Talk:Book - Contributions to Embryology Carnegie Institution No.28

From Embryology

By J. DUESBERG, Adjunct Professor of Anatomy, Faculty of Medicine, University of Liege, Belgium, Research Associate, Dcparttnent of Embryology, Carnegie Institution of Washington.

With two plates and five text-figures.



Material 49

Technique 50

Short survey of the process of spermatogenesis 51

First period 55

Second period 56

Third period 59

Fourth period 60

Chondriosomes 62

Sertoli cells 63

Spermatogonia 63

Spermatocytes 63

Spermiogenesis 64

First period 64

Second period 65

Third period 65

Fourth period 67

Discussion 69

Tlie apparatus of Golgi 7^

Di-icussion of apparatus of Golgi 75

IJihliography 82

Explanation of figures 84





The object of the present investigation was the study of the chondriosomes and of Golgi's apparatus in the seminal epitheUum of the opossum (Didelphys virginicma), the material consisting of nine animals. Five of these were full-grown; a sixth exhibited all stages of spermiogenesis but the verj^ last — i. e., the formation of the spiral filament at the expense of the chondriosomes; no spermatozoa were found in the epididj-mis. The three remaining animals were in a much less advanced stage of development. Whether spermatogenesis in the opossum is taking place throughout the year, or whether the testicle becomes active only before the period of copulation, could not be ascertained, as all the material was procured between August and January. It should be stated, however, that the animals which showed only early stages of spermatogenesis were distinctly smaller than those which had spermatozoa and were probably under one year old.

In the adult opossum, at least at the time of year when these animals were collected, the testicle exhibits all the characteristic featvtres of a typical mammalian testicle, particularly that striking regularitj' in the evolution of the germ-cells which was first revealed in the rat. The images are, however, not exactly super- posable, as the duration of the different stages is not the same in both species. I found, for instance, that in the opossum the migration of the ring (which in the rat takes place at the time the cells of the subjacent layer have reached the end of the period of maturation) occurs somewhat later, as the period of maturation is over and the spermatids have already differentiated as far as those represented in figures 17 or 32.

I was impressed by the large amount of fat — that is, such fat as blackens with osmic acid — usually present in the seminal epithehum. Fat appears in the seminal cells at the end of the growth period and a number of small fat droplets are con- stantly found in the dividing spermatocjiies (fig. 6) and in the young spermatids (fig. 7). It then disappears very quickly and none is to be found after that stage. The bulk of the fat present in the seminal epitheUum is contained in the Sertoli cells, and it is here that variations may occur. This fat is apparently a product of the degeneration of the c}i;oplasm (Regaud's "corps residuel") cast off by the spermatid at the end of its evolution.




Fragments of testicle were fixed in the following fluids: Hermann's, acetic sublimate, Ramon y Cajal's mixture of formalin and uranium nitrate, Bouin's, Flemming's, saturated sublimate, Altmann's, Meves's, Benda's, and Regaud's. Fragments of the ejndidymis were fixed in the latter two reagents. Smears of spermatozoa wore fixed either by the action of vapors of 1 per cent osmic acid in a moist chamlier for 20 to 30 minutes, or by immersion for 10 to 30 minutes in Regaud's mixture, then for 24 hours in 3 jier cent bichromate. Two adults were sacrificed for the purpose of studying the living cells, and were injected with a solution of janus-green (nro (tt)) i" 0.85 per cent salt solution. I am greatly indebted to Professor E. V. Cowdry for his help in carrying out these experiments.

Of the above-mentioned reagents Benda's, Meves's, Regaud's and Altmann's fluids were used for the purpose of fixing the chondriosomes. Perhaps the best preparations were obtained from Regaud's material. Regaud's fluid, however, is an exceedingly- one-sided reagent. While it often fixes the chondriosomes perfectlj^ well, other structures, such as centrioles, idiozome, axial filament, or "tingierbare Korner" are scarcely visible, if at all; the cell limits disappear and one no longer wonders at Regaud's peculiar conception, according to which the spermatogonia have no cell-bodies of their own (1908). As to the nucleus, there is no possibility of studying it during the growth or maturation period. In the later stages, however, especially with a good nuclear stain, such as methyl green, the same material proves very valuable. The worst feature in the action of Regaud's fluid is the dislocation it produces in the seminal epithelium, a dislocation due (in part at least) to the fact that the fat is not fixed and is consequently dissolved during subsequent manipulations.

A number of nuclear stains and (for the stud}^ of the chondriosomes) iron hematoxylin, Benda's and Altmann's method, and acid fuchsin-methyl green were used. Excellent preparations were obtained with the latter method, especially after fixation in Regaud's fluid. As it happened that my material would keep too much methyl green, I found it very useful to bring back the slides from the 95 per cent alcohol into distilled water, a procedure which eUminated immediately the superfluous green and improved the preparations greatly, then back to 95 per cent alcohol, and so forth. In order to bring into view the apparatus of Oolgi, I used, with the same .success as previously (1914), Ramon y Cajal's mixture of formalin and uranium nitrate. The best preparations were obtained bj' leaving the i)ieces in the fixative for 9 hours, in the silver nitrate 37 hours, and in the developer 14 hours. While undoubtedly this method almost unfailingly impregnates the appara- tus of Golgi, it is nevertheless a capricious one, inasmuch as it is hable to bring into evidence at the same time granulations, most probably mitochondria (fig. 26), fibrils in the connective tissue, cell limits, and eventually other structures difficult to identify.

Instead of other comjjlicated procedures, at the suggestion of Professor Cowdry I used the following method in the treatment of the sections: The slides were first


immersed on 0.1 percent gold chloride for 2 to 3 hours; then in hyposulphite for 20 to 30 minutes. The results were excellent. Probably any nuclear dye could be used as counter stain. I resorted to Ehrhch's hemato.xylin, to safranin, and to methyl green (Cowdry, 1916), which in the concentration of 0.5 per cent, apphed for 30 to 60 seconds, gives a very sharp constrast with the black color of the appara- tus and the pale background.


An accurate and profitable study of the chondriosomes and of Golgi's apparatus in the testicle can not be made without an intimate knowledge of the whole process of spermatogenesis, and especially of its last phase— spermiogenesis. In this con- nection von Korff's researches (1902) on the spermiogenesis of another marsupial {Phalangista vulpina) proved to be of great help. The process in Phalangida is so similar to that in Didelphys that it appears necessary to give a summary of von Korff's paper.

Since the work of Meves on the spermiogenesis of the guinea-pig, this phase in the evolution of the seminal cells has been usually divided into four periods: The first period extends to the formation of the so-caUed "Schwanzmanschette," which, following Oliver's example (1913), I shall call the caudal tube; the second ends with the disappearance of the same formation; the third extends to the time of the expulsion of the spermatozoa into the lumen of the seminiferous tubules and of the elimination of the major pan of the cytoplasm; the fourth period includes such changes as may take place subsequently — changes which, in many cases, are of minor importance. In Phalangista von Korff found that, as the caudal tube appears and disappears suddenly, it is better to use as a basis for the subdivision of spermiogenesis the modifications of the centrioles which coincide with the appari- tion and disparition of the caudal tube; i. e., their close relationship with the nucleus at the end of the first period and the beginning migration of the centriolar ring which marks the end of the second period. It appeared also expedient to sub- divide the second period according to the modifications of the nucleus: First, the spherical nucleus is transformed into an egg-shaped body whose long axis is per- pendicular to the axial filament; later it assumes its definite shape.

The young spermatid of Phalangista contains two granular centrioles, located first at the periphery of the cell, later moving toward one pole of the nucleus, while the idiozome moves toward the other pole. The distal centriole carries a thin filament and flattens out somewhat at the end of the first period. The modifica- tions of the idiozome are less complicated in Plmlangista than in other mammals. Instead of the numerous granules, each located in a vacuole as noted in other species (the guinea-pig for instance), a single vacuole without a granule is formed. This attains a considerable size and applies itself to the nucleus, while the remainder of the idiozome is eliminated. "\'on Korff mentions also a chromatoid body which later falls to pieces and is probably cast off at the end of the process.

The beginning of the second period is marked by the sudden appearance of the caudal tube. The nucleus flattens out and later assumes its most characteristic


shape. The proximal centriolc is inserted on the nucleus and very soon it becomes apparent that this insertion is not directly in the center of the nucleus, but toward one extremity {"latcmlc Insertion"). The distal centriole assumes the form of a disc, the middle portion of which, carrying the axial filament, later breaks off, leaving a ring and a small granule {distaler Cenlralkorperknopf). When the head has assumed its definite shape one finds this granule connected with the proximal centriole by a thin filament. Both are located in the notch of the head. The axial filament is very thin in the region of the future middle piece, while posteriorly it suddenly becomes much thicker. This part, which corresponds to the future main piece, exliibits a peculiar structure, a cross striation. The protoplasmic body has elongated and (contrary to what is found in many other mammals) extends farther back than the middle piece.

With the sudden disappearance of the caudal tube the third period begins. Most remarkable is the fate of the headcap, which is left by the spermatozoon in the protoplasm of the Sertoli cell. The centriolar ring migrates to the end of the middle piece, while the pro.ximal centriole breaks into two granules connected by a thin filament with each other and with the granular part of the distal centriole. The chondriosomes dispose themselves around the axial filament. Finally, most of the protoplasm, with the so-called von Ebner's "tingierbare Korner," is eliminated.

A striking change occurs during the fourth period, a change in the position of the head in relation to the tail. At first the axis of the head was at right angles with that of the tail, but in spermatozoa collected in the epididymis the head is found in the same long axis with the tail.

From this short summary it appears that the process is similar to what is observed in other mammals, the modifications of the centrioles, compared with those which occur in the guinea-pig, belonging to a rather simple type. The most striking feature of the process is the ehmination of the headcap from the spermatozoon.^

While it appeared necessary to give a summarj^ of von Korff's paper on account of the marked resemblance between his description and my own findings, and because of the close relationship between the species investigated, it is not my intention to review again the whole hterature relating to spermiogenesis in mam- mals. I refer the reader to my paper of 1908, and will hmit myself to an account of two papers which have appeared since and which are concerned with the process of spermiogenesis, or certain phases of it in mammals — one by Oliver (1913) on spermiogenesis in the fur seal, the other by Stockard and Papanicolaou (191G) on the modifications of the idiozome in the guinea-pig.-

Olivcr's conclusions agree in the main with those of Meves (1899), von Korff (1902), and the author (1908). There is, however, one point of difference. While Oliver admits, rightly in my opinion, the formation of the caudal tube at the expense of cytoplasmic material — here, as in the guinea-pig (Meves, 1899) its formation from filaments is quite apparent — in opposition to Meves, myself, and a

'A pnpcr by Bendn (19011) on ar>eriniUoKenesis in marsupials was not available in this country. I fan therefore only quote from my own former references to it (I'JOS. 1912).

' E. Allcn'H "Studies on cell-division in the albino rat III" (Journ. of Morph., vol. 31, 1918), appeared too late for diHruKKion here, but I should like to warn the reader that practically wherever I am quoted in that paper such quotations arc inaccurate.


number of other authors, he believes that the caudal tube eventually intervenes in the constitution of the middle piece. Since in the guinea-pig, the degeneration of the caudal tube is obvious, Oliver's description would lead one to b3lieve that the same structure can undergo a different fate in different mammals. I must say, however, that I am not convinced. OUver describes the process as follows: The caudal tube breaks off the nucleus just before the centriolar ring begins to migrate, then becomes increasingly narrower, surrounding more and more closely the axial filament. At the same time "the cell membrane covering the caudal tube has fused indistinguishably with the latter." The question which immediately suggests itself to me, but which, if Oliver's description be correct, I am unable to answer, is, where are the chondriosomes? These bodies are never, to my knowledge, found within the caudal tube, nor can they be located between the caudal tube and the cell boundaries, according to the quoted description. I would add also that the separation of the caudal tube from the nucleus appears to me, from Oliver's draw- ings, to be an accidental rather than a normal occurrence; and that between figure 32, in which the caudal tube is still far from being fused with the a.xial filament, and the next figure there is an important, unfilled gap.

Stockard and Papanicolaou have given a description of the fate of the idiozome (they spell it idiosome, after Regaud's proposal, 1910) in the guinea-pig. The present summary- is made from their communication at the meeting of the Anatom- ical Society (1916) and from the abstract published in the bibUographic cards issued by the Wistar Institute (No. 155). The idiosome of the primary spermatocytes consists of two spheres, one inclosed within the other. During the process of division the outer sphere (idioectosome) breaks into irregular pieces, while the inner one (idioendosome) forms a great number of granules (idiogranulomes). The pieces of the idioectosome and the idiogranulomes flow together in each of the secondarj^ spermatocytes and form a new idiosome consisting of a spherical idioec- tosome which contains a number of idiogranulomes. During the division of the secondary spermatocytes the idioectosome breaks up again, and its pieces are dis- persed in the protoplasm with the idiogranulomes. This process permits a uniform distribution of the very important idiosomatic material during the division. In the spermatids a new idiosome is reorganized, haAdng a new idioectosomatic sphere inclosing idiogranulomes, each granulome being surrounded by a small vacuole (idiogranulotheca) . The idiogranulomes and the idiogranulotheca flow together to form a large spherical body (idiosphserosome) inclosed mthin a large vacuole (idiosphserotheca) . The idiosphaerosome differentiates into an "upper" cap— the idiocalyptrosome, and a "lower" body — the idiocryptosome; "the idiosphaerosome secreting as soon as formed on its surface, furthest from the nucleus, a new sub- stance — the idiocalyptrosome" (1916). The idioectosome (called also idiophthar- tosome) is ehminated with the protoplasmic remains. The idiocr\iitosome and the idiocaljT^trosome persist in the spermatozoon as two caps, one beneath the other and inclosed wdthin the spermiocalyptra or idiocalyptrotheca, which is the trans- formed idiosphaerotheca.


