The Origin of Sex

"Sex is the queen of problems in evolutionary biology."
Graham Bell, The Masterpiece of Nature, 1982¹

"we are still far from a definitive answer to the question of why sexual reproduction is so common."
Rolf Hoekstra, Nature, 2005²

Summary

Most plants and animals reproduce by sex.  Two parents each pass on half their genes, through a special cell division called meiosis that halves the chromosome number and shuffles the parents' genes.  Many organisms skip it and clone themselves.  This article asks two separate questions and keeps them apart.  Why does sex persist, when cloning looks cheaper?  And how did meiosis and sex begin?  For each step it asks what has been observed, what is inferred, and what is assumed.

• Young-earth and design writers mostly argue about the origin. Walt Brown, Werner Gitt and Don Batten say male and female organs had to appear complete and at the same time. F. LaGard Smith says natural selection could not produce meiosis, and Jonathan Sarfati that its intermediate stages would be harmful. Brian Thomas of the Institute for Creation Research suggests animals keep sex because they cannot reproduce any other way. John Sanford says sex slows but cannot stop genetic decline. Harry Sanders of Answers in Genesis says no hypothesis for sex has "any observable evidence."³ Only a few separate the persistence question from the origin question.
• The cost of sex is real, and it has been measured. A sexual female spends half her offspring on sons, so a clone that makes only daughters should double its share each generation. In New Zealand mud snails kept in outdoor tanks, clones rose 1.6-fold in one generation, and the best estimate of the cost of sex was 2.14, with a 95% range of 1.81 to 2.55. The cost falls away where there are no males to pay for: in species with equal-sized sex cells, in hermaphrodites, and where sex is rare.
• Harmful mutations can pay for sex only in some species. The proposal of Alexey Kondrashov needs more than one new harmful mutation per genome per generation, and needs the mutations to harm more together than apart. Simple arithmetic shows it needs at least 0.69 even in the best case. Peter Keightley and Adam Eyre-Walker estimated about 0.06 to 0.07 in fruit flies, "some 20 times lower" than required in short-lived animals.⁴ The second condition is found in some tests and not in others. Muller's ratchet, the slow loss of the least-mutated clones, is real in the laboratory but too slow to matter in large populations.
• Parasites favour sex in the two species where it was tested best. In the mud snail, the share of males tracks parasite infection across 51 sites, and common clones were replaced by rare ones over 7 to 10 years. In the worm Caenorhabditis elegans, every self-fertilising population died within 20 generations when exposed to a coevolving bacterium, while outcrossing populations survived. But a general model by Sarah Otto and Scott Nuismer found that "species interactions typically select against sex" unless selection is strong.⁵
• Laboratory experiments show sex speeding adaptation, under conditions. Yeast and algae with occasional sex adapted faster in harsh or novel conditions, by about 10% per generation in growth rate in the best-measured case, and not at all in benign conditions in two studies. These organisms have no males, so they test the benefit of sex, not the twofold cost. Priced, an advantage of that size pays easily for sex once every 25 generations, and covers only about a seventh of the cost if sex happens every generation.
• Lineages that give up sex mostly do not last. Most asexual animal lineages are young offshoots of sexual relatives, and several accumulate harmful amino-acid changes faster than their sexual kin. The exceptions are real but fewer than they looked: the bdelloid rotifers, long called an evolutionary scandal, show signs of some genetic exchange after all, while ancient asexual mites show no excess damage. Laboratory crosses made a new clonal lizard species in one generation, so animals can switch to cloning, and Thomas's "no other way" does not hold.⁶
• How meiosis began is undemonstrated. The genes for meiosis are shared across living eukaryotes, including parasites never seen to mate, which leads mainstream biologists to infer that the last common ancestor of eukaryotes already had full meiosis. Nothing living shows an earlier stage. Four published proposals disagree about which step came first, and their authors call the origin an "enigma"⁷ that has "never been satisfactorily explained."⁸
• The claim that male and female had to appear together is wrong as stated. Sex does not need males and females. Single-celled algae mate as equal-sized cells of two mating types. In their many-celled relatives, a single mating-type gene, MID, now decides between sperm and egg, and switching it on in females made them produce sperm.
• Creationists mostly quote the scientists accurately but out of context. Bell's "queen of problems" is accurate, and its next sentence asks why sex is maintained, not how it began.¹ Smith misquotes it as "queen of evolutionary problems."⁹ Jonathan McLatchie of the Discovery Institute quotes Carl Zimmer calling sex "a recipe for evolutionary disaster"¹⁰ and drops Zimmer's next sentence, which gives the parasite answer. Calling the problem unsolved is fair: Otto wrote in 2009 that more data are needed "to settle the paradox of sex."¹¹ Saying there is no evidence is not.

Verdict.  Three answers.  On persistence the question is unfinished on both sides:  mainstream biology has demonstrated mechanisms that make sex pay in particular species, but not one that pays the twofold cost in general, and creationist writers have offered no mechanism at all.  On the origin of meiosis the creationist argument is right that no pathway is demonstrated, and wrong when it says one is impossible.  On the claim that both sexes had to appear at once, the creationist argument is wrong.

Part 1: Two questions, and what creationists and design writers say

The literature on sex asks two different questions.  The first is why sex persists.  Every sexual species could, in principle, be replaced by a mutant that clones itself, and the question is why that rarely happens.  The second is how sex began:  how meiosis, the halving division that makes eggs and sperm, came to exist at all.  An answer to one is not an answer to the other.  A benefit that keeps sex going today does not say how the machinery was first built, and a story of how meiosis began does not say why clones have not taken over since.

