This would have been no news to Queen Elizabeth I or Margaret Thatcher, each of whom was highly successful in dominating the clever men who surrounded them. And both of them made sure that no other woman could threaten their dominance.

The researchers, led by Charlotte Hemelrijk, theoretical biologist at the University, and assisted by her former PhD student, Dr. Jan Wantia and a Swiss anthropologist, Dr. Karin Isler, created a virtual world, Domworld, with which they could simulate interactions between monkeys, in order to discover how female dominance develops.

Domworld predicted females to be more dominant in a group with a relatively large number of males, and that this would result from 'self-organization' rather than from inherited size and strength; to verify this prediction, one that was unexpected by the researchers, they analyzed the literature for evidence, showing the predictions of the computer model to be accurate. Margaret Thatcher was smaller than every other member of her cabinet, by the way, and it was only on Spitting Image that she hand-bagged her ministers.

"This is an interesting way of conducting research," says Hemelrijk. "You discover something unexpected in the virtual world and then you test your findings in the real world."

What does this say about virtual reality worlds? Perhaps someone could study dominance relationships in virtual social groupings to see how it plays out for humans. There are some kickers, though: a male can have a female atavar, or vice versa, and there are other possible impersonations. But if behaviour is based on who you think you're dealing with, rather than the truth (doesn't that sound like real life?) maybe it wouldn't make a difference.

Do mostly male raiding parties in World of Warcraft sometimes have one or a small number of women members? How dominant are they?

Let me know! Perhaps we can develop a theory of virtual dominance behaviour.

This piece of not very earth-shattering news is wide of the mark, though.

It's certainly a fact that longevity is increasing rapidly, and ever faster than last year's estimates, and that people like pensions administrators and Goverment actuaries should start looking for new jobs as paralympic coordinators; but the answer is not far to seek.

In the extreme case, people will become immortal. It's not really in doubt; it's just a case of when, and it's obvious that such people cannot be offered pensions in the conventional sense of an annuity which is a mixture of capital and interest.

They will have to live off the returns of investment, whether that be cash, property, shares or whatever.

For most better-off older people, the State pension is already nearly irrelevant; perhaps it pays to feed the cats and for the odd evening out. But for a decent cheque every month, people know that they have to invest on a permanent basis.

A majority of people in the more advanced economies now understand that they have to provide for themselves in old age. Governments should give up the pretence that they invest pension contributions into some kind of mythical pension fund - it was never true, in any event - and compel short-sighted young people to save into approved investment schemes from the moment they leave school.

Social security contributions are nothing but a tax; if the equivalent amount of money had been invested into retirement provision by today's 65-year-olds, they would be getting three times the amount the government pays them.

The writing is now on the wall. Every year that governments hesitate simply worsens the crisis that will come, when they have to give in and stop lying to their citizens. Best to do it now!

The robot smells and acts like a cockroach, persuading the real insects, which rely on olfactory and tactile cues, to behave in unusual ways, for instance gathering in an illuminated shelter rather than in a dark one, which would be their normal preference.

'Insbot' is a wheeled robot about the size of a small matchbox, housing several computer processors hooked up to a camera and an array of infrared proximity sensors.

"If you don't put the pheromone molecules on them, the cockroaches get scared because they are afraid it is a predator," says roboticist Jean-Louis Deneubourg.

Deneubourg's team put 4 robots and 12 cockroaches in an enclosure with dark and lit-up areas. They programmed the robot to seek out real insects but also veer towards the lit area.

"It's important they prefer the light shelter, but not too much," Deneubourg says. "If they have too strong an attraction for the light they will go straight there and not interact with the real cockroaches."

The researchers plan to develop other robots that can socialise with animals and influence their behaviour in a similar way. They have already begun studying the group behaviour of sheep and chicken. "Chickens are a good example of a mixture of collective intelligence and leadership," Deneubourg says.

"It would be interesting to build our own intelligent societies of animals," says another member of the team.

Previous research at the University, led by Jean-Marc Amé and José Halloy, senior research scientist, tested cockroach group behaviour by placing larvae of the insects in a dish that contained three shelters. The test was to see how the larvae would divide themselves among the shelters. Cockroaches gather in sheltered locations, or resting places, so the question was how they go about forming groups that maximize protection but minimize overcrowding. "Resting places are a nice experimental setup to test collective decision-making," Dr. Halloy said.

