No. 435.] SHELL OF LITTORINA U TORE A. 



1 Si 



in both size and shape resulting from erosion. While the absolute 

 amounts of these changes seem small, they are relatively large 

 enough to falsify any comparisons that may be made between 

 the variabilities and types at different localities or at different 

 ages, unless a correction for erosion can be applied. 



For our measurements would show, not the normal variability 

 of growth, but the combined effect of variability of growth, 

 variability in power of resistances to erosive agents, and perhaps 

 variability of these agents themselves. To obtain data for the 

 correction of our results would involve an investigation of the 

 rate of erosion that would be too laborious to be undertaken. 



Our attention was drawn to this species especially by the very 

 interesting work of Bumpus ('98), who made a comparison of 

 the variability and types in shells from three English localities 

 with shells from ten stations on the New England coast. He 

 concluded that in all the characters studied the shells from all 

 the American localities (that is, those in the comparatively new 

 environment) were more variable than those from any of the 

 English localities. And he stated also that the American type is 

 " more enlongated, 1 lighter in weight, more bulky, and the color 

 markings are less pronounced." 



His tests of variability were: — (1) Ratio of breadth and 

 length, (2) Comparison of the extremes of this ratio, (3) Com- 

 parison of curves for different ages, (4) Weight, (5) Bulk, and 

 (6) Color. Our results show that the first five of these tests of 

 variability are rendered of doubtful value by the erosion. Or at 

 any rate, the factor of erosion must be taken into account in the 

 discussion of these results, for, as has been shown, the ratio of 

 breadth to height is very considerably affected by the loss in 

 height, the difference is the ratios amounting to as much as .05 

 or .06 ; while the total range of variation found by Bumpus was 

 .17 in the British and .24 in the American series. And the dif- 

 ference between the means of the two series was only .014, much 

 less than the probable effect of erosion in an apparently normal 

 shell such as is represented in Fig. 3. 