The origin argument.  Most young-earth writing on sex is about the second question.  Walt Brown's In the Beginning says that "the amazingly complex, radically different, yet complementary reproductive systems of the male and female must have completely and independently evolved at each stage about the same time and place," and that "just a slight incompleteness in only one of the two at any stage would make both reproductive systems useless."¹²  Werner Gitt, in a chapter Answers in Genesis republished in 2012, makes the same point:  "Sexual reproduction is only possible when both sexes have fully functional reproductive organs at the same time."¹³  F. LaGard Smith's 2018 book Darwin's Secret Sex Problem argued, in his own summary, that "Natural selection could not possibly have evolved even the most elementary form of sex by meiosis," and that it could not have supplied "the first sexually-compatible pair of any species."⁹  Jonathan Sarfati of Creation Ministries International grants that sex has advantages once it exists, and argues that advantages cannot explain how it "could be built from scratch."¹⁴  Jerry Bergman's review in Journal of Creation calls Smith's the only book he knows that "covers in detail the widely acknowledged problem of the evolutionary origin of sexual reproduction."¹⁵

The persistence argument.  Fewer creationist writers address the second question directly.  Harry Sanders III of Answers in Genesis does.  He says asexual reproduction "would appear to be the way to go," reviews the parasite hypothesis and rejects it because "the host would need to undergo radical genetic changes to produce radically different physical structures," and concludes that evolutionary hypotheses for sex lack evidence:  "Ultimately, none of them have any observable evidence to back them up."  He adds that "asexual lineages should be going extinct in the evolutionary paradigm."³  Jonathan McLatchie of the Discovery Institute, an intelligent-design writer, argued in 2011 that the costs of sex "are potent short-term disadvantages which should have mitigated against sexual reproduction evolving in the first place."¹⁰  Hugh Henry and Daniel Dyke of Reasons to Believe, an old-earth ministry, are the clearest on the split.  They ask how sex arose, and then say "The real question may be how do we keep sexual reproduction."  Their own answer is that "From the point of view of natural selection, there is no benefit to sexual reproduction."¹⁶

Sex as design.  Three writers say what sex is for in a created world.  Jean Lightner in Journal of Creation writes that "God designed meiosis in a way that naturally tends to increase diversity," as a source of variation within created kinds.¹⁷  John Sanford's Genetic Entropy holds that sex slows mutational decline and does not stop it:  "An unspoken theme of this book is that what is already known for clonal populations also applies to sexual populations."  He explains living clonal plants and animals by time:  "there has not yet been enough time for such populations to degenerate to the point of extinction."¹⁸  Sanford's argument that every genome degenerates belongs to the article on genetic entropy and is not re-argued here.  Brian Thomas of the Institute for Creation Research gave the simplest answer to the persistence question.  Writing about a beetle experiment in 2015, he suggested that animals keep reproducing sexually "simply because they have no other way to do it!"⁶

That last answer can be checked, and Thomas checked it himself four years earlier.  In 2011 he reported that laboratory crosses of whiptail lizards "produced four new parthenogenetic 'species'," lineages of females that clone themselves, and he credited the ability to design:  "It would make sense that a Creator would have endowed these egg-laying vertebrates with the potential to perpetuate themselves even in the event that a male was unavailable."¹⁹  If a vertebrate can switch to cloning in one generation in a laboratory, "no other way" cannot be the reason sex persists.  The question of why clones have not taken over remains, for the creationist as for anyone else.

What the two sides do not claim.  No creationist source found here denies that sex has a cost, and none claims that asexual organisms are better designed.  Several, including Sanders, say both modes of reproduction were created at the beginning.³  On the other side, no mainstream source read here claims that the origin of meiosis has been demonstrated.  The live disputes are narrower than the rhetoric:  whether the benefits of sex have been measured, and whether the first steps of meiosis can be described.

Part 2: What sex costs, and how the cost was measured

The argument.  John Maynard Smith set out the cost in 1978 with a one-generation sum.  Take a sexual species with equal numbers of males and females, and suppose a mutation makes some females produce only daughters like themselves.  Each mother has the same number of young, but the sexual mother's brood is half sons, who bear no young.  "Hence, in one generation, the proportion of parthenogenetic females increases from n/(2N+n) to n/(N+n); when n is small, this is a doubling in each generation."²⁰  He was careful about what the cost is a cost of:  "I believe that the twofold advantage of parthenogenesis is best seen as the advantage of not producing males."²⁰  The argument also holds "only if other things are equal," that is, if sexual and clonal females are otherwise equally fit.²⁰

How fast a clone wins.  A doubling every generation is fast.  Starting from a single clonal female, a clone reaches half of a population of a million in about 20 generations, and half of a billion in about 30.²¹  Even a clone 30% less fit than sexual females takes over a population of a million in about 41 generations.  So whatever keeps sex going has to work now, generation by generation.  A benefit that pays off over thousands of years, such as faster adaptation to a climate that has not changed yet, arrives too late.  Maynard Smith credited George Williams with the cleanest form of this point:  "If parthenogenetic and sexual methods of reproduction coexist in the same population, then there must be some immediate advantage to sexual reproduction to counterbalance its twofold disadvantage."²⁰

The measurement.  Amanda Gibson, Lynda Delph and Curtis Lively measured the cost directly in 2017.  The New Zealand mud snail Potamopyrgus antipodarum lives in lakes as a mixture of sexual females, males, and clonal females, which are mostly triploid.  The team collected young snails from Lake Alexandrina on New Zealand's South Island, stocked large outdoor tanks at a field station with them, and counted the share of clonal offspring one generation later, in six tanks a year for four years.  "The frequency of asexuals increased 1.60-fold (95% CI [1.48, 1.73]) from an initial frequency of 29% in the parental generation."²²  A 1.60-fold rise from 29% is close to what a doubling of the odds predicts when clones are already common, about 1.55-fold.  Fitting Maynard Smith's model to the counts, "The likelihood of model 3 was maximized at a cost of sex that slightly exceeds two (c = 2.14, 95% CI [1.81, 2.55])."²²  When the sex ratio was set at the expected 50:50, the fitted fecundity of clonal and sexual mothers was equal, "consistent with the all-else-equal assumption."²²

The same group found the cost larger when parasites were added.  In tanks dosed with the trematode Microphallus, "sexuals paid both the two-fold cost of sex and the cost of higher infection prevalence," and the fitted cost rose from 2.18 to 3.15.²³  In the lake itself the yearly estimates were lower:  "1.98 in 2012, 1.47 in 2013, 0.81 in 2014, and 1.39 in 2015."²³  The authors suggest that overlapping generations in the wild spread the cost over more time.

What is observed, what is inferred.  The cost is observed, in one species, over one generation, in tanks and in a lake.  That a similar cost applies to other species with separate sexes is an inference, though a short one:  it follows from the arithmetic of producing sons whenever sexual and clonal mothers are otherwise alike.

Where the cost does not apply.  The twofold cost is a cost of males.  Maynard Smith listed the exceptions himself:  "the twofold advantage of parthenogenesis does not apply to isogamous species, or to monogamous species with paternal care of the young."²⁰  Isogamous species are those whose sex cells are all the same size, such as yeast and many algae.  They make no sons, so a clone among them gains nothing by skipping males.  A species that has sex once in many generations pays the cost only in those generations.  Most of the laboratory experiments in Part 5 use organisms of these two kinds.  So they can show that sex has a benefit, but they cannot show that the benefit is large enough to pay for males.