The new and interesting part of this paper is the descriptif)ii of the behavior of the idioondosonie and its idiogranulomcs (or better, perhaps, idiogranulosomes) during mitosis, altliough it must be stated that the jiersistence of the idiogranu- lomes in the dividing spermatocyte has alreadj^ been reported by Niessing (1902), who represents them (fig. 12) in the prophase of the second division. Nothing was known, however, of the behavior of these bodies in later stages; that is to say, of their repartition between the daughter-cells. As to the constituents of the idiozome in the first spermatocytes, Niessing (1897) had already described it in the same species as formed by two layers. Numerous other authors have recognized the same structure in other species and applied to it various names. Stockard and Papanicolaou's idioectosome is nothing but Platner's Nebenkernstdbchen, Heiden- hain's Centralkapseln, Ballowitz's Centrophormien, Perroncito's, Terni's formazioni periidiozomiche, etc. This point will be discussed later. A process similar to the behavior of the idioectosome during mitosis has been described by Platner and called by Perroncito dyctiocinesis. Little is added to what we already knew of the fate of the idiozome during spermiogenesis. The idiogranulomes (Moore's archosomes, 1894), their idiogranulotheca (Moore's archoplasmic vesi- cles), the fusion of these bodies into one single vacuole and one single granule (w-hich is the acrosome or Spitzenkorper) , the presence within the acrosome of another granule, their relations with the nucleus, and the fate of the vacuole which ulti- mately forms the "Kopfkappe" or head cap (Stockard and Papanicolaou's spermio- calj^ptra or idiocalyptrotheca), have been described in detail for the guinea-pig by Meves (1899). Moore and Walker (1906) also have described two parts in the acrosome, and call the outer part "intermediate substance." None of these authors, however, have been able to distinguish these parts after the first period, but Niessing (1897) was able to follow them until advanced stages. They corre- spond most probably to the different zones w^hich I have represented (1910) in figures 62, 67, and 68.

A priori, it appears that the nomenclature of Stockard and Papanicolaou is by far too comphcatcd to be accepted. From the preceding considerations one may further conclude that it is also unnecessary since most of the things thus designated have been described before, and even prejudicial since some of them (the head cap and acrosome for instance) already have names to whicli but slight objection can be raised and which are generally accepted.

Furthermore, the description of these authors is inaccurate in several respects: (1) They describe the appearance of the granules only at the jirophase of the first divi.sion, whereas these exist prior to that time (see Meves, 1899, fig. 2, and perhaps also Niessing, 1897). (2) Similarly, the vacuoles, which they find only in the sper- matids, are already present in the second spermatocytes, according to ^Sieves (1899, fig. 3) and to Moore and Walker (1906, figs. 27 and 28), and indeed the last-named authors describe them in the first prophase (fig. 21). (3) The fragmentation of the idiozome does not necessarily take place before the first metaphase, as I found in 1910 (fig. 54), and again recently in preparations of the testicle of the guinea-pig with Cajal's method. (4) The outer substance of the acrosome can not hv consid-


ered as a secretion of this last body "on its surface furthest from the nucleus," as the two parts are already differentiated in the perfectly spherical acrosome before its connection with the nucleus. I refer to Meves's figure 9, and to Moore and Walker's figures 39, 40, and 41, and would call attention to the fact that Meves insists (pages 344 and 389) that this is not an appearance due to extracting the stain more or less, as both parts can be distinguished in unstained preparations. Finally, concerning the vacuolar structure of the outer part of the acrosome, I consider the point of Uttle importance, as its appearance varies considerably according to the fixing reagents and type of stain used. (See for example figs. 61-68, Duesberg 1910.) I come now to my own observations on the opossum and hasten to declare that I have Uttle to say here in regard to the spermatogonia and the sperma- tocji;es. Their nuclei I did not study, and their cytoplasmic constituents mil be described in the following chapters. Concerning the first spermatocji^es I would only mention that again, as in the rat, guinea-pig, and cat (Duesberg, 1910, p. 66), the growth period can be divided into two phases :

"Dont Tune va jusqu'a la formation des grosses travees chromatiques, tandis que I'autre est posterieure a ce stade. Au cours de la premiere phase, le spermatocyte de premier ordre est tres petit et s'accroit relativement peu. Pendant la seconde, protoplasme et noyau augmentent fortement de volmne, I'augmentation de volume du noyau etant par- ticulierement remarquable chez le rat. II y a done ici quelque chose de comparable a ce qui se passe au cours de la periode d'accroissement de la cellule sexuelle femelle, la seconde phase correspondant a la periode dite du grand accroissement (Gregoire, 1908) de I'ovule."

Mention should be made also of the idiozome as it appears after the action of reagents containing osmic acid. This bodj- consists of two parts, an outer, darkly- staining shell, and a hghter medullar substance. In the spermatogonia and in the young spermatocytes it is usually flattened against the nucleus, and in prepara- tions in which the chondriosomes are preserved it is often entirely covered by these bodies (figs. 2 and 4). In the older spermatocji;es the idiozome rounds out and then it is clearh^ apparent that the outer shell is missing opposite the nucleus (fig. 5). In the dividing spermatocytes it can be occasionally recognized as late as the metaphase, after which it falls to pieces. What becomes of these pieces is difficult to determine without the application of special methods. (See chapter on apparatus of Golgi) .

Concerning spermiogenesis I would say that no attempt was made to describe the process in everj^ detail. The study was undertaken in order to build bii a safe foundation mj^ researches on the chondriosomes and the apparatus of Golgi. I would also add that I was handicapped, especially in the studj- of the centrioles, by the lack of good ii-on-hematoxj'Un preparations. Enough could be elucidated, however, to make it appear worth pubUsliing. As to the subdivision of the process, my experience M^as similar to that of von Korff, and like him I have to adopt the modifications of the centrioles as a basis.


The young spermatid (fig. 7) contains, in addition to the chondriosomes with which I shall deal later, a spherical nucleus, an idiozome, and two granular centrioles


located at the periphery of the cell and perpendicular to its surface, with a thin filament in connection with the distal centriole.^ No chromatoid body was observed. A number of fat droplets are invariably found in one heap quite near the periphery of the spermatid; these verj'^ soon disappear (fig. 8). Their constant location at the periphery, as well as a special lighter aspect of the protoplasm in that region, sug- gests the possibiUty of an elimination in bulk. This process was not observed, however, and consequently one may just as well admit that they are simply digested. After a short period of growth the nucleus gradually diminishes in size while the condensation of the chromatin is taking place. There is no change in the form of the nucleus, but there is a change in its location. At the end of the first period it has moved to the periphery of the cell opposite the centrioles (fig. 9).

The idiozome is composed of the same constituent parts as described for the spermatocytes. Then, as in Phalangista (von Korff) a single vacuole appears within the medullar part (fig. 8) ; this increases to rather considerable dimensions, and coming in contact with the nucleus apparently exerts some pressure upon the nuclear membrane, as the latter shows a marked degree of flattening (figs, a, 8 and 17). The nucleus, however, resumes its spherical shape as the zone of contact with the vacuole becomes larger. Meanwhile, the cortical part of the idiozome has been detached; we find it at the end of the first period loose in the protoplasmic lobe. The head cap now extends over the anterior third of the nucleus, or thereabouts, and a small granule, a sort of acrosome, can be seen, although no trace of it was \dsil)le in the earlier stages (fig. 9). The centrioles, whose position, from the very beginning of the process, determines what it is customar}^ to call the posterior pole of the spermatid, migrate toward the nucleus, and consequently the axial filament increases in length. The proximal centriole comes in close contact with the nuclear 1; — . F'"- a.— spermatid in

... T/T- 1 1 t,^ — X\ ^ period. Zeiss

membrane and thus IS often very difficult to detect. [( » y\ apochr. imm. 2mm., oc.

mi Til i • 1 1 -1 -i J.1 11 I \ • •/^~A \ 12. Fixation and stain:

Ihe distal centriole exhibits then a small excre- Vlx J ) Benda. Thechondno- scence of granular form but somewhat elongated \. / noT been'^drawn'^"^^

in the direction of the mother-centriole (fig. a).

In order to make the description more precise, I might add that under the spermatids going through the process noted above one finds two discontinuous layers; one of spermatogonia, the other of first spermatocytes. The latter exhibit all stages of the first phase of the growth period and the beginning of the second. Above is a layer of spermatids in all stages from the end of the second period (fig. 12) to the elimination of the spermatozoa.


Here, as in Phalangista, the modifications of the nucleus would allow a sub- division of this period. First, there is a gradual condensation of the chromatin, until all structure disappears; at the same time the nucleus flattens out and assumes the shape of an egg whose long axis is perpendicular to the axial filament (figs, c, 10, and 33). Then other complicated changes occur which finally bring about a

'This filament is seen beating actively in the living cell.


peculiar form of the nucleus, roughly comparable to that of a swallow's tail. The process has been described by von Korff (pages 254-255), and is further illustrated to a certain extent by mj^ figures, so that it is not necessary to dwell upon it.

During the first phase of the nuclear modifications the rest of the idiozome is still found as a single, soUd body in the x_^^ /P^ F.os.Bandc.-Twosper-

protoplasmic lobe (figs, b, c, and 10). A \ lid] "j^V^a in the first phase

' ' ...' /\cOl /( ^ — / ** second penod.

Afterwards it disintegrates into particles 11/ / 1 / ^"^ apochr. imm. 2

r IP 1 mi r , • \ \ / 11/ mm., oc. 12. FUation

lormed oi granules, ihese iragments, in \ \./ \ \ / and stain: Benda. The preparations which show the chondrio- |o L-jLv p!ln>°o^iyno"^nrep!^

somes, are usually more or less covered by ^^ v_^ resented.

the latter bodies and are therefore difficult to detect. One of them, however, is generally conspicuous on account of its location. It is found near the opening of the caudal tube, sometimes even \\ithin it, and remains there throughout this period (figs. 11 and 12). The headcap now covers the anterior half of the nucleus (fig. b), and the acrosome, already referred to as appearing at the end of the first period (fig. 9), has become quite conspicuous. With the gradual flattening out of the nucleus the headcap becomes more and more closely attached to it and can be recognized only when the nucleus has shrunken somewhat (in figure c the nucleus exhibits a sUght amount of shrinkage); other\Adse only a slight thickening of the anterior edge of the head is noticeable. Later, with the beginning of the second phase of the nuclear modifications, a pecuUar appearance is observed. From the pointed extremitj^ of the head emerges a process w'hich projects as a sort of spur into the Sertoli cell (fig. 11).

In a still later stage (fig. 12) this structure has extended toward the other extremity of the head; thus the anterior edge of the nucleus appears, so to speak, duplicated. The two parts are connected at their extremities by an oblique fine, and between by two rather thick trails of a sharplj' stainable substance. This condition is plainly \asible after all fixations except inat of Regaud. Comparing figures 11 and 12, one gets the impression that something has become disjointed from the nucleus but still adheres to it by means of this stainable material. The two trails persist for a long time in the protoplasm of the SertoU cell, so that one may infer that the structure has not been put back into its original position. What becomes of it I could not ascertain. It seems probable, however, that here we have to deal with the ehmination of the headcap, a process which von Korff was able to observe more clearly in Phalangista. The disappearance of the structure described would then be due to its digestion by the SertoU cell.

We left the centrioles at the end of the first period as three granules: the proxi- mal centriole, the distal centriole, and its process. While I feel justified in assertmg that the further modifications of these granules are very similar to those described for other mammals, and belong to a tj'pe in many respects very close to those found in Phalangista, yet the process could not be followed in any great detail. The lack of good u-on hematoxyUn has abeady been mentioned. Three other factors also intervene to make the study difficult: (1) The eventual presence within the caudal tube (and in close proximity to the centrioles) of a fragment of the idiozome; (2)


the close contact of the jjioximal ccntriole witli the nucleus and later the location of the centrioles between the branches of the nucleus; (3) the excrescence of the distal centriole, which grows so large that it can cover the whole centriolar appara- tus. So much, however, could be seen. Very soon it appears that the centrioles are no longer in relation with the middle of the posterior side of the nucleus (von Korfif's "laterale Insertion," fig. b); then the distal centriole flattens out and finally breaks in two, a distal ring and a proximal granule which carries the axial filament (fig. c). Meanwhile, the centriolar excrescence described above undergoes con- siderable growth and becomes a spherical granule close to the other centriolar frag- ments (figs. B, c, and 10). When the head begins to assume its definite shape this arge granule is found always located toward its branched extremity. At the end of the second period — that is, when the ring is about to begin its migration — the same body becomes elongated and pear-shaped and its thickest part shows a differentiation in the form of a small, brilliant granule (fig. 12). At the same stage the proximal ccntriole and the anterior fragment of the distal centriole are more widelj- separated and connected by a thin filament. This arrangement is very similar to that found at the same period in Phalungista and represented bj' von Korff in figures 17a, 18a, and 19 of his paper, except for the presence of the large granule. Something similar to this latter body is found in other mammals, for instance in the guinea-pig. Meves describes in that species an excrescence of the anterior fragment of the distal centriole which grows to a considerable size. Later it breaks off, but is connected with one of the fragments of the proximal centriole by a thin filament. Whether any such connection with anj^ other part of the centriolar complex exists in the opossum I have been unable to ascertain. A cen- triolar process, formed by an anterior fragment of the distal centriole, but smaller than in the opossum and appearing onlj^ after the migration of the ring, is described by Oliver.

The axial filament is at first uniformly thin (fig. 11 and earlier). Toward the end of this i)eriod a differentiation takes place, inasmuch as on the posterior part a deposit of a somewhat Ughter substance appears (fig. 12). Thus, the future main piece can be distinguished from the middle piece. As in Phalangista, it is clear from that time on that the protoplasm extends farther back than the middle piece, a condition which differs from what is observed in other mammals, such as the guinea-i)ig, rat, etc.

