378 MY LIFE [Chap. 



have here, as I 4 am sure they would furnish some most interest- 

 ing results. Now, it seems to me that a person having no 

 special knowledge of the district would have no idea from 

 your paper that the species did not in almost every instance 

 occur on both banks of the river. In only one case do you 

 specially mention a species being found only on the north 

 bank. In other cases, except where the insect is local and 

 confined to one small district, no one can tell whether they 

 occur on one or both banks. Obydos you only mention once, 

 Barra and the Tunantins not at all. I think a list of the 

 species or varieties occurring on the south bank or north 

 bank only should have been given, and would be of much 

 interest as establishing the fact that large rivers do act 

 as limits in determining the range of species. From the 

 localities you give, it appears that of the sixteen species of 

 papilio peculiar to the Amazon, fourteen occur only on the 

 south bank ; also, that the Guayana species all pass to the 

 south bank. These facts I have picked out. They are not 

 stated by you. It would seem, therefore, that Guayana 

 forms, having once crossed the river, have a great tendency 

 to become modified, and then never recross. Why the 

 Brazilian species should not first have taken possession of 

 their own side of the river is a mystery. I should be 

 inclined to think that the present river bed is comparatively 

 new, and that the southern lowlands were once continuous 

 with Guayana ; in fact, that Guayana is older than north 

 Brazil, and that after it had pushed out its alluvial plains 

 into what is now north Brazil, an elevation on the Brazilian 

 side made the river cut a new channel to the northward, 

 leaving the Guayana species isolated, exposed to competition 

 with a new set of species from further south, and so becom- 

 ing modified, as we now find them. . . . The whole district 

 is, I fear, too little known geologically to test this supposition. 

 The mountains of north Brazil are, however, said to be of the 

 cretaceous period, and if so their elevation must have occurred 

 in tertiary times, and may have continued to a comparatively 

 recent period. Now if there are no proofs of such recent 

 upheaval in the southern mountains of Guayana, the theory 