Other costs.  Sex has costs besides sons.  Finding and courting a mate takes time and energy and carries risk.  Meiosis breaks up gene combinations that worked in the parents, so some offspring are worse than either parent.  Jussi Lehtonen, Michael Jennions and Hanna Kokko asked in 2012 whether the cost really is twofold and what its source is:  "males, genome dilution or both?"²⁴  None of these costs is measured as precisely as the cost of males.

Part 3: Harmful mutations as the reason for sex

Every genome copies itself with errors, and most errors that matter are harmful.  Two of the oldest explanations for sex say it helps a population get rid of them.  Both have been tested, and both work only under conditions that many sexual species do not meet.

Muller's ratchet.  In 1964 Hermann Muller pointed out that a clone can never get back to a cleaner genome than the cleanest one it has.  Joseph Felsenstein's 1974 paper quotes him:  "an asexual population incorporates a kind of ratchet mechanism, such as it can never get to contain, in any of its lines, a load of mutation smaller than that already existing in its at present least-loaded lines."²⁵  If the few individuals with the fewest mutations die out by chance, that class is gone for good, and the ratchet clicks once.  Sex can rebuild a clean genome by combining clean pieces from two parents.

The ratchet's speed depends on how many individuals sit in the cleanest class.  At balance between mutation and selection, a clone of N individuals has about N e^(−U/s) of them, where U is the number of new harmful mutations per genome per generation and s is the harm each does.²⁶  "The smaller the n0, the faster the ratchet is likely to be."²⁶  In a large population the cleanest class is huge.  At bacterial mutation rates of a few thousandths per genome per generation and s = 0.01, between half and nine-tenths of a million cells sit in it, and the ratchet does not move.²⁷  It moves when populations are small or mutation rates are high.

The laboratory agrees.  Dan Andersson and Diarmaid Hughes passed 444 lines of Salmonella through single-cell bottlenecks, the harshest possible, for about 1,700 generations.  "We can conclude that 5 of 444 lineages (1%) suffered a significant loss of fitness after 1700 generations."  They read this as showing "that Muller's ratchet indeed operates at a significant rate in bacteria."²⁸  Lin Chao saw fitness fall in 20 bottlenecked lines of an RNA virus, which mutates far faster.²⁹  Clifford Zeyl's group ran yeast at an effective size near 250 for about 2,900 generations, and two populations died, both of them strains with a raised mutation rate.³⁰  So the ratchet is real.  It is also slow wherever populations are large.  Peter Keightley and Adam Eyre-Walker summed up the consensus:  the ratchet "is unlikely to maintain obligate sexuality, since it operates very slowly in large populations and can be avoided with very little sex."⁴

Kondrashov's hypothesis.  Alexey Kondrashov proposed in 1988 a mechanism that does not depend on population size.  "If the deleterious mutation rate per genome per generation is greater than 1, then the greater efficiency of selection against these mutations in sexual populations may be responsible for the evolution of sex."³¹  The idea needs the mutations to interact:  two together must do more harm than the sum of each alone, which geneticists call synergistic epistasis.  Then selection removes many mutations with each death, and sex keeps reshuffling them into some individuals that can be removed.  Keightley and Eyre-Walker stated both conditions:  "If rates of deleterious mutation per diploid genome per generation (U) exceed 1, and mutations interact synergistically, then sexuals can overcome their inherent twofold disadvantage."  They added that "This is a minimum condition, since U needs to be rather greater than 1 if stochastic factors are taken into account."⁴

Priced.  There is a hard ceiling on what this mechanism can deliver, whatever the epistasis.  Motoo Kimura and Takeo Maruyama showed in 1966 that a clone at balance between mutation and selection has an average fitness of e^(−U) relative to its own best class, and that "under asexual reproduction, epistasis has no effect in reducing the mutational load."³²  A sexual population cannot be fitter than a genome with no mutations at all.  So the most sex can gain is a factor of e^U.  For that to reach two, U must be at least ln 2, about 0.69, even with the most favourable epistasis anyone could propose.³³  At U = 0.1 the ceiling is 1.11, a gain of 11% against a cost of 100%.

Measured U.  Keightley and Eyre-Walker estimated U from protein-coding genes in nine pairs of animal species.  Their estimates per generation ranged from 3.0 in human and chimpanzee to 0.49 in chicken and quail and 0.058 to 0.071 in three pairs of fruit flies.⁴  Even counting every point mutation in the fruit-fly genome as harmful gave 0.8.  For animals that breed monthly or faster, such as many insects, copepods and nematodes, they inferred rates "of the order of 0.05, some 20 times lower than the MD hypothesis requires."  Their conclusion:  "the hypothesis is not supported as a general mechanism."⁴  Two features of the method need stating.  It converts differences between species into rates per generation using divergence dates taken from fossils, so its numbers carry the deep-time timescale.  And it assumed 80,000 human genes, about four times the number later counted.  Keightley's own 2012 revision, which used the mutation rate measured directly in families, put the human rate from protein changes alone at about 0.35, "2–10 times lower than previous estimates," and the whole-genome rate at about 2.2.³⁴  So humans may clear the bar.  Flies, nematodes and copepods, which also have males and pay the full cost, do not.

Measured epistasis.  The second condition has been tested directly.  Santiago Elena and Richard Lenski built E. coli strains carrying one, two or three mutations and found interactions "antagonistic as often as they are synergistic," so that "These results do not support the mutational deterministic hypothesis for the evolution of sex."³⁵  A 2007 review by Roger Kouyos and colleagues said that the needed epistasis "is relatively uncommon."³⁶  A 2017 study of human and fruit-fly genomes by Mashaal Sohail and colleagues found the pattern synergistic epistasis predicts among rare gene-breaking mutations, and suggested it "may explain how human and fly populations persist despite high genomic mutation rates."³⁷  The effect they found was small, about a 5% reduction in variance, which shows the sign of the interaction but not that it is strong enough to pay the twofold cost.

