50 THE AMEKICAN MONTHLY [March, 



in that he gives every external evidence ot macrosmatic power : but by 

 tlie ahiiost total occlusion of his rhinocojle approaches structurally the 

 conditions found in the microsmatics. Physiologically he is macros- 

 matic ; morpliologically\\e. is microsmatic. 



Broca's studies were limited to the 7-atiers (rat-terriers) . He thinks, 

 however, that perhaps the cavity exists in other races of dogs. In some 

 cases he found an " apparent cavity," which he ascribed to the tearing 

 of the tissue by the knife used in cutting ; at other times there remained 

 a wider cavity " whose interpretation is doubtful." His specimens were 

 placed immediately into absolute alcohol for hardening. This he deemed 

 of especial importance since the substance forming his so-called ventric- 

 ular axis decomposed very rapidly, and after a few hours lost its con- 

 sistence, so that when sections were cut a cavity appeared which re- 

 sembled a true rhinocoele. 



In the present investigation, there were used the brains ot a rat-terrier, 

 spaniel, and of some half dozen other dogs whose breed was not deter- 

 mined. The results were practically the same in all cases. In addition 

 to the material just mentioned. Professor Wilder very kindly placed in 

 my hands for comparison the olfactory lobe of a coyote, Ca7zis latrans^ 

 and of a species of a wild Australian dog known as Canis dingo. 



The cavitiesof the brains were injected with the preservative in order 

 that the parts might harden uniformly and as nearly as possible in their 

 natural position. There should be no difficulty in detecting the presence 

 of a cavity in this region with a microscope ; and, if there is one, an 

 injection would make it very apparent and render assurance doubly sure. 



After some experimenting it was found that an alcoholic solution of 

 zinc chloride — 95'^o idcohol, lOO c.c. ; zinc chloride, 5 grains — was 

 the most efficient fluid for hardening purposes. 



Its action was rapid and even, and produced little or no distortion. 

 Sections hardened in this manner stain readily in either carmine or 

 hasmatoxylin. 



A transection of the dog's olfactory bulb at about the middle of its 

 length revealed a very small rhinocoele, whose boundary showed no in- 

 dication of having been torn by the knife, since the endymal cells still 

 retained their normal appearance. Very rarely did they preserve an 

 unbroken continuity. At one end there was usually an interruption 

 through which was projected a tongue-like extension of neuroglia. 



The process of occlusion appears to be the result of the intrusion of 

 the neuroglia into the rhinocrele, accompanied by the gradual growing 

 together of the walls. The lateral diameter being the smaller, the sides 

 come together first ; but the width of the cavity is not uniform, and there 

 may exist two or more ver}^ small cavities between the points of contact, 

 each surrounded with its endymal cells, except at the point of interrup- 

 tion. At the points of contact, the cells become obliterated, the neu- 

 roglia thrusts itself into the cavities, and ultimately fills them more or 

 less completely. 



Imbedded in the neuroglia are numerous small granules which take 

 the stain in the same manner as nuclei, but around which no cell-bodies 

 were seen. Small blood-vessels follow these intrusions of the neuroglia 

 and ramify in all directions. 



In the crus, as might be expected from the narrow canal-like exten- 

 sion of the rhinocoele, the occlusion first becomes complete. In the now 



