Things always tend to thin out towards the end of a conference, but a bunch of us were up bright and early to kick off a special symposium “Moving towards a science of evolutionary prediction”. After some opening words by session organizer Allen Rodrigo, Hamish Spencer led off with a talk entitled “Evolutionary prediction and adaptive landscapes”. He gave a nice overview of the problems and prospects for prediction in evolutionary biology, beginning with the idea that there is a conflict between popular conceptions of prediction and those entertained by evolutionary biologists. The public wants prediction that will presage outcomes; at the extreme some creationists argue that if evolutionary biology cannot make such predictions, it is not in fact a science. Spencer pointed out that most evolutionists do not hold out much hope for this outcome-based mode of prediction, and even argue whether evolution is a repeatable process. At one end of this debate are the ideas expressed through Gould’s “Tape of Life” metaphor, which suggests that repeated playings of the evolutionary process — even when restarted from the same place — would produce different outcomes. At the other end of the debate are evolutionists like Simon Conway-Morris, who argue that there are only a small subset of successful solutions to adaptive problems, resulting in predictable outcomes driven by convergent evolution. After framing this debate, Spencer went on to describe four factors which make evolutionary prediction difficult: contingency, chance, chaos/aperiodicity, and capriciousness.
Based on my experience modeling ecological interactions, I am a bit ambivalent on the prospect of promising prediction based on evolutionary models. There is a major gap between what evolutionary biologists are good at (producing testable predictions based on hypotheses about the evolutionary process) and what the public may want (producing reliable predictions of outcomes). Although I am not sure that I am right, the four sources of uncertainty outlined by Spencer seem to me more challenging than those often faced in making ecological predictions. I didn’t stay for the rest of the talks in this symposium, but I do think some tempering of expectations might be valuable, paired with a continuing improvement of how we educate the general public about uncertainty and risk.
Mutualism is an interesting form of cooperation because it occurs between species, necessitating that traits evolve in two separate populations to allow for mutual aid. This presents fundamentally different challenges than intra-specific cooperation because of the gene flow barrier that exists between species. Kristina Hillesland presented an interesting talk entitled “Rapid evolution of stability and productivity at the origin of a microbial mutualism” that explored how these mutualisms evolve using experimental evolution. She built a clever obligate mutualistic system by pairing two bacterial strains, neither of which could completely metabolize available nutrients in the absence of the other. While this is somewhat contrived (it artificially selects strains and pairs them in a very specific environment), it allows for the exploration of the coevolutionary process. What Hillesland found was that usually the paired strains increase their overall fitness as they coevolve, while occasionally the mutualism collapsed and both strains went locally extinct. She also was able to look at the effect of each coevolved strain on the overall increases in fitness by a clever recombination of ancestral and evolved partners. What was not answered by this study was how this coevolution increased the overall fitness of each strain. How was cheating prevented, both between strains (parasitic exploitation of the metabolic activity of the other strain) and within strains (freeloading off of the metabolic efforts of others)? Perhaps the environment itself eliminates inter-strain cheating, as both die if one ceases to do its metabolic part, but given that the benefits are diffuse (to my knowledge bacteria are not directly reciprocating from individual to individual) it is a bit of a quandary as to why cheating does not evolve within one of the strains. It may be that the two out of twenty-four coevolutionary arenas that collapsed are indicative of just such a tragedy of the commons scenario, and it seems like sampling of individuals during the latter phase of the decline could help detect cheating. What’s interesting is that there does not seem to be any population structuring going on in these fitness increases, suggesting that selection on individuals alone accounts for the observed fitness increases. How this happens (what the source of selection is) should be a direction of future work in such systems.