The second i)eriod is characterized by the sudden appearance and disappear- ance of the caudal tube. While very little can be said in this case about its differ- entiation and regression, it should be stated that no indication was found of its nuclear origin, a view which has found some supporters but which, in the presence of Meves's observations and more recently those of Oliver, can be safely discarded. Nor was there any indication of its particij)ation in the constitution of the middle piece, as Olivc^r supposes. In fact, the peculiar shape of the head and the great width of the caudal tube make such a participation hard!)' imaginal)le.

During this period of spermiogenesis the cells of the subjacent layer are found to be first spermatocytes in the second phase of the growth period, dividing primary


and secondary spermatocytes, and young spermatids up to the stage represented in figure 8. To be strictly accurate, I should add that below such cells as are repre- sented in figures 10 and 11 one finds spermatocytes that have reached the end of the growth period. The subsequent modifications of the head of the spermatid take place while the spermatocytes are dividing. The next generation of sperma- tids below the stage corresponding to figure 12 has progressed as far as the stage represented in figure 8.


At the end of this period the remnants of the idiozome have disappeared. The spermatid then contains a number of small granules in one or several heaps (figures 15a and 156), whose size and staining properties suggest that they may be geneti- cally related to these remnants. They are eUminated with the residual body.

Concerning the centrioles, the migration of the ring is the characteristic feature of this period. In figure 13 the ring is shown when the migration is half completed; in figure 14 at the end of it. The process is certainly a rapid one, as both stages are found side by side. From what has been said above as to the relationship between the protoplasmic body of the spermatid and the middle piece, it follows that the ring does not migrate as far as the posterior extremity of the protoplasmic lobe (fig. 14). During this period, in preparations made after Benda's method, the ring does not stain intensely, as do the other parts of the centriolar complex, and finally can not be seen at all (fig. 15a). Other changes in the centrioles are the following: Instead of two granules connected by a thin filament (fig. 12), we find in the next stage (fig. 13) three granules (a fourth one, which happens to be situated just below the ring, is a mitochondrium). A similar arrangement in Phalangista was met with by von Korff, who thinks that the two anterior granules are fragments of the proximal centriole, of which the posterior one moves gradually towards the distal centriole and finally comes to lie quite near it. By analog}^ one might conclude that a similar process has taken place here, although it was not actually observed.

The large, pear-shaped granule becomes more elongated as the ring migrates, as if the influence that carries away the ring had exerted itself on the former body also. It remains visible until after the deposit of the chondriosomes on the axial filament (fig. 14), but in the ripe spermatozoon it has disappeared (fig. 25). Inter- mediate stages frequently show one large granule or two smaller ones, which stain like centrioles and are found in the neighborhood of the head (figs. 16 and 24). To me it seems probable that they proceed from the centriolar granule. In describing a similar element in the guinea-pig Meves states that it gradually diminishes and is finally reduced to a very small size. The centriolar process described by OUver breaks off from the anterior part of the distal centriole; it persists in the sperma- tozoon and is connected with a corresponding fragment of the proximal centriole bj' a thin filament.

The sheath covering the axial filament in the region of the main piece increases very much in thickness during this period. On the anterior extremity of the main piece it appears as an annular swelUng (figs. 13 and 14); towards the posterior end it gradually becomes thinner. Within it the axial filament can usually be followed


for a certain distance. The sheath exhibits a deUcacy of structure, a transverse striation, not represented in my drawings. This has already been described by von Korff, but in my ojiinion is not so marked as liis te.xt-figurc 4 would indicate. It can be seen clearly in certain preparations only, and occasionally Ramon y Cajal's method gives very good images of it. A similar structure has been observed also by Retzius; not in Didelphys, however (1909, pages 125-126), but in the other marsupials studied by him (1906). In one of these {Bettongia) , instead of a trans- verse striation there is a beautifully developed sjnral. The existence of these con- ditions is interesting, as it shows that such structures are not necessarily formed by chondriosomes. I have also in mind, in tliis connection, the spirals, etc., descrit)ed by Retzius (1902, 1910) and by Koltzoff (1908); see also Duesberg, 1912, p. 687.

In the region of the middle piece the axial filament is very Uttle, if any, tliicker than at the end of the previous period (fig. 13). It should be recalled in this con- nection that in the guinea-pig, during the second period, Meves found a small vesicle in that region of the axial filament. The latter appears to be the direct pro- longation of the walls of this vesicle; furthermore, it is thinner within the vesicle than without. For these reasons Meves thinks that a thin sheath covers the axial filament in the region of the middle piece and that this sheath is in direct continuity with the much tliicker sheath that covers the axial filament in the region of the main piece. I found in the rat (1908) a similar arrangement and adopted Meves's conclusions. Von Korff, however, does not seem to be thoroughly convinced of the existence of this sheath on the middle piece, in Phalangista. He states (p. 252) :

"Meves nimnit ausserdem noch cine dritte Hiille an, die dem .\xenfaden direkt aufsitzt; das Bliischen des .\xenfadens, auf Grand dessen er ihre Existenz vermuthet, habe ich nur eimiml gesehen."

In the opossum I have never been able to find this vesicle; furthermore, the structure of the tail, as it appears in figure 13, makes it rather improbable that the sheath of the main piece extends onto the middle piece.

At the end of this period the major part of the protoplasm is eUminated. Here again there is a close resemblance to the same process in Phalangista. While in the guinea-pig (Meves, 1899) and in the rat (Duesberg, 1908) the protoplasm accumulates on the side of the middle piece and is there eliminated by progressive indentation, in the opossum (as well as in Phalangista) it flows to the anterior i)art of the spermatozoon (figs. 15a, 19, 20, and 21). The cast-off masses (Regaud's corps rC'siduels) constitute for a time a continuous laj'er between the expelled sj^cr- matozoa and the next generation of spermatids (fig. 16). These masses contain, besides the small granulations mentioned above, which are presumably remnants of the idiozome, a number of larger bodies and vacuoles which will be describeil in connection with the chondriosomes. The residual bodies are later taken up by the Sertoli cells and undergfi degeneration, the termination of wliich is verj- often a transformation into fat .


Besides the changes that affect the chondriosomes, described in tlie next chaj)- tcr, two imjjortant modifications take ])lace during this period: the r.)tation of the


head and the so-called "copulation" of the spermatozoa. The first phenomenon is described by von Korff (p. 2o4) as follows :

"Der Schwanz der Spermie liegt vor dcr Kopulation zweier Spermienkopfe quer zu den beiden Schenkeln, nach derselben dagegen langs zu ihnen und zwar mit dem vorderen conischen Ende des ^'erbindungsstiickes und dem Halse zwischen ihnen."

The copulation of two spermatozoa was first described by Selenka. Von Korff observed it in Didelphijs but not in Phalangista. However, as Selenka states that most of the spermatozoa found in the vagina of Phalangista are twin cells, von Korff suggests that the coi)ulation may take place only after the passage of the spermatozoa in the epididymis; i. e., in the vas deferens. Retzius (1906) does not mention the presence of twin spermatozoa in Phalangista, but found them in the epididj-mis of Didelphys (1909). Jordan apparently overlooked them at first (p. 54), but corrected his opinion afterward (note 4, p. 76). I myself also found the copu- lating spermatozoa in the epididymis of the opossum. Preparations fixed with osmic acid and stained with Benda's method show that a distinct Une of demar- cation between the two heads is visible onh^ in the posterior part (fig. 25). The fusion is, however, not as intimate as this appearance would lead one to beUeve; indeed the connection must be a rather loose one, as in smears the spermatozoa are often found separated. As to the significance of the copulation nothing is known. Some clue might be expected from a study of the fertilization in Didelphys; unfor- tunately, in Hartmann's paper on the subject (1916) the fact is not even mentioned.

A description of the comphcated head of the spermatozoon has been given by von Korff (p. 254) and by Retzius (1909). One detail of structure has apparently been overlooked by von Korff, but is described by Retzius. This author mentions the presence, in close proximity to the insertion of the tail, of —

"einer kleinen, bei der Osmium-Rosanilin-behandlung rot, nach der Sublimat-Hama- toxylinalaun schwarz fiirbaren Kugel, welche offenbar einem Centralkorper entspricht. Nach der Ablosung des Schwanzes nebst dieser Kugel bleibt gewohnlich ein heller Fleck an der Scheibe zuriick, welcher wohl als ein Griibchen aufzufassen ist, in welchem die genamite Kugel angeheftet gewesen ist."

No such structure could be found after Benda's stain (cf. figures on plate 1, and fig. 25), but it was brought into e\ddence after osmic fixation by iron hema- toxyhn (fig. d), and by acid fuchsin-methyl green after Fia.D.-Theheadsoftwo

fixation with Regaud's fluid (figs, 18, 19, 20, 22, 23, ^ , copulating spermato-

'^ • , 1 II (t) it ^°^- f"""™ ^ section of

and 24). As the latter series shows, the structure fi\l(7 the epididymis, stained

,..,,, iPii i I ^t with iron-hematoxylin

appears during the development of the spermatozoon, \j[y after Benda's Hxation. and is always to be found on the thick branch of the zcissimn:.2mm.,oc.i2.

nacleus; i. e., the one through which the twin spermatozoa become united (figs. 18, 19, 20, 22, 23, 24, and d). I never saw any evidence of the presence of a granule in this depression, as accepted by Retzius, and while I would leave this point un- decided one thing is certain : If such a granule exists it has nothing to do with the centrioles.

The centriolar ring, or "Schlusscheibe," is described bj^ Meves (1899, p. 360, footnote) and pictured by von Korff (text-figures 3 and 4) and b}- Retzius (1909,


figs. 7, 15, and 17) as a very conspicuous body. As already stated, it can not be seen in preparations after Benda (figs. 16 and 25), but appears very clearly in smears stained with iron hematoxylin. The centriolar granules located in the collar are represented l)y von Korff (text-figures 3 and 4) as three granules disjjosed very much like the same granules in my figure 13. Retzius, however, describes, instead of the two posterior granules:

"einen kleinen dunklen Halbring welcher wohl auch zum Centralkorperapparat gehort und dann als die vordere Abteilung des distalen Korpers zu betrachten ist."

]\Iy observations agree with von Korff 's, with this difference; that the two posterior granules, which are still apparent in spermatozoa taken from the testicle (fig. 16), appear usually more or less completely fused together in smears from the epididymis (fig. 25). SUght differences in the shape and length of the head and middle piece, noted between smears and sections, are probabl}^ due to a sort of capillary action exerted in the former (cf. figs. 16 and 25).


Since Jordan's surprising conclusions, a reinvestigation of the chondriosomes in the testicle of the opossum has always appeared to me as necessar5^ An effort made several j'ears ago to collect material from the zoological gardens in Europe proved unsuccessful, so that the present opportunity was gladly taken.

The details of Jordan's description will be discussed when the necessity arises, in connection with my own observations, and I shall Umit myself to a resume of his main conclusions. While he admits that chondriosomes are present in the spermatids, and that part at least of these bodies form the spiral filament of the spermatozoon, he denies their existence at certain stages of the process of sperma- togenesis. Concerning their absence in the spermatogonia and in the Sertoli cells, he does not express himself very definitely. He then continues :

"But my preparations leave no doubt respecting the absence of mitochondria during the early growth period of the primary spermatocytes. For this generation of cells, they first appear during the later growth period — and during a period coincident with a transi- tory' achromatic reticular phase of the nucleus. This observation is the more significant in view of the fact that both within and without the nuclear wall are similar darker- staining bodies. Subsequently such bodies (now deeply staining, sharply contoured spheres and dumb-bells) arc aligned on the nuclear membrane externally. All the evidence here hints to a nuclear origin of mitochondria, i. e., they appear to be tran.sformed chromidia (p. 59.)"

Of this nuclear origin Jordan is not, however, altogether sure. The main point upon which he insists is their discontinuity.

"I believe that the fact of their apparent absence in the young spermatocytes of the opossum (and possibly other forms) is one of the strongest arguments against the Bonda- Meves-Ducsberg theory of their continuity and hereditary significance" (p. (ii)).

There is no doubt that Jordan's conclusions, if verified, arc of considerable importance. I have from the beginning been aware of it, and have been anxiously awaiting an opportunity to study the same material. At the same time I could not


help expressing (1912) great skepticism, which the results of my present investiga- tions entirely justify. The main conclusion reached, in fact, is that there is no discontinuity in the chondriosomes of the seminal cells in the opossum. Chondrio- somes are present at all stages of spermatogenesis from the spermatogonia and the Sertoh cells to the ripe spermatozoa. This conclusion, arrived at through the study of the adult testicle, is corroborated by the study of the organ in young animals. I come now to the details of my observations.


Chondriosomes are exceedingly numerous in the SertoU cells (fig. 1) and exhibit a marked resistance to destructive influences, incluchng the action of acetic acid. One can obtain preparations in which most of the chondriosomes, especially in the early stages of spermatogenesis, are destroyed, while they are preserved in the Sertoli cells. In such preparations the cell-body, with its processes, is sharply brought into evidence. The nucleus is very darkly staining and, as in other mam- mals, shows indentations. Droplets of fat, blackened by osmic acid, are very numerous at certain stages, namely, immediately after the resorption of the resi- dual bodies. The chondriosomes are either granules or filaments. Some of these latter are very long and may extend into the processes, but are usually confined to the basal part of the cell, around the nucleus.


The resting spermatogonium (fig. 2) flattened against the basal membrane, contains, besides a relatively large nucleus, an idiozome and, notwithstanding Jordan's assertion to the contrary, numerous chondriosomes, but no fat. The chondriosomes are all mitochondria, massed wherever thej^ find space, mostly at the poles of the nucleus. The idiozome is usually hidden by them. During mitosis the cell rounds out and the mitochondria are found at the metaphase (fig. 3) scat- tered all around the spindle, and later between the daughter-nuclei.


