HUMAN SPERMATOGENESIS: A STUDY OF INHERITANCE. 19 



nuclear membrane. Meves (1897) was the first to describe early spermatids 

 with a pair of centrioles on the surface, and the flagellum connected with these; 

 later (1898), in another brief paper, he stated that the proximal centriole 

 comes to lie just behind the head, while the distal one, before becoming a ring, 

 buds off a little rod. I have paid particular attention to this last point, but 

 have found no division of the distal centriole. Broman (1901) figured clearly 

 the centrioles and cuff, and certain complex spherical bodies ("Korbblaschen") 

 that he holds are not homologous with mitochondria because there is no trace 

 of them in the spermatocytes and because they disappear before maturity. 

 Retzius (1909) has given the best figures of the later stages of the sperm iogenesis 

 but the accompanying description omits many of the interesting details shown 

 in the plates. He finds the mitochondrial mantle is derived from certain large 

 granules, six to seven in number, on each side of the middle piece; that the cuff 

 is an open cylinder and at no period fibrous ; and that a great lobe of cytoplasm 

 abstricts. His figures also show that not only the anterior portion of the proxi- 

 mal centriole may consist of two or three parts, but also the posterior portion of 

 the same centriole. Retzius' is the most important study, but it would have 

 much more value had he seriated the stages figured. 



An important side of any study of spermiogenesis is to determine just what 

 parts and substances contribute to the mature spermatozoon, and, accordingly, 

 to the fertilization of the egg. It is now established for a considerable number of 

 cases that the entire spermatozoon enters the egg, without omission of the tail. 

 But it is also known that in most cases only a portion of the spermatid contributes 

 to the spermatozoon. Thus in man as in most animals, with the possible excep- 

 tion of amphibians and some insects, the greater part of the cytoplasm becomes 

 abstricted and cast away. And that happens in man also with the greater amount 

 of the karyolymph, which we have described as being forced out of the nucleus 

 to produce the cuff substance which is later thrown off with the cytoplasm. 

 .Moves (1899), Duesberg (1908) and others generally hold that the cuff is a 

 cytoplasmic differentiation, but Van Molle (1905, 1910) found that the cuff is 

 derived from achromatic nuclear material, and my present account agrees with 

 his, except that I do not find the cuff is formed by buds off the nucleus, but rather 

 from a discharge of nuclear sap through the nuclear membrane. This explanation 

 of the genesis of the cuff coincidentally explains the sudden shrinkage of the 

 nucleus, and I have called attention (1911a) to the fact that this shrinkage is a 

 very general phenomenon. The spermatozoon then loses the greater part of 

 both cytoplasm and karyolymph, for which reason these substances can play 

 little part in inheritance. 



Another interesting matter concerns the history of the centrioles. ^ These 

 become large and very distinct about the middle period of histogenesis, near 

 its close they lose their staining power and become suddenly much smaller. No 

 trace of the distal one can be found in the mature sperm. In Euschistm it was 

 described by me (1911a) how at a certain period both centrioles seem to suddenly 



