Robert Plomin of King's College London has been plugging away at the genetics of IQ for decades. It's been frustrating, but now he thinks he's getting somewhere. The London Sunday Times reports:

Scientists have identified more than 200 genes potentially associated with academic performance in schoolchildren. Those schoolchildren possessing the "right" combinations achieved significantly better results in numeracy, literacy and science.

The finding emerged from a study of more than 4000 British children to pinpoint the genes and genetic combinations that influence reasoning skills and general intelligence.

One of its main conclusions is that intelligence is controlled by a network of thousands of genes with each making just a small contribution to overall intelligence, rather than the handful of powerful genes that scientists once predicted.

... There are potentially many millions of these variations, but the team restricted their search to looking at the million or so of the most common, to find out which gene variants were most frequently found in children with either a high or low level of academic achievement.

"Out of the gene variants we looked at, a couple of hundred are emerging which seem to have a small but significant relationship with ability in maths and English," said Professor Plomin.

... Research into height, for example, has picked out 300 genes that affect how tall people will grow, but even these genes can only explain 15 per cent of the total variations in human height. It implies that hundreds more genes must also play a part.

John Hawks points out a recent Nicholas Wade article in the NYT making a similar point the lessons of an experiment on fruit flies to breed for earlier hatching:

One well-known path to change is a heavily favorable mutation in a single gene. But it may be well known only because it is easy to study. Another path is exploitation of mildly favorable differences that already exist in many genes.

The question has considerable practical importance because if complex traits, including susceptibility to disease, are influenced by just a few genes, then it should be easy to develop treatments that target the few genes’ products. But if tens or hundreds of genes are involved in each trait, the task may be close to impossible....

The conventional view is that evolutionary change is generally mediated by a favorable mutation in a gene that then washes through the whole population, a process called a hard sweep because all other versions of the gene are brushed away. The alternative, called a soft sweep, is that many genes influence a trait, in this case the rate of maturation, and that the growth-accelerating versions of each of these genes become just a little more common. Each fly has a greater chance of inheriting these growth-promoting versions and so will mature faster.

In sequencing their subjects’ genomes, the researchers found that a soft sweep was indeed responsible for the earlier hatching. No single gene had swept through the population to effect the change; rather, the alternative versions of a large number of genes had become slightly more common.

The debate about whether evolution proceeds by altering one or many genes started 90 years ago among the three founders of population genetics, Ronald Fisher, Sewall Wright and J. B. S. Haldane. Haldane favored the single mutation mechanism, but Fisher and Wright backed multiple gene change. The fruit fly experiment ”is a total vindication of Wright and Fisher and a major defeat for Haldane and a lot of conventional geneticists who have sided with him,” Dr. Rose said.

The demise of the Haldane view ”is very bad news for the pharmaceutical industry in general,” Dr. Rose said. If disease and other traits are controlled by many genes, it will be hard to find effective drugs; a single target would have been much simpler.

So, it's not surprising that intelligence is dependent upon a lot of genes. That's generally true for a lot of complicated traits.