io THE VARIATION OF ANIMALS IN NATURE 



one we are limited in two ways. First, we are bound to give 

 some weight to statements not verified by these methods, when 

 the author alone has had, and perhaps can have, access to 

 the material. Secondly, many problems in the study of 

 variation appear at present to be outside the field of statistics, 

 because it is not yet possible to obtain sufficiently accurate 

 measurements for statistical tests to be applied, e.g. to differ- 

 ences in habit or to some of the finer structures. Often those 

 characters which are most easy to measure have no biological 

 significance, while those for which measurement is most needed 

 are least susceptible to it. Finally, all taxonomists are familiar 

 with differences between races and species which depend on 

 a general ' facies ' ; the individual characters which go to 

 make up this facies can be measured singly and the correlation 

 between any pair of them determined, but no single formula 

 can express the whole. 



We have laboured this point in order that at the offset it 

 may be amply clear that the study of variation within groups 

 is bound up with systematic procedure and is liable to errors 

 arising out of the inevitable defects of the latter. We do not 

 wish to minimise the risks to which theories of evolution are 

 liable through defective systematics. But although species and 

 other systematic categories are important reference points and 

 significant episodes in the course of evolution, with modern 

 intensive collecting-methods and the intensive study of large 

 numbers of individuals, the centre of interest is passing from 

 the systematist's species to the ' natural population ' from 

 which the species is abstracted. 



The term natural population (cf. Chapter III) is given to 

 any assemblage of individuals of a species living in nature irre- 

 spective of its systematic relationships, i.e. whether it is homo- 

 geneous or whether it contains diverse genotypic elements. 

 A ' population ' consists of a number of more or less geno- 

 typically similar individuals which are better able and have 

 more opportunity to interbreed with one another than with 

 the individuals of other populations. Such populations 

 considered taxonomically may be only a group of individuals 

 isolated topographically (e.g. on an island) from other struc- 

 turally identical individuals, or they may form a definite 

 variety, geographical race or species. The taxonomic name 

 given to the population depends on a variety of circumstances, 



