The Gametophyte, Embryo and Developing Sporophyte of Cardiomanes reniforme (Forst.) Presl.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 74, 1944-45, Page 196

The Gametophyte, Embryo and Developing Sporophyte of Cardiomanes reniforme (Forst.) Presl. By John E. Holloway, University of Otago. [Read before the Otago Branch, March 9, 1944; received by the Editor, March 15, 1944; issued separately, September, 1944.] Introduction. In a previous paper* By an error the magnification numbers of Figs. 14–21 in this paper (Holloway, 1930) were given at twice the correct size. the writer (Holloway, 1930) has described from cultures the earlier stages in the development of the gametophytes of Hymenophyllum pulcherrimum and of Cardiomanes reniforme, the latter under its old name Trichomanes reniforme. The gametophytes of the former species were then 3 ½ years old, but were only half-grown and did not bear sex organs. Those of Cardiomanes were considerably younger, but had advanced enough to make it seem probable that their further development would be along the lines of the Hymenophyllum ribbon and not of the Trichomanes branching filament. These particular cultures were continued, that of H. pulcherrimum for a total period of 13 years, and that of Cardiomanes for a total of 6 ½ years. The present paper deals with the results obtained from the study of the latter. All drawings (except those in figs. 3a, 5, and 35–37, which were plotted to scale) were made with the aid of an Abbé camera lucida. The gametophytes and embryos were sectioned throughout at 8 microns, and the sections were stained with Delafield's hæmatoxylin, combined in some cases with orange G. The drawings that illustrate the earlier stages in the embryogeny were checked against drawings made with the aid of an oil immersion objective. The Gametophyte. (a) General Development. Figs. 1–2 represent young ribbons after 12 months growth, Branching now commonly sets in, most of the gametophytes at 2 years taking the form of a tuft of elongating and widening ribbons growing more or less erect from the intact original procumbent region (figs. 3a and 3b). Later, by the decay of the basal region, these ribbons are set free as independent gametophytes. In among these tufts a number of gametophytes of an unusual, completely filamentous form were found. Fig. 4 shows one of these, taken when the culture was 3 years old. The further development of these is not known, but it is clear that the form of the gametophyte is very variable in its earlier stages of growth.

Both antheridia and archegonia appeared at or near the margins of the ribbons at 2 ½ years, when the latter were up to 7 millimetres in length. The antheridia soon became abundant, even on small ribbons, and were always more copious and widely scattered than the archegonia. The latter were confined to well-formed ribbons, and were in groups. The ribbons by this time had branched freely. After 6 ½ years growth they attained a length of 10–12 millimetres. Fig. 5 represents a well-grown, branching ribbon of typical form, bearing antheridia and archegonia, and also two young embryos on the hindermost branch. On the culture gametophytes of H. pulcherrimum adventitious branches (gemmæ) occurred in great profusion in successive crops, and by becoming detached aided in the formation of a dense turf of ribbons. On the other hand the Cardiomanes gametophytes produced no gemmæ, although the cultural conditions were apparently identical in the two cases. The two cultures differed also in the fact that no antheridia were ever observed on the Hymenophyllum gametophytes although repeatedly searched for; however, on two occasions a cluster of a few sporelings was found. (b) The structure of the ribbon. Over the greater part of its area the Cardiomanes ribbon is one cell layer in thickness, with localised marginal archegonial cushions (Fig. 7). It thus corresponds in structure with that of Hymenophyllum. The sectioning of many ribbons, however, revealed that not infrequently there are small areas well away from the margins that are two-layered. These areas vary in size from single cells that have divided in the plane of the ribbon, to patches of several such cells, and to more extensive areas. Archegonia may be borne on the larger of these areas, on either one or both surfaces (Fig. 8) there being then a still further increase in the thickness of the area to form a cushion. The writer has found these two-layered areas in the culture gametophytes of H. pulcherrimum also. The marginal cushions bear archegonia on one or on both surfaces (Fig. 7). The continued growth of these cushions, especially after fertilisation has taken place, commonly leads to an irregular shape and curving of the margin, with a consequent difficulty in orienting for sectioning the archegonia and the embryos that are present. (c) The sex organs. The development of the archegonium and of the antheridium was not followed. The mature archegonia (Fig. 9) have practically straight necks. The antheridium (Figs. 10–11) is considerably larger than that of typical leptosporangiate ferns. It corresponds with the antheridium of Hymenophyllum more closely than with that of Trichomanes. There is no cap cell, and there is a larger number of wall cells and of spermatocytes than in Trichomanes. The number of spermatocytes, as seen in a median vertical section of the antheridium, was found to be rather variable (cf. Figs. 10 and 11), in some cases being considerably greater than the number given by Bower for Hymenophyllum (1923, pp. 291–2).

