THE EVOLUTION OF SEX Sexual reproduction is so common we sometimes forget that many organisms reproduce asexually. That two fundamentally different modes of reproduction exist raises the question of the adaptive significance of sex and the need to explain why it is so common. We can assume that the first living organisms reproduced asexually since evidence of the specialized reproductive structures enabling sexual reproduction is found only in Cambrian fossils. Pre-Cambrian species most likely reproduced through mitosis. Because the origin of sex was an historical event without witnesses, we cannot know for sure when or why sex evolved. On the other hand, the question of why sexual reproduction persists can be studied by examining the conditions which seem to favor sexual over asexual reproduction today. By using the principle of uniformitarianism we can then hypothesize that this answer explains the conditions which led to the evolution of sex in the first place. Both forms of reproduction provide fitness benefits in terms of offspring which contribute one's genes to the next generation's gene pool, but sexual reproduction has costs associated with it which are not incurred by individuals which reproduce asexually. So substantial are these costs that they put sexual reproducers at a clear selective disadvantage in reproductive competition with asexually reproducing individuals. To resolve the problem of why sex persists today in such abundance we must examine these costs and attempt to find benefits great enough to overcome them. A satisfactory answer has yet to enjoy wide acceptance in the scientific community, so the reason why sexual reproduction is so common remains a major evolutionary puzzle. Costs of sexual reproduction The basic difference between sexual and asexual reproduction is that (1) sexual reproducers produce gametes containing only half their genes and (2) offspring result from syngamy or the fusion of two different gametes. This process imposes the costs described below. 1. Cost of meiosis. Some species of fish and lizards exist as all-female populations which reproduce parthenogenetically, i.e., without any genetic contribution from males. These species produce eggs through the process of mitosis which develop into females genetically identical to the parent. Parthenogenesis thus contrasts sharply with sexual reproduction wherein each parent contributes only one-half its genes to offspring. This cost of reduced genetic representation in each offspring is called the cost of meiosis and only exists as a liability if both sexual and asexual parents produce the same number of offspring per unit resource. Should the sexually reproducing individual produce twice as many offspring per unit resource as its asexual counterpart, the cost of meiosis would be zero. Since in the process of oogenesis most of the cytoplasmic content of the primary oocyte ends up in a single gamete, we can assume that the energetic cost per unit resource of producing eggs through mitosis and meiosis is the same. Thus, the cost of meiosis is a real cost imposed genetically on sexually reproducing females. 2. Cost of anisogamy. During meiosis four sperm are produced for every egg, but only one sperm is required to produce a zygote. The remaining three sperm are effectively wasted. This wastage of sperm imposes an additional cost on sexual reproduction - the cost of anisogamy or the cost of producing males. If we assume that the energetic cost of producing one egg equals the cost of producing four sperm through meiosis, then a population of asexually reproducing females would be far more efficient in converting resources into offspring than would a population of sexual reproducers. This being the case, why males exist at all poses a conundrum. 3. Cost of recombination. Recessive deleterious mutations express themselves only in the homozygous condition. Sexually reproducing heterozygotes run the risk of producing offspring less fit than themselves when syngamy results in recessive homozygotes. The only way an asexual heterozygote can produce offspring homozygous for recessive alleles is through mutation. The normal process of sexual reproduction regularly produces offspring less fit than their parents and this cost is called the cost of recombination. In an environment favorable to the genotype of the parents (otherwise they would not be reproducing!) asexually-produced offspring, genetically identical to their parents, will survive and flourish. The same cannot be said for sexually reproduced offspring. Gametic variation will inevitably produce some less-adapted offspring. 4. Cost of mating. The final cost of sexual reproduction arises when a sexual reproducer seeks a suitable mate. Courtship activity is costly both in terms of time and energy expended, and it often increases the risk of mortality through exposure to predators and inhospitable physical components of the environment. This cost of mating, however, is not intrinsic to sexual reproduction because some aquatic species simply release gametes into the water with little or no cost to themselves. Benefits of sexual reproduction For sexual reproduction to persist the benefits must be greater than the costs described above. Furthermore, since sexual reproducers are in competition with asexual reproducers, the benefits of sexual reproduction must not only exceed its costs, the difference between benefit and cost must be greater than the benefits of asexual reproduction. A number of potential benefits have been hypothesized for sexual reproduction and we will discuss these below along with the type of selection which promotes these benefits. Fisher-Muller hypothesis. One of the earliest explanations given for the advantage of sexual reproduction is that sex is an adaptation to