Further Notes on the Gametogenesis and Fertilization of Sponges

My previous work on the sponge Grantia was incomplete in several respects, because even after some years’ work all the required stages had not been found. I could not satisfy myself as to the nature of the granules in the choanocytes, other than the Golgi element, which was correctly identified. It was not possible to make a report on the early changes in the choanocyte and spermatozoon after the entry of the latter into the former. The oogenesis was not properly studied because in none of my successful Golgi preparations did eggs occur. The early and late stages of spermatogenesis were not discovered.

The discovery of the behaviour of the coarse granules found in the choanocytes of Grantia compressa, and figured by Minchin and Dendy, has enabled me to identify such granules as mitochondria. In my previous paper I called them ‘mitochondria and yolk ‘in an uncertain manner: at that time I could not feel satisfied that such irregular large granules could be mitochondria alone, because mitochondria of other invertebrates are rarely if ever of such an appearance. The difficulty was that one could not imagine such coarse irregular granules taking part in the cell divisions associated with rapid growth, whereby the few coarse granules might ultimately become segregated in a few cells, and be absent in others. On the other hand, no fine equal-sized granules had previously been noted in Granti a by any observer. Haeckel, Dendy, and the present author had stated their belief in the metamorphosis of collarcells into egg-cells. If such took place it was indicated that spermatozoa might be formed in that way too. The occurrence of the coarse granules in collar-cells did not make this likely, because at the spermatid stage some cells would contain very coarse granules, others none at all, or at best very few. In no animals hitherto described do spermatids occur with widely differing quantities of mitochondrial substance, for it is evident that if such occurred the resulting spermatozoa would be likely to have middle-pieces of vastly differing sizes. The similarity of all spermatozoa of a given animal, in respect of the cytoplasmic apparatus, is a well-known fact.

It was, as I have mentioned, the discovery of a peculiar fragmentation process of the choanocyte granules which solved these difficulties.

The only two papers devoted entirely to the gametogenesis of Grantia are those of the late Professor A. Dendy and the present writer. Dendy (1) showed that during oogenesis young oocytes engulf entire cells, which are brought to them by special carrier cells. He also described the remarkable appearance of the young eggs such as is shown in fig. 18, Pl. 21, of the present paper; the nucleolus becomes squeezed out to form lumps, which are extruded into the cytoplasm, and form what Dendy considered to be chromidia. In his fig. 52, Pl. 24, he gives a drawing of what we now know to be fertilization. Dendy tacitly accepted the interpretation of Jorgensen (6), that this stage is the feeding of the oocyte by means of nurse-cells. Jorgensen’s descriptions of this phenomenon had been accepted by the German cytologists, but my paper in 1919 showed clearly that this stage was the last period of entry of the swollen spermatozoon into the oocyte. Jorgensen and Dendy missed an interesting discovery by under-differentiating their slides.

Both Jörgensen and Dendy described a process whereby a part of one of the pronuclei occasionally became separated and lay in the cytoplasm as an accessory body or ¹ karyomere ‘.

In my previous paper, read in December 1919 before the Linnean Society (2), I showed that the spermatozoon of the collar-cell does not penetrate directly into the egg, but passes into the collar-cell, in which it swells up considerably. In Grantia compressa it was shown that the collar-cell becomes considerably modified by the presence of the spermatozoon.. In my paper at that time I made some attempt to describe the cytoplasmic inclusions. The Golgi body was correctly identified, but I was considerably puzzled by the other granules in the cell, and could not give a satisfactory account of their nature and behaviour. The embryology of the amphiblastula of Grantia was described in detail. The description by Haeckel (4) and Dendy of the metamorphosis of collar-cells into eggs was upheld. At the meeting of the Linnean Society on that date both Professor Dendy and Dr. Bidder accepted my interpretation of the stages in fertilization and spermatogenesis described by me. It was admitted that Professor Dendy’s and Dr. Poléjaeff’s sperm stages were either plant parasites or inquilines.

In my previous paper I depended on the fixatives of Champy-Kull, Hermann, or my own modification of Flemming, and Kopsch’s original method. I have since tried Nassonow’s and Ludford’s methods without much success. This work is based on the Champy-Kull fluid and Flemming without acetic (F.W.A.), followed by Heidenhain’s haematoxylin. I am trying Ludford’s methods again, for by them a more delicate control of osmium dioxide precipitation is obtainable. I have found that even after prolonged heating in osmic solution and hot water it is very difficult to get the osmium dioxide to precipitate.

