OSTEOLOGY OF BIRDS 9 I 



toe are very remarkable from the great reduction in length of the 

 third phalanx; and finally, the basal joint of the second toe is but 

 slightly longer than it is in Circus. 1 ' 



We next pass to the consideration of the osteology of two other 

 genera of the Accipitres, the genera Accipiter and Astur, which con- 

 tain three species and a subspecies. I have abundance of material 

 from the Sharp-shinned and Cooper's hawks ( A c c i p t e r v e 1 o x 

 and cooperi ) , but only a sternum and shoulder girdle (Astur 

 atricapillus) to represent the two goshawks. 



In their skeletons Accipiter velox and A c c i p t e r 

 cooperi are very much alike, and a glance at either one of 

 them is sufficient to satisfy us that they have a number of characters 



1 Through the courtesy of the Department of Comparative Anatomy of the United 

 States National Museum T have, since the above was written, come into possession 

 of a complete skeleton of Elanus, and this footnote will supply still others of its 

 characters in addition to those presented above. This specimen (no. 85256) has the 

 nasal septum entire; the accessory pieces to the lacrymals are large; a small vacuity 

 occupies the center of the interorbital septum; the postfrontal processes remind us 

 somewhat of the owls; the squamosal processes are completely aborted; the man- 

 dibular facets of a quadrate are narrow from before, backward; wide transversely. 

 This skull is nondesmognathous. The pterygoids are widely separated at their pala- 

 tine heads when articulated in situ. The palatines remind me of the palatines of a 

 goatsucker (Chordeiles), being broad, flattened horizontally, rounded posteroexternal 

 angles, and with very much reduced laminae. Beyond their out turned pterygoidal heads 

 they are in contact all along under the rostrum, and their prepalatine portions are 

 slender. The vomer is much as it is in Circus. The basipterygoidal processes are very 

 well seen, but they do not reach the pterygoids. Basitemporal region is broad and 

 horizontal, and the Eustachian tubes open separately in front, and they are not 

 patulous along their continuities. Very little difference in size is seen among the 

 sclerotals of an eye, and the Jiyoid arches agree very well with what we find in 

 Circus. 



Nineteen freely movable vertebrae are found between its skull and pelvis, and there 

 seem to be 13 in the pelvic sacrum. Seven more are found in the skeleton of the 

 tail, to which we must add the rather small pygostyle. Two last cervicals bear 

 free ribs; the next pair in this specimen are peculiar in that they are connected 

 with the small leading pair of haemapophyses by ligament only. This gives six 

 pairs of costal ribs articulating with the sternum upon either side; then in addition to 

 these there is a floating pair that has no pelvic ribs coming down to it from the 

 pelvis. The epipleural appendages are very broad at their bases, and they fuse with 

 their ribs. 



In the pelvis we find very little flexion of the postacetabular part of the bone 

 downward and forward as it occurs in Buteo and others. There is a wide inter- 

 ruption of the postpubic style, and a double row of interdiapophysial foramina occur 

 down the sacrum. Anteriorly, the long, narrow preacetabular parts of the ilia are 

 separated from each other by the broad sacral crista, the whole being considerably fused 

 together in this place. Nothing especially remarkable characterizes the bones of the 

 pectoral limb, beyond the great length of the humerus, and two segments of the 

 antibrachium, the ulna and radius. When the pectoral limbs are articulated in the 

 skeleton in situ, the elbows are opposite the first caudal vertebrae, while the first 

 metacarpals are opposite the ninth cervical vertebra. As in most Accipitres there is 

 a minute claw upon pollux digit. In the pelvic limb the femur is long and pneumatic, 

 nearly straight and only of moderate caliber. There is a great oblong pa 

 Pandion, which is obliquely grooved by the ambiens. Ossifications take place in 

 the tibial cartilage, and the fibula 'lacks considerable of being complete. The cncmial 

 processes at the head of tibiotarsus are much reduced. 



