New York Times 2003

They found that all ethnic groups harbor genetic signatures that protect against infection by prions, proteins in meat that lay waste to the brain. Prions can come from animals, as mad cow disease showed, but they kill off cannibals far more effectively.

The best-known example of that is the Fore people of Papua New Guinea, who from the 19th century to the late 1950's held "mortuary feasts" on deceased relatives.

From the 1920's to the 60's, an epidemic of kuru, a prion disease, killed more than 200 Fore a year, especially women, who ate brains while men ate muscle. A study of 30 elderly Fore women who survived the era showed that 23 had the prion-protection gene, which was far less common among Fore born later, suggesting that it had protected them.

The discovery of similar genes in ethnic groups around the world, the researchers said in the April paper, is powerful evidence that cannibalism had at some time made such protection important.

It should be noted that forms of human cannibalism still go on. "Muti murders" are well-known to the South African police. Each year they find a few bodies missing the lips, tongue or other parts used in traditional medicine muti in Zulu. Last year, a murderer was arrested in Krugersdorp trying to sell a head for $1,000, and journalists clandestinely filmed a morgue attendant offering two human hands for $400.

Eating human placentas has its fans, especially in Britain, among some vegetarians who think of it as meat that does not involve killing as well as women who give birth at home and believe it helps the uterus contract. In 1998, the British Broadcasting Standards Commission criticized a program featuring a mother making placenta pâté, and a Web site for older mothers,, still offers a recipe for roast placenta in tomato sauce.

There have been philosophical defenses of human cannibalism. The best-known, of course, is "A Modest Proposal," the 1729 satire by Jonathan Swift, who suggested solving the Irish overpopulation problem by judicious application of knife and fork.

Studying lower species is, at least, less fraught with political touchiness. Researchers who believe that the Anasazi people of the American Southwest widely practiced cannibalism have been accused of racism by their descendants, the Hopi.

But lower-order cannibalism, not driven by emotion or religion, is in some ways more of a puzzle.

Presumably, it confers some evolutionary advantage, but it seems such a backward-turning gear in the Darwinian drive train. After all, each cannibal has just as good a chance of being eaten as of eating. That ensures survival of the fittest, but ultimately of only one fittest not an ideal species-preservation goal.

Among robber flies so called because some with orange stripes sneak up to hives to snatch honeybees for food the females are so fast and vicious that they sometimes grab the smaller males and eat them without its being clear that they realize what they are eating. Male courting dances, done from a distance, seem to be a way of letting females know that their intentions are sexual, not culinary.

And there are powerful medical arguments against cannibalism. As the Fore showed, dining on kin magnifies disease even faster than marrying kin.

A 1998 study of cannibalistic tiger salamanders led by Dr. David W. Pfennig, a University of North Carolina biologist, tested that notion.

Four groups of cannibalistic tiger salamander larvae were fed, respectively, healthy tiger salamanders, healthy small-mouthed salamanders and salamanders of both types that had been sickened by letting them swim in a bacterial soup. The tigers that ate healthy prey or sickened small-mouths all survived and grew. But nearly half those that ate sick members of their own species died. The tigers, autopsies showed, had eaten bacteria best suited to killing tigers.

But they keep eating one another, despite the risks, because they breed in ephemeral snow-melt ponds. "Their pools dry up really fast, so they need to get out of the larval stage quickly, and the best way to do that is to get lots and lots of food," Dr. Pfennig said.

The biologist, who also studies spadefoot toad tadpoles in desert rainwater ponds, says such time pressures drive species toward cannibalism. "Disease is something you can potentially overcome, but starvation is pretty permanent," he said. Tadpoles, he said, can taste each other, and tend to spit out close kin.

In species that hatch many offspring, some serve as surplus protein, all of it presumably infection-free. Sharks eat their siblings in the womb, tiny copepod crustacean females eat their young, as do burying beetles, who match the number they leave alive to the size of the mouse corpse they bury with them.

But sometimes one's more mature brethren provide nutrients no one else can. Rattlebox moth larvae, which eat poisonous plants for toxic alkaloids that will protect them as moths, may eat alkaloid-laced siblings if they cannot get to the plants themselves.

Down at the bacterial level where species mutate rapidly and Darwin is the only god the mystery is why cannibalism persists.

"There's a big incentive to cheat, to evolve a strain resistant to the killing factor," Dr. Losick said. "But we don't see cheaters. And we still don't understand: What's in it for the siblings?"