248 MICHAEL F. GUYER. 



spermatids are to be found which are indefinite in nuclear make- 

 up and seemingly impaired in some way. 



It cannot be positively affirmed, of course, that such cells as 

 are represented in Fig. 204 are the product of the four-chromo- 

 somed class instead of the other. Such an interpretation is 

 merely an implication which grows out of what appears to be 

 other aberrancies in the behavior of the four-chromosomed class. 



While at first glance, this view that one entire class of speima- 

 tids degenerates without becoming functional may seem im- 

 probable to some, it should be borne in mind that such an 

 occurrence is by no means unique in spermatogenesis. In both 

 phylloxerans and aphids, for example, just such an aborting of 

 one of the two classes of sperm-forming cells occurs, to say 

 nothing of the even more remarkable conditions which obtain 

 in bees and in certain hermaphroditic forms such as Rhabditis. 



Figs. 200 and 203 represent stages in what I interpret as the 

 normal sequence of transformation. In such cases, by the time 

 the axial filament first appears the nucleus has already begun to 

 elongate slightly. Typically the chromosomes from the last 

 division seem to arrange themselves into more or less of a closed 

 ring around which the nuclear membrane forms. They then 

 tend to concentrate gradually toward one side of the nucleus 

 along the periphery until they form more or less of a crescent 

 which thickens and shortens as the process continues (Fig. 200). 

 The nuclear membrane, along the margin free from chromatin, 

 fades from view as the transformation progresses, leaving finally 

 an elongate chromatin mass rounded at one end and more 

 sharply pointed at the other. The pointed end is seen to be 

 provided with a definite head spine and the blunter end to be in 

 juxtaposition to the axial filament. The latter seems to develop 

 in the typical way from the divided centrosomes, one of which 

 has become knoblike and the other a ring. 



Instead of following what appears to be the simpler course of 

 development as depicted in Fig. 200, the nucleus seems fre- 

 quently to transform into the spermatozoon head by a process 

 of uncoiling, representative stages of which are shown in Fig. 203. 

 As is the case in other various types of spermiogenesis, fragments 

 of cytoplasm appear not infrequently to be cut off from the 



