186 



the result is unsatisfactory ; if they are ten to one the graduation 

 is not unreasonable, but the exact value when a result must be 

 discarded cannot be given. As, however, it is clearly impossible 

 to imagine any test which can fix an absolutely definite 

 standard, there is no reason for objecting to the particular 

 method because it fails to do so." 



The results of these calculations of P are tabulated below :- 



Table I. — Calculation of P on sub-samples of Manx herrings 

 taken during 1914, on characters D and l.cp.l. 



