THE ORIGIN OF SPECIES 



gone a similar adaptive radiation within a geologically brief time. The 

 Hawaiian Islands appear to have originated no earlier than the PHocene. 

 As the honey-creepers are forest-dwellers, their original ancestor could not 

 have reached Hawaii (by migration from Central America) before the 

 forests themselves were established. This might have occurred by mid- 

 Pliocene or later, thus giving a maximum of five million years for the dif- 

 ferentiation of the family in Hawaii. Lest one be tempted to think of this 

 as a long time, let us recall that many instances are known in which only 

 subspecific dififerentiation has been achieved after periods as long as 

 50,000,000 years. 



When the remote ancestor of the Drepaniidae first migrated from the 

 mainland to the Hawaiian forests, it found a rich field for which it had no 

 competitors. The result was a rapidly expanding population which soon 

 provided its own competition. That is, the population began to outstrip 

 the food supply available by the original method of feeding. This drepa- 

 niid progenitor may well have been similar to the living Loxops virens 

 chloris, which has a moderate-sized, slightly curved bill, adapted to feed- 

 ing upon insects in foliage. But the birds occasionally dig for insects in 

 loose bark, or probe flowers for nectar and insects. Now the development 

 of races or species with different feeding habits would permit the survival 

 of a much larger total drepaniid population. How this might have oc- 

 curred is perhaps indicated by living members of the genus Loxops. L. v. 

 chloris is widely distributed through the islands, but on Kauai, one of the 

 most isolated islands, this species is represented by another subspecies, 

 L. V. stejnegeri. Another species, L. porva, is found only on Kauai, and the 

 characteristics of the bills of these two Loxops representatives are most 

 suggestive. The bill of L. v. stejnegeri is somewhat larger, heavier, and 

 more strongly recurved than that of L. v. chloris. While it still has feeding 

 habits similar to those of L. v. chloris, it depends more upon digging for 

 insects in loose bark. It also visits flowers for nectar and small insects. But 

 the bill of L. porvo has deviated from that of L. v. chloris in just the oppo- 

 site way. It has become shorter and straighter. L. porva depends primarily 

 on insects on the surfaces of branches and leaves. Its bill is not well 

 adapted to digging, and it rarely attempts this. While it is not well adapted 

 to visiting flowers, it does frequently visit Acacia flowers. It seems prob- 

 able that, when these two species of Loxops were brought into competi- 

 tion, any variations which tended to adapt them to different sources of 

 food were strongly favored by natural selection. The result is that these 

 two species, so strongly divergent in bill structure, differ very little in 

 other respects. 



But the genus Loxops only begins to indicate the range of variations of 

 bills in the Drepaniidae (Figure 109). The genus Hemignathus, most 

 members of which have become extinct in recent times, showed a much 

 greater range of variation than usually characterizes whole families. 

 Hemignathus ohscurus, which was probably closely related to Loxops 

 virens, had a very long, slender, and strongly curved bill. It was adapted 

 for probing fine crevices in the bark of trees, and for visiting flowers. In 

 H. lucidus, the upper mandible is much like that of ohscurus, and it was 



310 



