i 5 4 THE VARIATION OF ANIMALS IN NATURE 



measure of geographical isolation is necessary for divergence 

 to begin. This view has been strongly maintained by Jordan 

 (1896, 1905), and is implicit in the ' Formenkreis ' theory of 

 Kleinschmidt and Rensch, according to whom geographical 

 races alone are the starting-point for new species. 



In the case of birds and mammals there would appear to 

 be good evidence for this idea. The lowest systematic cate- 

 gories (geographical races) never occur together except in a 

 minimal part of their range (cf. von Schweppenburg, 1924, 

 p. 143) and, generally speaking, only rather widely divergent 

 forms live together in the same habitat. It is true that the 

 geographical barriers between the races are not always abso- 

 lute, but imperfect barriers combined with the usually dis- 

 continuous occurrence of suitable habitats may be sufficient 

 to allow divergence. The chief lack at the moment is the 

 accurate study of the distribution and nature of the forms 

 occurring where two races meet. 



With insects the necessity of geographical isolation is 

 much more difficult to maintain, as might be expected from 

 the relative complexity of the way in which the sexes are 

 normally brought together. If selected cases are examined 

 (cf Jordan, 1896), it is easy to show the importance of geo- 

 graphical isolation, which in any case must always be operative, 

 even if it is not the only agency responsible for divergence. 

 Thus Jordan found in certain Oriental swallow-tail butterflies 

 that forms differing in colour, shape of wings or seasonal 

 occurrence never differ in genitalia unless they are restricted 

 to geographically separated areas. Since Jordan maintains 

 that mechanical isolation as a result of differences in the 

 genitalia is the chief means of making divergence permanent, 

 he argues that in these swallow-tails it is only the geographical 

 races and not variants which occur together in one locality 

 which will (or may) give rise to new species. It is possible, 

 however, in other groups to find examples which suggest the 

 opposite point of view. Thus species or races with genitalia 

 so similar as to differ from one another no more than do the 

 geographical races of swallow-tails, may occur together over 

 wide areas, as in the butterfly Satyrus huebneri (Avinoff, 1929), 

 in many Tortricids (compare male genitalia of species of the 

 genera Cnephasia or Epiblema, Plates v and xxiii (and p. 68) 

 in Pierce and Metcalfe, 1922) or in some Hesperiids (Warren, 