Interference in finite populations.  The theory that most reviewers now favour combines mutation, selection and chance.  In a population of limited size, a good mutation can be trapped on a genome carrying bad ones, and two good mutations in different individuals compete instead of combining.  This is called Hill–Robertson interference, after William Hill and Alan Robertson's 1966 analysis.  Keightley and Sarah Otto showed in 2006 that harmful mutations at many sites give sex "a stochastic advantage" that "increases with population size."³⁸  Otto's 2009 review shows a simulation in which genes for occasional sex spread against a twofold cost.  It also says plainly:  "An open question is whether the drift-based explanation is strong enough to account for the evolution of sex in the face of substantial costs of sex."¹¹  Its sketch of how sex could be common is hedged:  "It remains to be demonstrated but seems plausible that high levels of sex could evolve, given a combination of selective forces acting across the genome, even in the face of substantial costs of sex."¹¹

What is observed, inferred and assumed.  Observed:  the ratchet lowers fitness in small laboratory populations of bacteria, viruses and yeast;  measured rates of harmful mutation are below 0.69 per generation in short-lived animals and above it in humans;  synergistic epistasis is found in some tests and not others.  Inferred:  that interference in finite populations is strong enough to keep obligate sex.  That has been modelled, not measured.  Simulations run by John Sanford's own group found that truncation selection combined with sexual recombination can halt the build-up of mutations;  the genetic entropy article discusses that result.

Part 4: Parasites, the Red Queen

The Red Queen hypothesis, named after the character in Through the Looking-Glass who runs to stay in place, says that parasites keep sex going.  Parasites evolve to attack the commonest host genotypes.  A clone is a single genotype, so once it becomes common it becomes a target, while sexual offspring keep changing.  The advantage is immediate, which is what Part 2 showed any explanation needs.  It is also the explanation with the most direct field and laboratory support.

The snail in the field.  Curtis Lively surveyed the New Zealand mud snail in 1987.  The share of males, a rough measure of how much sex goes on, was higher in lakes than in streams and rose with the rate of infection by trematode worms across 22 lakes and 29 streams.  He was careful:  "These results suggest a role for trematodes, especially Microphallus, in the maintenance of sex in P. antipodarum, but they are not conclusive by themselves."  He tested one obvious artefact, that the worms might infect males more often, and found they did not.  His conclusion was that "the Red Queen hypothesis is the best supported of the hypotheses considered here."³⁹  Jukka Jokela, Mark Dybdahl and Lively then followed a mixed lake population of sexual and clonal snails.  "We found that, within 7-10 years, the most common clones were almost completely replaced by initially rare clones in two different habitats, while sexuals persisted throughout the study period."  The common clones had also become more susceptible to the local parasites, but not to parasites from elsewhere.⁴⁰  Over a longer run, the group found that in the early 2000s clones were more infected than sexuals, and "A decade later, the asexuals had declined in frequency and were less infected than sexuals."²³  That is the cycle the hypothesis predicts.

The worm in the laboratory.  The nematode Caenorhabditis elegans has hermaphrodites that fertilise themselves and males that mate with them, so it can be pushed toward selfing or toward outcrossing by mutation.  Levi Morran and colleagues exposed selfing, outcrossing and wild-type worms to the bacterium Serratia marcescens, either a fixed strain or one allowed to evolve alongside the worms.  "In the coevolution treatment, all of the obligately selfing populations became extinct within twenty generations."  In wild-type worms, "outcrossing rates increased from 20% to over 70%" in eight generations, and stayed high only where the bacterium kept evolving.  The obligately outcrossing worms, which pay the full cost of males, all survived.⁴¹  An earlier study by the same lab found that selfing worms also fared worse under an increased mutation rate:  "obligate selfing populations fixed significantly more mutations than did the obligate outcrossing populations."⁴²  The authors stated the limit of their result:  "the long-term maintenance of sex requires that the novelty does not wear off."⁴¹

Selfing is not cloning.  A selfing worm still goes through meiosis and shuffles its two chromosome sets.  But because it mates with itself, its offspring quickly become uniform, so selfing populations behave much like clones toward a parasite.  The experiment compares two sexual systems, one of which shuffles far less.  Jonathan McLatchie described it in 2024 as a contrast between self-fertilised and "sexually reproducing" worms;⁴³ the selfers were sexual too, only far less variable.

The theory's limits.  A general model gives a less friendly answer.  Sarah Otto and Scott Nuismer modelled interacting species across 13,608 combinations of settings and found that "species interactions typically select against sex."  They concluded that "although the Red Queen favors sex under certain circumstances, it alone does not account for the ubiquity of sex."⁵  In their model the Red Queen keeps much sex only "unless species interactions induce strong selection per locus," which needs parasites that are common, harmful, closely matched to particular host genes, and controlled by a few genes on each side.⁵  Lively's own 2010 review describes "an oscillation of opinion regarding the role of parasites in selection for sex."⁴⁴

What is observed, inferred and assumed.  Observed:  in one snail, clonal frequency and parasites cycle together over a decade, and common clones lose out;  in one worm, populations that cannot outcross die under coevolving bacteria while outcrossers survive, paying the cost of males.  Inferred:  that parasites do the same in most sexual species.  Theory says that needs strong selection, and how often nature supplies it is not measured.

Part 5: Sexual and asexual populations compared in the laboratory

The cleanest tests of whether sex helps are experiments in yeast, algae and rotifers that can be made to reproduce with or without it.  Their results are real, and they answer a narrower question than the one they are often cited for.

Yeast.  Matthew Goddard, Charles Godfray and Austin Burt grew eight sexual and eight asexual yeast populations, identical except that the asexual strain had two meiosis genes deleted, with a round of sporulation every 25 to 50 generations.  In a benign medium nothing evolved that they could detect.  In a harsh one, hot and salty, both adapted:  "In the asexual lines, relative fitness increased from 0 to 0.59, whereas for the sexual lines the increase was from 0 to 0.68.  These figures are equivalent to increases in population growth rate of 80% and 94%, respectively."  The difference, "about 0.1" per generation on a logarithmic scale, meant that "in the 25 mitotic generations between episodes of sex in our experiments, a sexual individual could expect to leave about 12-fold" as many descendants.⁴⁵  Jeremy Gray and Goddard repeated the design in 2012:  "Sex provided no advantage in the permissive environment where beneficial mutations were rare," while in the stressful one, "where detrimental mutations were likely of more consequence, and where beneficial mutations were apparent, sex provided an advantage."⁴⁶  Clifford Zeyl and Graham Bell had found the reverse pattern in 1997:  sex raised fitness in the sugar their populations were already adapted to, not in the new one, and they credited the gain to the removal of harmful mutations.⁴⁷  Michael McDonald, Daniel Rice and Michael Desai sequenced whole yeast populations in 2016.  With sex once every 90 generations, "sexual populations adapted significantly faster than asexual controls," and in asexual populations "significantly deleterious mutations hitchhike to fixation" on the backs of good ones.⁴⁸  That is Hill–Robertson interference seen directly.

