keep in mind that height & IQ correlate at about .27 (Deary, 2005). But Jensen (1998) claims it's an exogenous correlation.

But Jensen (1998) claims it's an exogenous correlation.

yes. someone always points to the heigh r^2 and someone always responds with the jensen contention. from what i recall the effect doesn't show up among siblings, right?

r^2 ~ 0 within familes for height & IQ.

This sounds like the type of thing parents would try to select in their kids. Unless of course that "Tall Tax" arrives to balance everything out.

This sounds like the type of thing parents would try to select in their kids.

mebee. the only thing is that you'd need to hit so many genes to get a big return. i guess lots of screened fertilized embryos would do it....

(and obviously you need more loci, i don't think that the ROI is going to be high enough for a few cms)

If I remember correctly, short girls hit puberty earlier(genetically short, not stunted) so there may be fecundity selection for short women.

so there may be fecundity selection for short women

Matt at Behavioral Ecology posted on this, and there are studies that show shorter women being most fecund:

http://blog.behavioralecology.ne...ls-tall-anyway/

Tall women are good as trophy wives to show off and increase your status / reputation, which would increase your mating value. But when you want to get the job done and make babies -- the only way we know how -- you're going for petite babes, obviously on the sly if you have a tall wife / gf.

I don't get the "Height is 90% genetic" thing at all. Whole populations (Japan, Holland) have increased greatly in height with better nutrition. An enormous proportion of all people are between about 4'10" and 6'4", and a high proportion of the variation within that range is environmental -- certainly more than 10%.

The extremes for normal humans (3'6" to 4' 10", and 6' 4" to 7'6") are probably mostly genetic, but few people are in the extremes.

I suppose that if you take the whole normal range to be 48" 10% would be 4.8", but even then, I believe that the average Japanese height has increased more than 4.8" since 1840.

Probably this is just a misunderstanding of terminology. A pop way of clearly expressing the different technical meanings should be developed, though.

Re: selection for short women again, maybe this is a consequence of adopting agriculture. We know that hunter-gatherers tend to be taller than agriculturalists, pretty much across the board. Part of that is due to the higher burden of disease in agricultural societies, which stresses the body and stunts growth.

But with agriculture, it's possible to space births closer together / have larger families, and it's clear that agricultural societies out-reproduced their h-g neighbors over time. The phenotypic variance in family size probably increased, with lots more losers and winners, but that just makes it easier for selection to see who to favor.

If being short makes a female undergo menarche a few years earlier, that could amount to an extra kid she can squeeze out.

Another blessing of agriculture: girls like Patty Mayo.
http://www.youtube.com/watch?v=x...h? v=xddmaryAiGU

that could amount to an extra kid she can squeeze out.

I think you're missing part of the puzzle: if short women come to reproductive age before tall ones, they may have a fitness edge even if they each produce same number of children as the tall ones. This is most obviously true in a situation where the population is expanding, but even if you have a situation where the population goes through boom and bust cycles you would end up with the same effect. Shorter generations mean a shorter doubling time whenever there is expansion.

I don't get the "Height is 90% genetic" thing at all.

Take each individual's height, subtract it from the mean height in the population, square the difference, and take the average of the squared differences over the population. This quantity is called the "variance" and can be apportioned to different sources.

If everyone in the population were genetically identical with respect to the loci affecting height, roughly ninety percent of the variance in height would disappear.

Secular increases in height (and IQ) seem to affect the population uniformly. That is, everyone in each successive generation experiences a boost of similar magnitude. The variance in height (and the proportion attributable to genetic variation) does not change under these circumstances (because adding a constant to the mean does not affect the squared deviation from the mean).

right, 9/10th of the variation in height can be explained just if you know the variation in genes. in a malnourished environment with a lot of variance in environmental inputs this is not so.

Shorter generations mean a shorter doubling time whenever there is expansion.

Ah, good point -- for some reason, the only thought that occurred to me on-the-fly was making more babies with shorties.

So within an otherwise uniform population, 90% of the variance is genetic? And that this does not apply to comparisons between the "same" population (e.g. the Japanese) in different centuries, nor to comparisons between contemporary populations with different living conditions? Or to national populations with enormous internal economic variation, e.g. Brazil.

