Embryos or Young Sporogonia: For very young embryos none of the above reagents were very successful. Chrom-acetic acid mixtures were useless, causing the tissues to shrink and even break up. This was more apparent in the foot region of the embryo than elsewhere. Various times were tried from 6–48 hours, but the results were all poor—the longer periods being worse than the short. Formalin-acetic-alcohol [formula (a)] gave good results with all embryos after the apical cell had cut off a few segments, but for earlier stages it was only fair. However, it was distinctly better than any other agent tried. The time for fixing is fairly short (6–12 hours), but some material collected in the North Island remained in this fluid for some three months and still gave excellent results. Carnoy's fluid was not tried for very young embryos, as only a limited number were obtained. For older stages, however, it was good, and almost equal to the formalin acetic alcohol. The time that material was left in the reagent affected it greatly. Embryos left for up to 18 hours gave good results, but after this collapsing began, and one lot left for two days was quite spoiled.

Older Sporogonia: Young capsules with sporogenous tissue just set apart gave good results after fixation in formalin acetic alcohol [either formula a or b]. In the later stages, however, the alcohol may have tended to harden the material a little, and capsules of this stage cut more successfully after fixing with “strong” chromacetic acid for 48 hours. Carnoy's fluid was good for the young stages of development of the sporogenous tissue and, of course, penetrated excellently. For the later stages, however, it was poor. Capsules dropped into the fluid about the “spore tetrad” stage usually burst near the apophysis and the contents was extruded in a long spiral.

Archegonia: Strong chrom-acetic acid and Licent's formula gave good results, although in a few cases a little plasmolysis occurred. Chamberlain (loc. cit.) recommends long periods for fixing in the chrom-acetic mixtures. This was accordingly followed, the usual time being 36–48 hours. Equally good results were obtained, however, by fixing for only 18 hours. Washing was done in running tap water. Carnoy's fluid was found to cause more or less plasmolysis if allowed to act for more than 18 hours, but with shorter times was quite good.

Antheridia: Young stages, up to the cutting off of the central spermatogenous tissue offer far less difficulty than do the later stages. In the latter shrinkage frequently occurs, due no doubt to the mucilaginous nature of the contents with poorly defined cell walls. Chrom-acetic acid (“strong” and “weak” above) and Carnoy's fluid caused plasmolysis in all but the young antheridia. Absolute alcohol caused no shrinkage even in the later stages, but formalin acetic alcohol again proved to be the best of the reagents tried. It caused no plasmolysis at any stage and staining with haematoxylins was brilliant. The time for fixing seems immaterial and objects may be left in the solution for months.

fallen tree trunks, etc. It is in general not very plentiful. As a rule Cyathophorum forms pure colonies often covering a few square feet —the largest observed by the writer was some six feet by three feet. Frequently it is found in clumps of a few score plants on boulders or rocks in creek beds, etc.

The factor limiting its distribution would appear to be the humidity and the generally damp condition of the substratum rather than light intensity. The species has been found in a great variety of situations, ranging from places where light intensity was very low to those where the plants grew in bright light—always, however, the station was decidedly damp.

The plants may actually grow with water falling on them, but this is unusual and only one such habitat was observed. All the plants here were sterile.

(b) Gametophyte. The mature gametophyte develops from a protonemal stage as is usual in Mosses. The spores are green when shed, with large obvious chloroplasts, oil droplets and practically no exospore.

To obtain mature spores, capsules were artificially ripened by drying slowly in a Wardian case. For culturing protonemata broken bricks were used. These were sterilised with 4% formalin, boiled in water for half an hour and then stood in glass dishes with distilled water. The spores were sown at room temperature (viz. about 15°.) and the cultures kept under bell jars. They were left alone for some weeks and as growth appeared slow, drops of Knopp's Solution were pipetted on, and allowed to run down the slope of the brick. This caused more vigorous growth and buds soon developed on the protonemata.

Spores were found to have germinated a day after sowing and development was quite normal. The spore swells slightly and a blunt protuberance is pushed out through the exospore which splits. With further growth and transverse divisions an “alga-like” filament of indefinite length is produced with plentiful oval chloroplasts—a distinguishing feature from the green algae that frequently occur with the plants in nature. The erect gametophores arise on the protonema in the usual way.

