« April 2008 | Main | June 2008 »

May 31, 2008

Traditional values in bees

The beehive was an early Mormon icon, symbolizing hard work and cooperation. To an evolutionary biologist, however, a beehive could symbolize reproductive skew, a situation where some individuals reproduce much more than others. Extreme reproductive skew is one of the defining characteristics of eusocial species, of which honey bees are a prime example. Reproductive skew can differ between the sexes. In honey bees, the queen lays most of the eggs, and most females don't reproduce at all. Polygamous species and groups show the opposite pattern: males vary much more in reproductive success than females do. Maybe an inverted beehive would have been a better symbol. Note that the cells in our bodies behave somewhat like a eusocial bee colony; any children we have are directly descended from a few sex cells, while brain cells and skin cells play the supporting role of worker bees.

This week's paper, "Ancestral monogamy shows kin selection is key to the evolution of eusociality" was published in Science by William Hughes and others. Like humans, some bees are monogamous, meaning that the queen mates with only one male, so her daughters (the workers) are all sisters. In other bee species, the queen mates with several males, so her daughters are half-sisters. Relatedness generally favors cooperation, although there are some possible complications, discussed below.

This week's paper asks how mating behavior affects the evolution of eusociality. They reasoned that, if mating system doesn't matter, then today's eusocial species could be descended from either monogamous, polygamous, polyandrous (each female has multiple mates), or promiscuous ancestors. Alternatively, eusociality may evolve more easily with one of these mating systems than with the others.

The authors used the method of ancestral state reconstruction, applied to nine groups of eusocial insects (bees, wasps, and ants). If a reader can suggest a web page with a clear explanation of this method, preferably with diagrams, I will add a link to it here. The basic idea is to work through the known family tree of a group of related species, trying to infer the traits of ancestral species from the traits of their descendants.

For example, consider the evolution of diet in a group of related animals. Suppose most species descended from X eat insects and only a few eat seeds. It seems more likely that there were a few switches from insects to seeds, over the course of evolution, rather than many switches from seeds to insects. So maybe X ate seeds. But what if, since X's time, insects have become more common where these species live, or seeds have become scarce? Then maybe eating insects could have evolved independently in multiple species. If insect-eating had evolved independently in each species, however, you might expect different species to catch insects in different ways. If the insect-eaters descended from X all use tubular tongues to catch insects, then maybe X also caught insects with a tubular tongue. The more arbitrary the similarities (in the sense that other methods would have worked equally well), the more likely it is that they are inherited from a common ancestor, rather than having evolved independently.

The authors used information on 267 species of eusocial insects. Most were monogamous and their analysis suggested that the exceptions, which include honey bees, represent an evolutionary change from the ancestral state of monogamy. In those cases when polyandry evolved, they found that "totipotency [ability of workers to lay eggs] was lost prior to or concurrently with the evolution of polyandry."

This last result seemed somewhat surprising to me, because I had thought that polyandry was a strategy by the queen to prevent reproduction by workers. In some species, workers can lay male eggs without mating. The more mates the queen had, the more likely a fellow worker is to be a half-sister rather than a sister. Therefore, multiple mating by the queen decreases relatedness of one worker to another worker's egg (relative to an egg from her mother, the queen). So, in polyandrous species, workers tend to eat eggs laid by other workers. In monogamous species "it is often the queen who does the policing, unless worker reproduction seriously decreases colony efficiency" (Beekman and Oldroyd 2008 Annual Review of Entomology 53:19). So if workers had already lost the ability to reproduce, what drove the evolution of multiple mating by queens? I hope we humans haven't been a bad influence!

