NEA.L: NEKVOUS SYSTEM IN SQUALUS ACAXTHIAS. 175 



The structure of the myelomeres in embryos of 40-50 somites is 

 represented in Figures 42 aud 43, Plate 6. As in Figure 39, only the 

 right half of the embryo is shown. The only evidence of the structural 

 peculiarities of neuromeres at this stage consists in an external constric- 

 tion opposite the myotome and the spinal ganglion (Fig. 43). In 

 sections dorsal or ventral to the one shown in Figure 43, even this con- 

 striction becomes lost (see Fig. 42, which is more ventral, occupying the 

 re'^ion of the ventral roots). All ti'aces of an internal dilatation and 

 constriction, and of the concomitant radial an-angement of cells, have 

 disappeared. In the head, on the contrary, the " neuromeres " still pre- 

 serve all the characteristics seen in the earlier stages. 



An examination of the structure of the myelomeres shows that the 

 conditions are easily explicable on the mechanical grounds stated. There 

 are no serial thickenings of the wall of the neural tube, as in the hind- 

 brain, and the radial arrangement of cells and nuclei shown in the frontal 

 sections (Fig. 39) presents no difficulty ; for the cells composing the 

 epithelium of the neural tube always have their long axes perpendicular 

 to the surface of the tube, so that, if the tube becomes constricted oppo- 

 site each somite, the cells will necessarily show a radial arrangement in 

 frontal sections. In view of this fact, it is difficult to understand how 

 investigators should have thought that the existence of a radial arrange- 

 ment of cells and nuclei was evidence sufficient to establish the morpho- 

 logical value of myelomeres, and their serial homology with hiudbrain 

 neuromeres. McClure ('90), for example, says, " The lateral walls of the 

 spinal cord are divided into neuromeres which, while less conspicuous, 

 have all the cellular characteristics seen in the typical neuromeres of the 

 hindbrain, and in fact are a continuation of the latter." That all of the 

 cellular characteristics seen in the typical neuromeres of the hindbrain 

 are also found in the myelomeres is demonstrably untrue for Squalus, as 

 may be seen by comparing the sections shown in Figures 38 and 39, Plate 

 6, both from the ventral half of the neural tube of the same embryo, one 

 in the head and the other in the trunk. The cellular arrangements are 

 decidedly unlike. In the head (Fig. 38) the cells and nuclei are crowded 

 in the region of constriction between neuromeres, while in the trunk, 

 if the cells are crowded at all, it is in the region of dilatation of the 

 myelomere. 



It has seemed a strong argument for the serial homology of myelo- 

 meres and hindbrain neuromeres that the former continue into the latter 

 gradually and in an unbroken series. For example, McClure ('90) 

 stated that " the constrictions of the myelon (in Lizard embryos) gradu- 