cope with long-term fluctuations in the environment. The variation generated by sexual reproduction enables a population to persist when the environment changes drastically (as it is sure to do sooner or later). Asexual populations require mutations to adapt them to new environmental conditions - a most unlikely happening given the low frequency and random nature of spontaneous mutation. There are two problems with this explanation. First of all selection acts on the here and now and not as a means of adapting to future environments. This explanation is possible only if the potential for future adaptation (preadaptation) is a by-product of variation adaptive to the present environment. The second problem is that the advantage of sex explained by this hypothesis accrues to the population, not to individuals whose variable offspring will not all be adapted to the existing environment. Because the Fisher-Muller hypothesis describes sex as giving a long-term advantage to populations at the expense of short-term disadvantage to individuals, it requires interdemic (group) selection rather than natural selection as its vehicle of evolution. Best man hypothesis. The Fisher-Muller hypothesis might explain why separate populations of sexual and asexual reproducers exist, but it cannot explain the permanent stable coexistence of both forms of reproduction in the same population. Many species which live in temporary environments reproduce asexually when conditions are favorable, but switch to sexual reproduction when the environment changes, e.g., in the spring and summer water fleas (Daphnia) reproduce parthenogenetically but lay sexual eggs in the fall which don't hatch until the following spring. Presumably, sexual eggs provide genetic variability to buffer any environmental change the following year. At least some eggs will be adapted to a different environment in the following spring and these will reproduce asexually until the next fall. The best man hypothesis, developed by George C. Williams, suggests that sex confers an advantage on individuals which live in environments which vary temporally over short periods of time. The problem with this hypothesis is that it simply is not consistent with the facts. In species with both sexual and asexual forms, the asexual forms prevail in temporally-varying environments, e.g., temperate regions, while the sexual reproducing forms are found in more stable habitats, e.g., the tropics. A further problem with this hypothesis is that while it can explain why sex might be advantageous to r-selected species, it cannot explain the existence of sex in K-selected species. Recall that r-strategists have such a high reproductive potential that only a few variants need survive to rebuild the population. K-strategists, on the other hand, produce so few offspring that even if a few were to survive in a new environment, their chances of rebuilding a population are very remote. Red Queen hypothesis. This hypothesis attempts to correct the deficiencies of the best man hypothesis while retaining its emphasis on natural rather than group selection. In particular, the Red Queen hypothesis explains sex as an adaptation to temporally-varying biotic rather than abiotic environments. Pressures imposed by competitors and predators require continual coevolution, thus a population must evolve continually just to stay in existence (hence the reference to the Red Queen). Biotic pressures are most common in the tropics, so one would expect sexually reproducing forms of a species to be most common here. Their asexually reproducing counterparts should be found in temperate and polar regions where biotic pressures are less important. Since this prediction agrees with the observed distribution of sexual and asexual reproducing populations, it is more appropriate than the best man hypothesis. A version of this hypothesis which specifies pathogens as the selective agent behind the continued existence of sex has considerable merit because it predicts that sex should be found only where pathogen pressure is high, i.e., warm, moist environments in which population density is high so that the pathogens can spread easily. This pathogen hypothesis not only predicts that sex should prevail in the tropics, but also that males should be rare in tropical environments with low density populations, e.g., islands, as well as in warm, arid environments (deserts) and cold, moist polar regions. The Red Queen hypothesis has the added advantage of explaining why sex is of value to K-strategists such as ourselves. Tangled bank or sibling competition hypothesis. This hypothesis, advanced by Graham Bell, explains the advantage of sex as an adaptation to spatial rather than temporal heterogeneity. In highly structured environments, e.g., tropical rain forests, the environment consists of a mosaic of structural features which favors the survival of diverse offspring, each of which is adapted to a slightly different environment. This spatial diversity prevents or reduces sibling competition. In homogeneous environments, e.g., the tundra, siblings would be in direct competition with one another and only those best adapted to that structurally simple environment would survive. Since asexually reproducing species cannot take advantage of the widespread structural diversity of heterogeneous environments, their offspring are exposed to intense sibling competition in the one facet of a structured environment to which they are all adapted. In any given habitat, the asexual form might eliminate the sexual form, but the sexual form would be better able to exploit spatially diverse habitats by virtue of its genetic diversity. Hitchhiking hypothesis. The last three hypotheses discussed explain the advantage of sex in terms of a fitness gain by individuals