256 THE VARIATION OF ANIMALS IN NATURE 



Finn (1909) draw attention to the same phenomenon in birds. 

 Possibly the frequency of such convergence is much greater 

 than is usually supposed, since it is much less likely to be 

 noticed than when the resemblance occurs between inhabitants 

 of the same country. Since the action of selection is out of 

 the question in these cases, we must assume that the number 

 of possible colour-patterns for one group of animals is not 

 unlimited, and that occasionally parallel evolution will lead 

 to striking resemblances. 



On a priori grounds the chance of this is the greater the 

 more nearly allied are the animals, and, when members of the 

 same family or genus are under consideration, it is quite 

 possible that parallel evolution should be fairly common. 



Species of the same genus, often, however, belonging to 

 different subgeneric groups, not rarely show resemblances 

 which have been claimed to be the result of selection. We 

 mention species of the genus Charaxes (Swynnerton, 1926), 

 Heliconius (Eltringham, 191 6), or of certain Pierine genera, 

 Mylothris and Phrissura (Eltringham, 19 10, p. 83). There is 

 no reason why some of these resemblances should not be due 

 to parallel evolution, quite unaided by selection. The chief 

 difficulty for such a hypothesis arises when the mimetic forms 

 have identical geographical ranges. This difficulty is more 

 serious when both species are polymorphic and in different 

 parts of the range the colour-patterns still go together : in 

 fact, evidence of this sort is far the most cogent argument in 

 favour of the view that mimetic resemblance is due to selection. 

 This geographical coincidence, however, is by no means fully 

 established in a large number of cases. Thus Eltringham 

 (1916, p. 141) states : 'To understand more fully the relation- 

 ships of models to mimics in Heliconius we require much more 

 information concerning geographical distribution, and also 

 as to comparative rarity of forms and other bionomic factors. 

 S. America is a very large area, and the commonest type of 

 data on our labels is " Upper Amazon," " Columbia," " Peru," 

 and even sometimes " Brazil." We might as well be told that 

 a certain insect occurs in Europe.' It is probably true in the 

 greater number of cases of mimetic resemblance that, though 

 the convergent forms have been shown to occur together in 

 certain localities, we have no knowledge of the exact range 

 of any one form. 



