THE CATEGORIES OF VARIANT INDIVIDUALS 69 



to whether intergrades occur between the groups. Sub- 

 species are groups between which intermediates occur only 

 rarely or not at all (see Dice, 1931 ; Merriam, 191 9, for 

 conflicting views on this subject). 



We have thus quite clearly established the recognition of 

 more or less distinct geographical groups on the one hand and 

 groups or types not spatially segregated, but appearing either 

 as individual variants sporadically throughout a population 

 or as larger local elements not segregated into geographical 

 units. We have now to inquire concerning other subdivisions 

 of this kind. 



Races. — The term geographical race is used as a complete 

 synonym for subspecies by several authors (cf. Alpatov, 1929). 

 But it is also used for a smaller unit not of the same dimensions 

 as the subspecies. Local race and local forms {cf. Duncker, 1896) 

 are used in the same loose way. In fact it will be readily 

 recognised that such a hierarchy might exist within the species, 

 that the boundaries of the various groups would be difficult 

 to draw and there would be some confusion of terminology. 



That such a hierarchy of local or geographical groups does 

 exist is, we think, quite clear. This is perhaps best seen in the 

 work of Schmidt (1920), who finds that the ^oarces population 

 is divided into numerous ' races ' and each of these can be 

 again split into still smaller elements. In this case (p. 114) 

 the averages of the smaller groups combined give the average 

 of the race. A similar example is seen in Duncker's studies 

 of the Flounder and Plaice (1896). 



In Sumner's investigation of the local variation oiPeromyscus 

 maniculatus it is quite clear that the local populations within 

 the three chief subspecies are not identical (1920, p. 388, fig. 2), 

 but exhibit significant statistical differences. He says (191 7, 

 p. 173), ' subspecies themselves are far from being elementary.' 

 They are composite groups comprising in numerous cases a 

 number — perhaps a great number — of distinguishable local 

 types. Similar groups which are the result of intense localisa- 

 tion in segregated populations are recorded by Gulick (I.e.), 

 Crampton (I.e.), Mayer (1902), Boycott (1919), Aubertin, 

 Ellis and Robson (1931) for ' colonies ' of land snails (general 

 discussion of the problem in the last-named paper). Many of 

 these colonies are found in valleys or on ridges. A still more 

 acute form of local differentiation is seen in the ' forms ' of 



