566 



NATURE 



[October 3, 1907 



exposed. The latest report of the U.S. Commissioner of 

 Education deals with the y.ear ending June 30, 1905. The 

 total income for that ye'ar of American institutions of 

 university rank, excluding benefactions, amounted to 

 8,355,000/., an increase of 289,200/. over the preceding 

 year, and of this amount 23-6 per cent, was from State 

 appropriations and 69 per cent, from Federal appropri- 

 ations. That is, more than 2,506,500/. was provided from 

 American public funds for higher education durinj; the 

 year with which the report deals. A very generous 

 estimate of the amount provided here from public funds 

 for higher education o( every kind, including the Royal 

 Colleges of Science of Londoti and Dublin, the universities 

 and the university colleges, would be to place it at a 

 quarter of a million pounds sterling, so that the British 

 case is in no way improved by importing the question of 

 the amounts provided for higher education. 



Sir W. T. Thiselton-Dyer is doubtful only about the 

 greater belief of American statesmen in the need for the 

 introduction of scientific methods in the solution of 

 problems of government, but of the need of scientific ways 

 of thinking on the part of our legislators he is quite con- 

 vinced, and that is reallv the important matter. 



A. T. S. 



The Interpretation of Mendelian Phenomena. 



Apropos of the discussion on the interpretation of 

 Mendelian phenomena, may I seek enlightenment on one 

 or two points from your readers? Mendelian phenomena 

 are possible onlv when reproduction is bi-parental. They 

 cannot occur, of course, when it is parthenogenetic. I 

 believe I am right in thinking that Mendelian workers 

 suppose or hope that they have found a master key to 

 the problems of heredity. Now, I am able to understand 

 that the study of alternative inheritance may ultimately 

 shed a light on the function of sex, but I find it ditificult 

 to conceive how it can shed a light on any other bio- 

 logical problem of importance ; for example, the problems 

 of the alleged transmission of acquirements, of the causa- 

 tion of variations, of the retrogression of characters which 

 have lost selection value, and of the mode of development 

 (whether or not by the recapitulation of the phylogeny). 

 All these problems are of at least equal importance to 

 the problem of sex. I have sought information from my 

 Mendelian acquaintances, but I am always told that we 

 must await the accumulation of data — a somewhat 

 Micawber-like attitude, as it seems to me. I hope I make 

 myself clear. The information I seek would be contained 

 in the answer to the following question : — If Mendelism 

 has a bearing on any biological problem save that of sex, 

 what is that problem? If, as I anticipate, no one is able 

 to name another problem, I venture to suggest that 

 Mendelians are engaged in nothing more than the investi- 

 gation of sex. 



Mendelian phenomena have been observed principally in 

 crossed artificial varieties of animals and plants. Crossed 

 natural varieties usually blend their characteristics. This 

 is conspicuously the case with man, the animal who, so 

 far as is kno\Vn, has crossed more often than any other, 

 and whose hybrids may be observed up to the tenth or 

 twelfth generation in South America and elsewhere. It 

 has been said that " human skin-colour is the only 

 character that is known to blend perfectly"; but this 

 statement is certainly incorrect. With the exception of 

 eye-colour, and possibly one or two other traits, such as 

 the Mongolian eyelid, human hybrids appear to blend every 

 char.-icter as perfectly as skin-colour. The transmission 

 of no character is Mendelian. Thus mulattos have the 

 black eye of the negro, and when they breed inter se con- 

 tinue to reproduce it indefinitely. There is no segrega- 

 tion. May I mention one other fact which is of con- 

 siderable interest, but which seems to have escaped the 

 attention of Mendelian workers? Crossed artificial varie- 

 ties usually reveal latent characters in abundance. I am 

 aware that the correctness of the term latency has been 

 disputed, but it will serve to indicate what I mean. So 

 far as I have been able to ascertain, no single instance 

 of a latent character resulting from the crossing of 

 natural varieties has been recorded. Certainly crossed 

 human varieties reveal no surh traits. A very signifi- 



NO. 1979, VOL. 76] 



cant passage bearing on this matter may be found 

 in " Animals and Plants " (vol. ii., pp. 24-5). It would 

 appear, then, that characters become latent only under 

 conditions of artificial selection, that is, when mutations 

 are selected. It has been maintained that nature also 

 selects only mutations, but, to say the least, this has not 

 been demonstrated as yet. 



Bearing in mind, then, the facts that latent characters 

 appear only when artificial varieties arc crossed, and thai 

 crossed natural varieties usually blend their characters, the 

 question arises whether Mendelians, so far from investi- 

 gating even the whole problem of sex, are engaged in 

 anything more than the investigation of those abnormali- 

 ties of se.xual reproduction which occur under conditions of 

 artificial selection. G. .Vrcmuall Reiu. 



Southsea, September 17. 



On Correlation and the Methods of Modern Statistics. 



Prof. K.arl Pe.^kson's letter in N.ulre of September >(} 

 gives me a welcome opportunity of explaining what was 

 not intelligible in the condensed report of my remarks in 

 the discussion at Leicester, on methods of modern 

 statistics. 



Prof. Pearson communicated to the Royal .Astronomical 

 Society (Monthly Notices, May, 1906) a paper by Miss 

 Winifred Gibson, giving an account of a research con- 

 ducted in the statistical laboratory of University College. 

 London. The first part of this paper discussed the rela- 

 tion between parallax and magnitude of the stars. I con- 

 fessed to some misgivings as to the astronoinical value of 

 the results, and raised two questions, first, as to the 

 method, and second, as to the matter. 



Prof. Pearson thinks that I am on safer ground in the 

 second than in the first. I will therefore examine first 

 his reply to my second point, which was that the parallax 

 material contained in Newcomb's table (appendix to " The 

 Stars ") is quite unsuitable for discussion by a general 

 statistical method, since it relates very largely to stars 

 selected for investigation because of abnormal proper 

 motion. 



Prof. Pearson " fancies that astronomers have been 

 guilty of a considerable amount of circular reasoning. 

 They start from the hypothesis that magnitude is very 

 closely related to parallax. . . . The fundamental hypo- 

 thesis that the brighter stars are much the nearer as yet 

 awaits statistical demonstration. . . . Surely the hypo- 

 theses of high relationships between magnitude and 

 pai^Ilax and proper motion are of suflicient importance 

 to deserve proof, rather than to be taken as axiomatic." 

 In this matter Prof. Pearson is under a misapprehension. 

 Astronomers do not believe that magnitude is very closely 

 related to parallax ; very obviously it is not. But they do 

 believe that parallax is somewhat closely related to proper 

 motion. There are seventeen stars in the sky brighter 

 than mag. 1-5, and their parallaxes have been determined 

 with the heliometers at the Cape and at Yale. Here are 

 the results. 



Seven have proper motions (on a great circle) less than 

 o".i per annum : — 



P.M. Parallax Mag. 



Canopus 



Deneb 



Rigel 



Betelguese 



Antares 



Achernar 



j3 Centauri 



Mean ... 004 ... 002 ... o'6 

 Six have proper motions between o".i and i".o : — 



Aldebaran 

 Regulus ... 

 Vega 



Capella ... 

 Pollux ... 

 Altair ... 



