1902.] NICHOLS — SPERMATOGENESIS ONISCUS ASELLUS LINN. 99 



latter group. It is possible, and I advance it simply as a tentative 

 hypothesis needing corroboration, that these strikingly dissimilar 

 forms have arisen from a primitive one, simple and amoeboid in 

 character like that of Polyphemus. 



The ingenious series of homologies drawn by Auerbach between 

 the head, tail, apex and middle-piece of the spermatozoa of Verte- 

 brates and Insects and similar structures in Astacus appears plausi- 

 ble. Since, however, the location of the centrosome and the 

 sphere substance remains undetermined, the homologies cannot be 

 said to be in all respects established. A more detailed and thor- 

 ough examination of the spermatozoa of the Crustacea, especially 

 of their behavior in fertilization, might extend these homologies. 

 If the spermatozoon of Oniscus be compared with the type most 

 frequently occurring in animals, the part immediately adjacent to 

 the nucleus, the delicate fibril shown in Figs. 77^, 79, corresponds 

 in location to the middle-piece. Whether this is in reality the 

 habitation of the centrosome might be discovered through a study 

 of its fate during fertilization. My observations on the spermato- 

 genesis throw no light on the question. 



The Isopods are unique among the Crustacea in the formation of 

 colonies of spermatozoa of a nature so close that they appear as 

 units. Concerning their origin in Oniscus, I can confirm M. 

 Gilson's statement that the formation of the bundle takes place in 

 a Plasmodium, cell boundaries being for a time entirely absent, and 

 with the main outlines of his account of the changes taking place 

 in the development of the spermatids into the mature colony I am 

 thoroughly in accord. 



The number of nuclei entering into a bundle, according to my 

 observations, is not invariably six, but may vary within considera- 

 ble limits. The number of cytoplasmic fibres is assumed by M. 

 Gilson to be equal to the nuclei, but in his Fig. 328, PI. VIII, they 

 are shown to be more numerous. As has been already said, I have 

 been unable to convince myself of a direct continuity between these 

 fibres and the nuclei. In his Fig. 320 (an immature spermato- 

 phore) the cytoplasmic fibres may be traced directly to the nuclei. 

 I have, however, riot been able to obtain images of equal clearness 

 from my preparations. Nor have I obtained anything at all similar 

 to the rings or vacuoles, shown in Gilson's Figs. 328, 329 and 330, 

 near the anterior end of the bundle. In Sphaeroma serratum, Gilson 

 states, the continuance of head and tail is very evident, forming an 



