But little detail need be given regarding the general structure of the testis in the pigeon, for it does not differ essentially from that of well known mammalian forms. The spermatogenesis of the rat is especially well known through the researches of Von Ebner {1. Von Ebner, V: "Zur Spermatogenese bei den Saugethieren." Arch. f. mikr. Anat., XXXI., 1888.}, Lenhossek {2. Loc. cit., p. 5.} and others. In the pigeon the testes are two elongated bean-shaped organs varying much in size with age and with individuals. The germinal cells are laid down in a very great number of convoluted tubules which wind back and forth throughout the interior of each testis and make up its main bulk. The tubules are much slenderer and more delicate than those of the rat, however, and can not be isolated so readily for study.

    In the arrangement of the germ cells in the tubule, there is no departure from that of the other warm-blooded animals. Next the walls are the spermatogonia or parent cells (Fig. 1, sg.), which by growth and division give rise to the various generations of germ cells lying inward toward the lumen. The adult spermatozoa are formed through the final transformation of the spermatids, or cells produced by the last division, which lie nearest the center of the tubule. The ripe spermatozoa pass out through the lumen of the tubule and into the ducts which lead to the exterior. As in many forms, the spermatozoa attach themselves to a supporting cell (Fig. 1, s.) for a period before their complete maturation and ejection from the testis.

    The usual four phases or types of the germinal cells are recognizable, viz : (1) spermatogonia (Fig. 1, sg.), a more or less regular layer of cells lying next to the membrane or wall of the tubule, each cell of which through division gives rise to two new cells. One or both of these may increase in size and become (2) primary spermatocytes (Fig.1, scy₁), or remain in the layer and continue as spermatogonia. The primary spermatocytes, after some interesting changes, divide to form, (3) the secondary spermatocytes (Fig. 1, scy₂), which divide again shortly to give rise to (4) the spermatids, through the transformation of which the spermatozoa are developed. The number of chromosomes in each type, as seen at the equator of the spindle before division is, in the spermatogonia 16 loops, in primary spermatocytes 8 rings or vesicles, and in secondary spermatocytes 4 rings or vesicles. It is not improbable that the socalled ring chromosomes, at least under certain conditions, are really small vesicles bearing the chromatin in the form of an internal shell. To decide this matter has proved the most tantalizing point in the whole investigation. Such a vesicle could easily take on the appearance of a ring when viewed by transmitted light and from the fact that from any point of view many of the chromosomes present in a given field appear to be rings one is inclined to accept this idea of their vesicular nature. On the other hand the chromosomes, of the primary spermatocytes at least, undoubtedly originate as rings although it is possible for them to become vesicles as they condense. What appears to be the elliptical edge of a thick ring is presented occasionally, and this has decided the writer to use the term ring in preference to vesicle, although it should be borne in mind that the apparent rings may be vesicles.

    Closely connected with the germinal cells proper are the supporting cells or Sertoli cells mentioned above, to which the spermatozoa become attached at one period of their transformation (Fig. 1, s.). They are irregularly disposed among the other cells and frequently do not lie in the spermatogonial- layer.

    It is difficult to get preparations showing only certain typical stages, for while certain kinds of cells may be present in one view, in another, some of exactly the same phases may appear side by side with other cells in entirely different stages of development. There seems indeed to be no very definite sequence of stages in a given part of a tubule as we find, for instance, in the rat. To be more specific, one may find spermatogonia, Sertoli cells with spermatozoa attached, and spermatids in one view. In another, where the spermatogonia are apparently in the same condition as before, primary and secondary spermatocytes may be present. Where one finds the primary spermatocytes, they are usually arranged in clumps, the individuals of which are at similar points of development. The same is true for the secondary spermatocytes. Fig. 2 shows both kinds of spermatocytes undergoing division in the same field of view.

    In the following account, the spermatogonia are considered first and the various other types are then taken up in the order of their succession in development into spermatozoa.

The spermatogonia lie in a more or less regular layer next to the tubule wall. Not infrequently the individual cells have the appearance of having been crowded out of place so that the layer appears irregular or double. There is considerable variation in the appearance and some variation in the size of the cells at different times. In early stages they are far apart and have small nuclei which are oblong with the long axis parallel to the tubule wall. The cells themselves are generally flattened more or less so that their long axes coincide with those of the nuclei (Fig. 1, sg.). In still earlier stages, the cell boundaries are very indistinct or seemingly absent, and gaps frequently intervene between the individual spermatogonia, so that they appear to have been left behind from a preceding set, or to have recently settled in their present position. Some of them resemble very much the wandering cells seen outside the tubule.

    In later stages, the spermatogonia are crowded together until they become more or less columnar in shape (Fig. 2). The nuclei increase in size and become very distinct. They are round or slightly oval in shape. When oval, the long axis is, as a rule, perpendicular to the tubule wall. In the closely packed cells, the nuclei come to lie eccentrically as they enlarge, and a condensation of the cytoplasm, or mass of sphere substance (idiozome of Meves), makes its appearance and gradually increases in size till it becomes a well defined area (Fig. 3, i). At first it seems to be entirely granular in nature, but later displays, at least in part, a fibrous or reticular structure. Near the center of the mass is generally a clear area in which a minute centrosome is discernible. As the sphere itself becomes better defined, the centrosome appears larger and the clear area more distinct (Fig. 3).

The primary spermatocytes or cells of the second spermatogenetic period originate, as has been indicated, from the cell-products of the last spermatogonial division through a process of growth. All transitional sizes between that of the spermatogonia and the comparatively large spermatocytes may be seen in practically the same order as that described by Lenhossek {Loc. cit., p. 5.} for the rat. In the new cells, the chromatin passes into the resting condition and an increase in bulk of both nucleus and the cytoplasm begins. The chromatin is arranged much as it was in the resting condition of the spermatogonia except that there is perhaps less distribution of the chromatin granules along the periphery of the nucleus and the linin fibres are coarser (Fig. 6).

