Thursday, April 22, 2010

The puzzle of European hair and eye color

I’ve been fascinated by a puzzle of modern human evolution: the diverse palette of hair and eye colors that has developed in some populations (Frost, 2006; Frost 2008). Hair may be black, brown, flaxen, golden, or red, and eyes may be brown, blue, gray, hazel, or green. Both polymorphisms are largely confined to Europeans, especially those from the north and east.

This is an evolutionary puzzle for several reasons:

1. Hair color and eye color diversified through two separate processes that involved several gene loci (principally at MC1R for hair color and at OCA2-HERC2 for eye color).

2. Both processes occurred within the same geographic area.

3. Both processes occurred within a relatively narrow time frame, i.e., after the arrival of modern humans in Europe c. 35,000 years ago. Current estimates place this evolutionary change quite late in time, perhaps during the last ice age (25,000 - 10,000 BP).

For some anthropologists, this palette of hair and eye colors is a side effect of the lighter skin of Europeans. This lighter skin is, in turn, due to relaxation of selection for dark skin at non-tropical latitudes and a resulting accumulation of ‘loss of function’ alleles that affect not only skin color but also hair and eye color.

Yet relaxation of selection could not have produced so many new alleles over so little time. If selection is relaxed at loci for hair and eye color, close to a million years must elapse to produce the hair- and eye-color variability that Europeans now display, including ~ 80,000 years for the current prevalence of red hair alone (Harding et al., 2000; Templeton, 2002). This is much longer than the c. 35,000 years that modern humans have been in Europe. Moreover, the presumed initial cause—the whitening of European skin—seems to have occurred long after the arrival date of 35,000 BP (Norton & Hammer, 2007). As a Science journalist commented: “the implication is that our European ancestors were brown-skinned for tens of thousands of years” (Gibbons, 2007).

The puzzle is not resolved if Europeans turned white because of positive selection for lighter skin, as opposed to relaxation of selection for darker skin. Such a scenario would not have caused hair and eye color to diversify. In fact, most of the new alleles have little or no relationship with skin color. Only red hair and blue eyes are visibly associated with lighter skin.

There must have been positive selection for diversity of hair and eye color in and of itself. And this selection must have been very strong, given the relatively narrow time frame.

I have suggested that the likeliest explanation is sexual selection (Frost, 2006; Frost, 2008). This kind of explanation is consistent with several general facts:

1. Sexual selection typically creates brightly colored traits.

2. Such traits tend to be on or close to the face, because this part of the body attracts the most visual attention.

3. Intense sexual selection can produce color polymorphisms.

But why would sexual selection have been stronger among northern and eastern Europeans than among other human populations? To answer this question, we must understand why sexual selection should have differed in intensity among ancestral modern humans. In general, the differences were latitudinal, i.e., sexual selection differed primarily along a north-south axis.

Latitudinal differences in the ratio of men to women on the mate market

In the tropical zone, a woman could gather or grow enough food for herself and her children with little assistance. Because the cost of providing for a second wife was very low (often negative, i.e., a net gain), a man’s optimal reproductive strategy was to have as many wives as possible. There were thus too many men competing for too few women.

The farther away ancestral humans were from the tropics, the more women needed food (meat) provided by men. This was especially so in winter, when opportunities for food gathering were scarce. The cost of providing for a second wife was thus high, making polygyny impossible for all but the ablest hunters.

Alongside this trend of increasing female dependence on male providers was another north-south trend: male mortality increased farther away from the tropics because of longer hunting distances and the resulting increased risk of death due to accidents, exposure, starvation, etc.

Continental Arctic: optimal conditions for sexual selection of women

These two trends culminated in the continental Arctic. Here, women had few opportunities for food gathering at any time of year. They and their children depended almost wholly on meat that men provided through hunting. Here too, hunting distance was at a maximum. Men hunted wandering herds of herbivores, mainly reindeer, over very long distances. The high rate of male mortality, combined with the low rate of polygyny, limited the number of males available for mating. Result: a corresponding surplus of unmated females and intense sexual selection of women.

