No. 526] VARIATIONS IN UROSALPINX 



total of 50,424 — it is possible to approximate a measure 

 of the variability of Urosalpinx as a species much more 

 nearly than is possible with smaller lots of 1,000. Such 

 a combination is shown in Table XII, which will be seen 

 to furnish the figures for a curve of considerable regu- 

 larity in which the arithmetical mean is 61.662 and the 

 standard deviation is 3.367 ±.0071. This standard devia- 

 tion is exceeded in but a single instance among the 

 smaller lots which make it up — namely, in the 1,664 

 shells from Staten Island which show a standard devia- 

 tion of 3.508±.0412. 



PerCent. 50 51 52 53 54 55 56 57 58 

 No. 4 15 32 120 289 675 1,510 2,450 3,812 



PerCent. 59 60 61 62 63 64 65 66 67 

 No. 5,052 5,491 5,861 5,515 5,115 4,225 3,647 2,357 1,714 



PerCent. 68 69 70 71 72 73 74 75 76 

 No. 1,161 737 373 181 54 21 8 4 1 



A. M., 61.662. e, 3.367. P. E., ±.007. Total No., 50,424. 



12. Summary.— 1. When two lots of 1,000 Urosal- 

 pinx shells each are taken from the same locality they 

 resemble each other sufficiently to indicate a character 

 typical for the locality. 



2. Lots of shells from different localities vary widely 

 enough from each other to be easily distinguished, indi- 

 cating thereby that the varying environment associated 

 with different localities exerts a measurable effect. 



3. Endemic Atlantic shells (with one exception noted 

 below) vary less than shells introduced into a new en- 

 vironment (California). 



4. The shells of Buzzard's Bay have a lower ratio of 

 greater shell-aperture to shell-height than those of 

 Vineyard Sound. 



5. When shells from the same localities in successive 

 fortnights are compared there is an increase in the 

 ratio of greater shell-aperture to shell-height (A. M.) 

 and also a slight increase in variability as shown by 

 the standard deviation, except in the case of the shells 

 from Penzance Point. 



