4l6 SNODGRASS 



On the whole, however, the evidence is against the supposition that 

 the ancestral race of Schistocerca lived on Duncan Island. 



Under (2) we may assume that the Duncan race is simply a 

 variety of either S. J7ielanoce7-a or S. literosa that has varied in the 

 direction of the other species. The great improbability of such an oc- 

 currence is sufficient to exclude this hypothesis. We may also assume 

 that the individuals on all the islands once varied as do now those on 

 Duncan, and that on Chatham, Hood and Tower the melanocera char- 

 acters were suppressed while on Albemarle, Indefatigable, James, 

 Charles and Barrington the literosa characters were suppressed. How- 

 ever, it appears highly improbable that islands so dissimilar as Tower 

 and Chatham should have produced the same form, and that islands 

 so much more nearly alike as Charles and Chatham should have pro- 

 duced forms so different. 



Case (3), assuming the Duncan form to be a hybrid race, has no 

 direct evidence in its favor. There is, however, nothing that can be 

 urged against it. The central position of the island would permit 

 stray individuals to be carried there by both wind and currents from 

 Charles, Hood, Chatham, Barrington and Indefatigable. If the race 

 is a hybrid then the incongruity of such a variable form inhabiting an 

 island of so limited extent and of such lack in diversity of conditions 

 is avoided. The assumption produces no discord with prevailing 

 ideas concerning the relationship of environment and variation. 

 However, as already stated, no direct evidence can be adduced in favor 

 of the hypothesis. It is simply the only one that explains the facts 

 and that cannot be disproved. 



If, then, we regard the Duncan race as a hybrid, the relationships of 

 the other races offers less difficulty. From a geological point of view 

 Chatham and Charles are certainly the oldest islands of the archipelago. 

 This is evident from (i) the disintegration of the surface lava, and (2) 

 the destruction, by erosion, of the central craters. In the upper parts 

 of the southwestern half of Chatham are large open fields of rich soil, 

 strikingly contrasting with the condition of all the other islands except 

 Charles and probably the upper part of Indefatigable. On Chatham 

 are numerous rounded hills, giving evidence of long continu'^d erosion, 

 and there is no recognizable remnant of a central crater. 



Charles Island appears newer than Chatham. The soil is thinner 

 and less fertile. The surface rises gradually toward the center of the 

 island where a number of high steep cone-like hills surround a central 

 valley. These hills stronglv suggest that they once formed the wall of 

 a crater. Nearly all the other islands have well preserved craters, and 