    The sphere first appears as an indistinct granular crescentic area closely applied to the nucleus, with the horns of the crescent so extended as to enclose more than half of the nuclear surface (Fig. 6, i.). As the young spermatocyte grows, the sphere also increases in size and become more and more rounded.

    From an early stage a minute centrosome is visible in the midst of the sphere substance. It is surrounded by a clear looking area which becomes more pronounced as the sphere grows older. Thus the developing cell gradually acquires characteristics of size, shape, and general appearance that differ considerably from those of the previous generations.

    At no point in this metamorphosis could I distinguish any nuclear element that might be identified positively as an achromatic nucleolus. Such a body, in fact, seems to be wanting in all stages of the spermatogenesis. One or more bodies of varying size were often present but there was no evidence whatever to show that they differed in any way from the ordinary linin material. Certainly no such well marked object exists as Moore {1. Moore, J. E. S.  On the Structural Changes in the Reproductive Cells During the Spermatogensis of Blasmobranchs. Quart. Journ., XXXVIII, 1893.} figures in the testicular cells of the elasmobranchs, or Lenhossek {2. Loc. cit., p. 5.} in those of the rat.

    In the growing cell the cytoplasm has the appearance of a granular mass throughout which fine ramifying fibres can be detected occasionally. The sphere at times exhibits a delicate fibrous structure but this was not observed frequently and may have been due simply to the action of the reagents.

As soon as the growth of the primary spermatocyte is completed, the nucleus, which is voluminous and apparently turgid with sap, undergoes a curious change. There is a marked drifting of the nuclear contents to that side in contact with the sphere (Fig. 7). This phenomenon is the same that Moore recognized in the maturation of Elasmobranch sperm cells and which together with the accompanying fusion of pairs of chromosomes he has designated by the term synapsis. A similar union of the chromosomes occurs in the, pigeon, for at the ensuing division only eight chromosomes appear. Thus, a pseudo-reduction in Ruckert's {3. Ruckert, J: Zur Eireifung bei Copepoden: An. Hefte, 1894.} sense occurs.

    In the collapsed and tangled mass of the nuclear elements that have passed to one side of the nucleus, no distinct arrangement can be recognized. The remainder of the nucleus seems to be absolutely devoid of any contents which will react to stains. Often one or more stiff bristle-like fibres or strands of what appears to be the linin material project out from the mass across the vacant area of the nucleus. The ends of shorter fibres are seen sometimes bristling out from all sides of the heap.

    If saffranin and gentian violet are used for staining, one observes granules and clumps of chromatin taking a deep red stain, mixed indistinguishably with a bluish or violet mass of linin. No distinct chromosomes as observed by Montgotnery {4. Loc. cit., p. 12.} in Pentatoma could be seen. Chromosomes could be distinguished in fact, only after the nucleus had passed through a spirem stage, when, as has been noted already, the chromatin appears laid down in half the original number of chromosomes; that is, the new chromosomes are of the bivalent type.

    As to the meaning of synapsis, we seem to be in almost total ignorance. Few suggestions seem to have been proposed to explain it beyond that of Moore {. 1. Loc. Cit., P. 16.} who mentions the possibility that it may mark an abortive attempt to bring about the formation of a tailed spermatozoon, due to a sort of historical reminiscence which marks this as a once final stage of development in the remote ancestral sex cells. It seems to the writer more probable that the fusion is a conjugation of maternal and paternal chromosomes, though why such is necessary, is not apparent. Some rather significant facts appear which may furnish a possible clew to the nature of some of the changes which occur during the period of nuclear collapse. A very careful determination of the size and shape of the nucleus and sphere substance respectively was made in a number of cases just before and during the contraction of the contents to one side. Immediately before the change, the large nucleus is nearly spherical, or if oval, its long axis may lie in any direction with reference to the sphere. After the contraction has occurred, the nucleus is always slightly flattened toward the sphere; that is, the long axis of the oval lies parallel to the sphere. Moreover, at times the nuclear membrane is a trifle wavy or uneven. The average diameter of the nucleus before is slightly greater than during the period of apparent collapse. A large number of measurements of the sphere reveals the fact that it is, on an average, almost a third larger after the chromatin has massed nearer it than before. This, together with the fact that the drifting of the nuclear contents is always toward the sphere, and that the nuclear membrane is indistinct in that region, leads to the conclusion that there has been a discharge of part of the nuclear material into the sphere. Then, too, occasional fragments of chromatin are seen scattered about in the sphere, indicating that they have been carried out, perhaps during a discharge of the material from inside the nucleus. Again, the sphere takes stain more intensely and appears more dense as soon as the above change has occurred and there are visible frequently minute star-like radiations of a lighter staining substance which spring out from the clear area around the centrosome (Fig. 7, i.). The radiations remind one of the clear-looking streaks which accompany the formation of the pulsating vaculole in some of the infuseria.

    What the interpretation of the above facts is, can be answered at present only conjecturally. The first thought that presented itself was that the centrosome originally lay inside the nucleus and that the above occurrence was a process of extrusion by which it reached the sphere. But careful examination of the cells reveals the presence of the centrosome in the sphere long before synapsis. Indeed it seems to originate from the sphere substance itself and to increase in size as the, sphere develops.

    It is possible that the passage of substance from the nucleus to the sphere may have something to do with the formation of the spindle, for immediately after synapsis the centrosome divides and the spindle fibres appear. It may be that some substance is required from the nucleus before the extremely heavy spindle of this division period can be constructed.

    The sphere in the germinal cells of the pigeon seems to be simply a very plastic area of the cytoplasm especially favorable for building up into such structure as spindles, asters, centrosomes, the apex of the head, and the axial fibre of the spermatozoon tail. That such a role on the part of the sphere is of very general occurrence in germ cells is so well known to workers in spermatogenesis as not to require specific reference. In the ovarian egg the sphere (idiozome) seems to be employed likewise in diverse ways. In one case of abnormal ovarian structure in the pigeon (Guyer {1. Guyer, M. F: Ovarian structure in an Abnormal Pigeon.--Zool. Bull. II. No. 5, 1899.}) where numerous multiple eggs were forming and extensive degeneration going on, the sphere seemed actively employed in building up cell membranes as well as in the formation of fibres. The large vacuoles which were present in many eggs, moreover, originated invariably in the center of the sphere. In other words, the sphere is, then, only an active condition of the cytoplasm in a given region.