Up to the age of 3 ½ years the cultures were watered only from below. Fertilisation was clearly thereby prevented, up to this point no embryos being observed, although both kinds of sex organs had been present on the ribbons for 12 months. At this stage of the culture the gametophytes were watered several times from above by thorough pipetting in order to induce fertilisation, with the result that it took place copiously. Within two months embryos were present in abundance. The Embryo. Over eighty embryos were studied in serial sections, ranging from the unsegmented fertilised egg up to the stage at which the cotyledon and primary root are elongating (Fig. 6). Since, according to the most recent accounts of the family, e.g., Smith (1938, p. 319), the embryogeny has hitherto not been known at all in any member of the Hymenophyllaceae, a somewhat detailed account will now be given of it in Cardiomanes. The features of especial significance for the purpose of comparison with the embryo of other Filicineae are the position of the first or basal wall, and the time and place of origin of the main body organs. (a) The early developmental stages. A considerable number of fertilised eggs were met with in the sections (Fig. 7). In determining the precise direction of the first or basal wall in the segmenting zygote it is necessary to visualise this wall in its relation to the embryo in its entirety, and not merely as seen in the plane in which the embryo happens to be cut. In the plane of sectioning the basal wall may appear in a single section to be transverse to the axis of the archegonium, whereas it may actually be inclined in the longitudinal plane at right angles to this, and thus must be described as an inclined wall. The behaviour of this wall in this latter plane can be ascertained by noting its position in the embryo in the successive sections of the series, when, if it is inclined in this plane, its position will be found to change from section to section. Figs. 12a-c illustrate three successive and more or less median longitudinal sections of a 2-celled embryo, in each of which the basal wall appears as transverse, whereas a comparison of the three figures shows that it is actually slightly inclined. Altogether seven 2-celled embryos were examined, and eight that were 3–6 celled as seen in median longitudinal view. In some of these the basal wall was found to be strictly transverse as observed in the two longitudinal planes, whereas in others it was more or less inclined. In none of them could it be described as truly vertical such as it is in the typical leptosporangiate ferns. In Figs. 13–14 are illustrated two 2-celled embryos in which the wall has a greater or less degree of inclintion. That shown in Fig. 14, in which the inclination was the greatest that was observed, was situated on the opposite side of the same cushion as that shown in Fig. 12. Thus in Cardiomanes the basal wall can in general be stated to be more or less transverse to the longitudinal axis of the archegonium, a point that is of importance in comparing the embryo of the Hymenophyllaceae with that of other Filicineae.

