« February 2008 | Main | April 2008 »

March 31, 2008

Learning vs. lifespan?

For my first 100 posts, I’ve ignored journals with “evolution? in their names, to make the point that evolution is at the heart of biology, rather than an “appendix.? Much of our DNA can be deleted without obvious ill effect, but taking evolution out of biology would kill it as an explanatory, hypothesis-driven science. Point made? This week I untie my hands and discuss a paper from one of the 30+ scientific journals that focus on evolutionary biology.

“Learning ability and longevity: a symmetrical evolutionary tradeoff in Drosophila? by Joep Burger, Munjong Koss, and others, will appear soon in the journal, Evolution.

The ability to learn is useful under a wide range of conditions, but is it always beneficial? If so, why do most species have limited learning ability? Is there some evolutionary constraint, such as head size, that prevents evolving greater learning ability? Apparently not. Artificial selection for learning ability has been successful in several species. When artificial selection imposed by humans achieves something in months that natural selection has failed to do in millions of years, that suggests that the “improvement? has some cost that exceeds its benefits, at least in nature. But could the ability to learn really have a cost that exceeds its benefits?

To find out, the authors of this week’s paper first looked at fruit-fly populations that had evolved previously in another laboratory, under selection for increased learning ability. Fruit flies don’t like the bitter taste of quinine, so they won’t lay eggs in a material that contains it. Earlier researchers, Mercy and Kawcki, put quinine in either orange- or pineapple-flavored substrate. After some of the flies learned to associate one of the fruit odors with quinine, they removed the quinine and let the flies lay their eggs. The next generation of flies was raised only from eggs laid in the substrate that had not had quinine during the learning period. Some of these eggs could have been laid at random, of course, but the assumption was that flies with more ability to learn and remember associations would be over-represented. This turned out to be true. After 30 generations of selection, the flies were better at learning. The authors confirmed that this was true for learning how to avoid shocks as well as quinine. In humans, increased learning ability could come from cultural evolution, but there was no evidence that the flies were setting up universities or apprenticeship programs, so this was strictly a genetic change.

The question was, did this increased learning ability have any cost? The authors measured various traits, including time to maturity, egg laying, and lifespan. Lifespan was 10% lower for males and 15% lower for females, compared to the less-learned control population. This could be because selection was based on egg-laying preferences, as described above.

They also did the reverse comparison, looking at fly populations selected (by Arking and colleagues) for longer lifespan – actually, for ability to reproduce later in life -- to see whether they had lost some of their learning ability in the process. Sure enough, the ability of long-lived flies to remember which substrate had the quinine was 39% lower than in control fly populations.

The authors suggest that there may be one or more genes that affect both learning and lifespan, but in opposite directions. One allele (version of the gene) increases learning at the expense of lifespan, whereas another allele does the opposite. They mention two particular genes that could have such an effect, but they haven’t yet tested to see whether there were differences in allele frequencies between control and selected populations. They also discuss alternative explanations, such as genetic linkage -- an allele for learning ability just happens to be physically linked with an allele that limits lifespan – or negative effects of inbreeding in small laboratory populations. They argue, plausibly, that such explanations are unlikely, but it would be nice to know for sure which genes are involved. Interestingly, the effect on learning was age-specific, with longer-lived flies learning less well when young, but better when old, relative to controls.

Tradeoffs are very common in evolution. Dolphins swim better than ducks, but they can’t fly. Changes in human learning ability and lifespan today are largely due to culture (education, pensions, etc.), but there must still be some genetic changes over generations as well. Should we expect any genetic increase in human longevity to be associated with decreased learning ability, or vice versa? Humans aren’t fruit flies, but at the cellular level we are surprisingly similar. On the other hand, even if some genes exist only in a low-learning/high-longevity version and a high-learning/low-longevity version, this tradeoff could be reduced by subsequent evolution of other genes. When microbes evolve resistance to antibiotics, for example, there is often an initial cost, such as lower growth rate without antibiotics, relative to control populations. Further evolution, involving different genes, often reduces those costs. Maybe, in the long-run, we can live long and learn.