    The greatest visible change is in the nucleus itself at this period of disturbance in the primary spermatocyte. The chromatin drifts to one side and during its stay must undergo some sort of a profound alteration in its arrangement, for when the chromosomes reappear they are the double or bivalent type. It is possible, however, that we emphasize this phase of the phenomenon simply because it is the more visible through the great staining capacity of the chromatin. That the pseudoreduction may occur in reproduction cells without the accompanying phenomenon of intra-nuclear collapse is well seen in the primary spermatocytes of the rat, where such a condensation seems never to occur, although only half the original number of chromosomes appears at the time of division.

    In one or two instances it seemed very probable that during the internal changes of the nucleus quite enough chromatin to form a complete chromosome or more was ejected into the cytoplasm. At least two distinct examples were seen where there was a very tiny spindle connected with chromatin material lying in the cytoplasm at considerable distance from the main spindle. In one case, there was a single chromosome attached to the dwarf spindle, and in a second there was one large and one very small chromosome. Unfortunately the number of chromosomes at the equator of the main spindle could not be determined in these instances. It is interesting to note in this connection that the chromatin apparently determines the development of the spindle, for where we have enough chromatin accumulated to constitute a complete chromosome, it would seem that a spindle, develops. Juel {1. Juel, H. O: Die Kerntheilungen in den Pollemutterzellen.-Jahrb. wiss. Bot., XXX, 1897.} in a recent paper records an apparently similar occurence in the pollen-mother-cell of Hemerocallis. On the other hand, the possibility exists, of course, that the small spindle may have been simply a readjustment of a part of the main spindle together with a chromosome which had become separted from it in some manner.

As soon as the phase of synapsis is past, the nucleus undergoes a process of reconstruction and the contents are once more redistributed throughout the whole internal area. It does not pass back into a resting condition, but proceeds at once to the development of the spirem. In reconstruction, the linin fibers are the first to spring out from the eccentric mass and gradually extend throughout the nucleus. The chromatin soon follows until a large well developed spirem is formed (Fig. 8). As in the spermatogonia, the chromatin substance seems to be laid down in the form of granules which are imbedded in the linin fibres. The granules spread until they are finally about equally distributed throughout the linin. In the early stages of spirem formation, chromatin nucleoli or net knots are present, but they gradually resolve into granules which spread until all portions of the spirem are approximately equalized in the supply of chromatin. Looked at from certain views, the spirem seems to be a continuous thread, but from others it appears as a number of closed loops. Whether these loops exist as such from the first could not be determined.

An adult spermatozoon as it exists in the vas deferens is shown in Fig. 27. The head is long and narrow and is intensely stained by nuclear dyes. Favorable preparations show faintly a spiral or vesicular arrangement of part of the chromatin in the interior. At the anterior end of the head is a slender finepointed head-spine. The head at its posterior end connects directly with the long cytoplasmic tail. No middle piece is visible. The tail and head-spine are very difficult to observe accurately, and but little of the details of their structure could be worked out. The only satisfactory way to gain a knowledge of the spermatozoon at all, in fact, is through a study of its development.

    Ballowitz {1. Ballowitz, E: Untersuchungen ueher die Struktur der Spermatozoon Arch. f. mikr. Anat., XXXII., 1888.} in his description of the spermatozoa of birds is somewhat in error in regard to the spermatozoa of pigeons (cf. his drawing Fig. 91,, Plate XVII.). He has seemingly mistaken the head for the tail.

    Before the spermatid begins its transformation, it is quiescent for a considerable period of time if one may judge by the large number that is nearly always observable in any section. When once the change begins, it is probably completed in a relatively short time. The spermatid, in the resting condition after the last division, has a very characteristic appearance (Fig.17). The nucleus is round and bears most of the chromatin at its center in the form of a large sphere which from optical section seems to be hollow. Fibers of linin radiate from the central mass to the nuclear membrane which is lined by a thin shell of chromatin. Not infrequently several small spheres of chromatin may persist at the center of the nucleus instead of one large one. The centrosome lies to one side of the nucleus imbedded usually in the sphere. The cell membrane is at times very indistinct.

    The first change to be observed is in the centrosome. It divides and moves out of the sphere and further away from the nucleus. A vacuole-like area persists in the sphere, marking the spot where the centrosome lay (Fig. 18, hs.). The two new centrosomes move apart for a short distance, but remain connected by a barrel-shaped sheath or mass of material (Fig. 18, c.). In the meantime the nucleus gradually approaches that part of the cell wall furthest removed from the centrosomes. 

   The next noticeable change is the disappearance of the barrel-shaped connective and the existence in its place of a very delicate fibril uniting the two centrosomes. One centrosome has enlarged, moreover, and if favorably situated for observation, can be seen to be a complete ring (Fig. 19). The connecting fiber just mentioned grows rapidly and is soon visible as a thread passing from the smaller centrosome back through the ring which it touches at one side, and continuing finally outside the cell (Figs. 20, 21). The vacuole or bubble-like area of the sphere gradually moves around the nuclear periphery in the meantime until it comes to lie at the pole opposite the centrosomes (Figs. 19-21, hs.). The centrosomes next approach the nucleus and so orient themselves that the smaller one points toward it. As they draw nearer a slight invagination appears in the nuclear wall and the smaller centrosome moves into it (Fig. 21, c.). The smaller centrosome is seen finally to lie inside the nucleus. The ring becomes closely applied to the nuclear wall where the invagination occurred. The filament or axial fiber of the tail, as it will ultimately be, continues to elongate.

    A close examination of this stage discloses the fact that the invading centrosome has seemingly carried with it into the nucleus an enveloping mass of extra-nuclear material which, judging from its reaction to stains, is of the same nature as the sphere substance. Taking this into account together with the fact that in the adult spermatozoon there is no visible middle piece, it seems probable that this internal centrosome together with its surrounding mass of cytoplasm may perhaps later develop into what corresponds to the middle piece of other spermatozoa. But here, since the material lies within the nucleus, it is obscured by the rearrangement and crowding back of the chromatin mass when the long head begins to form. Nothing further could be determined concerning the centrosomes as they become lost to view at this point.

