STUDIES ON INSECT SPERMATOGENESIS. 335 



variously described, that a comparative analysis is out of the 

 question in this paper. However, in view of the fact that I have 

 been able to clear up some of the doubtful points in Mont- 

 gomery's ( T n) description of Euschistus, a very brief statement 

 of the facts may be made as I have found them in Murgantia. 

 In the very young spermatids the centriole is not single as Mont- 

 gomery stated, but distinctly double (Fig. 22). This is especially 

 clear at the time when the centrioles are migrating from their 

 original anterior position to their definitive position between the 

 nebenkern and the nuclear membrane. At this time they are 

 always placed in a line normal to the nuclear wall, so that the 

 proximal centriole touches the nucleus while the distal one gives 

 rise to the axial filament of the tail (Fig. 22). Arrived in their 

 position at the future base of the head, they undergo a change in 

 orientation and for a short time their duplicity is lost, and only a 

 single granule-like centriole can be distinguished. A little later, 

 when the halves of the nebenkern have started to elongate, the 

 two centrioles can again be made out in favorable cases. They 

 appear in the form of two short rods, joined at one end to form 

 a "V," at the vertex of which the axial filament is inserted (Fig. 

 26). At first the rods lie tangentially (or nearly so) to the 

 nuclear membrane, and as seen from the side give the impression 

 of a large granule to one edge of which the axial filament is at- 

 tached, an asymmetry noted by Montgomery. Then the stray 

 chromatin masses which have not taken part in the formation of 

 the chromatin layer on the inner surface of the nuclear mem- 

 brane begin to collect in the vicinity of the centrioles where 

 they eventually form a compact and conspicuously staining body 

 (Fig. 29). This body I have called the pseudo-blepharoplast 

 because it has always been regarded as the centriole of the sper- 

 matozoon. Indeed the pseudo-blepharoplast so easily obscures 

 the true centrioles, that lacking a study of earlier stages its real 

 nature would certainly be overlooked. The rod-like centrioles 

 eventually straighten out parallel to the long axis of the sperm 

 (Fig. 29), and somewhat later the pseudo-blepharoplast fades 

 away, as though it were dissolved in the nuclear sap. This phe- 

 nomenon was correctly described by Montgomery who was natu- 

 rally puzzled by the apparent solution of the spermatid centriole 