I suspect that the original statement (90%) made perfect sense to people in the field, but would almost always be misinterpreted by non-specialists.

So within an otherwise uniform population, 90% of the variance is genetic? And that this does not apply to comparisons between the "same" population (e.g. the Japanese) in different centuries, nor to comparisons between contemporary populations with different living conditions? Or to national populations with enormous internal economic variation, e.g. Brazil.

these really high heritability numbers often come out of studies using modern scandinavians. so well controlled for population substructure & nutrition. brazil might be not-so-good as an example because despite the inequality, i assume most poorer brazilians can eat enough (mexico is the second fattest country in the world after the USA last i checked). africa or south asia is probably a better example in the modern world, where different environments due to SES explains more of the height variance. there are issues comparing different populations in terms of heritability; obviously some of the background parameters might differ. but i think with height it probably won't be a big deal. the issues between-populations are going to be simply the outcome of differences in the shape of nutritional regimes.

I suspect that the original statement (90%) made perfect sense to people in the field, but would almost always be misinterpreted by non-specialists.

yes. the sciencedaily story was imprecisely worded...but these generally are. heritability is a tricky concept, even for those with a biological background.

It's worth pointing out explicitly that the populations used to derive the heritability estimates (as razib says, mainly well-nourished northern/western Europeans) were essentially the same populations used for the genome-wide scans described in the article.

So in populations where we know that the vast majority of height variance is genetic, we still can't find more than a trivial fraction of the genes responsible.

So in populations where we know that the vast majority of height variance is genetic, we still can't find more than a trivial fraction of the genes responsible.

...though perhaps larger QTLs may show up in admixture studies? there could be private alleles fixed in different populations which exhibit larger effect, but won't show up cuz they don't count within population? i wouldn't bet on it, but just mebee?

In practice, when you hear "90% genetic", it is usually safe to add "after all other variation factors have been silently eliminated by the design of the study" :) (very slight exaggeration)

If height is a strongly epistatic, with deceptive epistasis (i.e. having a certain set of alleles has the opposite effect of having just a sub-set of these alleles, not just larger or smaller), then current designs for GWAS are bound to fail. You can't find large QTL because there is no QTL - only QTLGs (Loci Groups).

To detect whether an allele at a certain locus is involved in deceptive epistatic interactions, you could look at the total distribution of people with this allele, not just the average value. If this distribution shows several "peaks" on both sides of the mean (where "peak" means "above the Gaussian curve"), then there is deceptive epistasis (some people with this allele are shorter than expected, while others are taller than expected).

The question is whether any signal would be drowned by noise. Also, how would you test the significance of such a result? I suppose some smarter people already thought about it.

If height is a strongly epistatic

don't you think this might have been picked up in earlier linkage studies? (focus on the "strong" part)

Heritability is always relative to a specified population. Heritability of height is probably greater in the UK today than in Victorian times, when there was a difference of about 4 inches between social classes.

perhaps larger QTLs may show up in admixture studies? there could be private alleles fixed in different populations which exhibit larger effect, but won't show up cuz they don't count within population?

Makes sense - there may well be some low-hanging fruit there for admixture mapping. Curiously, I can't seem to find anyone who's done it yet, even half-heartedly.

Which populations would be the best ones to use?

If the "height" genes are so hard to find, perhaps there are some strange multiple mechanisms causing variation. For instance, has any study looked at the genetics of appetite, and the effect of appetite to height? I think a bad appetite is a VERY common phenomenon among kids and teenagers and perhaps it would explain some variation in height, not growth genes per se.

I am still amazed how important obsession height appears to be in the US... Even in youtube people comment on the celeb's height (OMG she is only 5'2''!!!!).

Has anyone made controlled studies about height and several traits: energy consumption, power, running speed, agility, battle ability... (all sorts of traits that might influence fitness)? Now everyone seems to believe that height is useful in most purposes (now I am not talking about possible sexual selection), but is there some proof about this? I heard that the Unokai (warriors who have killed an opponent) among Yanomamo are on average, average height. I also heard about a quite old ergonomics study about the most succesfull loggers/lumberjacks (or how do you call the guys who are running the trees down the rivers). Their working speed was correlated, not with body height or mass but with the circumference of their elbows and knees. This correlation was found probably because these circumferences were giving a rough image of total robustness of the body. (I bet there would be a lot of old ergonomics body measurement data and correlations between different tasks.)

