606 H. GRAHAM CANNON 



form ii continuous thread. Tliis process may be possi?jle ; but 

 what Gatenby has overlooked is the fact that the circles in his 

 first figure, and })resumably the drawn-out loops in his second 

 figure, do not actually represent loops of thread at all but 

 surfaces. 



I do not wish to maintain that it is impossible to obtain from 

 a system of mitochondrial bodies, such as Gatenby describes, 

 a mitosomal spireme. If it is imagined that the chromophilic 

 substance disappears from the interfaces between the adhering 

 bodies and remains in the interspaces then some sort of a 

 network of chromophilic substance would be obtained which 

 miglit be described as a spireme, but this would not be a con- 

 tinuous single thread but a nuich-branching system of threads. 



In viewing the mitosome, at any level of focus, one sees 

 a coiled thread-like mitosome just as Gatenby figures, but on 

 focusing up and down one is able to see that without any 

 doubt it is actually a plate work" foi'med by a system of vacuoles. 



Summary. 



The main results of the examination of cytological prepara- 

 tions of the testes of the horse-louse (Ha emat opinus 

 asini), of the dog-louse (Lignognathus pi lifer us), and 

 of Ha emat opinus consoln'inus may be summarized 

 briefly. 



1. In all main points the spermatogenesis of these three 

 species of louse agrees with that described for Pedi cuius 

 corporis and Pedi cuius capitis by the late Professor 

 Doncaster and the present author. Miss Foot's account of the 

 spermatogenesis of Pedi cuius vestimenti is criticized. 



2. In the elongating spermatid of the horse-louse, the 

 nucleolus appears for a short period as a chromatic mass 

 adhering to the nuclear membrane, projecting partly in and 

 partly out of the nucleus. 



3. The acroblast of the horse-louse is usually a double body 

 consisting of two ' bun-shaped ' halves which are sometimes 

 of equal size, and which are separated with flat sides together 