The epibasal and the hypobasal cells next subdivide, either of them being the first to do so (Figs. 15–16), to form an embryo that is 4-celled as seen in median longitudinal view (Figs. 18–19). Two abnormal examples of a 3-celled embryo were found, one of these (Fig. 16) being of a “giant” nature, and the other (Fig. 17) having undergone two successive transverse divisions. Comparison of the two embryos in Figs. 20–21 indicates that there is no regularity in the sequence of cell divisions that immediately follow. The outline of the hypobasal portion of the second of these embryos indicates that it has been cut somewhat obliquely. Eleven embryos of about the same size as the three shown in Figs. 22–24 were sectioned, and in none of them did a careful scrutiny of the entire series of sections reveal any cells that could be interpreted as apical cells. It would be very difficult to recognise apical cells as such for certain until they have begun to segment, unless, as in the case of the advanced type of leptosporangiate embryo, they originate at a fairly definite and early stage in the embryogeny. The size of the embryos illustrated in Figs. 22–24 can be held to show that in Cardiomanes the cotyledon and the other body organs are set apart relatively late. (b) The origin of the main body organs. It has already been stated that difficulty was experienced in orienting the embryos for sectioning. Not only do the marginal cushions usually curve over irregularly, but also, on account of the varying position of the basal wall, the cotyledon and root apices have no constant position relative to the neck of the parent archegonium or to the longitudinal direction of the gametophyte ribbon. Thus to obtain median sections of a globular embryo that show the origin of its growing points is a matter of chance. Nine embryos were sectioned in which the apical cell of the cotyledon alone was present. The three that have been selected for illustration (Figs. 25–27) indicate the size of the embryo at this stage. Figs. 25 a-b represent two consecutive sections through an embryo in which the cotyledon apical, which appears in Fig. 25a cut somewhat obliquely, has just begun to segment. Figs. 26 a-b represent two non-consecutive sections of an embryo of an extra large build, at a similar stage. Fig. 26a shows the full size of its epibasal region and the presence of the cotyledon apical cell, and Fig. 26b shows the full size of the hypobasal region. In Fig. 27a the cotyledon is at a slightly later stage of development as indicated by the arrangement of the cells at this point. The original basal wall is clearly to be seen in Figs. 26–27, as also in the earlier Figs. 22 and 24, lying more or less transversely to the direction of the neck of the archegonium. In these three embryos it was doubtful whether or not the apical cell of the primary root had as yet been set apart. No indication of such an apical was found in the embryo shown in Fig. 26. In Fig. 25 the rather symmetrical arrangement of the cells around the one marked X may possibly indicate that the latter is the apical cell of the root in an endogenous position and cut transversely. Fig. 27b represents one of the series of sections through the embryo shown in Fig 27a, on the opposite side of the epibasal region to the cotyledon, cut below the surface. Here also it is possible to interpret

the cell marked X as a segmenting root apical cut more or less transversely. On the whole, however, it may be concluded that the root originates a little later than the cotyledon. Figs. 28 a–b are two non-consecutive sections of an embryo that has developed a little further than those just described. The position of the original basal wall is not clear. The apical cells of both cotyledon and root have cut off several segments, the cotyledon being somewhat more advanced in growth than the root. A cell at the base of the cotyledon away from the position of the root may be interpreted as the stem apical. In older embryos (e.g., Fig. 32) the position of the stem apex is always clearly indicated by overlapping hairs, and in this particular case young hair-like protrusions of surface cells were to be seen immediately below the stem apical. Thus in this embryo, the cotyledon, stem and root all come from the region that lies nearest to the archegonium neck. If it can be assumed that in this embryo the basal wall was transverse or only slightly inclined, it would follow that the three organs originate from the epibasal region, the hypobasal region constituting the foot. The present study, however, has not succeeded in determining precisely the point of origin of these organs. In the embryo shown in Fig. 29 the root apex points directly towards the neck of the archegonium, the basal wall having been probably strongly inclined. The root is apparently endogenous in origin. Fig. 30a is a general view of an embryo similar to the last. The cotyledon and stem apices are cut medianly and are shown in greater detail in Fig. 30b. In Fig. 30a the dotted line separates the large-celled region of the embryo from that which bears the three growing points. The position of the root apex which appeared in another section cut obliquely, is indicated by an X. Figs. 31 a-b are two consecutive transverse sections through the apical cell of the root at about this age, showing its form and mode of segmentation. Figs. 32 a-b are two consecutive sections of another embryo in which the stem apical has commenced to segment. Here, as in Fig. 29, the root apex seems to be endogenous. A number of embryos were sectioned illustrating the succeeding stages of growth. The elongation of cotyledon and root in opposite directions (Figs. 34 and 6) is at first fairly fast as shown by observation of the cultures, but further growth is very slow. The cotyledon remains in alignment with the root, usually without any curvature. As in other ferns the stem is from the first especially slow in developing. Its conducting strand is formed later than that which connects cotyledon and root, and joins up laterally with this. Considering the slender nature of the Cardiomanes gametophyte, it was surprising to find how many of them bore a plurality of embryos. In one case there were as many as seven on one ribbon. There were several cases of two, and one of three well grown embryos on the same cushion. It was noticeable that starch frequently accumulates in and around a half-grown embryo. The foot is large, but in the older embryos it tends to be in direct contact with the cushion only at its base, so that these embryos project strongly (Figs. 34 and 6).