Other recent papers:
Residual reproductive value and male mating success: older males do better

MHC-mediated mate choice increases parasite resistance in salmon


The first hominin of Europe


Density-Dependent Cladogenesis in Birds


Loss of Egg Yolk Genes in Mammals and the Origin of Lactation and Placentation
Discussed by PZ Meyers

Phenotypic Mismatches Reveal Escape from Arms-Race Coevolution
Discussed by Ed Yong


Cooperative problem solving in rooks (Corvus frugilegus)

March 28, 2008

Which explains the origin of the earth?

That was one of the questions in a recent poll by The Economist. People in the US and the UK were asked to choose among these answers:
1) the theory of evolution
2) The Bible
3) "Intelligent design"

That's easy. Of the three choices, only The Bible even attempts to explain the origin of the earth. A broad definition of the theory of evolution may include possible explanations for the origin of life -- narrower definitions are limited to explaining how life has changed since its origin -- but "the origin of the earth" is the province of astronomy or geology, not biology. What I've seen of "intelligent design" is mostly whining about alleged gaps in the theory of evolution, rather than attempts to develop scientifically testable explanations of the origin of the earth or anything else, so that's out.

"None of the above" was not an option, so that leaves The Bible. It doesn't provide enough detail to allow other deities to replicate the creation process, however, so its "explanation" wouldn't be publishable in a scientific journal. (For example, the first two chapters disagree about whether birds were made before or after humans. This sort of thing can happen when a manuscript is pasted together from earlier papers and grant proposals, but it would certainly have been caught and corrected by peer review.)

Also, the origin of gods isn't explained. Given strong selection for benevolence, potence, and science, maybe gods capable of creating planets (or at least life) could evolve from a selfish, impotent, and unscientific replicator. But who, or what, would impose that selection? And how did the first replicator originate? Peer review of The BIble by more experienced Creators would have ensured that these important details were included.

March 23, 2008

Oestrus Island

"A struggle for existence inevitably follows from the high rate at which all organic beings tend to increase... It is the doctrine of Malthus applied with manifold force to the whole animal and vegetable kingdoms" -- Charles Darwin, Chapter 3, The Origin of Species
This week's paper is more about ecology and sustainability than evolution per se. In recognition of Easter, a holiday that originally honored Oestre (the goddess of spring, who also lent her name to oestrus), and which, at least in the US, retains its association with fecundity in the the egg-laying Easter Bunny, I will discuss "The simple economics of Easter Island: A Ricardo-Malthus model of renewable resource use", written by J.A. Brander and M.S. Taylor and published in 1998 (Am. Econ. Rev. 88:119).

Although this paper focuses on Easter Island, it also discusses many of the same societies in Jared Diamond's 2005 book "Collapse." The book includes much that is not in the paper, but the paper has the advantage of being shorter and of supporting specific points with specific citations, in contrast to the diffuse "Further Reading" approach used in Collapse.

The authors attempt a quantitative explanation of human population growth, resource depletion, and population decline on Easter Island from initial settlement around 400 AD to the arrival of slave-taking, smallpox-infected outsiders around 1860. They recognized the need to test their hypotheses using comparisons with other Pacific islands. They developed a mathematical model of the interactions between humans and their environment that resembles a standard predator-prey model, with humans playing the role of "predators" and natural resources being the "prey." That is, they assumed that human population growth increases with the availability of natural resources (particularly forests, which include trees that can be used to make canoes to catch fish), whose ability to recover from harvest is mathematically similar to the population growth of a prey species.
EasterModel.jpg
This relatively simple model matched archaeological and historical estimates of population growth and decline on Easter Island reasonably well. The human population overshot the level at which nature could replenish resources as fast as they were harvested, leading to faster degradation of the resource base and a population collapse. The same model, with appropriate changes in parameters, also apparently worked fairly well in correctly predicting sustainability on other islands, where populations leveled off rather than crashing. One major difference is that the wine palm that grows on Easter Island can take 60 years or more to produce fruit, whereas the coconut palms that grow elsewhere in Polynesia take as little as 7 years. This was modeled as faster "reproduction" of the resource base on other islands, but you could think of it as the difference between planting a tree for your own use versus planting a tree that your grandchildren might use, assuming that your family controls the land and manages to hold onto it long enough.