Today, this kind of environment is confined to the northern fringes of Eurasia and North America, but during the last ice age (25,000 – 10,000 BP) it lay further south and covered more territory. This was especially so in Europe, where the Scandinavian icecap had pushed the steppe-tundra zone down to the plains stretching from southwestern France through northern Germany and into eastern Europe. These temperate latitudes permitted a high level of bioproductivity and a comparatively large human population—the ancestors of today’s Europeans.

Sexual selection and color traits

When sexual selection is weak, the adaptive equilibrium is dominated by selection for a dull, cryptic appearance that reduces detection by predators. As sexual selection grows stronger, the equilibrium shifts toward a more noticeable appearance that retains the attention of potential mates, typically by means of vivid and/or novel colors.

One outcome may be a polymorphism of brightly colored phenotypes, due to the pressure of selection shifting to scarcer and more novel hues whenever a color variant becomes too common. This frequency dependence has been shown in humans. Thelen (1983) presented male participants with slides showing attractive brunettes and blondes and asked them to choose, for each series, the woman they would most like to marry. One series had equal numbers of brunettes and blondes, a second 1 brunette for every 5 blondes, and a third 1 brunette for every 11 blondes. Result: the rarer the brunettes were in a series, the likelier any one brunette would be chosen.

Among ancestral Europeans, this selection pressure may have caused a proliferation of new hair and eye colors to the detriment of our species norm of black hair and brown eyes. The selection was partly for novel colors. A rare color engages visual attention for a longer time than does a more common color (Brockmole & Boot, 2009). It may be that color rarity stimulates a mental algorithm that scans the visual environment for new or unusual objects.

In addition to color novelty, there also seems to have been selection for color brightness. Hair is carrot-red but not burgundy red. Eyes are light blue but not navy blue. Maan and Cummings (2009) argue that brighter colors have a stronger impact because they deliver a stronger signal that is more readily learned and retained in memory.

In a mate market already saturated with high-quality females, these eye-catching characteristics—color novelty and color brightness—may have made the difference between success and failure in finding a mate.

Other evidence for unusually strong sexual selection of European women

Hair and eye color polymorphism coincide geographically with other unusual physical traits. There is, for instance, the extreme whitening of the skin, which we do not see in other human populations at similar latitudes and which may have been driven by male targeting of lighter skin as a female-specific characteristic.

There also seems to have been selection to accentuate female-specific traits. Women of European descent have wider hips, narrower waists, and thicker deposition of subcutaneous fat than do women of other geographic origins (Hrdlička, 1898; Meredith & Spurgeon, 1980; Nelson & Nelson, 1986). Even before birth, Euro-American fetuses show significantly more sexual dimorphism than do African-American fetuses (Choi & Trotter, 1970).

In the same vein, Liberton (2009) has found that face shape differentiated between Europeans and sub-Saharan Africans in part through a selective force that has acted primarily on women, and not on both sexes. This too would be consistent with the selection pressure that seems to have diversified European hair and eye color.


Brockmole, J.R. & W.R. Boot. (2009). Should I stay or should I go? Attentional disengagement from visually unique and unexpected items at fixation, Journal of Experimental Psychology, 35, 808-815.

Choi, S.C., & Trotter, M. A. (1970). Statistical study of the multivariate structure and race‑sex differences of American White and Negro fetal skeletons. American Journal of Physical Anthropology, 33, 307‑312.

Frost, P. (2008). Sexual selection and human geographic variation, Proceedings of the 2nd Annual Meeting of the NorthEastern Evolutionary Psychology Society, The Journal of Social, Evolutionary & Cultural Psychology, 2 (supp.), 49-65,

Frost, P. (2006). European hair and eye color - A case of frequency-dependent sexual selection? Evolution and Human Behavior, 27, 85-103

Gibbons, A. (2007). American Association Of Physical Anthropologists Meeting: European Skin Turned Pale Only Recently, Gene Suggests. Science 20 April 2007: 316. no. 5823, p. 364 DOI: 10.1126/science.316.5823.364a

Harding, R.M., Healy, E., Ray, A.J., Ellis, N.S., Flanagan, N., Todd, C., Dixon, C., Sajantila, A., Jackson, I.J., Birch‑Machin, M.A., & Rees, J.L. (2000). Evidence for variable selective pressures at MC1R. American Journal of Human Genetics, 66, 1351‑1361.

