Darwin’s theory was descriptive and qualitative. Sixty years after Darwin, A British mathematician named Ronald A. Fisher proposed the first fully quantitative model of how evolution might work.
In the 18th and 19th Centuries, the science of physics had advanced through reductionism: dividing systems into their constituent parts, and demonstrating that each part obeyed simple, universal mathematical laws. Fisher wished to do for biology what Newton and Maxwell had done for physics. Physics built everything from simple point particles. Fisher asked himself, what are the atoms of evolution – the irreducible elements from which he might construct a theory.
In populations of bacteria, Fisher might have taken the individual as an atom of inheritance. Individual bacteria make exact copies of themselves, and the number of copies might be taken as a a measure of the individual’s success in evolutionary competition. But in sexual populations, no two individuals are alike. What is the “atom” of evolution for a sexual species?
In search of the “Atom” of evolution
Genetics was yet a new science in 1920, as Gregor Mendel’s breeding experiments in the mid-19th Century had just been rediscovered, and their importance for evolutionary science was first appreciated. In 1909, Danish botanist Wilhelm Johannsen had coined the word gene to describe the theoretical unit of inheritance. (Of course, it would be another 20 years before Erwin Schroedinger connected the gene conceptually to a length of DNA on a chromosome.) Fisher took the gene to be the atom of evolutionary biology. His theory would be about genes.
In Fisher’s model, there is a single species with a fixed population. (In other words, exactly two offspring survive from each mating pair, so that the death rate exactly matches the birth rate, not just in the long run, but exactly in every generation.) Genes circulate freely among different individuals, as they are mixed and swapped and traded during sexual reproduction. Different genes do not interact with one another, but each one contributes to making the individual better able to survive and reproduce, or not. Those that make a positive contribution to survival and reproduction increase in gene frequency from one generation to the next. This just means that there will be more individuals carrying that gene in the next generation than there are in the present generation, because the gene has been successful in enhancing either survival or reproduction. This success, in Fisher’s model, is a property of the gene, and he identified the gene’s success in spreading through the population with what Darwin called fitness.
(Fitness is a property of the gene, not the indivdual animal or the population of animals or the species. What might Darwin have thought of this? Darwin’s was a very different scientific culture, built on a firm foundation of observation in the natural world, sparing of theoretical claims. His “theory” would not be called a theory today, because he never defined terms rigorously, or specified the mechanistic details of natural selection.)
Fisher successfully argues that his simple model is an accurate reflection of reality
From the beginning, there were questions about whether genes could be considered as free agents, with individual interests of their own. A gene only makes sense within a given environment. That environment includes other genes, other individuals, and other species. Systems in biology are just too interdependent to admit the kind of reductionism that worked for physics, argued Sewall Wright, who was Fisher’s intellectual sparring partner in the 1930s and 40s. Fisher was a brilliant mind and a commanding presence. He won the debates.
The contentious assumptions on which Fisher’s model was based were
- No interaction between different genes (= no epistasis)
- No interaction among indiivdual animals or plants of the same species
- No change in the environment, or in other species while the gene is propagating
- Genes spread through the population and are mixed at random, so combinations of genes that work particularly well together are no more likely to occur than combinations that work at cross purposes.
Fisher (who was trained in mathematics and not biology) was not so naïve as to think that these assumptions were literally true. But in defending his model through the following decades, he argued that the world is a big place, and that in the long run you can focus on what matters most, and everything else will average out. It is the genes that matter most, because it is the genes that are evolving.
The Dilemma of Altruism
Fisher never heard of the “selfish gene”. The term was coined by Richard Dawkins a few years after Fisher’s death in 1962. But selfish genes were exactly the subject of his model, and Fisher’s theory can fairly be called a “selfish gene” theory. Biologists looking at the real world see a lot of cooperation. Animals make sacrifices for the sake of their children all the time. Almost as common is the tendency of animals to nest together, to hunt together, to cooperate in groups that may or may not share the same genes. Often an individual will sacrifice its own fitness for the benefit of the community, and the community, in turn, makes possible a much more secure environment in which the individual may flourish. (Here’s a striking example from amoebas – which may not be the first image to come to mind when you think of the English word altruism, but which is a classic of evolutionary altruism.) But what prevents individuals from “cheating”? Why don’t they evolve to mooch off the community, taking all the benefits while shirking the sacrifice and communal responsibility. A gene that promoted this behavior would be successful, because its host would receive all the benefits of the community, and in addition would save energy by not contributing to the communal welfare, and so it would have extra energy to channel into its own success in propagating into the next generation. Why don’t cheaters evolve to undermine cooperative communities? Why do we see any cooperation at all in nature? Fisher’s model didn’t readily account for this, and evolutionary biologists looked for a way to expand his theory so as to explain cooperation.
Kin Selection and the Legacy of W.D. Hamilton
In the early 1960s, William Hamilton was a promising young grad student at the London School of Economics when he conceived an answer to this dilemma, which was quickly recognized as a seminal contribution to the field, and to this day remains the only mechanism that is universally accepted by the many scientists who work within the Fisher model. Hamilton’s idea: a copy of a gene acts within a particular individual, but brothers and children and cousins of that individual have a certain probability of carrying an identical copy of that same gene, because of their relatedness. It is the gene that is selfish after all, and not the individual. Genes will evolve to promote copies of themselves that exist in other individuals, even at the expense of the individual carrying that gene. Hamilton’s rule is a simple mathematical expression of the balance between effects that a gene’s action has on the self, and on relatives that might carry copies of the same gene. A quip attributed to Hamilton references the math: “I would lay down my life for two brothers or eight cousins.” What this means is that if I have a particular gene, then my brother has a 50% probability of having the identical gene, because each parent contributes half his or her genes to each child, randomly selected. Similarly, the probability that my first cousin shares a copy of this gene (which impels me to risk my life for a relative) is ⅛.
During the 1960s, the community of evolutionary scientists was assimilating the mathematical framework that Fisher had developed several decades earlier, and grappling with the question how cooperation can arise in a process of natural selection. Conceptually, there were two big ideas that became competing models. One was group selection, the idea that natural selection pits not just individual-vs-individual but group-vs-group. A gene for altruism may be defined as one that depresses the fitness of the individual that carries it, but that promotes fitness of the trait group within which the individual cooperates. If evolution operates on the level of groups, that solves the dilemma of altruism. Groups that have more copies of the altruism gene simply out-compete groups that have fewer copies. This is a Darwinian process analogous to the one that Fisher modeled, but operating at a higher level of selection. The second alternative was that Hamilton’s Rule explains all. The conservative assumption is that only closely-related groups can maintain their cohesiveness and avoid evolving “cheaters”. Group selection is an unnecessary complication, a fifth wheel of evolutionary theory. George Williams and John Maynard Smith, two bright young leaders in the fast-developing field of mathematical biology, argued that group selection was not only unnecessary but improbable as well. Look at ecosystems in the real world, they said. Individual deaths are common, but group exctinctions are rare. Gradual changes in gene frequency are common. Group selection requires one group to drive another out of business, and this doesn’t seem to happen very often.
This is the argument that carried the day, and among evolutionary scientists the idea of group selection came to bear a stigma. Over time, skepticism of group selection grew to be a prejudice that has outlived the memory of where it comes from and on what grounds it is based. It continues today as a scientific bias.
Are there many examples in nature of behaviors that Hamilton’s Rule can’t explain?
This post originally appeared here: http://joshmitteldorf.scienceblog.com/2013/11/18/the-selfish-gene-vs-multi-level-selection-aging-doesnt-fit/