Btw, the premise underlying this fact is a fallacy. It is the norm for population groups, even meaningfully differing ones, to have "more internal variation" than between groups.
There is more variation in height within sexes than between them, but despite this there is still a very (statistically and culturally) significant difference in average height between men and women.
It's even possible for two distributions to be completely separated by a given variable while still having more internal variance.