Pre-term Births

The model that Greg is dancing around suggests (1) that there is variation in mutation rate dependent on temperature or something correlated with temperature, (2) higher mutation rates cause a higher genetic burden in human populations, (3) leading to IQ reduction and other minor dings.  One interesting possible ding is premature birth, which affects 10 to 12 percent of pregnancies among  Whites and Hispanics in America and about 18 percent of pregnancies among US Blacks.  The trio of prematurity, low birth weight, and infant mortality is an especially conspicuous burden for the Black population.

It is easy these days to pull relevant data off the web, bearing in mind that data quality varies a lot in unknown ways.  The scatter plot  below  shows average temperature by state for the US, on the x axis, and rates of preterm birth on the y axis, for non-Hispanic Whites.


The simple correlation between temperature and prematurity is 0.55, not bad for social science.  There is no real correlation at all for the US Black population and only an insignificant hint of one for the US Hispanic population.

This number should not be regarded as an ordinary correlation because there is spatial autocorrelation among the states.  For example, should Mississippabama be treated as one state or two?  How long has the DNA of people in a state been in the state?  For example one suspects that much of the White population of Florida has only arrived within the last one or two generations.  There are a lot of other ways that one could pick at these data.

The overall pattern is roughly mirrored by state IQ estimates or their proxies like NAEP scores, by infant mortality rates, low birth weight rates, and so on.  In color coded maps I often notice a kind of V of lower IQ and worse health in US Whites, the base of the V in Mississippi with the right arm extending up the Appalachians to West Virginia and the left arm, less clearly defined, reaching to Colorado.  A nice source of such interactive maps is at this link.

PS:  The jpeg of the scatter plot is not very good: a direct link to a pdf is here.

This entry was posted in Uncategorized. Bookmark the permalink.

52 Responses to Pre-term Births

  1. Paul Jaminet says:

    If the black prematurity rate is almost double the white one, whereas the range among states is only 50%, then you should plot whites-only prematurity rate or blacks-only to see if ambient temperature is having an effect. This would help distinguish ancestral/evolutionary ambient temperature (higher in African ancestry) from contemporary ambient temperature.

    There are other factors besides mutational burden that can account for these patterns. Mosquito and tick populations are higher where the average temperature is higher; so are infectious disease burdens. Lifespan is negatively correlated with average temperature, and I would bet it’s mainly due to infectious disease burden.

    My guess is that contemporary ambient temperature would have little effect on prematurity rate after accounting for ancestral temperature (thus mitochondrial haplogroup) and infectious disease burden.

    • harpend says:

      The plot is Whites only.

      Prematurity seems to be sensitive to inflammation. We wrote a paper a few years ago suggesting that descendants of tropical Africans have the immune system “turned up” and are more prone to inflammation while descendants of Native Americans have the immune system turned down because of low pre-Columbian disease load. We suggested that this is a big cause of ethnic differences in prematurity rates in North America.

      The paper is
      title={Group differences in proneness to inflammation},
      author={Pennington, R. and Gatenbee, C. and Kennedy, B. and Harpending, H. and Cochran, G.},
      journal={Infection, Genetics and Evolution},

      • says:

        That would also explain the higher rates of autoimmune diseases such as Lupus among African Americans.

      • Paul Jaminet says:

        Thanks! Interesting paper.

        • harpend says:

          It will be interesting to watch for reports about epidemiology of Somali immigrants to the US. If we are right it will not be like that of the US Black population since they are not tropical Africans.

      • Steve Sailer says:

        Abyssinians from the highlands of Ethiopia might offer a good test of heat v. winter theories. Central Ethiopia is at about 5 degrees of latitude north, but at quite high altitude so the climate is not sweltering.

  2. spandrell says:

    Isn’t it just that hotter temperatures make people mature earlier? IIRC early age of menarche is also correlated with temperature.

