Such as finding some number of genetic variants at high frequency (whatever variants were carried by Adam and Eve), a small number of variants at very low frequency (the result of new mutations), and almost nothing in between. Instead, we see lots and lots of variants with a frequency distribution that falls off almost exactly as 1/f. That happens to be the frequency distribution that expect to find for a population that's been at the same size for a very long time. (In the case of humans, for ~50,000 generations or more.)
You can find more detailed explanations and the results of simulations
here, where I was contrasting observed data with what's see from a bottleneck 2000 generations ago. (I post there as glipsnort.)