Algae and rotifers.  Nick Colegrave found in the alga Chlamydomonas that "sex has a marked effect in large populations but little effect in small populations," with sexual lines about 10% fitter than asexual controls after eight rounds of mating and 50 further generations.⁴⁹  Lutz Becks and Aneil Agrawal, using a rotifer that switches between cloning and sex, found that "Sexually derived genotypes are more fit than asexually derived genotypes when adaptive pressures are strong, but this pattern reverses as the pace of adaptation slows."⁵⁰

What these experiments cannot show.  Yeast and Chlamydomonas make no males.  Rolf Hoekstra, commenting on the Goddard study in Nature, put it directly:  "yeast has no male-female distinction, so the twofold cost of sex does not apply, which prevents a straightforward generalization of the conclusions to most animals and plants."²  And in every design above, sex happened once every 25 to 90 generations, not every generation.

Priced.  Take the best-measured case, Goddard's advantage of about 0.1 per generation on the log scale, a growth rate about 10.5% higher.  Suppose an organism with males had sex once every 25 generations and paid the twofold cost only in that generation.  It would lose a factor of two once and gain a factor of about 12 over the 25 generations, a net gain of about six.  Now suppose it had sex every generation, as obligately sexual animals do.  It would pay a factor of two every generation and gain a factor of 1.105, which covers about a seventh of the cost.⁵¹  The experiments therefore show that occasional sex pays well when a population is adapting to hard conditions.  They do not show that obligate sex with males pays.  Sarah Otto's 2009 review drew the same line in its simulations:  genes for low levels of sex spread against a twofold cost, while genes for high levels "did not spread unless costs were much less than twofold."¹¹

What is observed, inferred and assumed.  Observed:  in organisms without males, occasional sex speeds adaptation under strong selection, often not at all under weak selection, and keeps harmful hitchhikers from fixing.  Inferred:  that the same benefits, at the size seen here, keep obligate sex in animals.  The arithmetic says that inference does not follow without a much larger benefit.

Part 6: Lineages that gave up sex

If sex persists because clones fail, lineages that give up sex should show the damage.  Two kinds of evidence bear on this:  how long asexual lineages last, and whether they collect harmful mutations faster.

Most asexual lineages are young.  Asexual animals and plants usually sit at the tips of the tree of life, as recent offshoots of sexual relatives, a pattern often called "twiggy."  It was long read as proof that clones die out quickly.  Tanja Schwander and Bernard Crespi showed in 2009 that the pattern does not prove that much:  low rates of origin or of speciation "can generate twiggy distributions of asexuality, in the absence of high extinction rates for asexual lineages."⁵²  Maurine Neiman and colleagues found that estimated ages of asexual lineages follow a smooth distribution, so that taxa viewed as "scandalously" ancient "merely fall on the high end of this distribution."⁵³  Every one of these ages is inferred from a molecular clock or a fossil date.  None is observed.

Faster damage, measured.  Several studies compare the same genes in sexual and asexual relatives.  In the water flea Daphnia pulex, Susanne Paland and Michael Lynch estimated that "mitochondrial protein-coding genes in asexual lineages accumulate deleterious amino acid substitutions at four times the rate in sexual lineages."⁵⁴  In the mud snail, Neiman's group found that the effect "occurs rapidly enough to be detected in recently derived asexual lineages."⁵⁵  In Timema stick insects, asexual lineages show the same excess, but Jens Bast and colleagues added that "there is currently no evidence that the average fitness of asexual Timema females is lower than the fitness of sexual females."⁵⁶  The direction is not universal.  In oribatid mites, which lost sex long ago by clock estimates, Alexander Brandt and colleagues found the ratio of harmful to silent changes "on average lower at asexual branches as compared to sexual branches."⁵⁷

The bdelloid rotifers.  Bdelloids are tiny freshwater animals in which no male has ever been seen.  David Mark Welch and Matthew Meselson called them "the largest metazoan taxon in which males, hermaphrodites, and meiosis are unknown," and reported in 2000 that gene copies inside each genome were far more different than allele pairs in sexual species, as long asexuality predicts.⁵⁸  Bdelloids became the standard "evolutionary scandal."  Later genomes complicated the story.  The genome of Adineta vaga in 2013 showed that its chromosomes are arranged in a way that "does not allow meiotic pairing;  instead, we find abundant evidence of gene conversion."  Its allele copies differ by only 4.4%, within the range of sexual species, and the authors wrote that "current evidence does not exclude that they may engage in sex on rare, cryptic occasions."⁵⁹  A 2020 study of eleven wild A. vaga found "evidence for interindividual genetic exchange and recombination,"⁶⁰ and a comparison of several bdelloid genomes found that features taken as signs of long asexuality "are not general to all bdelloid species, even within the same genus."⁶¹  The bdelloid case, which was the best evidence that animals can live without sex for tens of millions of years, now looks like a case of rare or unusual genetic exchange.

Jeffrey Tomkins of the Institute for Creation Research wrote about the bdelloids' foreign genes, and argued that "evolutionary biologists have resorted to fictional stories cloaked in technical terminology to escape the straightforward conclusion that rotifer DNA was purposefully crafted."⁶²  He did not address whether they are asexual.  No creationist source found here engages the work on asexual lineages beyond Sanders's claim, in Part 1, that asexual lineages "should be going extinct in the evolutionary paradigm."  The mainstream literature does not predict that every clone dies at once.  It predicts that clones are mostly young and that they collect harmful changes faster, and the measurements mostly agree.

For a young earth.  This evidence is friendly to a young-earth reading in one respect.  The ages that make some asexuals "ancient" come from clocks and amber dates, and the bdelloid signal of very old asexuality has weakened under better genome data.  Sanford's view that clones survive because there has not been enough time to degenerate fits lineages that are young.  It does not by itself explain why clones exist alongside sexual relatives in the same lakes, which the Red Queen data do.

What is observed, inferred and assumed.  Observed:  asexual lineages are usually close relatives of sexual ones;  in water fleas, snails and stick insects they carry more amino-acid changes;  bdelloid genomes cannot pair chromosomes in ordinary meiosis yet show some exchange.  Inferred:  ages in millions of years, and that the twiggy pattern reflects early extinction.  Assumed:  the clock rates used to date the lineages.

