136 LOWER AMAZONS—PARA TO OBYDOS.  Cnuap. VI. 
occurring at the extreme points of it. If local conditions acting 
directly on individuals had originally produced this race or species they 
certainly would have caused much modification of it in different parts 
of this region; for the upper Amazons country differs greatly from the 
district near the Atlantic in climate, sequence of seasons, soil, forest 
clothing, periodical inundations, and so forth. These differences, more- 
over, graduate away, so that the species is subjected to a great diversity 
of physical conditions from locality to locality, and ought in consequence 
to present an endless series of local varieties, on the view mentioned, 
instead of one constant form throughout. Besides, how should we 
explain the fact of H. Thelxiope and H. Melpomene both existing 
under the same local conditions ; and how account for the diversified 
modifications presented in one and the same locality, as at Serpa and on 
the ‘Tapajos ? * 
There is evidently, therefore, some more subtle agency at work in the 
segregation of a race than the direct operation of external conditions. 
The principle of natural selection, as lately propounded by Darwin, 
seems to offer an intelligible explanation of the facts. According to 
this theory, the variable state of the species exhibited in the districts 
above mentioned would be owing to Heliconius Melpomene having 
been rendered vaguely instable by the indirect action of local conditions 
dissimilar to those where it exists under a constant normal form. In 
these districts selection has not operated, or it is suitable to the con- 
ditions of life there prevailing, that the species should exist under an 
instable form. But in the adjoining moister forests, as the result shows, 
the local conditions were originally more favourable to one of these 
varieties than to the others. ‘The selected one, therefore, increased 
more rapidly than its relatives; and the fact of the entire absence of 
these latter from an area whence they are now separated only by a 
few miles, points to the conclusion that they could not there maintain 
their ground. ‘Those individuals of successive broods which were still 
better suited to the new conditions would for the same reasons be 
preferred over their relatives ; and this process going forward for a few 
generations, the extreme form of H. Thelxiope would be reached. At 
this point the race became well adapted to the new area, which we may 
suppose to have been at that epoch in process of formation as the river 
plains became dry land, at the last geological changes in the level of the 
country. In the higher and drier areas of Guiana and the neighbouring 
countries, H. Melpomene has been the selected form ; in the lower and 
more humid regions of the Amazons, H. Thelxiope has been preferred. 
An existing proof of this perfect adaptation is shown by the swarming 
abundance of the species; the derivation of H. Thelxiope from H. 
Melpomene is made extremely probable by the existence of a complete 
series of connecting links ; and lastly, its permanent establishment is 
made evident by its refusal to intercross with its parent form, or revert 
* As the action of external influences would be on the early states of the insects and 
not on the adults, it is well to mention that the broods of the Heliconii appear to be 
social ; the larvee feeding together and undergoing their last transformation on the 
same tree. This I observed with regard to the H. Erato, a species closely allied to 
H. Thelxiope, 
