Suppose that 90% of people are farmers, and 10% are in other occupations, occupations that on the whole have greater prestige and remuneration. Like being a tailor, or a banker, or a king.
There is reason to think that those upper classes , in most societies have somewhat higher intelligence – and somewhat higher genetic potential for intelligence. More complex jobs.
If we were examining the family background of people of unusual intelligence – say prominent mathematicians – we might notice that the upper 10% was over-represented among their parents. Partly, this would be because those more prosperous people were able to pay for books, higher education, etc. But their average higher genetic potential for intelligence also mattered.
Since the fraction of individuals that exceed a high standard drops off rapidly as the cutoff increases, it is even possible for the upper class to produce more people exceeding the cutoff than than the bottom 90% of the population, if the cutoff is high enough. So, the typical world-class European mathematician was likely to have a pretty bourgeois background, even though those bourgeois were a small fraction of the population. It was partly educational advantage from more money, but it wasn’t just that.
Consider running ability: a small population with higher average running ability may produce runners better than the best from a very large population ( one village in Ethiopia vs all of China).
Gauss was anomalous – about as anomalous as a kid two feet taller than his parents.