After receiving your letter this day a month ago, it occurred to me that I had better write an article in 'Nature' on Panmixia, pointing out the resemblances and the differences between Weismann's statement of the principle and mine. Shortly after sending it in, Weismann's answer to Vines appeared, and from this it seems that he has modified his views upon the subject. For while in his essays he says that 'the complete disappearance of a rudimentary organ can only take place by the operation of natural selection' (i.e. reversal of selection through economy, &c.), in 'Nature' he says, 'Organs no longer in use become rudimentary, and must finally disappear, solely by Panmixia.' Thus, the same facts are attributed at one time 'only' to the presence of selection, and at another time 'solely' to its absence.
Now, the latter view seems exactly the same as mine, if it means (as I suppose it must) that the cessation of selection ultimately leads to a failure of heredity. (How about stability of germ-plasm here?) The time during which the force of heredity will persist, when thus merely left to itself, will vary with the original strength of this force, which, in turn, will presumably vary with the length of time that the organ has previously been inherited. Thus, differences of merely specific value (to which you allude in your letter) will quickly disappear under cessation of selection, while 'vestiges' of class value are long-enduring. The point to be clear about is that the cessation of selection (in my view) entails two consequences, which are quite distinct. First, a comparatively small amount of reduction due to promiscuous variability round an average which, however, will be a continuously sinking average if the cessation is assisted by a reversal of selection; and second, later on, a failure of the form of heredity itself.
Touching the first of the two consequences you say that 'variations below or away from the standard would not be balanced by those above, because the standard was reached by the selection of such an extremely minute fraction of all variations which occurred.' But can variations in the matter of increase or decrease take place in more than two directions, up or down, smaller and larger, better or worse? (Read Wallace, 'Darwinism,' pp. 143-4.)
I write this in view of the lecture you say you are going to give, because I do not know when 'Nature' will bring out my article.
March 20, 1890.
It might perhaps be well for you to read the type-written reply which I have prepared to Wallace's criticism on 'physiological selection.' But this is for you to consider. He has fallen into some errors of great carelessness, not only with regard to my paper, but also to that of Mr. Gulick, whose theory of 'segregate fecundity' is the same as mine. On this account I am able to upset the whole criticism, and, bottom upwards, to show that it really supports the theory.
I see 'Nature' of this week contains my letter on Panmixia, and hope it will define in your and other minds the outs and ins of the matter.
Please return the enclosed, which I send as a fact that may interest you.
To Professor J. C. Ewart.