Also this week:

Searching for candidate speciation genes using a proteomic approach: seminal proteins in field crickets

Massive Horizontal Gene Transfer in Bdelloid Rotifers

Water Activity and the Challenge for Life on Early Mars

Sex differences in spatial cognition in an invertebrate: the cuttlefish

Individual genetic diversity correlates with the size and spatial isolation of natal colonies in a bird metapopulation

May 25, 2008

Pest control for ants


(Top) A small leafcutter worker atop a leaf guards her sister against attacks by parasitic flies. Ants carrying leaves cannot use their mandibles for defense, so they carry hitchhikers to ward off the parasites. (Bottom) The fungus garden in a nest of Atta leaf-cutter ants. Notice the diversity of ant sizes within a colony, from the large red soldier ants to the minute orange ants tending to the garden. Atta ants have some of the most sophisticated caste systems among the social insects. -- photos and captions from Alex Wild (mymercos.net)

This week’s paper, “Black yeast symbionts compromise the efficiency of antibiotic defenses in fungus-growing ants" by Ainslie Little and Cameron Currie, was just published in Ecology. Elsa Youngsteadt interviewed me, among others, for a story in Science about this research.

I’ve never done research on the fungal “farms" of ants and termites, but I’ve been interested in them every since a camera company bought a close-up photo (not Photoshopped like this one) of an ant carrying a leaf along a barbed wire “bridge" on its way back to its nest, from my mycologist father, William Denison. Dad was best known for pioneering research in the tops of tall trees, but never had to fight a shaman, as far as I know.

The ants feed harvested leaves to fungi, then eat the fungi, so a feedlot might be a better human analogy than a farm. (Fungi are more closely related to animals than to plants.) The fungal “crop" (or “herd") can be attacked and destroyed by a second fungus, Escovopsis. (Wolves?) But Escovopsis is controlled by antibiotics produced by bacteria that live on the bodies of the ants.

Ever since I first heard about this, I’ve wondered, “what benefit do individual bacteria get from making the antibiotics?" Sure, collectively the bacteria and the ants are interdependent. But mutant bacteria that don’t make antibiotics (thereby saving some energy) must arise all the time. If the benefits of a healthy ant colony are just as available to these mutants as they are to antibiotic-producers on the same ants, then we would expect “cheating" mutants to take over the surface of each ant, whatever the long-term consequences. Bacteria don't plan ahead.

This week's paper introduces a new player that may hold the key to this question: a yeast (single-cell fungus) that lives on the ants, along with the bacteria. In culture dishes, the yeast grew more when bacteria were present, and the bacteria survived less when the yeast was present. So it appears that the yeast is, in some sense, “eating" the bacteria.

In the short term, therefore, the yeasts may be harmful to the ants. Escovopsis caused more damage to the ants’ fungal gardens when the yeast was present.

But is it possible that yeasts have an evolutionary effect that benefits the ants? If antibiotics that kill Escovopsis also have some protective effect against yeast, then there would be an individual benefit to antibiotic production that would select against “cheating" mutant bacteria that don’t produce antibiotic. The fact that antibiotic-producing bacteria still have a net positive effect on the yeast doesn’t undermine this hypothesis; the question is whether mutant bacteria that don’t produce the (Escovopsis-killing) antibiotics are hurt more by the yeast than antibiotic-producers are.

There is a good precedent for this hypothesis. Some bacteria that live on plant roots make antibiotics that protect the plant from fungal attack. In 2003, some colleagues and I pointed out (also in Ecology) that this raises the same question as the antibiotic-producing bacteria on ants: “is there some individual benefit that selects against nonproducing mutants?" It turns out that the antibiotics protect individual bacterial cells from being eaten by protozoa, in addition to helping them compete against other bacteria. I suggest that black yeasts on ants might play an analogous role to the protozoa near plant roots: imposing individual selection for antibiotic production that also happens to provide a collective benefit.

The authors of this week’s paper don’t mention this hypothesis and there are may be other explanations. Ulrich Mueller, another expert on ant agriculture, suggested in the Science story that by transplanting only (the most productive?) part of a fungal garden, ants are “selecting on an entire community that has desirable characteristics." If so, this could be a genuine example of “group selection" in nature.

But the antibiotic-producing bacteria live on the surface of the ants; they aren’t part of the garden. Therefore, for group selection to counteract loss of antibiotic production, the unit of selection would have to be the ant. Do ants whose bacterial populations stop producing antibiotics remove themselves from the colony, for the benefit of their mother and sisters? Do their sisters kick them out? I expect we will have answers to these questions some day.

