256 BUMP US. [VOL. I. 



curve. The curve of growth for Casco Bay is also more 

 variable, as it is for Beverly, Nahant, Plymouth, Seaconnet, 

 and, indeed, for all the American localities, without a single 

 exception. 



The Fourth Test of Variability. 



If we weigh the empty shells of 200 snails from Tenby, 

 Wales, all of approximately the same height (18 to 19 mm.), we 

 will find that they vary in weight from 1.4 grams to 2.5 grams; 

 i.e., the ir.dex of their amplitude of variation according to weight 

 is i.i. Those from South Kincardineshire and those from the 

 Humber District vary from 1.5 grams to 2.4 grams, the index 

 of variation being in both cases .9. The curves of distribution, 

 according to weight, of the shells from the three British local- 

 ities are represented on Chart XIV by dotted lines, and similar 

 curves of weight for nine American localities are represented 

 by entire lines. (The shells from the St. Croix River were not 

 weighed, as they were cleaned in a manner different from those 

 from other localities, and their introduction would lead to 

 error.) 



In every case it will be noted that the length of the base 

 line of the curve of the American shells, even when there are 

 less than 200 (indicated by the figures at the left of the chart), 

 is equal to or exceeds in length the base line of the curve of the 

 most variable British shells. The index of the amplitude of varir 

 tion according to weight of the most variable British shells is 

 11, while the amplitude of five of the least variable American 

 shells is also 1 1 ; four American localities, however, have an 

 increased amplitude; viz., 12. The least amount of variation 

 among the British shells is 9 ; the least among the American 

 is ii. 



The lightest American shell weighs i gram ; the lightest 

 British shell, 1.4 gram. The heaviest American shell weighs 

 2.6 grams; the heaviest British, but 2.5 grams. The lightest 

 and heaviest shells are American. 



