326 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



more prominent (Fig. 84). and in the spermatid, it still' persists as a very 

 definite deeply stained granule, lying at the focus of the remnants of the 

 spindle fibres (polar radiations, as already mentioned, do not occur), and in 

 contact with the cell membrane, much as Gorich (:03\ :04) has described 

 for Sycandra. Moreover, I may state at once that there is no period 

 during the metamorphosis of the spermatid when the existence of the 

 centrosome cannot be demonstrated. Connected with the persistence 

 of the centrosome and of the remnants of the polar portion of the spindle 

 (Fig. 89) is the fact that the spermatid nucleus (Figs. 88, 89), in the 

 early stages, never takes up the position close to the cell membrane 

 which is characteristic of spermatogonia! and spermatocyte nuclei in 

 similar stages, but always lies more nearly central. The average total 

 diameter of the spermatid is 3.5 /*, that of its nucleus, 2 /*. The con- 

 nection of pairs of daughter cells by the interzonal bridge persists, as 

 after the first maturation division, for a considerable period, during 

 which this structure becomes more and more attenuated, and the cells 

 move further and further apart. But that this separation is due to a 

 mutual repulsion, as Blackmail (: 05, p. 61) suggests, I am by no means 

 prepared to admit. Preparations in which the cells have been crushed 

 show that the bridge is of truly fibrillar nature, for in such the indi- 

 vidual fibres are often more or less clearly separated from one another 

 (Fig. 88). Occasionally, even before the complete separation of the 

 daughter cells, a short, delicate fibre can be detected, arising from the 

 region of the centrosome (Fig. 88) ; this is, doubtless, the earliest stage 

 in the formation of the tail of the spermatozoon. 



The stage shown in Figure 90 may be considered the beginning of the 

 actual metamorphosis. The interzonal bridge has now broken down, 

 the nuclear membrane has reformed, f.nd the chromatin has the form of 

 a distinct, deeply stained network, the remainder of the nuclear area 

 being occupied by rather granular karyoplasm. In different cells the 

 condition of the network varies, and I have not been able to determine 

 whether there is a definite number of segments. The cytoplasmic body 

 of the cell is comparatively large, has a dense, finely reticulated appear- 

 ance, stains feebly with iron haematoxylin, though strongly with the 

 ordinary plasma dyes, and occasionally encloses darker masses or gran- 

 ules. From the fact that I have observed these in only a few cases I 

 am inclined to believe that they are not archoplasmic structures. The 

 remnants of the polar portion of the spindle, and of the interzonal 

 bridge still persist. But the fibrous appearance so characteristic of both 

 in the telophase is now hardly distinguishable in either; they now ap- 



