This is a typical conversation in my kitchen:

I'm currently watching the first season of Battlestar Galactica, usually while cooking dinner. The premise of the show is that an evil race of robots destroys almost the entire human race, leaving only a small population of 50,000 individuals alive.

This immediately prompted us to pause the movie and launch into a discussion of the population genetics implications of such a crash.

Now, contrary to conventional wisdom, there is a huge difference between the actual and apparent genetic size of a population. That is, even though there are 6 billion people on this planet, human beings are remarkably genetically uniform; in fact, they show the amount of variation one would expect from a much smaller ideal population. This discrepancy leads population geneticists to speak of a quantity called effective population size - in the case of humans, about 10,000 individuals. This is such a small number because, first, there is population structure that prevents truly random mating between all individuals in the population, and, second, the human population has undergone at least a few "bottlenecks" - instances of dramatic population collapses - and the ancestral population was probably considerably smaller than the modern population. Thus, the amount of actual variation in our population is low.

A bit of Googling around led us to a paper by Masatoshi Nei, wherein he describes simulations of population crashes and subsequent rapid expansion. The upshot is: even absurdly unrealistic population crashes (down to N=2 individuals) do not eliminate most of the variation. I think in that extreme instance, the reduction was only from 15% down to 8% - so long as your 2 individuals are randomly selected, of course. This is pretty remarkable, and it suggests that the immediate problem will not be fitness loss from low genetic variation - you'll lose most rare, private variations (variation that only exists in you & your immediate relatives, e.g.), but most of this stuff is unimportant or even harmful. What WILL be retained is the bulk of frequent variation - stuff that is either beneficial or neutral, and so has not been eliminated by purifying selection. Of course, it takes on the order of 1/(mutation rate) generations (10⁸) to recover your initial level of variation, but never mind that. We're merely concerned with survival, here.

The second problem is inbreeding depression. This is exactly what it sounds like: when you have babies with your parent or sibling, you're much more likely to encounter severe recessive phenotypes that drastically reduce fitness. Note that this is a DIFFERENT problem from low genetic variation - I am genetically not all too distant from anyone on earth, but inbreeding depression results mostly from the expression of rare, private variation that no one else has - except my relatives.

So, after a bottleneck, inbreeding depression will surely be a problem. This possibly results in 'purging', that is, the speedy elimination of deleterious variation via selection in a highly homozygous/inbred population. This puts pressure on the population - inbreeding load - fewer individuals are surviving and the general fitness of the population is lower. Populations may founder at this stage, although purging is believed to result in a rapid recovery of fitness.

At any rate, 50,000 individuals is, in genetic terms, a great many, and I'm pretty certain that inbreeding depression would not be a severe problem in such a population (especially if they're all free to mingle as the Galactica population would be).

Last night's conversation included: group selection in chickens and its eugenics implications, a recitation of some Hindu mythology (mostly the Avatars of Vishnu), some tales of the Buddha and other Zen masters, the story of Kalidas, a discussion of the Inquisition as viewed by the book Demon Lovers, and Hitler's vegetarianism. The evening culminated with us watching Triumph des Willens.^*

---

^* Which prompted the observation on my part, "Americans will never be able to take the German language seriously."