What you may call the hip or pelvis is actually formed by the joining of ilia, ischia, pubis bones to the sacrum and the coccyx. The shape of the human pelvis is unique amongst primates and part of the complex of anatomical changes which allow us for bipedal motility.
The Male Pelvic Girdle
Between males and females, the pelvis is significantly different. I’ll review some of the features that should be common knowledge to anyone with a forensic or physical anthropology background. For starters, where the two pubis bones meet, at the lower edges of the two inferior public rami, there is a feature called the subpubic angle. Males have an angle smaller than 90 degrees, while females have a larger one. You need not take a protractor to make this observation. If you got a pelvic girdle with subpubic angle wider than a right angle then you most likely have a female… anything smaller and you have a dude.
But there is more, under the posterior inferior spine of the illium, and above the ischial spine, exists a feature called the greater sciatic notch, a sort of passage way for the piriformis muscle and the sciatic and posterior femoral cutaneous nerves. In females the notch is broader than males. Another feature, the biiliac width, a metric measurement made from the widest point between the two ilia is also a sex determining feature. Relative to their overall body sizes, females have large biiliac widths. The two ilia seem to flare out wider in a female. With a wider biiliac width, comes a wider pelvic inlet or the circumference of the lesser pelvis forms.
The Female Pelvic Girdle
So why do females have larger pelvic inlet, width between the two ilia, and a larger subpubic angle? That’s because they give birth. A wider pelvis allows for better distribution of the added weight that comes during pregnancy. A wide pelvic inlet allows for more space to hold the baby in utero. Furthermore, passing the largest brain to body size mammal through a narrow pelvis would not only be painful, but poses a serious danger to both the mother and baby during childbirth.
The average female adult has a biiliac width of 28 cm. Certain populations, such as Greek women have biiliac widths of 27.5 +/- 2.29 cm, falling within the average (Steyn et al., 2008). Inupiat women have widths averaging 28.6 +/- 0.2 cm, Finns at 27.9 +/- 0.2 cm (Ruff et al., 2004). But, east Asian populations, such as the Japanese have smaller pelvises, with less variation. The average billiac width of women from Japan is around 27.2 +/- .02 cm (Ikoma et al., 1988).
This all makes sense, east Asian people are on average smaller than white people or people from Africa. In fact, anthropologists have regularly relied on estimating body size and mass from biiliac measurements. The average Japanese woman is 153 cm tall, while European women from Germany or the Netherlands average 166 cm in height. You can see such a distinction when comparing Finnish and Japanese pelvic girdles. Asian newborns babies are also have smaller weight at birth (3.2 kg) compared to white babies (3.4kg). A white woman with a wider pelvis can give birth to larger white babies.
So what happens when a white man, with big white genes, reproduces with a small Asian woman? Well, Razib pointed out a new study in the American Journal of Obstetrics and Gynecology which reviews the impact of such couplings. The paper, “Perinatal outcomes among Asian–white interracial couples,” documented that 33% of such couples surveyed had caesarean deliveries. The latest NIH data on the caesarean rate in the United States is 30%.
The authors suggest that the reason why such couples have 3% more C-section deliveries is that the smaller Asian pelvis is less able to accommodate babies of a certain size. The Asian-white couples had larger babies, with a median 3.36 kg for Asian-mother/white-father versus 3.21 kg for babies from Asian-Asian couples.
There’s a much larger discussion to be had than just reviewing a review of the anatomy and evolutionary history of such a study. Ever so recently, we hosted yet another post on the anthropology of race, which summarized that, “race does not exist in the world in any ontologically objective way.” If you’ve been a regular reader, you would know I’ve tackled this mantra many times. What could be anymore ontologically objective than such a study?
The nature of an Asian is on average smaller in body size than other humans. Of course there is variation. There are some large Asians, but the are very few. The majority are smaller in comparison to other humans. Studies like this show that Asian-mother and white-fathers produce larger babies and have increased rates of C-sectiond deliveries.