I discussed about this with a friend and he agreed that if we control for the size (that is, fat free body mass), taller people might even have a disadvantage in a battle. Perhaps height is an energy efficient way to exaggerate size - after all 6'7'' are quite rarely big and robust, they often tend to be skinny guys with longer legs. If you grow longer legs, you look bigger, but you won't consume that much more energy. If you would build muscle mass to look as big, this would be energetically very costly.

Which populations would be the best ones to use?

i'm thinking black americans. perhaps hapas in hawaii?

Can someone who is familiar with how to use the HapMap database plug in these height loci to see how they vary among ethnic groups?

I've lived in a number of countries and find it striking how different different populations are in average height and physique and have always wondered what the relative influence of genetics and environmental/dietary factors was. I recently read a paper by Koepke and Baten looking at height in Ancient and Medieval Europe and the Eastern Mediterranean in different regions. They find that beef and milk consumption and arable land per person were the main factors affecting height and that differences between Western, Southern and Northern/Eastern Europeans disappeared when they controlled for these factors. Link: http://www.ata.boun.edu.tr/ehes/ ...en_Istanbul.pdf

On the other hand, there seems to be a clear difference today in Europe between people West of the Rhine and South of the Alps on one hand and people North of the Alps and East of the Rhine on the other. I could buy that perhaps greater economic prosperity causes Dutch (184cm GDP PPA: $38,600), Norwegians (180cm; $55,600), Swedes (180cm $36,900), Danes (181cm $37,400), and Germans (180cm $34,400) to be taller than Brits (177cm $35,300), Irish (177cm $45,600), French (177cm $33,800), Spanish (176cm $33,700), Portugese (173cm $21,800), and Italians (177cm North, 174cm South $31,000), but it hardly seems likely that people in the Czech Republic (180cm $24,400), Serbia (180cm $7,700), Estonia (179cm $21,800) or the Dinaric Alps (186cm, Croatia - $15,500 Bosnia-Herzegovina- $6,600) would be better nourished and better off than Western/Southern Europeans.

Perhaps dietary factors like dairy consumption are responsible, but, as Koepke and Baten note (but are eager to dismiss), perhaps the genetic mechanism that causes the height differentail has to do with the ability to use various foodstuffs such as with lactose tolerance (e.g. - Perhaps on poor, high carb, low protein and dairy diets, all European populations are similarly short, but some groups will be taller than others given sufficently high levels of dairy and meat consumption?).

why don't you try to look for SNPs
http://www.ncbi.nlm.nih.gov/sites/entrez
(remember to select "SNP" under the search drop down on the top left)

i think some of the papers provide them. look for stuff like "rs123545" and shit.

PhillyGuy: I think your economic data is somewhat strange. Is this purchasing power GDP or what? Also your height measurements. Your mean heights are too high in some cases (you always took the highest estimate).

Anyway, you have a cluster of Nordic contreys which have a relatively similar culture and economic system. They are pretty similar also genetically. You your datapoints are not really independent.

I think you miss the point that economic prosperity in these high levels does not measure the consumed food that well. For instance, even in the USSR producing enough meat for the labourers was a high priority. Even if they were lacking prosperity, people had enough food generally.

No one bothered to comment on measuring the real effects of height in different tasks... It is really strange no one has actually measured these things, or I have not heard about it.

Thanks Razib.

@iematta: Yes, PPA is a typo. The estimates are 2007 per capital GDP at purchasing power parity. (I know that if would be better to use time lag data and look at per capita GDP PPP over the period when the subjects in the study were growing up, but hey, I have a day job. Also, it would be better to have data on diet in the countries in question, but I don't know of any easily findable source.

