4 HiCKLiNG, Variation of Planorhis vniltiforniis, Bronn. 



departs from the mean in either direction — and con- 

 sidering the small number of shells dealt with, and the 

 necessarily crude method by which they were sorted, the 

 regularity of the curve is very striking. It is a typical 

 " continuous variation " curve, or " curve of error," such as 

 would represent the frequency of height or of length of 

 span in man.* If the shells of different heights repre- 

 sented different species, or races, or "mutants," the chances 

 against the various forms being present in the regularly 

 distributed numbers in which they are actually found would 

 be almost infinitely remote. On the contrary, the simple 

 variations of any character in a single organic type are 

 necessarily distributed in these proportions, unless dis- 

 turbed by the selective action of external factors. Such 

 a variation-curve is the most conclusive proof possible 

 that the group of individuals from which it is derived 

 belong to one indivisible organic unit, z>., form one species 

 in the strictest sense. Systematists may apply distinctive 

 names to different forms for convenience of description if 

 they choose, but these divisions have no objective exist- 

 ence. In passing, it may be remarked that this method 

 of enquiry would afford a most satisfactory means of 

 settling disputed questions of the specific unity or other- 

 wise of any series of organisms. 



There appears to be one case only in which the evi- 

 dence of such a curve is not conclusive. It is possible to 

 get a number of related species which present exactly 

 similar variations in respect of a given character, and in 

 such a case the curve for that character will not be 

 affected if the species are mixed. Such cases occur both 

 as the result of convergence produced by a common 

 environment and as the effect of parallel development 

 from a common ancestry. It will be shown in the sequel 



• Certain peculiarities of the curve are dealt with below. 



