1891.] PLEISTOCENE BIRD-REMAINS. 471 



C. gamila. Fossil. 



Length of tibio-tarsus 0,044 0,043 



„ of tarso-metatarsus . . 0,024 0,0225 



These show that not only are the fossil bones rather smaller than 

 those of C. garritla, but that in the former the tarso-metatarsus is 

 shorter in proportion to the tibio-tarsus than in the latter. It appears 

 that the two sexes of the Blue Roller do not differ in size, and it is 

 therefore pretty evident that the fossil bones do not belong to that 

 species. They may, however, be referable to the Abyssinian Roller 

 (C. ahyssinica), of which I have been unable to see a skeleton. That 

 bird, of which only one specimen appears to have been recorded in 

 Europe, is common in parts of Abyssinia and likewise occurs in 

 Arabia. The fossils may, however, equally well have belonged to 

 the N. African G. pilosa. The tvoodcut represents the coracoid of 

 the Sardinian Eoller. 



Fig. 1. 



Ventral aspect of right coracoid of Coracias, cj. abyssiniea ; from Tavolara. \. 



IV. Passeres. 



Corvus corone, Linn. — Evidence of the presence of the Crow in 

 the Sardinian Pleistocene fauna is afforded by the coracoid from 

 Tavolara represented in fig. 5 of the Plate. This specimen, which 

 belongs to the right side, accords in every respect with the corre- 

 sponding bone of recent skeletons. 



Fringillidce. — Several species of F7nngUlidcB are indicated by 

 humeri and other bones, but from their small size, and the large 

 number of the existing species of these birds, the specific, and in 

 some cases even the generic, determination is difficult. It may be 

 observed that the humerus of this family is easily recognized by the 

 great development of its tricipital fossa, which is separated from the 

 subtrochanteric fossa merely by a thin septum of bone in its upper 

 half, the lower portion of the two fossse being confluent. In this 

 respect the FrinyilUd(B differ widely from the GurvidcB (near to which 

 they are usually placed), in which the tricipital fossa of the humerus 

 is scarcely developed at all. 



Of the specimens which can be more or less definitely recognized, 