    At this stage of development there is no trace of the sphere in the cytoplasm. The bubble-like structure mentioned above which originated in it, lies at what may now be designated as the anterior end of the developing spermatozoon. A flattening of the nuclear wall has occurred at the point of contact with the bubble and the latter becomes closely affixed or welded, as it were, to the nucleus (Figs. 20-24, hs.). The bubble-like portion develops ultimately into the head-spine (Figs. 22-27).

   Soon after the centrosome has penetrated the nuclear membrane, the elongation of the nucleus to form the head begins. At first only the anterior and posterior ends extend, but in a short time the whole nucleus begins to narrow. Concomitantly, the chromatic mass at the center begins to sprout out both anteriorly and posteriorly, and form a central core to the rapidly lengthening nucleus (Fig. 23, ch.).

    As the process of elongation continues in the nucleus, a narrowing of the sides occurs to some extent, but when one takes into account the enormous elongation that takes place together with the relatively slight diminution of the transverse diameter, it becomes evident that there must be considerable increase in the volume of the nucleus.

    The chromatin continues to lengthen with the extension of the nucleus and is visible as a heavy central chromatic filament. It extends entirely to the posterior end of the nucleus but falls a trifle short in front. After a time it becomes arranged in a more or less distinct wavy or spiral manner (Fig. 24, ch.). As the transformation progresses, the spiral design, although often very irregular, becomes more perceptible. Numerous linin fibers radiate out from this core of chromatin to the periphery of the nucleus, and frequently bridges or twigs of chromatin may be observed extending out along them to connect with the inner surface of the deeply stained nuclear wall. 

    A splitting of the central core occurs sooner or later as indicated in Fig. 25, so that the main chromatin mass thereafter exists as two threads laid down in an irregular double spiral. This condition persists through the succeeding changes.

    The elongation of the nucleus ceases at about the time this bisection of the central filament has been accomplished, and the nucleus displays itself as an enormously long sinuous head which may measure twice the length of the adult spermatozoon head. A dense protoplasmic mass encases it and extends backwards along the axial filament (Fig. 25).

    A shrinking or condensation of the nucleus follows, and as the this progresses, the double spiral of chromatin shortens and it widens until it attains to the characteristic appearance shown in Fig. 26. From this point, through the further condensation of the head, the openings of the spirals are crowded closer together and the exact relationship of the chromatin of the two filaments can no longer be determined. Whether or not the double spiral structure is finally lost could not be definitely determined, but if it persists, it is considerably obscured and the general aspect of the interior of the head becomes that of a series of vesicles (Fig. 27). That is, the chromatin seems to be arranged like the links of a chain, each link of which incloses a a clear area which in some preparations appears highly refractive. A remarkable fact is that the number of these links or vesicles is apparently the same as the reduced number of univalent chromosomes should be, namely, eight. Not infrequently there were only six or seven of the vesicles, but in such cases one or two were much longer than the ordinary ones and hence possibly equivalent to two. It will be recalled that in the spermatid only four chromosomes were to be seen, but that they were of the bivalent type or equivalent to eight ordinary chromosomes. There is no positive evidence that the vesicles in the head of the spermatozoon correspond to individual ten chromosomes, but the coincidence in number is at least very suggestive and it would not be surprising if later the fact develops that after the entrance of the spermatozoon into the egg, the vesicles resolve themselves into eight distinct chromosomes.

    Shortly before its development is completed, the head bends or curls backward toward the tail and lies in a dense cytoplasmic mass, thus passing apparently into a short resting or finishing stage. The typical shape of the head in this condition is that of a horseshoe (Fig. 26) but frequently it may be twisted into a series of loops, especially in the earlier stages of condensation, in much the same fashion as a string will curl if one end is held and the other twisted. By referring to Fig. 26 it will be observed that a sphere has reappeared in the cell. Its function is unknown. It may be used in the completion of the head-spine and the tail. The final structure of the tail is due apparently to the growth backward of the cytoplasm along the axial filament. A sufficient differentiation of the several parts could not be obtained to follow out the details of its formation. As the spermatozoon attains to its complete adult form, the cytoplasmic case around the head disappears, the head straightens out and is ready for attachment to the Sertoli or nurse cell. Usually a number of spermatozoa mature at one time and move toward one of the Sertoli cells until the head is imbedded in the cytoplasm (Fig. 1.). The attraction between the Sertoli cell and the spermatozoa seems to result in mutual movement, for as the latter approach, the cell, and particularly the oval nucleus, often moves inward toward the lumen of the tubule to meet them. A striated appearance of the remaining cytoplasm results from this inward migration on the part of the nucleus. In osmic acid preparations some of these striae can be resolved into what appear to be rows of minute oil globules, while others seem to have more the nature of filaments of connective tissue. In later stages, the cytoplasmic mass of the Sertoli cell diminishes. It has evidently been consumed by the spermatozoa. 

    The adult spermatozoa become free in the lumen of the tubule finally and pass out into the vas deferens and thence to the exterior.

The facts which appear to be of primary importance as set forth in the foregoing pages are briefly as follows:

1. The usual four types of cells exist in the genesis of the pigeon spermatozoon, viz :- spermatogonia, primary spermatocytes, secondary spermatocytes and spermatids.

2. Sertoli or nurse cells are present.

3. There is a curious ejection of part of the chromatin material from the nuclei of the spermatogonia just prior to their last division.

4. Sixteen loop-shaped chromosomes occur in the spermatogonia. They split longitudinally in division.

5. The last division of the spermatogonia results in the production of cells which, through a process of growth, become the primary spermatocytes.

6. Synapsis occurs in the primary spermatocytes through which a pseudo-reduction of the chromatin takes place. In the division of the spermatocyte only eight chromosomes appear, but they are in the form of heavy rings or vesicles and are evidently bivalent.

