Ordinarily, a Friday post includes film segments, but I’m struck low by a vile cold (possibly in retaliation for revealing too many parasites’ secrets last Friday) and am not suitable for visual-audio company. I may revise this post in the future to replace some of its parts with filmed bits.
The human is a brown animal. Like other mammals, we have two pigments, eumelanin and pheomelanin; since most of our skin is visible, that’s the structure I’m discussing; although both pigments are present in our skin, eumelanin is by far the more significant one. It’s brown. There isn’t any white, black, red, or yellow person.
Let’s quickly get some basics down…
- The melanocytes, or cell components which make melanin, are uniformly distributed across the area for a given region of the body. The regions differ in their typical number.
- The melanosomes which contain melanin, once it’s made, are what provide that region with its visible color, or rather, color based on pigmentation.
- Differences in in pigmentations among regions of the human skin surface are real, but except for a few places, not sharply bordered or as distinctive as may be found among many other mammals.
The range of visible human skin coloration is derived from the number of melanosomes – i.e., not the skin’s capacity to make the melanin, but rather, how much is made and contained in the melanosomes. If you like, “it’s not the paint, it’s the painters.” That’s why this post doesn’t concern albinism, which is due to lack of pigment rather than how much it’s distributed into melanosomes.
Now for the first problem: that your brain frequently doesn’t use your eyes. We both assign and associate to the detriment of genuinely observing what other people look like.
Here’s a look at assignment. How could anyone sensible consider the people in these photos to be differently hued, let alone call those in one photo “yellow” and the other “white?”
Association means seeing more similarity than is actually there based on one variable, because you “just know” that thing X carries things Y and Z along (when they don’t). It can go either way, for instance, seeing the same faces due to similar coloration; how could anyone sensible consider the people in these photos to be ethnically tagged with the same word “black,” or the same word at all?
Association can go the other way too, assigning coloration when the features are similar, such that two people can be perceived as having the same pigmentation when they don’t, preserving the belief that people with the same ethnic appearance must be pigmented the same.
These habits also manifest when we depict ancient communities. 19th century British and Germanic education frequently assigned the current population’s appearance to the ancient Romans and Greeks, and this exact assignment carries through today in our visual arts. This is persistent beyond belief … there are whole branches of scholarship trying to justify why the ancient Macedonians looked like Vikings, in blatant defiance of the contemporary depictions which show us completely ordinary eastern Mediterranean people. And although audiences might find it absurd for an English or Irish actor to depict, say, a Tibetan ruler of two millennia ago with no special makeup, we apparently expect it for the ruler in a purportedly-realistic TV show set in Rome of that time period.
Next question: What controls variation in human skin color? – despite the short article’s charming optimism, I think the general viewpoint is off on the wrong road. More and more fine-grained analysis won’t help if we’re not making sense about the level of analysis we already have. Let’s take a look at that, with some more biology vocab:
- Quantitative genetics = when a feature displays a smooth gradient of variation as opposed to separate and obvious alternatives
- Multiple alleles = when varying instructions from many genes provide redundant input to a single developing feature
Here is a terrifying handout if you’d like. In plain language, a lot of genes influence the melanosomes in our skin, and each of those tends to “say” or “do” the same thing. Each one says “just a little,” “a lot,” or “hey, very much lots,” and the ultimate result is the sum of what they do. So it doesn’t matter specifically which of these genes is doing what, so much as how much they collectively do.
[FYI – this is much easier to explain in speech; I’m definitely going to replace this part with video when I’m better]
Ages ago, these phenomena were typically taught as exceptions to the standard one-gene two-allele model for a given trait, probably because of the latter’s successful application to sickle-cell and certain other problems. However, now that we’re armed with the Devo Evo insights of the 1980s and 1990s, fellow biologists will quickly spot that the concepts fold easily into the newer perspective. It turns out that the majority of our features, and those of most critters, need this sort of analysis instead the standard model. This also means, for evolutionary talk, never mind what genes are “for,” we should be assessing what they merely do: developmentally, interactively, and statistically.
Time for the really technical material: the inheritance. Remember how I said each gene does three things? Well, that’s because it has two versions, called alleles. In this case an allele results in either a little melanosomes or a lot of them – and you have two instructions, hence little-little, little-lots, and lots-lots. Again, this is additive, not the dominant/recessive you learned about in school. It’s actually called incomplete dominance.
Therefore with all these genes doing this at once … let’s say, ten of them (no one knows how many are actually involved, for human skin pigmentation). That would mean you have twenty “genetic input” units, and each one is little-little, little-lots, or lots-lots. Meaning, twenty little-little (which is light-skinned person, not an albino) grading one “lots” at a time all the way to twenty lots-lots. Cue horrible handout #2.
