Take, for instance, a recent piece of research on cannibalism.

Anthropologists have been arguing for decades over the prevalence of cannibalism in our past: did any of our ancestors -- however long ago -- practise it on regular basis? Despite lurid tales from explorers and adventurers and suspicious tool marks on fossil bones, the evidence for habitual cannibalism (beyond the occasional ritual) has been elusive.

But now a group in England is claiming that the human version of mad cow disease, and the genetics of resistance to it, point to cannibalism as a routine of human life in the past.

The story brings together several lines of scientific evidence. First there is the link between cannibalism and diseases such as mad cow.

That connection was made 40 years ago by Carleton Gajdusek, who discovered the cause of a strange brain-wasting disease called Kuru that plagued the Fore people of Papua New Guinea. They routinely consumed the bodies of their recently deceased, and Gajdusek figured out that some agent in the brains of the corpses (he thought it was a virus) was infecting those who participated in the ritual feast, most of whom were women and children. As soon as this cannibalistic practice ceased, the disease vanished.

In the years since, the nature of Kuru was been clarified: it is thought to be related to a handful of other diseases, including scrapie in sheep, mad cow disease, and the human Creutzfeldt-Jacob disease or CJD.

There are now two versions of CJD, one sporadic, the other apparently the result of consuming meat from animals with mad cow disease.

The agent causing these diseases is thought to be a "prion," a protein molecule that can adopt two shapes. In one version it is just one of thousands of proteins playing a routine role in the brain -- we all have those prion proteins.

But if prions of the alternate shape gain access to our brains, disaster ensues. They cause the normal resident prions to flip orientations too, and the result is wholesale destruction of brain tissue.

Here is where the genetics comes in. We have two genes for everything: one from mother, one from father. John Collinge and his colleagues at University College London discovered that those people who come down with either form of Creutzfeldt-Jacob disease almost always have two identical genes for their prion protein.

This suggested that, by contrast, people who avoid the disease have two genes that differ very slightly, but apparently crucially.

Collinge then went back to Papua New Guinea and looked at the survivors of the Kuru era, those women over 50 who had taken part in the cannibalistic feasts and yet somehow survived. Most of them had two slightly different prion genes as well.

But younger members of the Fore tribe, who grew up after cannibalism was banned, showed the usual mix of pairs of identical and different genes.

Collinge concluded that as long as Kuru was rampant in the Fore tribe, having two different prion genes was protective against the disease, and those who didn't have them, died.

In this indirect way, the genetics was pointing to the presence of cannibalism in the past.

Collinge then went worldwide, sampling genes for prion proteins from people of all ethnic backgrounds, and found that all possessed the two versions of the prion gene.

Something must have been exerting an evolutionary pressure to maintain those two genes in the population; Collinge argues it was widespread consumption of human flesh.

Further genetic analysis suggested that this pattern very old, going back something like 500,000 years.

The age should relieve those who aren't keen to think that their ancestors were cannibals: 500,000 years ago our ancestors weren't even Homo sapiens yet, but an earlier hominid.

Watson and Crick were prescient guys: they had some sense of what they had achieved, but there's no way they could have imagined how far DNA would one day reach.