INTERSTITIAL CELLS OF TESTICLE IN DIDELPHYS. 1 87 



shortly be discussed. In most cases, it is quite apparent that 

 this peculiar formation does not invade the entire cell-body, but 

 leaves two parts free; first, a space located in the immediate 

 neighborhood of the nucleus, which it is safe, I believe, to con- 

 sider as corresponding to the idiosome; second, the periphery 

 of the cell. There are, however, exceptions to the latter rule, 

 as in a number of cases some trabeculse are seen very close to the 

 cell-boundary. 



The question arises: is there any connection between this 

 network and the intracellular accumulations of the secretion 

 product? In many cases it would appear as though both forma- 

 tions were in continuity, and that consequently the network were 

 merely an extension of the intracellular processes throughout 

 the cell-body. I am, however, of a different opinion. To me 

 the network is the product of a transformation of the chondrio- 

 somes under special conditions of imperfect preservation. I am 

 perfectly aware that no such forms have been described hereto- 

 fore. The only thing with which I can, to a certain extent, 

 compare these structures are the chondriosomes in the seminal 

 cells of Scolopendra cingulata as represented (in my opinion 

 imperfectly preserved) by Bouin (1905) especially in his Figs. 

 13, 14 and 15. The basis for this interpretation is that this 

 network is found only in material fixed with Benda's or Meves' 

 fluid (after Altmann's fixation the interstitial cells kept the acid 

 fuchsin with such tenacity that no definite conclusion could be 

 reached), but only in the deeper parts of the pieces; that is, 

 where the chondriosomes are not well preserved. It is not found 

 in material fixed in Regaud's fluid, but I assume that, as the 

 latter reagent penetrates better than those just mentioned, it 

 never gives the chondriosomes a chance to produce such distorted 

 figures. As a matter of fact, they are pretty well preserved all 

 through the pieces, and the only modified forms that appear in 

 that material are the vesicles the well-known first stages of 

 chondriolysis. On the other hand, the secretion product is 

 preserved by a number of reagents which do not fix the chondrio- 

 somes. The staining reactions are also different and point toward 

 the chondriosomal nature of the network. Generally speaking, 

 the network is stained in purple after Benda's method, that is, 



