530 LOCY. [Vol. XI. 



Salamandra maculosa he found only four segments in the head 

 region, while Kupffer found eight in the same region of Sala- 

 mandra atra (Fig. 117). All other observers have found eight or 

 more in the head region, and Froriep stands alone in identify- 

 ing so small a number of segments in the head. In Rana 

 palustris there are about four obvious divisions in the median 

 plate of the head (but, in addition, there are other segments in 

 the neural ridges to the number of ten in the same region). 

 This affords a suggestion that the segments observed by Froriep 

 (see Fig. 116) correspond to those I have observed in the 

 median plate of Rana palustris, and not to the segments in 

 the neural ridges, which are evidently homologous with those 

 in the same position in Amblystoma, Acanthias, and other 

 forms. 



Whether we find the median plate smooth in Amblystoma 

 or faintly segmented depends upon the stage at which the ex- 

 amination is made, and we recognize that the appearances in 

 any one ^^^ are not constant throughout the open groove 

 stage; and, further, that eggs of closely related animals are by 

 no means necessarily similar at corresponding stages. The 

 failure to take things of this nature (which are really common 

 enough) into consideration has, I think, given origin to many 

 differences of opinion and formed the basis of many a contro- 

 versy. This should make us careful in comparing results to 

 have the same material in precisely the same stages. 



In Diemyctylus the conditions are very similar to those in 

 Amblystoma. The neural folds show metameric divisions in 

 the stages with wide-open neural groove. In the three am- 

 phibian forms I have examined there are about ten pairs of 

 segments in the broadly expanded neural folds of the head. 



In Torpedo ocellata I have also noted the occurrence of this 

 metameric segmentation in several stages. The youngest 

 embryos of that animal I have had corresponds to the stage 

 designated " C " by the Zieglers, and beginning at that point 

 I have traced it along through several stages with a widely 

 open neural groove. This form is so closely related to the one 

 I have especially studied that we would naturally expect — as 

 is the case — marked similarity between the neural segments. 



