540 CAROLINE BURLING THOMPSON 



on the outside of the cup and are termed by Joneseu the "gan- 

 glion cells of the wall." 



The conditions described by Joneseu in the honey bee bear a 

 very close resemblance to those of ants. The three cell groups 

 of the bee, the median and two lateral, situated in the cup, 

 are evidently homologous with the Groups I, II, /., II, r., of 

 ants. The wide calyx of the bee contains three cell groups with- 

 in it, but from the narrower cup of the ant the cell groups II, L, 

 and II, r., have been crowded upward and outward. The fibers, 

 however, are completely homologous in the two insects; the cen- 

 tral fibers of the stalk, from the median cell group, in the bee 

 correspond to the fiber tract h in ants, and the more lateral 

 fibers of the stalk, from the lateral cell groups, in the bee corre- 

 spond to the tracts g of ants. It should be noted that Joneseu 

 has found the large median cells in the worker, for his figure 36 

 is from the brain of a worker bee. It would be interesting to 

 know whether the slight but constant differences observed between 

 the cells of the Groups III and IV in ants may not also exist in 

 the "ganglion cells of the wall" of the honey bee. 



c. Fiber tracts of the mushroom bodies 



Definite fiber tracts arise from the cell groups and can be traced 

 into the stalks either of the same or of the adjacent lobe, or out of 

 the stalks into the fibrous core of the protocerebral lobes. The 

 origin of some of these fiber tracts in the cell groups can be seen 

 with such distinctness that their efferent nature may be assumed; 

 but although afferent tracts are doubtless present I have not 

 been able to distinguish them as such. 



The tracts observed are the same in both right and left mush- 

 room bodies, but differ in the outer and inner lobes of a single 

 mushroom body. The detailed descriptions which follow show 

 that many tracts are present throughout the different castes and 

 genera, while others are confined to one caste, and still others 

 are variable. The tracts that are of constant occurrence in all 

 the castes of the three genera here described are: 6, fl, h, gl., 

 gr. B is really a protocerebral fiber tract, uniting the outer and 

 inner protocerebral regions, but is connected by fibers with the 



