Darwin's roulette wheel

A comment from a British professor of molecular genetics in the Guardian Unlimited: Mutation in full flight. Here's the conclusion:

So far there have been only about 200 cases of H5N1 influenza in humans, vastly fewer than the millions of cases estimated in birds, so the evolutionary dynamics of H5N1 are still firmly tied to their feathered hosts.

However, each human victim is effectively a Darwinian roulette wheel, with natural selection acting as the banker ready to reward any new viruses capable of crossing the species barrier.

We aren't there yet. The fact that the H5N1 strain has infected only a few humans is testimony to its low rate of infectivity. Given enough time and enough human hosts a species jump becomes almost inevitable. But nobody knows how much time, or how many hosts are needed to generate a pandemic strain.

An outbreak of the H5N1 disease in Hong Kong in 1997 was controlled by extensive culling of domestic fowl. Our best hope is that this outbreak will be similarly controlled or burn itself out before the evolutionary dynamics of H5N1 can shift to the human population.