which reproduced sexually. The hitchhiking hypothesis suggests that sex confers no advantage on the individual but instead is perpetuated through genic selection. Suppose gene A controls sexual reproduction and gene B controls asexual reproduction. Since genes reside on chromosomes, the adaptive value of genes A and B is linked to the phenotypic effects of other genes that share their chromsome. Because gene A can change the allelic combinations of the loci with which it is linked, it can produce new gene combinations which could have a higher fitness effect on the phenotype. Gene B, on the other hand, is stuck with the alleles existing at the other loci and so would lose in competition with gene A. According to this hypothesis sex is favored because it promotes the reproductive potential of gene A and not of the individual possessing gene A. The problem with this hypothesis, other than the fact that it relies on genic rather than natural selection, is that it might explain why gene A would have an advantage by producing a small amount of recombination, but high rates of recombination would produce more unfit combinations (high cost of recombination). Speciation hypothesis. So far we have viewed sex as the product of genic, natural and group selection which evolved because of its advantage to genes, individuals and populations, respectively. The speciation hypothesis suggests that sex is non-adaptive with respect to these units and persists only because it promotes a high rate of speciation. By producing diversity sex allows for existence of genetically different populations which, if they become isolated, could become new species. Given a clade consisting of both sexual and asexual species, the sexual species would be able to speciate much more rapidly and so eventually replace their asexually reproducing relatives through species selection. Thus, according to this hypothesis, sexual reproduction is more common than asexual reproduction because it contributes to a high rate of speciation and not because it confers any adaptive advantage on genes, individual or demes. We can summarize the proposed benefits of sexual reproduction by contrasting the different hypotheses with respect to the environmental attributes to which sex is supposedly adapted and the type of selection involved. Environmental attribute Type of selection Hypothesis Temporal variation long term interdemic Fisher-Muller short term (abiotic) natural Best man (biotic) natural Red Queen (Pathogen) Spatial variation natural tangled bank Non-adaptive genic hitchhiking species speciation GOULD'S HIERARCHICAL MODEL OF SELECTION All matter is organized structurally into a hierarchy with many small units grouped into successively fewer and more inclusive functional groups. Thus, subatomic particles are grouped into a smaller number of atoms which in turn are packaged into an even smaller number of molecules. With respect to life, molecules are grouped into cells, cells into organisms, organisms into populations and populations into species. As one ascends this structural hierarchy, the organizational units become both less numerous and more complex. In studying nature's hierarchy scientists recognize two possible viewpoints regarding the significance of this arrangement: reductionism which views the whole as nothing but the sum of its parts, and holism (antireductionism) which views the whole as unique and greater than the sum of its parts. The physical scientists tend toward reductionism and this view is considered by many to be the goal of science: reduce nature's complexity to the fewest possible general explanations. Many life and social scientists, however, find little value in this approach and are impressed with the new properties which emerge at the higher structural levels they study - which properties cannot be easily explained from a knowledge of processes acting on the lower units. These new properties emerge, not as a consequence of some spiritual force (vitalism), but rather as a consequence of interaction among the smaller structural units which comprise the whole. Holists argue that one cannot reduce the whole to the sum of its parts because simply understanding the processes which act at the lower level will not enable one to predict the emergent properties which characterize the whole and result from an interaction among its parts. Edward O. Wilson, the founder of sociobiolgy, suggests that scientific study should incorporate both views and that the well trained scientist should be familiar with three different disciplines: the discipline in which the scientist is specialized, the discipline which studies the next lower structural level (antidiscipline), and the discipline which studies the next higher structural level for which his/her discipline serves as the antidiscipline. Thus, the well trained biochemist should be familiar not only with biochemistry, but also with the fields of organic chemistry (biochemistry's antidiscipline) and molecular biology (the field for which biochemistry serves as the antidiscipline). In a trenchant analysis of Darwin's theory of natural selection ("Darwinism and the Expansion of Evolutionary Theory" in Science, vol. 216, 23 April 1982) Stephen Jay Gould argues that Darwinism as a reductionist explanation of evolution is incapable of explaining fully the natural world. By focusing on the individual as the unit of selection Darwinism cannot provide an adequate explanation of such phenomena as altruism and the prevalence of sexual reproduction. In so doing he adopts a holist position and suggests that a full explanation of evolutionary phenomena will have to include selection acting on other units, including genes (kin selection), populations (group selection) and even entire species (species selection). It