No. 5.] THE INTRODUCED L1TTORINA. 251 



Let us now examine the curves of distribution of the 

 American shells, and see if a new environment has wrought 

 any change. 



Chart IV. - -This chart is based on an examination of 1000 

 shells from the St. Croix River in Maine, and, while it does 

 not rest on the base line in an extreme position, its general 

 trend and the width of its base are very different from any of 

 the British curves. The index of stature varies from 84^ to 

 102^), giving an amplitude of 18 ($o c /o greater than that given 

 by the Humber shells), though the regular trend of the flaring 

 base is still an indication of some continuously active law. If 

 a larger number of shells of the complement have departed 

 from the ideal and have presented more extreme types of elon- 

 gation on the one side and of ventricosity on the other, the shells 

 of mean stature must have been depleted, and the St. Croix 

 curve is consequently less precipitous than any of the British. 



Chart V. - -This chart is based on an examination of 1000 

 shells from Casco Bay. The position of the curve indicates 

 a general tendency towards ventricosity. There are but twenty- 

 two shells the index of which is below SSfi, while there were 

 among the South Kincardineshire collection 440 examples. The 

 altitude and general configuration of the present curve is not 

 unlike curve IV, and the amplitude of variation of stature, as 

 indicated by the breadth of base, is but slightly greater; viz., 

 19 (ios--84= = 19). 



Chart VI. --On this chart 1000 shells from Beverly, Mass., 

 have been tabulated. Beverly is located at the mouth of a 

 small inlet, ninety miles southwest of Casco Bay. The curve 

 is typically American, the base indicating an amplitude of 

 variation of 20 (102 -- 82 = 20), an amount far in excess of 

 the most variable British shells. It is worthy of note that the 

 right limb of the curve is much less abrupt in its descent and 

 longer in its course than the left, a feature which also distin- 

 guishes the other American curves thus far examined. The 

 significance of this character will be considered later on. 



We have now examined 3000 British and 3000 American 

 shells, with the result that in every case the latter are more 



