I ) The morphological literature concerning the origin of the germ cells in the Vertebrates can be grouped as follows: a) The germ cells have a purely entodermal origin in early development. b) They originate from the mesoderm in early development. c) They derive from the germinal epithelium. d) They derive from embryonic entoderm and germinal epithelium. e) They originate from embryonic mesoderm and germinal epithelium. Many investigations point to an early entodermal origin of the germ cells in the Anures, although the theory of a second generation of germ cells from the germinal epithelium is defended by several authors. Respecting the Urodeles, the opinions of the various authors are altogether conflicting. 2) The experimental literature is in favour of an early entodermal origin of the germ cells in the Anures, and of a possible early mesodermal origin in the Urodeles. As to a secondary origin of germ cells from the germinal epithelium, the experiments are contradictory. 3) My own experiments were carried out with Urodeles ( Triturus species and Ambystoma mexicanum). 4) Vital staining experiments of the entoderm yielded results completely in agreement with those of VOGT, SCHECHTMANN, and others. The material, situated at the vegetative pole of the uncleaved egg, ascends to the blastocoel. It furnishes the ventro-caudal entoderm, which will form the nutritive yolk material. The superficial cells of the ventral half of the vegetative field of the gastrula develop into the dorso-caudal entoderm, which will form the entodermal epithelium of the intestine. 5) Removal of the plasma at the vegetative pole of the uncleaved egg has no influence upon the formation and differentiation of the germ cells. The same holds good when this material is removed at the early gastrula stage by defects in the bottom of the blastocoel. 6) Removal of the superficial cell layers of the ventral half of the vegetative field at the early gastrula stage causes a great reduction in the number of germ cells, accompanied by an impaired differentiation of these cells. The fertile region decreases in length with a decreasing number of germ cells. The distribution of the germ cells along this fertile region, however, remains regular. 7) Vital staining experiments of the posterior part of the mesoderm prove that the presumptive lateral plate mesoderm occupies the ventral and latero-ventral blastoporal lips at the early yolk plug stage. The nephrogenous cord mesoderm, situated for the greater part in the ventro-lateral blastoporal lips, forms a horseshoe-shaped band of cells round the presumptive lateral plate mesoderm. The presumptive lateral plate mesoderm diverges considerably to the lateral sides after passing the ventral blastoporal lip, and shifts dorsad by the movements of dorsal convergence. It occupies a narrow strip of cells at the lateral sides of the entoderm mass at the neurula stage. During the stretching of the embryo, it extends in a cranial and ventral direction (ventral convergence), and finally covers nearly the whole entoderm. 8) After removal of the presumptive lateral plate and nephrogenous cord mesoderm at the early neurula stage these parts are partially regulated from more dorsal mesoderm. Notwithstanding this regulation, several embryos show a complete absence of germ cells. The germ cells, left behind after incomplete sterilization, show a good differentiation, and are localized in the most cranial part of the "germinal region". Removal of the presumptive myotome mesoderm causes no alteration in the number and differentiation of the germ cells. 9) These observations prove that both the dorso-caudal entoderm and the posterior mesoderm (lateral plate and nephrogenous cord) have a rôle in the formation of the germ cells. The distribution and differentiation of the germ cells in these experiments are factors which favour the opinion of an origin of the germ cells in the mesoderm, and the presence of an inductive centre in the dorso-caudal entoderm. 10) After the removal of the whole entoderm at the early neurula stage, the larvae show a reduced number of germ cells localized in the cranial segments of the germinal region. Only that part of the mesoderm contains germ cells which was already invaginated at the time of operation and has, therefore, been in close contact with the dorso-caudal entoderm. These facts are an additional motive for assuming an inductive action between entoderm and mesoderm. 11) In the Urodeles the size, stainability, etc., cannot be used as species characters of the germ cells in heteroplastic transplantation experiments. Only the pigmentation of the germ cells forms a reliable difference between the germ cells of Tr. cristatus and the other Urodele species used. 12) After heteroplastic transplantation of the whole entoderm, the animals show only germ cells of that species which furnished the ecto- and mesodermal layers. The germ cells, reduced in number, are situated chiefly in the cranial half of the "germinal region". Their number and position cannot be explained merely by the inductive action of the host entoderm before the operation, but constrain us to assume a supplementary induction by the grafted entoderm. 13) Heteroplastic transplantations of the presumptive lateral plate and nephrogenous cord mesoderm prove that myotomes, Wolffian ducts, nephrogenous cords and entoderm do not form sources of germ cells. They prove that the germ cells originate from the presumptive lateral plate mesoderm only. This series lends again (cf. no. 8) support to the opinion that lateral plate tissue within the germinal region, formed by regulation from other mesoderm, cannot form p.p. germ cells. This fact is an argument in favour of a predetermination of the presumptive germ cells. 14) This predetermination may be brought about by the segregation of a special sort of protoplasm during the "chemodifferentiation" of the egg. This implies probably that only the presumptive germ cells have "reactive power" to the inductive stimuli exerted by the entoderm; in other words, the presumptive germ cells are already specific elements. Cf. BOUNOURE 1939. 15) Vital staining experiments of entoderm and mesoderm prove that the dorso-caudal entoderm and the presumptive lateral plate mesoderm come into direct contact with each other at the end of the gastrulation and during the first half of the neurulation period, so that an induction can actually take place. 16) Germ cells differentiate in the presumptive lateral plate mesoderm, invaginated a few hours before the removal of the entoderm at the early neurula stage. This fact proves that the determination of the germ cells occurs very rapidly. 17) Transplantation experiments of the pronephros rudiment at the early neurula stage prove that the Wolffian duct grows out from the caudal part of the pronephros rudiment to the cloaca, and that every cell of the Wolffian duct rudiment has the power of active migration. 18) Probably, neither the pronephros nor the Wolffian duct influences the formation and differentiation of the germ cells. 19) Transplantation experiments of the notochord render it probable that the notochord may even prevent the differentiation of the germ cells in its environment, 20) The fact that several organs from which inductive influences emanate and act upon other parts (notochord, Wolffian duct), cannot replace the action of the dorso-caudal entoderm in the determination process of the germ cells, points to a specific character of this action. 21) A disturbance of the invagination of the mesoderm by an active influence of a supernumerary notochord, coupled with a mutual displacement of the caudal entoderm and mesoderm by the stretching of the graft, had prevented the contact between caudal entoderm and mesoderm in several animals. In these animals the presumptive germ cells could not be determined. Moreover, the transformation by the notochord of the adjacent lateral plate mesoderm has prevented the germ cell differentiation. 22) Three phases in the germ cell migration in the Urodeles can be distinguished. In the first phase, they migrate from their place of origin to the neighbourhood of the Wolffian duct, in the second from this point to the median plane, and in the third, from the median germ cell ridge to the definitive germ ridges. 23) The displacements of the germ cells in the first phase occur, probably, by a combined active and passive migration. The first displacements are mainly passive (gastrulation movements), whereas the end of this migration path is covered by an active migration of the germ cells. In the second phase, a passive transport of the germ cells by the growing out of the dorsal parts of the lateral plates suffices to explain the displacements of the germ cells. Finally, in the third phase, we may conclude from the absence of extensive material displacements and the presence of an attractive action from the definitive germ ridges, that the migration is active again.