of the thallus—separating the “cover” cell, which immediately divides by a vertical wall parallel to the long axis of the thallus, and soon a wall at right angles to this follows.

The basal cell is now cut off at the base, and the central cell grows up between the cell-rows of the developing neck (fig. 13), and the primary neck cell is cut off (fig. 14), and later divides into two. When the neck is full-grown the ventral-canal cell is separated from the egg cell. When the egg is mature, and before fertilisation has taken place, the cells surrounding the egg are generally divided, so that a small-celled layer surrounds the egg (fig. 20).

Sometimes in Cyathea one or two cells break away on the opening of the neck.

The nucleus of the egg cell becomes very clear, and stains little just before fertilisation, and the nucleolus rapidly decreases in size.

Should an egg cell fail to be fertilised, the walls of the cells surrounding are rapidly cuticularised and turn brown. This process prevents bacteria and fungi from penetrating the soft walls round the egg (fig. 20). A similar cuticularisation takes place in prothallia attacked by fungi. A straight row of cell-walls becomes cuticularised, and the part invaded by the fungus is thus cut off.

Sporophyte.

Petiolar Wings.

The first few leaves have a bulky green thin-walled cortex in the petiole. But as the leaves become more robust the assimilating tissue is found only in lateral wings, and later still in clusters of thin-walled cells forming discontinuous streaks on each side of the petiole. These groups are cut off and die; a lignification of the f.v. bundles begins. They are probably for aeration of the developing leaf.

Stomata

Very numerous in first leaves, especially in Dicksonia. The mother cell is cut out from the acroscopic end of the elongating cells; auxiliary cells are absent. In the mature form (figs. 46, 47) an auxiliary cell is present, but there is much variation.

Slit of stoma parallel to line of greatest growth.

Petiole.

Dicksonia squarrosa.

In the first leaf there is a simple stele consisting of three or more tracheids grouped into a solid strand, and surrounded by two or three layers of parenchyma and an endodermis. The bundle is collateral, the few phloem elements being on one side, but the elements are more evenly distributed as we descent to the foot.

Diagram of Bundle at this Stage.
(a.) Phloem absent in the bay. (b.) Endodermis. (c.) Tracheids. (d.) Protoxylem group. (e.) Protophloom. (f.) Pericycle, with origin, with endodermis in a single original layer.

In later leaves the number of tracheids rapidly increases, and assumes the form of a shallow U, with definite spiral protoxylem in the centre (fig. 4.).

(a.) Phloem extends to here, and is not found inside the bay. (b.) Protcxylem group for insertion of pinnæ. (c.) Median p otoxylem group. (d.) Xylem elements.

A few leaves later the protophloem is broken up into three separate masses (fig. 5), but the xylem forms a continuous arc.

Later again the groups of tracheids formed round the protoxylem groups are not contiguous, and now the arc is ready to break up into three separate bundles (fig. 5).

When the stem is about ⅓ in. and the largest leaf 2 in. the petiolar bundle breaks up into three separate portions, but these three fuse together again before the pinnæ are given off.

Base of Petiole.

Just Below First Pinna

Differences between the petiole at this stage and when mature are unimportant, being only due to increase of size. In

Pinna from Petiole.

Dicksonia.

In the first leaf of Dicksonia the venation is generally dichotomous. In later leaves the successive pinnæ arise by segments, being given off from the free ends of the bundle arc. But when the bundle has three groups of protoxylem elements only the two lateral groups provide for the pinnæ.

A Series of Sections Showing the Derivation of the Pinnæ Bundle from the Petiolar Bundle

Later leaves show a similar process.

Roots.

Similar in origin in the embryo and in later development, and branching to the Polypodiaceæ.

Often in Dicksonia in slender plants there is only one root per leaf for eight or nine leaves. The first few roots hardly branch at all. In Dicksonia in the slender diarch strand there are few protoxylem elements, but in Cyathea (fig. 32).they vary

between two and five, the number partly depending on the branches given off.

When lignification of the cortex is taking place a few cells—especially well marked in C. Cunninghamii—Opposite the oligogenetic rows remain thin-walled for some time, probably as long as they are likely to produce lateral rootlets. The endodermis stains deeply in acid fuchsin, but the oligogenetic rows do not stain.

