THE ENDOSKELETON 169 



maining unsubstantiated by a parallel series of embryonic 

 stages, rests upon an insecure foundation and is liable to re- 

 ceive considerable modification through the discovery of ad- 

 ditional facts. In tracing these histories the anterior and 

 posterior limbs must be treated separately, since, although the 

 free limbs are plainly serially homologous, and often corre- 

 spond quite closely, part for part, the girdles, although equally 

 a portion of the appendicular skeleton, differ fundamentally 

 from one another and must have had a somewhat different 

 early history. 



Beginning with the posterior limb, which is more conserva- 

 tive than the anterior, and probably retains more primitive 

 characteristics, it may safely be supposed that at its origin as 

 a localized flap ofj^once continuous fin-fold, its skeleton con- 

 sisted of a series of spines or rays, independent of one another, 

 and somewhat longer at their bases, tapering to their free 

 extremities (Fig. 43, a). To insure strength and to gain a 

 concerted action, a very natural step would-be to widen these 

 at their bases still farther until they fuse, forming a piece 

 something like a comb, with long teeth far apart (Fig; 43, b). 

 As these organs become of still greater importance and need 

 a firmer support, the basal portions, corresponding to the 

 backs of the combs, would be likely to grow inwards until 

 they meet and fuse across the mid-ventraMine^ thus forming 

 a very primitive girdleTwitrT which the free part would become 

 movably articulated (Fig. 43, c). This last case, is, however, 

 not a hypothetical one, but drawn directly from the posterior 

 girdle and free limb skeleton of the dog-fish, a typical se- 

 lachian, in which the free limb consists of a basipterygium, 

 bearing a series of rays, the whole being movably attached to a 

 girdle in the form of a ventral band. Although there is at 

 present an impassable space between this free limb and that 

 of even the simplest amphibian, the girdles of the two forms 

 are not so far apart, since a broadening of the middle piece 

 into a plate, and the extension of its ends dorsally until they 

 come in contact with a pair of ribs, would convert the one 

 into the other. (Cf. Fig. 44, d.) 