Watch for a chapter on ant agriculture in my book, Darwinian Agriculture, which should be published in 2009.

May 16, 2008

Sex! Identity theft! Burying beetles!

The “Coolidge effect? – I would have named it for a different American president – is a tendency of some males to be more interested in a new sex partner than one they have mated with in the past. Males that don’t help care for young may have more descendants this way than if they put all their eggs (so to speak) in one basket. But to avoid remating with the same partner, one first needs to remember them all.

Most chimps and humans can probably remember all previous partners, but what about other species? One burying beetle looks much like another, or so it seems to us. In this week’s paper, Sandra Steiger and coauthors asked whether males of this species prefer novel mates and, if so, how they tell them apart.

Male beetles had a female introduced into their chamber every ten minutes, but the first four were actually the same individual. (So the police chief, whose assistant was bringing him the same newspaper every day to save money, says, “I know people think policemen are stupid, but look at this paper; a doctor has run over his own mailbox four days in a row!? Joke credit: Milan Copic.) The fifth female was different. The males took longer to mate with the first female each time she was introduced, but mated as quickly with the new female as when the first female was first introduced.

But how do they tell them apart? Female crickets avoid mating again with the same male by scent-marking them. Reasoning that two inbred brothers probably smell somewhat similar, the authors tested whether males readily mate with a female who had previously mated with their brother. They do, so scent-marking doesn’t seem likely.

On the other hand, males seemed somewhat reluctant (a slight delay, not statistically significant) to mate with the sister of a previous mate. Sisters could resemble each other in various ways, but biometric identification in insects is usually based on odors, rather than face recognition, for example. So the authors doused novel females with essence-of- previous-mate (extracted in solvent from four of her inbred sisters). These females were avoided, relative to control females doused only in the solvent.

Male burying beetles are less likely to mate with the same females repeatedly, apparently because remembered smells -- how long can they remember a smell? -- make them less interested. Before switching perfumes, remember that humans are not burying beetles.

May 10, 2008

Regulation of sex ratios in plants

“Under drought conditions,? says Bänziger, CIMMYT’s director for corn research, “the maize plant puts more resources into pollen formation and less into seeds.? From the plant’s point of view this makes sense. Pollen is much cheaper energy-wise for the plant to make, and, if the pollen manages to fertilize another plant’s seed, the drought-afflicted parent will still contribute 50% of its genes to the offspring. But this is of little help to farmers, who sell kernels, not pollen." -- Nature 452:273

Maize plants are hermaphrodites, having both male (pollen-producing) and female (seed-producing) flowers. Other plant and animal species have two sexes, such as males and females. From the title, “Density-dependent regulation of sex ratio in an annual plant?, I assumed that this week’s paper (by Marcel Dorken and John Pannell, published in American Naturalist) would be about how parent plants adjust the male:female ratio in their offspring, a topic I have discussed previously.

But no. Mercurialis annua is stranger than that. Its two “sexes? are male and hermaphrodite.

When a hermaphrodite plant pollinates itself or is pollinated by another hermaphrodite plant, all of its seeds grow up to be hermaphrodites. When a hermaphrodite is pollinated by a male plant, half its seeds are hermaphrodites and half are male.

Male plants don’t produce seeds, so an isolated population of only male plants will die out. An isolated population of only hermaphrodites will stay hermaphrodite-only. Real-world populations tend to be between 0 and 40% male. Maleness is genetically determined in this species, so it can evolve, changing the frequency of males over generations. But in which direction?

The authors hypothesized that males would become more common at high density, when lots of M. annua plants are crowded together, but less common at low density. One reason is that isolated hermaphrodites (the only sex that produces seeds, in this species) would mainly pollinate themselves, producing more hermaphrodites. At higher population density, however, hermaphrodites would receive more wind-blown pollen from nearby male plants. If the males produce enough pollen to swamp that from the hermaphrodites, up to half of the resulting seeds would be male.

