Friday, August 24, 2012

But but but, we don't even need to explain the evolution of menopause!

The latest of the adaptive arguments for why menopause evolved appears in the 22 Aug Ecology Letters ("Severe intergenerational reproductive conflict and the evolution of menopause," Lahdenperä et al.). A discussion of the work in this week's Nature points out that humans, killer whales and pilot whales are the only animals known to stop reproducing before they die.  So of course the question is why.

Previous adaptive explanations include the "grandmother hypothesis" which suggests that reproductive fitness is higher when women stop having their own children but can then take care of their children's children.  And the "mother hypothesis" which suggests that mothers gain more in terms of fitness by investing in their older children than by having new ones.  Alternatively, the increased risk of dying in childbirth for older women might be the explanation.

The Shattuck Family;
Aaron Draper Shattuck,
Brooklyn Museum
The first two are basically inclusive fitness explanations: what I contribute to the group is good for me.  (Yep, karma.)  That's because relatives share the same genes you have, so if you help them it's like helping proliferate your own genes.

Inclusive fitness is a hot topic these days, because the basic argument is shaky in terms of how and when it actually applies, and we won't go into that here.  But we will say that the mother and grandmother hypotheses have run into mathematical issues, as fitness differences at late age in populations with little survivorship to that age aren't enough to explain the trait given that grandmothers share 1/2 the genes of their own children, but only 1/4 of their grandchildren's genes.  The extra fitness, if it were even measurable and and and were due to some major gene conferring old-age survivorship, would be so low that drift would over-ride it.  There just would not be much, if any, selection pressure to live to an old age for that evolutionary reason.  These are well-known issues, though answered by simply ignoring them by the behavior-evolution community.

In this new study, Lahdenperä et al. made use of a 200 year data set from the Lutheran Church in Finland. Their sample included 653 women born during 1702-1823, who gave birth to 4703 children, of whom 1736 then had 9164 between 1757-1908, but they drew a smaller sample to look at intergenerational reproductive overlap, defined as grandmothers giving birth within 2 years of the younger mother. They specified 2 years, before or after a birth to a mother/mother-in-law or daughter/daughter-in-law, because they assumed that was when the mothers/mothers-to-be would have the most conflict over resources. This left them with 209 mothers who gave birth to 613 offspring, and of the offspring, 342 were mothers who gave birth to 824 offspring.

They performed a number of analyses to determine the effects of numerous variables on fitness (details in the paper, which is open access), controlling for age, sex of offspring, maternal age, birth intervals (number of years between births), as well as for "potential effects of maternal presence, living area, social class and birth cohort." Of course, sample sizes would have been trivial for detecting whether this potential opportunity for selection was realized in any genetic terms.

 They then calculated inclusive fitness, a woman's total fitness counting her own surviving children and the fitness benefits she accrues from helping her relatives. Mind you, reproductive overlap was rare in pre-industrial Finland; only 6.6% of the 556 mothers who had at least 2 births gave birth within 2 years of a grandchild. Of these 30 or so women, offspring survivorship was not affected when mother/daughter pairings were concerned, but survivorship to age 15 of the overlap offspring in mother-in-law/daughter-in-law comparisons was found to be statistically significantly lower. "These results suggest that intergenerational reproductive conflict is low among related mothers and daughters, but is substantial between unrelated in-laws."

Evolutionary non-sense, and hence nonsense!

This study is from modern, even if pre-industrial, pre-contraceptive times, and is absolutely irrelevant to the argument trying to explain human longevity, the necessary argument for the hypothesis (because only if you live longer can you take care of your grandchildren). It is absolutely irrelevant to evolutionary conditions that have to have been in Africa more than 100,000 years ago (because all modern humans share the trait). It is not just menopause that must be explained. And Jim Wood and colleagues showed more than 10 years ago that there is no specific issue about human menopausal age. The process is essentially the same in humans as in mice. Beyond that, of course, not many people survived to experience menopause, or put another way, to provide selective pressure not to stop ovulating.

The picture is even worse. Not too many years ago, a story appeared in The Lancet that showed that it is grandparents who are economic burdens on their grandchildren, quite the other way round from this study and the evolutionary hypothesis. Should it not be at least as general in hunter-gather times, that those too feeble to care for themselves not burden the resources of their grandchildren--even if the former occasionally do some baby-sitting? Since very-elders were rare, this would seem comparably, if not much more, the story. But even here, the Finnish story is from modern-times, and wholly irrelevant to any evolutionary speculations.

This is a fanciful evolutionary hypothesis, and cute and consistent with relentless Darwinian selectionism as it is, it is non-sense when it comes to any actual evolutionary evidence, including for the reasons given above. The urge to explain the evolution of menopause is beloved of darwinian determinists, but it has never had legs to stand firmly on. It is so difficult to support genetic causal arguments for complex traits like survival and menopause, and so difficult to find specific adaptive evidence of this sort in genomes, and the relevant energy and resource expenditure arguments so vague, that really, this simply makes no scientific sense.


Bjørn Østman said...

Thank you for that! I am so frustrated when I hear people talking about menopause in evolutionary terms.

Ken Weiss said...

There's more. For a mammal of our size, we do not even live particularly long or unusually long. This fact is entirely relevant because it shows there is no need for an adaptive explanation for our longevity.

There are many details, and I have written papers about this---more than 20 years ago!--but the press to find human exceptionalism is so strong that such papers are simply ignored. It's incovenient for those who want this hypothesis to be true.

Scott Garren and Heather Shay said...

Most discussions of inclusive fitness in humans completely miss the crucial issue of the evolution of social fitness. Humans don't live and die alone. The most fit individual in a family or clan that does not function well cooperatively will not out reproduce a middling individual in a highly functional group. Grandparents are important in the transmission of cultural traits that help the group function cooperatively. Even if an individual in a poorly functioning group does reproduce prolifically the group as a whole will be out competed by more functional groups for valuable resources in the environment and in the long term such genes will not prosper.