Another great social evolution talk entitled “Experimental coevolution between two strains of the social amoeba Dictyostelium discoideum” was given by Brian Hollis. “Dicty” is an appropriately-famous amoeba that can be social, aggregating to form a “slug” that can travel significant distances. One of the most extreme social behaviors that Dicty displays is the formation of a fruiting stalk. This stalk forms in order to facilitate the release of spores from the fruiting body and escape from an unfavorable environment. But it also represents a form of cooperation, as those individuals who form the stalk fail to disperse. Cheating has been detected in Dicty, manifested as a over-representation of particular strains within the fruiting body (relative to the stalk). Hollis wanted to know how this coevolution might occur, so he established two protocols to allow for the coevolution of two strains of Dicty, one of which was transgenically induced to produce green fluorescent protein (GFP). One protocol allowed for serial coevolution by cycling between social growth and fruiting stages, allowing only those individuals who made it into the fruiting body to enter the next experimental arena. The second protocol attempted to select for cheating by combining surviving spores from one strain (subjected to coevolution with other strain) with individuals of the other strain grown in monoculture. Because of the GFP being produced by one strain, the systems could be assayed for relative population densities during the growth and fruiting stages using a flow cytometer. Although Hollis has not completed the full experiment, his early results seem to indicate that their is competition and exploitation going on in both the vegetative and social growth stages.
If social evolution (in particular the evolution of cooperation) has been my favorite topic of this Evolution meeting, sexual selection and the evolution of sex have been a close second. Today I heard a number of interesting talks in this area, including Laurie Stevison’s “Male Drosophila influence crossover rates in females”. Apparently recombination rates can vary both within the genome of an individual (from chromosome to chromosome) and within a population (from individual to individual). Stevison was interested in asking whether males, who are apparently somewhat limited in their ability to recombine their chromosomes and therefore may be limited in their ability to produce diverse offspring, are able to induce females to recombine more readily. I was not entirely clear on this point, but I am guessing that males may exert a post-copulatory influence on female meiosis, which is possible because Drosophila females store sperm. Stevison did find that male treated with hormones as well as males from different global subpopulations show distinct variation in their effect on the recombination rate of clonal females. Although no mechanism has been discovered to explain these trends, they suggest an interesting new area for exploring the nature of sexual conflict.
Amanda Izzo presented a really neat talk entitled “Females gain direct benefits in a non-economic, lek-based mating system”. Her work focuses on the potential economic value of female choices that are often considered non-economic. When males offer some resource to the female in exchange for mating opportunity, female choice will be selected for accuracy in assessing resource value (so-called “economic matings”). But when males do not offer an assessable resource, females must choose based on indirect signals of value. Traditionally it is assumed that females are selected to be attracted to proxy traits that indicate good genes, but Izzo believes that sometimes these proxy traits may actually advertise male resources. Specifically, she believes that female Polistes wasps (who overwinter before founding a new nest in the spring) receive benefits from drone males in the form ejaculate compounds. Ejaculate compounds have been implicated in sexual conflict as a means of manipulating females, but Izzo suggests that these compounds might also be a means of provisioning females. This makes sense because males perish before winter, so anything they can donate to the females that would increase offspring viability or abundance would be of potential value to male reproductive fitness. In previous work Izzo has shown that male wasp coloration patterns are an environmentally-mediated indicator of male condition and that females preferentially mate with high-condition males. The work she presented demonstrates that mating in general and mating with high-quality males in specific confers overwinter survival benefits on females. The next step is to analyze the ejaculate of males of differing condition to identify a candidate compound or compounds which might be providing this advantage to females.
Bart Pollux gave a thought-provoking talk entitled “Sexual selection and evolution of the placenta in the fish family Poeciliidae”. Fish in this family mostly provide their young with nutrition through a yolk (lecithotrophy) but some females have evolved a placenta, allowing them to directly and incrementally feed their gestating embryos (matrotrophy). The placenta has independently evolved in many lineages of these fish, suggesting strong convergence. Pollux was interested in whether matrotrophy, which allows for parent-offspring conflict, affects the intensity of sexual selection. Using a large-scale survey of the various species in this family, he found that males in matrotrophic species are less likely to be dichomatic, more likely to be substantially smaller than females, and more likely to have a longer gonopodium. These findings all indicate that pre-copulatory sexual selection is less prevalent in the placental species.