The Developing Sporophyte. The following account is based upon the study of a large number of sporelings obtained from the cultures and, in the case of the later stages, from several forest localities. They have provided a continuous developmental series in stem and frond form and structure. The most advanced of the culture sporelings were about 3 years old, and bore on a short, erect stem 4 or 5 linear fronds up to ¾ inch long. The youngest of the field sporelings were similar to these, those that were older bearing on the primary erect stem and the subsequent young rhizome fronds that showed a progressive transition from a linear to a lobed form, and finally to a form closely comparable with the characteristic reniform frond of the mature sporophyte. (a) The form and mode of growth of the frond. The growth of the cotyledon is altogether apical, but the mode of its growth passes through several distinct changes. It at first elongates as a cylindrical organ without lamina wings. Early in this stage the original apical cell changes to one of a dome-shaped form by the progressive flattening of its cutting faces, so that it now segments by walls parallel to its flat base (Fig. 33). Next, narrow one-layered lamina wings begin to develop at the apex owing to the dome-shaped apical changing to one of a wedge-shaped form with two cutting faces. The actual origin of this change was not observed. When the cotyledon is a few millimetres long its apex becomes occupied by a marginal series of segmenting cells which divide anticlinally and periclinally, the lamina wings now progressively increasing in width. In his study of frond rudiments of the mature sporophyte of Cardiomanes, Bower (1889, p. 331) describes the transition from a single wedge-shaped apical to a marginal series as taking place by a T division of the former. The cotyledon preserves the linear form, but occasionally exhibits a weak single forking which was observed to originate as a true dichotomy of the apex. The fronds next formed either remain of a simple linear form or are once forked. In tracing the ontogenetic sequence in frond form in the field sporelings comparison will be made between fronds that are fully grown, such a condition being indicated by the complete cessation of cell-divisions in front of the vein tips. The fronds that succeed show a progressively greater extent of vein-forking accompanied by lobing of the forward region of the lamina (Figs. 35–36). As they succeed one another this lobing becomes more symmetrical (Fig. 36), then it becomes less pronounced (Fig. 35), until finally it is completely absent (Fig. 37). Fronds with a continuous, unlobed lamina show no lobing at any stage of their growth. From this point in the ontogeny, all succeeding young fronds take the form of an elongated petiole terminating in a very small entire lamina which increases in extent by growth all around the margin of the forward part. Reference must also be made to the growing points during this sequence. Where lobing is present, meristematic activity remains apical, the several growing points being thus well separated. In the later fronds the progressive elimination of lobing, together with the increasing number of dichotomies of the growing points, brings

