Introduction

The basic principles of Mendelian Genetics and almost all those of population genetics are based upon the assumption that an individual's genotype is formed by genetic information drawn from two different sources, with each source making an approximately equal contribution.  In other words, that the mode of transfer of genetic information is via some form of sexual reproduction.

This assumption is certainly valid because sexual reproduction is so prevalent, and, in most cases, the mode is biparental.  Clearly, the prevalence of sexual reproduction would argue that it must be useful.  Therefore, it is often surprising to those outside of evolutionary biology that there are serious questions regarding the evolution of sexual reproduction.

It will be helpful to review the basic features of sexual reproduction before we address the evolutionary issues.  Before we begin, however, you need to keep in mind that sexual reproduction may be very different for prokaryotes and eukaryotes.  Genetic recombination for prokaryotes often involves the unequal transfer or genetic information, and it may be further complicated by plasmid vectors.  Discussion of prokaryotic sexual reproduction should include these qualifications.

Sexual reproduction in eukaryotes involves the process of Meiosis at some stage.  Furthermore, Meiosis includes a number of phenomena that are part of the principal issues regarding sexual reproduction.  First, Meiosis often includes the phenomena of synapsis and chromosomal recombination (also known as "crossing over") during Prophase I.  This means that homologous DNA sequences become closely associated and that blocks of genetic information are redistributed among chromosomes.  Thus,  there is the opportunity for both the "comparison" and rearrangement of genetic information from two sources.

The next feature of Meiosis is Independent Assortment.  This happens when the homologous chromosomes are separated, with the only restriction being that a member of each homologous pair is represented in the subsequent cells.  The result is that, with the exception of chromosomal systems such as translocation heterozygotes, chromosomes are reshuffled every generation.
These are the basic aspects of sexual reproduction.  What, however, is the problem with sex?  The answer comes when we examine the ramifications of sexual reproduction for individual
fitness.
 

The Costs of Sex

The costs of sexual reproduction are considerable when they are examined in detail.  Their total effects on individuals must be counterbalanced by significant benefits in order to explain both the origin and prevalence of sexual reproduction.

Sexual reproduction takes time at the cellular level.  For example, the process of Meiosis takes considerably longer than two mitotic divisions (the cellular equivalent of the number of meiotic divisions).  Also, the processes of syngamy (gametic fusion) and karyogamy (combining of gametic nuclei) require a significant period to accomplish.  An asexual lineage could complete several division cycles - doubling its population with each cycle - in the time that it would take a sexual lineage to complete a single sexual cycle.

The additional costs of sexual reproduction are more obvious when examined with reference to multicellular species, but the same arguments may be applied to unicellular organisms.  The major features of Meiosis (chromosomal recombination and independent assortment) mean that allelic combinations may be disrupted on a haphazard basis both within and between chromosomes.  Thus, it is likely that high fitness genotypes will be destroyed by the process, and that the mean fitness of the offspring will be lower.  This problem is only compounded by the overall mating process.

The existence of sexual reproduction introduces additional costs.  For example, individuals must find and/or attract mates, and this results in additional expenditures of time and energy.  At the same time, individuals may encounter predators, and the contact between individuals during mating is an ideal mechanism by which parasites and pathogens may be transmitted.

The cost of sexual reproduction that is most often cited is found in those species characterized by anisogamy, in which there are two distinct classes of gametes (e. g., eggs and sperm), and they are often produced by two distinct classes of individuals - males and females.  This cost has been termed the "cost of meiosis", "the 50% cost", or the "cost of genome dilution".  This disadvantage of sexual reproduction may be summarized by the phrase "males are useless" (some of you may have reached this conclusion already).

The idea is really quite simple.  It is based on the observation that the sex which invests in the cytoplasmic sustenance for the embryo (the egg producing females) invests far more than the sex which only provides chromosomes (the sperm producing males).  In other words, females produce offspring, while males only contribute their genes.  Furthermore, the sexually reproducing female contributes only 50% of her genes to the egg, and this contribution is "diluted" in its influence by the male contribution to the offspring.

