 If the contracting earth hypothesis were true, then what you'd expect is a random distribution of elevations of the Earth's surface. So on this plot, I've got elevation on the y-axis, and that thick line represents c-level, and then I've got sort of an arbitrary frequency axis on the x. So here's what a random distribution would look like. In fact, the mean elevation on Earth is actually about two kilometers below c-level. And so if the contracting earth hypothesis were right, this distribution is what we'd expect to find. But in fact, this isn't the real distribution. The real distribution looks like this. It's a bimodal, and there are actually two peaks. One is near c-level, which is the average elevation of continents, and then the other one, the peak at about almost negative five kilometers, corresponds to the abyssal c-floor. So statistically what this tells you is that the process by which these elevations are distributed was not random. In fact, there's probably two different processes, one that makes continental crust and one that makes oceanic crust. And once you take that into account, you realize that the contracting earth hypothesis doesn't really work very well.