( 429 ) 



form a discontinuous seriet-. We can exju'ess this phenomenon in other words as the 

 law of independent variation of organs as op])osed to the correlative variation 

 of organs. 



There are many species in which not only the individuals form discontinuous 

 series, but which exhibit also one or more discoutinuously variable characters ; and it 

 is especially tlie appearance of such discontinuity which Bateson so amply illustrates 

 in Materials for the Study of VariaHon (London, 1894). In this case the specimens 

 belonging to a species can be arranged into groujjs according to a discontinuous 

 character, each group containing individuals in which that character is continuous or 

 nearly so, while from one group to the other that character exhibits a more or less 

 wide gap. In diagnostic zoology the groups of individuals of polymorphic animals 

 have often been mistaken for distinct species. The study of independent discontinuous 

 variation of two or more organs can, however, often help us to distinguish poly- 

 morphism from specific distinctness. In the following diagram the specimens are 

 first an-anged into grou[is according to character «, and then into groups according 

 to character b : — 



n. C D E 



J' A= h^ 



A H I 



In the first arrangement (T.) A B C form one, D E F another group; in the 

 second arrangement CDE belong to one, and A H / to another. In each case the 

 individuals A B C D E F II I form two groups not connected by intergradations, each 

 group, therefore, being conformable to what we species-makers call " species " as a 

 rule ; but in the first arrangement the specimen C belongs together with A and iJ to a 

 " species," while in the second arrangement it stands together with D and E, which, 

 accoi-ding to arrangement I., are specifically different from j4,i?, and C The danger 

 of arriving at erroneous results by taking into account onlv one organ is so evident 

 and so well known that one must wonder how it is possible that nevertheless a classi- 

 fication — for example, of Lepidoptera — based upon wing-markings alone, or upon 

 neuratiou, can be expected to be an exact expression of the blood-relationsliip of the 

 classified forms. 



Which characters are correlated and wliich independent in respect to variation 

 can only be found out by the .study of individual variation; where this study is 

 neglected correlated characters are often taken to be independent diagnostic char- 

 acters. The length of the wing and the breadth of the wing-bands are often correlated 

 in Lepidoptera, and so is in many cases the lesser or greater concavity of the outer 

 margin of the forevring of butterflies depending on the length of the wing ; to say in 

 such cases that a certain form of butterfly differs from another form in the wing being 

 longer, externally more concave and liaving nanower liands, would amoimt to the 

 same as if in the case of two black forms with yellow spots we should say tluit the one 

 is differentiated from the other by the yellow spots being restricted and by the black 

 ground-colour being extended. 