it about that the latter remain in progressively closer juxtaposition to one another. They are placed around the periphery of the forward region of the lamina, the margin of the hinder cuneate region consisting of permanent tissue. In Fig. 37 is represented a half-grown unlobed frond from a sporeling rhizome. The finely crenulate forward margin is occupied by a considerable number of closely spaced growing points, each lying in a shallow sinus in front of a vein-tip. With the increase in size of such a frond the growing points undergo further dichotomies with the result that thy come to lie so close together as to be practically continuous. In subsequent fronds dichotomies of the growing points follow one another in quick succession as soon as the lamina begins to form at the tip of the petiole. The hinder cuneate region of the lamina thus becomes less well developed, and the strong growth of the forward region leads to the well-known reniform shape of the frond as a whole. (b) The structure of the lamina. In the Hymenophyllaceae the lamina of the frond of the mature sporophyte is typically one cell-layer thick. Certain exceptions to this are known, the several-layered lamina of Cardiomanes and of H. dilatatum being well-known cases. Bower (1889, p. 342) states that in some other species the lamina is several-layered near the veins. The sporelings of Cardiomanes afforded an opportunity for tracing the manner of origin of this several-layered condition and of certain other features of interest. In the upper part of an occasional well-grown cotyledon, the vein is flanked on both sides by a narrow several-layered belt of lamina tissue of irregular width. This belt is readily distinguished in surface view, sections showing that it is a 3-layered extension outwards of the parenchyma that surrounds the vein. In the fronds that immediately follow this belt is a usual feature, and is somewhat wider in extent. Fig. 38 represents a transverse section of such a frond where the belt is present on one flank of the vein, but absent on the other. In fronds that are beginning to show “webbing” of the lamina lobes the 3-layered condition progressively extends over that part of the lamina that lies between the veins, from the hinder part of the frond forward into the lobes, following each vein towards its tip as a narrowing belt (Fig. 36). Fig. 39 shows in section this extension taking place at the point indicated on the frond in Fig. 36. Finally, in completely webbed fronds, the condition is reached in which the entire lamina is 3-layered except for a continuous marginal zone. In the forward region of such a frond this zone is very narrow (Figs. 37 and 40), lying beyond the position of the vein endings, but it widens out in the hinder cuneate region. Thus the several-layered structure originates in the neighbourhood of the veins, and extends progressively outwards from these by cell-divisions taking place in lamina tissue that had previously been inactive. For the purpose of comparison, the writer has examined the frond structure in all the New Zealand species of Hymenophyllum and ofTrichomanes and has found that, in addition to Cardiomanes,

only H. dilatatum and H. scabrum show any departure from the one-layered condition of the lamina. Field sporelings of each of these two species were available for study. The several-layered condition of the lamina originates in the sporeling frond, at a late stage in the ontogeny of H. scabrum but much earlier in H. dilatatum, in precisely the same way as in the sporeling of Cardiomanes, viz., as a 3-layered belt of irregular width flanking the veins of the frond segments. In progressively later fronds it extends outwards until in H. dilatatum, but not invariably in H. scabrum, it reaches to the margins. In tracing still further the ontogenetic sequence in frond structure in Cardiomanes, it was found that in the entire fronds that next arise on the young rhizome a modification of the mode of segmentation of the marginal cells sets in that results in the formation of a strengthening rib around the forward part of the frond periphery. Fig. 41 shows the actual origin of this in a young frond, and Fig. 42 shows the extent to which the rib has developed by the time such a frond is fully grown. With the inception of the rib further extension in the size of the lamina is restricted to intercalary cell-divisions in the zone within the margin. The marginal rib is a prominent feature of the frond of the mature sporophyte (Fig. 43). In sterile fronds it is continuous around the lamina to its base. In fertile fronds it is present at the places where sori are absent. There is a well-defined cuticle over the rib surface, the lamina cuticle being more delicate even on old fronds. The lamina is several-layered right up to the rib, the number of layers finally becoming four by the subdivision of the central large-celled layer. The youngest fronds examined of mature sporophytes, in which the lamina was still very small, were found to possess a fairly well-formed rib. Bower states with a figure (1889, p. 344 and Fig. 51) that the “young” frond of the well-grown sporophyte of Cardiomanes has a marginal series of hemispherical segmenting cells whose segments almost immediately subdivide to form a several-layered lamina. He does not mention the marginal rib. The present writer did not observe the condition described by Bower, probably because it occurs only in the lamina while still circinnately inrolled. Such a condition would be comparable with that which is present at a much earlier stage in the ontogeny. A marginal series of this type is a characteristic feature of the Hymenophyllaceae, but is only transitory in Cardiomanes. The rib is no doubt of importance in preventing tearing of the broad lamina. It is well known that Cardiomanes can occur in more or less exposed positions, and that under dry conditions the lamina inrolls strongly from both sides. (c) The stem stele. On the short, erect primary stem of the sporeling the phyllotaxy is spiral. With the sideways curvature of the stem apex to give the horizontal rhizome the fronds become more widely spaced apart, and are borne laterally along the two sides of the rhizome, so that the latter is dorsiventral in habit.

