October 17, 1907] 



NA TURE 



615 



hold that this relatively low value is the second point her 

 paper has indicated for the first time. Meanwhile, I leave 

 Mr. Hinks to consider whether the proven correlations 

 of both proper motion and magnitude with paralla.K 

 (under 0-4) have or have not any significant bearing upon 

 the differential method of determining parallax, and upon 

 the fact that more than 20 per cent, of negative parallaxes 

 can be found. 



(c) The third point in Miss Gibson's paper was the 

 statement that colours (and probably spectral classes) were 

 more highly correlated with magnitude than distance. 

 Mr. Hinks takes this point as one which will fully justify 

 his criticisms at Leicester. I am of opinion that it is 

 peculiarly , a case in which he would have done well to 

 have tempered his judgment by previously asking whether 

 there was no method in our madness. He charges us 

 with three grievous offences : — (i) using a highly selected 

 material ; (2) omitting to take into consideration the 

 "white stars"; and (3) deducing from such material 

 sweeping conclusions about the stars in general. He 

 further, charges me, on the basis of this investigation, with 

 asserting " that colour and magnitude are related at least 

 as closely as parallax or proper motion and magnitude." 



In my letter, when making my statement, I made no 

 reference to Miss Gibson's published work, but the fact 

 that I cited the value of the correlation of magnitude and 

 spectral class which is not . given in the published paper 

 might have warned Mr. Hinks that we 

 held other reductions in our hands. 

 Mr. Hinks asserts that our results 

 would have been modified had we in- 

 cluded the " white " stars. Using the 

 2834 stars of the Harvard Catalogue, 

 of which roughly one-quarter are white 

 (" Annals," vol. xiv.). Miss Gibson 

 worked out more than a year ago the 

 contingency of' colour and magnitude ; 

 the value was 0-27 ( + 6-oi), as com- 

 pared with the 0-30 ( + 005) of the list 

 in the Cape Observatory " Annals " 

 previously given by her. Omitting the 

 white stars from the Harvard data, 

 the value is 0.297, agreeing absolutely 

 with the result obtained from the Cape 

 data. Thus we see that Mr. Hinks 's 

 suggestion that the Cape Catalogue is 

 worthless, owing to selection of special 

 stars, has no validity at all when we 

 turn to the relationship of colour and 

 magnitude, and, further, the inclusion 

 of white stars produces, as we 

 had logically anticipated, no sensible 

 effect. 



But I wilt go a step further, and reveal another con- 

 clusion, which I should naturally have preserved for 

 the present, as the research is as yet incomplete. The 

 mean magnitude of the white stars is almost identical 

 with the mean magnitude of all the remaining truly 

 " coloured " stars; the white star has not a preponderance 

 of any special part of the colour spectrum, and if we 

 wish to investigate the relationship between luminosity 

 and colour we must logically leave out the white stars. 

 The accompanying diagram gives the Harvard stars 

 classed according to colour, with (a) the mean magnitudes 

 of each colour group, and (b) the corresponding luminosity 

 on the assumption that the light of a tenth-magnitude 

 star is unity. It will be seen that the stellar luminosities 

 form a curve very similar to the light-intensity curve of 

 the solar spectrum, but shifted towards the violet end of 

 that spectrum, possibly owing to the fact that the average 

 star is hotter than our sun. On this scale there is clearly 

 no place for the white stars, and the essential feature is 

 that stellar /nagnitude takes its place in a continuous and 

 definite relation to stellar colour. 



I had no intention of anticipating work not yet com- 

 pleted, but Mr. Hinks's contemptuous reference to our 

 omission of the white stars needed to be dealt with. 

 Their inclusion or exclusion makes no difference from the 

 standpoint of the statistical constant ; their exclusion is. 

 however, justified by the physical considerations which I 

 have here suggested. 



NO. 1981, VOL. 76] 



I should wish to say one word, albeit I am afraid it 



must be a strong one, about Mr. Hinks's further treat- 

 ment of Miss Gibson and myself. In the paper, to use 

 its own words, a suggestion, " even if it be only of the 

 vaguest kind," is made that the bulk of the lucid stars 

 may belong to a differentiated system. Mr. Hinks asserts 

 that the basis for " this far-reaching suggestion " is one 

 in which the white stars had no frequency in the record. 

 Will the reader believe that this suggestion, which the 

 writer herself describes as of the " vaguest kind," is not 

 based on the colour correlation at all? Can Mr. Hinks 

 really have criticised the memoir and supposed that even 

 this vague suggestion was based on the 159 Cape stars? 

 The suggestion, such as it is, is based on counts of all 

 stars, and results from showing that a continuous curve 

 can be found which describes with remarkable closeness 

 the counts up to the sixth and seventh magnitude, but 

 that beyond this magnitude any formula hitherto proposed 

 fails even approximately to describe the frequency. This 

 result, reached by other investigators, is confirmed by 

 Miss Gibson, and in association with changes in other 

 stellar chaiaclers, which occur about the same magni- 

 tude, does suggest, I venture to think, that in the vaguest 

 kind of way some differentiation of the stellar system may 

 possibly exist beyond the bulk of the lucid stars. I think 

 Mr. Hinks owes us an explanation of what his state- 

 ment, that a far-reaching suggestion has been based on 



stellar statistics from which all white stars have been 

 excluded, really is intended to convey. 



In conclusion, may I say that I have learnt from my 

 experience with biologists, craniologists, meteorologists, 

 and medical men (who now occasionally visit ' the 

 biometricians by night I) that the first introduction of 

 modern statistical methods into an old science by the lay- 

 man is met with characteristic scorn ; but I have lived 

 to see many of them tacitly adopting the very processes 

 they began by contemning. Mr. Hinks is at present in 

 the first stage ; but may I remind him that even astro- 

 nomy owes something to the layman, snd express my 

 hope that he may quickly reach a more understanding and 

 sympathetic frame of mind? Karl Pearson. 



Biomctric Laboratory, University College, London. 



The Body of Queen Tii. 



Judging from the letter addressed to Nature of 

 .September 26 (p 545), Mr. Hall (like Prof. Sayce in the 

 Times of September 17) has been thrown into a state of 

 doubt in regard to the real sex (? and age) of the mummy 

 supposed to be " Queen Tii " by a letter from Mr. Theo- 

 dore Davis-, calling in question the accuracy of my state- 

 ment that the mummy suopospd to be an old lady of at 

 least fifty years is the skeleton of a young man of about 

 half that age. 