Multiple genes with identical or near-identical effects, and incomplete dominance in each, yields a wide and finely graded spectrum of effect, with many genetic ways to get to the broad middle range. A person doesn’t have enough genetic range to contain this full inheritable potential all by himself or herself, and since he or she provides one allele per gene to a given offspring, and in this case for many genes regarding this feature …
- Given parents who are both toward one of the far ends of the range, their children land near the corresponding end
- If one parent is at one of the far ends, and the other is at the other, their children land square in the middle of the range
- For any of the other possible pairings of two human parents, their children may land anywhere across the broad middle range, with the mode (for those parents) depending on the exact distribution of alleles across the relevant genes
Now for some more evolutionary talk based on one of the concept’s most powerful components: biogeography. A fine variable for mapping historical dispersal is mitochondrial DNA sequences. Many human features track this history very well: facial proportions, frequencies of blood type, and more.
Here’s one of the earlier branches for groupings as we see them today:
This diagram has been extensively reassessed and the African sections, especially, have undergone refinement. But it’s correct enough for you to scan up and down the named groups, maybe run some searches if you want to get a visual or two, and to discover the following observations.
- The human is indeed brown. Furthermore, most of us are very similarly brown, in the broad middle of the human coloration range, regardless of which dispersal branch we’re talking about … and that habits of assignment ruin our observation of how medium-brown most people really are.
- Each identifiable dispersal branch of our species history includes a wide range of pigmentation across its smallest, i.e., most recent branches. I can’t stress this enough: that either far-end of the human coloration range is no indicator of origin for any cluster or grouping of humans based on historical ethnicity. It’s a “twig,” not a “root.”
- The darkest and lightest human skin colors are found at the margins of distribution and/or in isolation.
Got it? Human skin coloration does not track the dispersal! It’s a scattershot effect within each branch instead.
To punch these points further: the two (and only) wide distributions of these extremes – lighter across Europe, darker across Africa – have nothing at all to do with the dispersal branches or with origins; they are much more recent phenomena based on population expansions in the past 2000 years. Use your eyes and you’ll see a lot more pigment diversity in each of those places too, and not just due to recent immigration.
Which suggests that the most-familiar and frankly, uncritical phrasing that skin pigmentation is “for” protection against UV radiation, needs a stern assessment. At the very least, I think that selection favoring any such thing occurs atop a number of other population phenomena which strongly affect allelic diversity. The two most important in this case are …
- The founder effect means that when a subpopulation is isolated or at least significantly separated from a parent population, it may include different frequencies of alleles than the whole original population did.
- In serial, the populations are dispersing farther and farther across a landscape, such that each successive group is a subset of the previous one, thus beginning each one with less and less of the original group’s genetic diversity.
- Inbreeding, in non-derogatory terms, means that mating occurs among even somewhat related individuals – no need for social incest, just a fair amount of second-cousins. The effect is to increase homozygosity (AA or aa, at the expense of Aa) across the population.
Each of these effects is much, much more likely if the newly-dispersed population(s) is or are small – to the extent that, for numbers typical for historical examples of dispersal, in the dozens or low hundreds. Now put them together and …
Really? Merely going somewhere, in typical numbers that wouldn’t raise an eyebrow, can slam the local population’s pigmentation one way or another? It can go either way? With no reference to fitness or selection? Yes, really. If the founder effect starts the population with a slight bias toward one end of the range or another, and with even just a couple of generations of technical inbreeding, such an effect is much more likely than otherwise.
Here’s where I break with the common narrative that skin pigmentation is”for” protection against UV radiation. First, just because a health problem is involved, doesn’t mean it has a history based on reproductive success. Selection doesn’t exist to rid us of health problems. Second, just because something is important to us psychologically doesn’t mean that exact variable has been acted upon by selection. Third, I’m not claiming that there has never been any selection regarding human skin color, but rather that the frequency and distribution of alleles upon which it may act is already deeply affected by circumstances.
Therefore it may be that humanity prior to our own species dispersal was relatively more pigmented than the median of our current whole-species range, based on selection for protection against UV radiation … but to make this claim more than maybe/maybe-not/so-what one must assume (i) that this environmental factor was and is a consistent significant threat to reproductive success, and (ii) that current pigmentation distribution across Africa reflects this incredibly ancient selective situation more than it reflects recent population expansions and interactions. I don’t find these assumptions to be especially strong.