7.  During the synaptic phase there is a marked drifting the of the chromatin to the side of the nucleus in contact with the sphere. Some of the contents of the nucleus apparently pass out into the sphere.

8. During division the eight chromosomes which are incased in capsules of linin break transversely, and as they move apart they remain connected by threads of the linin casing. These threads form the interzonal fibres.

9. Intermediate bodies at the equator of the interzonal fibres are present after all divisions.

10. The secondary spermatocytes result from the division of the primary spermatocytes. Some, at least, go into a resting stage, but it is of very short duration.

11. In the division of the secondary spermatocytes only four chromosomes appear. They are of the same size and shape as those of the preceeding division. It is suggested that they are quadrivalent.

12 . If chromosomes are to be considered as differentiated into qualitative areas, then a qualitative reduction apparently occurs. It is possible that the chromosomes of the secondary spermatocyte are quadrivalent. This would probably result in a greater amount of variation in the offspring than would follow from the division of an ordinary bivalent chromosome.

[Comment: On page 33 of the printed thesis, for unknown reasons the numbering jumps from 12 to 17. The printed version of the thesis does not have 13 to 16. I do not know whether this reflects the numbering in the original thesis typescript, and why those numbers should be missing.]

17. In the transformation of the spermatid, the first perceptible change is in the centrosome. It divides and one of the resulting centrosomes enlarges and becomes ring-shaped.

18. The axial filament of the tail first appears as a thread connecting the two centrosomes and later continues backward through the ring-like centrosome and out of the cell.

19. The smaller centrosome together with material of cytoplasmic origin finally comes to lie inside the nuclear membrane and perhaps gives rise ultimately to a middle piece which becomes obscured by a covering of chromatin, and consequently appears to be absent in the adult spermatozoon.

20. The nucleus elongates to form the long head. It has a central core of chromatin in the form of a spiral filament which later splits to form a double spiral.

21. The head, during the later stages of development, undergoes a very great contraction but the spiral arrangement of the chromatin still persists in a modified form, constituting a series of eight vesicles apparently which may possibly be eight univalent chromosomes.

22. The head spine originates from a bubble-like mass of material which arises in the sphere.

It is a remarkable fact that no attempt has been made so far to investigate carefully the spermatogenesis or ovogenesis of hybrid forms. In all the mass of literature discussing or touching upon hybridism, so far as I have been able to ascertain, there has been in no instance an approach to a thorough study of the germ cells. Yet almost every writer states that through the study of hybrids, we have perhaps the best opportunity for gaining a clew to many of the most vital points in the great problem of heredity. A number of investigators have remarked that in certain instances the anthers, ovary, or testes as the case might be, were defective, and have let the matter go at that.

    The term hybrid is used in conformity with the definition given in the Century Dictionary which pronounces as hybrids such animals or plants as result from a cross of two forms noticeably different. Unsatisfactory as the definition is, it seems to be as accurate as it can be made. The objection, of course, is that it is uncertain how noticeably different two forms must be to be regarded as having individualities distinct enough to produce hybrid offspring when crossed.

    In the pigeon, some crosses are fertile, others are not. The sterile birds show a greater or less degeneration of the germinal cells. In the hybrid forms studied, the general rule seemed to be that the more divergent the parent forms, the more marked was the degeneration of the germinal cells.

    From parents which differ very widely in structure or habits, there is greater difficulty in securing female hybrids than male. The meaning of this is as yet a mystery. So far, I have been able to get but one female for microscopical examination. On the other hand, I have had six males, the offspring of very distinct species. These were all sterile forms. 

[Comment: Haldane's rule for hybrid sterility, enunciated 22 years later, noted that the heterogametic sex, which is the female in birds, is less evident among the offspring of crosses between diverging lines. See Haldane's Rule].

    From the hybrid offspring of the common ring dove (Turtur risorius) and the white ring dove (Columba alba), a large number of sections were made for microscopical study. These two forms are perfectly fertile when crossed and the fertility of their offspring seems in no wise diminished. The latter are both fertile one with another, and with the parent species. The germ cells show some of the same phenomena as those of the the sterile birds, only in a much less marked degree.

    Offspring of the common ring dove when crossed with the white ring dove are brown in color. One member of the result pair is frequently a few shades lighter in color than the other. In the next or third generation there is generally a return to the original colors of the grandparents; one of the young is white, the other brown. Occasionally both of the young are brown or, less frequently, both white. There is a marked tendency for the white ones to be female and the brown ones male.

    Inasmuch as doves produce but two young at a brood, it would require a long period of time to secure numbers sufficient to arrive at satisfactory conclusions regarding the percentage of actual reversion to the original species. From the one character of color one might be led to infer that in the third generation there is in the majority of cases a reversion to the grandparent types, since the offspring are ususally one white and one brown. This conclusion does not necessarily follow, however, for we have no means of knowing whether the brown one gets its color from a return to the brown grandparent or directly from the parents of the second generation, both of which are brown. So far as the writer has carried his experiments, the indications are that on the whole there, are more brown than white birds in the third generation, and this points to the conclusion that in the brown birds we may have both intermediate forms like the hybrids of the second generation and forms which have reverted to the brown grandparent, as the white doves have seemingly returned to the white grandparent. 

[Comment: Guyer may have achieved the classical Mendelian ratios, with brown being dominant to white!]

The abnormalities in mitosis are in the nature of multipolar spindles and asymmetrical division and distribution of the chromosomes (Figs. 28-39). These may exist independently one of another, or both may occur together in the same cell. They are more pronounced in sterile birds but may at times be seen in the fertile forms. In very many of the divisions of the primary spermatocytes one or the other, or both of these phenomena are seen. It is a curious fact that the multipolar spindles seem to be confined largely to the primary spermatocytes, and one is prompted immediately to associate the fact with the pseudo-reduction or formation of bivalent chromosomes which occurs normally at this stage of spermatogenesis. The irregularities in chromatin distribution are also seen for the most part in the primary spermatocytes.

