20 



SCIENCE 



[N. S. Vol. XXX. No. 757 



ther than Galton, and precisely confirms my 

 own estimate, made in 1902, of about nine 

 tenths for heredity. Barrington and Pearson 

 have recently found that the influence of en- 

 vironment on sight is nil* Thus the predic- 

 tion is already strongly in favor of future 

 investigators arriving at a similar result, that 

 the ordinary influance of environment on the 

 higher human attributes is at most but trifling 

 when the heredity factor remains the same, 

 or when the heredity factor can he measured, 

 or eliminated from the discussion. 



Usually it is not possible to separate hered- 

 ity from environment. We often merely find 

 some correlation which may be explained as 

 due to either or both of these forces. An 

 example of this sort of correlation is the one 

 I found to exist between general superior 

 achievement in the United States and city 

 birth. I have already explained in my other 

 article why some correlation is to be expected 

 from inheritance alone, while from environ- 

 ment it may or may not be expected. No one 

 is in a position to speak on the latter question 

 because no one knows whether, on the whole, 

 the good and bad sides of city life strike a 

 balance in favor of the city against the good 

 and bad sides of country life. And even if 

 the answer to this complicated problem were 

 known, we should not then know if there were 

 anything efiicacious enough to produce a meas- 

 urable result, as I have explained above. 



The failure to find a higher ratio for the 

 cities would have been a serious blow for 

 heredity. The finding of a higher ratio for 

 cities merely wards off a possible attack. It 

 is a purely negative defense, and this is all I 

 have claimed for it. Now, in his second letter 

 Mr. Spillman positively asserts : " Dr. "Woods's 

 own figures prove the effect of environment as 

 against heredity." In the light of the expla- 

 nation in my former letter, which I have here 

 just now, in other words, repeated, I ask. 

 How can he possibly know this ? How can he 



* " A First Study of the Inheritance of Vision 

 and the Relative Influence of Heredity and En- 

 vironment on Sight," Eugenics Laboratory Mem- 

 oirs, v., 1909. 



know that the figures are even beyond the 

 expectation from the heredity factor alone? 



Mr. Spillman then complains that the per- 

 sons listed in " Who's Who in America " do 

 not represent leaders. I do not wish to enter 

 into a discussion on the use of words, and for 

 the sake of the argument will grant that they 

 shall be called merely " competent workers," 

 but I can not refrain from saying in passing 

 that " leaders " seems a very fair word to 

 apply to a group so small as 16,000 out of a 

 total population of some 80,000,000. I should 

 think of the colonel of a regiment of a thou- 

 sand as the leader of that thousand and it 

 would be in no less just a way and with as 

 good a proportionate sense to call the higher 

 railroad officials, greater bankers and relatively 

 few doctors, lawyers, etc., whose names are 

 included in this same volume " leaders " in 

 their special fields of activity. But even if 

 we are to call them mere " competent workers," 

 are not just such " competent workers " to be 

 desired, whatever be the cause of their com- 

 petency ? 



I knew that " Who's Who " itself would be 

 criticized, therefore I forestalled this criticism 

 by the following: 



Some will not be willing to accept conclusions 

 drawn from a list which like this doubtless has 

 certain flagrant omissions, and where he sees 

 names that he considers should not have been 

 included. If he will stop for a moment and think, 

 he will see that the very objection he raises only 

 argues in the other direction from what he sup- 

 poses. If, for instance, I flnd a marked correla- 

 tion between city birth and more or less notable 

 subsequent achievement, drawn from an imperfect 

 list, the correlation would be even higher were the 

 list of names ideally perfect. 



Mr. Spillman makes no reference to this. 

 The same applies whether the list be lower in 

 standard than it should be owing to errors 

 within itself, or whether it be in general a list 

 showing a low standard of selection. Pro- 

 viding of course that the standard be above 

 the general average of the population, then it 

 follows that the higher the standard the higher 

 would be the correlation, at least as far as 

 mathematical expectation is concerned. 



Mr. Spillman then takes a prop from the- 



