THE OSTEACODA. 



269 



Against both the foregoing possibilities may also be urged the argument from constant 

 association. Of the 65 comparable hauls in the epiplankton, 48 contained what I term 

 Conchoecia magna *". Of these 48 hauls, 81 per cent, contained at least two stages, and 

 45 per cent, contained all three ; the association was therefore very constant. 



The third explanation — almost the same as the first, but in other words — is that the 

 species is unstable, is beginning to split up into a larger and a shorter race. So long as 

 a species is stable it seems probable that any character measured and plotted graphically 



The two lower curves on the right hand of the ahscissa are of both sexes of " A " larvae mixed. 



will yield a curve of frequency with a marked mode ; but if the species tends to vary in 

 respect of that character into two races, the curve will ultimately show a double hump 

 with a central depression between the two modes about which the measurements are 

 grouped; the depth of this depression has even been used as an "index of isolation." 

 But before the two races have been so far established as to show the double-humped 

 curve, there will presumably be a transition period at which the curve, instead of 

 presenting a tall mode, will show a flattened apex, and this condition is apparently 



Of the remainiug 17, 11 contained Halocyprids but not magna. 



44* 