There are serious health issues with C-section deliveries, and thus serious, tangible biologically race related issues when people from two different populations mate and increase their chance of having a C-section delivery. The health issues I mention are the increased childbirth mortality rate. On average 1 in every 10,000 women who gives a natural birth will die during childbirth, while 1 in every 2,500 women who undergo C-sections will die during childbirth. In otherwords, women who give birth via C-section are 4x as likely to die. Furthermore, caesarean deliveries increase the risks for malpresentation, placenta previa, antepartum hemorrhage, placenta accreta, prolonged labor, uterine rupture, preterm birth, low birth weight, and stillbirth in their second delivery. The paper also outlines the higher prevalence of gestational diabetes for such interracial couplings.
That all being said, race is not just a social construct. How we interpret biological differences, such as pelvis size, skin color, etc. are not socially constructed but real observations, made from quantiative analysis. There are distinct anatomical, genetic, even behavioral differences that are not derived soley from stereotypes.
- Michael J. Nystrom, Aaron B. Caughey, Deirdre J. Lyell, Maurice L. Druzin,Yasser Y. El-Sayed (2008). Perinatal outcomes among Asian–white interracial couples in American Journal of Obstetrics & Gynecology 199 (4), (385.e1-385.e5) DOI: 10.1016/j.ajog.2008.06.065
12 Comments
October 6, 2008 at 1:15 pm
OMG. you are *SO* evil!
October 6, 2008 at 3:11 pm
Maybe it would help you understand if instead of using “social construct” we used the term “classificatory construct”. Pelvic width is a continuous variable – if we took a random sample of humans from across the globe we would have a continuous range of widths. Now if we then looked at whether there was a correlation between geographic location and width we might find some regions tend to group closer to one end or another. However, because the variation is continuous and so is geography it is up to the analyst to specify any “groups” and give them names like “Asian” – i.e. YOU draw the boundaries. Hence “classificatory construct”.
In medical studies like the one you cite no such sampling is actually done – in fact data is derived from large census databases in which ‘race’ is one of the fields. In such cases this involves ticking a box from a few restricted categories – i.e. it is predetermined rather than uncovered through quantitative analysis. Often filling out the ‘race’ field relies on self-reporting by the patient. This means that lots of social effects can get caught up in the classification, and so, many medical articles use the phrase “race/ethnicity” with the (implicit) caveat that the pattern or association shown relates only to individuals who self-report in a specific way. Practically this means a doctor should ask the patient what their race/ethnicity is, and then afterwards they can advise on risks etc associated with people who self-report in that way. The doctor should not act on their own analysis of what the patient’s race/ethnicity is.
With regard to the particular study you cite it is also worth reiterating the issue noted by Razib in his post: The authors discover a difference in caesarean rate and from this adduce that there is a difference in pelvic width. They do not actually report any data on pelvic width, or whether this is also correlated with the caesarean deliveries in their sample. To quote Razib “if the authors corrected for pelvic size on an individual level they might not need to make recourse to population level generalizations.”
You say: “How we interpret biological differences, such as pelvis size, skin color, etc. are not socially constructed but real observations, made from quantiative analysis.”
But in fact it is EXACTLY how we “interpret” those variations that is socially constructed. The variations exist, what we choose to draw from them, group them, or do with them is up to us.
October 6, 2008 at 4:20 pm
I think you don’t know what you’re talking about Tim, especially in regards to normal distributions. What you suggest is that there’d be an even frequency of biiliac widths across all people. There is not a continuous range of widths amongst all human populations. Certain populations have distinct clustering that set them apart from others when comparing.
Using the data from the papers I have cited above, I’ve constructed a chart that illustrates the specificity of biiliac width between Inupiat, Japanese, and Finnish women, check it out.
Kambiz
October 6, 2008 at 4:59 pm
OK, I can see that my phrasing has led you to misunderstand me. Any population, whether objective or arbitrary will have a mean and standard deviation, a particular distribution. Depending on how you define your populations the distributions may differ as your diagram demonstrates. My point however is that it is the definition of ‘populations’ which is problematic here. Imagine a 3D plot in which along one axis we have pelvic width, along another we have latitude and on the other longitude. You then measure the pelvic widths of a random sample of humans from the entirety of the Earth’s population. Would you expect to see Objective boundaries in your plot? I say boundaries rather than clusters because this is the primary difference between ‘race’ and ‘population’.