In the sporeling erect stem the stele is a typical protostele, with a central solid mass of tracheids surrounded by phloem and pericycle (Fig. 44). As seen in transverse section the protoxylem is not distinguishable. Fig. 45 shows the appearance of the stele of an erect stem at the point of departure of a frond trace, it being evident that the stele is undergoing a more or less equal dichotomy. The inception of forking is indicated by the oval form of the stele as a whole, the interruption of the phloem ring, and the presence of parenchyma breaking across the xylem mass. In the horizontal rhizome there is a distinct and sudden change in stelar form. This is illustrated in Fig. 46, which represents the stele of a very young rhizome about 5 millimeters long, on which no fronds had as yet become apparent. The ring of phloem is well developed, but the tracheids are few in number, those placed centrally probably being the protoxylem. After further growth the rhizome stele begins to show an arrangement of the tracheids comparable with the two-banded arrangement in the mature rhizome. In Fig. 47 is shown the stele of a rhizome that was 4 centimetres in length and bore several fronds, the section being taken about halfway along the rhizome. The metaxylem takes the form of an upper and a lower band, with the protoxylem in a central position embedded in parenchyma. Boodle (1900) has described and figured the stele of the mature rhizome of Cardiomanes and of various other members of the family. He states that in the larger species of Hymenophyllum and in Cardiomanes the frond and root traces come off from the ends of the xylem bands at the two lateral lines, and that there are no leaf gaps. The rhizome stele is thus of protostelic nature, although of a modified form. Discussion. In his account of the Hymenophyllaceae Bower (1926, p. 248) has set out the arguments in favour of regarding this family as “relatively primitive,” while showing features of specialisation to a hygrophilous habit particularly along the lines of reduction. He regards Cardiomanes, together with certain large species ofHymenophyllum, as on the whole best representing the ancient stock of the family. It remains now to discuss briefly certain features in the life history of Cardiomanes, described in the present paper, in connection with these views. The gametophyte is of the Hymenophyllum ribbon type. The altogether filamentous gametophyte characteristic of Trichomanes is commonly regarded as being much specialised, so that in accordance with this view Cardiomanes and Hymenophyllum have retained a more primitive form. Bower, however, (1923, p. 294) expresses a less decided view as to which of these types is the ancestral one, holding it to be tenable that the ribbon is the derived form. At the same time he considers that the filicinean gametophyte does not provide sufficiently trustworthy material for phyletic argument. In the form of its gametophyte Cardiomanes stands nearer to Hymenophyllum than to Trichomanes. It may be held that since this similarity involves a number of distinct points of detail it is more likely to be due to affinity than to homoplasy.

