In the recent paper on the genetics of Philistines they had good quality DNA from 10 individuals. Some archaeologists have criticized over-generalizing from such a small dataset. Naively I think this is a good caution. But we have many many ancient DNA results from humans now, and I think this naive objection needs to be tamped down some. Additionally, 10 samples in a genomic sense have a lot more information than that “10” might imply.

Genome-wide data is such that you can take one individual, and infer their ancestral lineage and so capture the history of many upstream in the genealogy. Additionally, when it comes to f-4 statistics and what not they used 20,000 markers.

Also, we’ve got some experience now with the “first” individual from given populations, and how representative and informative they were as more data came in. The Loschbour sample was the first of what we later called “Western Hunter-Gatherers” (WHG). Later WHG are all pretty similar to this individual, with only minor differences (the late Pleistocene “Villabruna cluster” prefigured it). The reason that the Loschbour sample worked so well is that human metapopulation dynamics seem to be characterized by rapid range expansions and population turnovers, especially in some regions of northern Eurasia. The genetics of “Cheddar Man” was surprising to no one within the field (actually it would have been a bigger publication if he was not so WHG).

Think of what Ma’lta and the first Neolithic farmers in Europe have taught us, and how little further samples from these cultures told us.

One of the things biologists like to say about humans is that we’re a young species that went through a bottleneck. In fact, there have been serial bottlenecks. That means there’s a lot of homogeneity in many groups across geographies. This doesn’t even take into account endogamy. Representativeness still matters…but the reality is that humans across a huge region don’t vary that much.

The main exceptions seem to be due to cultural barriers. Endogmany in South Asia, religious differences in the Near East, and variance in mode-of-production in Africa can mean that who you sample matters a great deal. The last was clearly operative early in the Holocene (early farmers often did not intermarry with hunter-gatherers), but I doubt the first two were particularly important until very complex literate polities emerged.