London

ANYONE who supposes that evolution doesn't happen, or doesn't matter, should spare a thought for H5N1, the virus causing avian flu. If we're unlucky, this virus will give us a nasty demonstration of evolution in action.

Viruses are among the simplest parasites. They are essentially tiny parcels of genes that are mailed from one organism to another, either directly, through sneezes, feces, semen and the like, or indirectly, through carriers like insects. But these tiny parcels can mean big trouble: viruses reliably feature on nature's roster of top killers.

The influenza virus that caused the infamous Spanish flu pandemic of 1918 had only eight genes - but it brought about more than 20 million human deaths. And alas, its lethality cannot be blithely attributed to wartime deprivation. For one thing, it was particularly deadly in young, healthy adults. For another, in a remarkable feat of genetic engineering, a team of biologists recently reconstructed the 1918 virus and used it to infect mice. The results are sobering. The 1918 virus is far, far more lethal in mice than are other human flu viruses.

H5N1 also has eight genes (by way of comparison, humans have about 20,000). So far, the virus's effects have been more modest than those of the 1918 influenza: it has killed a lot of birds and about 60 people. That's still worrying, however, because it has killed more than half of the people it has infected. For a virus, that is a high death toll.

At the moment, the virus cannot pass easily from one person to another. But there are a couple of ways it could evolve to do so.

The virus might infect someone already sick with a strain of human flu, and the two viruses could have sex, thus creating a new virus that contains some genes from each. Such viral hanky-panky is thought to have led to the flu pandemics of 1957 and 1968. Or the virus could mutate - acquire accidental changes to its genetic material - in such a way that it becomes able to travel between people. Mutations to an avian flu virus are thought to lie behind the 1918 pandemic.

Sex and mutation: these are not special processes reserved for viruses. They are two fundamental mechanisms of evolutionary invention. Mutations alter the information content of genes; sex shuffles the pack, generating new gene combinations. They sound simple, and they are - but don't let that deceive you. Simple processes can have great power. After all, a few mutations to a bird virus could - in the absence of a vaccine - mean the difference between 60 people dead and several million.

Now that we can sequence genes and genomes, we know precisely how evolutionary changes accumulate. We know the differences between a fruit fly and a mosquito, between a human and a chimpanzee, between a virus that kills chickens and a virus that kills people. We can see which genes have been changing quickly and which have hardly changed at all. We can see which genes cause populations to diverge and then split into new species.

What is more, with genes and genomes we can supersede the often patchy fossil record to look back in time. One day, when the crocodile has joined the chicken in having had its genome sequenced, we'll be able to compare birds and crocodiles - the two closest living relations of Tyrannosaurus and company - and conduct evolutionary detective work, using their genomes to infer the genome of a dinosaur.

But the most important point is this: viruses and other pathogens evolve in ways that we can understand and, to some extent, predict. Whether it's preventing a flu pandemic or tackling malaria, we can use our knowledge of evolutionary processes in powerful and practical ways, potentially saving the lives of tens of millions of people. So let's not strip evolution from the textbooks, or banish it from the class, or replace it with ideologies born of wishful thinking. If we do, we might find ourselves facing the consequences of natural selection.

Olivia Judson is an evolutionary biologist at Imperial College in London.