One might ask "what were they thinking when they cut down the last canoe tree?", but the authors point out that the destructive trends on Easter Island would not have been obvious. Population growth never exceeded 1% per year (less than the current world average) and loss of forest cover never exceeded 5% per human lifetime (also less than the current rate of deforestation). If carbon dioxide in the atmosphere were increasing 5% in a lifetime, rather than 50%, would we be worrying about it?

The paper has lots of interesting discussion (with citations) on other societies, from Mesopotamia to Rwanda, that suffered breakdowns linked partly to environmental degradation. Could we face similar collapses on a larger scale, today? The relationship between population growth and resource supply is more complicated today than assumed in their model. Global trade and migration might help us survive problems that only affect a small area, such as the 50-year drought that led to the collapse of Chaco Canyon. But are current world grain stocks, equal to only two months of consumption, sufficient reserves to buffer the possible effects of a climate-altering volcanic eruption, major war, or crop disease epidemic?

The paper also discusses social factors that affect willingness to make the changes needed to prevent societal collapse. In particular, "institutional change is more likely to occur when the individuals that must make the change are confident that they will be among the beneficiaries" This principle seems to be broadly applicable, even to challenges like healthcare reform.

This paper would make a great centerpiece for a seminar class on sustainability, with supplemental readings from both papers it cites and later papers that have cited it, including those with alternative points of view.

The other logical choice for the holiday would have been a discussion of mammals that lay eggs, such as the platypus. This "missing link" makes a brief appearance in a great recent post on Pharyngula, which discusses some of the evidence that we are descended from egg-laying ancestors.

March 15, 2008

100th post: reversing evolution II: mimicry in snakes

This is a kilometerstone of sorts: my 100th post! Also, cumulative visits passed 10,000 this week. I know some blogs get more hits than that in only one day, but I used to spend hours preparing a lecture for 25 students, so I guess it's worthwhile to write a blog post for 10,000/100=100 readers. My readership trend over months seems to be slightly downward, however; I hope that's due to other blogs are getting better and readers having limited time, rather than my posts getting worse. Maybe I should be spending the time on my research or my Darwinian Agriculture book instead.

I recently wrote about mimicry in butterflies, then saw an interesting paper on how natural selection and migration affect mimicry in snakes. Selection and migration ("gene flow") are two of the four main processes responsible for evolutionary changes in the frequency of alternative genes in populations; the other two are the random ("drift") processes that can have a big effect in small populations but get smoothed out in large populations and, of course, mutation.

Selection and gene flow often act in opposite directions, because animals migrating into an area (or seeds or pollen blowing in) tend to be less well adapted to their new home, relative to animals or plants that have been evolving there. This general rule held up in this week's paper, as evident from the title: "Selection overrides gene flow to break down maladaptive mimicry", written by George Harper and David Pfenning and published in Nature.

The range of the nonpoisonous scarlet kingsnake overlaps with that of the poisonous coral snake. Snake-eating predators have learned or evolved avoidance of coral snakes. Where coral snakes are common, predators also tend to avoid king snakes, because they look similar.

But where there are no coral snakes, predators are not afraid to attack kingsnakes. Without natural selection for similarlity to coral snakes, kingsnake populations evolve, becoming less similar to coral snakes in appearance. In theory, this divergence could happen in either of two ways: natural selection (acting differently than where coral snakes are present) or random drift.

Consider random drift first. In a small population, one pregnant female might survive by chance when the remainder of a small local population is killed. Or she might migrate into an uninhabited valley and found a dynasty. Or a male snake might get lucky, for reasons unreleated to his genes, and father most of the baby snakes in some small area. Either way, that individual's genes will be strongly represented in that area, whether or not they would have been favored by natural (or sexual) selection.

