PROCEEDINGS OF THE NATIONAL MUSEUM. 571 



in the owls. This difference in number might have been something- of 

 a dittienlty in sliowing any connection between the two gr<)ni)s were It 

 not for the interesting discovery tliat MicropaUas possesses only ten, 

 a fact which seems to have previously escaped notice. Another thing 

 of interest in regard to the tail is that while in both Striges and Capri- 

 mulgi it is the rule for the middle pair of rectrices to be longest, there 

 is an exception in each group; amofig the former Sfrix, and in the 

 latter, Chordeiles, has the outer pair longest. 



CONCLUSIONS. 



Having thus compared in detail the i)terylographical characters of 

 both Caprimulgi and Striges, as far as the material at hand would 

 allow, I may justly be i^ermitted to draw a few inferences from the 

 facts before nie. It can hardly be denied that these facts indicate a 

 certain degree of affinity, and although to me this relationship seems 

 quite close, of course it is well understood that conclusions based on 

 one set of facts will often be overturned by another set and are, there- 

 fore, unreliable. .Judging from the wings and tail, the united sternal 

 and ventral tracts, and the striking longitudinal arrangement of the 

 feathers on the crown, there can be little doubt that the Caprimulgi 

 are the more decidedly modified of the two groups. In each one of 

 these particulars, moreover, there are owls almost as fully modified, yet 

 they do not combine them as do all of the Caprimulgi. Thus, MicropaUas 

 has only ten rectrices, but the head is uniformly feathered and the ven- 

 tral tract is distinct, while Strix shows a partial union of the sternal and 

 ventral tracts, but has twelve rectrices and a closely feathered crown. 

 We may thus reasonably conclude that the common ancestors of the 

 two groups were rapacious birds much more like owls than goat- 

 suckers and probably more or less nocturnal in their habits. They 

 had a well-feathered head, a widely forked upper cervical tract, separ 

 rate sternal and ventral tracts, a peculiarly situated femoral tract, 

 twelve rectrices, of which the middle i^air were longest, and twenty-six 

 or more remiges, of which eleven were on the hand. The plumage prob- 

 ably possessed an aftershaft and the oil gland was bare, while the tarsus 

 may have been feathered. From this extremely hypothetical stem, there 

 soon arose birds more crepuscular than nocturnal and insectivorous 

 rather than carnivorous. These were the immediate ancestors, of the 

 Caprimulgi and soon lost the first primary, one pair of rectrices, and 

 several secondaries. In them also the ventral and sternal tracts fused 

 and the feathers of the head condensed into longitudinal rows, while 

 the feathers of the tarsus began to disappear. Thus the general 

 Caprimulgine pterylosis arose by what will at once be seen as a process 

 of condensation, possibly due to the need of greater lightness and 

 speed for the capture of their insect prey. Phahvnoptilus shows the 

 most perfect development of this Caprimulgine form and so is the 

 most modern descendent of these hypothetical ancestors. Chordeiles 