Part 7: How meiosis began

Meiosis is the step that makes sex possible.  Before it, a cell copies its chromosomes once.  Then matching chromosomes from the two parents find each other, pair along their length, and swap segments, which requires deliberately cutting both strands of the DNA and repairing them.  Then the cell divides twice, so that each of the four products carries one set instead of two.  At least four things are new compared with ordinary cell division:  pairing of the matching chromosomes, crossing over between them, holding the sister copies together through the first division, and skipping the copying step before the second.

What is observed.  The genes that run meiosis are shared across living eukaryotes.  Marilee Ramesh, Shehre-Banoo Malik and John Logsdon found a core set of them, including five used only in meiosis, in the parasite Giardia, which had never been seen to have sex.⁶³  A wider survey in 2008 found 27 of 29 such genes in Trichomonas vaginalis:  "Although meiosis has not been observed in T. vaginalis, our findings suggest it is either currently sexual or a recent asexual."⁶⁴  One key meiotic enzyme, Spo11, which makes the deliberate breaks, is related to a topoisomerase found in archaea.⁶⁵  Some archaea fuse and exchange DNA:  the salt-loving Haloferax forms cells "containing the full genetic repertoire of both parental cells,"⁶⁶ and Sulfolobus swaps DNA between cells when damaged by ultraviolet light.⁶⁷  Neither halves its chromosome number in a controlled division.

What is inferred.  From the shared genes, Malik's group concludes that each one "also must have been present in the common ancestor of all eukaryotes."⁶⁴  That inference depends on common descent:  genes shared across lineages that branch deep in the tree are taken as inherited from their common ancestor.  On that reading, meiosis was already complete in the last eukaryotic common ancestor, the same ancestor the endosymbiosis article shows was already complex.  Dave Speijer, Julius Lukeš and Marek Eliáš call sex "ubiquitous, ancient, and inherent," while admitting that "entire protist lineages cannot be considered sexual with any certainty yet."⁶⁸  The oldest fossil taken to be a sexual organism is the red alga Bangiomorpha, whose separate spore and gamete forms suggest sex;⁶⁹ its age was revised in 2018 to about 1.05 billion years by the conventional timescale.⁷⁰  The fossil shows a finished sexual organism, not a stage on the way to one.

What is proposed.  Four published accounts of the first steps disagree about which came first.

1. Adam Wilkins and Robin Holliday propose that meiosis came from ordinary division, and that the first new step was pairing of matching chromosomes, which at first served to limit crossing over, not to promote it. They begin by saying the origins of meiosis "have never been satisfactorily explained," and end by admitting "There is, of course, no direct way to test the basic hypothesis presented here since the cells in which meiosis first originated existed well over 1 billion years ago."⁸
2. Thomas Cavalier-Smith proposed that meiosis arose to correct accidental doubling of chromosome sets, by a change that delayed the splitting of the chromosomes' attachment points. He wrote that the link his model needs "has not yet been demonstrated," and that if it exists, "meiosis could have originated by a single key change."⁷¹
3. Harris Bernstein, Richard Michod and others proposed that meiosis grew out of DNA repair, the recombination bacteria use to fix broken chromosomes.⁷² Elvira Hörandl and Speijer tie it to damage from oxygen made by the new mitochondria, and call their own version "much debated and not yet (?) universally accepted."⁷³
4. Thomas Lenormand and colleagues review all of these. They call the gradual origin of meiosis "among the most intriguing evolutionary enigmas," and object to the repair account that "prokaryotes (including archaea) have efficient repair mechanisms that involve recombination, but not meiosis."⁷

What is missing.  No living organism has a partial meiosis that looks like an ancestral stage.  Protists show many variants of meiosis, but every lineage that has one has the full core machinery or has lost parts of it.  No laboratory has evolved any step of meiosis.  The proposed steps are ordered differently by different authors, and none of the authors claims a demonstration.

The creationist claim, tested.  F. LaGard Smith says natural selection "could not possibly have evolved even the most elementary form of sex by meiosis."  Jonathan Sarfati of Creation Ministries International argues that "simply having advantages doesn't remotely explain how they could be built from scratch," because "The hypothetical transitional forms would be highly disadvantageous."¹⁴  The first half of this is right.  No pathway is demonstrated, and the mainstream authors say so in their own words.  The second half is a claim no one has shown.  No creationist source found here computes the fitness of any proposed intermediate, and the archaeal cases show that cell fusion and DNA exchange can exist without a two-sex system or a reduction division.  "Impossible" is not established.  "Undemonstrated" is.

Part 8: How males and females began

The creationist claim.  The most common young-earth argument about sex is about the two sexes.  Brown says the "complementary reproductive systems of the male and female" had to evolve together, since "just a slight incompleteness in only one of the two at any stage would make both reproductive systems useless."  Gitt says sex "is only possible when both sexes have fully functional reproductive organs at the same time."  Don Batten of Creation Ministries International asks how chemistry could "invent the complementary apparatuses needed at the same time," since "non-intelligent processes cannot plan for future coordination of male and female organs."⁷⁴  The argument goes back at least to William Stroud in Creation Research Society Quarterly in 1970, who wrote that unless the male and female systems "arose simultaneously from a population that heretofore was reproducing asexually, then there was no copulation, no progeny, no evolution."⁷⁵

Sex does not need males and females.  Many single-celled eukaryotes that have sex do it with sex cells of equal size.  The alga Chlamydomonas has two mating types, plus and minus, which look alike and differ in a small region of one chromosome.  Two cells of opposite type fuse, and the fused cell later undergoes meiosis.  There is no egg, no sperm, no male and no female.  Maynard Smith drew the conclusion in 1978:  "the essential characteristics of eukaryotic sex, meiosis and syngamy, almost certainly were first evolved in an isogamous population," so that "when considering the origins of sexual reproduction in eukaryotes, one need not allow for a twofold disadvantage."²⁰  So the claim that both sexes had to appear at once is not what any evolutionary account proposes.  Male and female, on that account, come after sex, not with it.