During the first phase of the growi;h period the chondriosomes keep their granular form. They are gathered mostl}' around the idiozome at one pole of the nucleus (fig. 4) . Jordan denies the existence of such a distribution, because, having overlooked the chondriosomes in this stage he endeavored to find the same condi- tion when it no longer existed. As a matter of fact, this distribution of the chon- driosomes is of common occurrence in the first phase of the growth period in mam- mals. Later on the location of the chondriosomes changes and they are found all around the nucleus (fig. 5) . Their shape likewise is modified ; most of them are now short, rather thin filaments, retaining this form well into the period of spermio- genesis. They never look hke the dumb-bells represented by Jordan in figure 24, nor hke the granules in his figures 25, 26, 27, 29, 30, 31, 32, 33, 34, 36, 37, and 38. Nor are they located in the cell as his chromidia shown in the same figures. As to the metachromatic granules, "which are tentatively interpreted as the earlier


forerunners of the mitochondria," it should be noted that they are already dei)rived of their main interest, since it has been estabUshed that chondriosomes are present in all preceding stages. A special effort, however, was made to ascertain what they could be. It appears probable that the nuclear condition to which Jordan refers is the appearance one constantly finds at the periphery of pieces fi.xed with reagents containing osmic acid; in fact, the typical image of a nucleus after strong osmication, as shown in figure 5. That the masses of chromatin there represented are expelled into the protoplasm is an assumption in favor of which no evidence could be found.

There is little to say regarding the beha^aor of the chondriosomes during the mitoses of maturation. The process is similar to what has been described in other mammals, and identical in both divisions, so that it appeared useless to give more illustrations than figure 6, which represents the metaphase of the first division. The chondriosomes are found all over the cell. Later they occupy the space between the daughter-nuclei and are segregated in equal quantities, or approximately so, between the daughter-cells.


The chondriosomes of the young spermatid have still a filamentous form (figs. 7 and 8). Immediatelj'^ after the stage represented in figure 8 in Benda's jjrepara- tions, and still a little later after Regaud's fixation (see figure 17, which represents a somewhat more advanced stage than figure 8), one finds only small vesicles, elongated at first (fig. 9), later perfectly spherical. These vesicles have been observed by Jordan. As it is well known that chondriosomes when poorly pre- served have a tendency to swell, the interpretation suggested itself that this appear- ance was due to defective fixation, notwithstanding that during the interval between the end of the first period and the beginning of the second no other form of chon- driosomes was found in fixed material. The study of living cells confirms this view; such stages as are represented by figure 9 are readily recognizable in teased prepara- tions of seminiferous tubules, and no vesicles (only solid granules) can be seen in them. It is a fact, however, that during this particular period the chondriosomes are especially sensitive to the action of fixing reagents. At the same time they acquire a considerable power of resistance to the dissolving action of acetic acid. From this time on the}' can be found in nearly any material. I have seen them in I)ieces fixed with Flemming's, Bouin's, or Hermann's fluids, all of which hold 5 per cent acetic acid. They do not, however, appear quite so clearly as in Benda's or Regaud's preparations and, furthermore, they retain a swollen appearance even in these later stages, during which, in Benda's or Regaud's material, they reappear as solid granules or rods. This increase in the resistance on the jiart of the chon- drio.somes to acetic acid in the last stages of si)ermiogenesis is nothing new and seems to be of general occurrence. Only lately (191<S) I d(\scribed another instance of it in the testicle f)f Fundulun. It is wortli while, however, to empliasize it in tliis


case, as Jordan has endeavored to meet the reproach that his technique may have been defective by saying:

"Even a poor technique that, however, reveals them (the chondriosomes) clearly and typically at a certain stage and subsequently, should reveal them at every stage (when present) represented by cells in the same tissue (page 69)."

It is, in fact, not surprising if Jordan, in his material, could find chondriosomes in the spermatids and not in earlier stages.


At first the chondriosomes have still a vesicular appearance. Verj'^ soon, however, at least in the most peripheric parts of pieces fixed in Benda's or Meves's fluid, they assume the shape of rather large, solid granules (fig. 10). Even this appearance is somewhat artificial and due to a certain amount of swelling, as the granules in the living cell are distinctly smaller. The territory of the caudal tube is entirelj' free of chondriosomes. Since, in the opossum, the caudal tube fills the w^hole width of the spermatid, all chondriosomes are separated from the nucleus by the entire length of the tube. In the protoplasmic lobe the chondriosomes do not show an}' special arrangement ; no stage was found during this period as in the guinea-pig (Duesberg, 1910, fig. 46), nor during the preceding period as in the rat (Regaud, 1908), and again as in the guinea-pig (Duesberg, 1910, figs. 57 and 58) in which all the chondriosomes are collected at the periphery of the cell.

In a somewhat later stage, when the head begins to assume its definite shape (fig. 11), the chondriosomes are constantly found, both in Regaud's and in Benda's material, in groups of thin, bacillus-shaped rods. Still later (fig. 12) thej^ appear as granules, forming several heaps in which the mitochondria are crowded very closely together. The same stage is marked by the fiirst appearance of the so-called "von Ebner's tingierbare Korner" in the form of two or three granules or droplets, which usually are so close together that they fuse, assuming thus a shape different from that of the chondriosomes (fig. 12, upper left corner). These bodies can be easily distin- guished from the chondriosomes, notwithstanding Jordan's opinion to the contrary. Even after Benda's method, which stains them purple, or iron hematoxylin which stains them black, their shade is different from that of the chondriosomes. Their form is also different. These bodies are fixed by all reagents except Regaud's. They can be brought into evidence in a specific way, even in material in which the chondriosomes are well preserved, by staining the sections with a nuclear stain — for example, with safranin. An interesting and convincing experiment consists in unstaining a preparation made with Benda's method after having drawn a given cell and having carefully noted its location by means of the mechanical stage; then staining the preparation with safranin and redrawing the same cell. "VMiiie the chondriosomes do not take up the safranin, the "tingierbare Korner" do so with great avidity.


At the time the ring begins to migrate the chondriosomes are still granules, but are scattered all over the cytoplasm (fig. 13). It has aheady been pointed out


that the migration of the ring is very rapid. The next change, i. e., the collection of the chondriosomes on the axial filament, also takes place with remarkable rapidity, for stages Like those represented in figures 13 and 14 are found side by side, while intermediate stages are very scarce. It would seem that after the ring has traveled a short distance on the axial filament a sort of attraction is exerted on the chondriosomes. This impression is rather strengthened by a study of preparations fixed with Regaud's fluid, in which rods are found instead of granules (fig. 18).

As neither the "tingierbare Korner" nor the small granules which I have regarded as remnants of the idiozome are fixed by Regaud's fluid, these prepara- tions are especially convenient for the study of the chondriosomes in the last stages of spermiogenesis. In figure 18 (and also in figure 14) the middle piece appears covered with chondriosomes, while a number of these bodies are still scattered in the cytoplasm. These latter ones will never find a place on the axial filament. Figure 19 shows how they are carried away by the protoplasm flowing toward the head to form the residual body; they are finally accumulated below the nucleus (figs. 20 and 21). At the same time they dispose themselves in a peculiar and quite characteristic manner close to the periphery of the protoplasm (figs. 15a, 20, and 21), a disposition which is still more evident on cross-section (fig. 15b). Finally they are eliminated. I therefore agree on this point with Jordan, but nevertheless doubt very much that his conclusion was supported by his own observa- tions. Had he seen the process just described he certainly would have mentioned it, but neither in his drawings nor in his text does he give any indication of it. In my opinion, what Jordan probably saw was the elimination of the "tingierbare Korner" and other granules especially well preserved in fixing reagents containing osmic acid.

All the protoplasmic granules and detritus are very conspicuous in prepara- tions after Benda's method. The "tingierbare Korner" (figs. 13, 14, 15«, and 156) are somewhat larger than in the preceding stage (cf. fig. 12). Next to these are the small granules interpreted as remnants of the idiozome (figs. 15a and 156). In the residual body cast off by the spermatozoon, there appear numerous vacuoles and large granules (fig. 16), some stained in purple, others in brown. The granules, or part of them at least, are probably formed by the eUminated chondriosomes which degenerate. Romeis (1912) has described a "Verklumpung" of the chondriosomes in degenerating spermatozoa found in the so-called "poche seminale" of Ascaris; the masses formed by the chondriosomes retain for a time their original stain, but later, in Benda's preparations, take up the suLfaUzarin.^ Regaud's material shows in the residual bodj' more vacuoles than Benda's (fig. 21), the former reagent appar- ently dissolving some of the granules preserved in the latter.

In connection with the eUmination of chondriosomes at the end of the spermio- genetic process in mammals, I would recall that Regaud (1908) described its occur- rence in the spermatid of the rat, while I would not admit it, any more than in the

'A paper by Bang and Sjovall (1916), which I would have liked to consult in this connection, was not available.


guinea-pig (1910). The present experience, however, has led me to reconsider this opinion, and I contemplate reinvestigating both the rat and the guinea-pig at an early date.

The number of the chondriosomes eUminated varies, but is usually very small (figures 15a, 156, 19, 20, and 21). On the other hand, the number of chondriosomes on the middle piece is fairly constant, as will be shown later. The quantity elimi- nated is consequently a function of the quantity held by the spermatid, all the chondriosomes (no matter how numerous) that do not find a place on the middle piece being eliminated. To the significance of this ehmination Jordan seems to attach much importance :

"The loss of a considerable amount of mitochondrial substance in the cast-off portion of the spermatid * * * during the process of metamorphosis, militates against the interpretation of mitochondrial continuity and a hereditary role of this substance (page

70). "1

Setting aside the fact that the amount eliminated is not considerable, I fail entirely, as stated before (1912, page 624), to see what theoretical importance it can have.

To return to the chondriosomes surrounding the axial filament, let us note their further evolution in Regaud's preparations. While the superfluous chondriosomes are migrating toward the anterior part of the spermatid to be later eliminated, the others become gradually distributed with considerable regularity (figs. 19, 20, 21; also fig. 15a). When the elimination is completed (fig. 22) all the Uttle rods are arranged obUquely in relation to the axial filament and at the same time are laid out in regular longitudinal rows suggesting a fir-apple or an ear of corn. The same regularity appears also in preparations after Benda, although, as already men- tioned, instead of rods we find in these rather large granules (fig. 15a). The num- ber of elements in each longitudinal row can be estimated up to seven or eight, and in cross-section, seven (fig. 156); the total number of chondriosomes would con- sequently be 49 to 56. At the anterior part of the middle piece preparations after Regaud show invariably some irregularity in the disposition of the chondriosomes, inasmuch as some of them (usually two) are disposed almost perpendicular to the others (figs. 21 and 22). These, I assume, are going to form the first winding of the spiral.


From the description given above it results that, at the time of the elimination of the residual body, no spiral has been formed. Its formation belongs entirely to the fourth period; in other words, it takes place in the spermatozoon eUminated in the lumen of the seminiferous tubule. First, the rods dispose themselves perpen- dicular to the axial filament and appear somewhat thinner; the two anterior ones seem to have fused together into one curved filament (fig. 23). Then, by confluence of the rods the spiral is formed. In material fixed in Benda's fluid it is perfect in its regularity (fig. 16) ; after Regaud's fixation it is somewhat u-regular (fig. 24).

•A similar opinion was again expressed by Jordan in 1914 (p. 167).


As clearly shown in figures \oa and 16, there is a great discrepancy lietween the size of the chondriosomes and the thickness of the spiral after Benda's fixation. I do not doubt that the ajJi^earance represented in figure 15r; is som(;\vhat artificial and due to a certain amount of swelling. Certainly the chondriosomes after Regaud's fixation are more similar in form to those in the living cell than they are after Benda's fixation.

The formation of the spiral filament at the expense of the chondriosomes, in the spermatozoon of marsupials, was first described bj' Benda (1897) for Phalan- ffisia. ^'on Korff came to the same conclusion both for Phalangista and Didelphys; it must be stated, however, that liis representation of the sjiiral filament in the last- named species (text-figure 3) is somewhat schematic. In a later paper Benda (1906) confirms his former description for several marsupials. Jordan also has observed the spiral in the opossum. Finally, in the present paper I have followed its formation, step by step. In fact, it is such a conspicuous constituent of the spermatozoon that it is hard to understand how Retzius failed to see it. That author has pubhshed two papers on the spermatozoon of marsupials. In 1906 he studied Betlongia cimiculus, Macropus hilliardieri, Petrogale penicillatn, Onycho- gale lunata, and Phalangista vulpina. The structure of the middle piece is in all these species the same; it is covered by a relatively small number of rather large granules, disposed in longitudinal rows. For Phalangista Retzius expres.sly states that the regular disposition of the granules simulates a cross-striation, but he fails to compare his results with the entirely different ones of Benda and von KorfT, although he mentions them. The next paper (1909) deals exclusively with Didel- phys. Here again the middle piece is found covered with granules :

"Diese Korner liegen in geraden Reihen, mit zehn Kornern in jeder Langsreihe; von dcT Seite oetrachtet zeigt das Verbindungsstiick drei solche Langsreihen. Nach dieser

Berechnung diirfte die .\nzahl der Korner sich auf etwa 40 belaufen Sie liegen

auch am reifen Spermium nicht in spiraliger Ordnung, sondern regelma.«sig der Quere nach, und .sie verwandeln sich jedenfalls nicht zu einer Spiralfaser. (p. 12.5)."

