174 bulletin: museum of comparative zoology. 



compared by him with the "neuromeres" of the medulla.^ Such sym- 

 metrical cross furrows on the widely expanded neural plate of embryos 

 of Salamandra atra were described by Kupfifer ('86), and considered by 

 him as remnants of a primitive segmentation. Since Kupffer believed 

 that at this stage there was no trace of mesodermic somites, he regarded 

 the segments as "primary," i. e. not formed secondarily in adaptation to 

 the mesodermal segmentation. Both Froriep ('92) and Wiedersheim ('92) 

 have, however, declared that mesodermal somites are present at the stage 

 described by Knpffer, and that the segments could be explained as the 

 passive result of the pressure of these somites. Locy ('95) finds in the 

 trunk region of embryos of S. acanthias with a widely expanded cephalic 

 plate that the lines of division between his " neural segments " may be 

 traced inward toward the median furrow, probably as the result of the 

 appearance of somites in that region. As stated on page 160, I have 

 failed to find this exact correspondence between the neural segments 

 of Locy and the somites. 



When the neural plate has closed to form the neural tube, the regions 

 of elevation between the furrows become constrictions, which however 

 affect only the ventral half of the tube, i. e. that portion against which 

 the somites lie (see Fig. 41, Plate 6). Neither frontal sections nor 

 cleared specimens give evidence of constrictions in the doi'sal half of the 

 tube. The constrictions in the ventral half of the tube are most clearly 

 max'ked in the early stages, when the mesodermal somites are most rounded 

 in form, and they disappear as this rounded form disappears.^ 



Figure 39, Plate 6, represents a frontal section in the ventral half of 

 the myelon of an embryo with 28-30 somites (Balfour's stage H). The 

 right half of the neural tube and of the mesoderm is shown. It is seen 

 that the wall of the neural tube shows a rounded constriction opposite 

 the somite, while opposite the cleft between two somites, and conform- 

 ing with it, an outer ridge and an inner rather sharp groove are seen. 

 This section affords evidence more favorable to the contention that 

 "neuromeres" exist in the spinal cord than that seen at any other stage 

 of development, or in any other plane of sectioning. In dorsal sections 

 of the same series the constrictions disappear, as do the somites also. 



1 Marshall (78) had previously stated that in the chick "the cord is slightly 

 constricted opposite the centres of the protovertebrag, and slightly dilated opposite 

 the intervals between successive protovertebrse." 



2 Miss Piatt ('89) has said with regard to the chick, " Here [in the trunk region] 

 as in the medulla, the segmentation is more manifest in the ventral region than in 

 the dorsal." 



