2 HiCKLING, Variatioji of Planorbjs multiformis, Bioiin. 



tinuous variation of a single species. On the contrary, if 

 several species or sub-species were present, there would 

 be no reason why the various types should be represented 

 in any definite proportion. It therefore appeared desirable 

 to ascertain exactly the relative numbers of individuals 

 of various types, since the variation-curve so obtained 

 would give the most decisive possible answer to the 

 question of the existence of one or more specific units. 



The problem of constructing a variation-curve is com- 

 plicated by the fact that the shells vary independently in 

 several characters, of which the height of the spire is 

 merely the most conspicuous. The existence of these 

 concomitant variations makes it almost impracticable to 

 apply any system of measurement to the shells, while an 

 elaborate statistical enquiry is neither necessary nor justi- 

 fied by the relatively small amount of material available. 

 It was therefore deemed sufficient to investigate the 

 variation in the " height-ratios " of the shells 

 /. , . height of shell \ 



\ ' diameter of base/ 



by sorting the shells in respect of this character into ten 

 grades by eye-estimation. The first step was to select 

 ten typical shells, differing in height as nearly as possible 

 by equal amounts, to form a scale of comparison covering 

 the whole range of variation. The whole of the material 

 was then sorted into ten grades by comparison with this 

 scale. 



The shells examined were of various stages of growth, 

 and they were accordingly divided initially into three 

 groups, according to size : in the smallest, the diameter of 

 the base is from 008 to 012 inch ; in the median set, 012 

 to 0'20 inch ; in the largest, 020 to 0"28 inch. This was 

 done to facilitate both the sorting and the interpretation 

 of the results. A similar type set of comparison-shells 