The mature gametophore of Cyathophorum consists of a stem from two to ten inches in length bearing leaves arranged dorsiventrally (Fig. 1). There are two rows of large dorsal leaves extending laterally, and a single row of small ventral leaves. Usually the stem is unbranched, but a number of instances were seen where branching had occurred. This is an abnormal feature and will be referred to later.

The base of the stem ends in a thickish rhizomatous axis which can creep to some extent and turn erect. This is apparently a method of vegetative reproduction, as the apex, when it does turn erect, gives rise to a new leafy axis.

(c) Sex Organs. Cyathophorum is strictly dioecious. The sex organs are borne terminally on small branches in the axils of the dorsal leaves (Fig. 2) and are overlapped by the leaves of the ventral

row. Both antheridial and archegonial branches have an investment of small leaves enclosing the sex organs and paraphyses. The latter are much more abundant in the antheridial tufts, where they have somewhat swollen heads. Paraphyses do occur among the archegonia, but there are very few and they are much smaller than those in the male tufts.

Archegonial plants seem to be much more plentiful than the antheridial, and frequently whole colonies—in some cases feet across—were found composed entirely of the former.

A similar example is quoted by Johnson (23) for Plagiochila adiantoides. He advances the explanation that vegetative propagation from a few original pioneers, all female, is responsible for these large collections of one-sexed colonies.

Antheridial plants are found in similar patches, but here they are usually limited to a few dozen plants.

In most of the localities listed above Cyathophorum is plentiful. Yet in spite of this fact the percentage of plants found to be bearing sporogonia is surprisingly small—probably not more than about 10%.

The explanation seems to be due primarily to the separation of the sexes as mentioned above. There is no lack of either male or female plants, but the occasions on which they are intermixed in such a way to allow fertilisation to take place seem to be by no means frequent. This conclusion is borne out by the fact that when plants of different sexes are found together practically every archegonial plant bears several sporogonia.

The following instance may be given in illustration of this:—Sporogonia at several stages of development were found on all of a small clump of plants (about 10–20), growing on a vertical rock bank by a waterfall. No male plants were present among these, and they appeared quite isolated. However, two or three feet above this were found a number of antheridial plants with water dripping from them on to the female plants below. Clearly fertilisation had been most efficient. In contrast with this, hundreds of archegonial plants on a bank alongside had not a single sporogonium.

A longitudinal section of a sex branch shows a definite central zone of elongated cells which run into the cortex of the stem but do not attach themselves to the main strand in the latter. As already noted, a much-branched main axis is occasionally met with. This condition seems to arise from the antheridial or archegonial branches failing to form sex organs and the apical cell of some of these branches growing on to give leafy branches of some length. In all cases examined these laterally branched plants were completely sterile. Other cases were observed where instead of sex organs there was a large tuft of secondary protonemata. These tufts usually occurred in the axils of half a dozen or so of the leaves nearest the apex of the stem.

Considerable attention was given to ascertaining the season when the sex organs mature, and it can be stated that for New Zealand as a whole, sex organs and sporogonia can be found in any

off two rows of segments in the usual regular succession (Figs. 7–9). Obviously, owing to the absence of any pedicel, the stalk of the antheridium in the early stages of development cannot be distinguished definitely. All that can be said is that the lower segments (i.e., the first cut off by the apical cell) constitute the basal part of the stalk, while its upper limits are indefinite until periclinal walls appear in those segments destined to give rise to the body of the antheridium (Figs. 11–12).

Before these appear there is often a certain amount of subdivision of the lower segments cut off by the apical cell although the extent to which this occurs is variable and in general affects the lower portion of the stalk rather than the upper.

The number of segments cut off by the apical cell varies, but is usually about eight. Before the full number is attained secondary divisions begin in the lowest of the segments and extend upwards.

As seen in longitudinal section (Figs. 10–11) these divisions appear as periclinal walls cutting off an inner series of cells, the forerunner of the spermatogenous tissue, from a peripheral series which constitutes the antheridial wall.

How this is accomplished can only be made out clearly by a comparison of longitudinal with transverse sections of the antheridium. Before periclinal division sets in, a transverse section through a young antheridium shows two segments separated by their bounding wall. (Fig. 12.) The divisions which follow are best described with the aid of a diagram (I).

The original wall separating adjacent segments is represented a–a. The advent of periclinal division is marked in transverse section by the appearance of walls b–b which divide the two segments into cells of unequal size. The walls b–b are approximately parallel and meet the primary wall a–a towards opposite ends. The third series of walls c–c meets the other walls a–a and b–b as shown in Diagram I.

