In the remote predecessor of Anura, Amphibamus of the Pennsylvanian, the pectoral girdle was less substantial than in many of its contemporaries, but it contained the primitive median interclavicle in addition to the clavicle, cleithrum, and scapulocoracoid. (The figure of Watson, 1940, and that by Gregory, 1950, are of individuals of different ages, the latter being older.) It is clear that the paired elements of such a girdle were held rigid by their attachment to the interclavicle, via the clavicles. Subsequent elimination of the interclavicle in the Anuran line of descent, and decrease of ossification, left a girdle like that of Protobatrachus, Notobatrachus, Ascaphus and Leiopelma. But in several advanced families a more rigid median "sternum," of one or two bony pieces plus cartilage, is developed secondarily, possibly (as Cope, 1889: 247, suggested) in correlation with axillary amplexus.
Among Urodela no dermal bones occur in the pectoral girdle. There is usually a scapulocoracoid ossified as a single piece, from which a thin cartilaginous suprascapula extends dorsally and a broad cartilaginous coracoid plate extends medially, overlapping the one from the opposite side; a precoracoid lobe of this reaches forward on either side, and a median, posterior "sternum" of cartilage may make contact with the edges of the two coracoids. In Siren and Amphiuma two centers of ossification are found for each scapulocoracoid, and in Triton and Salamandra three. Probably the more dorsal and lateral of these represents the primitive scapula and the other one (or two) the primitive coracoid.
Comparing the girdle of a salamander with that of a frog, the closest similarity can be seen between Ascaphus and a salamander in which the scapula and coracoid ossify separately. Both have the median "sternum" in contact with the coracoid plates. The major difference, of course, is the lack of clavicle and cleithrum in the salamander.
CARPUS AND TARSUS
In Ascaphus (Ritland, 1955a; cleared and stained specimens of nearly grown males) distal carpals 1, 2, 3 and 4 are present and separate, increasing in size in the order given (Fig. 9). A prepollex rests against centrale 1; centralia 2 and 3 are fused; the radiale fuses with centrale 4, and the intermedium fuses with the ulnare; radius and ulna are fused with each other as in other frogs. The digits (and metacarpals) are considered by Ritland to be 1-4, in addition to the prepollex, rather than 2-5.
Fig. 9. Skeleton of fore foot of Notobatrachus (after Stipanicic and Reig, 1956, terminology revised) and Ascaphus (after Ritland, 1955 a); all ×5. For explanation of abbreviations see Fig. 3.
In the Jurassic Notobatrachus Stipanicic and Reig (1956) have shown the carpus with surprising clarity (Fig. 9). If their nomenclature of the parts be revised, we obtain a fairly close resemblance to Ascaphus, except that centralia 2 and 3 are not fused, distal carpals 1 and 2 do not show (which would easily be understood if they were of the size of those in Ascaphus, or not ossified), and the intermedium remains separate from the ulnare.
In Salamandra (Francis, 1934; Nauck, 1938) distal carpals 1 and 2 are fused in both larva and adult, and 3 and 4 are separate; the radiale, intermedium and ulnare are separate in the larva but the latter two fuse in the adult; centrale 1 (labelled prepollical cartilage by Francis) and centrale 2 are separate. Francis considers the digits (and metacarpals) to be 1-4. Apparently the arrangement here indicated for the larva is characteristic of other larval salamanders, except where further reduced, and reduction below the number given for the adult is common in other terrestrial forms. The radius and ulna are, of course, separate.