FINS OF ELASMOBRANCHS. 469 
attachment of the shoulder-girdle. Now, inasmuch as we have seen that the pectoral 
fins must have an antero-posteriorly narrow point of support, so it is also necessary for 
that point of support, if it seeks a firm basis, to do so by extending, not antero- 
posteriorly, but vertically ; and thus it has come to abut against its fellow of the oppo- 
site side ventrally, and against the ventral column (as in Raia), or against the skull (as 
in the Teleostei) dorsally. It must also grow in two directions, and so develop into a 
limb-girdle, as, had it but one oblique attachment, it would not be firmly fixed. 
In this way the mystery of the limb-girdles seems to be satisfactorily explicable ; 
they are neither modified hypaxial nor paraxial parts. They have, as I have before 
contended, neither the nature of branchial arches, nor of modified ribs, but are parts 
respectively sud generis, due to the ingrowth of originally superficial structures—exo- 
skeletal hardenings which have grown inwards, and become endoskeletal. 
It might possibly be objected against these views, that in fishes which have no 
pectoral fins, such as Amphipnous, Monopterus, and Symbranchus, there is, none the 
less, a pectoral arch, and one, moreover, which, coming, as it does, just behind the 
branchial arches, and being of very small size, has much the appearance of, and 
structure serially homologous with, these arches. 
But the absence of paired fins in these fishes is manifestly due to degradation, since 
their shoulder-girdle, instead of being a nascent rudimentary structure, has its normal 
Teleostean form, each lateral half consisting of three bones, the uppermost of which, 
in Monopterus and Symbranchus, forks proximally in the usual fashion. Therefore 
these fishes are in a similar case with the legless lizards, which are provided with a 
small, but normally developed, shoulder-girdle. 
As to the pelvic girdle, it might perhaps be expected that it would in some fishes 
take on a development and adhesion to the spinal column approximating to what we 
find in tailed Batrachians. But, in fact, no fish known to me uses its ventral fins at all 
in the way in which these Batrachians use theirs; and it is therefore not wonderful 
that the structure is never such as theirs. 
In some fishes, however (e. g. Cotylus cephalus, Chorisochismus dentex, and Sicyases 
sanguineus), the pelvic arch may acquire considerable complication. But this compli- 
cation is peculiar, and does not present any approximation to higher, non-piscine, pelvic 
structure. In Chimera andin Callorhynchus, however, the pelvis, as we have seen 1, does 
assume much the appearance of the pelvis of air-breathing vertebrates (Plate LXXIX. 
fig. 5), though it does not acquire a direct fixation to the spinal column. 
Altogether, I conclude then that the limb-girdles are ingrowths from the paired fin- 
skeleton, and that such skeleton is the modified remnant of a longitudinal series of 
similar radial parts, like those of the ventrals of Polyodon, formed primitively in a 
continuous lateral fold, such fold being similar, save in situation, to the folds which 
form the azygos fins, which therefore, together with the limbs, form one distinct genus, 
of parts of which the paired limbs are species. 
* See ante, p. 456. 
