THE DISTRIBUTION OF VARIANTS IN NATURE 87 



largely an arbitrary one and in practice merely implies that 

 the differences between individuals are sometimes so slight 

 that they can be arranged (in a graph or diagram) in a more 

 or less imperceptibly graded series. Similarly ' discontinuity ' 

 merely implies that there is a more or less perceptible break in 

 such a series of variates. The sizes of the steps in a continuous 

 series and of the breaks in discontinuous series are of course 

 incapable of standardisation. It is largely held that differences 

 of environment (e.g. the amount of nutrition received by 

 individuals) contribute very largely to ' continuous ' varia- 

 bility, though it is now known (e.g. from the work on Drosophila 

 or that on Ephestia kiihniella (Kiihn and Henke) ) that the 

 smallest and least sharply distinguished variants may have a 

 discontinuous hereditary basis. 



One of the most important applications of elementary 

 genetics to the field of taxonomy is to break down the distinction 

 between ' continuous ' and ' discontinuous ' variation. This 

 is still insufficiently realised by taxonomists. When we are 

 dealing with a single character, the occurrence of continuity 

 or discontinuity is determined by how two contrasting charac- 

 ters happen to interact in a particular species. It is well 

 known that the expression (as opposed to the inheritance) of 

 hereditary characteristics may depend on the environment 

 to which the individual is exposed. Thus with a given heri- 

 table basis deciding the main lines of, e.g., colour-pattern, its 

 actual degree of development may depend on the environment, 

 heredity determining only the mean. This principle is doubt- 

 less very important in considering the numerous examples 

 of pairs of species having a different mean but overlapping 

 range in some character. Where a complete gradation can 

 be found over a certain range of variation, it is not sufficiently 

 realised by taxonomists that very simple statistical treatment 

 will often demonstrate that the continuous range of variation 

 really masks a fundamental discontinuity. Taxonomists 

 usually content themselves with saying either that ' inter- 

 mediates are rare ' or that ' the forms are connected by all 

 intergradations,' in each case deciding summarily to separate 

 or ' lump ' together the two forms. If one makes a table 

 showing the frequency with which the character appears in 

 different degrees of development (e.g. as prepared by Sumner, 

 1923), the true nature of the variation-range may become 