The mature roots of C. medullaris are more robust and more variable than the others. Triarch and even tetrarch bundles are sometimes found (De Bary). This calls to mind the polyarch bundles in the Hymenophyllaceæ.

The Vascular System of the Stem.

The tracheids are scalariform in the foot of the embryo, but become spiral in first leaf and root.

Figs. 25–29 show the changes in the stele at this stage as we ascend from the root (fig. 25).to the protostele above the foot. The tracheids, at first extended in a line (fig. 26), become clustered as the foot is reached (fig. 27).and turn into a horizontal position. They turn into the vertical position again, and now the phloem is clustered to one side in the collateral bundle of the petiole.

The tracheids of the second leaf fit directly on to those of the first, and so a solid strand is found. But there is generally a change from the protostele to the tubular form of stele before the third leaf is given off. But the time is very variable, and in Cyathea dealbata especially the protostele may persist for five or six leaves. Sometimes the transition took place between the foot and the insertion of the first leaf (figs. 85–88). Here a few parenchymatous cells appear among the xylem elements (fig. 86), and rapidly increase in number (fig. 87), and then the segment is given off to the leaf. Figs. 79–81 show the third leaf given off in C. dealbata from a protostele. Here a parenchyma cell appears in preparation for the giving-off the leaf, as in Dicksonia. But generally the transition in Cyathea is more irregular. Figs. 37–41 show the process in C. Cunninghamii. The sections are of the internode between the first and second leaves. The number of tracheids remains almost constant during the change.

Diagram of Xylem, Showing Transition.
(a.) Part directly below third leaf. (b.) Cauline part. (c.) Part below second leaf. (d.) Third leaf given off here a little above. (e.) Second leaf now given off here.

Figs. 82–84 show the change in C. Cunninghamii at the base of an older plant (between first and second leaf). It will be noted that there is a considerable increase in the number of tracheids over a series in a younger plant (transition also between first and second leaf). In the younger plant there is almost constantly a single layer of tracheids on the ring; while in the same internode, if the plant has now seven or eight leaves, there are two or three layers of tracheids in a similar transition region. But without a great number of series it could not be stated that there is a late differentiation of tracheids outside the primary ring.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

After the siphonostele is attained the stem increases rapidly in breadth. A well-defined endodermis is not present till the stem is about 1/10 in. long.

Sieve-tubes are ill defined in the first petiole, and it is only after six or seven leaves have been formed that the tubes begin to assume the characteristic form. Distinct sieve-tubes do not appear inside the tubular stele for a considerable time.

Fig. 43 shows typical solenostelic structure, but at once the leaf-gaps begin to elongate, and persist throughout an internode.

[In the running stems which take their origin from buds formed early in the life of the plant a robust solenostele is found till the runner nears the surface of the ground and leaves are crowded again.]

Change takes place gradually till the mature form is reached: the leaf-gaps elongate, the number of orthostichies is increased, the outline of the stele becomes wavy, and the lips project to give off the leaf f. v. bundles.

The medullary bundles of the Cyatheas do not begin to be formed till the pith is fairly broad.

Near the apex, where the developing ring is still, meristematic groups of cells are separated by parenchyma from the ring, and these give the medullary bundles (fig. 42).

Mucilage.

No signs of a mucilage system in the early stages: mucilagecells appear after the tubular stele is established: in the petiole especially they form regular rows.

Diagram Showing Origin of a Mucilage-Cell Row. (Longitudinal section of leaf.)
(a.) Apex of leaf. (b.) Mucilage row.

In the petiole the rows follow the protoxylem groups rather closely, the rows being generally in the bays of the vascular arcs.

Protoxylem.

The spiral elements of the petiole just join on to the stem, but the elements of the stem are scalariform.

In fig. 41 the two cells px are the protoxylem of the next leaf; these cells die out in the section lower down. Fig. 44 shows the stem protoxylem, but, as in Loxsoma (Gwynne Vaughan), these elements are scalariform.

Stelar Structure.

