298 MR. W. p. PYCRAFT ON THE [Apr. 15, 



long spine, as in Falco. The sternvim varies mneh both in its 

 relative length and the foi'm of its posterior border. Thus, this 

 last may be either notched or fenestrated, or both, or it may be 

 entire. As these are points which may vary in members of the 

 same genus, they are of comparative little value save as specific 

 characters. 



The stei-num with the relatively largest caiina of this groxxp is 

 that of Pandion. The carina generally has its anterior border 

 sloped backwards, but in Accipiter this border turns forwards and 

 upwards, thus greatly increasing the length of the keel. The 

 carina varies much in its development, generally passing insensibly 

 into a sternal plate before leaching its hinder border, but in 

 Accipiter and Peroiis, for instance, the carina is cariied back 

 almost to the extreme free border of the sternal plate. 



The sternum of the Accipitres resembles that of the Striges. 

 In many cases this resem.blance is rather close. The following 

 chaiacters will be fovmd useful in determining between sterna 

 belonging to these two veiy different gi-oups. The posterior 

 border of the sternum in the Striges is never entire and never 

 fenestrated, but always notched. With the exception of the 

 steina belonging to the Strigidse, and the sternum of Htihua 

 orientalis of the Bubonidfe, there are two pairs of notches. The 

 single pair of notches of If. orientalis are of great size. They lie 

 on either side of the metasternum, and extending forwards to 

 beyond the level of the middle of the posterior lateial pi-ocess, 

 cause the sternum of this bird to closely resemble that of 

 Microhierax. The great difference in size, however, renders any 

 possibility of confusion on account of this resemblance impossible. 

 The single pair of notches of the Strigidse resemble those of some 

 Accipitres, e. g. Elanoides, in that they are very shallow, so that 

 the posterior lateial processes pass almost insensibly into the 

 metasternum, being divided therefi om only by a sinuous line. But 

 the processes are relatively much longer in the Stiigidse than in 

 the Accipitres, and the steinum is narrower. FTuthermore, the 

 sternum of the Owls of this group may be distinguished by the fact 

 that the spina externa is obsolete. The anterior lateral processes 

 in all the Owls are small. The ai'ticular surfaces for the sternal 

 ribs never extend beyond the middle of the sternal plate, rarely so 

 far as this. There is no spina interna. 



The coracoid presents two forms : — (1) That which obtains in 

 the Falconidee only, wherein there is a distinct hook-like pro- 

 coracoid process curving downwards till it touches the furcula ; 

 and (2) that in which this process is wanting. 



In the Falconidee we can distinguish the coracoid of the 

 Falcones from that of the Polybori ; inasmuch as in the former 

 there is no trace of the foramien supracoracoideum in the dried 

 skeleton, this being bounded inteinally by membrane only. In 

 the Polybori this foramen is present. .' 



Whether the foramen in the coracoid of the remaining Accipitres 

 is leally the foramen sujyracoracoideum, or a foramen pierced 