The misshapen spermatozoa that come under this heading were present only in the sterile hybrids. In such forms there is a curious varicosity or swelling about the middle of the spermatozoan head (Fig. 40) that attracts the attention immediately when the objects are examined under the microscope. This enlargement seems to be almost universal among the spermatozoa of sterile hybrids and is sufficient of itself to produce sterility, for such a malformation would prevent its possessor from entering the egg. In a very few instances what appears to be a normal spermatozoon can be observed among the deformed ones and it is possible that if these reached a suitable egg, fertilization might result. Although the odds against them are very great, there is no reason apparent why they should not occasionally reach an egg and fertilize it.

Degenerative processes were in progress in the testes of all the sterile forms, but were most pronounced in hybrids between very divergent species, or between unlike hybrids from birds which were themselves descendants of fertile hybrids. There were some such extreme cases of degeneration that only the layer of cells lying along the wall remained in the tubule. Where such a degree of degeneration exists there is, of course, no approach to the formation of spermatozoa. There is often a strong invasion of wandering cells into the tubules, especially where the degenerative activity has become extensive. The interspaces between such tubules are also usually packed with cells which have much the same appearance as white blood corpuscles. Little clumps and globules of deeply staining cytoplasm are scattered about among the cells within the tubules. The oval nuclei of the Sertoli cells are also generally to be seen in varying numbers.

It appears that in crosses of very divergent species the degenerative processes in the germ cells are at a maximum. In closely related forms like the brown and the white ring doves fertility is not diminished and the testes seem to be normal except for occasional irregularities in mitoses.

    The formation of multipolar spindles in division and the unequal distribution of chromosomes seen in many instances are among the most interesting phenomena presented. Such irregularities, however, occur to a slight extent in normal birds.  A very careful study reveals the fact that they are more prevalent in the ordinary dove-cot pigeons than in pure breeds of doves which have bred true for many generations.

    A hybrid offspring is really a compound of two very different individual plasmas, hence conflicting tendencies must necessarily have been induced within its body. Most of the abnormality in division is in all probability but an attempt of each plasma to assert its individual activity. But why does this effort become apparent only in the primary spermatocytes? The answer appears to be simply that there is no necessity for fusion of the components of the chromatin of cells except in the germ-cells at one stage of their maturation. Investigators have found that in maturation of germ-cells, a reduction of the ordinary number of chromosomes to one-half occurs. Before this actual reduction there is ordinarily a so-called pseudo-reduction in which the chromosomes fuse in pairs so that when the cell is ready for division, although only half of the regular number of chromosomes appears, each is really double (bivalent)and equivalent to two of the simple (univalent) type. In spermatogenesis this change generally comes about in the primary  spermatocyte. 

    Now, in hybrids it may be supposed that in the ordinary cells of the body, the chromosomes from the paternal and the maternal species lie side by side and carry on the customary functions of the cells but when it comes to an actual fusion of chromosomes to form the bivalent type necessary for reduction, the incompatibility of the two different plasmas renders the union incomplete or prevents it entirely. 

    That the ordinary somatic cells of hybrids, either plants or animals, are under the influence of two distinct tendencies is well shown in the decided mosaic-like structures which frequently occur, the classic figure being to liken hybrids to warp and woof. To cite but one example, Macfarlane {1. Swingle, W. J. and Webber, H. J: Hybrids and Their Utilization in Plant Breeding. - Yearbook, Dept. of Agric., 1897.} found that a hybrid of the gooseberry and black currant instead of being a strictly intermediate type, really possessed side by side, organs characteristic of each parent; the leaves bore both the shield-shaped, oil-secreting hairs of the currant and the simple hairs of the gooseberry, though each hair was but half the size of the parent type. We may infer then that in Hybrid pigeons the univalent chromosomes from each of the parents may lie side by side in the ordinary cells of the body and divide normally, but when it comes to the period of fusion in the germ-cell, they will not unite to form the bivalent type or else they unite incompletely. The result is that in the primary spermatocyte instead of one spindle bearing eight bivalent chromosomes, a multipolar spindle, or not infrequently two separate spindles, bearing two groups of univalent chromosomes may appear. In cases where both large and small chromosomes are seen, it is necessary to suppose that a loose union has occured in some chromosomes. Whether the bivalent chromosomes formed under such conditions consist of chromatin from only one parent, or whether both parents are represented, there is no means of determining. The unequal divisions of the bivalent chromosomes of hybrids indicate that such chromosomes have in some way been rendered very unstable.

    In chromatin we have a substance which from all we know of its nature and actions, seems to be intimately bound up in the phenomena of inheritance. It is reasonable to suppose, therefore, that it constitutes at least a part of the material basis for variation in the germ. The question then arises as to whether there is any correlation between the distribution of chromatin as seen in the germ-cells of hybrids, and the marked variability which characterizes the offspring of fertile hybrids.

    If we accept the view that chromatin is a substance capable of varying in qualities in the different regions of the chromosome, then in fertile hybrids where irregular mitoses occur, the different germ-cells will certainly not be qualitatively similar after division and one would expect the offspring produced from such cells to be variable. That the chromatin of each parent species often retains its individuality, is indicated by the fact observed in many primary spermatocytes where two separate groups of the small or single type of chromosome exist. The division of such a cell into three or four, as the case may be, results in the formation of new cells, some of which will manifestly contain chromatin from only one of the original parent species and some, only from the other. Some of the spermatozoa then, will bear chromatin from only one of these species. In the offspring from such a cell one would expect a much closer return to whichever one of the parent forms it represented than in the offspring of a "mixed" spermatozoon.

    In discussing irregular divisions, however, it must not be forgotten that many apparently normal divisions of the primary spermatocytes also occur in all hybrids, and constitute by far the predominant kind of division in hybrids from closely related forms. Unequal distributions of chromatin cannot therefore play the most important part in variation or reversion.  There seems to be no other interpretation, indeed, than that in the many normal mitoses of the bivalent chromosomes which occur, the chromatin of the father and of the mother is set apart so that the ultimate germ-cells are what might be termed "pure" cells; that is, a given egg or sperm-cell contains exclusively or at least predominantly qualities from one parent. The offspring from fertile hybrids of the same parentage might then be similar to the mixed type of the original hybrid, or revert to one of the grandparent types, dependent upon the chances of the various cells for union at fertilization. If a spermatozoon and an egg containing characteristics of the same species unite, then the reversion will be to that species; if a sperm-cell containing the characteristics of one species happens to unite with an ovum containing characteristics of the other species, then the offspring will be of the mixed type again. By the law of probability the latter will be the more prevalent occurrence, because there are four combinations possible, and two of the four would result in the production of mixed offspring, while only one combination could result in a return to one of the ancestral species.