But even your Inupiat, Japanese, and Finnish populations are your own construct – they are predetermined by variables outside your analysis, rather than revealed by an analysis of worldwide pelvic width data. Are your Finnish pelvii different from Swedish or NW Russian pelviii? If not why are you treating them as a population?
October 6, 2008 at 5:24 pm
I don’t have the data for Swedish pelvii or Russian ones. But on average, including the standard deviation, the pelvis and overall body size of a person from Europe is larger than that of someone from east Asia. Let me remind you that there is a positive correlation to pelvic width and body size.
I acknowledge your argument of defining populations. It has always been rather troublesome to do so, especially nowadays with people of Asian ancestry living in Europe and vice versa. But, I’m comparing women from Finland to women from Japan and Alaska! These are people from different geographic locations.
If we factor in latitude, longitude, and altitude along with pelvic girdle width, I would expect to see some boundaries between populations. But pelvic girdle with is just one marker. Looking at many different morphological markers, we can see that isolated populations more closely resemble each other anatomically than outsiders. One example that comes to mind is the short, stocky barrel-like shape of the thoracic cavity of Andean populations, and others which reside in high altitudes. It is believed that the chest evolved that way as an adaptive response to hypoxic conditions.
Like I mentioned, I don’t have the data right now to answer your last set of questions. I’d venture to say, of course, there will be more gray area (overlap of the distribution and standard deviations) when you try to morphometrically compare Swedes to Russians to Finns. Some Russian measurements may even overlap with Asian ones. But when you compare a population from Europe, to one from Asia, you can see distinct differences.
Kambiz
October 7, 2008 at 2:36 pm
Kambiz, my point is simply that your population “Finnish” is a social construct. It is derived from the geopolitics of the modern nation-state. We know this because that area of land has been part of Sweden and part of Russia during various periods of history. You did not objectively discover this population from your analysis of pelvic width data.
I am well aware that if you compare a bunch of Europeans with a bunch of Asians you will see differences – it is blatantly obvious that human variation exists. This kind of comparison is after all exactly what happened during the Colonial period in which the concept of Race was developed – in their travels Europeans encountered people who looked systematically different from themselves. But this in itself tells us nothing about the objective reality or scientific utility of the concept of Race.
The point I have been trying to make seems to be too subtle for you, so lets try a different tack. The kind of population studies you are drawing on to argue for the objective/scientific validity of Races, are actually powerful antidotes to Racial thinking. They treat human variation as a statistical problem of clusters and fuzzy scatters, rather than one of bounded groups. They recognise that no objective boundaries exist in most lines of human variation at a global level. The most we can arrive at is polythetic statistical sets – but these are not by any means equivalent to Races, and cannot be used to argue for the objective validity of that concept.
October 8, 2008 at 10:24 am
tim, you’re a CAFT. btw, you do know that your individuality as a physical entity is arbitrary? you’re just a compound of nuclei and electron probability distributions. where does your distribution end? where do you draw the boundary? what is the appropriate cut-off value? arbitrary!
October 8, 2008 at 10:34 am
Tim,
I think we may both agree that Finns are people who speak Finnish, and 90% of people who live in Finland are recognized as being ethnically Finnish. That being said, people who speak a different language and have a different culture are a socially constructed group.
Last month, Razib reviewed a study which also showed that Finns have a distinct genetic population structure that is not observed in neighboring regions. How then are people who have a specific genetic composition a socially constructed group? How then are people who share a specific range in biiliac widths a socially constructed group?
Regardless, I see your point. I should have never called my groups by where they are geographically found. Imagine that I never called my population from Finland as “Finns.” Instead, I called them “Population A.” If I measured the pelvic widths of people from “Population A,” compared them to people from another population, and observed a difference, then that would be quantifiable distinction… A difference between the two populations. Noting the existence of biological differences between groups of people is not an antidote to racial thinking.
