Natural selection describes how small variations in plants and animals lead to the remarkable fit between their traits and the environments they live in. It does not, however, always work. The limits to natural selection receive less attention than the potential it has to explain all manner of adaptation and diversity. I argue here that reflecting on these limits would lead to a stronger and more predictive theory.
Chiefly, there are three limits to natural selection:
- limited contribution to the origin of species;
- limited prevention of the extinction of species; and
- limited prediction of the results of selection.
The Origin of Species
Darwin’s intention in his 1859 classic was not to explain slight adaptive differences between varieties but to explain the origin of differences so large that it makes interbreeding impossible. This definition of a species – a group of plants or animals able to breed with one another – is still in use today; even though many of the groups we consider species (e.g. lions and tigers ) are able (but not likely) to interbreed. Darwin’s cousin, Francis Galton was one of the first to point out that natural selection could not produce a distinct species if individuals with new traits continued to breed with the larger population. The new traits would be swamped by the abundance of typical members of the population. Later, the persistence of rare traits was explained by combining Darwin’s ideas with Mendel‘s gene theory but modern theory still requires natural selection to operate on small, isolated groups to develop a new species.
It should be clear that reproductive isolation, whether it is physical (e.g. the flooding of lowland areas to create islands), behavioural (e.g. preference for a particular type of mate) or genetic (e.g. doubling of chromosome number), is the starting point for the origin of a species and indeed the most important part of the process. Natural selection can offer some trimming of the groups formed by isolation and allow them to be better adapted to the (slightly?) different environments in which they find themselves but it is not required. What is required is that isolation last long enough for random genetic differences to accumulate to the point that the two groups will no longer be able to breed together. The differences between closely related species would then be random differences in color and form and would not have any particular function. Indeed, this is what we generally see with closely related species.
In this view, reproductive isolation sets the pace for evolution and is the only necessary requirement. Differences between new species need not be adaptive, that is, correlated in some way with their environment. They only need to make the possibility of interbreeding between the two groups extremely remote.
The Extinction of Species
The great majority of species that have existed on earth are now extinct. This is a startling commentary on the success of natural selection in maintaining the fit between a group of breeding individuals and their environment. Not only are these species extinct but in most cases there is no evidence of new, slightly adjusted species to take their place, nor of a cataclysmic change in environment that would explain this limitation in natural selection. Mass extinctions aside, there is a background rate of extinction in the fossil record that points to the regular failure of natural selection to give plants and animals the traits that would have made the difference in their survival. Of course, no one ever said that natural selection is all powerful. There are limits, especially in multicelled plants and animals to which traits can appear in a population. Big changes in structure will mostly not survive the complicated process of developing from a fertilized egg. The background extinction rate gives us a baseline of how often natural selection fails. I suggest that those periods of evolutionary history with lower background rates represent times and circumstances where natural selection is more vigorous and relevant.
My biggest concern about natural selection as a scientific theory is the lack of any independent measure of fitness. The idea is an old one and was observed well before Darwin connected it to having more babies. William Paley published Natural Theology in 1802 and attributed the wonderful suitability of each plant and animal to its environment as a proof of the existence of God. Yet today we assume natural selection has been in operation and automatically associate fitness with long life and many offspring. Without an independent idea of the correlation between traits and environment and how far any given individual strays from that perfect fit, natural selection is incomplete. Newton gave us an equation for the source of gravity as well as a set of equations for predicting the consequences for bodies subject to gravity. Without a comparable source law for fitness, natural selection is doomed to be something we use to explain rather than something we use to predict.
It could be argued that we will never know enough about a species’ biology and environment to predict ahead of time which individuals will succeed. If that is so, despite best efforts, I would argue that either:
a) fitness is an illusory and ill-defined concept, or
b) natural selection is a weak or infrequently occurring phenomenon.
In my next blog I will offer my own attempt to develop such a source law for fitness.