Because drift operates in random directions, changes in the appearance of kingsnakes due to drift would occur differently in different parts of their coral-snake-free range. But that's not what Harper and Pfennig found. Instead, there was a consistent trend towards less black and more red, as distance from coral-snake country increased. So apparently we are seeing selection rather than random drift. But why more red? Wouldn't that make the snakes more visible to predators, rather than less? This is the opposite of what happened with the butterflies, which evolved to be less apparent when the distasteful species they mimic was absent. Even some poisonous snakes are well-camouflaged, like the one below I almost stepped on. snake.jpg
Could kingsnakes in areas free of coral snakes be mimicking some other distasteful or dangerous species? For example, could baby snakes benefit from mimicking the poisonous red eft salamander (also mimicked by some other salamanders!), with red color in adult snakes an unfortunate side-effect?
RedSnake.jpg
Photos from Cindy Tong and from Harper and Pfennig (2008).

If there's no selection to resemble coral snakes, where real coral snakes are absent, why hasn't natural selection (or perhaps drift) eliminated the resemblance altogether? Evolution operates across generations, but there have been a lot of snake generations since the glaciers retreated from the snake's territory. The authors hypothesize that gene flow could maintain the moderate resemblance to coral snakes. Gene flow would result from migration of snakes from coral-snake country into regions where coral snakes are absent, carrying in their DNA the instructions for making a body that looks somewhat like a coral snake.

If so, who is migrating, males or females? The authors looked at various genes to find out. The few genes in the genome of mitochondria (descendants of bacteria that moved into the cells of the common ancestor of all animals, a billion years or so ago, and now provide all animal cells with energy) are inherited only from mothers. They didn't find any evidence that these genes were moving, so apparently female snakes don't migrate much. Genes from the cell nucleus, inherited from both parents, did show evidence of migration. So male snakes are the travelers.

This reminds me of an earlier post, in which similar data provided evidence that male aurochs mated with domesticated cows, but not the other way around. Or at least, if any domesticated bulls mated with female aurochs, their descendants died out when the aurochs did.


March 14, 2008

Why we need peer review

Most scientists also volunteer their time as "peer reviewers" for scientific journals, checking submitted papers for serious flaws, such as lack of appropriate controls. Reviewers also make good papers better by, for example, suggesting alternative interpretations of results. My own papers have been greatly improved by this process, which makes up for the few times I've thought a paper was rejected unfairly. (Fortunately, there are plenty of good journals, and the odds are against getting the same incompetent or biased reviewer twice.)

As a minimum, reviewers try to make sure that the paper describes what was done and what the results were, clearly and unambiguously. Which brings me to two recent sentences from the New York Times that probably wouldn't have made it through peer review:

And now add to the lengthening list Gov. Eliot Spitzer, husband, father of three teenage daughters, who authorities on Monday said had been involved with a ring of prostitutes.

Police found the soldier, who was still in the vicinity, shortly after 11 p.m., using a helicopter with a thermal camera.

March 9, 2008

Tricky parasites winning the evolutionary arms race

Two papers this week describe recently discovered sophisticated adapatations of two different parasites: Gall insects can avoid and alter indirect plant defenses, published in New Phytologist by John Tooker and colleagues, and Parasite-induced fruit mimicry in a tropical canopy ant, published in American Naturalist by Steve Yanoviak and colleagues (if you're in a hurry, skip to the end for amazing photos).

Various plants recruit "bodyguards" when attacked by insects. For example, when caterpillars start munching on corn (maize) plants, the plants (including uninjured leaves) release gaseous chemicals called terpenoids. These terpenoids attract parasitic wasps, which lay their eggs into the caterpillars. This eventually kills the caterpillars, which presumably benefits the plant. But what if the caterpillars could prevent the plant from signaling to the wasps? As far as I know, caterpillars haven’t evolved this trick (yet), but there are apparently some insects – the Hessian fly, Mayetiola destructor (say) – that do not trigger signaling when they feed on wheat plants. There are at least two possible explanations…

1) Maybe the plants don’t detect flies as well as they do caterpillars. This was a plausible hypothesis, because the plants detect specific chemicals in caterpillar saliva, not just the damage they cause.
2) Maybe the flies actively suppress signaling somehow. If so, then you might expect reduced production of volatiles from plants attacked by caterpillars, if they were attacked by signal-suppressing flies first. However, plants attacked by caterpillars increased total daytime production of volatiles about the same amount, whether or not they were also attacked by flies. Some specific chemicals apparently decreased with flies+caterpillars, however, relative to caterpillars alone, so that might affect signaling.

