294 MR. W. p. PYCRAFT ON THE [-^W- ^^} 



either side of, rather than between, the capitular and tubercular 

 glenoid surfaces. 



In the Falconidfe, excepting only Herpetotheres, and in Poly- 

 boi'idfe, as ah^eady remarked, the thoracic vertebrte 1-4 are anky- 

 losed into a single mass. But this mass also includes the last 

 cervico-thoi^acic vertebra, so that the whole is made up of five 

 vertebrsp. The penultimate thoracic vertebra is free, the ultimate 

 is fused with the synsacrum ; so that this free vei-tebra serves as a 

 hinge in the middle of the back. In some examples of Polyhoms 

 the antepenultimate vertebra may show traces of its former 

 existence. Such specimens are probably not quite adult. 



It is probable that the last tvi^o cervico- thoracic vertebrse have 

 not long been transferred to the cei'vical series, for the dorsal 

 segment of the ribs belonging thereto has as yet undergone no 

 shortening. Those of that vertebra which has ankylosed with 

 the thoracic still retain their vmcinates. 



In the Falconidfe and Polyborid?e only one, and in other 

 Accipitres two thoracic vertebrae appear to enter into the syn- 

 sacrum. 



The synsacrimn includes from 12 to 14 vertebrae. In Serpen- 

 tarius it is composed as follows : — 2 thoracic, 4 lumbar or cruro- 

 sacral, 2 lumbo- sacral or ischiadic, 2 sacral, and 4 caudal vertebrae. 



The anterior renal fossa (fossa ischiadica) in Serpentarms is 

 very lofty and narrow ; when seen from the ventral aspect, recalling 

 that of Gariama. The posterior renal fossa is long and narrow. 

 This last fossa is bounded by a very broad planum anale, and this 

 appears to be peculiar to Serpentarms, amongst the Falconif ormes. 



In the Accipitres, the foim and size of these fossfe vary much, 

 as is indicated by the following illustrations. It is impossible, 

 however, to do more than indicate the general nature of the 

 form of these fossae, for they vary in details in almost every 

 species. The anterior renal fossa forms a narrow and lofty 

 chamber, longer than the posterior fossa, in BusarelhijS, Circaetus, 

 HaliaMus, Haliastur, Lophaehos, Gi/pohierax, and Aoitenor, 

 It is narrow, lofty, and shorter than the posterior fossa in 

 Parabuteo, Aquila, and Spizaetiis. It is wide, lofty, and longer than 

 the posterior fossa in Falco, Chimachima ; wider, lofty, and shorter 

 in Elanoides and Polyhorus. In Pandion both these fossse are 

 exceptionally large, and of the type seen in Elanoides. In 

 Spizaetus the 2^osterior renal fossa has reached the minimum limit 

 of reduction, including but a single caiidal vertebra. 



It -v^ill thus be seen that the form of these fossae is of little or 

 no value for taxohomic purposes other than the identification of 

 species. The lofty and nari'ow type appears to belong to the 

 more specialized, the wide and shallow to the more primitive 

 members of the respective groups to which they belong. 



In Pandion, there are three and sometimes four lumbar or 

 cruro-sacral vertebrae, and these are more sharply defined than in 

 the other Accipitres. The lumbar possess well- developed dorso- 

 lateral processes. The sacrals are, in the skeleton, indistinguish- 