I disagree with this description, first, in the estimate of the number of rows; second, as to the form of the chondriosomal sheath of the rii)c spermatozoon, where, like Benda, von Korff, and Jordan, I found a spiral filament. This difference might be explained in two ways: First, the species studied by Retzius may not have been Didelphys virginiana; this hypothesis, however, is rather improbable and would not account for the discrepancy between his conclusions and those of von KorfT and Benda concerning Phalangista. The second hypothesis, which would explain all differences, is that the material studied by Retzius was poorly ])resor\ed and showed a chondriosomal sheath that had fallen to pieces.

While writing on the structure of the mammalian six'rmatozot)n 1 wish to correct certain errors ajjpearing in the second edition of Bonnet's Embrj'ology (1912, J). 28), in the rejiroduction of some of my drawings showing the development of the spermatozoon of the guinea-pig (1910). What Bonnet calls "rordcrc Ilals- knblchcn" in figures d and c, is really the jiosterior edge of the headcap; while his "hinterc H alsknutchen" are the fragments of the proximal centriole and should be


designated as "vordere Halsknotchen." In figure g this last term is correctly used, but in the same figure, as well as in figure/, the chondriosomal sheath of the middle piece is referred to as " Spiral-faden und -hulle;" in the guinea-pig, however, the chondriosomes do not form a spiral filament. I must add that I am not by any means responsible for these mistakes.


As to the ultimate fate of the chondriosomes in spermatogenesis, the present investigation leads to the conclusion, readily foreseen, that the chondriosomes build a part of the spermatozoon — ^in this case a spiral filament surrounding the middle piece.

As to their origin, Jordan's assumption of their discontinuity and their nuclear nature in the seminal cells of the opossum can, in my opinion at least, after this reinvestigation be considered as a failure. To me this conclusion is no more sur- prising than the first one. It must be stated, however, that in later j^ears the theory of the nuclear origin of the chondriosomes has again been taken up by AlexeiefT, Walton, and K. E. Schreiner.

As far as I know, Alexeieff has pubhshed a number of notes on the subject, all deaUng wath protozoa (1916). He finds in these organisms bodies which he calls "mitochondries," and which he beheves to be of nuclear origin; hence he proposes calling them chromidia. I must say that nowhere can I find any argument in favor of this author's conclusions.

Walton (1916) thinks he has demonstrated that the chondriosomes of the seminal cells in Ascaris canis Werner are formed at the expense of nuclear material in the spermatocytes, and he draws therefrom far-reaching conclusions. As chon- driosomes do, however, exist in the spermatogonia of Ascaiis (Duesberg 1912, p. 638, and Faure-Fremiet, 1913), Walton's premises are incorrect, and any further discussion is unnecessary. The explanation of liis failure is very simple: he fixed with strong Flemming's and Carnoy's fluids, neither of which can be trusted for the preservation of the chondriosomes.

More serious appears Schreiner's attempt. So far he has, to my knowledge, pubhshed observations only on the fat-ceUs of the subcutaneous tissue of Myxine. He promises to deal in subsequent papers with pigment-cells, blood-cells, and cells of the connective tissue, with glandular and seminal cells, the study of which brings him to the same conclusion as the study of the fat-cells. In these he finds a num- ber of rods, "Plasmastabchen," stainable wdth acid fuclisin (after Altmann's or Altmann-Kull's method), or with iron-hematoxyhn, "welche zur Bildung der Fettvakuolen Anlass geben." Schreiner's view on the formation of fat is a confirmation of those already expressed by a number of authors, among them Metzner, Dubreuil, and Hoven. He differs from them, however, inasmuch as, according to him, the "Plasmastabchen" are formed from nuclear substance, not from chromatin, as in Goldschmidt's chromidial theory, but from nucleolar sub- stance. He concludes that liis observations formally contradict what he calls "die Meves-Duesbergsche Lehre."


Before undertaking the critique of Schreiner's paper, a few remarks of a general character would seem not to be amiss. Schreiner's assertion of the nuclear origin of his "Plasmafiiden" recalls to my mind two other papers in which a similar asser- tion was made: one by WassiUeff (1907) on the seminal cells of Blnlln germanica, the other by Jordan on Didelphys. In both cases it was claimed that the chon- driosomes were formed in the spermatocytes only and in both cases I found (a positive result against a negative one) that the chondriosomes were present in the spermatogonia and were transmitted during mitosis to the spermatocytes. For Blatta the formation of chondriosomes at the expense of nuclear material was sup- posed to appear so clearly that Goldschmidt speaks of "die schonen Befunde von WassiUeff, deren unbedingte Beweiskraft fiir den, der die Praparate kcnnt, die noch viel klarer sind als die Zeichnungen, keinem Zweifel unterhegen kann (1909, p. 110)." When, however, I studied the same material (1910), I could find no evi- dence to substantiate the nuclear origin of the chondriosomes and, so far as I am aware, no answer has ever been made to my criticism of Wassilieff's conclusions. In fact, the sharpest critics of Goldschmidt's theory have since been found in his own laboratory, a point to which I shall return later.

These two experiences, together with the increased knowledge I have acquired of cytoplasmic structures, have made me very skeptical of such claims as Schreiner's. As to the present case, I might add that (like others, who have not forgotten the story of the enumeration of chromosomes in Zoogonus niirus) I am not inclined to accept Schreiner's assertions as "ready money." When one attempts, however, a specific and detailed consideration of his present observations, it readily appears that a thorough discussion is hardly possible, as Schreiner's communication is only a preliminary one, in which he repeatedly refers to his future paper and on points of no minor importance. Let us take for example the question of the seriation of the different aspects of the "Plasmafiiden." A priori, there is no reason why we should accept that figure 20 represents a stage of fragmentation of these bodies, while certain granular filaments in figures 4, 5, 6, and 7 represent their formation. A correct seriation is, in fact, very difficult, for in most cells not only smooth fila- ments, but also granular ones, which Schreiner supposes to be here in process of formation, there of fragmentation, are present (see pages 159 and 164). The author himself admits the difficulty; one of his criteria is the condition of the nuclei:

"Auch der Kern hat in den Zellen, wo die Segmentierung stattfindet, pin von demjcnigen ganz verschiedenes Aussehen, das wir von den Zellen konnon, innerhalb deren Cytojilasma die Stiibchen gebildet werden. Betreffs dieses letzteron Punktes muss auf nieine ausfuhrliche Arbeit verwiesen werden (p. 163)."

As I have not seen Schreiner's completed paper, I can not express any opinion as to the value of his arguments and will therefore Umit myself to pointing out the difficulty. I might state incidentally that there is a contradiction (at least what ai>pears to be a contradiction in the face of available data) between figure 22 (wliich shows smooth filaments only and a nearly spherical nucleus) and the description on page leO, where we read that "diejenigcn Fettzellen, die in ihrem Cytoplasma zahlreiche grosse Kiigclchcn enthalten, sphiirische Kerne mit ebenfalls runden


Nukleolen aufweisen, wiihrend diejenigen Zellen, deren Plasmakiigelchen sich zu Stabchen zu entwickeln angefangen haben, in der Kegel gelappte Kerne besitzen." As to the crucial point of the origin of the "Plasmafaden," I would suggest as a possible cause of error the fact that Schreiner has made use of only such methods as stain nucleoU and chondriosomes alike — i. e., iron-hematoxyUn and acid fuchsin — and not of Benda's method, which stains them differently. It should also be recalled that other authors who have studied the fat-cells have declared themselves for the cytoplasmic nature of their chondriosomes; for instance Dubreuil, whose completed paper (1913) Schreiner has overlooked.

Taking for granted, however, that Schreiner's description is accurate, nothing proves yet that liis interpretation is correct, for the process described as an expul- sion of substance from the nucleus might just as well be the reverse. And even if we accept Schreiner's interpretation, we find that the question of the origin of the chondriosomes has really not been touched, since these bodies are present, according to Schreiner's own description, in all cells before the process he describes takes place.

These are the points wliich the preliminary account of Schreiner suggest to me. I wish to add that the bibUograpliic review contains a number of omissions and errors. Among the first I note, aside from the one already mentioned (Dubreuil's paper), Schreiner's denomination of the theory of the cytoplasmic nature of the chondriosomes as the "Meves-Duesbergsche Lehre." As a matter of fact, prac- ticalh^ all cj^tologists agree on this point, one of the last to so express himself being Maximow (1916). As to errors, we read, for instance, on page 148 that "die Anhan- ger der Chromidialtheorie stimmen mit denen der Plastosomentheorie darin voU- kommem uberein, dass die Chromidien und Plastosomen die namhchen GebUde sind" — an opinion never expressed by me (see Duesberg, 1912). Further, we find that the criticism of Retzius (1914) is very highly praised. I have already pointed out why Retzius can not be considered as an authority in this matter (Duesberg, 1915, pages 62-63) and would refer the reader also to Meves' answer (1914, 2) to Retzius. On page 169, Schreiner writes:

"Da ich das Material, bei dem Goldschmidt die Beobachtungen machte, welche die Grundlage seiner Chromidialtheorie bilden (die Gewebszellen der Ascariden) nicht aus eigener Untersuchung kenne, wage ich zu den verschiedenen Meinungen, die iiber die Natur seiner Chroroidialstrange von verschiedenen Seiten (Vejdovski, BUek, Duesberg) geaiissert sind, keine Stellung zu nehmen. Doch muss ich gestehen, dass es mir schwer fallt zu glauben, dass sich Goldsmidt von seinen Praparaten dermassen hat tauschen lassen, wie die genannten Autoren behaupten."

To this, I would answer that I have expressed no definite op'nion on Gold- schmidt's "Chromidialstrange," as I recognize "dass es kaum mogUch ist, sich ohne eigene Kenntnis des Objektes kategorisch auszusprechen (1912, p. 907)." I insisted, however, that all authors agree on one point, i. e., that Goldschmidt was mistaken; and I would emphasize here that these authors are not only the ones mentioned by Schreiner, but also Sjovall and Lundegardh (who declare themselves unconvinced by Goldschmidt's description), Hirschler, EhrUch, Jorgensen, and


von Kemnitz (the last three working in Goldschmidt's laboratory) who, like Vej- dovskj' and Bilck, speak from personal experience. In all fairness it should be added that Jorgensen and Ehrlich agree with Goldschmidt as to the nuclear origin of some of the "Chromidialstriinge," but in a way which can hardly be satisfactory to him: "In den von Goldschmidt beschriebenen Fiillcn handelt es sich um durch das Messer herausgerissene Nukleolen oder Chromatinbrocken des Kernes selbst." I shall not dwell further upon the chromidial theory, for the whole case was thor- oughly exposed bj' me in 1912 and, as I like to recall, my criticism has never been answered.

Note. — Through the courtesy of Prof. E. B. Wilson, I had the opportunity, after the above was written, to read a recent paper by Schreiner: "Zur Kenntnis der Zellgranula. Untersuchungen iiber den feineren Bau der Haut von Myxine glutinosa. Erster Teil, erste Hiilfte. Arch. f. mikr. Anat. Abt. I. Vol. 89." I can not in the present paper, discuss at length Schreiner's article, but I would state that in my opinion all I have said before concerning (1) the accuracy of Schreiner's description, (2) the correctness of his interpre- tation, and (3) the legitunacy of his conclusions, still holds. As to the second point, it should in all fairness be stated that Schreiner himself is aware of the difficulty and discusses it at some length. On page 140 I find the following argument in favor of his view:

"Schon der Umstand dass die feinen Verbindungsfaden zwischen den kleinen f. en Plasmakorn- chcn und dinn Nukleolus innerhall) des Kernes oft eine betrachtliche Lange haben konnen, und dass die Nukleolarsubstanz sich in diese Faden nicht selten allmahlich fortsetzt, scheint zugun- stcn der ersteren Erklarungswcise zu sprechen. Eine in dieser Hinsicht noch grossere Bedeutung wird man aber der Tatsache beimessen mtissen, dass in einigen Zellen vom Nukleolus ahnliche Faden ausgehen, die an der inneren Wand der Kernmembran endigen, ohne mit irgend einem Plasmakornehen in Verbindung zu treten. Solche Bilder wird man schwer auf andere Weise deuten konnen, als dass wir hicr eine Vorbereitung fiir die Ausstossung der Nukleolarsubstanz tlurch die Kernmembran vor uns haben."

How the conditions described should rather speak in favor of Schreiner's opinion than against it, I fail entirely to see.



The entire literature concerning the apparatus of Golgi, up to the spring of 1914, was reviewed by me at the meeting of the Anatomische Gesellschaft in Inns- bruck. Recent contributions to the subject have been made by Addison (1916), Basile (1914), Birck (1914), de Castro (1916), Cowdry (1916), Deineka (1914),' Hortega (1914), Monti (1915), Pensa (1915), Pappenheimer (1916),' Ramon y Cajal (1914), Sanchez (1916), and Speciale (1914).- The most interesting points in these contributions will be reported at the end of this chapter. For the pre\'ious Uterature I take the liberty of referring the reader to my review and will confine myself at this time to recalling the observations of Sjovall (1906), Perroncito (1910), and Weigl (1912), which are closely connected with my subject.

By means of a special method Sjovall brought into evidence a number of rods at the periphery of the idiozome, in the spermatocji^es of the mouse. In the sper- matids the same method stains that part of the idiozome which is not used in the formation of the headcap and which is finally eliminated.

Perroncito was the first to apply the silver impregnation (Golgi's method with acidum arseniosum) to the testicle. In the spermatogonia of the rabbit he found, at one pole of the nucleus, a typical reticular apparatus. In the spermatocj-tes the apparatus is somewhat smaller and sometimes irregular, inasmuch as from the reticulum a long process may be sent out into the cell. As to its behavior during mitosis, Perroncito does not express himself definitely, although he thinks that there are some indications of a process similar to that which he finds in the sperma- tocytes of Paludina (dyctiocinesis) . In the spermatids the reticulum is even smaller than in the spermatocji^es. What becomes of it he could not ascertain, but is incUned to believe that one part of it is eliminated, while the rest remains in the spermatozoon.