Up to the last few years consideration of the stele has been on the lines laid down by Van Tieghem, but lately more attention has been paid to the vascular structure of ferns, and a study of the ontogenetic development has modified the old standpoint. For instance, Jeffrey evidently considers the polystelic structure to be derived from the protostele through the siphonostele. For in an abstract (Proc. Roy. Soc.) of a paper (full paper not seen) which appeared in the Phil. Trans. Roy. Soc. there is the following: “Starting from the conception that the polystelic structure does not originate by the repeated bifurcation of the epicotyledonary central cylinder, but that the latter first becomes a concentric fibro-vascular tube, with gaps for the branches alone…” And in a note in the “Annals of Botany”—“Lindsaya, a new type of fern stele” (Tansley and Lulham): “Thus Lindsaya seems to furnish a phytogenetic link hitherto wanting between the protostele and solenostele, and this view is distinctly supported by the occurrence of the same stage in the ontogenetic series.”

Thus the old views are being modified. The single strands no longer make us overlook the conducting system as a whole. The internal parenchyma is excluded from the stele (Jeffrey); the endodermis is no longer regarded as of great morphological importance; and a study of the ontogeny is held to be necessary for the right understanding of any form (cf. Farmer and Hill—Angiopteris: “It would appear to be probable that no right understanding of a difficult vascular structure is possible apart from a study of its ontogenetic development”).

The presence, then, of the protostele in the early stages of modern types, and the persistence of the protostele in forms like Gleichenia and Schizæa, point to the protostele as the earliest form of stele. But there are two questions—(1.) Is this protostele made up of leaf-traces, or is it partly cauline? (2.) And how did the transition to the solenostele take place?

(1.) In forms with crowded leaves like Cyathea and Dicksonia it would be easy to agree that differentiation of the stele followed the differentiation of the petiole bundles; and in the earliest

change is somewhat similar in Helminthostachys (Lang, 1901, “Annals of Botany”).

Perhaps it will not be out of place to refer to the running stem given off from the leaf-base in Lomaria procera. The stele is at first solid, and this may grow for some distance, and even branch dichotomously. But sooner or later a weak strand of parenchyma cells appears in the centre of the xylem, and rapidly increases in bulk. An island of sclerenchyma then appears in the centre of this parenchyma, and this a little later is surrounded by an endodermis; and now phloem elements are clearly visible inside the xylem ring. The runner now presents a robust solenostelic structure. Later, when leaves begin to be given off, the leaf-gaps elongate, and typical polystely results.

From a hurried study of Aspidium aculeatum plantlets, it seemed that robust plants with a strong protostele had parenchyma cells among the xylem, and small weak plants had a small solid strand. The transition is similar to Dicksonia.

Only a study of the early stages of a large number of ferns will show whether there is any constancy in the method in which the transition is made—constancy in groups of related ferns, or even in the same fern with the sporelings under varied conditions of nutrition. I incline to think that the method of change from solid strand to tubular stele is dependent somewhat on the rapidity of growth. If growth is rapid and the stem broadens quickly, some of the elements of the xylem strand will not need to function as wood elements, and so will remain undifferentiated. This will be the beginning of the pith. It was due in the early history of the stele to broadening of the stem, and consequent loss of function of some of the more deeply placed water-carriers, and these remained undifferentiated; then the stem widened further, and the segment of the xylem cut out for the leaf extended right to the pith; and then phloem elements would extend down into the pith, because the pith, now it is not cut off from the leaves by the xylem ring, can be advantageously used for storage of starch.

Polystely is only a well-marked variety of the tubular stele. Here the continuous ring is broken up by gaps other than those above the leaf-insertion. The change from the tube to the extreme polystely of some Polypodiums—c†. P. serpens and P. novæ-zelandiæ—is due to change of stem-habit. When the rhizome becomes thick because it is used for water and starch storage, and a creeping habit necessitates no mechanical strengthening, then only those wood elements of the primitive ring are differentiated which are needed for water-carriage. The ring could have been widened and attenuated, but this would not serve so well as the network that represents the tube.

Conclusion.

The study of the structure of the few tree-ferns examined, and their comparison with other forms, makes me feel that the form of the stele is too directly adaptive to prove relationship. Among the modern ferns the function of the stem decides the form of stele. If the stem is a creeping one, and not too bulky, then a tubular stele is found—cf. some species of Pteris, Hypolepis, Polypodium punctatum, runners of Dicksonia and of Lomaria procera.

