IN the 1920s and 1930s, scientists from both the political left and right would not have found the idea of designer babies particularly dangerous - though of course they would not have used that phrase. Today, I suspect that the idea is too dangerous for comfortable discussion, and my conjecture is that Adolf Hitler is responsible for the change.

Nobody wants to be caught agreeing with that monster, even in a single particular. The spectre of Hitler has led some scientists to stray from "ought" to "is" and deny that breeding for human qualities is even possible. But if you can breed cattle for milk yield, horses for running speed, and dogs for herding skill, why on Earth should it be impossible to breed humans for mathematical, musical or athletic ability? Objections such as "these are not one-dimensional abilities" apply equally to cows, horses and dogs and never stopped anybody in practice.

I wonder whether, some 60 years after Hitler's death, we might at least venture to ask what the moral difference is between breeding for musical ability and forcing a child to take music lessons. Or why it is acceptable to train fast runners and high jumpers but not to breed them. I can think of some answers, and they are good ones, which would probably end up persuading me. But hasn't the time come when we should stop being frightened even to put the question?

Richard Dawkins is Charles Simonyi Professor of the Public Understanding of Science at Oxford University