324 EDWIN E. REINKE. 



\Yhen the bundle of axial fibers, by its growth, has increased 

 the length of the cell by about one third, it splits and the halves 

 begin to move to either side of the cell (Figs. 10 and u). This 

 movement is caused by a twisting of the secondary bundles, due 

 probably to the unequal growth of the fibers. This process 

 continues until they push out the cell membrane on either side to 

 form the undulating membranes (Figs. 12 and 13). Fig. 12 is an 

 oblique section through the middle of a spermatozoon which has 

 not quite matured; later the membranes become flatter and 

 wider as shown in Fig. 13, which is a more anterior section of a 

 still older spermatozoon. As is indicated in Fig. 12, the con- 

 tinued growth of the fibers without a compensating increase in 

 the length of the cell causes several even folds to occur throughout 

 the length of the membranes. Some of the axial fibers do not 

 participate in the formation of the membrane. As the bundles 

 begin to evaginate the cell membrane, a few of the fibers begin 

 to migrate and finally come to lie longitudinally across the cell 

 and just beneath its membrane (Figs. 12 and 13). 



The albuminous bodies are formed in the same way as Kuscha- 

 kewitsch ('n) has briefly described in Vermetus gigas. Large 

 vacuoles appear in the cytoplasm, first in the anterior portion of 

 the cell, and these gradually become filled with an albuminous 

 substance. When such a vacuole has been almost filled but 

 before a membrane is formed, narrow strands may be seen 

 connecting the albumen with the surrounding cytoplasm (Fig. 1 1). 

 The cytoplasm which is not displaced by the formation of these 

 bodies becomes fibrillar, the greater part of it forming a core down 

 the center of the spermatozoon (Figs. 12 and 13). 



As the spermatozoon develops, the vesiculated nuclear frag- 

 ments which were scattered throughout the cell continue to 

 degenerate. They gradually become more condensed and darkly 

 staining and undergo further fragmentation. They may dissolve 

 in situ, but they may also go to form the many small granules 

 that lie in amongst the fibers composing the undulating mem- 

 branes (Fig. 13). These granules are probably mitochondria 

 and while at present it cannot be definitely asserted that they 

 are thus of a direct nuclear origin, there is considerable evidence 

 in favor of this view. The mitochondria which originally sur- 