If the creeping stem is extensively used for storage of starch and water, then extreme polystely will be found. If the stem is upright and the leaves crowded, a tubular stele, with leafgaps, will result, as in the tree-ferns, and in a less developed form in large forms of Polypodium pennigerum and Aspidium aculeatum.

The transition from the solid strand to the tubular form in any particular fern now is not important from an historical point of view. Perhaps the idea that in the ferns function insures differentiation, and unless there is functioning to be done no differentiation follows, suggests how the parenchyma appeared in bulky stems in the first place; and the same tendency results in extreme polystely in some ferns now.

But as far as the relationship between Dicksonia and Cyathea is concerned, though no single similarity will prove anything, yet the similarity of means employed in the young plants in overcoming the environment at a great many points does point to a similar inherited constitution.

Explanation of Plates I-V.

Fig. 9. Transverse section, immature petiole. × 250. px., protoxylem; i.v., young tracheid; d.l., dense layer of cells surrounding the protoxylem, and growing in to form the cavity parenchyma.

Fig. 10. Bundle of leaf of Dicksonia, near end of leaflet. × 250.

Fig. 11. Bundle of leaf of Cyathea dealbata, near end of leaflet. × 250.

Plates II, III.

Figs. 12–18. Vertical (microtome) sections of prothallia of Dicksonia squarrosa parallel to longitudinal axis of thallus. The sections show the development of the archegonium. × 250.

Fig. 19. Sections parallel to surface, showing cells cut off in the parenchyma surrounding the egg cell. × 250.

Fig. 20. Similar section, showing cuticularisation of walls of venter. × 250.

Figs. 21–24. Surface views of young prothallia and their first archegonia. The shaded cells are the archegonium mother cells (C. medullaris). × 250.

Figs. 25–29. Transition from stele of root (fig. 25) to just below foot (fig. 27) to protostele of stem (fig. 29). × 250.

Fig. 30. Transverse section, first root C. dealbata. Characteristic thickened layer. × 250.

Fig. 31. Mature root D. squarrosa. c, compressed tissue. × 120.

Fig. 32. Part mature root C. Cunninghamii, showing separated protoxylem. × 120.

Figs. 33–36. C. dealbata. Four successive transverse sections near apex, showing insertion of protoxylem elements of the petiole x₁,x₂, on to those of stem s₁-s₅ s₃ and s₅ are connected with next leaf. × 250.

Figs. 37–41. Transition protostele to siphonostele in C. Cunninghamii, between first and second leaves. × 250.

Fig. 42. Early stage, medullary bundle, C. Cunninghamii. × 250.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Fig. 43. Solenostele in a Dicksonia, 1/10 in. long. × 60.

Fig. 44. Transverse section near apex of runner of D. fibrosa, showing the scalariform irregularly disposed first-formed xylem. × 80.

Fig. 45. Transverse section, stem, mature Dicksonia. Tracheids in rather regular rows, with parenchyma between. Welldefined layer of sieve-tubes. × 60.

Plate IV.

Figs. 46, 47. Epidermis developing leaf, C. dealbata and D. squarrosa.

Fig. 48. Apex leaf, longitudinal section.

Fig. 49. Stoma, nearly mature, seen from below.

Figs. 50–59. Developed prothallia, D, squarrosa.

Figs. 60–68. Cyathea Cunninghamii. Figs. 60–63, abnormal forms, due to overcrowding; figs. 64–68, antheridia on filamentous prothallia.

Fig. 69. Vertical section, embryo, with basal and quadrant walls darkened. c., apical cell, first leaf; st., stem quadrant; f., foot quadrant.

Fig. 70. Later embryo. Only root and first leaf have grown much.

Figs. 71, 72. Embryos dissected out and mounted whole.

Fig. 73. Dicksonia, six leaves; longitudinal section, showing apical cell.

Fig. 74. Transverse section, similar stage.

Fig. 75. Transverse section, mature apex of C. dealbata. Segments out-off in order (s₁, s₂, s₃).