Marc Johnson presented a really compelling story in his talk “Life After Sex: The Evolutionary Consequences of Losing Recombination and Segregation in Plants”. He started with a comprehensive overview of sexual reproduction, which often seems to be ubiquitous, setting up a “paradox of sex” wherein it is difficult to quantify the benefits that outweigh more obvious and substantial costs of mating with another member of one’s species. A question that still has not been clearly answered is “what benefits do recombination and segregation provide?”. Presumably asexual organisms pay two major costs for forsaking sex: the gradual loss of adaptive genes through mutation (Muller’s rachet) and/or a reduced rate of adaptation to novel environments. Given the predominance of sexual reproduction in nature, it seems as though these costs of asexuality are high, and some have suggested that asexuality may be an evolutionary dead end. What Johnson pointed out was that although there’s been lots of conjecture about the origin and maintenance of sexual reproduction, few systems allow for real hypothesis testing. He works in a system that allows for a comparative approach to determining the costs and benefits of sexual and asexual reproductive strategies, the evening primrose. Of the 260 species of evening primrose, 84% are sexual reproducers while the remaining 16% reproduce asexually. Interestingly, those that are asexual are also diploid, as they still reproduce through the fusion of sperm and egg (not via apomixis, the plant equivalent of parthenogenesis). The asexual primroses maintain a permanent translocation heterozygosity (PTH) due to a mutation which causes a meiotic ring during gametogenesis; this leads to genetically identical sperm and eggs and clonal offspring. Johnson took advantage of the presence of this genetic quirk to compare how herbivory affects the sexual and asexual primrose species. In both laboratory and field experiments, he found that the asexual species were more vulnerable to generalists herbivores. But when he tested these same species by exposing them to specialist herbivores, he found that the asexual species were more resistant to herbivory. He seems to have discovered a classic tradeoff between the benefits of sexual and asexual reproduction: Johnson speculates that the mutational load experienced by asexual species is actually an advantage against specialists, which often exploit secondary defensive compounds. In periodically losing their ability to make these compounds, asexual species are more vulnerable to the generalists but less vulnerable to the specialists.
My class in Human Evolution tackles the issue of female pelvic width evolution, so I was excited to hear Heidi Schutz’s talk “Functional trade-offs and pelvic sexual dimorphism in mammals”. In humans there is pelvic sexual dimorphism, with females having wider pelves than males, presumably due to a tradeoff between offspring cranial size and locomotor efficiency/durability. Schutz was interested in the question of whether or not pelvic dimorphism in carnivorous mammals correlates with a variety of factors including means of locomotion, mating system, and dimorphism in body mass. Using a model selection methodology, she was able to connect the degree of dimorphism to both male/female body mass ratio and mating system. Interestingly, the relationship between pelvic dimorphism was negative for multimale/multifemale mating systems but positive for polygynous (harem) mating systems.
Julie Meachen-Samuels gave an interesting talk entitled “Convergent evolution of the prey-killing arsenal of saber-tooth predators”. Saber-toothed carnivores differ from modern cats in the shape of their teeth: as the name suggests, their teeth were thin and oval in cross section whereas non-saber-tooth cats have conical teeth. Meachen-Samuels wanted to test the hypothesis that different teeth morphologies correspond with different prey capturing tactics. Whereas living felids kill their prey with a suffocating chokehold to the neck that only requires moderate grappling of the prey, saber-toothed carnivores are thought to have delivered a throat tearing bite after knocking down their prey, bleeding the victim to death. The chokehold technique requires strong teeth, as the bite itself restrains the prey; tearing the throat of one’s victim requires weaker but sharper teeth and strong forelimbs to restrain the prey. This leads to the prediction that the musculature of saber-tooth forelimbs should be greater than that of cats with conical teeth. Interestingly, the saber-tooth strategy is convergent, having independently evolved both within and outside of the felid family. This allows for multiple comparisons between saber-toothed and conical-toothed predators that rule out phylogeny as a cause of trait similarity. In looking at a number of fossil markers that indicate degree of musculature, Meachen-Samuels did indeed find that saber-toothed carnivores have greater forelimb strength than their conical-toothed relatives.
I was able to attend this meeting thanks in part to funding from the Pratt Institute Faculty Development Fund.