On the Conditions for Natural Selection

Olivia Judson's piece in today's New York Times makes the important (and chastening) point that too many contemporary biologists have not actually read Darwin in the original.  As usual, she is terrific and generally right on.  I do have one small issue with a comment she makes:

Natural selection will operate whenever all of three conditions are met. These are: (1) some of the differences between individuals are inherited differences, not due to differences in their environments; (2) more individuals are born than can survive; and (3) part of the reason at least some of the survivors make it is owing to the traits — a longer-than-average beak, say — that they inherited from their parents. For natural selection, then, what is important is that some differences are inherited; and this, Darwin could show.

The problem with this statement is that it's not actually correct.  But it's incorrect in an interesting way, given Judson's point that there is value in understanding the history of our science.  In the Origin, Darwin was clearly influenced by another foundational document, Thomas Mathus's Essay on the Principle of Population. Indeed, the Malthusian insight that populations, human or otherwise, can easily outstrip their resource base proved to be a critical link in Darwin's formulation of the idea of natural selection.

The key about Judson's statement is that point (2) is not necessary.  If it were true, it would imply that natural selection can not act in growing populations.  The necessary and sufficient conditions for natural selection acting on a trait are: (i) variation in the trait, (ii) heritability of the trait, and (iii) differential reproductive success as a function of the trait. So a trait can increase even in the absence of differential mortality as long as there is differential fertility.  Darwinian fitness is a relative concept. A trait with higher fitness is one that increases in frequency relative to other traits.  This increase in relative frequency can come about because of differential mortality or differential fertility or a combination of the two.  

It turns out that models of selection are more mathematically tractable when one assumes only viability selection and so this is what is most commonly employed in theoretical work.  This relates to the fact that fertility selection works at the level of the breeding pair and not the individual.  The equations describing selection in this context thus become rather more complex.  But this is more technical than most sane people would care to get.  A good review of models of fertility selection can be found in Feldman et al. (1983).

Judson is essentially correct though.  We should all have a better understanding of the history of Evolutionary Biology (and, I should add, Anthropology).  A major part of this is reading the classics: Darwin, Malthus, Wallace, Fisher, Dobzhansky, Wright, Lorenz, Tibergen, Hinde... I could go on for a while, of course.  When I teach my class in life history theory, almost all of the readings are (20th century) classics.  My experience in the field is that lots of people cite Hamilton (1966) or Cole (1954), for instance, but very few people have actually read them.


Cole, L. C. 1954. The Population Consequences of Life History Phenomena. Quarterly Review of Biology 29 (2):103-137.

Feldman, M. W., F. B. Christiansen, and U. Liberman. 1983. On Some Models of Fertility Selection. Genetics 105 (4):1003-1010.

Hamilton, W. D. 1966. The Moulding of Senescence by Natural Selection. Journal of Theoretical Biology 12:12-45.

Leave a Reply

Your email address will not be published. Required fields are marked *