92 NEW YORK STATE MUSEUM 



quite like Circus, but still others that have an evident tendency 

 in another direction, and what this latter may be can only 

 be decided later on, when other genera are examined. As 

 to the sternum of the goshawk at my hand (Astur atricapil- 

 lus), and its shoulder girdle, I can dismiss it here by saying that 

 the latter agrees in its characters with the shoulder girdle of 

 Accipiter cooperi. This also applies to the sternum, with 

 the exception that in the goshawk it is 2-notched posteriorly, while 

 in Cooper's hawk it is 2- fenestrated. Accipiter cooperi has 

 the keel extending back further than it does in the goshawk. An- 

 teriorly the bones are quite similar, though the goshawk has but six 

 facets upon either costal border, and Accipiter cooperi has 

 often seven. The characters, however, which I have pointed out 

 as differences, we now know often vary among the Accipitres, 

 especially the ribs, and the notching and fenestration of the xipho- 

 sternum. Farther along we shall see that the anterior part of the 

 sternum of an Accipiter is quite unlike the corresponding part of 

 that bone in Circus. 



Upon comparing the skull of Accipiter cooperi with the 

 skull of Circus hudsonius, I find that in all the essential 

 details they agree exactly, Accipiter cooperi having the 

 basisphenoidal processes quite as well developed as they are in the 

 harrier. The only differences worthy of mention are the form of the 

 external narial apertures, these being in the Accipiter rather more 

 circular; the septum narium is somewhat more complete above; the 

 superohorizontal portions of the 1 aery mats with their " accessory 

 pieces " are longer and more prominent ; and finally, the temporal 

 fossa at the lateral aspect of the cranium is slightly more extensive, 

 though the superoposterior margin of the bony entrance to the ear 

 is not as conspicuously outstanding as it is in Circus [text fig. 19; 

 pi. 8, fig. 13, 14]. 



The mandibles also practically agree, as does also the hyoidean 

 apparatus, and the independent ossifications of the ear, eye, and 

 windpipe. Except in point of size, the skull of Accipiter 

 v e 1 o x agrees with the skull of its larger congener A c c i p t e r 

 cooperi. 



The remainder of the trunk skeletons in these two genera also 

 practically agree, with the exceptions that in Accipiter, we find that 

 the scapular process of a coracoid is not pierced by a foramen; the 

 scapulae are relatively longer, narrower, and more pointed pos- 

 teriorly; likewise the sternum is comparatively somewhat longer 

 and narrower, with its carinal angle more produced and directed 



