« August 2008 | Main | October 2008 »

September 26, 2008

Social intelligence

The other day, my brother was telling me about playing mandolin on songs he didn't know, with a band he'd just met. Then there are people who can play several chess games at once or write crossword puzzles for the New York Times. How did humans get so smart? Over most of our evolutionary history, mandolin players and crossword writers probably had a hard time making a living.

This week's paper tests the hypothesis that natural selection favored intelligence largely because it was important in social interactions. If that's true, then species that interact with more different individuals should evolve greater intelligence. Federica Amici and colleagues tested this hypothesis using apes and monkeys and reported their results in Current Biology. Their paper is titled "Fission-fusion dynamics, behavioral flexibility, and inhibitory control in primates."

Chimps, gorillas, orangutans, bonobos, macaques, and two species of monkey were given the same battery of tests, including the "delayed reward" test, where they could get more food by waiting. When each species' performance was compared with that of related species (chimps vs. gorillas or spider monkey vs. capuchin monkey, for example), the species that did best was the one with the most complex social interactions in the wild. Chimps, which live in fission-fusion bands that get together in ever-changing combinations, did much better than gorillas, which live in small stable families. When they developed a "family tree" based on test results, rather than known relatedness, those with more complex social interactions were grouped together (bonobos with spider monkeys, for example). Orangutans were more similar to chimps than gorillas, which seemed a bit inconsistent with the overall pattern, as they are mostly solitary. Or so I thought.

They propose testing other species, including sperm whales. I'd like to see how they do on the delayed gratification test.

September 20, 2008

Who suffers from stress?

Recently, I wrote about how grooming each other can reduce levels of stress hormones, for example, in baboons and birds. But I asked, “why should natural selection allow excessive levels of this stress hormone?�
This week’s paper shows one way that natural selection can lead to harmful levels of stress hormones. The question, of course, is “harmful to whom?�

Writing in American Naturalist, Oliver Love and Tony Williams report that stressed mother birds pass stress hormones to their offspring. (Passing your stress on to others seems to be popular in humans also.) These hormones increase the risk of chicks dying, especially male chicks. But they may also increase the mother’s lifetime reproductive success.

How might this work? Birds that work too hard feeding their chicks may be less likely to survive to breed in future years. If conditions were the same every year, natural selection could adjust brood size over generations until it got the right balance between current and future reproduction. But conditions are not the same every year, so birds need some way to adjust brood size, decreasing it in bad years. One solution would be to lay fewer eggs in bad years (analogous to birth control). Stress hormones in the eggs could be an alternate solution, if those hormones reduce nestling survival (analogous to infanticide).

To test this hypothesis, they simulated bad years by trimming feathers on half of the birds. Trimmed and untrimmed birds had access to the same food supply, but trimmed birds had to work harder to bring the same amount of food to their chicks. The trimmed vs. untrimmed treatments were further divided into hormone treatments: half of the eggs were injected with stress hormone.

If their mother was untrimmed (simulating a good year), most chicks survived. If the mother was trimmed and the eggs were injected with stress hormone, then many chicks died. If the mother was trimmed but eggs were not injected with hormone, results were intermediate.

So far, stress hormone in the eggs looks harmful, not only to the chicks but also to the mother’s reproductive success. But those are only the results for the first breeding cycle. Feeding a nest-full of chicks is hard work, especially for a mother with trimmed feathers. Those mothers lost more weight than those in any of the other treatments. In particular, they started the second breeding cycle smaller than those in any of the other treatments. As a result, they had lower chick survival in their second brood and they themselves were less likely to survive to the next year. Trimmed mothers who had fewer chicks to feed in their first brood (due to hormone treatment of eggs) had greater chick survival in their second brood and were more likely to survive until the next year.

What about overall reproductive success? This was measured over two broods in one year and one the next. Mothers that were not trimmed had the most surviving chicks, of course. Within the untrimmed treatment, those without stress hormone added to their eggs did best. (Presumably, birds in good condition have low stress hormone levels and pass little stress hormone to their eggs.) But for birds whose ability to feed their first brood was reduced by trimming, overall reproductive success was higher if the first brood was smaller, as it was when hormone was added to their eggs.

As the authors put it, mothers with a full-size brood but reduced ability to feed them (due to clipping, analogous to a bad year) were “mismatched� the demands of their brood. They ran themselves ragged trying to feed to many chicks, which had long-term negative effects on lifetime reproductive success. Stress hormone in the eggs essentially sacrifices some chicks to enhance their mother’s lifetime reproductive success.

Which chicks were sacrificed? Mostly males. The sex ratio at laying was close to 50:50, but hormone treatment caused a small decrease in the percent of male eggs that hatched. This trend in sex ratio continued, with more male chicks dying than female. In species where males compete for mates or territories, small males may not reproduce, whereas every female can reproduce. Therefore, a bird with only enough resources to produce small offspring should concentrate on females. This appears to be an important secondary effect of stress hormones in eggs. See a previous post on control of offspring sex ratio in mice.

September 12, 2008

Evolution of mental illness

I usually only discuss papers with original data, but I'm going to make an exception this week.

"Battle of the sexes may set the brain" was published in Nature by Christopher Badcock and Bernard Crespi. It's labeled "Opinion" but it is based on facts as well as theory. Their central hypothesis is that mental illness in humans is often the result of conflicts between genes inherited from the mother and father.