  3. Timtoc says:

    One of the physical correlates of prematurity is the length of cervix as measured by ultrasound during pregnancy. Shortened cervix is much more common in African American women than white women or Asians. There may be a gradient in cervical length that accounts in part for differences in pre-term births and prolonged labor among the racial groups.
    The Length of the Cervix and the Risk of Spontaneous Premature Delivery
    N Engl J Med 1996; 334:567-573February 29, 1996

  4. Jason Malloy says:

    Harpending sayeth: “The model that Greg is dancing around suggests (1) that there is variation in mutation rate dependent on temperature or something correlated with temperature, (2) higher mutation rates cause a higher genetic burden in human populations, (3) leading to IQ reduction and other minor dings.”

    See these posts from February and April for the conceptual background.

    While not fully or explicitly articulated, this is the first New Big Theory of race differences in quite a while, and an interesting alternative to the reigning sociobiological models available since the 1980s. In the latter models intelligence and reproductive differences are seen as consequences of natural selection in divergent latitudes, but this new model replaces natural selection with accumulated mutational burdens. The differences at lower latitudes are not selectively advantageous, but dysfunctional.

    Dr. Cochran notes that complex adaptive systems, involving the functioning of many genes, should be the most vulnerable to genetic load, so this would obviously be the brain and probably reproductive physiology. So in addition to higher general mortality, dysfunctions associated with mutational burdens might include:

    – Lower intelligence
    – Higher retardation
    – Higher mental illness

    – Lower birthweight
    – Higher premature births
    – Higher infertility
    – Higher reproductive deformities
    – Higher miscarriage (and general obstetric complications)
    – Lower sperm quality

    Of course there is a difference between establishing population differences in genetic load, and proving that this is related to population differences in socially valued traits. I’m not sold on this as a replacement for sociobiological models, although there are aspects that make it useful and attractive in different ways. For example, I recently found that ethnic differences in rate of homosexuality are inversely correlated with latitude. Since theories of selectively advantageous homosexuality fall flat, this theory seemed like a better fit.

    • Steve Sailer says:

      Another odd data point is that the surplus of males over females at birth among African Americans is only about 3% versus 4 or 5% for other groups. This is especially strange because black males from age 0 to 18 die at relatively high rates, leaving a surplus of black females during the mating years, which isn’t supposed to happen. I’m never known what to make of this data, but it has always seemed like it might turn out to be an important clue.

      • Jason Malloy says:

        Several years ago I laid out an extended theory of sex ratio at Peter Frost’s blog. When there are more women relative to men, the social dynamics generate a more male-biased mating system where men aggressively compete with other men for multiple, short-term mating opportunities (There is a large empirical literature supporting this). Low female dependency and high parasite burdens in Sub-Saharan Africa facilitate a male-biased mating system, and various physical and behavioral adaptations guide this along. Female-biased sex ratios are one of these adaptations; when you compare children from black-white interracial pairings, the sex ratio at birth is determined by the race of the father.

      • Steve Sailer says:

        Jason, thanks. Now, can your theory of sex ratios be integrated with Greg’s theory of heat-caused mutations?

      • TWS says:

        Isn’t supposed to happen in a monogamous system. In a polygamous situation it makes no difference maybe it’s even ‘better’ because more ‘big men’ are getting more females.

      • Jason Malloy says:

        Sex ratio is part of a larger sociobiological package. And maybe. Greg’s theory predicts lower functioning on certain traits, but not others (cf. complex adaptive systems’), although I’m not entirely sure which ones and why. The sociobiological theories have tended to focus on traits that blacks excel with, like athleticism and creativity (e.g. cultural influence), related to the mating system I described above. Greg might argue that these traits wouldn’t take the same dings as intelligence (and others like blogger n/a would probably challenge most evidence for black strengths vis a vis temperate populations to begin with). But there are other traits that it probably should effect. For example, blacks appear to have higher sperm quality (and a dozen traits that correlate with it), and this appears to result in faster pregnancies. Which you would expect from a population with a promiscuous, male-biased mating strategy (conversely, monogamous East Asians have terrible sperm quality). Granted, though, on most of the vital reproductive processes listed above, blacks have a lot of problems that sociobiology doesn’t really address. Rushton hand-waves away black fertility problems by pointing to African birthrates; fitness-reducing dysfunction is as fitness-reducing dysfunction does, no? He also argues that most of these problems are caused by STDs. Although do blacks have more STDs simply because they are more promiscuous or are they more susceptible to STDs because of sub-optimal reproductive functioning? (I tried to spin this one in favor of Greg’s theory, but he seemed reluctant to agree that genetic load had much to do with black STD rates.)

        Another issue I’ve raised is that if blacks have higher genetic load they might have specialized adaptations to counter-act its effects. For example, blacks have higher bilateral symmetry than other populations, which is what sociobiologists equate with “good genes” or developmental stability. One example of this is height. Sub-Saharan Africans are much taller than they should be given the conditions in Africa. They don’t get much taller when they are transported to healthy Western environments, and the variation in height between African nations doesn’t seem to correlate much with the variation in conditions. Intelligence may actually be another example, if Wicherts et al. are right about Sub-Saharan Africans having intelligence scores not much different than African-Americans (though my own estimate is closer to 75 instead of 80). It would mean that, although the African IQ is low, it doesn’t develop much lower in abject poverty than it does in white adoptive households in Minnesota.

        Dr. Cochran, however, is skeptical that high developmental stability says much about the effects of mutational burden. I confess that my intuitive grasp of the theory begins to drop here.

      • Jason Malloy says:

        (I tried to spin this one in favor of Greg’s theory, but he seemed reluctant to agree that genetic load had much to do with black STD rates.

        By the way, I apologize if this was unclear. I did not discuss this theory with Rushton. His arguments about black fertility problems were taken from his two 1989 exchanges with Michael Lynn in the Journal of Research in Personality. It was Greg himself who didn’t seem to think his theory held much relevance for black STDs relative to good old fashioned promiscuity. (However I have long suspected that genetic differences play a much more important role here, and find Greg’s theory a potentially fruitful explanation.)

      • n/a says:

        “For example, blacks appear to have higher sperm quality”

        Jason, as far as I know, this is not the case. The little direct evidence I’m aware of tends to point in the opposite direction. E.g.:

        Could the differences in sperm concentration in various geographical areas be related to the ethnic origin of the studied populations? From the results shown in Table 3 it seems that the MSC of healthy men might be lower in Africa and South East Asia compared to other parts of the world.

        All sperm parameters were significantly lower in the small subgroup (n = 7) of African-American men compared with other men in this population (p-values for sperm parameters, < 0.001 to 0.016).

        This finding would be consistent with either higher genetic load or an evolutionary history with higher rates of true polygyny (gorillas have small testes) among blacks, though I don’t necessarily subscribe to either explanation.

        “and this appears to result in faster pregnancies.”

        Might be more easily explained by lower rates of breastfeeding and/or less-effective use of contraception.

      • teageegeepea says:

        At my prodding (due to Malloy’s writing) Robin Hanson dug up some studies on the relation between sex-ratio and violence. He concluded the evidence was unclear.

    • Steve Sailer says:

      One potential problem with this theory is that it sounds like we ought to see blacks having a long, fat right tail to their bell curve, with a bunch of people out at 145 or 160 IQ who got lucky and didn’t have many deleterious mutations. It’s like how in the physical realm, we see African Americans having higher than average rates of prenatal and infant health problems, but also producing more NBA power forwards, who, presumably, were the ones who didn’t suffer from physical problems.

      Instead, though, we just don’t see a lot of extremely high IQs among African Americans. It’s more like how there aren’t a lot of Japanese players in the NBA: they’re just a little shorter on average.