It has already been mentioned that my recently prepared slides showed clearly that the polymorphic granules already figured in Grantia compressa were mitochondrial in nature. This conclusion is based on the following facts. In two areas of the sponge I have noted that much activity of the inclusions occurs, namely on the outside where new choanocyte chambers are in process of formation, and in rarer cases towards the centre of the sponge substance. In fig. 11, Pl. 20, is drawn, at a fairly low magnification, a part of a flagellated chamber. Normal choanocytes are at NA, while at AT are collar-cells containing more granules than is usual, an intermediate zone being marked IA. Closer examination of the choanocytes at AT shows that many contain peculiar spheres made up of from twenty to fifty minute granules. In fig. 25, PL 21, one such cell is drawn, the sphere (SPH) being a remarkable object.

Now in fig. 19, Pl. 21, another type of flagellated cavity is shown. None of the flagellated cells has a collar, and on the basement side in every case is found a single sphere of minute granules, such as is drawn with a higher power in fig. 22, Pl. 21. In rare cases in this area other granules are found in the cell.

One can soon find areas of collared cells which show clearly how these remarkable spheres arise. In fig. 20, Pl. 21, is a normal choanocyte containing a few polymorphic granules; in Grantia practically all collar-cells will be found to be of this type. In fig. 26, Pl. 21, is another collarless cell from a flagellated cavity. It contains four granules, a, b, c, and another. Granules a and b are breaking up, the latter forming a number of minute granules: in fig. 24, Pl. 21, is another collarless choanocyte showing the same phenomenon. In fig. 21, a, b, and c, Pl. 21, are three stages in the fragmentation of the large granules of the collar-cells. Such stages are very clear in my material of Grantia, and I can hardly be mistaken about it.

It should be pointed out that, while none of the cells in the flagellated cavity in fig. 19, Pl. 21, had a collar, the changes in the granules which bring about the formation of the spheres in figs. 22 and 25, Pl. 21, often take place in cells with collars, as in fig. 25, Pl. 21. The appearance of the granules does not therefore herald the disappearance of the collar, except in the special eases to be described below.

All sponge eggs I have studied contained a fair number of chromophile granules. Sponge eggs, however, never contain so many granules as most other eggs. The chromophile granules of Grantia are often remarkable crenulated objects, as shown in fig. 8, EG, Pl. 19, and fig. 12, Pl. 20. In Sycon ciliatum such granules are shown in the oocyte in fig. 17, Pl. 21. Dendy and others have figured these granules, the first worker calling them ‘chromidia ‘.

The origin of these remarkable-‘nucleolar emissions ‘of Dendy is a matter of importance, because there is little room for doubt that these granules correspond to the mitochondria of collar-cells. Dendy quite accurately described the peculiar condition of the oogonia entering the oocyte stage, so far as concerns these granules, and since both Dendy and I considered that these bodies were nuclear in origin, it seemed interesting to re-examine the question whether a nuclear origin of mitochondria had at last been proved. Some new Sycon material has been helpful, in an unexpected manner, for I could not go any further in my investigations with the new sections of Grantia compressa.

In fig. 13, Pl. 21, an example of a Grantia oogonium, showing the shrunken nucleus and the apparent connexion be tween the latter and the remarkable ‘nuclear extrusions ‘at OG. In Grantia some such stage occurs in every oogonium entering growth. Both Dendy and I have concluded that large nucleolar lumps are thrown out of the nucleus at this stage. There is little doubt that these lumps later on constitute the well-known egg-granules of sponge ova. I conclude that these granules are mitochondria, because the granules within the choanocytes of the amphiblastulae are their lineal descendants, and give rise to all the granules in the choanocytes of the full-grown sponge (2). These choanocyte granules, as I have shown in this present paper, have the power of fragmenting and of becoming arranged in the choanocyte, in the form of a granulated sphere.

It may be concluded by the reader that it is very doubtful whether the nucleolus could form mitochondria. I may say I do not like the idea myself, but a study of Grantia at present leaves me with no choice but to suggest that this is what happens in this sponge.