is important to note that he is not claiming that natural selection is unimportant (on the contrary, he posits that natural selection acting on individual variation does explain much of nature); he simply suggests that natural selection is too restrictive in focus as an evolutionary mechanism. He emphasizes Darwin's genius in arriving at the concept of selection, but argues that this concept must be expanded to include other structural units to correct the limitations of natural selection theory. Darwinism has taken such hold on modern evolutionary biology that everytime an alternative form of selection is proposed as an explanation for a certain phenomenon, it is opposed on the grounds that natural selection can provide an equally valid explanation. Consequently, alternative forms of selection are dismissed as either not operating in nature or of minor importance as regards frequency in explaining evolution (theoretically possible but with little practical force). Gould criticizes this mentality by claiming that evolution is too complex to be explained solely by a process which acts only upon individuals in a single population. Alternative units of selection must be considered as viable options, not summarily dismissed by Darwinian reductionists. Only when the concept of selection has been expanded to include several structural units can an important, holistic question be asked: How might different units of selection interact in producing evolutionary phenomena? As a process of differential survival and reproduction selection can operate on genes, demes and even species in addition to individuals. The concept of selection operating at the species level (species selection) has recently been advanced by punctuationalist paleontologists who suggest that some existing species are prevalent, not because their individuals are better adapted than those in other species which became extinct, but rather because certain species in a clade (group of descendent species from a single ancestor) were more prone to the process of speciation and so left more descendent species than their relatives. If extinction is a random process, then the ancestor which leaves more descendent species has a higher probability of its lineage being represented at some future date than has a less prolific clade member. The concept of species selection has a very interesting corollary: species existing today which owe their existence to species selection might contain traits which are not adapted to any environmental conditions. This idea is contrary to the Darwinian view but is a logical consequence of random factors which act during the process of speciation (founder effect, genetic drift, and neutral mutations which occur during the genetic revolution). This is not to say that these species are not adapted, but only that their phenotypes might contain some nonadaptive traits which evolved through chance mechanism associated with the process of speciation. Gradualists, who adopt the Darwinian position, would seek adaptive value in all the phenotypic traits of a species. The main thrust of Gould's paper is the suggestion that a holistic approach to selection might provide insight into some puzzling evolutionary phenomena which cannot be explained through reductionism. Different levels of selection might interact in either a positive manner to reinforce one another, or in a negative manner with one level restricting traits favored by a lower level. Why do some species contain much more nonfunctional DNA in their chromosomes than other species? Gould suggests that selection at the genic level favored the proliferation of genes which contribute nothing to the phenotype (and so are neutral with respect to natural selection) but the process is eventually restricted by natural selection when the amount of such DNA begins to interfere with cellular processes (negative interaction). Alternatively, some phenomena might be very common simply because they are favored at all levels of selection (positive interaction). A case in point is the prevalence of sexual reproduction over asexual reproduction. No theory has yet been advanced to explain to everyone's satisfaction why sexual reproduction is so common since it has costs associated with it which are not experienced by asexual reproducers (e.g., only one half of a sexually reproducing individual's genes are passed on to offspring). Gould suggests that sexual reproduction is favored by genic selection because a gene which caused the chromosome on which it existed to recombine would be able to take advantage of beneficial mutations on its homologous chromosome; by individual selection because variable offspring would be better able to exploit a spatially heterogeneous environment; by group selection because populations which reproduced sexually would contain a sufficient reservoir of genetic variation to cope with a temporally varying environment; and species selection because, by facilitating the speciation process, sexual reproduction would be carried along with lineages represented in the future simply by virtue of their numerical superiority. Hypotheses which explain the prevalence of sex at each level have problems which mitigate against their full acceptance (hence the controversial nature of the problem) but the positive interaction of selection at different levels alleviates these problems and so offers a better explanation. Thus, claims Gould, a hierarchical or holistic view of the selection process would expand the narrow focus of Darwinism and provide greater insight into evolutionary problems that up to now have defied explanation. He predicts that as evolutionary study advances Darwin will be remembered and honored not for his concept of natural selection, but for his concept of selection itself!