No. 553] DISTRIBUTION OF THE GHjBSTOGNATHA 41 



Finally we find that Group 4 of similar distributional 

 species consists of S. serratodentata, K. snbtilis, and E. 

 hamata, and by referring to the key (p. 36) we see that 

 they are separable only by the fact that S. serratttdentata 

 occurs to some extent above 25 fathoms, while E. hamata 

 never occurs above this depth, and K. subtilis never 

 occurs above 50 fathoms. Yet, we have as members of 

 this group three species belonging to three genera, so 

 that there can be no question regarding their funda- 

 mental morphological difference. 



In what way then do these facts answer our question : 

 "To what extent are morphological differences between 

 species proportional to, or correctable with, their distri- 

 butional differences?" It is obvious that the only reply 

 permitted by our data is that there is a very definite cor- 

 relation, but one that is the exact reverse of what would 

 a priori be expected on the basis of the Darwinian theory 

 of "natural selection"; namely, that the morphological 

 difference between two species is inversely proportional 

 to their distributional difference, or, to state it otherwise, 

 the coefficient of correlation between morphological and 

 distributional differences among species approximates 

 closely to — 1. 



Relation between Species Obtained in the Same Hauls 

 with Respect to Sexual Maturity 

 Under this head it is proposed to briefly consider the 

 evidence of physiological isolation or coincidence between 

 species relative to their maturity in those cases where 

 two or more were obtained in a single haul. It is obvious 

 that open vertical and vertical closing net hauls do not 

 yield data relevant to this question, for the reason that the 

 vertical distance covered is so great (25 fathoms or more) 

 that it is impossible to tell whether the specimens of two 

 or more species were taken from the same depth or not. 

 Concerning the horizontal hauls, however, this objection 

 can not be made, and when they are examined we find that 

 only 14 out of 148 surface hauls and 23 out of 108 closing- 

 net hauls obtained more than one species. 