Diagram I.

This results in the formation of centrally placed cells, surrounded by peripheral ones. The latter from now on divide only radially, so giving rise to the one-layered wall, while the former divide in all planes and eventually produce the mass of sperm cells. The central cells from the first stand out clearly in stained sections, having large nuclei and very dense cytoplasmic contents. The spermatogenous tissue can thus be distinguished at a very early stage from those cells which constitute the wall, since in the latter the cytoplasm is much vacuolated and the nuclei somewhat smaller.

The lower of the periclinal divisions, as seen in longitudinal section, occur while the apical cell is still actively cutting off segments. According to Campbell (13, pp. 196–7), “The number of these segments is limited, in Funaria not often exceeding seven, and after the full number has been formed the apical cell is divided by a septum parallel with its outer face into an inner cell, which, with the inner cells of the segments forms the mass of sperm cells, and an outer cell which produces the upper part of the wall.”

The present writer finds that the apical cell in Cyathophorum shows no such division, but apparently retains the ability to divide further. As already noted the spermatogenous tissue can be recognised early, and when this is apparently all formed, and much sub-divided, the apical cell is still intact (Figs. 20–21). In fact, the fully mature antheridium usually shows one or two divisions at the apex directly resulting from the action of the apical cell (Fig. 23). These are not isolated instances but were seen in dozens of antheridia.

This appearance is explainable in two ways: Either the periclinal walls do not form in the last few segments nearest the apex, or else the apical cell cuts off segments subsequently to the formation of the last of the central cells. Whichever interpretation is accepted, the fact remains that the apical cell functions to a late stage and can be recognised even in the mature antheridium.

The spermatogenous cells divide very regularly in all planes, and the original walls separating adjacent primary segments can be recognised at a late stage (Fig. 22). However, eventually all the cell walls break down and the mature antheridium consists of a large number of sperms embedded in mucilage, which still shows traces of cellular structure.

The stalk is relatively long. In median longitudinal section the original divisions due to the “two-sided” apical cell can be traced, although with increase in age the regularity of the early segmentation is much disturbed. However, it is quite clear that the stalk is derived from segments cut off from the two-sided apical cell, and not due to a series of transverse divisions, formed by the original papilla (as in Andreaea) prior to the setting apart of the two-sided apical cell. The wall is one layered and its cells, except for a few at the apex, contain chloroplasts.

Ripe antheridia open in a few moments if placed in water or very dilute sugar solution (0.4%)—the latter was found to cause much more vigorous movement in the sperms when they were attempting to escape from the mass of mucilage in which they were extruded from the antheridium.

A rupture occurs between the opercular cells, of which there are several, and the complete contents of the antheridium emerges through a constricted opening. This mass of sperms, still enclosed in mucilage is often two or three times the length of the antheridium from which it came, giving an indication that the swelling of the mucilage, by inbibition of water, is one of the main factors causing the bursting of the antheridium.

Cyathophorum this early separation of the endothecium is an uncommon variation, and that normally the separation is not complete until the third series of walls is formed. This is also the case in other Bryineae which have been described by other writers.

When the endothecium has been separated, the structure shown in Fig. 59 remains constant throughout a considerable length of the upper region of the embryo (viz. for a distance of 96–120 microns in the particular examples studied). This structure as seen in longitudinal section is shown in Fig. 48. The embryo as a whole at this stage is shown in situ in Fig. 49. It has much increased in size and the swelling that eventually causes the rupture of the venter is evident. For purposes of convenience the account of the development of the young capsule is carried on from this point.

(b) Young Capsule. The nuclei of the four cells of the endothecium at this stage are very large and stain deeply. (Figs. 59–60.) There is slight variation in the divisions which follow on from the stage shown in Fig. 59, but by far the most usual case is for the amphithecium to become separated into two layers by periclinal divisions. (Figs. 60–61.) Occasionally the endothecium begins to divide before the amphithecium, but this is quite unusual, and as a rule by the time the former begins to divide the amphithecium is at least two and maybe three or four-layered.

