72 EDWARD MCCBADY, JR. 



several facts seem to indicate that only a few of these cells 

 actually contribute to the ganglia. In the first place the cells 

 thus derived are perfectly continuous with the mesenchyme of 

 the whole adjoining branchial arch. It is not until stages 25 

 and 26 that the ganglionic rudiments become fairly distinct 

 from the mesenchyme. In the second place, when the ganglia 

 do become distinct they are composed of a relatively small 

 number of cells much smaller than the total number which 

 must have been proliferated during stage 24. And finally, 

 there is the fact that until this proliferation begins there is 

 no recognizable mandibular or hyoid mesenchymatous mass. 

 There is mesenchyme in the head region from stage 20 on, 

 but only a thin and even layer. Simultaneously with the pro- 

 liferation of neural crest in stage 24 the mesenchyme becomes 

 so abundant anterior and posterior to the first branchial 

 pouch that it shows up in transmitted light as two dark masses 

 (fig. 20, totomount), which thereafter may be followed through 

 later stages as the mandibular and hyoid arches. 



It seems necessary to believe, therefore, that a considerable 

 portion of the mesenchyme of these branchial arches is de- 

 rived from mesectoderm proliferated at the sites of origin of 

 the gasserian ganglion and the acoustico-facial complex. It 

 should also be mentioned that the visible evidence in the 

 opossum does not support the view that the otic placode con- 

 tributes any cells to the acoustico-facial complex, but in the 

 light of experimental evidence from Amblystoma this visible 

 evidence must at least be held inconclusive. 



Posterior to the second branchial pouch the neural crest 

 is again fairly abundant at the site of origin of the glosso- 

 pharyngeal ganglion and the third branchial arch. In fact, 

 in figure 19 a third opacity is quite distinctly recognizable 

 in all three photographs. 



In the cordal region of 17153 (early stage 24) there is very 

 little neural crest proliferation, but in later specimens as 

 17166 and 17139 it is well formed. A consistent difference be- 

 tween the neural crest in the cranial and spinal regions seems 

 to be that the former arises from a broad region extending a 



