318 



THE EFFECTS OF CONDITIONS 



ter to the observations made by Weldon * on the corre- 

 lation between certain dimensions in local races of the 

 shrimp. The degree of correlation was found by him 

 to be practically constant, but the mean measurements 

 themselves show distinct differences in the various local 

 races. The variability or range of variation of the 

 measurements about their mean shows much greater 



differences still. As we see in the accompanying 

 table, the total carapace length (expressed in terms 

 of the body length, taken as 1000) varies in different 

 localities from 247.33 to 251.38, or by 1.6 per cent. 

 However, the probable error of variation of this dimen- 

 sion varies from 3.29 to 4.55, or by no less than 38.3 

 per cent. The mean post-spinous carapace length 

 varies in its extreme limits by 2.06 per cent., and its 

 probable error by 38.1 per cent. Now in that the 

 shrimps differ so little in their average dimensions, they 

 cannot be very divergent races, and hence one must 

 conclude supposing, of course, that the samples meas- 

 ured were fair ones, collected under similar conditions 

 that the differences in the variability of the shrimps 

 obtained from the various regions are due chiefly to the 

 * Proc. Roy. Soc., vol. li. p 2, 1892. 