The best way to examine this cost is to view it in demographic terms.  Suppose you have an asexual population of 100 individuals in the same environment as a sexual population of 100 individuals, but with a 50:50 sex ratio.  All of the asexual population can produce offspring, but only 50% of the sexual population.  The sexual population would appear to be at a 2:1 demographic disadvantage.

Obviously, this demographic disadvantage may be offset by increased fecundity in the sexual species, or reduced fecundity in the asexual species, or some form of parental investment by the males.  In fact, some combination of these possibilities have been observed in comparative studies.  It is unclear, however, if such differences can completely offset the disadvantages of sex, and there is no way to compensate for the dilution of the parental genotypes.
 

The Benefits of Sex

At this point you have undoubtedly concluded that cloning is the best option for you, and it may be under specific circumstances.  The widespread occurrence of sexual reproduction, however, argues that it must have value at both the individual and population levels.  The important thing to keep in mind is that few, if any, of the hypotheses discussed below are mutually exclusive, and the history of Biology is full of cases where a combination of factors best accounts for the existence of a particular phenomenon.  Having made these points, we will now look at the major hypotheses that have been proposed to explain the prevalence of sexual reproduction.

"For the Good of the Species"

Perhaps the oldest explanation for the prevalence of sexual reproduction is based on the fact that one thing that sexual reproduction clearly does is that it results in genotypic variation among offspring.  Sex guarantees that every generation will be a diverse sample of all the genotypes that could have been produced by the previous generation.  Presumably this genotypic variation will enable the population to respond to future selection pressures.

Sexual reproduction can be especially beneficial in finite populations.  For example, the combined effects of directional selection and genetic drift can create a high level of linkage disequilibrium.  This will reduce the number of potential genotypes and inhibit further response to selection.  It is a common observation in artificial selection studies that a relaxation of selection for a period can be followed by further, significant response to selection, and most of this "new" variation is due to sexual reproduction.

The problem with this proposed advantage of sexual reproduction is that the benefits accrue at the level of the population or species, while the costs are carried by the individual.  This argument is basically a group selection argument in which individual and group selection are opposed.  Theoretical studies indicate that group selection is possible, but that its effect must be strong in order to counteract the more immediate effects of selection on individuals.  To put it simply, the advantages of genotypic variation through sexual reproduction would appear to be insufficient to prevent the increase of an asexual mutant.  It should be noted, however, that comparative studies indicate that asexual species are more recent, have shorter durations as extant species, and have restricted ranges when compared to closely related sexual species.

"Muller's Ratchet"

Another proposed advantage of sexual reproduction has population level aspects, but it emphasizes one feature of sexual reproduction and its effect on genetic variation.  This hypothesis was articulated by H. J. Muller, but additional ramifications were explored independently by R. A. Fisher.  Some authors refer to it simply as Muller's Ratchet, while others refer to it as the Fisher-Muller Model.

Muller was interested in mutations, their causes, phenotypic effects, and impact on the average fitness of the population.  It is not surprising, therefore, that he saw one advantage of sexual reproduction in terms of its effect on the "load of mutations" in a population.

The concept of Muller's Ratchet is really quite simple.  Asexual lineages will accumulate mutations independently of each other.  Some lineages will have few, if any, mutations, while others will carry a greater number of deleterious mutants.  Given that mutation is a pervasive phenomenon, the relative frequency of lineages with no mutations will decrease over time.  Furthermore, in a finite population there is a high probability that the mutant-free class will be lost.  Given that a succession of "backward" mutations is unlikely, the frequency of deleterious mutations within and among lineages will increase.  As a result, using the analogy of a ratchet (a mechanical device that only allows forward movement), the population of asexual lineages will carry a steadily increasing load of deleterious mutations.

The proposed advantage of sexual reproduction is that recombination of chromosomes and random mating will produce recombinant genomes with fewer mutant alleles.  Simultaneously, genomes would be produced with such a large number of deleterious mutants that they would be eliminated by selection.  In fact, there is the theoretical possibility that the mutant-free class could be reproduced each generation.  The expected equilibrium frequency of deleterious mutations would therefore be much lower in a sexually reproducing population.

