STUDIES ON INSECT SPERMATOGENESIS. 33! 



in the middle prophases that the results become clearly evident. 

 Then it appears that the Golgi bodies, moving from all parts of 

 the cell, have been gradually forming a narrow belt or girdle en- 

 circling the nucleus in what subsequently proves to be the equa- 

 torial region. Viewed from one pole, the discs form a more or 

 less complete ring (Fig. 8) and are in general so oriented that 

 their flat surfaces are turned toward the nuclear membrane, the 

 edges with the characteristic " split " being thus directed toward 

 the observer. The middle prophase now passes rapidly into the 

 late prophase and in this stage the ring of Golgi bodies undergoes 

 a most characteristic change. This consists of an autonomous 

 fragmentation of all the Golgi bodies so that in the place of the 

 ring of large discs, there is now a throng of much smaller bodies 

 (Figs. 9 and 10). In some, a composition similar to that of the 

 growth period discs can often be made out, but usually from this 

 point on the morphological distinctions become more or less ob- 

 scure owing to the small size of the bodies and faults in impreg- 

 nation. In the small cells each Golgi body appears to be divided 

 into two parts, but I do not wish to emphasize this point. In 

 the large cells each of the Golgi elements fragments into many 

 pieces, and the process of fragmentation can actually be fol- 

 lowed. It is certain that the division is always at right angles to 

 the "split," and the latter, therefore, however well adapted it 

 may appear for purposes of division, never serves that end. 

 Further, in the large cells the division clearly includes the cen- 

 tral portion of the disc, resembling very strongly the figures given 

 by Gatenby as representing the division of a Golgi body (see 

 Gatenby ('19^), Fig. 14). The nuclear membrane now fades 

 away and the metaphase figure is rapidly formed. 



Before taking up the actual division stages, a word may be said 

 concerning the nomenclature of the Golgi elements. The scat- 

 tered condition of the Golgi apparatus throughout the growth 

 period renders the use of the term Golgi " bodies " a most con- 

 venient one. Their fragmentation introduces a complication 

 which for practical purposes I propose to avoid by the use of the 

 term dictyosomc, a word coined by Perroncito (see especially 

 his paper of 1910) to denote the pieces into which the Golgi 

 apparatus was divided during mitosis. If its meaning be re- 



