THE EMBRYOLOGY OF THE HONEY BEE 161 



so closely associated with the ganglion cells of the ventral cord 

 and so like them in appearance. It is still less probable that they 

 owe their origin to the median cord since at the time they are 

 first evident the median cord is just separating from the hypo- 

 dermis, and moreover its cells are still relatively large and pre- 

 serve their original elongate form. There accordingly remain 

 two possibilities : either the neurilemma cells are, as Heymons 

 suggests, split off from the dermatogenous layer ; or else they are 

 merely transformed ganglion cells. The former view has much 

 to commend it, from a theoretical standpoint and, as far as the 

 writer's observations go, can not be totally excluded. The evi- 

 dence at hand however seems to favor the latter view. Prior to 

 Stages XI-XII the line of separation between the lateral cords 

 and the future hypodermis is coincident with the outer boundary 

 of the neuroblasts, in other words there is no direct evidence that 

 any considerable number of dermatogenous cells are split off 

 from the peripheral layer; on the other hand at the time the 

 lateral cords are definitely split off a rearrangement of the cells 

 takes place in the intraganglionic regions of the lateral cords, at 

 which time a number of small cells appear on their ventral sur- 

 faces. These cells closely resemble the ganglion cells. It seems 

 probable therefore that in the honey bee the neurilemma owes its 

 origin to the ganglion cells themselves. 



E. Corpora Allata 



The origin and development of the corpora or ganglia allata in 

 the honey bee corresponds very closely to the account given by 

 Heymons (1895) for Forficula. In both insects these bodies arise 

 as ectodermal ingrowths located in front of the bases of the first 

 maxillae. In the honey bee this is also the location of the tubular 

 invaginations which fprm the apodemes, of the adductor muscles 

 of the mandibles. Each of these ingrowths at Stage X has already 

 become a long hollow finger-like structure curving dorsad and 

 caudad. The mouth of this invagination is wide, invading the 

 base of the maxilla on its anterior and lateral sides. At Stage 

 X there may be found near the outer and caudal end of each of 

 these openings a small tubular invagination extending mesiad 

 into the base of the maxilla. These invaginations are the rudi- 



