294 ON THE CLASSIFICATION OF BIRDS. 
the meadows long after the sun has set ; and Dr. 
Richardson mentions that, although it often sits for a 
long time on the boughs of a tree, it is frequently seen 
skimming over swampy pieces of ground, and hunting 
for its prey by the subdued daylight, which illumines 
even the midnight hours in the high parallels of lati. 
tude. These two forms we consider as typical of the 
buzzards ; and so perfectly does the first, which belongs 
to the subgenus Circus, represent the Fissirostres, that 
Dr. Richardson’s account of its manners is precisely 
applicable, as every one must perceive, to the common 
swallow. Fissirostral types either hunt for their prey 
upon the wing, like the swallow family, or watch for it 
from a fixed station, like the tyrant fly-catchers and 
the orilinary buzzards ; while, to show that these op- 
posite modes of feeding are quite consistent with the 
same primary type, we find them actually united in the 
Buteo lagopus, which thus intervenes between Circus 
and the common buzzard, and represents at once both 
the typical and the aberrant Fissirostres : for all these, 
in fact, capture or seize their prey upon the wing. 
(241.) The relations of the two next groups, the 
kites and the Tenuirostres, from our comparative igno- 
rance of the habits of both groups, cannot be traced in 
their economy ; but this need not be much regretted, 
since they possess, in common, certain peculiarities of 
structure which are not to be found in any other birds 
of their respective circles. The humming-birds have 
wings fully as long as the swallow.s, joined to a peculiar 
elongation of the upper mandible, which they, or any 
other perching family of slender-billed birds, do not 
possess. Their feet, moreover, are tlie shortest of all 
the Insessores ; so sliort, indeed, that the tarsus is fre- 
quently shorter than the hinder toe, while the middle 
toe scarcely exceeds the lateral ones. Now, upon look- 
ing to the true kites, comprehending the subgenera 
Cyniindis, Nauclerus, and probably Elanus*, we ob- 
* We have not an example of this genus before us, but the published 
figures give an idea that the tarsi arc remarkably short. 