The antheridium is more like that of Hymenophyllum than of Trichomanes, in its general size, the relatively large number of wall cells and of spermatocytes, and in the absence of a cap cell. The antheridium of Hymenophyllum shows a closer similarity with that of such ancient forms as Osmunda and Gleichenia than with that of the advanced Leptosporangiatae. That this represents in the Hymenophyllaceae a persisting ancient feature is emphasised by the fact that in the Cardiomanes antheridium there is an even larger number of spermatocytes than in that of Hymenophyllum. In the embryogeny of Cardiomanes primitive features are especially in evidence. The first or basal wall is more or less transverse to the longitudinal axis of the archegonium as in certain eusporangiate ferns, and not parallel with it as is usual in the leptosporangiates. This is remarkable, since even such ancient leptosporangiates as Osmunda and Gleichenia have the latter type of embryo segmentation. La Motte (1937, p. 702) has shown that in the very young embryo of Isoetes the position of the basal wall can experimentally be made to change in response to changes in the direction of the gravitational stimulus. In Cardiomanes the basal wall varies somewhat in position. It is possible that this may be due to variation in the position of the young embryo with respect to the direction of the gravitational stimulus, since the archegonia are not oriented in any fixed position. Even if such be the case, the fact still holds that a considerable number of young embryos were studied in all of which the basal wall was more or less transverse. In the relatively late differentiation of the main body organs the embryo of Cardiomanes corresponds with that of the eusporangiate ferns and of Osmunda, and not with that of the advanced leptosporangiates. It preserves a globular form for a considerable time, and is a fairly massive body by the time the cotyledon is set apart. The primary root apparently arises endogenously, as in the more ancient fern types. On account of the varying position of the basal wall it has not been possible to determine definitely from which region of the two-celled embryo the several body organs are derived, but in the opinion of the writer the cotyledon, primary root and stem all come from the epibasal cell, i.e., the cell nearest to the archegonial neck, the hypobasal cell forming only the foot. The developmental stages figured, with the exception of the earliest, correspond very closely with those figured by Cross (1931) for Osmunda cinnamomea. The sporeling possesses, at some or other stage in its development, such frond features, usual in the family, as the more or less dissected form, the one-layered lamina, and a marginal series of cells that segment in the manner characteristic of the Hymenophyllaceae. These features become progressively replaced by others that finally lead to the very specialised frond of the mature sporophyte. Thus with respect to these features Cardiomanes provides a good illustration of the doctrine of recapitulation. Bower (1926, p. 238) has stated the view that the occurrence of a several-layered structure of the lamina in Cardiomanes does not represent the persistence in this genus of an ancestral character, but rather a tertiary return to it. The ontogenetic facts can be held to substantiate this

view. It is noteworthy that two other members of the family, namely, H. dilatatum and H. scabrum are similar to Cardiomanes both in the possession of a several-layered lamina and in the manner of its origin. The typical protostele of the erect primary stem must also be referred to. It is well known that in many ferns in which the mature stele is of a relatively complex form, there is an initial protostelic stage which is regarded by Bower (1923, chap. 7) as a persisting primitive feature. There can be no doubt that the Cardiomanes sporeling does recapitulate certain frond characters that are typical of the family, but absent from its own adult frond. Such frond characters are thus ancestral for Cardiomanes. It is possible to regard the protostele of its erect stem as a persisting type of stele that originally gave rise to the various modifications that Boodle has described (1900) in different members of the family. The evidence of recapitulation, however, is not so cogent here as in the case of the frond characters referred to. Thus certain facts with regard to the life history of Cardiomanes are in accordance with the view that the Hymenophyllaceae is a relatively primitive family. The structure of the antheridium, possibly also the primary protostele, and especially the embryogeny, bear this out. It remains to be seen whether or not other members of the family have the same type of embryo. It appears that Cardiomanes has closer affinities with Hymenophyllum than with Trichomanes with respect to the type of gametophyte and the structure of the antheridium, as well as with respect to the number of spores per sporangium and the stele of the mature stem as other writers have shown. At the same time it has developed specialised frond characters that enable it to withstand drier conditions than those usual for the family, these characters having developed in a frond that at an earlier period showed the same type of specialisation in frond-lamina structure as has the family as a whole. Literature Cited. Boodle, L. A., 1900. Comparative Anatomy of the Hymenophyllaceae.Ann. of Bot., vol. xiv, pp. 455–96. Bower, F. O., 1889. The Comparative Examination of the Meristems of Ferns as a Phylogenetic Study. Ann. of Botany, vol. iii, pp. 305–92. —– 1923. The Ferns, vol. i. Camb. Univ. Press, London. —– 1926. The Ferns, vol. ii. Camb. Univ. Press, London. Cross, G. L., 1931. Embryology of Osmunda cinnamomea. Bot. Gaz., vol. xeii, pp. 210–17. Holloway, J. E., 1930. The Experimental Cultivation of the Gametophytes of Hymenophyllum pulcherrimum Col. and of Trichomanes reniforme Forst. f.Ann. of Bot., Vol. xliv, pp. 269–84. La Motte, C., 1937. Morphology and Orientation of the Embryo of Isoetes. Ann. of Bot., vol. i (N.S.), pp. 695–715. Smith, G. M., 1938. Cryptogamic Botany, vol. ii. McGraw-Hill Book Co., New York.