After much "consultation", through antenna probing, touching and more, the cockroaches divided themselves up perfectly. For example, if 50 of the animals were placed in a dish with three shelters, each with a capacity for 40, 25 of them grouped together in the first shelter, and 25 in the second, leaving the third shelter empty. If each shelter had capacity for 50 inmates, all the cockroaches moved into one shelter.

Halloy says that, in the absence of aural communications, cockroaches use chemical and tactile signals, plus visual cues, to communicate with each other. "When they encounter each other they recognise if they belong to the same colony thanks to their antennae that are 'nooses', that is, sophisticated olfactory organs that are very sensitive," he says.

Altruism - benefiting fellow group members at a cost to oneself - and parochialism - hostility toward individuals not of one's own ethnic, racial, or other group - are common to human nature, but we don't immediately think of them as working together hand in hand. In fact the unexpected combination of these two behaviors may have enabled the survival of each trait according to Bowles and Choi.

They show that the two behaviors - which they term "parochial altruism" - may have in fact coevolved. On the face of it joining parochialism to altruism is puzzling from an evolutionary perspective because both behaviors reduce one's payoffs by comparison to what one would gain by avoiding them.

Aggression consumes resources and risks death; altruism, particularly toward those with whom we have no direct relationship, has the effect of helping other genes advance at our expense. But parochial altruism could have evolved if parochialism promoted intergroup hostilities and the combination of altruism and parochialism contributed to the success of these conflicts.

Using game theoretic analysis and agent-based simulations Bowles and Choi show that under conditions likely to have been experienced by late Pleistocene and early Holocene humans neither parochialsim nor altruism would have been viable singly, but by promoting group conflict, they could have evolved jointly.

"But even if a parochial form of altruism may be our legacy," said Bowles, "it need not be our fate." He pointed to the many examples of contemporary altruism extending beyond group boundaries, and the fact that hostility toward outsiders is often redirected or eliminated entirely in a matter of years.

The new study by Dr Lucy Bates and Professor Richard Byrne found that African elephants reacted with fear when they detected the scent of garments previously worn by men of the Maasai tribe - whose young men are known to demonstrate their virility by spearing elephants. The elephants also responded aggressively to red clothing, which is characteristic of traditional Maasai dress.

On the other hand the elephants showed much milder reaction to clothing previously worn by the Kamba people, agriculturalists who pose little threat.

Using evidence of elephant behaviour gathered over 35 years by researchers from the Amboseli Elephant Research Project in Kenya, the psychologists expected that elephants might be able to distinguish among different human groups, according to the level of risk that each presents to them.

They said, "We were not disappointed. In fact, we think that this is the first time that it has been experimentally shown that any animal can categorise a single species of potential predator into subclasses based on such subtle cues."

The researchers first presented elephants with clean, red clothing and with red clothing that had been worn for five days by either a Maasai or a Kamba man. They found that Maasai-scented clothing motivated elephants to travel significantly faster in the first minute after they moved away. The elephants also travelled farther in the first five minutes, and took significantly longer to relax after they stopped running away.

They then investigated whether elephants can also use garment colour as a cue to classify potential threat - and found that the elephants reacted with aggression towards red but not to white cloth; suggesting that they associated the colour red with the Maasai.

The researchers believe that the difference in the elephants' emotional reaction to odour versus colour might relate to the amount of risk they sense in the two situations, encouraged by a particularly keen sense of smell. "With any scent of Maasai present, fear and escape reactions seem to dominate anything else," said Dr Bates.

The tendency of the elephants to flee at the mere whiff of a Maasai person may have other implications. Professor Byrne explained, "While elephants can undoubtedly be dangerous when they come into conflict with humans, our data show that, given the opportunity, they would far rather run away, even before they encounter the humans in person.

"We see this experiment as just a start to investigating precisely how elephants 'see the world,' and it may be that their abilities will turn out to equal or exceed those of our closer relatives, the monkeys and apes," he added.