Hrdlička, A. (1898). Physical differences between White and Colored children. American Anthropologist, 11, 347‑350.

Liberton, D.K., K.A. Matthes, R. Pereira, T. Frudakis, D.A. Puts, & M.D. Shriver. (2009).
Patterns of correlation between genetic ancestry and facial features suggest selection on females is driving differentiation. Poster #326, The American Society of Human Genetics, 59th annual meeting, October 20-24, 2009. Honolulu, Hawaii.

Maan, M.E. & M.E. Cummings. (2009). Sexual dimorphism and directional sexual selection on aposematic signals in a poison frog, Proceedings of the National Academy of Sciences (USA), 106, 19072-10977.

Meredith, H.V., & Spurgeon, J.H. (1980). Somatic comparisons at age 9 years for South Carolina White Girls and girls of other ethnic groups. Human Biology, 52, 401‑411.

Nelson, J.K., & Nelson, K.R. (1986). Skinfold profiles of Black and White boys and girls ages 11‑13. Human Biology, 58, 379‑390.

Norton, H.L. & Hammer, M.F. (2007). Sequence variation in the pigmentation candidate gene SLC24A5 and evidence for independent evolution of light skin in European and East Asian populations. Program of the 77th Annual Meeting of the American Association of Physical Anthropologists, p. 179.

Templeton, A.R. (2002). Out of Africa again and again. Nature, 416, 45-51.

Thelen, T.H. (1983). Minority type human mate preference. Social Biology, 30, 162-180.


Eugene said...

Is there an association between skin color and eye color? Is there a specific point somewhere on the skin color scale beyond which eye color is always brown? Thanks

UncleTomRuckusInGoodWhiteWorld said...

I have a problem with some of this...

"Alongside this trend of increasing female dependence on male providers was another north-south trend: male mortality increased farther away from the tropics because of longer hunting distances and the resulting increased risk of death due to accidents, exposure, starvation, etc."

Historically Africa has been underpopulated until the 17th-18th century or so, it is thought that high population density was impossible due to tropical disease, especially Malaria. European settlers found this out as they tried to construct European style settlements early on (well before any medicines existed) and they died off quickly. This is one reason Africans tended to live in sparse settlements.

This is one reason why colonialism did not really make a huge foothold in Africa until the 19th century, although Europeans had direct contact with SubSahara Africa since the 15th century.

This suggest to me that the death rate was high, far higher than in Eurasia, which the population in Europe, India, and China boomed.

There also seems to have been selection to accentuate female-specific traits. Women of European descent have wider hips, narrower waists, and thicker deposition of subcutaneous fat than do women of other geographic origins "

I have never heard or read anywhere that European women have more subcutaneous fat than East Asians. The fat makes East Asians have a slight yellowish tinge, whereas Europeans are very often ruddy (because you can see their capillaries through their skin).

Tod said...

Eye color Predicts Disagreeableness in North Europeans: Support in Favor of Frost (2006)

Tod said...

Genetics of human iris colour and patterns

Anonymous said...

Dragon Horse, I think Peter's argument is that in Europe male mortality was higher than female mortality. Disease mortality may have been higher in Africa, but it affected both sexes equally.

Peter Frost said...


There is a weak association between skin color and eye color, especially when eye color is blue. But there is no physiological reason that prevents dark skin from co-existing with light eye color. I've met Jamaicans of mixed origin who have dark skin and blue eyes.

Dragon Horse,

We're not talking about the same period of human history (or rather prehistory). I'm talking about the hunter-gatherer stage and the early stages of agriculture. During that period, human population density was higher in the tropical zone than in the temperate and arctic zones. It was the development of agriculture, in particular the ability to store crop surpluses for the winter and early spring, that enabled humans to achieve high population densities in the temperate zone.

In any case, a society can have a high rate of male mortality and still maintain a large population (and conversely a small population may be characterized by a low rate of male mortality). Total population size is dictated by other factors as well.