      • gcochran9 says:

        You have to look at the distribution of mutations. Most mutations in a coding gene, one that makes a protein, are synonymous – they don’t change anything. Some do: they’re called nonsynonymous. Most such mutations have mild negative effects: they don’t ruin the protein. Some do: stop codons, frame shifts, etc. More mild changes occur than severe changes… Perhaps five times as many.

        The unobvious thing is that the mild mutations cause just as much trouble as the severe changes, because they persist so much longer. Because their deleterious effects are mild, it takes a long time for selection to remove them. At equilibrium, the class of mutations that cause a 1% drop in fitness exists at a level 100 times the per-generation creation rate of such mutations. On the other hand, the equilibrium number of lethal mutations is exactly the same as the per-generation creation rate.

        Most of the bad effects of genetic load ought to be caused by the numerous, mildly deleterious mutations. Some are caused by more severe, rarer mutations. The effects of those more severe mutations should vary more: fewer big dice, instead of lots of little dice. If that is so, then as you go farther into the low end of the bell curve, more and more of the kids will have one or a few seriously bad alleles. Those alleles may well be pleiotropic, which means that they push your phenotype in an odd direction, as well as lowering your IQ. You’re a funny-looking kid. If you just had considerably more of the slightly deleterious mutations, which are so numerous that they don’t really push in any particular phenotypic direction other than inefficiency (law of large numbers), you could have the same low IQ score without being funny-looking…

        Someone from group A, with a substantially larger genetic load than group B, would almost never have as low a load as a 3 std low-load person from group B.

      • Steve Sailer says:

        Okay, Greg, but why don’t we see a situation with intelligence the way we do with physique? Presumably, LeBron James — 6’8″, 260 pounds, and highly nimble — happened to not inherit many mutations that undermine his body. But where are the black LeBron’s of the intellect?

      • Chuck says:

        Using VES, Gottfredson et al. investigated the mutant load hypothesis in “Does a fitness factor contribute to the association between intelligence and health outcomes?” The analysis included only Whites. One way to test Greg’s hypothesis would be to see if Blacks fell on the same regression lines as White for IQ and the various health outcomes. The authors of the above paper might have run this analysis (I would have) — so you could contact them and inquire.

    • random mutation says:

      Jason Molloy said:

      However I have long suspected that genetic differences play a much more important role here, and find Greg’s theory a potentially fruitful explanation.

      While I think there are problems with the notion of differential genetic load idea, especially since it seems to imply that different groups somehow have different amounts of genetic load in different areas, I wanted to say that I have begun to think that there has been selection for genetically-based behavioral differences with respect to certain cultural complexes, like religion.

      It would not surprise me if gene-culture interaction has produced different engines that are each more in tune with the different cultural complexes out there.

  5. harpend says:

    Thanks, Jason, I wish I had said that. (This after my 16 year old kid just chewed me out for bad writing and lack of context for my post.)

  6. random mutation says:

    By similar reasoning, the reason that whites are not so hot at sprinting is due to a higher genetic burden among whites.

  7. wijjy says:

    What is the mechanism?
    Is heat itself is the problem? You need to check month to month data
    Does disease removing resources from pregnancy? There may be rural urban differences
    Or is it just a correlate with wealth? Correct the plot for income

    • harpend says:

      The mechanism may be something along the lines that Paul Jaminet suggested in the first comment, or it could simply be temperature. If the mutations at issue are old, thousands of years old perhaps, then month to month data would not help at all.

      Almost certainly PMB is a correlate of wealth, but is it not more likely that health and IQ cause the wealth differences we observe? If we “correct” for atmospheric pressure then Death Valley is at the altitude as Pike’s Peak.

      • wijjy says:

        If there is a crucial period in pregnancy when temperature matters then it may be detectable by comparing births in each month, collating over years. So compare every january birth to every July birth say.

        Something similar was done to investigate temporatal patterns of cot death (mainly winter).