From mating types to sexes, observed.  The volvocine algae, relatives of Chlamydomonas that live as colonies, include species at every stage:  equal sex cells, unequal sex cells, and eggs and sperm in Volvox.  Sa Geng, Peter De Hoff and James Umen tested the gene MID, which makes a Chlamydomonas cell the minus type.  In Volvox the same gene now decides sex.  "Transgenic female V. carteri expressing male MID produced functional sperm packets during sexual development," and males with the gene knocked down "produced functional eggs, or self-fertile hermaphrodites."⁷⁶  Takashi Hamaji and colleagues compared an isogamous and an anisogamous species that bracket the first step.  Both carry a small mating region with "only two sex-limited genes," the male MID and the female FUS1, and they concluded that "the first step to anisogamy in volvocine algae presumably occurred without an increase in MT size and complexity."⁷⁷  Theory explains why unequal sex cells would be favoured once they vary in size:  Geoff Parker, Robin Baker and Vic Smith showed in 1972 that "a marked disruptive effect can be generated" that favours both very small and very large gametes.⁷⁸

The limits.  The volvocine series is a set of living species, not a line of ancestors.  Reading it as a sequence assumes they share common descent.  Geng and colleagues themselves wrote that "the origins of dimorphic sexes and their relationship to mating types in unicellular species are not understood,"⁷⁶ and Jussi Lehtonen and Hanna Kokko that "The historical sequence of events leading from isogamy without mating types to anisogamy with mating types is unclear."⁷⁹  How mating types themselves began, and how the fusion of two cells came to be coordinated with meiosis, is part of the unanswered question of Part 7.

Verdict on this claim.  The argument that a male and a female system had to arise complete and together is wrong as a statement of what evolution requires.  Sex exists without sexes, in living organisms that can be grown in a laboratory.  In one group, a single mating-type gene now controls sperm and egg, and moving it switches a female to making sperm.  This does not show how the volvocine sexes arose step by step, and it says nothing about animals, whose sex-determining systems are different and vary widely.  But it removes the simultaneity premise on which the creationist argument rests.

Part 9: How creationist writers use the scientists' own words

Creationist writing on sex leans heavily on mainstream scientists who called sex a problem.  The quotations checked here are mostly accurate.  The way they are used is often not.

Graham Bell, 1982.  "Sex is the queen of problems in evolutionary biology" is the first sentence of Bell's The Masterpiece of Nature.  Sanders quotes it with the right page, and McLatchie's epigraph reproduces the paragraph exactly.  F. LaGard Smith's version, "Queen of evolutionary problems," is a misquotation printed inside quotation marks.⁹  The more important point is what Bell meant.  The same paragraph ends:  "The largest and least ignorable and most obdurate of these questions is, why sex?  Or, to put this more technically, what is the functional significance of sexuality, which leads to its maintenance under natural selection in biological populations?"¹  Bell's queen of problems is the persistence question.  Smith and McLatchie use it for the origin question.

George Williams, 1975.  Walt Brown quotes the first sentence of Williams's preface accurately:  "This book is written from a conviction that the prevalence of sexual reproduction in higher plants and animals is inconsistent with current evolutionary theory."  The next sentence says what the book is for:  "My purpose is to propose minimal modifications of the theory in order to account for the persistence of so seemingly maladaptive a character."⁸⁰  Williams was naming a problem in order to solve it, and the problem was persistence.

John Maynard Smith, The Theory of Evolution.  McLatchie quotes Maynard Smith's conclusion that the first stages of sex arose "under the influence of selective forces quite different from those which are responsible for the maintenance and spread of sexual processes once they were erected."  The book reads "were responsible" and "perfected."⁸¹  McLatchie's two changed words match exactly the version printed in David Swift's 2002 creationist book Evolution Under the Microscope,⁸² which suggests the quotation was copied from Swift rather than from Maynard Smith.  The meaning survives the slip.  And the sentence itself makes the distinction this article makes:  Maynard Smith separated the origin of sex from its maintenance, and said a long-term benefit could not explain the origin.  On that point the creationist use is fair.

Carl Zimmer, 2002.  McLatchie quotes Zimmer's Evolution: The Triumph of an Idea:  sex "ought to be a recipe for evolutionary disaster," and "any group of animals that evolves sexual reproduction should be promptly outcompeted by nonsexual ones."  He ends the quotation at "Why is sex a success, despite all its disadvantages?"¹⁰  Zimmer's next sentence is the answer he goes on to give:  "Scientists have recently been gathering support for a surprising hypothesis: sex fights off parasites."⁸³  McLatchie then argues against the parasite hypothesis without telling readers that the author he quoted had raised it in the next breath.

Experiments misdescribed.  Sanders says evolutionists "were stunned to discover a stickleback fish that was undergoing parthenogenesis and carrying unfertilized eggs."³  The paper he cites says the opposite:  "Genetic and physiological examination of the parent fish and offspring ruled out development by parthenogenesis and hermaphroditism, therefore implicating internal fertilisation by a male stickleback."⁸⁴  Thomas's "four new parthenogenetic 'species'" were, in the authors' words, four lineages that together "constitute a laboratory-generated vertebrate species."⁸⁵  McLatchie's account of the worm experiment turns self-fertilising worms into an asexual group, as Part 4 noted.

Fair and unfair.  It is fair to say that biologists have not agreed on why sex persists.  Hoekstra wrote in 2005 that "we are still far from a definitive answer," and Otto in 2009 that more data are needed to "settle the paradox of sex."¹¹  It is fair to say, with Maynard Smith, Wilkins and Lenormand, that the origin of meiosis is unexplained.  It is not fair to say, as Sanders does, that the hypotheses lack "any observable evidence."  Parts 4 and 5 describe the evidence, much of it published before he wrote.  And it is not fair to quote the 1970s and 1980s statements of the problem as if no work had followed them.

Objections and responses

"Sex could not evolve gradually, because male and female had to appear together."  This is the argument of Brown, Gitt, Batten and Stroud.  It rests on the premise that sex needs two sexes.  It does not:  isogamous algae and fungi mate as equal cells of two mating types, and in the volvocine algae a single mating-type gene now controls sperm and egg (Part 8).  The argument is wrong as stated.  Its narrower form, that no one has shown how mating types and meiosis began, is right (Part 7).

"Evolutionists have no observable evidence for any hypothesis about sex."  Sanders makes this claim.  Clones replacing one another under parasite pressure in a lake, selfing worms dying under coevolving bacteria while outcrossers survive, yeast with occasional sex adapting faster and shedding harmful hitchhikers:  each is observed (Parts 4 and 5).  What is missing is evidence that any one mechanism pays the twofold cost of males in general.

"The twofold cost is a myth;  sexual and asexual females are not really equal."  A skeptic might argue that the cost is overstated.  In the one system where it was measured, the mud snail, the fitted cost was 2.14 with a confidence range from 1.81 to 2.55, and the all-else-equal assumption held at the expected sex ratio (Part 2).  In the lake it was lower in some years, and it does not apply to species without males.  Where it applies, it is about two.

