Slashdot Mirror
Pants Were Optional, 100,000 Years Ago
RobertB-DC writes "German scientists have used differences in the DNA of lice to determine when humans started wearing clothes. It seems lice are highly specialized -- head lice lay their eggs only on hair, while body lice hide theirs in the folds of clothing. Using the differences in the two species' DNA and a "standard" mutation rate, the scientists determined when clothing-specific lice (and by extention, clothes) came into existence. No comment, though, from Calvin Klein."
washing clothes wasn't invented until 5,000 years ago.
They're not optional now? That probably explains a few things...
comes as no supprise to CmdrTaco, as he has long known that:
Pants are optional, but recommended for you.
This sig intentionally left blank.
You must be a programmer.
Best Slashdot comment ever
> The article states that the "scientists" calculated one metronome per 30,000 years and thus concluded that body lice branched off from head lice about 72,000 years ago. What?!?!? How likely is it that mutations really occur on average without much of a deviation from the mean that regularly?
You do have to be careful about that sort of thing. For example, attempts to apply the same logic to language (glottochronology) are not generally accepted by linguists because so much of language change is driven by social factors rather than blind mechanistic processes.
> For all we know, mutations occur in leaps and bounds. It might be very similar to those annoying studies of amortized cost in my algorithms classes. Sure, great, probability theory is great and all, but what about reality?
I'm certainly no expert on this, but there are several things that appear to be working in favor of using mutations as a biological dating system. For one the molecular mutations that form the basis of the method do appear to be the result of blind mechanistic processes - at least if you can avoid the error of measuring parts of the genome that are subject to pressures for or against preservation or change. For another you've got the Law of Large Numbers working in your favor, both in terms of the size of the DNA molecule and the number of generations. Unless we're missing something these factors should conspire to give us an expected value for the mutation rate and an opportunity to average it over a very large number of events (length of DNA * number of generations), allowing us to apply standard statistical methods to calculate an expected value for the number of mutations and a confidence range for it.
Also, we live plenty long enough to measure mutation rates between generations of organisms today, including humans as well as lice and other species. It should be pretty easy to calculate an expected value of number of mutations per unit length of DNA per generation, and see how much variance there is in that number today, and compare that to what our calculations suggest for the evolutionary timescale. (I'm not familiar with the literature on the actual numbers but I know at least some of it has been done, because a few years ago I read something about the typical number of mutations observed in human babies. I'm not citing the number because I don't trust my memory on it. But the fact that they could even name a number goes far toward establishing the kind of model you need for this kind of dating.)
A big problem for the method would be if mutation rates have changes significantly over time, e.g. due to radiation or environmental chemistry, and this kind of stuff is hard to check directly. However, science is "convergent" in the sense that we expect our various theoretical approaches to give the same answer if they are in fact correct, so the fact that this study produces a number that matches the previously established number for when our ancestors left Africa and moved into regions where clothes would be necessary, all adds up to a satisfyingly consistent model of what happened to cause all the relevant observations.
There is of course the epistemological problem of the inability to prove anything in the empirical sciences, but since that problem is unsurmountable we more or less ignore it and take our supportable results as "true" - but not as "Truth" - so long as the explanation seems to work and converges with all our other models for what's going on in the universe. If we discover later that we did something wrong we simply have to revise our results when that time comes, but that's an unavoidable risk we have to take; the only other alternative is to throw up our hands in despair and not try to understand the world at all.
If you want a more expert analysis of any of this you might want to post it as a question on talk.origins, which is inhabited by all manner of biologists, mathematicians, etc., who can daze you with more than you want to know about virtually any topic.
Sheesh, evil *and* a jerk. -- Jade