Weigl (1912) studied the ripe spermatozoon in the guinea-pig by means of Golgi's, Ramon y Cajal's, and Kopsch's methods. In the protoplasmic sheath of the collar he found some rods and granules, which he is inchned to consider as an apparatus of Golgi. He wonders whether the apparatus is carried by the sper- matozoon into the egg and has, perhaps, something to do with the formation of the apparatus in the embr3^onic cells.

In the spermatogonia of the opossum I find very often two kinds of bodies impregnated by Ramon y Cajal's method (fig. 26). There are first a number of granules which resemble very much the mitochondria of these cells. There is, further, a much more voluminous body flattened against one pole of the nucleus. The central part of this body is formed by a fight substance and its periphery is

' While Pappenheimer felt the necessity for "collating the widely scattered and rather inaccessible Uterature," he merely reproduces the data collected by me up to 1914 and gives only an incomplete account of the recent papers.

' Voivov (1916) should probably be added to this list. He descrilws, in the spermatocytes of GruUotaliM, a body which is stainable by the chondriosomal methods and which can also be impregnated by the silver methods. Single in the young spermatocytes, it later separates into four parts which come to rest against the nucleus. During the mitoses of maturation these bodies are equally segregated between the daughter-cells, so that each spermatid gets one of them. This body is elimi- nated at the end of the spermiogenesis.

There is undoubtedly a striking resemblance between this body on one side, and Platner's "Nebenkernstabchen and Perroncito's "dyctiosomes" on the other.


marked by a sharp, black line. Although no such reticulum as Perroncito figures for other mammals could be brought into evidence, I do not doubt that this darkly impregnated substance corresponds to his apparatus of Golgi. That the method was apphed with full success is demonstrated by the presence of a beautifully developed reticulum in the interstitial cells (fig. 34). Further, the Sertoli cells in my preparations show a structure similar to that represented in Perroncito's figure 83. Tliis pictures a Sertoli cell of the guinea-pig, containing a system of anasto- mosed rods, which Perroncito considers a typical reticular apparatus. I can con- firm this observation for the guinea-pig; in the opossum a similar condition is met with, the difi"erence being that the filaments are somewhat tliickcr and often appear hollow.

In the spermatocytes studied during the first phase of their evolution (fig. 27) the apparatus has increased considerably in size and shows certain details in struc- ture which maj' have been present in the spermatogonia, but which could not, perhaps, be seen on account of the small dimensions of the body. Sections per- pendicular to the apj)aratus show that the dark envelope is missing in the middle part of the side toward the nucleus (fig. 27, left). Tangential sections (fig. 27, right) reveal the fact that this dark envelope is not a soUd shell, but is formed of a number of filaments of varying thickness. Some of these filaments are anastomosed, yet the impression given is never that of a typical reticulum, as Perroncito figures in other species. Sometimes the apparatus is formed of two lobes reunited by a filament.

During the second part of the growth period the change in the shape of the apparatus apparently is in relation to the increased size of the spermatocj^te; having more space, the apparatus rounds out (fig. 28). During the prophase of the first division, and as late as the metaphase (fig. 29), the apparatus appears as a sort of unrolled coil located anj^where in the cytoplasm. In the anaphase (fig. 30) a num- ber of granules, or clumps of granules, may be found scattered between the daughter- nuclei, the thread having apparently fallen to pieces. The reconstitution of the apparatus in the daughter-cells takes place graduallj'; there is a stage during which two apparatuses are found in the second spermatocytes (fig. 31). Jordan, who figures and describes a division of the "sphere" in what he considers as the first prophase, must have mistaken the second generation for the first, as a fragmenta- tion of the idiozome does not take place, as demonstrated in my description, before the first metaphase.

If we compare the behavior of the apparatus during mitosis in the testicle of mammals with its behavior in invertebrates, as first described from silver jirepara- tions by Perroncito for Paludina, and much earlier by Platner for Helix — for literature see Duesberg (1914), pages 34-35 — we find that in mammals the process lacks the regularity exhibited in the lower forms. The apparatus behaves, as far as I can judge from the accounts given, like the idioectosomc of Stockard and Papanicolaou.

In the young spermatid the apparatus looks more like a reticulum than at any other stage. Then a differentiation takes place; a vacuole appears which increases


in size until it finally comes in contact with the nuclear membrane. In the mean- time the reticular structure is gradually collected on one side of the vacuole (fig. 32), and finally detaches itself from the latter, to pass into the protoplasmic lobe. There it appears as a granular body up to the stage represented in figure 33, which corresponds approximately to figure 10, drawn from Benda's preparations. A similar arrangement, although in a somewhat younger stage, is represented by Perroncito for the cat in liis figure 89. Later, the apparatus breaks into several smaller bodies of the same structure. Finally, only small granules are found which are eliminated with the protoplasm. Thus it appears that the apparatus does not take any part in the constitution of the ripe spermatozoon, in contradistinction to Perroncito's and Weigl's suggestion reported above, and that similar structures found in the embryonic cells either are derived from the egg's apparatus or are formed de novo.

If we compare the silver preparations with those made from material fixed with. Flemming's, Hermann's, or Benda's fluid, we find that in the resting cells (cf. figs. 5 and 28) the images are similar. The apparatus is nothing but the outer, darker- staining shell of the idiozome — Stockard and Papanicolaou's idioectosome. The points that stand out most clearly in silver preparations are: (1) the behavior of the apparatus during mitosis; (2) that it is identical with the so-called "Idiozomrest."

Finally, I wash to point out the complete similarity between my conclusions and those of Sjovall, arrived at by entirely different methods; and to add that a study of the testicle of the guinea-pig by means of Ramon y Cajal's method has given me identical results.


In my review on the apparatus of Golgi (1914) I endeavored to clear up the question of relationship between tliis element and other constituents of the proto- plasm. It would perhaps be of interest to reexamine these conclusions in the light of my present and other recent observations.

A point that I especially desired to settle, as far as possible, was that of the relationship between the apparatus and Holmgren's trophospongium. I came to the conclusion (pages 60-61) that two categories of cells should be distinguished: The neurones and the non-nervous cells with a localized trophospongium on the one side, and the non-nervous cells with a diffuse trophospongium on the other. As to the latter, the identity of both formations can be rejected without further discussion; for, while the trophospongium extends all over the cjloplasm, the apparatus of Golgi is locaUzed at one pole of the nucleus.^ As to the former, both formations appeared to me to be identical. The difficult}', however, is to reconcile Holmgren's opinion (according to which the trophospongium is in communication with the outside) with that of a large number of authors who hold that the appara- tus of Golgi is limited to the cell. I expressed the view that Holmgren must have confused the apparatus with the exogenous processes which are known to pene-

' Exception made for the lutein cells.


trate certain nerve cells (since then another example of this penetration has been reported by Ross, 1915), an opinion which is strongly supjjorted by the publica- tions of Nusbaum's pupils.

At the time I was making mj- report at Innsbruck, Holmgren (1914) pubUshed a paper in which he reaffirmed the continuitj' of the "Trophozj'^ten" and their processes with the intracellular apparatus of the gangUon cell, his conclusions being based especially upon the study of preparations made after Kopsch's method. In another paper (1915) he insists again upon the correctness of his views and expresses at the same time his dissatisfaction with the opinion given in my review, an opinion wliich I have just summarized. Regarding Holmgren's paper I wish to say this: When I undertook mj^ review I was entirely unprejudiced. My conclusions were based, first, upon a thorough study of the Literature (in fact Holmgren can not reproach me with any gaps or misrepresentation of his views — quite an achievement considering his prolixitj^ and versatihty) ; second, upon a study of preparations of my own; thii-d, upon a number of Holmgren's preparations. Details concerning the latter can not be given, as the preparations themselves have been returned to Holmgren and the notes I took at Liege are not available. I recall very clearly, however, that, notwithstanding the persuasive notes which Holmgren sent with the preparations, I failed to be convinced. My skeptical attitude towards Holmgren's theor}', therefore, is well based, and he himseK is in part responsible for it. I can not help wondering why, if he really had the facts, he did not come to Innsbruck and show his preparations instead of writing articles, for I had already informed him of my conclusions.

Since my review and Holmgren's first article (1914), a paper by Ramon y Cajal (1914) has appeared, in which the author discusses the same question. His opinion on most of the points is entirely in accord with my own. He calls the intracellular apparatus aparato tubular de Golgi-Holmgren, meaning two things: First, that the two formations are identical, a view in agreement with my own, exception being made, however, for the non-nervous cells with diffuse trophos- pongium, for which it can not hold; second, that these formations are a system of ducts, an opinion I must regard with some skepticism. Speaking of the possible connections of the apparatus with processes of trophocytes, Ramon y Cajal expresses himself as follows (p. 211):

"Lo que importa notar particularmente es que, si positivamente en ciertos elementos ganglionares de los vertebrados, existen condustos radiados para alojar apendices de los trofocitos, estos apendices no sc hallan en continuacion substaucial con el aparato de (jolgi. Ni acabamos de persuadirnos de que las cistadas celulas nutritivas representen disposicion general. A nuestro juicio, no es posible descubrir el menor resto de ellas en los epithelio!?, ni en la inmensa mayoria de las neuronas centrales, ni en los elementos del embrion di uno a dos dias, donde el aparato de Golgi esta bien diferenciado."

Another point of great interest is the question of relationship between the apparatus and the chondriosomes. In 1914 I thought it safe enough to conclude that in nerve cells (p. 18), as well as in other cells (j)]). 35-30), the apparatus is a structure different from the chondriosomes, although the possibility of genetic


connections was to be taken into consideration. Since that time three interesting papers have appeared which deal with this question. Two of them, one bj^ Deineka (1914), the other by Rina ^Slonti (1915), favor the identity of the chondriosomes and the apparatus, while Ramon y Cajal (1914) takes the opposite stand.

Deineka's conclusions are based upon his study of the connective tissue during the process of ossification. The method used was that of Golgi, with acidum arseniosum, but the duration of the fixation was reduced to about 30 minutes, instead of 6 to 7 hours. In the osteogenous tissue this method impregnates, besides a num- ber of granules and filaments which are obviously chondriosomes, some thicker filaments localized at one pole of the nucleus. These correspond to the apparatus, but for Deineka they are of the same nature as chondriosomes, of which they repre- sent only one part. In the osteoblasts very numerous short filaments and granules are accumulated on one side of the nucleus, leaving free amid them a clear space which is obviously the "sphere." In the bone-cells the image changes with the age of the cell. In none of these cases could Deineka make out a difference between an apparatus and the chondriosomes.

Ramon y Cajal, who opposes Deineka's conclusions, beUeves (p. 158) the latter's method does not impregnate the apparatus. In osteoblasts he shows it (fig. 23) precisely around that clear spot ("sphere") described by Deineka. To me it seems possible that the thick filaments localized at one pole of the nucleus, which Deineka represents in the cells of the osteogenous tissue, and perhaps in some of the bone-cells, correspond to the apparatus; while the other elements in the cells of the connective tissue are undoubtedly chondriosomes. The fact, however, that aU these bodies are or can be impregnated at the same time, does not prove that they are the same. Deineka himself (1912) has shown that under certain condi- tions of fixation the apparatus alone is impregnated, while his present method apparently preserved the chondriosomes better than the apparatus.

While Deineka's standpoint can hardly be defended, Monti's opinion appears to me worthy of consideration. This author has studied the nerve-cells of inverte- brates and vertebrates by means of the chondriosomal and silver methods. In her opinion the chondriosomes of the adult neurone and Golgi's apparatus are one and the same thing. The differences between the two sets of preparations are due to the following causes : First, that the latter is a method of impregnation, while the former consists of a progressive differentiation. Second, that in the first case thin sections, in the second thick ones, are used. Monti, however, is far from admitting the identity of Golgi's apparatus with the chondriosomes in all cases. In young nerve-cells she finds two reticula; a large one extending all over the cell-body and a small one localized at one pole of the nucleus. The former is the same as one finds in the adult cell and is formed by the chondriosomes; it is similar to the reticulum described first by Pensa, in cartilage cells, which is also formed by chondriosomes;^ the other is identical with the small reticulum described by von Bergen in the

'That the apparatus first described by Pensa (1901) corresponds to the chondriosomes can not be doubted; but I do not believe that the chondriosomes of these cells form a reticulum (see Duesberg, 1914, p. 23), even if it may appear *n nftfr the silver impregnation.


cartilage cell and other similar formations and has the same behavior during mitosis. It does not exist in the adult cell in which the centrosphere has disappeared and the reproductive activity has ceased.

If Monti were correct, all the difficulties concerning the existence of a relation- ship between the so-called apparatus of the adult nerve-cell and the cellular centers (see below) would be settled at once, as the structure that exhibits such a rela- tionship — the small reticulum — would exist only in the young cell. At the same time a revision of the nomenclature would appear necessary. Which formation should we persist in calling Golgi's apparatus — the small reticulum or the chon- driosomes? There is no doubt that the latter, or any similar term would be pre- served by the majority of authors, and we would find ourselves in the rather awk- ward position of no longer being able to speak of Golgi's apparatus in the adult nerve-cell, where it was first described. To be entirely logical, the term Golgi's apparatus should be given up altogether, as otherwise it would perpetuate an error, and some other denomination be substituted for the small reticulum whose main characteristic in the resting cell is its topographical relation with the cellular centers. I must state, however, that although Monti's opinion appears to me interesting, I am not entirely convinced that the morphological differences between chondrio- somes and reticular apparatus can be satisfactorily explained as she proposes. I have in mind a number of figures published by different authors which can hardly be reconciled with her opinion, including some of the figures given by Cowdry (1912) and Ramon y Cajal (1914).

