I criticized the Zietsch and Santilla paper on the female orgasm. Now one of the authors has responded.

One response he makes is that some of the limitations to the study that I pointed out were also explicitly recognized in the paper. This is true; however, my purpose in mentioning them was to highlight the fact that they make it impossible to draw even the tentative conclusions the authors do…which obviously is not something that was done in the paper. Admitting that assessing orgasmic function with self-reports, for instance, is a limitation doesn’t really change the fact that extremely weak evidence was published to support a particular hypothesis.

Another problem I raised is that the comparisons between male and female orgasmic response were inappropriate. They compared the timing of male orgasm to the likelihood of women having an orgasm at all. My objection is two-fold: they are using phenotype as a proxy for a genetic difference, which is problematic in a trait so strongly responsive to an environmental difference, and it treats two parameters, timing and likelihood, as equivalent in men and women. I don’t think these are necessarily directly connected at all. Zietsch’s reply emphasizes that he does think this is a valid comparison.

Indeed, we measured susceptibility to orgasm in response to sexual stimulation (let’s call it ‘orgasmability’) by assessing the likelihood of orgasming during sexual activity in women and the time taken to orgasm during sexual activity in men. That’s because during sex, men tend to reach orgasm faster than women and generally cannot continue once it is reached. Even when women reach orgasm faster than their man, they can generally continue intercourse until he reaches orgasm, sometimes achieving more orgasms. As such, women’s orgasm during sex is time limited – if she doesn’t reach it in relatively quick time, she might not have it at all, whereas men can go until they finish. That’s why we measure likelihood of orgasm in women and time to orgasm in men, consistent with countless other studies and definitions of orgasmic ‘dysfunction’ in men and women in DSM-IV.

How odd. If sexual activity is limited mechanically by the maintenance of the man’s erection, then yes, it would be true that women’s orgasm during sex is time limited. However, given that vaginal intercourse and female orgasm are only weakly connected, isn’t this an unfortunately male-centered perspective? I will confess that personally, if my sexual performance were measured only by time to orgasm, I’d be considered a pathetic lover (admit it, all you guys reading this: it’s true for you too), but somehow my sexual encounters go on considerably longer, to the delight of both participants. Human sexual activity is considerably more complex than wham-bam-thank-you-ma’am, as I’m sure Zietsch knows.

It’s also a self-destructive assumption. Was there ever a time in our evolutionary history when sexual interactions were limited by male time-to-orgasm? I suspect not; caveman/cavewoman sex probably involved a fair amount of courting and cuddling and playing, just as it does nowadays. And if it didn’t — if it really were nothing but 3-to-5 minute intromission and ejaculation sessions — then there was no opportunity for female orgasm to be selected for. So this is really all an irrelevant objection.

Also…extremes in variation in the timing of the male sexual response are considered dysfunctional because they deviate far from a solid norm. A man who takes a half-hour of focused stimulation to achieve orgasm is probably experiencing some real problems. A woman, on the other hand, who takes that long is not that unusual at all; she is not ‘dysfunctional’. She is normal. It’s a problem when your study assumes that a stable, normal, healthy condition in a woman is comparable to a dysfunctional condition in a male.

Now one central explanation I offered was that male orgasmic response was strongly canalized — that is, there had been selection for multiple genetic processes keying on a strong environmental cue, the presence of testosterone, that made the male response much more robust. The byproduct hypothesis postulates that the female orgasmic response uses the same genetic circuitry, but is more weakly expressed because the cue is largely missing. Unfortunately, we seem to be arguing past each other.

