477 



Immediately surrounding the nerve strand is a single 

 layer of very small cells (fig. 227). They appear to be also 

 neuroglia cells, forming a support for the fibres of the nerve 

 strand which they enclose. At the time of defaecation they 

 are undergoing the same changes as are taking place elsewhere 

 at this period. Some of the large nucleoli are degenerating 

 and being absorbed ; others are in mit osis, and are evidently 

 going to replace theria. 



The absorption of larval cells, and the proliferation of 

 the imaginal elements within the nerve cord take place then, 

 at the time of defaecation, and are complete about eight to 

 twelve hours later. But at about the time of pupation other 

 changes which affect the gross anatomy of the ventral nerve 

 cord commence. These are the changes which Newport first 

 investigated, using Sphinx ligustri as his subject, and con- 

 sist in a remarkable migration of ganglia, resulting in the 

 fusion of these in groups to form the concentrated nervous 

 system of the adult. 



The regenerated nerve cord is composed of twelve ganglia, 



t connected in front by the circumoesophaeeal connectives with 



the brain (figs. 225a, 231). The first ventral (suboesophageal) 



ganglion fuses with the brain and will be considered in con- 



! nection with that structure. 



A little after pupation the sixth and seventh, and also 

 the ni nth and tenths ganglia fuse, so that the number of 

 v entr al ganglia has been reduced to ten (fig. 225c). In the 

 pu^a four hours old the eighth ganglion has merged into 

 the fused ninth and tenth, the number being now reduced to 

 nine (fig. 225d). By the fusion of the ganglion of the pro- 

 podeal (first abdominal) segment with the third thoracic 

 ganglion in the pupa eight hours of age and the absorption, 

 at^about this same period, of the eleventh (second last) ventral 

 ganglion into the fused eighth, ninth, and tenth, the number 

 of ganglia becomes finally reduced to seven. As the first 

 ventral ganglion becomes merged into the brain it is no longer 

 |i recognizable as a distinct ganglion (fig. 225e), and the ventral 

 I nerve cord cannot therefore be said to consist of more than six 

 ' ganglia. In this condition we find them in the imago. The 

 first three are large and lie one in each thoracic segment; 

 connecting the last thoracic ganglion (fused fourth and fifth) 

 with the first abdominal (fused sixth and seventh) is a par- 

 ticularly long nerve strand. The first abdominal ganglion is 

 rather small. The next two, especially the last, are much, 

 larger and supply the hinder and ventral region of the 

 abdomen. 



I have not observed the cellular activities which underlie 

 these migrations of ganglia; there can, however, it seems. 



