Trinkaus on mosaicism at the archaic-modern boundary

In a review article on the earliest modern humans (more on this later), Erik Trinkaus has this short, suggestive paragraph that stands apart from the rest of the discussion:

There has been little consideration of the biological implications of what we perceive as mosaic morphology. These were functioning organisms. The combinations of features observed paleontologically must have been biologically integrated to some degree. Thus, the mosaics can be investigated profitably in terms of their paleobiological implications (Trinkaus 2005:221).

Sometimes in an article like this, people include ideas that occur to them as interesting, but without having given them much thought.

Other times, they include a slight foreshadow of a substantial research project already underway.

For Trinkaus, I'm not sure which this represents. But there is some interesting subject matter here for someone. Why do some features of archaic humans persist longer than others? Why are some still around -- including many that distinguish different regional populations from each other today? And what were the selective gradients that allowed the full phenotypic complements of those features to fade away, gradually replaced by the emerging package of modern humanity? It's the sort of stuff that keeps you awake at night...