70 
Birds of Celebes: Falconidae. 
habits while the causes of the evolution of the coloration have undergone no change. 
One feels compelled to seek some natural selective agent to explain why the 
coloration did not develop in very different directions in the two birds, 
instead of in an identical direction. 
The second hypothesis — that the change of structure and the evolution 
of the coloration proceeded quietly together through perhaps a long period of 
time, — is more acceptable because it makes a less abrupt and violent appeal to 
the imagination, but there seems to be nothing further in its favour. 
The third hypothesis — that the present adult dress had been acquired 
before ever the structural ditferentiation into Pernis and Spizaetus (not to speak 
of other genera) had taken place — admits of discussion from several points 
of view. 
It is certain (A) that small structural changes may take place without caus- 
ing any marked changes of the adult coloration: witness many of the King- 
fishers, where the type of coloration of the four-toed Alcedo may be found in 
certain species of the three-toed Ceyx (the more recent form in respect of the 
foot) and others, or certain Flycatchers like Siphia, Muscicapula (pis. XIII — XIV) 
and others, with one or two types of coloration running through closely- allied 
genera, or the Flowerpeckers Acmonorhynchus and Pachyglossa, much alike in 
coloration, but the former (the more recent form in respect of the wing) has 
nine primaries, the latter the customary ten. 
(B) The hypothesis involves the assumption that the land of origin of the 
genera Pernis and Spizaetus was ancient Celebes. There is, we believe, a form 
of prejudice, against which we ourselves have to struggle, which pre-supposes that 
ancient Celebes (whatever sort of island, or islands, or adjunct of Asia it may 
have been) has simply been a receiver of birds from other parts, a land for 
colonisation, which has kept to itself the bird-forms it received. In the main 
this view is very likely right, but it is probable that in some cases Celebes may 
have sent out bird-colonists of its own. 
The genus Pernis is a link between Celebes and Asia, not being found in 
the Australian Region; only one aberrant Spizaetus^ S. gurneyi, is found east of 
Celebes, and the two genera may, of course, have become differentiated structur- 
ally from one another in Celebes as well as anywhere else. There is some- 
thing to be said for this view. In the Great Sunda Islands and the Philippines 
as far as India, both the Honey-buzzard {Pernis ptilonorhynchus) and the Hawk- 
eagle (Sp. limnaetus)^ instead of possessing — as seems to be the case in Ce- 
lebes — a fixed adult dress, agree in being wonderfully variable. Now, it appears 
that long ages in a given locality, with the same climate, food, dangers, predi- 
lections and regular routine of habits, and, perhaps, other more hidden influ- 
ences, lead to the assumption of a more or less fixed, stable coloration in birds, 
as is shown by the species of all tropical islands sufficiently remote from one 
