There is considerable evidence to suggest that Dawkins' example is an instance of
self-delusion. It may be adaptive for women to
appear more "
faithful" than men, but what is chiefly adaptive is to find at least one man (and that usually means
only one man) who is willing to carry some of the burden of
protecting,
feeding and
sheltering children.
Evidence from closely related primates has shown that a fairly high percentage of offspring among species that tend to practice pair bonding behavior similar to that seen in humans are not the biological offspring of the pair-bonded male.
Perhaps the male derives other benefits from pair-bonding, aside from simple reproductive success? Perhaps being a pair-bonded male with offspring signals to a larger number of females little more than that this is a reproductively competent male (not only that, but they can look at the children and decide whether the traits they suspect he has passed on to them are traits they desire in some of their own offspring?)
Reproductive success is almost incontestably a significant factor in shaping human behavior, at least in the aggregate. But accepting this notion doesn't necessarily have to lead to the first or most naive set of conclusions.