28 
Notes on the History of Co?Telation 
(2) The next important work to be considered is that of August Bravais. It 
is entitled " Sur les probabilites des erreurs de situation d'un point." It was 
published in the Memoires presentes par divers savants d V Academie royale des 
Sciences de I'lnstitut de France, T. ix. Pai'is, 1846, pp. 256—332. It appears, how- 
ever, to have been reported favourabl}' upon in 1838*. Bravais was in many 
respects a remarkable man. Essentially a geologist he wrote also on astronomy, 
physics, meteorology and the theory of probabilities. He made a voyage to 
Lapland for geodesic purposes and took the opportunity of measuring a number of 
Lapp skulls ! He had a width of action most sympathetic to the biometrician. 
Writing in 1895 of the history of correlation I said : 
" The fundamental theorems of correlation were for the first time and almost 
exhaustively discussed by Bravais [Title as above of his memoir] nearly half a century 
ago. He deals completely with the correlation of two and three variables." 
Then speaking of Galton's coefficient of correlation I say: "This indeed appears 
in Bi'avais' woi-k, but a single symbol is not used for it. It will be found of great 
value in the present discussion. In 1892 Professor Edgeworth, also unconscious of 
Bravais' memoir, dealt in a paper on ' Correlated Averages ' with correlation for 
three variables {Pliil. Mag. Vol. xxxiv. 1892, pp. 194 — 204). He obtained results 
identical with Bravais', although expressed in terms of ' Galton's functions 
[i.e. coefficients of correlation]. 
Again later, p. 287, in giving the fundamental equation for the correlation of 
three variates I wrote : " This agrees with Bravais' result, except that he writes for 
''i, ^'a, I's the values ^{ijz)l{na.,a:^ etc., which we have shown to be the best values 
(see loc. cit. p. 267)." Again on p. 301 I write before proving the general theorem 
of multiple correlation : " Edgeworth's Tlieorem. We may stay for a moment over 
the results above to deduce Professor Edgeworth's Theorem," with the footnote, 
" Briefly stated with some rather disturbing printer's errors in the ' Phil. Mag.' 
Vol. XXXIV. p. 201, 1892." 
Now all these statements if they were correct would indicate that Bravais dis- 
covered correlation before Galton and that Edgeworth first published the form of 
the multiple correlati(jn surface. They have been accepted by later writers, notably 
Mr Yule in his manual of statistics, who writes (p. 188): 
" Bravais introduced the product-sum, but not a single symbol for a coefficient 
of correlation. Sir Francis Galton developed the practical method, determining his 
coefficient (Galton's function as it was termed at first) graphically. Edgeworth 
developed the theoretical side further and Pearson introduced the product-sum 
formula." 
Now I regret to say that nearly the whole of the above statements are hope- 
lessly incori-ect. Bravais has no claim, whatever, to supplant Francis Galton as the 
discoverer of the correlational calculus. For the most part he is simply taking 
a very special case of the Gaussian analysis, and nowhere on p. 267 of his memoir 
can I now find that he has used the expressions for the correlation symbols without 
* Cdutptes i-okIuh, T. vii. p. 77. 
