62 ROBERT H. BOWEN. 



(Fig. 278). To other workers this stage has looked like a thread- 

 work of some sort, an appearance due possibly to the disturb- 

 ances of fixation or the failure to stain clearly (but compare also 

 with Gatenby's description). Thus, compare Henneguy ('96) on 

 the " filamentous " nebenkern in P\rrhocoris, and other early 

 workers. Many observers have also noted the curious disposi- 

 tion of the " layers " of the nebenkern to take the stain differently. 

 This is particularly common after the osmic acid fixatives (Flem- 

 ming), the results being often very fantastic and unexpected. 

 Gross ('07) has figured a variety of such results in Pyrrhocoris, 

 and I have seen much the same thing in the pentatomids. The 

 most interesting of these phenomena is that of alternative stain- 

 ing, in which the different layers on one side of the mid-line are 

 stained in various degrees while in the other half of the nebenkern 

 the degrees of staining are exactly reversed. (See Gross ('07) 

 Figs. 100 151, 154, 155, and 156.) The meaning of these pecu- 

 liar phenomena is entirely unknown. 



The structural foundation of these patterns will be made some- 

 what clearer by a comparison of the nebenkern as seen from the 

 side (Fig. 5) and from one pole (i.e., in cross-section) (Fig. 6). 

 It is evident, in the first place, that the substance of the nebenkern 

 has become sharply differentiated into two substances of very 

 different staining capacities. One of these stains little, or not at 

 all, and forms a more or less homogeneous matrix in which is em- 

 bedded a mass of material that takes the usual mitochondrial 

 stains (crystal violet, for example). Further, this stained sub- 

 stance is arranged in very characteristic concentric rings, the 

 successive rings being separated by zones of the non-staining 

 matrix. These rings might possibly be taken for a tangled thread- 

 work comparable to the "spireme" of Gatenby ('17), but by fo- 

 cusing up and down it is clearly proved that these rings are not 

 threads at all but more or less extensive, shell-like plates. This 

 is made even more clear by Fig. 6 in which the nebenkern is cut 

 at right angles to that of Fig. 5. It is quite obvious that the ap- 

 parently concentric rings are in reality the optical cross-sections of 

 a more or less complete plate-work, of which the imperfectly 

 spherical plates are arranged one within the other in a concentric 