Dendy, who called these extrusions ‘chromidia ‘, was the first to claim that they originated from nuclear emissions.

I have mentioned that I could not go any further in my investigations on Grantia compressa than had Dendy, and I have tried to examine another sponge.

Some of the stages in Sycon ciliatum are wonderfully clear. For instance in fig. 18, Pl. 21, there is a Sycon oocyte at the leptotene stage; in the cytoplasm are remarkable spherical objects which are well preserved and stained by F.W.A. and iron alum haematoxylin. Now, in none of my Sycon material have I found collapsed nuclei seemingly in the stage of nucleolar extrusion. On the contrary, these granules in Sycon have been quite easily traced back to a single juxta-nùclear structure, shown in fig. 1, Pl. 19. I have fpund many examples like fig. 1, with two, three, and more granules, according to the size of the oogonium. I would not care to suggest that these granules were nuclear in origin. Yet they are unquestionably the granules of the adult egg (fig. 17, Pl. 21), corresponding to the granules of the Grantia oocyte.

Are we then to conclude that the remarkable collapsed oogonial nuclei of Dendy and myself do not represent nuclear extrusion of some kind ? I am unable to answer this question at present, but I feel that in view of the clear results obtainable with Sycon there may be some doubt about it.

The various stages leading to fertilization, which were previously described, have been mentioned above. In some slides of Sycon several examples of spermatozoa within collar-cells were found. On fig. 5, Pl. 19, and fig. 10, Pl. 20, are two normal choanocytes which have been penetrated by spermatozoa. In fig. 5, Pl. 19, the spermatozoon lies across the collar-cell and the nuclear (SN) and middle-piece (MAM) parts of the spermatozoon are recognizable. The majority of cases resembled fig. 10, Pl. 21, where the spermatozoon is curled around the nucleus and inner part of the choanocyte, and only small areas of the spermatozoon are in focus at one time. Few of the examples which were found were as clear as fig. 10, Pl. 20.

Turning to Grantia we have such cases as are drawn in figs. 3, 4, 6, and 8, Pl. 19, which represent most of the stages missing from my previous paper.

Fig. 6, Pl. 19, probably represents the earliest example found in Grantia, though the specimen in fig. 3, Pl. 19, may have been inside the collar-cell longer. The interest of fig. 6, Pl. 19, lies in the fact that the large choanocyte nucleolus has begun to fragment, for in the majority of sperm nurse-cells formed from choanocytes, the single central nucleolus has broken down partly to form one large and several smaller irregular particles. The chromophility of the nucleoplasm had become less intense, and the whole cell had assumed a different appearance from that of its neighbours. The flagellum and collar were still present. The sperm itself possessed a long middle-piece, and the head had already expanded somewhat, without becoming reticulate. In fig. 3, Pl. 19, another specimen is shown, the future sperm nurse-cell being indistinguishable from its immediate neighbours. The staining of this example was more precise, and a crescent could be noted in form of the sperm head—the former being the acrosome, which becomes so clear later on.

A contrast to these already described examples is provided by fig. 8, Pl. 19, where the sperm nurse-cell had changed completely, having lost collar, flagellum, and assumed the reticulate nucleus. The position of the sperm nurse-cell with reference to the neighbouring choanocytes has altered, the immediately neighbouring collar-cells tending to cover the sperm nurse-cell. It is a peculiar fact that the latter should have changed so completely in this example, whereas spermatozoa in figs. 3 and 6, Pl. 19, are as ‘old ‘as those in fig. 8, Pl. 19, judging by the amount of modification of the sperm-head and middle-piece.

The next stage is shown in fig. 4, Pl. 19, where both parts of the spermatozoon have swollen up. This specimen is interesting because the sperm-tail could still be seen clearly. Only part of the sperm nurse-cell is shown in this section.

The next stage is shown in fig. 9, Pl. 20, where the entire egg and the sperm already penetrated are to be seen. This stage has been described in my previous paper in the Linnean Society’s ‘Journal ‘, but my drawings there suffered from being too crowded. The present drawing is made from much better sections. Various points of interest will be mentioned below.