The endothecial cells now subdivide to give a group of four central cells (c. Fig. 63) surrounded by about eight peripheral cells (p. Fig. 63). The regularity of the divisions is usually disturbed, and instead of the theoretical eight there may be ten or eleven cells in this layer (Fig. 64 p). It is from these cells that the archesporium and the inner spore sac develop, the four central cells subdividing further to make up the tissue of the columella. The peripheral cells (p. Fig. 64) divide first by radial, and then by periclinal division, into inner and outer cells. The inner cells constitute the inner spore sac (i.s. Fig. 77)—actually the outer layer of the columella—while the outer cells by a number of further radial divisions form the archesporum (Fig. 77 sp.).

As seen in transverse section the amphithecium consists in its primary stage of eight cells (Fig. 59). The first divisions in these cells are usually periclinal, cutting off smaller inner cells from larger outer ones (Fig. 60). These periclinal divisions separate what will become the wall of the capsule from the outer spore sac. The cells of the latter divide first radially (Figs. 62–64) and then periclinally so that at this stage the spore sac is two layered (Fig. 78 o.sp.s). Further periclinal divisions set in as a rule, giving an outer spore sac which may be three or four cells thick when mature (Fig. 82).

Longitudinal sections of embryos of increasing ages show how the capsule originates. Figs. 48–49 show that some distance behind the apex the cells undergo rapid division, resulting in a swelling of the young sporogonium at this point (Fig. 49 c). This swollen region is the forerunner of the lower portion of the capsule, while the tissues above this swelling give rise to the remainder of the capsule and operculum, although, of course, its limits cannot be defined as the embryo is still growing fast. The foot region at this

stage (Fig. 49) is relatively very large, and in fact it does not further-increase much in size even in the mature sporogonium. The outer layer of cells has obvious nuclei and rather dense contents, no doubt due to the rapid intaking of food from the gametophore, while the central cells are elongated. (Figs. 49 and 71.)

Where the young sporogonum emerges from the much swollen “venter,” it shows a distinct collar, very obvious in longitudinal sections (Fig. 67, semi-diagramatic). This collar appears to function in the forcing off of the calyptra—helped of course by the general elongation of the young sporophyte—and is still evident in the mature sporogonium where the seta enters the vaginula. (Fig. 68.) Rapid division in the tissues immediately above this collar initiates the formation of the seta which carries the young capsule, with its enclosing calyptra, upwards. (Fig. 68.)

Transverse sections taken through the young capsule show the development of the archesporium as already described, and also the origin of the air space. The layer of cells (1. Fig. 77) immediately outside the outer spore sac, shows very characteristically as noted by Campbell (13). They are narrow radially but rather extended laterally, and it is between these cells and the spore sac that the air space develops. The walls of the cells which will border the air space when it forms take the stain very deeply, enabling the position of the space, as seen in longitudinal section (Fig. 69) to be traced for some distance above the point where the cells have actually come apart. By the time the air space is formed the outer spore sac is two or even three layered, but the sporogenous tissue still remains in its one-layered condition for a time. (Fig. 77.) Eventually, when the full extent of the air space has formed and the general lay-out of the capsule has been completed, the fertile tissue becomes two layered—divisions appearing first at the base and extending upwards. (Figs. 72–74.) The result of these divisions is to produce the spore mother cells which come apart from one another and lie free in the space between the inner and outer spore sacs. Here they undergo the usual “tetrad” division, each producing four spores. (Fig. 84.)

A transverse section of the seta at this stage shows that the conducting strand has developed. In transverse section this is seen to consist of a few thin-walled cells, while the walls of the cortical and epidermal cells in the seta region become much thickened especially at the angles, and turn brown. (Fig. 80—full thickness of walls in cortex not shown.)

In Cyathophorum the operculum presents a characteristic appearance in longitudinal section. The apical cell functions for a very long time as already mentioned, and can frequently be recognised even in a mature capsule. (Fig. 76.) From the apex a central “core” of tissue with very regular divisions extends down to the top of the columella. The peristome arises from the fourth or fifth layer of cells from the outer margin of the operculum. In longitudinal section this layer is seen to consist of cells whose radial walls are much extended. (Fig. 81.) It is by thickening of these radial, and also peripheral, walls that the peristome teeth are formed.

Thickening is first seen on the peripheral walls and it gradually extends inwards along the radial ones until the lumen of the cell is practically filled—the nucleus remaining obvious till a very late stage in the process. (Fig. 75.) From this stage on it was found very difficult to get complete sections through the peristome, as the greatly thickened teeth usually tore away from the thinner walled tissues below.

