My first surprise was to find the facts in Caretta closely like those described by Allen for Chrysemys. The germ-cells are segregated in the entoderm from about the second to fifth day near the lateral border of the area pellucida, from a point near the cephalic tip of the pro nephros to the caudal extremity. Up to the fourth day occasional cells occur also in the area opaca near the border of the area pellucida. This indicates the first step in the migration, namely, from area opaca to lateral border of area pellucida caudally; since germ-cells disappear later from the area opaca. In the area pellucida the germ-cells then become fairly closely aggregated into two bilateral cords. These are, however, not of uniform thickness throughout, but have an interrupted or segmental character, suggesting "gonotomes." At about the fifth day germ-cells wander into the splanchnic meso derm and migrate medially. When the hind-gut becomes closed (from the sixth to the tenth day) germ-cells become included in its mucosa and in the more peripheral portion of wall. From here many migrate through the mesentery across the angle of the ccelom to the differen tiating sex-gland at the seventh day, a few apparently degenerating within the gut-wall. At the thirty-second day they are distributed among the peritoneal epithelium of the developing sex-gland and in the subjacent mesenchyma. Only very occasionally is a cell found extra-regionally at this stage, and apparently no sex-cords have as yet begun to form; but neither is there the slightest evidence of a dif ferentiation of the so-called germinal epithelium into genital cells. In the earlier stages a few germ-cells may be seen in mitosis and an occasional cell may apparently divide at any stage.
My second surprise was to find practically no germ-cells in any vascular channels; only two exceptions were noted. Nor does there seem to be any possibility of confusion of sex-cells with blood-cells, judging from Swift's figures or on the basis of my sections of the logger head-turtle embryo. Swift's observations have now been confirmed by von Berenberg-Gossler for chick and duck. Von Berenberg-Gossler was unable, however, to find similar conditions in the case of the lizard embryo. The condition of abundant intravascular germ-cells in the chick embryo and the duck embryo is so far unique. Conditions respecting the germ-cell history in birds are evidently different from those obtaining in reptiles and certain other amniotes, and depend upon the cephalic (pro-amniotic) site of origin of the germ-cells and their close relation to the vascularizing mesoderm.
Considering the initial very close relationship (both spatial and genetic) between the entoderm and the mesoderm at the time when the latter is in the early stages of vascularization and the considerable migratory capacity of the primordial germ-cells, the presence of these cells outside of the usual migration route, even in the blood-vessels, is nothing extraordinary. In the chick the initial source of origin is
in the entoderm at the anterior extremity of the blastoderm in a region originally free of mesoderm. The cells begin their first migration about the time the vascularizing mesoderm invades this region, and so readily enter into the blood-vessels by which they are transported in large part caudally to the medial splanchnopleure, whence they pass via the mesentery to the gonads (Swift). In forms like the turtle and the clog-fish, where the germ-cells are originally scattered caudally in the area opaca, they do not in large numbers come into so intimate early relation with the young blood-vessels, and in consequence reach the medial splanchnopleure largely via the visceral layer of the mesoderm by their own amceboid activity, meanwhile passing through a tempo rary stage of sharp segregation into the paired cords of the area pellu cida. But even here it is apparently an easy matter for the germ-cells to become involved with the blood-channels either through active migration or passive inclusion. There is nothing, therefore, inherently contradictory or unique in the condition of the germ-cell migration as it obtains in the chick and duck. The difference is one of degree rather than kind, and depends upon the difference in initial location of the primordial germ-cells with respect to the mesoderm and the head end of the embryo.
The very careful work of Allen, in which he counted the number of germ-cells at the different stages of development in different areas in Chrysemys, clearly shows that the total number of germ-cells in different individuals of a species varies within wide limits (302 to 1,744), and that the time and rate of migration also vary considerably. I there fore made no attempt to approach the question of the route and manner of migration on the basis of germ-cell counts. The fact that at progressively later stages in development the germ-cells were found in the entoderm near the lateral border of the area pellucida, then in the visceral plate of the lateral mesoderm and more medially in the entoderm (later also in the closed gut and enveloping mesenchyma), then in the mesentery, and finally across the caelomic angle among the peritoneal epithelium of the sexual gland (genital fold), indicates that this represented the actual route of migration. The further facts that only an occasional germ-cell was seen in mitosis throughout the series (second to thirty-second day), that only a few were seen in conspicuous stages of degeneration, and that practically none appeared in blood vessels (and only a few in the periaortic mesenchyma) seems to prove the indicated route as the actual and substantially exclusive migra tion path. In the case of Cymalogaster, also, Eigemnann found that the sex-cells do not divide from the period before the first somites have formed—at which time they are first distinguishable—up to about the period when the young fish reaches a length of 7 mm.