Sta/istirnl lnra<tiii(iti<nia 3fl3 



groups suffice to indicate that the model must have been made about the 

 time of his piiner of 1 875. It is, I think, sufficient evidence that fialton denlt 

 with tlu! corrt'lution of ranks Ix-fore h«M'vpn rniched til' ' ' '" ' '^. 



iind the cluini that it is a contrihution of tho psyci i 



forty years later to the conception of correlation does not seem to me vahd. 

 (ijilton, we may with high prohalnlity suggest, had satisfied himst-lf t' *' 

 correlation of ranks was more c»nnlx^rsom»> than the correlation of 

 because in the simplest case, that of the normal distribution, it fails to provide 

 linear regression, but gives a non-integrable curve, which can only be plotted 

 by aid of the prol)al)i]ity integral table. 



Galton undoubtedly tirst attacked the problem of correlation from the 

 standpoint of ranks, and it natin'ally did not lead him in any simple expres-^iun 

 for the relation U'tween relatives. In his .]fetnories (p. ;J00) ho tells us iiow : 



"As these lines are l>eing written, the ciroutnstancttH under which I first clearly grasped the 

 importAnt gencralisntion that the laws of herodity were solely concerned with (levin* '  

 expressed in statistical units [for Galton the quartile values] arc vividly recalled to my nui 

 It was in the grounds of Naworth Cjustle', where an invitation had l)een j;iven to ranihle i 

 A ten)]>orary shower drove nie to seek i-ufugo in a retldish recess in tho rock hy the side • 

 pathway. There the idea ilashml across mo, and I forgot everything else for a moment in uiy 

 great deliglit." 



That 'recess' deserves a commemomtive tablet as the birthplace of the true 

 conception of correlation! 



CJalton continued his campaign in favour of marks for physical efficiency. 

 He published a paper in the Nineteenth Century, Feb. 1889 (pp. 303-8), 

 entith'd, "The Sacrifice of Education." The title does not seem well chosen, 

 and 1 expect it was an editorial choice; Galton himself has written "Testa 

 of Physical Capacity" above the paper and so indexed it. The object of the 

 pa]>er is to show the ea-se with which certain anthropometric measurements 

 can l)e made, and the cost of making them. With most of the fiicts .stilted in 

 the paper the reader will be already fiuniliar. The tests are described, the time 

 taken in making them, and the cost of running an anthropometric lalniratory. 



"The problem is to give marks for physical qualifications just a-s they are now given for 

 intellectual ones, in order to pas-s those candidates who being a little under par intellectually 

 are far above par botlily ; conversely to weetl out those other candidates who, not being [larticu- 

 larly tit in respect to their brains, are at the same time of decidedly inferior physi(jue. The 

 relative weight to be lussigned for intellectual and bodily excellences is a question of detail, most 

 important no doubt, but one that need not be discussed here." 



Galton states the various criticisms of the proposal that have been made, 

 but points out that they apply not only to marks for physical capacity, but 

 to all exaiuinatioits what.soever. We can never test all the faculties even of 

 the mind, and again we can only test the examinee on the special occasion on 

 which he presents himself It is admitted that the examination of any facidty, 

 physical or mental, is a difficult art, and one not to be ])erfected ofi'hand. 



' Naworth Castle is in Cunilwrland, north-east of Carlisle. The Oaltons went north on 

 Aug. 12 and lodged at Wetheral in the neighlx>urhoo<l of Naworth; later, on Sept. 10, they 

 travelled to Newciustle and .stayed with the J^pence Watsons at Bensham Grove during the 

 B.A. meeting at which Francis (lalton read his pajier on marks for physical etfioiency. "Frank 

 has l)eeu busy all the year on problems connected with correlation but has publislu>d nothing yet 

 about them." (L. G's Jiecord under 1889.) 



V ci n 60 