  8. billswift says:

    >If the mutations at issue are old, thousands of years old perhaps, then month to month data would not help at all.

    That would depend on whether it was a longstanding mutation for a hot climate, or whether it was an interaction between genes and current heat.

  9. bg says:

    Here’s what I don’t get.. If Africa is so bad for the evolution of intelligence (either via mutational load, or via natural selection), then why did the greatest intelligence leap ever (human intelligence) evolve there?

    Is it because the climate of Africa in the past was much more hospitable? Or is there a more complex theory whereby its a necessary place for evolution to human intelligence evolve, but after it’s emerged it thrives more up north?

    • Paul Jaminet says:

      A good question. Here are my thoughts:

      First, higher mutational load would have accelerated evolution in the Paleolithic, since hominid populations were so small that it took over a hundred thousand years to sample mutation space. So beneficial mutations would have appeared first in Africa in the Old Stone Age. Even if average fitness was impaired, you’d still expect to see evolution in brain size originate there.

      Second, there’s no good reason to think that the genetic advances of the later Paleolithic occurred in Africa. There’s just as strong a case for them occurring in Arabia or the Levant, where gene flow from Asia and Africa would have met to create interesting combinations. Certainly the major demographic expansions seem to have come largely from that area in the last 100,000 years maybe even the last 500,000.

      So I would expect that as long as hominid populations were small, mutational advances would occur first in Africa. Once hominid populations grew toward 1 million, then the locus of beneficial evolutionary innovations would move toward the Middle East.

  10. ironrailsironweights says:

    What about immigrants from tropical but non-African locations, such as SE Asia, India or parts of Latin America? Do they have greater rates of prematurity that might be attributable to climate-based mutations?

  11. random mutation says:

    Aren’t you simply assuming that 40 weeks (266 days) is the human gestational period?

    Could it be that Sub-Saharan Africans have a 38-week gestational period on average?

    I note that on Wikipedia Chimps are listed as having a gestational period from 230-250 days. However, humans are listed as 266 days. That seems unlikely.

    • harpend says:

      Yes, you are right of course. And US Black birth weights are somewhat lower than those of US Whites, and for a given lower weight US Black babies are at lower risk. IOW the optimum for US Blacks is somewhat lower.

      On the other hand the complex of prematurity and low birthweight is reflected in the much higher infant mortality of the US Black population.

      • A Lady says:

        African immigrants have higher birth weights than US black mothers, in total (not just college-educated immigrants, but refugees and other immigrants with potential for more marginal health status in pregnancy).

        I suspect the variance in risk for US-born black mothers is affected by maternal age distribution being quite different to the US-born white maternal age distribution.

      • Timtoc says:

        This is a interesting study of birth demographics from New York City with statistics complied from 1958 to 1961: There are Bell curves for duration of gestation and birthweight and the curves differed for White and Non White for nearly 650,000 deliveries with almost 135,000 Non White deliveries. That time period certainly featured inequalities in access to prenatal and natal care but is just before the technologic advances in health care would make those access issues as dramatic in outcomes.. The paper comparing birthweights in African immigrants and native born Blacks and Whites is here: There are significant differences in maternal age and married status for the immigrants.

  12. As I understand it:
    1) about 20% of genes
    are related to brain, so it
    would offer a much
    bigger target for
    mutation than \’physical fitness\’.
    2) the advantages in
    athleticism is confined to
    a small enough area of
    the genome to escape
    significant LOF mutations.

    • ziel says:

      that was my thought – for example, a boost in athleticism can be gained pretty easily via chemical intervention. East German ladies did so routinely in the 60’s. But chemically induced improvements to cognitive functions are elusive – temporary at best.

      Plus, high levels of physical fitness and performance are traits that have been perfected over 50 million years of mammalian evolution = or is it 250my of vertebrate evolution? Abstract thinking, on the other hand, only originated in humans, and thus resistance to the cumulatively baneful effects of random mutations has not been well developed.