Imagine a gene -- let's call it IGF2, since that's it's name -- that causes a fetus to grow faster, resulting in somewhat higher birth weight. Babies with slightly higher birth weight tend to be healthier, up to a point, but they may endanger the health of the mother, at least a little. So most women shut down the copy of IGF2 in their eggs. Therefore, most babies get an active copy from their father and an inactive copy from their mother. Occasionally, though, the mother's copy is also active in the fetus, resulting in larger-than-normal babies. It turns out that these babies have a higher risk of autism.

Their reasoning was as follows. Over most of our evolutionary history, males had the opportunity to have children with more than one female. Therefore, mutations in males that increase the growth and survival of one of their children would be favored by natural selection. Most such mutations, however, would spend half their time (over generations) in females. Their, selection would also favor growth and survival of children, but not at the expense of other current or future children. So genes like IGF2 get shut down by most mothers but not by fathers.

The authors suggest that this and other undiscovered genes may have effects that increase the lifetime reproductive success of one parent at the expense of the other. Mothers would (usually) shut down genes they were passing to their children that made one child demand excessive resources, either as a faster-growing fetus or as an overly demanding child. Fathers would activate those genes and deactivate any genes with the opposite effect. So autism may be an example of excessive male influence on a child's genes.

The theory does not assume that autism is actually beneficial to anyone. Rather, it is the result of too many doses of male influence, which (in smaller doses) would benefit the child at the expense of the mother and her other children. They suggest that some other mental illnesses, including depression, represent the opposite extreme: excessive maternal damping-down of "selfish genes." Children with Prader-Willi syndrome are undemanding but prone to depression when they get older. This trait is linked to a maternal effect on a particular part of chromosome 15.

These ideas are far from proven, but they seem worth testing. In fact, the data on the association between autism and having two active copies of IGF2 was apparently inspired by an earlier paper by these authors.

Mothers and fathers have a shared interest in the survival of their offspring, but their interests are not quite identical. This is true of plants as well as animals, of course. Do corn plants have genes expressed in pollen that make one seed grow at the expense of others? If so, does the mother plant have ways of countering these selfish genes?

September 6, 2008

Conflict builds cooperation

I just heard an interesting talk by Joan Silk on lasting friendships among female baboons, in which grooming and mutual support during conflicts are both important. Here’s a link to some of her papers. This week’s paper is on a somewhat-related topic, but in birds rather than apes.

“Duration and outcome of intergroup conflict influences intragroup affiliative behaviour� was just published in Proceedings of the Royal Society by Andrew Radford, of the University of Bristol.

Woodhoopoes are African birds (videos here) that live in small groups, typically a breeding pair and some close relatives. Conflicts over territory with neighboring groups (mostly yelling at each other) are common, often more than once a day. Neighbors rarely take over each other’s territories, but if they win the shouting match they stay and forage for awhile. Do such conflicts and their outcomes affect group solidarity?

Radford observed 12 groups in the field over several months. His main measure of positive social interactions within groups was the frequency with which they preen each other’s feathers, especially those the birds could have preened themselves. Apparently, getting the back of your head preened is an inalienable right critical to health, but body preening is a perk. (Similarly, in Joan Silk’s baboon studies, grooming is a key measure of social interactions.)

Groups that spent more time in conflicts with neighbors (up to 10% of the time!) preened each other more. This was especially true after conflicts a group lost. Interestingly, breeding pairs contributed less to territorial defense, but did more preening, especially after their group lost a conflict. Sort of like George Bush’s “"Brownie, you're doing a heckuva job" statement after New Orleans was devastated by Hurricane Katrina.

In addition to the direct benefit of parasite removal, grooming in apes can benefit health by reducing levels of cortisol. (But why would natural selection allow excessive levels of this stress hormone? Because it encourages social interactions with other benefits?) So there could be similar benefits to birds, particularly when feathers being preened are somewhere the bird itself can reach. Being preened after conflicts may reduce the chances of a bird leaving the group or increase their willingness to participate in the next conflict. Anyone want a medal?

Brief note on thumbs and junk DNA

I was going to write about this paper about a gene that evolved rapidly in humans since our lineage split from that leading to chimps. But Ed Yong at Not Exactly Rocket Science has already done a great post on it, including a picture showing its likely link to thumbs.

Comments on Ed's blog and a more complete treatment on Carl Zimmer's "The Loom" (both favorites of mine) point out the fallacy of some popular press coverage claiming this is the first evidence that "junk DNA" isn't junk after all. They both make the important point that we've known for decades that some DNA that doesn't code for protein is nonetheless very important.

On the other hand, lots of our DNA really does seem to be junk. Much of it is the product of "jumping genes" that copy themselves and insert themselves into existing DNA. These are common because they copy themselves, not because they do us any good (although, just by chance, they may occasionally be beneficial).

About 5% of DNA that doesn't code for protein is nonetheless "highly conserved", as if it were somehow beneficial and therefore maintained by natural selection. But a paper I reviewed earlier showed that much of this conserved noncoding DNA can be deleted without apparent ill effects. So if it's beneficial, it's not very beneficial. Or maybe it's beneficial only under special circumstances.