334 ROBERT H. BO WEN. 



" no evidence was found for the casting off of any substance by 

 the sperm," it is abundantly clear that Montgomery failed to ob- 

 serve the final steps in the differentiation of the sperm head. In 

 a stage following shortly on the last one described by Mont- 

 gomery, the head undergoes a characteristic change resulting in 

 what appears to be a complete vacuolization of the chromatic 

 lining. Then the chromatin collapses toward the axis of the 

 head, and a thin, compact thread is formed running throughout 

 the length of the head and enclosed in a protoplasmic envelope 

 which appears to correspond to the outline of the old head (Fig. 

 33). So far as I know, this stage in the transformation of the 

 Hemipteran sperm has not been noted heretofore, but the de- 

 scription given by Faust ('13) of the mature sperm in Anasa 

 tristis corresponds very closely to the stage just described in 

 Murgantia. Possibly this phenomenon may prove to be much 

 more generally distributed among the Hemiptera than has been 

 suspected. That this is not a mature sperm is quite clear in my 

 material, for subsequently the outer protoplasmic envelope dis- 

 appears and the head of the completed spermatozoon is merely a 

 rod of chromatin (Fig. 34) corresponding- to the descriptions of 

 many authors. 



Such in brief is the course of events in the small cell genera- 

 tions of spermatids, but in the large cells it is quite different. In 

 these, the layer of chromatin lining the nuclear membrane, so 

 characteristic of the small spermatid, is from the first more or 

 less vacuolated, and soon after the head begins to elongate (see 

 Montgomery ('n), Fig. 141) the layer breaks down completely. 

 The head now rapidly elongates, greatly exceeding the small 

 sperm in this particular, while the chromatin content remains in 

 an indefinitely vacuolated condition. Finally, the chromatin con- 

 denses into a more or less completely spiral thread which ulti- 

 mately forms an axial rod as in the case of the small sperm, but 

 of very much greater length (Fig. 35). 



In brief, the two sizes of spermatids in Murgantia yield sperms 

 of very different lengths by processes of nuclear differentiation 

 which, while superficially unlike, prove to be fundamentally 

 similar. 



The centrioles in the Hemipteran spermatid have been so 



