332 
Merisni and Sex in " Spinax Niger'' 
position of the girdle), the larger the number of post-girdle nerves associated with 
that position. 
A considerable discrepancy will be noticed in the two sexes for the value of the 
correlation between the 1st g. p. n. and the collector nerves. The explanation is 
probably somewhat as follows. It has been shown by Braus ('Ol, p. 620) that the 
number of collector nerves in Spinax is greater during earlier than in subsequent 
stages of development. Even in early stages this rostral extension of the collector 
area has its limits. It is natural to suppose that with forward shifting of the 
girdle the more rostral of these branches will persist as the rostral part of the 
nervus collector. As however no nerve whose serial number is less than 21 seems 
to take part in the formation of the collector during early stages, any marked 
backward homoeosis in the pelvic area will be associated with a correspondingly 
marked reduction of the collector nerves. As the girdle shifts forwards the 
reduction in the number of collector branches will become disproportionately 
greater. Consequently, where the rostral pelvic shifting is less, we should expect 
to find a smaller correlation between the 1st g. p. nerve and the collector nerves. 
And this is actually what takes place. The $ s, with their more caudally situated 
pelvics and greater average number of collector nerves, show a correlation value 
of only 'S8, whilst that for the {/s, where the girdle is more rostrally placed 
and the number of collector nerves somewhat fewer, amounts to '55. 
(4) It will be noticed in Table 5 b that certain correlation values are very low. 
This is the case for those between the anterior spine and half vertebrae, between 
the total number of segments and the collector nerves, and between the fin ra3's 
and the pelvic nerves. On the hypothesis that some uniform process of excalation 
is going on the mean correlations for the above characters ought not to differ 
greatly from the mean of the rest, which is somewhere between "5 and "6. 
Especially should there be a high correlation between the number of fin rays and 
the fin nerves. On the side-fold theory of limb origin the morphological connection 
between these structures is so intimate that a high correlation between the two 
would naturally be looked for, and the fact that there is practically no correlation 
whatever tells strongly against this theory. Moreover the little that is known 
on this head for Acanthias, the only other Elasmobranch similarly investigated 
('Ol, p. 13), is in accordance with the foregoing facts. 
(5) A well-marked correlation exists between the number of the whole 
vertebrae and their total length relative to the whole vertebral column. As will 
be seen (in Appendix, Table 36) the length of the series of whole vertebrae 
varies between "605 and '695 of the whole vertebral column. The range is 
consequently "09, i.e. about l/7'5 of the length of the series, whilst the range for the 
number of whole vertebrae is 6 in a series of 48, i.e. \. The range of variation in 
respect of length therefore bears approximately the same relation to the relative 
length as the range for number does to the total number of whole vertebrae. The 
value of the correlation is fairly high but should be very much closer to unity if 
the number is entirely dependent upon the length. 
