November 23, 1906.] 



SCIENCE. 



665 



in the separate sciences — usually falling within 

 the probable error of position. But the acad- 

 emy has no method of comparing performance 

 in different sciences, and if one science has 

 less than its proper representation, the dis- 

 parity is likely to increase rather than to de- 

 crease. Thus there are in the country about 

 haK as many astronomers as botanists, but 

 there are twice as many astronomers in the 

 academy. The second principal variation in 

 the membership of the academy is due to the 

 fact that men do not always retain the posi- 

 tions that they hold when elected. Apart 

 from the somewhat greater accuracy, the su- 

 periority of this list consists in the assignment 

 of probable errors of position. Thus the prob- 

 able error at the close of the first hundred is 

 about 25 places, that is, there are about 25 

 men not in an ideal academy of a hundred, 

 whose chances of belonging there are at least 

 one in four. A list such as this would also 

 give us academies of any desired size — the 

 sixty most eminent m.en of science, as in the 

 Paris Academy, the hundred or thereabouts 

 as in the National Academy, or the 450 or 

 thereabouts, as in the Eoyal Society. 



While under existing conditions of senti- 

 ment, the publication of a list of our thousand 

 leading men of science in the order of merit 

 with the probable errors would not be toler- 

 ated, I have indicated those who are included 

 in the thousand in my ' Biographical Direct- 

 ory of American Men of Science,' a work of 

 reference that may be regarded as a by- 

 product of this study. I did this with some 

 hesitation, but it seemed best to place on 

 record those who were the subjects of this 

 research, more especially as this could be done 

 without any invidiousness. The probable 

 error toward the end of the list is about 100 

 places, so there are one hundred others who 

 have at least one chance in four of belonging 

 to this group. Further, several scientific men 

 of standing were omitted from the lists as 

 originally drawn up, and were not considered 

 in making the arrangements. Consequently, 

 while each of those indicated in the Biograph- 

 ical Directory is probably one of the leading 

 thousand American men of science, there 

 are others not indicated who belong to this 



group. This, however, is a minor factor, and 

 we have with sufficient accuracy for statistical 

 purposes a group of the leading thousand 

 American men of science arranged in the 

 order of merit with the probable errors of posi- 

 tion known. J. McKeen Cattell. 

 Columbia University. 



a note on assortative mating. 



In the natural selection and topographic 

 isolation theories combined there is offered a 

 plausible explanation of the means whereby 

 species may be derived from other species, 

 granted that the derived species do not occupy 

 the same geographic (topographic) range as 

 the parent form. Where they do, some new 

 aid to natural selection in place of topographic 

 isolation must be invoked to explain how 

 slight variation may be progressively increased 

 until differences of selective worth exist be- 

 tween parent and splitting-off types. De- 

 terminate variation and physiological isola- 

 tion are two such aids proposed. The latter 

 (with which Romanes's name is familiarly as- 

 sociated) assumes that among the members 

 of a species living in the same locality there 

 may be among the inevitable slight fluctuating 

 variations some of such a character as to lead 

 to assortative mating, *. e., that individuals 

 of certain like variation may tend to mate 

 together, either because of mutual attraction 

 between like, or of mutual repulsion between 

 unlike forms. This tendency to selective or 

 assortative mating between like individuals 

 may come to result in time in such an increase 

 of differentiation among groups of individuals 

 of the species, although these groups may live 

 side by side or confusedly mingled with each 

 other in the same locality, that mating between 

 unlike groups will become physiologically im- 

 possible. That is, that these groups will con- 

 stitute distinct species. 



The facts of observation or experiment ad- 

 duced to support this theory are very few. 

 Indeed, I do not recall any at the present 

 moment. Nevertheless, the need of an aid to 

 selection capable of bringing slight continuous 

 fluctuating variation up to a life-and-death 

 selective value, and the generally plausible 

 character of this theory of Romanes (and of 