At the junction of the operculum and theca there appears a ring-like depression which runs completely round the capsule. This marks the position of the annulus. In Cyathophorum the form of the annulus is rather different from that figured by Campbell (13) for Funaria. In longitudinal section there is one cell very much larger than those above and below, which still shows its nucleus and contents after the surrounding cells have lost theirs. Thus, as thickening of the walls of this cell occurs late, and is never well developed, it represents a place of weakness where the operculum joins the theca. (Fig. 81.) Eventually when the capsule is mature the rupture occurs at this level, and the operculum is shed. It seems that the peristome teeth must be effective in forcing off the operculum and causing the rupture of the annulus. No doubt the general drying and collapse of the central tissues will help, but the main factor seems to be the extremely hygroscopic nature of the teeth which respond to humidity changes even when protected by the operculum. The fact that when the operculum in a mature capsule is removed gently, the teeth of the peristome immediately curve outwards as if they had been held under tension, is further evidence in the same direction.

The apophysis in Cyathophorum is very poorly marked. Its cells are loosely arranged with wide air spaces, and except for the central conducting portion most of them have chloroplasts. Stomata are present on the surface, thus allowing communication between the atmosphere, the intercellular spaces and the actual air space of the capsule.

At the stage when the sporogenous cells are undergoing their final division the columella is very densely stocked with large starch grains, and the cells bordering on the spore mass appear to function as a “tapetum.” Their nuclei are large and the whole contents stains deeply. (Fig. 84) At a somewhat earlier stage, when the fertile tissue is in the condition shown in Fig. 70, starch grains are much less abundant and very much smaller. There are also fewer chloroplasts in those cells bordering on air spaces than in the later stages. Thus it seems reasonable to suppose that the bulk of the starch is the result of photosynthesis on the part of the sporogonium itself.

The calyptra is not large compared with that in some Mosses, and its development can be easily traced. The venter of the fertilised archegonium undergoes a large amount of secondary cell division and eventually becomes bell-shaped with a distinct incurved base. (Figs. 65–68.) A transverse section shows the outer cells to be much larger than those bordering on the space containing the young

two-sided apical cells, before any pedicel is apparent. (This also occurs in Mnium, Holferty 21.) The segmentation of the antheridium to cut off central spermatogenous cells is somewhat different from that described by Goebel (17, p. 13); and Ruhland (29, p. 66). The succession of walls given by them is shown in Diagram II. In the species described by these authors the walls b–b both meet the primary wall a–a at the same point. Similarly with the third series c–c. In Cyathophorum (see Diagram I) the walls b–b are approximately parallel and meet the primary wall towards opposite ends. In Funaria hygrometrica (Campbell 13) the succession of walls corresponds to that given for Cyathophorum (Diagram I). The account of the apical cell has already been given.

Diagram II.

(b)Archegonium. As already noted, there appears to be some confusion as to the exact manner in which the archegonium grows and forms the axial row. Campbell, Goebel. Holferty and others distinguish the archegonium of the Musci from that of the Hepaticae by the fact that the cover cell is active in the former, adding both to the cells of the neck and to the axial row. It thus seems fairly clear that in Musci the archegonium grows in length as a result of segmentation by the three-sided apical cell. How far this apical cell or cover cell is responsible for the cells of the canal row is however not so clear.

According to Campbell (13) in Funaria hygrometrica the terminal mother cell of the young archegonium divides transversely, giving rise to an upper cell, corresponding to the cover cell of the Liverworts, and an inner cell which produces the primary neck canal cell, the egg and the ventral cell. The cover cell functions as an apical cell cutting off lateral segments which give rise to the outer cells of the neck, and also segments parallel to the base to form the cells of the neck canal. Thus in this account all the neck canal cells arise from the apical cell, with the exception of the primary neck canal cell. This appears to persist as the lowest of the cells in the canal row lying immediately above the ventral canal cell. Campbell (loc. cit.) states definitely that “the canal cells so far as could be determined do not divide after they are first formed.” In Cyathophorum however this is evidently not the case.

G. M. Holferty (21) for Mnium cuspidatum gives the same general sequence of events except that he holds that the primary neck canal cell undergoes intercalary divisions. The cells cut off from the apical may also undergo intercalary division. Thus here the uppermost cells of the neck canal series are different in origin from those towards the base.