But what if the hermaphrodites change their investment in pollen vs. seeds with changing conditions, as maize plants apparently do under drought? The authors hypothesized that hermaphrodites would produce less pollen at high than at low density. If so, then other plants (some of them males) would be responsible for a larger fraction of pollination at high density. More pollination by males, rather than self-pollination by hermaphrodites, would lead to a higher percentage of male seeds. So high density could lead to more males, in two apparently independent ways.

To test this hypothesis, the authors collected seeds from several locations and grew each batch of seed at high density (closely-spaced plants) and low density (widely-spaced plants). They measured pollen production and the sex ratio in seeds produced.

Consistent with their hypothesis, “hermaphrodites produced 3.6 times as much pollen per unit biomass under low densities.? Most populations also evolved in the predicted directions, increasing in percent male at high density and decreasing at low density.

The paper doesn’t seem to explain why wider spacing would increase pollen production by hermaphrodites. I can see that this would be beneficial if a hermaphrodite needed to supplement its neighbors’ pollen with more of its own, but wouldn't closer neighbors also increase the success rate for any pollen it sent drifting on the wind? Also, how can plants tell how many other members of their species are nearby? Are they just responding to the amount of nonself pollen they receive, or are other signals (perhaps gases) being exchanged?

Your homework assignment is to explain the implications of these findings for human sexual morality. Blank sheets of paper will receive full credit.

May 7, 2008

Real vs. fake controversy

I liked this essay comparing areas in evolutionary biology where there is genuine controversy -- i.e., where people who are actually collecting data and publishing on a topic disagree -- vs. the phony controversies imagined by creationists. Group selection may still almost qualify as a controversy, a question I may address in a later post, but age of the earth, common ancestry of all species (at least those studied so far!), and the power of natural selection to solve difficult problems are not at all controversial among those actively publishing on related topics.

The question of how much exposure high school students should have to genuine scientific controversies seems a bit more complex to me. I agree that helping students get enough of the basics to understand active controversies in any depth is a big challenge. On the other hand, I've been amazed how many high school students (and their parents) think that the only definition of "research" is looking up information in a library or on the web. If we want students to understand that scientific research is an exciting, ongoing activity, some kind of exposure to areas where scientists disagree seems essential. Areas of research that are easier to understand, like the mindless screening of drugs, don't convey the intellectual excitement of real science.

Here's a seminar class I've thought about for either high school seniors or first-year college students. First, let's set the minimum standard for a scientific controversy as: at least two conflicting points of view, each represented by data-containing papers from at least two nonoverlapping groups, in journals with an impact factor of at least 1.0. Each week we consider one question, such as:
1) What causes AIDS?
2) What is killing amphibians around the world?
3) How old is the earth (within 10%, say)?
4) What living species is the closest relative of chimpanzees?
Students get points for showing that each topic was controversial, at least at one time, with a big bonus for whoever shows controversy most recently. Then we could make a time-line, showing when each question was settled (pending new data, of course!).

May 3, 2008

Sharing diseases with relatives and neighbors

Not enough people voted on the Reader’s Choice, so this week’s paper is “Phylogeny and geography predict pathogen community similarity in wild primates and humans? by Jonathan Davies and Amy Pedersen, published in Proceedings of the Royal Society.

Many humans diseases, from flu to AIDS, come from other species. Similarly, diseases from dogs are an increasing threat to lions, while cat diseases kill sea otters. Are there general rules that predict how likely two species are to share diseases?

To find out, the authors analyzed several large data sets on diseases of humans and 117 other species of primate (apes, monkeys, etc.). They hypothesized that species are more likely to share diseases if they live near each other and/or if they are more closely related, that is if they share a more recent common ancestor. This is similar to how we define relatedness in humans: brothers and sisters have more recent common ancestors (parents) than cousins do (grandparents). Fortunately, the family tree for primates is relatively uncontroversial, at least among scientists.

Overall patterns were consistent with both of their hypotheses. For example, pairs of species whose last common ancestor lived 20 million years ago were more than twice as likely to share diseases, relative to less-related pairs of species, whose last common ancestor lived 40 million years ago. Within each relatedness class, pairs of species were about three times more likely to share diseases if they live near each other.

