316 
Merism and Sex in. "Spinax Niger" 
(a) On the hypothesis that the half vertebrae do not extend to the end of 
the tail, but that at some point the whole vertebral condition is resumed, one 
would naturally look for a transitional portion such as occurs more rostrally 
between the whole and half vertebrae. No such portion is however to be found. 
The centra of the half vertebrae become gradually and uniformly smaller till 
they cease. 
(/3) Occasional variations are to be met with in which an unusually long 
centrum occurs among the half vertebrae (cf. Plate I, Fig. 8). Such a variation 
is easily explicable on the assumption that it is a whole vertebra which has failed 
to split into two half vertebrae. Such centra I have reckoned as equivalent to 
two half vertebrae. 
(7) Ridewood states that in Acanthias vulgaris the transition from half 
vertebrae back to whole vertebrae occurs at about the 24th vertebra from the 
caudal extremity. This statement led me to examine the position of the spinal 
nerves in three specimens of this species. The last few segments have no spinal 
nerves issuing out from them but the attenuated myotomes in this region receive 
their innervation from nerves passing out more rostrally. By opening the neural 
canal and using osmic acid the minutest caudal roots were evident. Their mode 
of exit was uniform to their end, i.e. over every even-immbered half vertebral 
centrum. Counting gave the following results : 
Total number 
Last halt vertebra 
of half 
over which a nerve 
vertebrae 
root passes out . 
Specimen A 
67 
46 
B ... 
63 
50 
C 
65 
48 
In these three cases there could not have been more than 21, 13, and 17 whole 
vertebrae respectively caudal to the half vertebrae, whilst from the arrangement 
of the myotomes Ridewood claims that there are 24. Taking all the evidence 
together, and admitting that the myotomes may show some difference of attach- 
ment at the end of the tail, there can, I think, be very little doubt but that 
the centra throughout the caudal region belong to the same class, viz. that of 
half vertebrae. 
3. Total segments. The number of these was obtained by adding the number 
of the whole vertebrae to the number of the half vertebrae divided by two. Odd 
half vertebrae were counted as a whole segment, e.g. a specimen with 44 whole 
vertebrae and 41 half vertebrae was reckoned as possessing 44 + 21 = 65 segments. 
4. Anterior spine. The position of this was determined by a sagittal section. 
It articulates most commonly with the basidorsal and interdorsal cartilages of a 
single vertebra (Plate I, Fig. 1 h). In such cases it has been reckoned as belonging 
