98 THE GENESIS OF SPECIES. [Chap. 



of Borneo and Java on the one side, and of the Moluccas on 

 the other, are ahnost entirely al).sent from Celebes. Their 

 place seems to be supplied l»_v tlie cater])illar-catchers, of 

 which six or seven species are known I'lom Celebes, and 

 are very numerous in individuals. AVe have no positive 

 evidence that these birds pursue butterflies on the wing, 

 but it is highly prnl)able that they do so when other food 

 is scarce. Mr. IJates suu^ested to me that the larger dracjon- 

 flies prey upon butterflies, but I did not notice that they 

 were more abundi;nt in Celebes than elsewhere." ^ 



Xow, every opinion or conjecture of ^Ir. Wallace is 

 worthy of respectful and attentive consideration, but the 

 exjtlanation suggested and above referred to hardly seems 

 a satisfactory one. AMiat the past fauna of Cehibes may 

 have been is as yet conjectural. ]\[r. AVallace tells us that 

 now there is a remarkable scarcity of fly-catchers, and that 

 their place is supplied by birds of v»diicli it can only be 

 said that it is " highly probable " that they chase butter- 

 flies " when other food is scarce." The quick eye of 

 !Mr. Wallace failed to detect them in the act, as also to 

 note any unusual abundance of other insectivorous forms, 

 which therefore, considering ^Ir. AValhace's zeal and powers 

 of observation, we may conclude do not exist. ^lureover, 

 even if there ever has been an abundance of such, it is by 

 no means certain that they would have succeeded in pro- 

 ducing the conformation in question, for the efiect of this 

 peculiar curvature on flight is far from clear. AVe have 

 here, then, a structure liyi)otlietically explained by an 

 uncertain jtroperty induced by a cause the presence of 

 which is only conjectural. 



Surelv it is not unreasonable to class this instance with 

 1 " Natural Selection," p. 177. 



