BLACKMAN: THE SPERMATOGENESIS OF SCOLOPENDRA. 65 



On the contrary, enlargements occur throughout its course, and these 

 often alternate with constrictions which, however, never amount to com- 

 plete interruptions of the thread. An extreme case of this condition is 

 shown in Figure 89, where the fibre is typically monilifonn. The other 

 extreme exists in the cell from which Figure 86 was drawn. Here the 

 axial filament, with the exception of its extreme proximal end, is cleat- 

 cut and of uniform calibre. 



The archoplasmic mass always envelops the proximal end of the axial 

 filament, and in many cases is at first so elongated as to enclose the 

 greater part of it. Before the formation of the filament has proceeded 

 very far, however, its distal end extends beyond the archoplasm and 

 stretches out through the cytoplasm toward the cell membrane (Figs. 86, 

 87). As a rule its course is at first tangential rather than perpendicular 

 to the surface of the nucleus. The filament seems to have no fixed rela- 

 tion to the reticulations of the cytoplasm, appearing sometimes to follow 

 the walls of the meshes, sometimes to pass through the meshes of the 

 network (Figs. 86-89). The distal end of the fibre may be either 

 embedded in the substance of the reticulum (Figs. 87, 88) or lie free in 

 the hyaloplasm (Fig. 89). In Figure 87 the part of the cytoplasmic 

 reticulum which is a direct continuation of the distal end of the filament, 

 is somewhat more dense than neighboring parts of the reticulum, and the 

 same is true to a certain extent in Figure 88. However, after careful 

 study of many cells at this stage I feel confident that the growing axial 

 fibre shows no constant relation to the cytoplasmic reticulum. 



Now, as regards the relation of the centi-osomes and axial filament. 

 When these structures have reached the vicinity of the nucleus, the 

 centrosome has given rise by fragmentation to from three to seven deeply 

 staining granules, which seem to bear a fairly constant relation to the 

 proximal end of the filament. In perhaps ninety per cent of the cells 

 examined the number of centrosome fragments found detached from the 

 axial filament was only two. Of the remaining ten per cent only a very 

 few cases exhibited more than three. When there are only two centro- 

 somes they bear a constant relation to the axial fibre as shown in Figures 

 87, 88, 90, 93, 94, and in numerous later stages. They are almost 

 symmetrically located on opposite sides of the fibre. Very often this 

 end of the axial filament is considerably enlarged, and this I interpret as 

 evidence that a portion of the centrosome still retains its direct connec- 

 tion with the filament (Figs. 86, 87, 88, 90 et seq.). As we shall see 

 later, these conditions persist throughout nearly the whole period of 

 transformation, being seen even in the nearly mature spermatozoon. 



