Well, again, according to Guns, Germs, and Steel, you have to take into account the fact that there were large conquests/migrations that blur those lines one way or the other.
That map shows "indigenous populations", but that basically only means "what the population was like a century or so ago".
The darker-skinned Africans from just south of the Sahara actually made a major invasion into southern Africa, relatively soon before the Age of Exploration brought colonialization, in much the same way that Saxons invaded England or the Huns invaded what became Hungary.
You have to take into account the fact that any culture that trades together will intermingle, and you'll occasionally get large population shifts when you're talking about why skin colors stratify in different areas.
The Arabs traded from the east coast of Africa to as far as China, and the Chinese (often led by the ethnicity from northeastern China) traded back across essentially the same range by sea - they were relatively lighter-skinned and traded and intermingled freely with the peoples of Southeast Asia, which would keep them lighter-skinned, as well.
Conquest can play some role in creating that "selection" idea finka was talking about, as well, however. In India, the caste system was essentially based upon skin color - the conquering paler-skinned people put themselves in the top caste, and gradiated society by what skin color they had. Of course, that could also play the other way around, with a preference for darker skin when you deal with conquering darker-skinned Equatorial Africans invading lighter-skinned Africans from further south.
Of course, I would also kind of point to how that map paints the palest-skins with too wide a brush - all of Europe, half of Asia, and even a fair chunk of North Africa are all just plain "white", which implies that Sicilians, Mandarin Chinese, and Scandinavians are all as pale as one another, which kind of hurts attempts to look at that and find where the population migrations took place.