WCP4141

[1]

Parkstone, Dorset

December 1st 1893

My dear Galton,

Thanks for the papers which I return. The Committee Programme seems excellent, but I do not think it would be any use my joining it as I could never attend, though I am greatly interested in the proposed work. I am however, so poor a mathematician that I cannot follow the formulae & tables in their mathematical form. The only things I can clearly comprehend are the diagrams and curves showing variability in various ways. It seems to me (though it may be [2] quite wrong) that the mathematicians treatment of the subject does not bring out some of the most interesting points as regards evolution by nat. selection. For instance, what may be called irregular deviations from the mean are I think of great importance for nat.[ural] select.[tion]. The variations of some organs for instance will be something like this showing

[a sketch of a graph appears here]

that great numbers of individuals which ^{in some years or localities,} vary considerably both in excess & defect of the mean value. Now if by taking more individuals or in [3] other years the mean value were (as usual in other cases) considerably in excess — the resulting curve might come out more thus, and the fact of the

[a sketch of a graph appears here]

great amount of material occurring for modification by natural selection would be obscured. It seems to me important that the measurements of every individual should be represented by a dot in its proper place, leaving the whole mass of dots, even when several thousand to determine the curve of variation & the amount of divergence in each [4] direction. This would be understood by the most mathematical intellect! Another point is the correlative variation of parts in each individual. This is quite a different thing from the law of correlation determined by you. This ^Your law may show a relation between the means of several hundred individuals, while in many of these individuals the variations may be in opposite directions. While, for instance, the cubit may increase in definite proportion of stature — as a mean — there may be, & probably are, large numbers of individuals in every generation in which one is stationary while the other either decreases or increases — this furnishing materials for any combination of variations. I do not see that this is brought out in any of the mathematical results, but rather obscured.

Yours very truly | Alfred R Wallace [signature]