In most other Vertebrates the pectoral girdle has shifted backwards along the vertebral column, by a "metastasis" (Geoffroy) similar to that by which the pelvic fins in many fish have shifted up close to the pectoral girdle. The scapula (with supra-scapula) is the pleurapophysis, the coracoid the haemapophysis, of the occipital vertebra. The clavicle is homologised with the slender bone in fish now known as the post-clavicle, which shows a connection with the first or atlas vertebra of the vertebral column, forming, according to Owen, the haemapophysis of the atlas. Owen considers it no objection to this view that in other Vertebrates the clavicle is anterior to the coracoid--"its anterior position to the coracoid in the air-breathing Vertebrata is no valid argument against the determination, since in these we have shown that the true scapular arch is displaced backwards" (_On the Nature of Limbs_, p. 63, London, 1849). In the pelvic girdle the ilium corresponds to the scapula, the ischium to the coracoid, the pubis to the clavicle. Hence the ilium is a pleurapophysis, the ischium and pubis are both haemapophyses. The fore-limb is the developed "appendage" of the occipital vertebra, the hind-limb the developed "appendage" of the pelvic vertebra. They are serially homologous with, for example, the uncinate processes of the ribs in birds (see Figs. 5 and 6). The fore-limb is a simple filament in _Lepidosiren_, and presents few joints in _Proteus_ and _Amphiuma_; in other air-breathing Vertebrates it shows a more complete development, the humerus, radius and ulna, and the bones of the wrist and hand becoming differentiated out. As the fore-limb is equivalent to a single bone of the archetype, it is said to be, in its developed state, "teleologically compound" (p. 103). Since in the archetype every vertebra has its appendage, more than two pairs of locomotory limbs might have been developed. "Any given appendage might have been the seat of such developments as convert that of the pelvic arch into a locomotive limb; and the true insight into the general homology of limbs leads us to recognise many potential pairs in the typical endoskeleton. The possible and conceivable modifications of the vertebrate archetype are far from having been exhausted in the forms which have hitherto been recognised, from the primaeval fishes of the palaeozoic ocean of this planet up to the present time" (p. 102). It is not of the essence of the vertebrate type to be tetrapodal.