I don't think I was injudicious in picking the height estimates. I tried to pick recent samples of young adult males that were old enough to be finished growing (i.e., at least 19 or 20) and that were actually measured (rounded to the nearest cm), as opposed to self reported data, since people tend to overreport their height (e.g. an Australian study found that young males 18 to 25 said they were 179.9cm but were actually only 178.4cm when measured). For Norway, Sweden, Denmark, and Germany I used national conscript data from the respective national statistical agencies. For Portugal, I used conscript data presented by Padez in J. Hum. Ecol., 22(1): 15-22 (2007). (173cm is a typo, actual height is 172cm.) For England I used the height of 20 year olds in the Survey of Health for England and and for the Czechs I used the 2004 Czech Health Survey. For Italy I used the data for 20 year olds in a growth study by Cacciari et al. in the European Journal of Clinical Nutrition, February 2002, Volume 56, Number 2, Pages 171-180. For the Dutch, I used data for 20 year olds (all ethnically Dutch) in a Fredriks et al. in Pediatric Research, Vol. 47, No. 3, 2000. (The same team also looked at Morrocan and Turkish childrens who grew up in the Netherlands and found they were 175cm and 174cm, respectivley). The data for the Dinaric Alps comes from C R Biol. 2005 Sep;328(9):841-6 by Pineau, JC, et al. and the data for Estonians comes from Economics & Human Biology, Volume 4, Issue 1, January 2006, Pages 89-103 by Lintsi and Kaarma (and since the subjects in both studies were only 17 when measured, the results probably understate their final heights). The data for the Serbs is for 18 year olds from Novi Sad in a study by Bozić-Krstić VS, Pavlica TM, Rakić RS., 2004. As for the French, Spanish and Irish, I had to rely on ECHP data for men over 20 born between 1975 and 1980. Since ECHP data is self reported, the real heights are probably shorter.

I agree with you that the Nordic countries are
relatively similar in culture, economic system, and genetics, but there are notable differences when you compare them to peoples from Central Europe (Germans and Czechs), the Balkans, the Eastern Baltic or Western and Southern Europe. I also agree that economic prosperity at European levels does not necessarily measure the consumed food that well (but I would think that at least in the Balkans, where living conditions are much poorer than in Western Europe, I would be surprised if they have a nutritional and health care advantage over, say, the UK or France) which is exactly what I was pondering and do not know the answer to: Are East and North Europans taller than Western and Southern European because of what they eat, or is it in their genes?

I've read a fair bit from economic historians using anthropometric measures to assess living conditions (e.g., Koepke, Baten, Steckel, Komlos, Floud, etc.) and find their work interesting, but I am somewhat skeptical of them because their motiviations seem to be at least partly political. You may have seen articles every once in a while talking about how Americans are shrinking and falling behind Europeans because of income inequality, junk food, and lack of socialized medicine citing Steckel and, particularly, Komlos. It is clear that Komlos desparately wants nationalized healthcare and more wealth resdistribution for the US. However, if you look at the NHANES data from the NCHS, Americans are not shrinking (at least controlling for demographic changes) and whites are actually getting slightly taller. While it is true that US whites are slightly shorter than some Northern and Eastern European whites, they are taller than Southern and Western European ones. Komlos conveniently ignores wealthy countries with socialized medicine like the UK, Ireland, France, and Italy because they don't fit his narrative and instead focuses on Scandinavia, Holland and Germany. Whenever I read papers on economics and anthropometric measures, I am struck by how often one sees the a priori presumption that there can be no significant, functional genetic differences between ethnic groups. It reminds me very much of the whole ethnicity/IQ controversy (and we now know that, at least in the case of DTNBP1, different SNPs of genes that influence cognitive functioning, according to the HapMap database, are not distributed exactly the same in all populations. I see no reason why height would not be similar unless there is some selective factor constraining it. After all, Europeans differ in their average facial features and pigmentation, whey shouldn't they also vary in stature and physique. I think it is eqally plausible that perhaps white Americans are shorter than people from the taller European countries for genetic reasons since many of their ancestors also come from shorter European countries.

Well, I was looking at wikipedia and there were also some lower estimates for height.


"Americans are shrinking and falling behind Europeans"

And why is this important? Are we going to compare penis sizes next? (It is interesting that apparently french penises are the largerst in Europe, did they undergo a stronger sexual selection at some point?) I understand the concern of dropping height in Argentina where is not very high in the beginning (so the food conditions must be quite poor already). But does it make ANY difference if national average height is 180 cm or 175 cm? In a trench war 180 cm is worse at least... On the other hand a population of average height of 175 cm will produce worse basketball players... :)

I think it is rediculous to make this a social debate; I think the height of Bosnians is not about the diet but about their genes. Eating habits are usually cultural anyway so you cannot force people to eat certain foods.

