more rostral position of the girdle on the side-fold theory, we might
expect to find some trace of this in the number of post-girdle vertebrae
that is, we might reasonably look for a larger number of whole
vertebrae behind the girdle-piercing nerve when the girdle is situated
more rostrally than when it is more caudally placed. A glance at
Table XIII will show that this is only to a very small extent the
case. Whilst the girdle shows a difference in position amounting to
five segments, there is a difference of less than one between the
average number of post-girdle whole vertebrae associated with its
extreme positions.
On the migration theory such difficulties are not encountered. The comparative constancy in the number of post-girdle whole-vertebrae for all positions of the girdle is to be explained by supposing that there exists some relation between the position of the girdle and the point where that more flexible portion of the body the tail with its half-vertebrae commences. The position where the " Anlage " of the pelvic fin is laid down must be supposed to determine, probably for mechanical reasons (cf. Gadow (6), p. 195), the point where the half- vertebrae shall start. Moreover, supposing the migration to be secondarily in a rostral direction helps us to understand why the number of post-girdle whole- vertebrae tends to be rather greater for the most rostral positions of the girdle, than for the intermediate or most caudal positions. We must imagine that the conversion of whole into half-vertebrae tends on the whole to keep pace with the rostral migra- tion of the girdle. As the migration of the girdle attains its maximum the formation of half-vertebra? tends, so to speak, to lag behind, and as a consequence we find that in the most rostral positions of the girdle a rather larger average number of post-girdle whole-vertebra?.
Some Embryological Data.
The material used consisted of Acanthias embryos which had been preserved in corrosive sublimate and acetic acid. Horizontal longi- tudinal sections were cut through the pelvic area. These were treated with gold chloride and formic acid as described in a previous paper ((12), p. 344). Such portions of the plexus Avhere the inter- communications of the nerves was required, were re-constructed on millimetre paper. The results may be tabulated as follows :