How can species that don’t live near each other share the same disease-causing microbes? Maybe they lived near each other some time in the past, or maybe the microbes were spread by species that travel widely: migrating birds, say, or ecotourists. Wash the soil off your boots before you travel to the Galapagos, OK?

For protozoa (microbes bigger and more complex than bacteria), host relatedness was most important. But for viruses, host relatedness was less important. Because viruses evolve faster than protozoa, they can more easily adapt to new hosts.

Guns, Germs, and Steel suggested that the European diseases that killed so many Americans after 1492 originally came from livestock. Consistent with the results in this paper, these diseases were mostly caused by viruses, which could more easily adapt to new hosts.

Cows and sheep are much less related to humans and chimps than humans and chimps are to each other, but I've never gotten a clear answer from creationists as to where they think the boundaries of "created kinds" lie. Since they like to caricature evolution as a frog "turning into" a cow -- an even larger change that would take millions of generations -- maybe apes, or even mammals, are all the same "created kind"?

Also this week:

Environment-contingent sexual selection in a colour polymorphic fish

Encoding choosiness: female attraction requires prior physical contact with individual male scents in mice

Subtle cues of predation risk: starlings respond to a predator's direction of eye-gaze

Local resource competition and local resource enhancement shape primate birth sex ratios

May 1, 2008

Science fair secrets 3: The $250 science lab

This is part of a series (copyright R Ford Denison) on the secrets of winning science fair projects. Click "science fairs" under Categories (at right) for more.

It is quite possible to do good experimental science fair projects using only everyday materials (rulers, paper cups, etc.). However, a small investment in inexpensive scientific apparatus can greatly expand the range of feasible experiments. For a fraction of the cost of a desktop computer, you can measure weight (mass), volume, temperature, acidity (pH), and light, all with sufficient accuracy to generate useful data. Unlike a computer, this equipment won't be obsolete in two years, or in twenty. These prices are old, so it might be a $275 science lab by now. On the other hand, these are all new prices; used would be cheaper. Items earlier on the list are most widely useful. Add an inexpensive microscope and you'll be about as well-equipped as Darwin. Aside from the boat, gun, greenhouse, and assistants, of course.

Compare with this much more ambitious home lab. Before spending that kind of money, I would wait and see what direction my research was going.


Triple beam pan balance (600 g x 0.1 g).......$ 93.00
Graduated cylinders (100 mL).................2 for 11.00
Multitester (use with sensors below)..............24.99
Mini-hook adaptors for above.............................2.59
Thermistors, for temperature (2 @ 1.99)...........3.98
Photocell assortment, for light............................1.98
Red-fluid thermometers (2 @ $6.50)................13.00
Acid/base pH indicator paper...........................13.30
Range extension set for balance (2 kg) ........24.95

Safety alcohol burner......................................16.80
Ring stand (support rod)..................................14.40
3" ring for above................................................8.10
Small object clamp for above.............................8.30

Total:...........................................................$ 246.39

Sources: Edmund Scientific, Radio Shack, Fisher

An example of an experiment you could do with only the above, plus a metal or Pyrex container from the kitchen, would be a comparison of the heat output of two different alcohols.

1) Put the alcohol burner on the balance, to measure fuel use as loss of weight.
2) Put the container (with water) above the burner, on the 3" ring on the ring stand.
3) Suspend a thermometer in the water, using the small object clamp.
4) Measure the water temperature as it heats up, and graph it over time. (While you're at it, stick a thermistor in there and measure how it's resistance changes, using the multimeter; plot these data together with the thermometer temperature and you've got a calibration curve for the thermistor, for future use. An extra data graph in your lab notebook, or even in your display, never hurts!)
5) Repeat with the second fuel (ethanol vs. propanol, say). Remember to test each fuel at least twice.

The slope of the temperature vs. time curve is proportional to the heat from the burning fuel. Each 1 C increase in temperature requires one calorie of energy from the fuel for each gram of water in the container. As a second measure of heat output, measure the rate of evaporation, once water reaches boiling. Each g evaporated takes 540 calories.

I plan to use this example again in a future Science Fair Secrets entry on how a little simple math can really add up.