Yes, I read the first paper, but I'm not really convinced by the authors' theory. I don't see how sexual selection of women would favor an overall increase in disagreeableness. It might favor a lower threshold for disagreeableness among women in a context of female-female rivalry for mates. But this kind of disagreeabless would have to be very context-specific.

Anonymous said...

Moreover, as Peter says, in Europe polygyny was economically less feasible than in Africa, so even if the sex ratios had been similar, selection pressures would have been different in Africa and Europe.

UncleTomRuckusInGoodWhiteWorld said...


Okay I got you, that's my fault I overlooked the specific period.


I don't know about that, I think the polygamy thing changed after the advent of agriculture. I mean...what would a "spinster" or a widow do? Just die? Have their father and brother live with them forever?

I don't know though, Polygamy was not at all uncommon in China and the Middle East. But we have little knowledge of these areas in pre-historic times. I know they are at lower latitudes, but the Middle East is not exactly a "bread basket", I realize at different times it much more green than now. I know far more about China. The ancestors of the Han, the Huaxia tribes are thought to have come from the Northwest into the Yellow River area of Central China proper. It does get cold there in the winter, there are seasons, even if there is little snow and the winter is shorter. It is not as if you can go pick berries all year or something. Although the selection pressures may not have been as high as Europe, women were dependent on men to hunt, especially in the winters, this would have been more true in Mongolia, Korea, Manchuria (modern day Dongbei region)...although there is light skin, there is not much variation in eye color of skin color. you can see "brown" and what they call "black eyes" (dark brown), you can also see people with brown hair, especially in childhood, but often turns black as they get older. Still, I guess, North Korea is the same latitude as Bulgaria, although one can argue Bulgaria has a more temperate climate due to modern air currents. Mongolia is about the same Northern Italy/Switzerland. Still from what I read light hair and eyes is thought to have originated along around the Baltic? Which is way North, into the Russian Fareast.

Tod said...

IMO the strongest evidence against the Baltic agriculture hypothesis being the cause of hair colours is that red hair is vastly more common in Scotand and Ireland than in the Baltic. It really won't do to say:-

"Blonde hair and blue eyes were other infantile traits that were just swept along accidentally."

As red hair goes with much lighter skin in a way that blond hair does not redheaded people should be more common in the Baltic than West Europe. Furthermore red hair ought to to be vastly more common than blond everywhere if light skin colour was being selected for. If any recessive alleles were be swept along it would those linked to skin lightening, ie red hair ones. Blond hair wouldn't be swept along because it does not affect skin lightning, furthermore it's recessive and couldn't have increased by drift as quickly as it did.

The Sami are white and have a diversity of eye and hair colours despite living on meat (Meat consumption reduces the risk of nutritional rickets and osteomalacia.) and vitamin D rich fish.

UncleTomRuckusInGoodWhiteWorld said...


This might be far more simple. The lack of selective pressure from intense sunlight in the tropics allowed for various types of mutations to spring up for skin and hair. Ireland and Scotland are basically on or part of islands. Those people could have developed the red hair mutation and by chance is became dominant as people there selected for it, maybe blondism came later. It so happened that red hair and light skin are more closely linked. Also, I could be off, but I believe the British Isles, Ireland, etc are far more overcast than the Baltic. I read a theory once that red hair is more reflective in cloudy weather than blond.

On the European continent, it could be that blond hair was dominant (no bottleneck going to an island) and was selected for primarily due to nothing else but aesthetics. It does not have to be a zero sum game. It could almost be an issue of convergent evolution, in this case "selection for light hair in general, what ever color that is".

Light skin

Aaron Davies said...

I don't suppose there's any way (Greek/Roman commentary? Bog mummies?) to tell if Celts were as prone to red hair three thousand years ago, when they covered Europe almost completely, as they are today…

Eugene said...

Thanks for the response regarding skin/eye color association and Tod's article on genetics.

Anthropologist Satoshi Kanazawa (author of "Why Beautiful People Have More Daughters") has tried to explain why blond hair and light eyes are "attractive."