A complementary version of the Fisher-Muller model emphasizes the role of sexual reproduction in the incorporation of favorable alleles.  Beneficial alleles can only be introduced into asexual lineages by successive mutation.  Similar combinations of favorable alleles could be produced in a single generation in sexual populations.

Muller's Ratchet is a powerful argument in favor of sexual reproduction, and it may help to explain the patterns of recombination in certain types of viruses (Chao, 1992).  The problem with it is that it is effectively a group selection argument, and it is doubtful that it is sufficient by itself to offset the costs to the individual of sexual reproduction.

"The Tangled Bank"

Other approaches to the question of the usefulness of sexual reproduction have emphasized the interaction of the organism with aspects of its environment.  Most of these models employ scenarios where biological interactions are intense and there is a hypothetical advantage to producing variable offspring some of which may experience a reduction in the effects of biological interactions.

A good example of such a scenario is Graham Bell's "Tangled Bank" concept (which derives its name from the closing passage of the Origin of Species).  His view proposes an advantage for sexual reproduction in saturated environments where biological diversity provides ecological opportunity and successful individuals are those who can produce diverse offspring.  There is clearly a frequency-dependent element to this scenario, so that the resulting dynamic equilibrium continues to favor sexual reproduction.

Sibling Rivalry

A special case of the ecological diversity scenarios emphasizes the effect of competition among offspring.  The reasoning in this case is that asexual offspring will utilize the environment in exactly the same manner, with the result that, in a resource limited environment, they will reduce their average fitness via intraspecific competition.  Sexually produced offspring, in contrast, will have slightly different resource requirements, and the impact of intraspecific competition will not be as great.  A corollary is that sexual populations will have broader niches, which might make them less vulnerable to extinction.  The problem with this latter point is that it is a group selection argument which requires that extinction must be a comparatively frequent event.

Lottery: Winners and Losers

The previous scenarios assume a spatially heterogeneous environment, but another model proposes a world with more temporal heterogeneity.  This scenario often uses the analogy of a lottery.  The argument is that asexual reproduction is like buying a number of tickets with the same number, while sexual reproduction is a diversified approach to wagering.  To continue the analogy, the mean "payoff" to each strategy may be similar, but the variance for asexual reproduction should be greater.  In biological terms this would translate into a higher rate of lineage extinctions for asexual species.

The problem with lottery models is that the conditions under which sexual reproduction is advantageous appear to be fairly restrictive.  To date, most of the attempts to model the "lottery" have concluded that environmental variability must occur in a particular pattern in order offer sexual reproduction a consistent advantage.

The Red Queen

Most of the scenarios under which sex is hypothesized to be advantageous involve either changing genomes (through mutation) or changing environments.  That both are characteristics of the real world is not in dispute.  The difficulty comes in determining how much change is necessary to offer sex a clear advantage.

One scenario that has received increasing attention is the "Red Queen" hypothesis.  This concept takes its name from the Alice in Wonderland character who had the opinion that one must run as fast as one can in order to stay in the same place.  When applied to sexual reproduction, the Red Queen refers to the frequency-dependent relationship between hosts and parasites.

The scenario is really quite simple.  Parasites and pathogens are assumed to cause a significant reduction in host fitness, and susceptibility to a given parasite is a function of genotype.  Novel, parasite resistant genotypes will have a high fitness when they first appear, but their selective advantage will decline as they increase in frequency and the parasite evolves to circumvent their resistance.  Obviously, any mechanism which produces new genotypic variants will be favored.  Sexual reproduction, therefore, becomes an adaptation to this intimate species interaction.  There is also a complementary argument that favors sexual reproduction by parasites faced with genotypically variable hosts.

The Red Queen scenario is popular because parasitism is an almost universal phenomenon.  It does not depend upon specific mutation rates or environmental heterogeneity.  It does, however, have the problem that we have not determined the level at which sex has a clear advantage.