The next stage is shown in fig. 12, Pl. 20. This remarkable specimen shows the sperm centrosomes, SON, the nucleus having already revolved; the middle-piece has beenleft behind at MAM. The subsequent history of the middle-piece was described in my previous paper, and will not be mentioned further. The sperm nurse-cell is still in situ at see.

With the exception of the stage of the unchanged ripe sperm within the collar-cell, which were not found in Grantia, but are described in Sycon, all stages of fertilization in Grantia have now been found.

It remains to allude to those changes in the sperm nurse-cell, which I have been able to follow more carefully in my latest materia].

In fig. 7, Pl. 19, is a specimen of a sperm nurse-cell containing a spermatozoon at the stage before the latter enters the oocyte. The egg lies towards the top of the plate, the sperm nurse-cell now lying in a concavity in the former. The parts of the spermatozoon are clearly visible, the acrosome (AC), the nucleus (SN), the middle-piece (MAM), and the nucleus of the sperm nurse-cell at see. The cytoplasmic inclusions of the sperm nurse-cell are the same as those in the collar-cell, the Golgi body (GA), the mitochondria (M) being shown, but they stain less well. It is this quality which causes the apparently clear appearance of the sperm nurse-cell when compared with normal collar-cells. Possibly the difficulty in staining the granules in many of the sperm nurse-cells is due to the temporary absence of some substance from their inclusions. During such times as cell division the cytoplasmic inclusions of many cells are found to stain faintly. After carefully examining many examples of enlarged spermatozoa within sperm nurse-cells, and in eggs, I am convinced that the Golgi elements themselves do not exist in the ripe Grantia spermatozoon. That Golgi apparatus derivative, the acrosome, is, of course, very distinct in such sperms (fig. 7, Pl. 19).

In my previous paper I gave a diagram showing the place of penetration of the spermatozoon into the egg: the majority of sperm nurse-cells were in the position drawn in fig. 9, Pl. 20, in the present paper, i. e. exactly opposite the middle of the oval oocyte. I found few cases in which penetration took place far from this region. In my recently prepared material of Grantia I found several examples where penetration took place on the opposite side, where the egg abutted against an open space, and not a flagellated cavity. Such an example is shown in fig. 2, Pl. 19. It is quite certain in this case that the cell containing the spermatozoon must either have wandered into the substance of the sponge, or that a spermatozoon must have penetrated right through the collar cavity into the underlying tissue, in which it found some amoebocyte to enter. The latter suggestion does not appeal to me as much as the former.

The spermatozoa in the collar-cells of Sycon drawn in fig. 5, Pl. 19, and fig. 10, Pl. 20, were not near oocytes, though all the larger oocytes in my Sycon material had spermatozoa within nurse-cells, in much the same position as in Grantia. In Sycon, however, the spermatozoa were often found much to one side of the oocyte, and I think it likely that these may have been carried some distance by the sperm nurse-cell. The reasons for this conclusion are that the majority of Sycon oocytes had sperm nurse-cells towards one side of the middle line as shown in fig. 17, Pl. 21, and that many collar-cells far from eggs contained spermatozoa. In Grantia, however,

I have never found sperms in such quantity away from the oocytes. In the hundreds of examples I have examined in Grantia the sperms, at all stages, are found towards the middle line as in fig. 9, Pl. 20. I feel certain that Sycon is différent in this respect. Other species of sponges are still being examined with reference to this interesting point.

The process of swelling of the sperm within the collar-cells is very interesting because it shows that the preliminary stages of male pronucleus formation can take place in cells which are not egg-cells. This has been known for many years since Jacques Loeb’s work on the swelling of cockerel sperms in egg-white, a piece of work which should be re-investigated by modern methods. In all other cases of fertilization I have read about, or personally investigated, a membrane around the swelling spermatozoon is never present. I refer to the membrane and space shown on fig. 7, Pl. 19, and disintegrating in fig. 9, Pl. 20, sw. This membrane is very clear in the late sperm nurse-cell stage, and I am not certain whether the membrane belongs to the sperm or is merely a limiting structure caused by the production of fluid, around, or by, the swelling spermatozoon. As shown in fig. 12, Pl. 20, the later stages of fertilization are normal, the space and membrane having disappeared. One must mention, nevertheless, that in the late sperm nurse-cell stage there appears to be definite formed membrane, which has been thrown into relief by the fact that fluid has come between it and the sperm it covered.

