THE EMBRYOLOGY OF THE HONEY BEE 139 



degeneration in the Orthoptera (Viallanes 1891, Wheeler 1893, 

 Heymons 1895) an d also in the Dermaptera (Heymons 1895). 

 In Donacia, according to Lecaillon (1898), they simply disappear. 

 Burger (Carriere and Burger 1897), says (p. 370) in regard 

 to Chalicodvma: "I do not think that the cells I have designated 

 neuroblasts degenerate. Later they can not be distinguished 

 from the cells which they have produced." In the honey bee 

 the neuroblasts do not degenerate before hatching, since at Stage 

 XV they are conspicuously visible at the periphery of the ganglia 

 (and brain as well) and are frequently dividing mitotically 

 (Figs. 37A, 370, Nbl f see also Fig. 41). 



The development of the median cord, unlike that of the lateral 

 cords, appears to differ considerably in different insects. That 

 it is derived from a median strip of the ventral ectoderm, forming 

 the floor of the neural groove, seems to be at least certain. The 

 ultimate fate of this strip is less uniform. All investigators of 

 the subject with the exception of Wheeler agree with Hatschek 

 (1877) that the intragangl ionic sections of the median cord con- 

 tribute at least a large part if not all of the median portions 

 of the ganglia, including the transverse commissures. To this 

 opinion, however, Wheeler (1893) takes exception. This in- 

 vestigator, while admitting that in Xiphidium the intersegmental 

 regions of the median cord the progeny of the median neuro- 

 blasts are taken up into the central portions of the ganglia to 

 form functional ganglion cells, does not believe that the median 

 cord cells in the intrasegmental regions became ganglion cells, 

 but believes that they are used up in the formation of neurilemma. 

 In the same group to which Xiphidium belongs, the Orthoptera, 

 Heymons (1895) later found that the anterior and central median 

 gangliomeres were actually formed by the median cord, much as 

 in other insects. With this exception, the differences in regard 

 to the development of the median cord center principally about 

 the fate of the intersegmental (interganglionic) sections. Hat- 

 schek (1877) stated that these remained in connection with the 

 ectoderm and contributed nothing to the ganglia. 



Graber (1890) found that in Melolontha the median cord was 

 separated from the hypodermis throughout its entire length, but 

 that the intersegmental portions later divide transversely, each 



