174 Journal of Comparative Neurology. 



to the relation of the glomerulae to the surrounding cells is cor- 

 rect and even though some of my preparation are so strongly- 

 impregnated with bichromate of silver as to present consider- 

 able resemblance to a meshwork, I am inclined to believe other 

 preparations showing a tuft-like system of branching illustrate 

 the true condition (fig. lo and i6, PI. XXI), and that the idea 

 of connection by contact holds as well here as elsewhere. As 

 already pointed out cases might occur in such prepara- 

 tions in which the darkened fibers might appear perfectly con- 

 tinuous, so as not even to be explained by Lenhossek's sugges- 

 tion of a silhouette of superimposed fibers, and still not be 

 fused or continuous. Happily such deceptive appearaces 

 seem to be rare except in too heavily impregnated preparations. 

 When impregnated with bichromate of silver the glom- 

 erules strongly recall the appearance of the olfactory glomerulae 

 figured by Retzius ( 92) for the vertebrates, and may perhaps be 

 very well called olfactory glomerulae. But exactly what their 

 relations may be with the surrounding cells and the incoming 

 fibers I am not able to say definitely. In none of my prepara- 

 tions is an unbroken connection shown between them and the 

 cells clothing the lobes, the fragmentary appearances, however, 

 amply warrant the supposition that the processes of the cells 

 make their way between the bodies and after passing some dis- 

 tance in the middle mass of fibers send a branch into one of the 

 glomerulae and another one into the proto-cerebron (fig. 10, 16 

 and 24, PI. XXI). Whether the process from a single cell sends 

 branches to more than one body does not appear, but if one 

 may judge from the form of the cells of the mushroom body, 

 probably it does not. 



THE root of the ANTENNO-SENSORY NERVE. 



Besides the fibers just described there are to be distin- 

 guished two other kinds. One of these composes the root of 

 the antenno-sensory nerve, which in preparations with osmic 

 acid or by the copper-haematoxylin method may be traced from 

 its entrance into the lobe on the lower anterior surface nearly to 

 the posterior upper surface. Throughout its course it gives off 



