BIRDS OF PREY. HARPAGUS. 
299 
nunculus of Dr. Horsfield. For the present, however, 
we shall consider all these variations from the typical 
structure as sectional groups ; the more so, because 
deviations, equally marked, exist among the Thanino- 
phelince, or bush shrikes, another group of predatory 
birds, although belonging to the DentiTostTes. Now, as 
these modifications occur in closely connected species, 
and of the same countries, it is quite clear that they do 
not characterise either natural or geographic groups; in 
other words, that they are aberrant species, and not sub- 
(245.) The next principal type of Falco appears 
to be that of Bidens of Spix (which we discovered in 
Brazil previous to that traveller), but which nanae has 
been properly changed for Harpagus. In this the 
wings are less pointed than in F<ilcOj but not so rounded 
as in the hawks ; the bill has a peculiar thickness, and 
instead of one very strong tooth in the upper mandible, 
there are two smaller ones ; the structure of the feet 
accords with that of the generahty of true falcons, but 
the scutellation of the tarsi are very different ; instead 
of numerous small scales, disposed in a reticulated or 
net-like manner, those in front of the tarsus are large, 
broad, and transverse. The passage between Harpagus 
and Falco is marked by a well known and beautiful 
little bird, the Falco ceerulesccns of Linnsus, which we 
class as an aberrant HarpaguSj with which it agrees in 
having a doubly toothed bill, united with the more 
pointed wings of the subgenus Falco* This species, not 
now before us, does not appear to be otherwise different 
from Harpagus; for, by the description we quote from *, 
the tarsi appear to be similar to those of the typical 
species, all of which inhabit Brazil. The third sub- 
genus is represented by a most beautiful crested falcon 
from India, the Falco lophotes'^, of which only one 
specimen, now in the Paris Museum, is known. Like 
the latter, the bill of this also is bidentate; but the feet 
• Zoological Journal, vol. i. p. 328. 
f Teinminck, Planches Color^es, pL 10. 