DNA Repair

The final hypothesized advantage for sexual reproduction is based on the observation that meiotic, chromosomal recombination is always associated with sexual reproduction.  In fact, chromosomal recombination still occurs in species that have abandoned mating and fertilization (e. g., parthenogenesis), but interbreeding never occurs without recombination.  This has caused some researchers to propose that DNA repair through recombination is the primary function of sexual reproduction, and all of the other aspects of sex are merely ancillary effects.

There are some additional observations that seem to support the DNA repair hypothesis.  First, there is a high degree of similarity between the enzymes involved in DNA repair during replication and those utilized during chromosomal recombination.  Second, the breaks in DNA sequences during recombination are ideal for signals associated with the standard DNA repair process.  Third, the process of chromosomal recombination appears to be far more complex - and involves far less actual recombination - than would be expected if its only function was the physical exchange of chromatid segments (Bernstein, et al., 1988).  Finally, there is the intriguing observation that there is often a positive correlation between the frequency of DNA lesions and recombination.

The basic idea of the DNA repair hypothesis is that "sex" evolved because of the benefits derived from bringing together two similar, functional segments of DNA for the purpose of comparative repair.  These DNA segments could be used as complementary templates for excision, recombination, and restoration of damaged sequences.  There are some similarities to the benefits assumed under Muller's Ratchet, with the implication being that heavily damaged chromatids would be eliminated by selection.  The rest of sexual reproduction simply becomes the machinery necessary to bring the chromosomes together.

There are several variations on the theme of DNA repair.  For example, it has been noted that normal development leads to changes in gene activation through DNA methylation.  It has been proposed that meiotic recombination will help to reverse this process, thereby returning the genome of gametes to an undifferentiated, totipotent state (Holliday, 1988).  Another suggestion is that recombination helps to rid the genome of "parasitic" DNA.

The DNA repair hypothesis is certainly a useful proposition that helps us to focus on the molecular events associated with sexual reproduction.  The main criticism of the hypothesis is that it is difficult to accept that the process needs to be this complex if its primary purpose is the conservation and restoration of genomic DNA.

Recent Events:  The Red Queen Meets Muller's Ratchet

It should be clear by now that there is probably no single explanation for the maintenance of sexual reproduction (although there may be a single phenomenon responsible for its origin).  Howard and Lively (1994) have described a model that combines the effects of parasitism and mutation, and offers an example of a pluralistic solution to the problem.

When viewed separately, Muller's Ratchet requires significant mutation rates and the Red Queen model implies considerable mortality due to parasitism in order for sexual reproduction to be clearly advantageous.  Howard and Lively used a range of values for mutation rate and parasite mediated selection in order to determine the range of values over which sexual reproduction was advantageous. Using computer simulations, they found that moderate effects of parasites and reasonable mutation rates could combine to produce conditions under which a sexual population could not be overwhelmed by an asexual invader.

Miscellaneous Observations

All of the hypotheses discussed above are difficult, if not impossible, to test at this time.  The reason is that sexual reproduction is a basic fact of life on this planet, and it is almost impossible to conduct either controlled experiments or systematic observations where the only factor is sexual vs. asexual reproduction.  There have been attempts, however, and the following observations are presented for your consideration:

Parthenogenetic insects have a lower hatching rate than that of sexual species.

Apomictic (parthenogenetic) plant species often produce more seeds than sexual varieties.

Asexual plant taxa tend to have larger geographical ranges, range into higher latitudes, and are often found at higher elevations than sexual relatives.  The are also very often found to be polyploids, while their sexual relatives are diploids.

Parthenogenesis is rare among multicellular plants and animals.  When it does occur, it is found in small, isolated taxa, whose nearest relatives are all sexual.  Among protists and prokaryotes, however, sex is rare.

Parthenogenetic animals tend to occupy recent, novel, or disturbed environments.

In a freshwater snail in New Zealand, the frequency of parthenogenesis is negatively correlated with the level of trematode parasitism.

In a survey of 24 species of mammals, there was a positive correlation between generation time and the frequency of chromosomal recombination.