Figs. 1–11.—The gametophyte of Cardiomanes. Figs. 1–2.—Early setting in of the ribbon habit, x40. Fig. 3a.—Tufted form at 2 years, x17. Fig. 3b.—Basal region of the same, x40. Fig. 4.—Unusual filamentous form, x40. Fig. 5.—Typical ribbon at 6 ½ years, x5. Fig. 6.—Surface view of a marginal cushion bearing a well-grown embryo, x40. Fig. 7.—Vert. sect. across a marginal cushion, x120. Fig. 8.—Vert. sect. across a ribbon showing a centrally-placed cushion, x80. Fig. 9.—A mature archegonium, x160. Fig. 10–11.—Two mature antheridia in median vert. sect., x160.

Figs. 12–27.—The earlier stages in the embryogeny. All the embryos are cut in a plane more or less parallel with the neck of the parent archegonium. All Figs. 220. Fig. 12a-c.—Three consecutive sections of a 2-celled embryo. Figs. 13–14.—Two 2-celled embryos showing variation in inclination of the basal wall. Figs. 15–16.—Two embryos at the 3-celled stage, that in FiG. 16 being of an abnormal gaint size. Fig. 17.—Another abnormal 3-celled embryo. Figs. 18–19.—Two 4-celled embryos. Figs. 20–21.—Two embryos somewhat further developed, the latter cut obliquely. Figs. 22–24.—Three older embryos still lacking growing points. Figs. 25–27.—Three embryos in which the cotyledon apical has been set apart: Figs. 25a-b are consecutive sections, but Figs. 26a-b and also Figs. 27a-b are not consecutive; the cell marked x in Figs. 25 and 27 may possibly be the root apical.

Figs. 28–33.—Later stages in the embryogeny. All Figs. (except 30a) x175. Figs. 28a-b.—Two non-consecutive sections of an embryo showing the cotyledon, root, and young stem apicals. Fig. 20.—Embryo with cotyledon and root apices further advanced, the latter being apparently endogenous. Fig. 30a-b.—A similar embryo: Fig. 30a in general view x88: Fig. 30b, details of cotyledon and stem apices. Fig. 31a-b.—Two consecutive trans. sects. through a root apical at this age. Fig. 32a-b.—Two consecutive sections of an embryo showing the position of the three apicals: the root is apparently endogenous. Fig. 33.—The dome-shaped apical cell of the cotyledon at a still later stage.

Fig. 34.—Embryo showing elongation of both cotyledon and root, x150. Fig. 35.—A sporeling showing erect primary stem and young rhizome, with sequence in frond form, nat. size. Fig. 36.—A lobed sporeling-frond showing outline (dotted) of 3-layered region of the lamina, x2. Fig. 37.—A young entire sporeling-frond showing outline (dotted) of 3-layered region, x2. Fig. 38.—Vert. sect. across a linear sporeling-frond showing several-layered belt of lamina on one side of the vein, x57. Fig. 39.—Vert. sect. across a lobe of the frond in Fig. 36, at the place marked. showing extension of the 3-layered region outwards from the veins, x57. Fig. 40.—Vert. sect. across margin of the forward region of the frond in fig. 37 showing marginal growth and extent of the one-layered region, x57. Figs. 41–42.—Vert. sects. across margin of forward part of a later entire sporeling-frond, when young and when fully grown respectively, showing formation of marginal rib, x57. Fig. 43.—Vert. sect. across margin of a well-grown sterile frond of a mature sporophyte, showing full size of marginal rib and absence of a one-layered region, x57. Fig. 44.—Trans. sect. of the protostele of the erect stem of a sporeling, x365. Fig. 45.—Trans. sect. of the protostele of the erect stem of a sporeling, at the point of departure of a frond trace, x365. Fig. 46.—Trans. sect. of the modified protostele of a very young sporeling rhizome, x365. Fig. 47.—Trans. sect. of the xylem region of the modified protostele of an older sporeling rhizome, x365.