He suggests that blond hair is a marker of youth for women. He observes that (1) hair tends to darken with age -- most blondes' hair turns light-brown in middle age, and (2) blond hair is desirable in women, but not in men, reflecting the different premiums each gender places on youth.

Thus, a woman who is currently blonde can "honestly advertise" herself as young and fertile, something that dark-haired women can't do as effectively. He suggests the "dumb blonde" stereotype could have evolved as a result of blondes' youth and immaturity.

With respect to blue eyes, he suggests that a woman with light eyes makes it easier for the male to judge the size of her pupil, which dilates when the person is interested. In contrast, a woman with brown eyes makes it more difficult for the man to gauge her interest. Brown-eyed people, particularly women, therefore appear to be somewhat "mysterious."

Ben10 said...

But you can only select for mutations that already happened. Being attracted by blue eyes won't 'create' the mutation for blue eyes.
I have two questions.
1)Did these news alleles appeared in few individuals localized around the Baltic or did they appeared in more than one place and time in Europe (to accelerate the spreading).
2) Many (all?) of these alleles are recessive. They must have been even more recessive and unstable at the time they first appeared. Difficult to imagine how they can spread if they are selected at the homozygous state. Is it possible that the heterozygotic state gives the carrier some advantage of some sort ?

Jack said...

I do not have any scientific background other than reading websites such as this one, so I may be completely off base with this, but it is an idea that has occurred to me...

In lightly pigmented people with blonde hair, the fundus inside the eye also has a blonde pigmentation. I first became aware of this about 20 years ago when an optometrist told me that I have a blonde retina (my hair was blond at the time, though it is now more of a light brown). This is a trait that makes people more sensitive to bright light, but which I think would allow people to see better in low light conditions.

In addition to any selective pressure that might exist for blonde hair, a blonde fundus might have been a definite advantage in Northern Europe, where the high latitude and thick cloud cover result in a darker environment during much of the year. Even a slight advantage in seeing in low light conditions might have been an advantage to a hunter whose survival depended on spotting game at a distance on a cloudy winter day.

If this is correct, then the selective pressure that existed might have favored those mutations that caused a light fundus in the eye. Light hair, though not providing much of an advantage in and of itself, might have evolved as a side effect.

A. Rivedercci said...

P. Frost : "Current estimates place this evolutionary change quite late in time, perhaps during the last ice age (25,000 - 10,000 BP)."

I think it occured at least several milleniums before 10,000 BP.
The region of north Africa where the blond/blue-eyes/light skin occurs the most frequently is apparently in the Moroccan rif (mountains). Interestingly enough, bones of about 12,000 BP cro-magnoids * from this region have been genetically tested and most of their mtDNA haplogroups are majoritarily similar/close to mtDNA European haplogroups (H, U, V (but also a few JT, U6 and one that was probably M1 or L3)). It can be rattached to the ibero-maurusian culture that seems to have come from present Spain. "West Eurasian" mtDNA haplogroups are found relatively frequently in north Africa ** but IIRC they are particularly frequent in ths region, so we can assume these traits are a remaining tracks of these paleolithical populations that came from Europe.


** for instance : (Saami and Berbers—An Unexpected Mitochondrial DNA Link)

It think it could prove indirectly that these phenotype characteristics were already quite spread before 12,000 yrs BP.

Peter Frost said...


There is a project in Germany to determine the hair color of neolithic Europeans, but I don't know what they've found so far.


There are two problems with Kanazawa's reasoning:

1. It doesn't explain the overall increase in the number of alleles for hair or eye color.

2. It doesn't explain why there wasn't fixation of blond hair and light eyes in at least some human populations. Even in Sweden, most people are not blond or blue-eyed.


Mutant alleles are relatively rare occurrences, and it is even more unusual for the same mutation to occur more than once (although this does happen). How often are redheads born in populations that do not normally have redheads?


Your hypothesis would make sense if light hair and light eye color were due to a single genetic variant at a single gene loci. But this isn't the case.


I believe that these changes in hair and eye color occurred after the separation between proto-Europeans and proto-East-Asians (Mongoloids). A recent paper has dated this split to 20,000 BP. This date would also coincide with the glacial maximum, when the ice sheet and glacial lakes along the present-day Ob formed a barrier to east-west gene flow.