"And why is this important? ...I think it is rediculous to make this a social debate..."

Agreed. Unfortunately, my impression of a lot of the economics/social science literature is that the authors want to make it one, and in some cases they present data in a misleading way to serve their ends and make mountains out of millimeters. (I've also noticed that English language journalists are pretty good at selectively using 2.54 and 2.5 in converting centimeters to inches in press coverage of these studies so as to exaggerate small height differentials.) If we actually could work out the genetics of height, we would be on much firmer ground in using it as an indicator of human welfare between advanced economies. I actually think such research is quite silly and peripheral to the question of the merits the US adopting or not adopting more redistributive income policies and socialized medicine, but I am amazed at how many times I've seen articles about this topic in magazines like the New Yorker and in many newspapers in recent years. (In fact, this media coverage is how I heard about this field of research in the first place.)

Well, in Europe preventing obesity is seen as relevant, not height. No one is talking about national height, I haven't heard a single time about it. I wonder if there is a cultural difference how people think of height and obesity and what is seen as bad. I think height does not influence for instance political carisma in Europe much at all (many/most South European top politicians are short), whereas even a relatively slight obesity most certainly will. I say this because I heard that usually the taller of US presidential candidates is selected. Finland had an overweight president a few years back (Ahtisaari) and he was made fun of all the time he was in power because of this.

here is the former "fatso-president"
http://en.wikipedia.org/wiki/ Ima...i_Ahtisaari.jpg

"Well, in Europe preventing obesity is seen as relevant, not height...I wonder if there is a cultural difference how people think of height and obesity and what is seen as bad."

I think you're on to something about cultural differences re appearance, but I'm not too sure about obesity. Obesity is also a big no-no for politicians here. There are a few exceptions, but they are often ridiculed for it. I'm an American, but I lived 4 years in the UK and 3 in Germany and I was struck at how much less importance people seemed to place on their politicians' appearance compared to the US. (I can't speak for Romance language countries, as I've never lived in them.) Take Kohl, Schroeder, and Merkel. The first is a huge, fat, bald man with glasses, the second is a dumpy, if well groomed, little man, and the third is a frumpy looking woman who brings images of a cow to mind. None of them would be electable in the US due to their neglect of or shortcommings in their appearance (Heck, even HR Clinton is trying to do the best with what she's got). In the US politicians are expected to be tall (if not, they often wear lifts or avoid standing next to taller opponents and demand platforms to stand on behind their podiums in debates cf. Dukakis), handsome or pretty, when they get old they have cosmetic surgery and botox, they often color their hair (and wear toupes if bald), and usually wear contacts instead of glasses. It probably has to do with an obsession with youth culture and the "hollywoodiztation" of politics that have been reduced mostly to soundbites. Very rarely do you ever see unscripted interaction between candidates and the public or each other. Even the the debates are mostly prescripted speeches given in respones to the moderators and not arguements between the candidates themselves. In Germany I was very impressed to see high level politicians appear on political talk/debate shows and really lay into each other and speak intelligently, responsively and extemporaneously. In the US we seem to just want someone who is out of central casting to play a politician on TV. Sad isn't it?

Another cultural factor may be the nature of the sports people like. Our big three sports are American football, basketball and baseball. In most positions in these sports (apart from the field positions in baseball and a few of the backfield positions in football), height and/or muscle bulk provide a huge advantage. By contrast, in soccer, which is dominant in Europe, height and bulk are far less important (though not always irrelevant). Hence, most of our sporting icons, and by extension, the popular image of masculinity, is of a huge, imposing physique. Just contrast David Beckham with Lebron James, Michael Jordan, Barry Bonds, or Tom Brady and the Manning brothers. These American sporting stars are all very large men. (We do however have a few shorties like Jimmy Rollins (Baseball NL MVP) and all-time great football RB Barry Sanders (both around 1.7m) but even these two, while short, are very compact, powerful men (Sanders was about 92-3kg and could squat and bench something like 300-350kg and 230-250kg, respectively).