370 MR. W. p. PYCRAFT — CONTRIBUTIONS [Api'. 9^ 



gained by the sharp downward turning of the spine-like postero- 

 inferior angle of the ischium which transforms the closed ischio- 

 pubic fissure into a foramen of considerable width. 



vii. The Pectoral Limb. 



Were it possible to examine the wing-skeletons of a large 

 number of genera of any of the Families herein described, data 

 might be found for use in systematic work. So fai-, however, 

 such a collection is not to be met with in any museum. From 

 the material at my jjresent disposal it is possible to deal only with 

 the main features of the wing, and these do not display any very 

 remarkable characters. 



The humerus only is pneumatic, except in the Hirundinidee, 

 wherein all the wing-bones are non-pneumatic. But this 

 segment of the limb differs, in the several genera, chiefly in 

 the relative development of the deltoid crest, and of the ect- 

 epicondylar process : a relation, in so far as the deltoid crest is 

 concerned, which varies as the power of flight. Where this is 

 strong, as in the Hirundinidje and Gymnorhinfe for example, 

 this ci-est is of considerable size, having an oval contour, pro- 

 jecting well beyond, and extending down the shaft, for about 

 one- third of its length. Where the flight is weak, as in the 

 Tyrannidae, this crest is barely developed. Curiously, however, 

 the size of the ectepicondylar process does not appear to depend 

 on the flight to the same extent, since in Tyrannus, for instance, 

 it is relatively larger than in Gymnorhina. But this appearance 

 is deceptive, and is due to the relatively more slender shaft of 

 Tyrannus. Nevertheless the ectepicondylar process does not 

 appear to be so directly under the influence of flight as the deltoid 

 crest. In Tylas, for instance (in the ' Hand-list of Birds ' placed 

 with the Pycnonotida^, but probably one of the Prionopida?, or 

 Dicruridae), the power of flight, judging from the size of the 

 carina sterni, does not seem to have undergone any serious re-' 

 duction ; this process is small, and surmounted by a small, but 

 sharp spike. Similarly, in the Campophagidse, this process is 

 gi'eatly reduced, but a small spike remaining ; and here again 

 full powers of flight are certainly retained. In this reduced size 

 of the ectepicondylar process, these birds resemble the Pittidse 

 and Eurylfemidfe, so that yet another alternative may be pre- 

 sented — the character may be a primitive one. In the Swallows 

 there is a large entepicondylar process lying behind the ulnar 

 condyle, and projecting directly backwards. 



The forearm is relatively long and straight in all, though the 

 radius is slightly broad save for its distal one-third which is 

 straight. The ulna has prominent tubei-cles for the attachment 

 of the secondary remiges, and a pointed olecranon process, which 

 interlocks with a corresponding entepicondylar spur on the 

 humerus. 



The manus is the most variable segment; of the limb, inasmuch 