      On the other hand, since I thought of these myself, my default assumption is that they’re wrong:)

    • Jason Malloy says:

      But is there good reason to believe, e.g. the normal range of intelligence differences we find in human populations involve a lot more genes than, say, the normal range of height differences? “Hundreds of genes of small effect” is the standard phrase found in modern genetics papers to describe the architecture of quantitative traits.

  13. Correction. From ‘Typos’ :
    “IQ would decrease, probably
    more so than many other
    fitness traits, since more
    than half of all genes are
    expressed in the brain”

  14. Steve Sailer says:

    People from the Ethiopian highlands might be a good test of the competing theories of Heat v. Lack of Winter: Addis Ababa is only 5 degrees north of the Equator, but the temperature is not high at all. The warmest month is March with an average high of 77 f. The coldest month might be January with an average low of 43.

  15. Nador says:

    Shouldn’t one assume an ideal rate of mutation to exist? (Obviously that rate would depend on time and space). I mean: a lineage with an extremely high rate of mutation would likely die off, but wouldn’t that be the fate of a lineage that doesn’t change at all (assuming we are not talking about rotifers).
    The Red Queen hypothesis for sex is the need to change (mix) so that the fast reproducing parasites can not adapt too well to the host. If this is applied to the higher parasite load of the tropics, one should expect higher optimal rate of mutations there. It seems a more plausible explanation for the higher mutational load then heat, since mammals have a rather constant body temperature. (Not to mention the lower idle metabolic rate of Africans then Eurasians.)

  16. random mutation says:

    It is worthwhile going to some place where you can see African American kids and Chinese kids near each other so you can see the differences.

    In parts of Silicon Valley you can routinely see them.

    The first thing I noticed was that AfAm kids have small heads on top of robust bodies. Chinese kids, on the other hand, have enormous heads on top of dinky little bodies.

    Isn’t there also evidence that AfAm kids are physically more competent than white or East Asian kids at the same age, on average?

    It sure looks to me like different life histories resulting from different selection pressures.

  17. ttt says:

    Does anyone look at these mutational load issues together with miscarriage rates and reasons? I remember being told that in our modern age miscarriages are often/usually because of defects in the baby. I’m sure there have been some studies on different rates among countries and groups; I’m wondering more if there have been large studies of genetic defects in miscarried babies and whether those may say more about different groups.

  18. j says:

    Nowadays. the most common length of a pregnancy in the U.S. is 39 weeks and not 40. There is also a dramatic rise in late preterm births, or babies born between 34 and 36 weeks gestation. The trend is clear. Or is it “Random Mutation”?

    • Sideways says:

      My first thought would be changes in population distribution, with so much more of the country in the Sun Belt now

  19. Matt says:

    It occurs to me that one prediction from this model, though, would be decreased mutational load in Neanderthal and Denisova cave humans, who would have been resident at cooler latitudes for a very long time, compared to Homo Sap, of tropical to subtropical origins.

    Unless perhaps population bottlenecks in these populations would lead this to not be the case(?).

    If that is the case, it would be surprising for Neanderthal not to show any signs of being particularly smart(?). And that we aren’t more descended from Neanderthal than African Sap.

  20. hbd chick says:

    @henry – “that there is variation in mutation rate dependent on temperature or something correlated with temperature….”

    if correct, implications for the out-of-africa theory are…? (seeing as most of africa lies in the tropics.)

  21. Eric Rasmusen says:

    Wouldn’t a higher mutation rate also imply faster evolution via favorable mutations? Hence, I would have predicted lower average IQ— due to bad mutations— but higher top IQ– due to good mutations— and perhaps higher average IQ in th long run (tho with a steady-state bigger number of very low IQs).

  22. Pingback: Evidence for Cochran’s & Harpending’s Theory of Genetic Load: The Link Between IQ and Attractiveness | JayMan's Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s