40 PROCEEDINGS OF THE MALACOLOGICAL SOCIETY. 



exist, and five of the ten possible differences in diameter. 

 Evidently, therefore, a beech-wood such as I am dealing with is 

 not a homogeneous locus qua the size of CI. bidentata. The data 

 for Ena ohscura, unfortunately with less ample material, given in 

 tables V and VI show that the same differentiation is shown 

 by this species, though to a less degree ; seven of the ten possible 

 differences exist in altitude, diameter, or both, 



Table IV. — SHOWING THE SIGNIFICANT DIFFERENCE IN 

 ALTITUDE ( + ) AND DIAMETER (o) FOR THE FIVE AREAS 

 COMPARED WITH ONE ANOTHER, GL. BIDENTATA. 



Table VI. — SHOWING THE SIGNIFICANT DIFFERENCE IN 

 ALTITUDE ( + ) AND DIAMETER (o) FOR THE FIVE AREAS 

 COMPARED WITH ONE ANOTHER, E. OBSCUBA. 



A B G D E 



+ 



§ 6. Our beech-wood, then, does not form a locus in the sense 

 that it is similar in all its parts as regards the size of the two snail- 

 shells we have considered. The snails which live in different parts 

 of it clearly differ in size, and on the basis of their differences the 

 wood, which in a general way is homogeneous, may be dissected into 

 many loci. It is an obvious question whether such varieties as these 

 are correlated with variations in external circumstances,, or whether 

 they should be regarded as fortuitous results of relative isolation ; 

 clearly snails living several hundred yards apart cannot be suspected 

 of much interbreeding. Bateson ^ says very truly that we have in 

 the past been too ready to find the explanation of local differences 

 in the localities rather than in the organisms. The present data 

 may, I think, throw some light on the point. If the local differences 

 arise from mutation within the organisms, the variations in 

 Clausilia bidentata should have little or no relation with those in 

 Ena obscura ; if, on the other hand, they are caused by differences 

 in environmental circumstances it is possible that the variations 

 in the two species would run more or less parallel. Such proves 

 to be the case in the present instance, for if we arrange the loci 

 in descending order we get : — 



' Problems of Genetws, I9I3, p. 131. 