This week’s first paper looked at a similar phenomenon in goldenrod (Solidago) attacked by various insects. Three of the four insect species triggered no increase in volatiles. Caterpillars did trigger volatile release, but this signaling response to caterpillars was less if the plant was also under attack by the gall-forming fly Eurosta (no relation to the train, for those of you with Boston accents). Does preventing signaling in response to caterpillars reduce predation on helpless fly larvae, developing inside galls? If, so it is interesting that another gall-forming species did not suppress signaling in response to caterpillar damage. Gall-formers manipulate the biochemistry of their plant hosts so much anyway, to make them form galls, that effects on volatile signaling could be a side-effect, rather than the result of active manipulation of the “distress signal? pathway.

I don't know whether anyone has worked out the evolutionary history of plant-to-bodyguard signaling. My guess is that all insect-damaged plants release some volatile chemicals, which predatory insects (or parasites that lay their eggs in plant-eating insects) evolved the ability to detect and follow to their prey. That then selected for plants that produce even more of the volatile chemical -- "Hey! I've got more caterpillars than those other plants!" (whether or not that's true) -- at which point it seems reasonable to call it a signal. Someone has genetically engineered plants to produce these signals even when they're not under attack, as a possible form of ecological pest control. This isn't necessarily a good idea, for reasons I'll explain in my book on Darwinian Agriculture.

Parasites manipulate their hosts in other ways as well. Host with brains may have their behavior manipulated. For example, Toxoplasma gondii infects rats, but then needs to move to a cat to complete its life cycle. So it manipulates the rats (presumably by producing a hormone-like chemical) so that they are less afraid of cats. T. gondii also infects humans, and there may even be cultural differences among societies, depending on the frequency of infection. It sounds like science fiction, and Biology in Science Fiction (among my favorite blogs) has additional examples.

But there's nothing fictitious about this week's second paper. In this case, the parasite is a tiny worm-like nematode that infects ants. The ants actually infect their own larvae by feeding them a diet that includes bird droppings containing the nematodes. But how do the nematodes get into the bird droppings? By manipulating the ants so that they get eaten by birds. They somehow make the ants rear section turn red -- it's normally black -- resembling a ripe fruit. They make the ant's skin transparent, and the hind section is full of nematode eggs that look red. Fruit-eating birds are fooled into eating the ants and then excrete the parasite in their droppings. The droppings are collected by ants and the cycle repeats. Seems like a lot of trouble. You would think that the ants would pass the nematodes in their own feces and reinfect each other that way, but apparently their digestive system doesn't work that way.

Both papers this week are examples of what Dawkins called the "extended phenotype." Normally we think of the phenotype as a physical, biochemical, or behavioral trait of an organism that is controlled by that organism's genes. But, Dawkins points out, if beaver genes control the behavior that controls the shape of a beaver dam, then the shape of the dam could be considered an extension of the beaver's phenotype. In this case, ant rear sections that look like fruit are the phenotype resulting from one or more nematode genes. It would be interesting to mutate the nematodes to see which genes are involved.
plant_5708.jpg
Above: infected ant, with real fruits for comparison. Below: normal and infected ants. (All photos by Stephen Yonoviak)
Normal_ant.jpg
Infected_ant.jpg

March 1, 2008

Knowing when not to cheat

This week’s paper is Facultative cheater mutants reveal the genetic complexity of cooperation in social amoebae published in Nature by Lorenzo Santorelli and colleagues at Rice University and Baylor College of Medicine, both in Texas.