One of the points emphasized in my review is the close topograpliical rela- tionship between the apparatus and the cellular centers, a relationship to which Ballowitz (1900) was the first to call attention, and whose importance can not be overestimated, as it is Ukely to constitute an excellent criterion. For details of the bibliography I refer the reader to my review, and especially to pages 37-39. I would call attention, however, to certain conclusions embodied therein.

The non-nervous cells can be divided into three groups: (1) A group in which the relationship between the apparatus and centers is estabhshed; (2) a group in which the same relationship appears extremely probable; (3) a group in wliich the relationship appears possible. Concerning the nerve-cells, I pointed out that this relationship is clear enough in the embryonic stages; not so, however, in the adult cell. For these one could admit that the apparatus gradually outgrows the centro- theca and surrounds the nucleus, as we know it does in certain cells. But even in the adult its complicated form does not exclude, as I suggested, such a relationship, for we know of cases in which the centrotheca undergoes somewhat similar changes. This last hypothesis does not, however, seem to be supported by recent investigations.

Hortega (1916) has recently published a number of data concerning the cen- trioles of the adult nerve-cell; the comparison of his results for the Purkinge cells, for instance, with those obtained by Sanchez (1916) for the reticular apparatus of of the same cells is certainly not in favor of any close relationsliip between appara- tus and centrioles. Monti's opinion, as indicated above, settles the difficulty, but it can not be accepted without further investigation. Ramon y Cajal (1914), for


whom the chondriosomes and apparatus are two different things, believes that "en las neuronas adultas, la despolarizacion del aparato de Golgi y su difusion peri- nuclear coincide probablemente con la atrofia y desaparicion de la esfera y cen- trosoma (p. 209)." This opinion is plausible only if, as Cajal himself indicates, it is Hmited to mammals, and further, if it is well understood that "esfera" and "centrosoma" do not mean "centrioles," a confusion which is too often made.

As to the non-nervous cells Ramon y Cajal (1914, p. 208), after extensive observations on the apparatus in embryonic, cartilaginous, epidermic, and glandu- lar cells, in odontoblasts and osteoblasts, in fat and goblet cells, subscribes whole- heartedly to my opinion:

"Desde las clasicas investigaciones de Ballowitz, efectuadas en el epitelio posterior de la cornea, y confirmadas despues para otros tejidos por numerosos autores (Negri, Pensa, Barinetti, Terni, Perocito, Deineka, etc.,) quedo perfectamente estahlecido que, en toda celula portadora de un reticule endocelular polarizado y concentrado, existe en el centre de este un hueco donde se aloja la esfera atractiva. Numerosos indicio cenfirma- torios de esta conexien hemes consignade tambien nesostros al describir la disposicien del aparato de Golgi en las grandulas, osteoblastos y osteoclastos y durante las fases onto- genicas de las neuronas y corpuscules epitelicos. En este punto suscribimos de buena gana el pensiamento de Duesberg (1914), para quien el aparato de Golgi estaria prunera- mente ligado al sistema de la esfera atractiva "

A further confirmation of the same opinion is found in Basile's researches (1914) on the modifications of the apparatus in the renal epithehum after nefrec- tomy. While normally the apparatus is located between nucleus and lumen, after the operation it has moved to the basal part of the cell, and with it the centrioles.

The relationship between the apparatus and the centers of the resting, non- nervous cell appears, therefore, more and more safely estab- hshed.^ The structure of a young resting cell could, in my opinion, be adequately represented by the accompanying schema. It should be well understood, however, that the ap- paratus is not always a reticulum. Another remark of im- portance is that, as a rule, in epithelial tissues the apparatus is located in that part of the cell which Cajal (1914) has called the "polo mundial o de relacion exterior" (see Duesberg, 1914, p. 21 .)^ As to the chondriosomes, their form and location vary.

During mitosis the apparatus falls to pieces and its frag- ments are scattered all over the cell-body, or else the primitively isolated parts of the apparatus behave in the same way. There appears to be in some cases con- siderable regularity in the shape of these fragments, their number and their distri- bution between the daughter cells. The behavior of the chondriosomes during mitosis is again variable, as I emphasized recently (1917, p. 478 etseq.; see also Meves, 1914, 1).

'Cowdry (1916) does not agree with me in my "attempt to define the apparatus in terms of its relation to the cen- trosome, because our knowledge of the centrosome itself is so deplorably inadequate (page 40)." If this applies to the non-nervous cells as well as to the nerve-cells, I certainly do not agree with Cowdry's skeptical attitude, either towards the relationship of the apparatus to the centers, or towards our knowledge of the centrosome.

2 The similar location of the centrioles is, since the researches of K. W. Zimmerman and Heidenhain, well known.


Ill older cells the apparatus maj' lose its connections with the cellular centers; for instance, in certain epithelial cells, the lutein cells, cartilage cells during the process of ossification (for Uteraturc see Duesberg, 1914, pages 40-41, and Ramon y Cajal, 1914), and finally (if we do not accept Monti's interpretation) in the nerve- cells. Then, instead of being localized at one pole of the nucleus, the apparatus surrounds this body more or less completely.

The nomenclature of these cj^toplasmic constituents is unfortunately exceed- ingly compUcated. As to the chondriosomes, I have already explained why I con- sider the term "mitochondria," taken in a general sense, as illogical and confusing (1917, pp. 469-470). The mass of cUfferentiated protoplasm which in many resting cells (young ovocytes, seminal cells, cells of the connective tissue, cartilage cells, etc.) incloses the centrioles, constitutes another object of confusion. How little some appreciate the difference between this mass and the original "sphere attractive" of van Beneden, is best illustrated by the following quotation from a recent paper by Shaffer (1917, p. 416):

"There is no essential difference between the 'attraction sphere' of van Beneden, the 'centrosphere' of Strasburger, and the 'astrosphere' of Fol and Boveri; and so fa,r as I have been able to ascertain, there is no fundamental difference between these last-named structures and the 'idiozome' of Meves. One thing is clear — that these structures all refer to the achromatic substance of the spindle, situated at the poles and usually inclosing the central corpuscles."

This is, of course, entirely incorrect. The idiozome has nothing to do with the "achromatic substance of the spindle situated at the poles." The necessity of dis- tinguisliing between the mass of protoplasm which surrounds the centrioles in many resting cells (idiozome or centrotheca) on one side, and the attraction sphere of van Beneden on the other, has been emphasized long ago by von Erlanger and by INIeves. (For hterature see Meves, 1897 and 1914, 1). What term, then, should we use? I agree with those authors who beUeve that the term "sphere" should be rejected in order to avoid any confusion with the "attraction sphere" or "astro- sphere" of the dividing cell. To avoid this confusion Meves has proposed the term idiozome, and later, in order to emphasize the relationship of the body with the centrioles, the term centrotheca. To the first term Regaud has objected (the same objection could be made to "centrotheca") for the reason that the relationship with the centers does not persist during mitosis nor (in seminal cells) during sper- miogenesis. He proposes the term idiosome, which Stockard and Papanicolaou have adopted. There is no doubt that this name, although rather vague (or per- haps because of tliis), has much in its favor, if only investigators could agree upon it. It should be added that the term "centrosome" is also used in a very loose manner, and a great number of authors unfortunately take it as a synonym for "centrioles," Shaffer, for instance.

Still more compUcated and confused is the nomenclature of that body which surrounds the idiozome. It is now well established that it corresponds to what was formerly called "Xebenkern" in the seminal cells of Helix, by Platner (Hermann's "Archoplasmaschleifcn"), to some of Van der Stricht's "pseudochromosomes," to


Ballowitz's "Centrophormien," and to Heidenhain's "Centralkapsel." More com- plete bibliographical data will be found in Duesberg (1914) and also in Kuschake- witsch and Terni. In recent times the same body has been called frequently "Golgi's apparatus;" other denominations are "von Bergen's reticulum" (Monti), "forma- zioni periidiozomiche (Terni), "dyctyosomes" (for the fragments during mitosis, Perroncito), "idioectosome" and "idiophthartosome" (Stockard and Papanicolaou). Kuschakewitsch proposes calling it "Spharotheca" and "Spharosomen" (during mitosis), while the complex formed by the "Spharotheca" and the "idiozome" is designated as "Statosphare."

Since, as noted above, I have rejected the term "sphere," I can not accept Kuschakewitsch's nomenclature. To these terms, which emphasize the relation- ship between the discussed body and the centers or the idiosome (Centralkapsel n, Centrophormien, formazioni periidiozomiche, idioectosome), the same objection may be raised as to the terms "idiozome" and "centrotheka," for such a relation- ship is not durable. Other denominations, such as "Nebenkern," "von Bergen's reticulum," "idiophthartosome," etc., are obviously bad. Personally, I would prefer the term "Golgi's apparatus," or rather "Golgi's intracellular apparatus" (Golgischer Binnenapparat, as proposed by Nusbaum), which is already very widely used and does not prejudge any special morphological appearance. The only difficulty is the one emphasized above, when discussing Monti's conclusions.


Addison, W. H. F., 1916. The C.olgi apparatus in the cells of the distal glandular portion of the hypo- physis. Proc. Araer. Assoc. Anat. Anat. Rec, 1917, vol. 11, p. 317.

Alexeieff, a., 1916. Mitochondrics chez quelques pro- tistes, Mitochondrics, glycoplastes ct adipoplas- tes. Caractftres g(5n(5raux des mitochondries. C. rend. Soc. Biol., t. 79, p. 1076.

Ballowitz, E., 1900. Eine Bemerkung zu dem von Golgi und seinen Schiilprn beschriebcnen "Appa- rato reticolare intomo" dor Ganghen- und Driisenzellen. Anat. Anz., Bd. 18, p. 177-181.

Bano, I., und E. Sjovall, 1916. Studien iiber Chon- driosomen unter normalen und pathologischen, Bedingungen. Bcitr. zur path. Anat. u. aUg. Path, (quoted from A. Guillierraond).

B.\siLE, G., 1914. SuUe inodifieazioni deU'apparato reticolare intcmo di Golgi neU'epitclio renale di animali nefrectomizzati. Internat. Monatschr. f. Anat. u. Phys., Bd. 31, p. 1-9.

Benda, C, 1897. Neuere Mitteilungen iiber die His- togenese des Saiigetierspermatozoon. (Verhandl. d. physiol. Gesellsch., Berlin) Arch. f. Anat. u. Physiol. Physiol. Abt., p. 406^14.

1906. Die Spermiogencso der Marsupialier.

•Semons Zoologische Forschungsreisen in Aus- tralien. Jena, Fischer (not available).

BiRCK, L., 1914. Demonstration. Verhandl. d. anat. GeseUsch. Innsbruck, 1914, p. 279.

Bonnet, R., 1912. Lehrbuch der Entwicklungsgesch- ichte. Zweite Aufl. Berlin, Parey.

Castro, F. de, 1916. Nota sobre la disposicion de aparato reticular de Golgi en los botones gusta- tivos. Trab. Lab. Invest, biol. Universidad Madrid, vol. 14.

Cowdry, E. v., 1912. The relations of mitochondria and other cytophusmic constituents in spinal ganglion-cells of the pigeon. Internat. Monatschr. f. Anat. und Physiol., Bd. 29, p. 473-504.

1916. The structure of the chromophile cells of

the nervous system. Contributions to Embry- ology, (Carnegie Inst, of Washington), No. 11, p. 27-43.

DilNEKA, D., 1912. Der Netzapparat von Golgi in einigen Epithel- und Bindegeweszellen wiihrend der Ruhe und wiihrend der Teilung derselben. Anat. Anz., Bd. 41, p. 289-309.

1914. Beobachtungeii iiber die Entwicklung des

Knochengewebes mitteis der Versilberungsme- thode. Anat. Anz., Bd. 46, p. 97-126.

DuBREUiL, G., 1913. Lc chondriome et le dispositif de I'activiti; sdcrdtoire aux diffr'jrents stades du ddveloppement des 616ments cellulairas de la lignde connective, descendants du lymphocyte (globules blancs mononucl6(5s de la lyinphe et du sang, celhiles connectives, cartilagineuses et osseuses). Arch. d'Anat. raicr., t. 15, p. 53.

DoE-SBERfj, J., 1908. La spermiogen^e chez le rat (Mus decuraanus, Pall., var. albinos). Arch. f. Zellf., Bd. 2, p. 137-180.

1910. Nouvelles recherches sur I'appareil mito- chondrial des cellules sdminalcs. Arch. f. Zellf., Bd. 6, p. 40-139.

DuESBERG, J. 1912. PIasto.somen, "Apparato reticolare intemo" und Chromidialapparat. Ergebn. d. Anat. u. Entwicklungsgesche (Anat. Hefte, 2. Abt.), Bd. 20, p. 567-916.

1914. Trophospongien und Golgischer Binnen-

apparat. Verhandl. d. anat. GeseUsch. Inns- bruck, 1914, p. 11-80.

1915. Recherches cjrtologiques sur la fdcondation

des ascidiens et sur leur ddveloppement. Con- tributions to Embryology (Carnegie Inst, of Washington), No. 8, p. 35-70.

1917. Chondriosomcs in the cells of Fish- embryos. Amer. Jour. Anat., vol. 21, p. 465-489.

1918. Chondriosomcs in the testicle-cells of

Fundulus. Amer. Jour. Anat., vol. 23, p. 133.

Faure-Fr6miet, E., 1913. Le cycle germinatif chez 1' "Ascaris megalocephala." Arch. d'Anat. micr., t. 15, p. 435-757.

GoLDSCHMiDT, R., 1909. Das Skelet der Muskelzelle von Ascaris, nebst Bemerkungen liber den Chromidialapp.arat der Metazoenzelle. Arch, f. Zellf., Bd. 4, p. 81-119.

Gregoire, v., 1909. Les ph6nomenes de I'etape synap- tique reprdsentent-ils une caryocinese avortde? La Cellule, t. 25, p. 85-99.