    From the fact that in cases of apparently complete return to one parent type, characteristics of the other parent may nevertheless crop out from time to time in succeeding generations, it is evident that all of the germ-cells are not absolutely "pure." The occasional inequalities in the division of individual chromosomes, as already mentioned may account for this fact. Not infrequently in division more than half of one chromosome goes to one cell; thus, what might otherwise have been a pure cell will contain a few characters of the foreign species and these may appear later in offspring as variations of the intermediate type. Thus the actual number of intermediate forms will doubtless be augmented somewhat beyond the proportion indicated in the probability of combination where all the mature germ cells are pure. It is also possible that such mixed cells as result at one end of a tripolar spindle will introduce yet another factor for variation though tripolar spindles are too few in most fertile hybrids for this element to enter to any great extent. It is probable that the irregular distributions of chromatin, where such occur, have more to do with such succeeding offspring as show variation, and less with those which return to the specific types. In the latter case the chromatin of each species has frequently remained entirely distinct (in marked hybrids), or has normally separated again at the sundering of the bivalent chromosomes (in mild crosses) into the two original plasmas. Thus it is very obvious that most of the variation seen in the offspring of fertile hybrids is due to the union again of two "pure" germ-cells each of which represents a different one of the original parent species, but it is also an evident that inequality in the distribution of chromatin may enter as a factor.

    That irregular divisions cannot account entirely for reversions to grandparent types is very evident in the crosses of brown and white doves, where the irregularities are by far too few to equal the percentage of reversions. There is but one ultimate conclusion, then, namely, that the irregularity in division of the primary spermatocytes which appears in hybrids between very different species, is but an index to what occurs in ordinary crosses. In the latter, instead of separate spindles and non-fusion of chromosomes, a true union occurs, but the bivalent chromosomes ultimately divide in such a way that the respective plasmas occupy different cells. There is a separation of the paternal and the maternal chromosomes which had fused during synapsis.

    The above interpretations are offered with the hope that they may perhaps lead to some clew concerning the real nature of the material basis of heredity. If the conception proves to be a true one, then it doubtless affords a key, among other problems, to the long-standing one as to why many plants will come true from slips or grafts, but not from seed. The reason may be sought in the pseudo-reduction period of the germ-cell. Plants such as the apple, for example, which do not come true from seed, are practically multi-hybrid. In the germ-cells there will be numerous incompatibilities due to the fact that the plant has been miscellaneously fertilized for a number of generations. In propagation by means of slips, the chromosomes lie side by side and divide in the ordinary way to construct and maintain the new body, so that it is practically a continuation of the old one; but when the time comes for maturation of the germ-cells, the lack of harmony between the various plasmas asserts itself, with the result that bivalent chromosomes are formed, which divide in such a manner as to segregate different sets of ancestral qualities. The resulting combinations in fertilization will give rise to seed many of which may possess dissimilar sets of qualities. 

    Concerning the other abnormalities met with in the spermatogenesis of hybrids, about all that can be said is that the whole phenomena show lack of vigor in the development of the germ-cells, whatever this may mean. The deformed spermatozoa indicate want of sufficient vitality to push the development through to completion. The germ-cells start out apparently to perform their functions normally, but later succumb to the conflicting forces at work within their boundaries. In most instances where the species are very distinct, the hybrids can be obtained in the first place only, as it were, under protest [Comment: Hybrid inviability]; the breeding season must be at its height and everything in the most favorable condition possible.

    As to why the reproductive organs should be more susceptible to abnormal changes than other regions of the body, we appear have no clew. Darwin has pointed out repeatedly the curious parallel between crossing and the change produced by physical conditions. Animals and plants removed from their natural conditions are extremely liable to have their reproductive systems affected. Still he recognizes that sterility is incidental and not a necessary concomitant of hybridism. In some fertile hybrids, indeed, there seems to be an improvement on the parent form [Comment: Hybrid vigour.]. Thus Vernon {1. Vernon, H. M : The Relation between the Hybrid and Parent Forms of Echinoid Larvae. - Phil. Trans. Series B. CXC, 189.} found two species of sea urchins, which when crossed showed greater fertility and produced larger, stronger larvae than in direct fertilization. 

The above conception of the chromatin does not necessarily imply that it is of exclusive importance as the basis for hereditary transmission. There seems to be no sufficient reason for not regarding the cytoplasm likewise as an important factor.

    Connected with the possibility of the qualitative differentiation of the chromatin there arises another possibility, namely, that only the individual and the more variable characteristics are conveyed by the chromatin, while back of these in the cytoplasm is a general stock or plasma from which the more stable and constant form of an animal is derived. 

    In other words, that there is a sort of double background of qualities to every organism which might perhaps be designated as general and individual. The cytoplasm provides the general substratum upon which all the later individual traits controlled by the chromatin are gradually built up. 

    To illustrate in pigeons: what is recognized as the pigeon type or the general characteristics that mark a particular bird as a pigeon would constitute the first or general type [Comment: species specific characteristics], and those characteristics by which one recognizes different groups of pigeons or different individuals, would represent the second or individual type..[Comment: classical Mendelian characteristics]

    Such a conception, among other things, would simplify our interpretation of reversion to a distant ancestral form. With reference to the actual occurrence of such reversion it may be said that there are sufficient well authenticated cases on record to establish beyond a doubt its existence. Among others, Ewart {1. Ewart, J. C: The Penycuik Experiments., 1899.}, in a recent paper, gives some beautiful corroborative evidence.