After examining a great deal more material I am satisfied that the sperm nurse-cell, after the expulsion of the sperm into the egg, merely relapses into a choanocyte form. This was the impression I had already gained during my previous investigation.

Up to the present time the only observer who has seen all stages in spermatogenesis is the German worker Gôrich (3), who described them inSpongilla. I have collected and sectioned much Spongilla material in both Oxford and Dublin: I have never succeeded in finding any sperm stages. In my previous paper I described some structures which are unquestionably Grantia spermatids. They coincide with what Gbrich described in Spongi 11 a.

During the years which have elapsed since my last paper was read I have tried to find other stages of spermatogenesis in Grantia, Spongilla, Halichondria, and Sycon.

There appears to be some mystery about sponge spermatogenesis, because every breeding sponge among the scores I have examined contained fertilized eggs, while only one contained one stage of spermatogenesis. In such cases I can think of only one explanation of this difficulty—it is that spermatogenesis must take place sporadically and rapidly.

Some time before his death the late Professor Dendy kindly allowed me to examine all his sections of sponges which might have contained sperms. Only one specimen of Euspongia showed some stages which might have been nests of spermatids. Except for Gorich’s paper, which is cytologically imperfect, the questions surrounding the appearance of sponge spermatozoa are still largely unanswered.

In the material used in this investigation I found stages which I can only interpret as spermatogenesis. They are, however, something quite different from what I described in Grantia before, but I think that it is this very difference which is the key to our difficulties in finding Grantia and Sycon specimens showing spermatogenesis.

In my previous paper the sperm stages were in definite nests or pockets lined by large cells. These were definitely not plant or animal parasites, and resembled the spermatids of other animals. Only the spermatid stage was discovered.

In the present paper the stages I have described as spermatocytes are derived directly from collar-cells, and are in fact metamorphosed collar-cavities. The previously described cellnests lay beneath the collar-cavities and were almost certainly derived from cells of the sponge matrix. Such bodies could undoubtedly be called testes. The new areas to be described are, as I have said, something quite different, and I believe represent the rapid and sporadic production of spermatozoa from collarcells.

In fig. 19, Pl. 21, is a flagellated cavity in which every cell is like that in fig. 22, Pl. 21, that is, flagellated without collar containing a reticulate nucleus exactly like that of spermatocytes in other animals, and provided with a remarkable sphere (SPH) of equal-sized granules situated at the base of the cell. Now these pockets of modified collar-cells can be found in all stages of formation from collar-cavities. Their origin cannot be in doubt, even if their nature may be.

It may be suggested that they represent preparatory stages in division of the choanocytes. This view does not appeal to me at all, because I know not only from Minchin’s work but from my own observations, that cell division in collar-cells is not a process which all cells in one area undergo synchronously, but is sporadic and affects single cells hero and there. It is obvious that fig. 19, Pl. 21, represents something quite different. Collarcells, too, undergoing division do not appear to shed their collars. In a few cases I found cells in the stage drawn in fig. 23, Pl. 21, which would almost certainly be described as a prophase of the heterotypic division. It should be mentioned that the appearance of the spheres within the cells in fig. 19, Pl. 21, may be taken as some evidence that these cells are spermatocytes, but it cannot be looked upon as conclusive in any way, for such spheres are found in undoubted collar-cells (fig. 25, Pl. 21) and exist commonly in regions where the sponge is growing rapidly.

Until such time as more material can be procured I feel justified in suggesting that sponge spermatogenesis may occur in two ways—in testis-like areas in the sub-epithelial matrix of the sponge, and sporadically from collar-cells, according to the process shown in figs. 22,’ 23, 24, and 26, Pl. 21.

The evidence brought forward in this paper shows that, far from being inert yolky masses, the granules in the cells of the sponge Grantia are capable of auto-fragmentation and presumably of growth by some form of accretion. Binary fission and fragmentation of mitochondria and Golgi elements have been described now in a large number of animals from different Orders.

The fragmentation of the granules of sponges takes place actively in areas of growth, and is associated with active cell divisions. The larger granules within the collar-cells fragment individually, and all the smaller granules so produced gather together to form remarkable spheres. These spheres are found either towards the flagellum, or towards the basement membrane. Specially modified areas of collar-cells have been described, in which all the cells have lost their collars, their nuclei have become reticulate, and the formation of the peculiar spheres lying at the basement side has taken place.

