136 PROCEEDINGS OF THE AMERICAN ACADEMY. 



The results are as follows : — 



Length hinge joint 

 Locallty - Anteroposterior length ' 



Eddystone Light .507 



Irish Sea .483 



Firth of Forth .473 



Thus the Firth of Forth shells have relatively the shortest hinge 

 joint ; those from Eddystone the longest. 



We have now to consider the portions that the anterior and posterior 

 ears, respectively, bear to the entire hinge joint. This proportion is im- 

 portant because in many species, e. g., Pecten varius, the posterior ear is 

 only half as long as the anterior, or even less ; that is, the proportions are 

 33 per cent and 67 per cent, respectively. In Pecten opercularis, also, 

 the posterior ear is the shorter ; the proportions fpr the three localities 

 are as follows : — 



That is to say, near the Eddystone Light Pecten opercularis has the pos- 

 terior ear proportionately shortest (46:54), while in the Irish Sea the 

 two ears are most nearly equal (48:52). 



b. Variability of the Ears and their Correlation with the Antero- 

 posterior Diameter. — The correlation between the antero-posterior di- 

 ameter on the one hand and the anterior ear and the posterior ear on the 

 other has been worked out for 541 lower valves from the Irish Sea. 

 The correlation is sensibly the same in both cases; for the anterior ear, 

 r — 0.768 ± .009, and for the posterior ear r = 0.779 ± .009 (Appendices 

 I, K.) 



The standard deviation for the length of the anterior ear is 1.375, and 

 for the posterior ear, 1.378 ; the coefficient of variation is in the two 

 cases 7.97 ± .16 and 9.30 ± .19, respectively. Since these are graduated 

 variates, the coefficient is the more significant index of variability.* The 

 coefficient of variation is 17 per cent higher in the posterior ear than in the 

 anterior ear. So far as this case goes, then, it confirms for a mollusc 

 the results of Brewster for man and mammals and Heincke for fish. 



* See my 1900 paper, p 962. 