"Long-term benefits like faster adaptation explain sex."  A clone with no deficit takes over a population of a million in about 20 generations.  A benefit that only shows over thousands of generations arrives too late.  This is why Maynard Smith, Williams and the modern reviews all look for immediate advantages, and why the parasite evidence matters more than the adaptation-rate evidence.

"Animals cannot change their way of reproducing, so they keep sex because they must."  Brian Thomas proposed this in 2015.  New clonal lineages arise often.  Laboratory crosses of whiptail lizards made one in a single generation, as Thomas himself reported in 2011, and clonal mud snails, water fleas and stick insects have arisen repeatedly from sexual relatives (Parts 1 and 6).  If cloning were impossible, the cost of sex would not need explaining;  it is possible, so it does.

"Asexual lineages should all be extinct if evolution is true."  The mainstream prediction is not that every clone dies at once.  It is that clones arise often, collect harmful changes faster, and mostly do not last.  The measured pattern fits:  asexual lineages are mostly young, and several carry excess amino-acid changes.  The oribatid mites are a real exception that the mainstream account has to explain (Part 6).

"Ancient asexual bdelloids show that sex is unnecessary for tens of millions of years."  A skeptic of the mutation theories might use the bdelloids this way.  The genome data now show gene conversion, foreign genes, and signs of exchange between individuals, and the authors of the first genome wrote that rare sex is not excluded.  Their age rests on amber and clocks.  The bdelloids no longer show that animals can do without any genetic exchange.

"Shared meiosis genes prove that meiosis evolved once from a common ancestor."  The genes are observed;  the common ancestor is inferred.  A young-earth biologist would read the same pattern as a shared design for a shared function, as with other conserved machinery.  That reading predicts the sharing.  It does not by itself predict why parasites such as Giardia keep meiosis genes without visible sex, which the mainstream reads as cryptic sex or recent loss.

"Genetic entropy says sex cannot stop decline, so sex is irrelevant."  Sanford's position is that sex slows decline and is designed.  That fits the measured advantages of sex over cloning in Parts 3 to 6.  Whether any population can halt decline belongs to the genetic entropy article.

"Creationists explain sex by fiat;  that is not an explanation."  The skeptic is right that no creationist source found here offers a mechanism or a testable prediction about the persistence of sex.  Sanders says both modes of reproduction were created at the beginning.  That is a statement about origin.  It does not say why clones and sexuals coexist in the same lakes and why parasites shift the balance, which is the persistence question.  On persistence the creationist side has offered critique, not a model.

Issues and unanswered questions

1. The cost has been measured in one species. The mud snail estimates are good, but no comparable direct measurement exists for an insect, a vertebrate or a plant with separate sexes. Gibson and colleagues' lake estimates, lower than the tank estimates, show the cost can vary with how generations overlap.
2. No mechanism has been shown to pay the twofold cost in general. The parasite evidence is strongest, and it comes from one snail and one worm. Otto and Nuismer's model says it needs strong selection per gene, and how often nature supplies that is unknown. The finite-population theory is the reviewers' favourite, and its own advocates say whether it is strong enough "is an open question."
3. The harmful-mutation rate is uncertain where it matters. The fruit-fly and nematode figures that sink Kondrashov's hypothesis come from 2000-era divergence methods; the human figure was revised several-fold downward for protein genes and stands near 2.2 for the whole genome. Direct family-based estimates for short-lived animals with males would settle it.
4. Epistasis is measured in sign more than in size. Sohail's human and fly result shows synergistic interaction among rare damaging mutations, but at a size far below what would pay the twofold cost on its own.
5. The yeast results disagree on when sex helps. Goddard and Gray found benefits only under stress; Zeyl and Bell found them only in the familiar medium; McDonald found them in a benign medium. The designs differ, and no one has reconciled them.
6. The ages of asexual lineages are clock estimates. Every "ancient" asexual is dated by molecular divergence or amber. A young-earth model needs to say how much of the observed divergence between allele copies in bdelloids it attributes to creation and how much to change since. None has done so.
7. No intermediate stage of meiosis is known. Every proposed origin of meiosis is a scenario without a living or fossil intermediate, and the four main scenarios disagree about the order of steps. Neither side has a model of the first meiosis that makes a testable prediction.
8. The creationist side has no model of persistence. No creationist or design source found here proposes a mechanism by which sexual populations hold off clones, or a prediction that could be tested against the snail and worm data. "Created that way" answers the origin question for a creationist; it does not answer the persistence question, which is about why the balance between clones and sexuals in a lake shifts with parasites.
9. Sources not obtained in full. Kondrashov 1988, Lehtonen and colleagues 2012, Keightley and Otto 2006, Jokela and colleagues 2009 and Lively 2010 were read in abstract only; Salathé and colleagues' 2008 review of the Red Queen's conditions and Maynard Smith's 1971 paper on the cost were not read. None is quoted beyond its abstract.

Verdict

Why sex persists:  unfinished on both sides.  The cost of males is real and measured at about two.  Mainstream biology has shown, in particular species, that sex pays:  parasites drive clones down in a lake and kill self-fertilising worms in the laboratory, and occasional sex speeds adaptation and keeps harmful mutations from hitching a ride.  It has not shown that any of these pays the full cost of obligate sex with males across the species that have it, and its own reviewers say so.  Harmful mutations alone cannot pay it in short-lived animals, by a margin of about twentyfold.  The creationist claim that there is "no observable evidence" is wrong.  The creationist claim that the problem is unsolved is right.  And the creationist side has offered no mechanism of its own.

How meiosis began:  the creationist argument is right that it is undemonstrated.  The genes for meiosis are shared by living eukaryotes, which mainstream biology reads as inheritance from an ancestor that already had full meiosis.  Nothing living or fossil shows an earlier stage.  The four published scenarios disagree about which step came first, and their authors call the origin unexplained and untestable by direct means.  Calling it impossible goes further than the evidence:  no one has shown that the proposed intermediates would be harmful, and archaea show cell fusion and DNA exchange without meiosis.

Whether male and female had to appear together:  the creationist argument is wrong.  Sex exists without sexes, and in the volvocine algae one mating-type gene now sets sperm and egg.  How mating types and the coordination of fusion with meiosis began is part of the unanswered origin question, not evidence that both sexes had to arise at once.

The sibling articles on multicellularity and on nervous systems and animal body plans take up the transitions that follow;  sex and complex multicellularity appear together in the fossil record from Bangiomorpha onward.

Sources