It has been suggested that in sponges spermatozoa may be formed from these areas by a rapid process of direct metamorphosis of collar-cell into spermatocyte, as was believed to be the case with the oogenesis, by Haeckel, Dendy, and the present writer.

The early stages of fertilization in the sponge Grantia have been described, and the fully ripe spermatozoa of Sycon within collar-cells figured. It has been shown that long before the spermatozoon within the collar sperm nurse-cell has swollen up to form the characteristic oval body which later passes into the oocyte, the sperm nurse-cell has been formed from the collar-cell (fig. 8, Pl. 19).

In a number of examples of very slightly changed spermatozoa, the degree of metamorphosis of collar-cell into sperm nurse-cell varies within wide limits. In no case found in Grantia does the collar-cell entered by a sperm remain unaltered, though in Sycon the degree of alteration is never quite so remarkable.

Two or three cases in Grantia and Sycon have been found which indicated that the sperm nurse-cell may have travelled some distance with its burden in order to find an egg. Such examples are very rare in Grantia, but may occur more commonly in Sycon.

AC, acrosome; AT, area of transformation; CAV, cavity of flagellated chamber; CH, choanocyte; CNT, space in connective tissue; CN, centrosome; COL, collar of choanocyte; E, embryo; EG, egg granule (mitochondrion); GA, Golgi element; GAX, supposed accessory Golgi element; ax, supposed mitochondrial body; IA, intermediate choanocyte area; MAM, mitochondrial middle-piece of sperm; N, nucleus; NA, normal choanocyte area; NCH, nucleus of choanocyte; ON, oocyte nucleus; ONE, oocyte nucleolus; OTE, oocyte; s, sperm; see, sperm-carrying cell; SON, sperm centrosomes; SF, sperm-tail; SN, sperm nucleus; SPH, mitochondrial sphere; sw, wall between sperm and egg cytoplasm; sx, supposed sperm.

Fig. 1.—Small oogonium of Sycon compressum. Later stage in fig. 18, Pl. 21.

Fig. 2.—Fertilization of oocyte in Grantia on side away from choanocyte chamber.

Fig. 3.—Sperm beginning to undergo changes within choanocyte. Grantia.

Fig. 4.—Later stage. Grantia.

Fig. 5.—Ripe unchanged sperm within collar-cell of Sycon compressum.

Fig. 6.—Ditto, sperm more changed. Sycon.

Fig. 7.—Sperm of Grantia, fully changed within collar (sperm nurse-) cell. “

Fig. 8.—Stage just after fig. 3. Collar-cell nucleus much altered. Collar and flagellum lost.

Fig. 9.—Typical stage in fertilization of Grantia showing spermatozoon at stage of fig. 7, Pl. 19, entering oocyte.

Fig. 10.—Sperm in collar-cell of Sycon. See also figs. 5 and 6, Pl. 19.

Fig. 11.—Part of choanocyte chamber showing changes in mitochondrial granules. Grantia.

Fig. 12.—Stage of fertilization in Grantia showing breaking up of sperm and formation of male pronucleus. Centrosomes present.

Fig. 13.—Oogonium of Grantia showing nucleolar extrusions of Dendy.

Fig. 14.—Golgi bodies in growing egg of Gr anti a.

Fig. 15.—Early oogonium of Grantia with flagellum.

Fig. 16.—Grantia choanocyte with included sperm (?).

Fig. 17.—Fertilization in Sycon—stage corresponding to fig. 9, Pl. 20.

Fig. 18.—Oocyte of Sycon at leptotene. Remarkable cytoplasmic granules.

Fig. 19.—Metamorphosed choanocyte cavity of Grantia. Enlarged cell drawn in fig. 22.

Fig. 20.—Normal choanocyte of Grantia.

Fig. 21.—Stages in fragmentation of choanocyte granules.

Fig. 22.—Supposed spermatocyte formed from choanocyte of Grantia.

Fig. 23.—Ditto, pachytene stage.

Figs. 24 and 26.—Formation of sphere shown in fig. 22.

Fig. 25.—Choanocyte with sphere.