The researchers reverted a population of lab-bred fruit flies to their natural state, state from which their ancestors were removed in 1975, and observed if and how "reverse evolution" happened through 50 generations. The results are that some traits and genes did reverse fast, while others did slowly and yet others did not de-evolve at all. Overall the flies got only like 50% of their genome back to their ancestral state only, while the phenotype instead in some cases fully reverted to the feral state.
The first conclusion is obvious: history cannot be fully reversed, not even in the field of genetics. Historical accidents may change the future indepently of enviromental conditions.
The second conclusion is the somewhat shocking fact that different genetics can encode the same phenotype.
This is most interesting also in regard to humans, I'd say, because people with different genetic backgrounds may happen to converge to a similar phenotype, which is not necesarily just the average of the genes involved. This phenomenon, if my memory doesn't fail, has been noticed in some areas of significative admixture, like Mexico and Sudan, where the resulting phenotype of admixed populations was not merely the simple "hybrid" type but a rather new local phenotype of its own (a new "race" so to say). In different contexts it is very possible that people of varied genetic backgrounds, say Askhenazis and Germans, may converge to similar phenotype without massive admixture (I don't mean here to say that Askhenazis and Germans/Poles are not mixed with each other, they obviously are, but that the resulting phenotype modification of so many milennia of co-residence may be much greater than the mere genetic input, whatever the exact enviromental reasons (climatic, social, whatever). .