50 



differences incidental to its stao-e. The above description 

 should be compared with the appaiently similai' one in 

 Sebastolobus (Starks, op. cif.), and with the very dissimilar 

 one occurring in the herring.* 



If the last distinct vertebra (fig. 14, 7.41], 19, F.42) 

 be now examined, it will be seen that the neural spine 

 {N.SA'i, 42) resembles the haemal spine {H.SM, 29) in 

 structure, but that both are peculiar. Each consists of a 

 partly cartilaginous shaft behind and a thin laminate 

 portion in front. The posterior shafts so closely resemble 

 the succeeding epural and hypural bones respectively as 

 to suggest that an epural above and an hypural below have 

 fused on to the laminate portions, which latter are un- 

 doubtedly similar to and perhaps represent the neural and 

 haemal spines in front. As, however, we have no positive 

 evidence of such a fusion, the spines in question are here 

 described as simple neural and haemal spines. 



Wedged in between U + Hjj. -^ and N.S. 48 (fig. 14) 

 are two partly cartilaginous spines (Kp. 1 and 2) which 

 are closely connected by ligament with each other and 

 with the last neural spine, but which are not sufficiently 

 long proximally to reach the vertebral column. The 

 gradual increase in length of these spines as the animal 

 grows older suggests that they may in senile forms become 

 connected with the vertebral column. That they develop 

 in the same way as the neural spines seems certain, 

 although there are no vertebra ostensibly belonging to 

 them. One is appreciably larger than the other, and there 

 is a big gap between the second and U. + Hj). 3. They 

 I'epresent the epural or epiural bones of other authors. 



The second hypural bone (H/>. 2) is of the same shape 

 and structure as the upper piece (U + Hp. 3), both being 

 cartilaginous distally and strongly ossified proximally. 



* Duncan Matthews, Fishery Board, Scotland, Report v., 1886. 