The evolution of cooperation is a central problem in the history of life. Darwin explained how sophisticated adaptations -- “the structure of the beetle which dives through the water… the plumed seed which is wafted by the gentlest breeze? -- can evolve in a series of small improvements over generations. But some of the major transitions in evolution are harder to explain, because It seems that they should have been opposed, rather than supported, by natural selection. The origin of multicellular life is a good example. It’s not that hard to imagine independent cells working together in loose groups for mutual benefit – huddling together for defense, say – but why would a cell give up the ability to reproduce, as most of the cells in our bodies have done?

Supernatural intervention, perhaps? If Intelligent Design were a scientific field, rather than a religion, its researchers would be doing experiments to test this hypothesis, rather than making movies and whining. The problem, of course, is that "test" means "subject to possible disproof."

Cooperation among a group of genetically identical cells is easy to understand. Then, kin selection will favor some giving up individual reproduction, if that increases their collective reproductive success. But, when multicellular life first evolved, how often would cells in a cluster all be the same strain? “Hi, I’d like to join your cluster! How about if I reproduce while the rest of you fight off predators?? Cells that behave this way are known as “cheaters? because they don’t cooperate themselves but benefit from cooperative activities of others. They don’t really talk like that, of course.

dicty.jpg
Dictyostelium life cycle (David R. Caprette, Rice University)

Cooperation in the “social amoeba? Dictyostelium can be disrupted by such cheaters, just as we assume early multicellular cooperation could have been. These amoebae live much of their lives as independent cells. But, when starved, they form resistant spores, which eventually resume growth as the next generation of independent cells. The spores are held at the end of a stalk, which presumably helps dispersal by animals to environments that may be more favorable.

The problem is that the cells in the stalk are left behind and die without reproducing. (I previously discussed a paper showing that when “slugs? composed of multiple Dictyostelium cells crawl together through the soil, some cells sacrifice themselves to protect their clonemates.) Suppose you have a 50:50 mixture of two Dictyostelium strains. A “cheating? strain would be one that manages to get mostly into the spores, letting the other strain form the stalk. (This is how the researchers measured cheating, using a fluorescent version of the control strain so they could see what percent of the spores it made.) So would cheaters become more common in each generation? Maybe not, if the cheater wasn’t able to form a good stalk on its own. Then, its evolutionary success would be entirely dependent on infiltrating other strains.

But what if cheating is facultative? That is, what if a strain forms stalks to hold up its own spores, but relies on the stalks of another strain when it can? I would have guessed that such a sophisticated form of cheating would be beyond the capabilities of an amoeba. (Similarly, in our own research, I have assumed that rhizobium bacteria that “cheat?, by providing their host plant with less nitrogen, don’t check first to see whether other rhizobia on the same plant are taking up the slack.)

It was certainly worth checking for facultative cheating, however, and that’s what the researchers did. They let 10,000 different strains compete over ten “generations? (life-cycles; see diagram), all mixed together. This mixing would allow cheaters to become more common, via spores, at the expense of cooperative strains that contribute more cells to stalks. Then they tested a couple thousand strains – try doing that with meerkats! -- to make sure they could form spore-bearing stalks on their own. Only 1% of the strains couldn’t. So if there were any cheaters in the other 99%, they were facultative, only cheating when they could get away with it, i.e., when another strain was forming a stalk they could use.

Such facultative cheaters were apparently fairly common. When the cooperative control was mixed 50:50 with the evolving population, only 40% of the resulting spores were of the control strain. Testing 40 different strains, they found that 31 were cheaters (making more than their share of spores), 5 were “losers? (making less than their share of spores), and the rest were neutral.

Is there a cheating gene? For these amoebae, cheating clearly had a genetic basis, but more than one gene was involved. In fact, mutations in any one of more than a hundred different genes could convert a cooperator into a cheater. Does this mean that, when cooperation among cells first evolved, it would have required hundreds of mutations to happen all at once? No, that would be another version of Behe’s mousetrap fallacy. Just because a complex system has a lot of parts that are now essential, that doesn’t mean that a simpler version couldn’t have worked with fewer parts than we imagine to be necessary. Still, if even a small fraction of these genes are essential to cooperation, that could help explain why multicellular life took so long to evolve.