A. Rivedercci said...

P. Frost : 2. It doesn't explain why there wasn't fixation of blond hair and light eyes in at least some human populations. Even in Sweden, most people are not blond or blue-eyed.

This is because of the "later" migrations.
I think these blond, blue eyes, milky skin phenotypes were overwhimingly present among hunter-gatherers (some of the Y-DNA haplogroup I and mtDNA haplogroup U5/U4 ?) of the north half of Europe (maybe they were all like that). It's the population movements from souther that started around 7,000-6,000 YBP, if I remember correctly, that altered this "nordic" phenotype, I think.

Ben10 said...

"Mutant alleles are relatively rare occurrences, and it is even more unusual for the same mutation to occur more than once (although this does happen"

It depends. I am thinking about one of the most prevalent genetic disease, Autosomal Dominant Polycystic kidney Disease, affecting one in a 500 individuals (usually late in life). Despite its name, the disease is purelly RECESSIVE at the cellular level, i.e, you need BOTH alleles to be mutated to have the phenotype. Strangely, affected individual are born heterozygous for it and they should not have the disease. But it is suspected that some surrounding DNA sequences (introns, gc rich or duplication) cause a genetic instability that translate into somatic mutations. When these somatic mutations, called 'second hit' occur in the kidney, they cause the apparition of the phenotype at a frequency sufficiently high to make the phenotype look 'dominant', that is, heterozygous individuals who devellop a recessive phenotype!. It is possible that this genetic instability is exacerbed because the responsible gene has been duplicated recently in evolution (only in higher primates, including Man). We can imagine that in thousand years, when the instability settles, the disease would return to its 'recessive' state and becomes much less common.
In our case here, to explain the apparition of the blonde phenotype in the first white europeans, we could imagine a highly prevalent 'second hit', which, when it occured in neural crest cells, the precursors of the melanocytes, would cause the blond phenotype even in heterozygous individuals. That way, we don't have to postulate a high level of incestuous endogamy to explain the positive selection of a 'recessive' phenotype. Genetic instability is probably associated with every individuals of a new species or subspecies that has recently branched off the evolutionary tree, that would be the case of these early white european people. The genetic for these alleles then quickly stabilized and return the phenotype to its modern recessive occurence, but not before the allele became more prevalent, since in this particular case it was positively selected. This is pure speculation of course and I don't say that it happened this way, but it is to indicate that genetic can be complex, especially in newly formed species and maybe that can be used to explain an initial boost of frequency of an allele in a recently formed population.

Peter Frost said...


Your explanation would make sense if there were only two alleles for hair color or eye color. In that case, there would have been intermixture between a southern population with an allele for dark hair and a northern population with an allele for light hair.

In reality, there exist at least 7 alleles for hair color that are specific to European populations. An intermixture model cannot explain how ancestral Europeans went from having only one allele to seven alleles over a relatively short time period.


Good point! The dominant/recessive model of inheritance is a simplification of reality. Many 'recessive' alleles have some phenotypic activity even in the heterozygous state, particularly if some other factor is present.

Andrew Oh-Willeke said...

Suppose that there was positive selection for light skin (e.g. due to diseases like M.S. which strike young and are sun exposure related), and suppose that paleolithic people believed on the basis of this that light skin was a signal for fitness in their environment.

They don't know (and we barely have a sense now) how eye and hair color are related to skin color genetically. The crude signal that they may have had is "light color = greater fitness" and this believe, even if wrong, could have driven sexual selection (on both sides of the gender divide).

I don't see why the story has to be any more elaborate than that.

Rez(A) T. said...


i'm really having a hard time finding this hair color I saw some place.

ummm its closest to the dirty blond I guess.

the color is like the Nescafé instant coffee mix (3in1) suff.

very light, brown, but not actually brown, you know ...

could you tell me what it is?

Anonymous said...

I don't understand why so many of these traits would be linked to females. Why wouldn't females also want males with light skin, wider hips, blonde hair, etc?

Luke said...

Dr. Frost,

Which European population(s) have the very darkest (black) hair at the highest rates?

Thank you.