Brian Josephson (Letters, 12 January) misunderstands Richard Dawkins. The selfish gene theory doesn't require "a direct correspondence between specific behaviours and specific genes". A gene's direct effect is to synthesise a protein, which indirectly affects physiology and behaviour. A specific behaviour pattern can be statistically more common with a particular gene than without it. This is all that is required for that gene to be selected; if the behaviour aids the survival and/or reproduction of that gene, it becomes more numerous in the gene pool.
The term "a gene for" is often used, which is not genetic determinism but shorthand for "a gene making [a specific behaviour] more frequent than it would be without the gene".
In fact, such "genes for [a specific behaviour]", aren't usually real, known genes, but hypothetical genes, used to show how any DNA "causing" the certain behaviour would spread throughout the gene pool.