6O B. F. LUTMAN. 



able free space between them. This free space is what makes 

 the task of following them, and their boundaries, so easy. About 

 two thirds of the primary cells, spermagonia, or "Ursamenzellen" 

 as they are variously called, have divided to form groups of 2, 4, 

 8, 1 6 or 32 cells while the other third is undivided. 



A closer examination of these undivided cells will now show 

 that in addition to the larger germ-cell there is lying closely 

 appressed against it a smaller kidney-shaped nucleus (Fig. 4). 

 The chromatin in both nuclei is distributed in the form of little 

 flecks rather regular in size and number and united by strands. 



The origin of this nucleus that lies in the epithelium of the 

 cyst is not clear. There is no such differentiation in the cells in 

 the testes at the preceding stage. If these epithelial cells were 

 in the testes at that time it was impossible to distinguish them. 

 It may be that it is an epithelial cell that has made its way in 

 some fashion into the interior of the testes and there surrounded 

 the germ-cell. It is a difficult question to decide and one on 

 which my material gives little evidence. 



The further development of the cysts of the follicle can be 

 readily followed in the two- (Fig. 5), four- (Fig. 6), eight- (Fig. 7), 

 sixteen- and thirty-two-celled (Fig. 8) stages. The cell-walls are 

 difficult to distinguish at these stages as the plasma membrane 

 which is all that surrounds the cells is very thin and the cells 

 are pressed tightly against each other. The same difficulty was 

 noted by St. George (12) who found, however, that he could dis- 

 tinguish the walls in material fixed in "quick acting reagents." 

 At any of these stages the nucleus of the cyst epithelium will 

 show. In the older cysts there appear quite frequently to be t\v<> 

 such nuclei, but in the younger stages at least I have observrd 

 only one. These nuclei do not seem to divide. 



The divisions in the secondary spermagonial cells are vrr\ 

 regular, all the cells in one follicle dividing at one time (Fig. 9). 

 All these divisions are typically mitotic. Sutton (14) found th.ii 

 the division spindles were largely tangential to the walls ot ilu- 

 cyst, owing to pressure in those directions. This caused the c> -4 

 to grow in length. As the cysts of the caddis-fly are spherical, the 

 divisions occur in equal numbers in .ill planes. It can be readily 

 determined that there are regularity 32 cells in each cy>t ot the 



