How Hamilton mathed up senescence

2 minute read

This is one of the most beautiful openings to a paper, ever:

Consider four hypothetical genes in man. Suppose all are limited in their expression to the female sex and also age-limited in the following way: each gives complete immunity against some lethal disease but only for one particular year of life. Suppose the first gives immunity for the first year, the second for the fifteenth, the third for the thirtieth, and the fourth for the forty-fifth. What are the relative selective advantages of these genes?
If for further simplicity parental care is ignored and it is assumed that the menopause always comes before age 45, it is at once obvious that the fourth gene is null, whereas all the others do confer some advantage. It is also fairly obvious that the third gives less than the second. But how much less? Does the second give a maximum becasue it occurs at the age of puberty? Does the first give less than the second?
The importance of questions of this kind for an evolutionary theory of senescence has been realized for some time. Most of the answers that will be given in this paper agree with the theory of Williams (1957). Although perhaps not obvious, they are so simple that it is surprising to find almost no indication that they had been realized earlier. Several writers have in effect answered the last two questions in the affirmative, which is for the one inexact and the other wrong.

If anybody in biology ever wrote like Nero Wolfe talks, it was Bill Hamilton. That intro is from "The Moulding of Senescence by Natural Selection," (Hamilton 1966:12-13), which ought to be required reading, if it isn't.

Later in the paper Hamilton presents almost fully-formed his idea of "sibling replacement" to explain why subadult mortality should be concentrated in infants:

Suppose that the catching of a disease in the immature period is ineveitable and that hte first infection has only two possible outcomes: death or survival with perfect immunity against second infection; and suppose that the probabilities do not change according to the age of the first infection. Then if there is any degree of sibling replacement at all, a gene bringing forward the expected age of first infection will be selected, for it can easily be seen that the more commonly the gene is appearing in a progeny the larger its expected completed size will be, while at the same time the expected frequency of the gene within the progeny is unchanged.
If the bringing forward of the susceptibility involves a disadvantage, for instance by a slight increase in the chance that death is the outcome of infection, the situation is more complex and will require mathematical analysis to delimit the possibilities. This is because it is then no longer true that the proportion of the gene is unchanged by the amount of replacement that goes on; positive selection could fail even if it was guaranteed that the amount of replacement could more than compensate for the extra mortality. This is in effect a problem of a more "altruistic" versus a more "selfish" trait (ibid:39).

And so the theoretical power of inclusive fitness is made plain: it encompasses even death itself in the service of first-order relatives.