THE DISTRIBUTION OF VARIANTS IN NATURE 1 1 7 



islands than on the larger ones and on the mainland, suggesting 

 that isolation has been the most important factor. It is curious 

 that some of the races occur on more than one small island. 

 Admittedly these islands are usually close to one another, 

 but not always closer than other islands which bear distinct 

 races. Further, the most similar races do not usually inhabit 

 the closest islands. Taking the islands as a whole we see 

 a progressive change in colour from the mainland form, but, 

 as the various changes are scattered at random amongst the 

 islands, it is unlikely that the series represents the actual line 

 of evolution, which was probably polyphyletic. 



In considering geographical races it is a matter of some 

 importance to examine the normal size of the racial population. 

 Many races of course exist over enormous areas and include 

 millions of individuals, but in the case of smaller units taxonomic 

 practice becomes somewhat arbitrary. It is evidently con- 

 venient to have a name for any race which covers a large area, 

 even if structurally it is little differentiated from its closest 

 allies. But in more localised races a higher degree of divergence 

 tends to be demanded. Thus a statistical examination of the 

 populations of a species inhabiting a number of small islands 

 might show that each had a different mean character, but it 

 might be taxonomically very inconvenient to give a name to 

 each. On the other hand, unnamed variations tend to be 

 ignored, and in making any such survey as the present only 

 the most general information about such forms can be obtained. 



We may give examples. Perhaps a record for smallness 

 of racial area is held by Lacerta simonii (Cott, 1932), which 

 inhabits a small rock with a surface of perhaps 1,000 square 

 yards in the Canaries. Cott estimates the total population 

 at not more than a few scores of individuals. The Skomer 

 Vole is confined to an island only a few square miles in extent, 

 and the same is true of many other island races. Isolated 

 colonies of the Rabbit (Oryctolagus cuniculus) are known which 

 are quite distinct in colour, e.g. a mouse-coloured race on 

 Sunk Island in the Humber (Barrett-Hamilton and Hinton, 

 I.e. pp. 196-9). The Skomer Vole is given a name because 

 it is a relict form whose nearest allies live in the Hebrides, 

 while the Rabbit is unusually variable and there are too many 

 trifling local variants for a name to be given to any one. In 

 the moths of the genus J^ygaena, particular colonies have often 



1 2 



