Although people know better, we see papers or discussions daily that effectively refer to findings of, or searches for the gene(s) 'for' some particular trait. The gene 'for' obesity, or autism or aggression. Or the gene responsible for the evolution of the trait. The gene 'for' what makes us human. When the trait is reasonably simple, and a genetic variant's effects in some population is fairly strong, this makes literal sense. From a standard Darwin evolutionary perspective, or a standard developmental perspective the ideas also seem to make sense. There are in fact traits and diseases that are basically attributable to single genes, as Mendel showed with his work on peas in the 19th century, and many have found thereafter. But this kind of thinking can be misleading and often -- or maybe usually -- is.
Many, many areas of the genome contribute to a trait. In a population or species, most if not all regions of the genome are subject to mutation and usually vary at least in some individuals. Some are 'genes' in the usual sense of coding for a protein, but other genome areas are inv0lved in different kinds of function, such as controlling the expression of protein-coding genes.
With lots of variation in lots of factors, and the complicated patterns of cells in an embryo or individuals in a species, it's likely that many, rather than one or even just a few, of these varying elements contribute to variation in an embryo or variation in a species over evolutionary time. A single gene many contribute to a trait, but they are often misleadingly thought of as being 'for' it, as if they play a special role.
This becomes all the more important when it is considered that most elements in the genome, and in particular most protein-coding regions ('genes' in the usual sense of the word), have many different uses even within the same species -- even within the same organ or system at any given time and/or during its development. So is a gene expressed in teeth, limbs, and gut a gene 'for' teeth? And if the homologous gene in flies -- that 'tooth gene' -- is expressed in (say) limbs and brains in vertebrates, is it now a limb gene? a brain gene?
Gene-for thinking is often decried by authors of reviews and overviews (including ourselves). But it's a hard habit to break, and few of us can claim to have made that break, that conceptual break, as cleanly as we ought (though Ken was just at a meeting with a very prominent microbiologist who said he hasn't used the word gene when talking about genetics in class in 20 years -- he sets the bar high). There are many ways to consider a genetic element that contributes to many different things in a given individual and/or in different species. Altering the gene may affect multiple traits, and its effects may be more important in some individuals or species than in others. Likewise, it may change in evolution because of selective effects on only one of the traits in which it's used. Disentangling these issues is not easy. But it's clear that we need to, and that linear thinking of DNA as beads on a strings is a limited way of thinking -- in many ways, we're prisoners of Mendel. Yet, thinking in the many higher dimensions of development and evolution is by no means easy, though it provides exciting prospects for those who enjoy the challenges of biology -- on all its time scales.