Kambiz
October 8, 2008 at 4:02 pm
Razib: a CAFT? What’s that? Coalition Against the Fur Trade? Customer Automated Financial Transaction? Huh? Although I have no problem with your characterisation of my body, I would point out that statistically the problem of its boundaries is a lot easier to resolve than the boundaries of human variation. What’s the Fst?
Kambiz, you ask how are people who share a specific genetic composition or a specific range of bi-iliac widths a socially constructed group. My answer is that they are not. Because in each hypothetical case the group is defined internally by the data under consideration – ie. by a “specific” range in that data. My question to you though is this what you are doing with your population of “Finns” or “population A”? I would argue that it is not. Instead what you do is select an arbitrary group of people (classified perhaps by some other measure: national boundaries, language, etc) and compare them to another arbitrary group. You have no idea whether they constitute a bounded breeding population or not. The difference you uncover is Relative, not Absolute – i.e. it is constructed by your comparison. You have sectioned two ends of a cline and compared them to each other. Hence this is not supportive of the concept of Race. It is merely supportive of the notion of human variation. And please note no-one has ever argued that human variation is ’socially constructed’. Be careful not to conflate this with Race.
I will now bow out of this debate, since I don’t seem to be able to make my point clearly to you. Most of the things I have been saying are well reviewed in a Wikipedia article which summarises much of the literature on these issues. It is the page called “Race (classification of human beings)”. I try not to recommend Wikipedia articles to my students, but since this is a blog and the page in question is informative I think it’s OK here.
If you want something peer reviewed please read this article in Nature Genetics by Keita et al “Conceptualizing human variation”: doi:10.1038/ng1455
This quite clearly explains why the population studies you cite are antidotes to racial thinking. With regard to population genetics you might consider the following quote:
“Y-chromosome and mitochondrial DNA genealogies are especially interesting because they demonstrate the lack of concordance of lineages with morphology[15] and facilitate a phylogenetic analysis. Individuals with the same morphology do not necessarily cluster with each other by lineage, and a given lineage does not include only individuals with the same trait complex (or ‘racial type’). Y-chromosome DNA from Africa alone suffices to make this point. Africa contains populations whose members have a range of external phenotypes. This variation has usually been described in terms of ‘race’ (Caucasoids, Pygmoids, Congoids, Khoisanoids). But the Y-chromosome clade defined by the PN2 transition (PN2/M35, PN2/M2) shatters the boundaries of phenotypically defined races and true breeding populations across a great geographical expanse[21]. African peoples with a range of skin colors, hair forms and physiognomies have substantial percentages of males whose Y chromosomes form closely related clades with each other, but not with others who are phenotypically similar. The individuals in the morphologically or geographically defined ‘races’ are not characterized by ‘private’ distinct lineages restricted to each of them.”
October 8, 2008 at 7:32 pm
Although I have no problem with your characterisation of my body, I would point out that statistically the problem of its boundaries is a lot easier to resolve than the boundaries of human variation. What’s the Fst?
so what if it’s *easier* how do you draw a line. it’s arbitrary and subjective.
October 8, 2008 at 8:21 pm
Razib: let’s say we subscribe to a quantum field theory and agree that I am just a probability distribution – how would I report this? Well I might report ‘me’ to 2 standard deviations and say the rest is too uncertain. This is of course arbitrary, but it is not equivalent to “drawing a line” (which I did not mention by the way): it is a statistically accurate descriptive convention. So no I wouldn’t draw a line.
October 8, 2008 at 9:00 pm
Sorry got interrupted there. By ‘easier to resolve’ I meant it is easier to agree on those arbitrary, but meaningful statistical conventions because of the steepness of the falloff in ‘my’ density – I might for example simplify my model and rely on the Solid, Liquid, Gas, Plasma, boundaries used by Particle Physicists. The problem with Racial differences is of a much greater order – conventional measures like fixation indices do not show values approaching anything like what we see in other non-human species for example, being well under the (also arbitrary) conventions used by wildlife biologists when they define non-human races. Which I guess is the crux of the matter.