276 ir. LYNDHUKST DUKE. 



spindles is a black mass of deeply staining matter at the 

 extreme poles. In some early spindles shown in fig. 18 a, the 

 earliest actual daughter spindle-stage to hand, this polar 

 mass is seen as two adjacent granules or centrioles lying in a 

 definite centrosphere showing radiations. In fig. 18 h these 

 two granules are connected by a deeply staining link. This 

 I interpret as the early division of the centrosome, occurring 

 almost before the daughter-nuclei, which in the figs. 18 a and h 

 are distinguishable as faint vesicles, are free from their parent 

 spindle. In this connection it is of interest to note that the 

 daughter-nuclei always appear provided with two centrosomes. 

 I have not been able to discover any with a solitary centro- 

 some. This is in keeping with the above suggestion as to 

 the early division of the centrosome in the history of each 

 daughter-nucleus. As the daughter-nuclei become smaller 

 their division-figures become less complicated, while the 

 chromatin becomes arranged as a single mass rather than as 

 separate particles. Some of the smallest spindles still show 

 occasionally distinct chromatin elements near their poles, but 

 the majority do not. There appear to be no definite astral 

 rays distal to the terminal mass of chromatic substance 

 (figs. 21 d, 23, and 24 h). Finally all traces of spindle- 

 formation disappear, and the nuclei are reduced to mere 

 masses of chromatin about 1 to 1'5 ^ in size. These are 

 arranged on the periphery of masses of protoplasm, after 

 the fashion of a typical so-called Perlenstadium, and the 

 protoplasm soon becomes mammillated round each nucleus 

 with the formation of gametes (fig. 25). 



That part of the protoplasm which does not take part in 

 the formation of the gametes — the Restkorper — contains 

 a few nuclei which have not kept pace with the general 

 division (fig. 25 b). These laggard nuclei are present here and 

 there in all sections of the later daughter-divisions, and are 

 noticeable in that they are larger than their more numerous 

 companions. Similar nuclei have been noticed by Leger and 

 Duboscq in Hoplorhynchus (13). Scattered throughout 

 the later cysts are also seen a number of round clear bodies 