[Comment: James Cossar Ewart was a friend of George John Romanes, who supported his application for the Chair of Natural History at the University of Edinburgh, which he occupied from 1882-1927. Ewart crossed a male zebra with a mare to produce offspring with stripes. The same mare on subsequent crossings with males of her own type never again produced striped offspring. No big deal now, but in the 1890s the idea that male characteristics from an earlier cross might emerge in later crosses ("telegony") had to be disproven.]

[Comment: There are two issues here - first the duality of species-determining characters and individual-specific characters, and second the idea that the agencies responsible for these different characters might have different locations. The duality was noted by Romanes in 1886, and by William Bateson, who was much impressed with Guyer's work. Bateson postulated a "residue" or "base" that carried fundamental species-determining characters. He was reluctant to ascribe both species-determining and individual characters to the chromosomes.] 

    A supposition that would be required to supplement the above conception is that the qualities of the nucleus after long continued recurrence must finally become stamped upon the cytoplasm, else there could be no evolution or development of the type itself. Our knowledge of nucleus and cytoplasm so far is, of course, too crude to give a justifiable basis for such an assumption beyond that of mere suggestion, but from what we do know of the interrelation of cytoplasm and nucleus, it would seem that perhaps this difficulty may not, in the end, prove insurmountable.

The points of special interest in the spermatogenesis of hybrid pigeons are as follows:

1. The general plan of spermatogenesis is not essentially different from that of normal pigeons.

2. From forms not too distantly related, fertile hybrids are easily obtained.

3 Infertile hybrids, resulting from the crossing of very distinct species, are more difficult to secure. Females are rare.

BALLOWITZ, E : Untersuchungen ueber die Struktur der Spermatozoan. - Arch. f. mikr. Anat., XXXII, 1888.

BOVERI, TH : Ueber die Befruchtung der Eier von Ascaris megalocephala. - Sitz. Ber. Ges. Morph. u. Physiol. Munchen II, 1887. Id.: Ueber partielle Befruchtung.-Ibid : IV, 1888. Id.: Ueber die Befruchtungs-und Entwickelungs-fohigkeit kernloser Seeigel-Eier. - Arch. Entwm., II, 1895.

DRIESCH, H. and MORGAN, T. H : Zur Analysis der ersten Entwickelungsstudien des Ctenophoreneies. Arch. Entwm., II, 1895.

DRIESCH, H : Ueber rein-mutterliche Charaktere an Bastardlowen von Echiniden. - Arch. Entwm., VII, 1888.

VON EBNER. V : Zur Spermatogenese bei den Saugethieren. - Arch. f. mikr. Anat., XXXI, 1888.

EWART, J. C : The Penycuik Experiments., 1899.

GALEOTTI, G : Ueber experimentelle Erzeugung von Unregelmassigkeiten des karyokinetischen Processes. - Beitr. z. patholog. Anat. u. z. Allg. Pathol., XIV, 2, V, Jena, Fischer, 1893.

GUYER, M. F: Ovarian Structure in an Abnormal Pigeon.- Zool. Bull. II, No. 5, 1899.

HANSEMANN, D: Karyokinese and Cellularpathologie. - Berl. Klin. Wochenschrift, No. 42, 1891.

HENKING, H: Erste Entwickelungsvorgange in den Eiern der Insekten. - Zeitschr. f. Wiss. Zool. LL 1891. 

HERLA, V: Etude des variations de la mitose chez 1'ascaride megalocephale. - Arch. Biol. XIII, 1893.

JUEL, H. O: Die Kerntheilungen in den Pollenmutterzellen. - Jahrb. wiss. Bot., XXX, 1897.

LENHOSSEK, M. v: Untersuchung ueber Spermatogenese. - Arch. f. mikr. Anat., LI, 1898.

LUSTIG, and GALEOTTI, G: Cytologische Studien u.eber pathologische menschliche Gewebe.--Beitr. Path. Anat., XIV, 1898.

MCCLUNG, E. C: A Peculiar Nuclear Element in the Male Reproductive Cells of Insects. Zool. Bull. II, 4, 1899.

MCGREGOR, H: The Spermatogenesis of Amphiuma. - Journ. Morph. XV, Suppl., 1899.

MONTGOMERY, TH. H: The Spermatogenesis of Pentatoma. - Zool. Jahrb. XII, 1898.

MOORE, J. E. S: On the Structural Changes in the Reproductive Cells during the Spermatogenesis of Elasmobranchs. - Quart. Journ. XXXVIII, 1893.

Von RATH, O: Zur Kenntniss der Spermatogenese von Gryllotalpa vulgaris. - Arch. f. mikr. Anat., XL, 1892 .

RUCKERT, J: Zur Eireifung bei Copepoden. - An. Hefte, 1894. Id.: Ueber die Selbstandigbleiben der vaterlichen and mutterlichen Kernsubstanz wahrend der ersten Entwickelungs des befruchteten Cyclops-Eies. - Arch. f. mikr. Anat., XLV, 1895.

STANDFUSS, M: Handbuch der palarktischen Grossschmetterlinge. - 2nd., edition, Jena, 1896.

SCHOTTLANDER, J: Ueber Kern and Zelltheilungsvorgange in dem Endothel der enzundeten Hornhaut. - Arch. f. mikr. Anat., XXXI, 1888.

SWINGLE, W. J. and WEBBER, H. J: Hybrids and their Fertilization in Plant Breeding. - Yearbook, Dept. of Agric., 1897.

VERNON, H. M: The Relation between the Hybrid and Parent Forms of Echinoid Larvae. - Phil. Trans, Series B, CXC, 1898.

WILCOX, E. V : Spermatogenesis of Calloptemus femur-rubrum and Cicada tibicen. - Bull. of the Mus. of Comp. Zool., Harvard Coll., XXVII, i, 1895. 

ZOJA, R: Sullo independenza della cromatina paterna a materna nel nucleo delle cellule embrioli. - Anat. Anz. XI, 1895.

ZUR STRASSEN, O: Ueber die Riesenbildung bei Ascaris-Eiern. - Arch. Entwm., VII, 1898.