There’s a persistent myth that you may have seen before in the guise of news. It claims to cite a study from the World Health Organization (WHO) showing that naturally blond hair will be a thing of the past within a few hundred years.
When we dig into the history of this story—or, more accurately, when we read the results of Snopes digging into it—we find that the same story with slight variations has been making the rounds for more than a hundred and fifty years! People have been claiming the imminent extinction of blonds since at least the American Civil War.
Why does this seem believable enough to keep repeating it over and over? Part of the reason is probably that many of us remember only a few things from our assorted biology classes, and that likely includes the concept of dominant and recessive traits.
A simple but distorted way of thinking about it is that hair color is controlled by two copies of a gene, one from each of our parents. Many genes have alleles, or multiple versions resulting in different hair colors, and these alleles can be dominant or recessive. A dominant allele is expressed if it shows up in one or both copies, but you have to have a recessive allele on both sides for it to be expressed. So if a hypothetical mom has blond hair but the hypothetical dad has brown hair, there’s another genetic dead end for blonds, right?
Fortunately for genetic diversity, this isn’t really the way it works. We’ll keep it simple and consider just brown and blond hair (an obvious simplification, since redheads aren’t dying out either). We’ll also ignore the facts that many traits aren’t determined by a single gene and that dominance isn’t always complete enough to exclude the recessive trait.
Take my example, with a blonde mother and a dark-haired father. Those are their phenotypes, or the way their genes are expressed in their physical development. Blondness is recessive, so we know the mother is homozygous, with two alleles of the same type. Let’s also say the father got dark-haired alleles from both his parents, so he’s also homozygous. Here are our genotypes:
Based on their genotypes, we can construct a Punnett square with the father’s genotype on the horizontal axis and the mother’s on the vertical axis. The resulting grid shows the possible combinations of the single alleles they’ll give to their children.
This one’s kind of boring, since each of them can only give one type of allele. All their kids get one dark and one blond allele, so they have the dominant phenotype of dark hair. Note, however, that their blond alleles haven’t gone anywhere! They’re hanging out, just waiting for their chance to hop into the gene pool.
At this point, we can talk about what’s called Hardy-Weinberg equilibrium, with a few tweaks to our situation. One is that genetic statistics can’t be done at the level of one couple, or even a couple couples. To have numbers large enough to work for population statistics, our hypothetical parents would have to shatter the world record for the number of offspring born to one couple (69 kids, courtesy of the Vassilyevs of 18th-century Russia, if you’re curious). We’ll just take their heterozygous kids and make them into a population of, say, several thousand.
The other change is that half dominant and half recessive alleles is a boring setup, so we’re going to make it a 60-40 split.
…yes, it’s pretty much an entirely different scenario. I just wanted us to feel like we had built a connection with our imaginary breeding population, OK?
Anyway, let’s look at the population. Let’s say that out of all the alleles floating around, a certain percent code for dark hair (A) and a certain percent code for fair hair (a). We can label those as p and q, expressed as fractions of 1, representing all the alleles in the population. So we have
so that if 40% of our hair color alleles are dominant (brown) and 60% are recessive (blond), we can say
These fractions also give us the chance that any randomly selected allele is either A (p) or a (q). For our simple scenario, p and q are our only allele fractions, so we can say
When we set up a Punnett square, we can look not only at the combinations of A and a, but also at the statistical chance of each combination. We get these by multiplying the fractions for the two alleles that combine in each block, so that an AA combination has a (0.4) x (0.4) = 0.16 or 16% chance of occurring.
For Hardy-Weinberg conditions to apply, we have to assume that hair color alleles aren’t significantly sex-linked, so that p and q are the same and A and a are distributed equally for men and women (another reason not to go with our original homozygous parents as a basis). That means that our Punnett square has the same values on the vertical and horizontal axes, so it comes out like this:
This is a setup that looks very similar to binomial expansion, which I covered a while back. I’ve illustrated our genetic scheme with the same kind of square I drew then to show the terms of the binomial expansion, which shows us that in this case we get those same binomial terms for the fraction of offspring that has each possible genotype:
or, more specifically,
To see how this actually works out for brown and blond hair, remember that the middle term is heterozygous (one of each allele) so the brown will dominate (again, a simplification compared to how it works in some cases). Here’s how the phenotypes add up:
With this first generation of offspring, we come to the reason this is called the Hardy-Weinberg equilibrium—this distribution of allele combinations will be the same for every generation from this point forward, regardless of how the first generation started out. The 60-40 split we used could have come from an initial population that was 60% homozygous blond and 40% homozygous dark-haired, but from now on (statistically speaking) it’s going to be 64% dark-haired and 36% light-haired.
Note that dark hair is more prevalent, even though more than half of the alleles are for blond hair! I chose the numbers for this scenario specifically to illustrate that dominant traits don’t necessarily have more alleles in the population. They’re not always the majority, either—if q for blond hair had been higher (say, 0.75), q2 would have been high enough for blonds to be a majority:
It seems that blonds are probably safe, considering the size of our human population. Of course, our conclusion depends on a whole list of assumptions, including the assumption that there’s nothing in particular keeping blonds from passing on their genes. If significantly fewer blonds were having kids, then yes, the number of a alleles in the mix would drop significantly. As far as I know, though, there’s no reason to suspect that’s a danger.
Now that your worst fears are allayed, read up on the details of the Hardy-Weinberg principle and decide for yourself how well it actually applies to the human population! Real genetic inheritance for many human traits is much more complicated than A or a, and it’s amazing to see how our diversity comes about. Whether your eyes are brown, blue, hazel, or an unsettling inferno of all-seeing fire, enjoy your reading!