Myers makes some other points that suggest that although he read the paper, he didn’t read it very carefully, since he misses its main point. The by-product theory, as described in detail by Lloyd (2005), says that female orgasm is currently (and always has been) maintained by ongoing selection on the male orgasm. Selection can only operate on additive genetic variation, so if the male orgasm has zero heritability (i.e. zero additive genetic variation) as Myers suggests, then there is no selection on it and therefore no indirect selection on female orgasm. (Hidden genetic variation with no phenotypic effects in males but expressed in females, which Myers alludes to, is irrelevant here because it’s invisible to selection, assuming no direct selection on female orgasm). He goes on to talk about males and females sharing orgasm-related circuitry and genetic apparatus (which nobody denies) – but, to be repetitive, selection only acts on genetic variation – for selection on the male orgasm to act on the female orgasm, additive genetic variation in male orgasm needs to correlate with additive genetic variation in female orgasm. If such a correlation exists there would be a correlation between opposite-sex siblings, and that’s what we tested.

Well, male orgasm has almost zero heritability. There is a low frequency of dysfunction that can be selected against. But largely, it’s true, males hit puberty and they’re generally sprouting erections and ejaculating frequently; selection has done its job and given us guys a remarkably reliable physiology in that regard (and I for one say hooray for evolution). We males are so good at that part of sex that it’s unlikely that there is currently much selection going on on our side of the sexual divide to make orgasm more likely, and so you can remove us from the equation right now and for a long time in the past — we’re done, and all the evidence suggests that that part of our evolution was established at least since mammals evolved.

So Zietsch and Santtila went looking for some kind of significant variation in the male population, and found one in the self-reported variation in time to orgasm. Could there be natural genetic variation in that parameter? Sure. But my objection is that it probably is not significant (was selection for sexual performance ever so strong that males who ejaculated in 3 minutes had an advantage over males who took 5 minutes? I doubt it), and that a self-reporting survey on such a charged question would not produce valid results. 35% of their sample reported spending more than 10 minutes in active intercourse before orgasm, which ought to set off alarm bells right there.

But of course, we males are only half the population. There is known variation in the frequency of orgasm in women, Zietsch and Santilla found the same thing in their survey (note again, though, the unreliability of self-reporting), and we could imagine selection working on that variation. Males are done, as I said, with robust testosterone-dependent developmental mechanisms that assemble a reliable orgasm-generating machine, but there could be, for instance, selection for non-testosterone-dependent orgasm pathways in women, since there is variation in the population.

Only there doesn’t seem to be. There doesn’t seem to be a pattern of women who can orgasm 3 minutes after a penis touches their vagina being more reproductively successful than women who take 20 minutes of clitoral stimulation, nor is there any reason to think faster orgasms would make a woman more fertile. That’s the basis of the byproduct theory — a lack of evidence that selection can or does actually operate on the range of variation in the female half of the human population.

The heart of Zietsch and Santilla’s argument above, though, is this weak one, that unreliable self-reported data shows a lack of correlation in time to orgasm in males and frequency of orgasm in their female siblings. I argue that the variation they describe in the males is unreliable;it is also not significant, even if true; and they haven’t shown that the genetic basis of any variation is even relevant to the genetic basis of orgasm frequency in women. While the neural and physiological basis of orgasm may be shared in men and women (that is a foundation of the byproduct theory), the details of the regulation of the expression of the phenotype are also likely to be dependent on different genetic circuitry. I’d argue that there are multiple pathways in development leading to the formation of the orgasm response, and that all of them contribute to the male pattern, but a major contributor, testosterone-dependent development of the brain and reproductive system, is largely absent in women, leading to a greater reliance on auxiliary systems.

If you want to show that the byproduct hypothesis is false, one good way would be to find, for instance, an estrogen-dependent developmental process that contributes to the female orgasm. That’s what I’d like to see: evidence of a parallel pathway that would only be under selection in females. Showing that would at least be evidence of historical selection for activation of orgasm in women.

One more uncomfortable problem that I’m sure was unintentional: A male scientist writing about female physiology has his work criticized by a number of bloggers; he responds to two of the male critics but ignores a female critic. Again, it’s probably just chance, but it’s an omission that doesn’t leave a good impression. I’ll assume it was just because my argument was so much more magisterial by virtue of my entirely non-sexist authority that he had to reply to me.

(Also on FtB)