No Virginia, evolution isn't ending

Updated: Follow up post End Update

I've already the covered Steven-Jones-evolution-is-ending story at my other weblog. I notice that John Wilkins has also objected to Jones' exaggerations. When I initially read the quotes from Jones in The Times I was alarmed, but wondered if his position was being taken out of context or misinterpreted. I emailed a prominent evolutionary biologist who I suspected would know Jones well enough to clarify this issue. My correspondent responded that Jones really does believe this, and he finds Jones' ideas as ludicrous as I do (adding for good measure he doesn't get the sense that Jones has a really good grasp of population genetics).

Since the quotations in The Times were rather spare, and the comments somewhat inchoate I didn't really air many of my criticisms in my other post. So below are some extended thoughts....

First:

Professor Jones added: "In the old days, you would find one powerful man having hundreds of children." He cites the fecund Moulay Ismail of Morocco, who died in the 18th century, and is reputed to have fathered 888 children. To achieve this feat, Ismail is thought to have copulated with an average of about 1.2 women a day over 60 years.

One anecdote does not an argument make. The genetic and anthropological data suggest that polygyny is somewhat common in our species, but Moulay Ismail is a once a century phenomenon.

Jones' point here is that in the past old men with many mutations produced many offspring. Today, young men with fewer mutations produce fewer offspring. All things controlled the power of natural selection to drive evolutionary change is proportional to extant genetic variation. That variation is ultimately driven by new mutations injected into the population every generation. As I noted above, it really isn't established that old men had so many more children in the past. Secondly, one could imagine a scenario where most potentially beneficial mutations are in "excess." Consider a situation where you have a phenotype X, and there are 10 beneficial mutations on that phenotype per generation within a population. If one of them is picked up by positive selection, then the fitness of the others may decrease because fitness is measured against the genetic background. In other words, the response to selection might be rather the same for a population with 5 vs. 10 mutations. Of course, you need large populations to have so many good mutations produced per generation, but as Jones observes we do have large populations now.

Next:

Another factor is the weakening of natural selection. "In ancient times half our children would have died by the age of 20. Now, in the Western world, 98 per cent of them are surviving to 21."

First, the high mortality rates that Jones alludes to makes me wonder why he's so confident about the ubiquity of fecund gerontocrats. But in any case, evolution does not occur unless there is reproduction. For evolution driven by natural selection to occur there needs to be a correlation between genetic variation and a trait and fitness due do variation on that trait. In other words, death is sufficient to reduce fitness, but it is not necessary. You might notice that many Westerners are now childless. Why? It may be that personality traits which are genetically controlled correlate to a great extent with reproductive variance; then evolution will occur.

To make this explicit, imagine a society with a fertility below replacement. But, within this population there is variance. Some individuals for whatever biobehavioral reason prefer to have many children as possible as opposed to investing their income in consumption. Over a few generations the personality types which result in the production of more children will spread at the expense of those which do not result in reproduction. That is evolution.

Additionally, the fixation on death seems overdone. Consider the wave of advance across the American West during the 18th and 19th centuries. Among siblings some stayed put, while others moved West. Because of surplus land one might posit that those who left for the frontier had more offspring. As the frontier hit the Malthusian limit there might be more migration West, and again selection for "bold" personality types. And so on. Allele frequencies would change even though the population was rapidly expanding because there was a bias in the expansion (also, Mark Ridley would make the case that there's plenty of mortality, it is just pre-natal. Human spontaneous abortion rates are very high).

The last part is really hard to unpack:

Decreasing randomness is another contributing factor. "Humans are 10,000 times more common than we should be, according to the rules of the animal kingdom, and we have agriculture to thank for that. Without farming, the world population would probably have reached half a million by now - about the size of the population of Glasgow.

"Small populations which are isolated can evolve at random as genes are accidentally lost. World-wide, all populations are becoming connected and the opportunity for random change is dwindling. History is made in bed, but nowadays the beds are getting closer together. We are mixing into a glo-bal mass, and the future is brown."

It seems that here Jones is alluding to the shifting balance theory. Basically the model is that adaptive evolution will occur more rapidly when there is population substructure because then the full adaptive landscape can be explored. Random genetic drift can shift populations across fitness "valleys" before ascending up to a new peak. In a large population stochastic forces are much weaker vis-a-vis selection so that drift could never drag a population across a fitness valley.

There are many problems with coherency and clarity in regards to the shifting balance model. It is an interesting and perhaps illuminating model as a heuristic, but it has never truly been well fleshed out. Sewall Wright, the originator of the idea, wasn't always clear about what he meant, and what the adaptive landscape was. So if Jones is alluding to the shifting balance to argue for the end of evolution, it is a weak leg to stand on. It is an interesting idea, but it is certainly not mainstream consensus theory.

But perhaps Jones doesn't really have the shifting balance in mind. If that is so, I have to point to his contention that "Small populations which are isolated can evolve at random as genes are accidentally lost." Yes, this is true. But this is also true of large populations according to neutral theory. In fact, one of the more counter-intuitive findings at first blush of neutral theory is that the rate of substitution simply equals the rate of mutation. Substitution is the process where at locus A allele 1 is replaced by allele 2 as the fixed variant in the population (between which there is a transient state where polymorphism exists). The reason for this is simple: drift is more powerful at churning new alleles in small populations, but large populations have many more new alleles! In other words, if the probability of fixation of a new mutant in a population of 100 is 1 out of 100, and in a population of 10000 is 1 out of 1000, there are likely to be an order of magnitude more mutants in the latter population than the former. So they cancel out.

Neutral theory is just a place to start. It isn't the be-all-and-end-all, as some people would say. But the last sentence really jumps out at me and suggests that Jones hasn't thought the whole issue through, "We are mixing into a glo-bal mass, and the future is brown." It is true that there is a great deal of variation in skin color, and that admixture leads to a average outcome which can be modeled roughly in an additive fashion. That is, black + white → brown. But, there is a large variance around this expectation; so homogeneity is not the future. Skin color variation is controlled by only a few discrete genes. The phenomenon of black and white fraternal twins shows that the expected variance is large enough that the extremes of the tail can show up with enough "trials." But all you have to do is talk to Brazilians to comprehend that admixture does not lead to elimination of variation (if the logic extended to hair color people in European villages should all have the same hair color because they are intermarried with each other).

The last is somewhat at a remove from the bigger picture of Jones' assertions. But it goes to the point that for a "leading geneticist" he hasn't thought about the population genetics in great deal.

I could really go on. For example, large population sizes were critical in the story of adaptive acceleration. But you don't need to hinge your argument on just this model. An interconnected demographic model might also produce many more gene-gene interaction effects. Additionally, there is more extant genetic variation which is a normal part of the population distribution from which to select for. I'm sure readers can offer up other rejoinders, so I'll leave it at that.