THE DISTRIBUTION OF VARIANTS IN NATURE 95 



(i) Each colony has a rather wide range of variation 

 in the two characters, rather more in respect 

 of altitude. In altitude the ' spread ' is very 

 wide, i.e. no one class is very frequent. In 

 diameter there is a distinct tendency for a 

 high grouping about one phase. This is 

 very well seen in the figure of ' polygons of 

 variation ' (op. cit. pp. 59 and 61). 

 (ii) In general the series from neighbouring loci are 

 more or less alike, but the converse is not true. 

 No two loci have exactly the same mean. 

 The shell-characters are not correlated with 

 the ecological characters of the various loci. 



(b) Boycott (191 9, 1927) also studied the shape of the 



shell in C. rugosa. He found the same amount of 

 variation in each locus and that there was no relation 

 between the former and the character of the locus, 

 though he suspected some relation between shell- 

 altitude and environment. ' Significant ' differences 

 were found in 5 out of 6 pairs of contrasted characters. 



(c) Aubertin (1927) studied a number of colonies of 



Cepea nemoralis and C. hortensis both for shell-colour, 

 etc., and anatomical characters. The former only 

 are considered here. The number of specimens used 

 is rather low. 



(i) C. hortensis. — Each colony has a wide range of 

 variation and in three out of four ' adjacent 

 colonies ' there was good ' spread ' for ground 

 colour. In one (' Hedge Lane ') yellow was 

 90 % of the total. For three types of banding 

 the spread as between type 12345 an d 00000 

 was equal. Some colonies lack a particular 

 ground colour-class altogether. 



Adjacent colonies tend to be different signifi- 

 cantly in ground colour, less so in banding. 

 A Buckinghamshire colony closely resembles 

 one Wiltshire colony, though it differs in the 

 absence of a colour-class found in the latter. 

 (ii) C. nemoralis. — In colour some colonies lack 

 certain classes altogether, as in (i) ; but this 



