The point still stands, though, that race is not a single characteristic; it's a combination of many different characteristics, which are hereditary, which is to say, genetically determined. Not being a geneticist, I can't quote offhand the genes or gene clusters that encode for, say, fuzzy/wiry hair. But these genes, wherever they are, are common to Africans but not generally found in Europeans or Asians. They must be there, otherwise this characteristic would not exist, right? And this characteristic is just one of many shared by Africans and not generally shared by non-Africans; of course, as Leroi says, there are characteristics shared by people from some parts of the continent not shared with Africans from other parts, and there are genetic markers that don't encode for visible characteristics that can nonetheless be identified by genome sequencing. (Again, please point out anything I've said here that's incorrect.)
So that, in a nutshell, is what I mean by 'race' - not discrete, well-demarcated 'breeds' of mankind, or representative 'stages' of human evolution, or groups of people who are inherently better than other groups, or any of the other patently untrue and obsolete ideas that have been put about in the past by people who studied, or purported to study, race in the past. Hey, the meaning of scientific words changes: the atoms of Democritus were a lot different from the atoms of modern physicists and chemists, but we still use the same word, right? If the very word race is too loaded and contentious for modern use, then we can use 'genealogical ethnicity' or something instead if you really want to.