GuiLLiERMOND, A., 1917. Sur la nature et le role des mitochondries des cellules v6g(5tales. C. rend. Soc. Biol., t. 80.

Hartmann, C. G., 1916. Studies in the development of

  • the opossum Didelphys virginiana L. I. History

of the early cleavage. II. Formation of the blastocyst. Jour. Morph., vol. 27, p. 1-83.

Holmgren, E., 1914. Trophospongium und Apparato reticolare der spinalen Ganglienzellen. Anat. Anz., Bd. 46, p. 127-138.

1915. Die Trophospongien spinaler Ganglien- zellen. Arkiv for Zoologi., Bd. 9.

Hortega, P. DEL Rio, 1914. Alteraciones del sistema ncrvioso central en un caso de moquillo de forma paralitica. Trab. Lab. Invest, biol. LInversid. Madrid, vol. 12.

1916. Estudios sobre el centrosomo de las

celulas nerviosas y neuroglica-s de los vertebrados, en sus formas normal y anormales. Trab. Lab. Invest, biol. Universid. Madrid, vol. 14.

Jordan, H. E., 1911. The spermatogenesis of the opossum (Didelphys virginiana) with special reference to the accessory chromosome and the chondriosomcs. Arch. f. Zellf., Bd. 7, p. 41-86.

1914. The spermatogenesis in the mongoose.

Papers from the Tortugas Lab. of the Carnegie Inst., vol. 5.

Koltzoff, N. K., 1908. Studien iiber die Gestalt der Zelle. 2. Untersuchungen iiber das Kopfskelett des tierischen Spcrmiums. Arch. f. Zellf., Bd. 2, p. 1-65.

v. KoRFF, K., 1902. Zur Histogenese der Spcrmien von Phalangista vulpina. Arch. f. mikr. Anat., Bd. 60, p. 233-260.

KoscHAKEWiTSCH, S., 1913. Studien iiber den Dimor- phismus der inannlichen Goschlechtseloniente bei den Prosobranchia. Arch. f. Zellf., Bd. 10, p. 237-323.




Maximow, a., 1916. 1. Sur les m^thodes de fixation et

de coloration des chondriosomes. C. rend. Soc.

Biol., t. 79, p. 464. 1916. 2. Sur la structure des chondriosomes.

C. rend. Soc. Biol., t. 79, p. 465. Meves, Fr., 1897. Zellteilung. Ergebn. d. Anat. und

Enticklungsgesch. (Anat. Hefte, 2. Abt.), Bd. 6,

p. 284-397. 1899. Ueber Struktur und Histogenese der

Samenfaden des Meerschweinchens. Arch. f.

mikr. Anat., Bd. 54, p. 329-401. 1914. 1. Die Plastochondrien in dem sich teilen-

den Ei von AscarLs megalocephala. Arch. f.

mikr. Anat., Bd. 84, Abt. 2, p. 89-110. 1914. 2. Was sind die Plastosomen? Antwort

auf die Schrift gleichen Titels von G. Retzius.

Arch. f. mikr. .\nat., Bd. 85, Abt. 1, p. 279-302. Monti, R., 1915. I condriosomi e gh apparati di Golgi

neUe cellule ner\'ose. Arch. ital. di Anat e dl

Embriol., vol. 14, p. 1^5. Moore, J. E. S., 1894. Some points in the spermato- genesis of mammalia. Intern. Monatschr. f.

.\nat. u. Phvsiol., Bd. 11, p. 129-165. Moore, J. E. S., and Walker, C. E., 1906. The maiotic

process in mammaha. Liverpool Institute for

Cancer Research (first report on the cytological

investigation of cancer). Thompson, Yates &

Johnston Laborat., vol. 3, p. 75-88. NiEssiNG, C, 1897. Die Beteiligung von Centralkorper

und Sphare am Aufbau des Samenfadens. Arch.

f. mikr. Anat., Bd. 48, p. 111-142. 1902. Kurze Mitteilungen und Bemerkungen

iiber Spermatogenese. Anat. Anz., Bd. 22, p.

112-118. Oliver, J. R., 1913. The spermiogenesis of the Pribilof

fur-seal (Callorhinus alascanus J. and C). Amer.

Jour. ,\nat., vol. 14, p. 473-496. Pappenbeimer, a. M., 1916. The Golgi apparatus.

Personal obser\-atioiis and a review of the

literature. Anat. Rec, vol. 11, p. 107-148. Pensa, a., 1901. 1. Osservazioni sulla struttura delle

cellule cartilaginee. Boll. Soc. Med.-Chir. Pavia,

p. 199-205. 1901. 2. Demonstration von Praparaten iiber

einen Netzapparat in den Hyalinknorpelzellen.

Verhandl. d. anat. Gesellseh., Bonn, p. 205. 1901 . 3. Observations sur la structure des cellules

cartilagineuses. C. rend. Assoc. Anat., Lyon, p. 185. 1915. Ancora sulla struttura della ceUula car-

tilaginea (a proposito del Referat di J. Duesberg

"Trophospongien und Golgischer Binnenap-

parat"). Anat. Anz., Bd. 47, p. 627-631. Perroncito, a., 1910. Contribution k I'^tude de la

biologic cellulaire. Mitochondries, chromidies et

appareil r^ticulaire interne dans les cellules

spermatiques. Le ph^nomene de la dictyokinese.

Arch. ital. de Biol., vol. 54, p. 307-345. R.VMON T C.UAL, S., 1912. Formula de fijacion para la

demonstracion facil del aparato reticular de

Golgi y apuntos sobre la disposicion de dicho

aparato en la retina, en los nervios y algxmos estados patologicos. Trab. Lab. Invest, biol. Universidad Madrid, vol. 10, p. 209. 1014. Algunas variaciones fisiologicas y pato-

logicas del aparato reticular de Golgi. Trab.

Lab. Invest, biol. Universidad Madrid, vol. 12.

Regaud, Cl., 1908. Sur les mitochondries de I'dpitheUum seminal. 2. Les mitochondries de la lign^e spermatique. C. rend. Soc. Biol., t. 65, p. 607.

1910. Etudes sur la structure des tubes s<5mini-

fSres et sur la spermatogenese chez les mammi- f6res. Arch. d'Anat. micr., t. 11, p. 291-4.33.

Retzius, G., 1902. Ueter einen Spiralfaserapparat am Kopfe der Spermien der Selachier. Biol. Unter- such., N. F., Bd. 10, p. 61.

1906. Die Spermien der MarsupiaUer. Biol.

Untersuch., N. F., Bd. 13, p. 77-86.

1909. Die Spermien von Didelphys. Biol.

Untersuch., N. F., Bd. 14, p. 12.3-126.

1910. Weit«re Beitrage zur Kenntnis der Sper- mien mit besonderer Beriicksichtigung der Kem- substanz. Biol. Untersuch., N. F., Bd. 15, p. 6.3-82.

1914. Was sind die Plastosomen? Arch. f. mikr.

Anat., Bd. 84, Abt. 1, p. 17.5-214.

RoMEis, B., 1912. Beobachtungen iiber Degeneration- serscheinungen von Chondriosomen. XachUnter- suchungen an nicht zur Befruchtung gelangten Spermien von Ascaris megalocephala. Arch. f. mikr. Anat., Bd. 80, Abt. 11, p. 129-170.

Ross, L. S., 1915. The trophospongium of the ne^^'e- cell of the cray-fish (Cambarus). Jour. Comp. Neurol., vol. 25, p. 523.

Sanchez, M., 1916. Recherches sur le reseau endoceUu- laire de Golgi dans les cellules dc I'ecorce du cervelet. Trab. Lab. Invest, biol. Univers. Madrid, vol. 14.

Schreiner, K. E., 1915. Ueber Kern- und Plasmaver- anderungen in Fettzellen wahrend des Fettan- satzes. Anat. Anz., Bd. 48, p. 145-171.

Sh.atfer, E. L., 1917. Mitochondria and other cj-to- plasmic structures in the spermatogenesis of Passulus comutus. Biol. Bull., vol. 32, p. 407-434.

Sjovall, E., 1906. Ein Versuch, das Binnennetz von Golgi-Koijsch bei der Sixjrmato- und Ovogenese zu homologisieren. Anat. Anz., Bd. 28, p. 561-579.

Speclale, F., 1914. Sulla fine struttura delle cellule endoteliali dell'endocardo e delle cellule che tappezzano le fenditure di Henle. Arch. f. ZeUf., Bd. 12, p. 513-515.

Stockard, C. R., and G. N. P.apanicolaou, 1916. The morphological changes of the idiosome during spermatogenesis of the guinea-pig. Proc. Amer. Assoc. Anat., Anat. Rec, vol. 11, 1917, p. 412. {See also Bibhographic cards, Wistar Institute, No. 155, Dec 1917.)

Terni, T., 1914. Condriosomi, idiozoma e formazion, periidiozomiche nella spermatogenesi degU anfibi (Ricerche sul Geotriton fuscus). Arch. f. Zellf. Bd. 12, p. 1-96.

VoiNov, D., 1916. Sur une formation juxta-nucl^aire dans les Elements sexuels du Grj-llotalpa vulgaris, caduque 4 la fin de la spermiogenese. C. rend. Soc. Biol., t. 79, p. 542-544.

Wassilieft, a., 1907. Die Spermatogenese von Blatta germanica. Arch. f. mikr. Anat., Bd. 70, p. 1—40.

Weigl, R., 1912. Vergleichend-zytologische Untersu- chungen iiber den Golgi-Kopschschen Apparat und dessen Verhaltnis zu anderen Strukturen in den somatischen Zellen und Gieschlechtszellen verschiedener Tiere. Bull, intemat. Acad. Sc. Cracovie. Cl. Sc. math. st. nat., p. 417-448.


All figures were outlined with a Zeiss camera lucida, at the level of the stage of the microscope. Zeiss apochromatic immersion 2 mm., ocular 12. Artificial hght (gas).

Plate 1.

Fixation, Benda; stain, Benda.

Fio. 1. A Scrtoli-cell. The three big granules in the basal part, near the nucleus, arc fat-droplets.

Fio. 2. Resting spermatogonium.

Fig. 3. Dividine spermatogonium, metaphase.

Fio. 4. First spcrmatoryto, paehytrn sttidium

Fig. 5. First spermatocyte in the second phase of the growth-prrio<l.

Fig. 6. Dividing first spermatocyte, metaphase. The small granules to the right of the spindle and belo\v the equator

are fat-<lroplets. Figs. 7, 8, 9. First period of spcrmiogcncsis.

Fig. 7. Young spermatid. The granules in the upper left corner are fat-droplets. Fio. 8. More advanced stage.

Fio. 9. More a<lvanced stage. Note the small acrosoine on anterior part of the nucleus. Fio8. 10, 11, 12. Second period of spermiogenesis.

Fio. 10. The head has assumed the form of an egg-shaped body, whose long axis is perpendicular to the axial filament. Flo. 11. The head begins to assume its definite form. First indication of disjointing of headcap. Note an idiozomic

remnant near the opening of the caudal tube, on the left. Fig. 12. Further stsige of the disjointing of the he;idcap. Idiozomic remnant near the opening of the caudal tube.

In the upper left corner, first appearance of "ttngierbare Korner." Note the sheath developing around

the axial filament in the future main piece. Fios. 13, 14, 15,a and h. Third period of spermiogenesis. Flo. 13. Migration of the ring. The granule right below the ring is a mitochondrium, not a centriole. In the proto-

pliismic lobe, remnant of the idiozome (to the right of the ring), mitochondria and "tingierbare Korner,"

darkly stained and fused together. Fig. 14. Deposit of mitochondria on the axial filament in the middle piece. "Tingierbare Korner" to the right

of the main piece. Flo. 15a. Elimination of protoplasm with some mitochondria, idiozomic remnants (?) and "tingierbare Korner"

covering part of the head and the anterior part of the middle piece. P^G. 156. Transverse section through the region of the middle piece showing the peripheric disposition of the chon-

driosomes which are going to be eliminated. To the right, cross-section of the middle-piece; to the

left and somewhat below, "tingierbare Korner;" in the lower right corner, idiozomic remnants (?). Fig. 16. Fourth period. The spiral filament is formed. Between the spermatozoa, expelled into the lumen, and

the seminal epitheUum, a layer of residual bodies with big granules and vacuoles.

Plate 2.

Fios. 17-24. Fixation, Regaud; stain, acid fuchsin-methyl green.

Fio. 17. Spermatid in a stage somewhat more advanced than the cell represented in figure 8. Note the difference in size after the action of two different reagents.

Fios. 18-21. Third period of spermiogenesis. Successive stages of the deposit of chondriosomes on the middle piece and the elimination of the residual body, with some chondriosomes.

Fios. 22-24. Fourth period of spermiogenesis. Successive stages of the formation of the spirale.

Fio. 25. From a smear of sperm from the epididymis. Fi.xation vapors of osmic acid. Stain, Benda. Copulating spermatozoa.

Fios. 26-34. From preparations after Cajal and handled further as described in the tex-t. Counterstain: methyl- green for figures 26-30, 33 and 34; EhrUch's heraatoxyhn for figures 31 and 32.

Fio. 26. Spermatogonium with impregnation of mitochondria (?) and of the intracellar apparatus.

Fio. 27. Two first spermitocytcs in the first phase of the growth-period. To the left a perpendicular section, to the right a tangential section.

Fio. 28. First spermatocyte, second phase of the growth-period.

Fio. 29. First division, metaphase.

Fio. 30. First division, anaphase.

Fig. 31. Second spermatocyte, with two apparatuses.

Fio. 32. Spermatid in a stage of development corresponding approximately to figures 8 and 17.

Fio. 33. Further stage of <icvclopment of the spermatid, corresponding approximately to figure 10.

Fio. 34. Interstitial cell.