88 Clark, Pterylosis of Swifts and Hummingbirds. [.'an 



of the swifts, as a whole. A comparison of figures 4 and 5 of plate 

 II, with the wings of Cficetura and Trochilus, will emphasize how 

 really intermediate between the two swifts, in this point, the hum- 

 mingbird is. There is no essential difference in the humeral 

 tracts, though those of the swifts are more obviously connected 

 with the dorsal and femoral tracts, than are those of theTrochili. 

 Pterylosis of the Bod;/. — In the form of the dorsal tract swifts 

 and hummingbirds differ. In the latter it broadens on the mid- 

 dle of the back, extending downward to the sides, and then narrow- 

 ing again to a point at the oil gland; were it not for its obvious 

 connection with the femoral tracts on each side, it would thus be 

 an almost perfect diamond in shape. In the swifts, on the con- 

 trary, the tract does not extend downward on the sides nor docs 

 it directly connect with the femorals; its shape therefore is that 

 Of an elongated ellipse. In each case, however, there is the well- 

 marked spinal apterium, a very important point of resemblance. 

 In the swifts, the femoral tracts are well developed, extend for- 

 ward on the sides to the posterior part of the humerals, and back- 

 wards to connect with the upper tail-coverts. In the humming- 

 birds, however, the femorals are weak and only extend towards 

 the dorsal with which they connect. It is evident therefore, that 

 in the swifts the dorsal tract has become quite sharply defined, 

 and the femorals extended to form a part of the covering for the 

 sides, while in the hummingbirds the dorsal tract remains more 

 extensive, has not become separated from the femorals, and pro- 

 vides all the covering for the sides. Thus the dorsal pterylosis 

 of the swifts indicates a greater specialization than that of the 

 hummingbirds, possibly associated with their extraordinary power 

 of flight. The ventral pterylosis of the Cypseli is essentially 

 identical with that of the Trochili, for the apparent absence of 

 ventral tracts in the latter is a difference of degree, not of kind, 

 while the separation of a lower cervical apterium from the ventral, 

 in some swifts, is also an unimportant character, present in only 

 a few species. 



' Conclusions. 



In bringing to a close this account of the pterylography of swifts 

 and hummingbirds, it is worth while to consider two general ques- 



