Among the "just-so stories" popular with scientists who seek genetic explanations of human behavior, few are so odious as the idea that males are genetically predisposed to kill their stepchildren.

The idea is that such behavior would have been adaptive for our Stone Age ancestors. Males who carried genes pushing them to kill their stepchildren, goes the theory, would have left more kids themselves because the murder would have freed up their new mate for, well, mating (a female nursing a small child is much less likely to conceive). Also, a Stone Age man who cared for and supported only his biological children, rather than stepkids, would leave more descendants than a man who cared for his stepchildren. Such murderous males would have left more descendants than males who tolerated, let alone supported, their stepchildren; we, their descendants, would therefore also carry the stepchildren-killing gene.

This bleak view of human nature has become widely accepted, even to the point of excusing stepfathers who abuse their stepchildren. ("My genes made me do it.") One textbook states that kids living with a parent and a stepparent are about 40 times as likely to be abused as those living with both biological parents. But as David Buller's brilliant 2005 book "Adapting Minds" showed, the data refute that. Why? Because reports of abused kids don't always say who the abuser was. Some children are abused by their biological mother, so to ascribe all stepchild abuse to the stepdad is specious. Also, a child's injuries are more likely to be called abuse when a stepfather is in the home, Buller finds, but to be called accidental when a biological father is. "There is no substantial difference between the rates of severe violence committed by genetic parents and by stepparents," Buller concluded.

Despite the flimsy data, the dogma that males are genetically programmed to commit infanticide of other males' offspring is still widely believed. Infanticide is generally considered a male trait, for this "just so" reason (kill mate's kids, get access to mate; get access to mate, leave more of your own kids--all carrying the "kill stepkids" gene). Which is why a new study of chimpanzees is so striking. Researchers at the University of St. Andrews, Scotland, who are part of the Budongo Forest Project in Uganda are reporting several instances of female infanticide, as they describe in tomorrow's issue of the journal Current Biology.

This isn't the first time scientists have observed female chimps killing the offspring of other females. But when Jane Goodall observed Passion and her daughter Pom cooperating to kill and cannibalize at least two infants of other females, in the 1970s, it was dismissed as pathological. But the St. Andrews researchers report three instances in which females led infanticidal attacks. In two, the females took infants from "stranger" females--not members of their group. One old male tried to stop the females, to no avail. The scientists conclude that female infanticide is not isolated or pathological (in the sense of being rare and aberrant), but a standard part of female behavior in chimps.

What triggers it? Killing the infant of another female does not increase the killer's reproductive potential, as it supposedly does for a male who kills another male's offspring, so it's hard to make the case that there is a genetically-based circuit for infanticide. As best the scientists can tell, it might be a response to pressure on females competing for foraging areas. In the previous 10 years, the chimp population in the region had risen from 42 to 75 in 2006, and 13 females with dependent offspring had moved in since 2001.

Why is it so hard to accept that sometimes behavior--by humans, chimps and every other creature with a brain--just goes off the rails? There need not be a genetic explanation for everything--and certainly not a genetic excuse.