31 
the fibers frequently follow the boundary between the clear and gran- 
ular portions of the intermediate substance (Calandruccio, 1890a). Ac- 
cording to Mertens (1892) the fibers do not attach to the outer mem- 
brane, since this may be split off and the intermediate layer remain 
undisturbed. The peripheral portion, at least, of the intermediate 
substance seems thus to possess a certain solidity as though bounded 
externally by a membrane, which must, however be very thin, separat- 
ing it from the outer membrane. Mertens (1892), indeed, states that 
Fig. 43. — Eggs of H. nana and spherical bodies, perhaps eggs in course of development. Enlarged. 
(After Senna, 1889, fig. 2.) 
he has observed at times a very thin homogeneous layer between the 
intermediate substance and the outer membrane, and I have occasion- 
ally noted a similar appearance. It frequently happens that the inter- 
mediate substance shrinks awa^" from the outer membrane, or from 
the inner membrane, or from both. This results in the appearance of 
a third membrane intermediate between the other two, as occurs under 
similar circumstances in the egg of II. dhmnuta^ (p. 93). As in II. 
diminuta^ also, there are 3 or 1 comparativehMarge nuclei (visible in 
Fig. 44.— Eggs of H. nana. Enlarged. (After Senna, 1889, fig. 1.) 
sections) present in the intermediate substance which seem to play a 
part in the formation of the shell. W. H. Ransom (1856) and Senna 
(1889) describe a concentric striation of the intermediate substance. 
Grassi (1886b, 1887d) was the first to observe the filaments which he 
described as a long convoluted thread apparently attaching to the 
poles of the inner membrane. Blanchard (1891a) considers that Grassi 
(1886b, 1887d) and Senna (1889) were in error regarding the presence 
of filaments, and that they had erroneously interpreted folds or wrinkles 
of a third membrane. That filaments are present is a well-established 
fact, however, attested b}" a number of oliservers. The three pairs of 
