INTRODUCTION 13 



systematic data is the continued break-up of populations, 

 the divergence of the groups thus formed along different paths, 

 and the replacement of groups having one kind of constitution 

 by other groups having a different constitution. What we 

 have to account for is not only the changes in single characters 

 or groups of characters in single individuals, but also the means 

 by which they become characteristic of populations. We 

 stress this obvious and generally accepted truth, because in the 

 generalisations based on experiments and observations in the 

 laboratory, or in the genetical and mathematical treatment of 

 the subject, emphasis is usually laid on the origin of new 

 characters and their chances of survival and the fact of group 

 formation are neglected. Moreover, various authors (e.g. 

 Kinsey, 1930, pp. 34-35 ; Hogben, 1931 ; Guyenot, 1930, 

 p. 211 et seq.) have suggested that any mutant might spread, 

 if it was not actually harmful to its bearer. Darwin also was 

 evidently of the same opinion and seemed to think that ' neu- 

 tral ' characters might survive. Haldane and Fisher, however, 

 have clearly shown that the mere fact of re-emergence from 

 a cross does not confer on mutations the power to spread 

 through a population. The spread of variants is, indeed, one 

 of the most crucial problems in the study of evolution. 



We will now proceed to formulate what we believe to 

 be the chief problems which a study of natural variation 

 raises. 



(i) A population inhabiting a definite area may gradually 

 change in the course of time, or two populations, originally 

 similar and practically homogeneous, but inhabiting different 

 areas, may diverge so as to become two distinct groups. The 

 two processes are probably much the same, though in the 

 latter case it may be possible to point out definite differences 

 in the environment of the two areas to which the divergence 

 might be due. In either case we have to explain the origin 

 of the new characters by which the diverging groups differ 

 from those they used to resemble, i.e. we have to consider the 

 causes of variation. 



(ii) As indicated on pp. 8-1 1, variants are not found dis- 

 tributed chaotically but in groups of various kinds. It is 

 necessary to define what these groups are and how they occur in 

 nature. 



(iii) It is evident that our definition of the term ' population ' 



