An essay by Michael Balter in Science
This question seems to have an obvious answer. If you're in Africa and thinking about going somewhere else, you're going to have to go through the North part to get anywhere. South Africa seems like a really bad place to look for a "launch pad" of human migrations.
Lots of people have written about South Africa as a "cradle" for modern human behavior. The density of high-quality archaeological sites explains this focus. The Howieson's Poort and Still Bay industries are genuinely interesting, and we can further examine a broader sampling of Middle Stone Age discoveries such as early adhesive use, heat-treated pressure flaking or cereal gathering. Still, most of these developments are very late in the game. If we're looking at things in South Africa 70,000 years ago, that's substantially later than the key events leading to human diversification within Africa.
Balter reports on work that has during the past few years uncovered old dates for Aterian sites in North Africa. This regional variant of the Middle Stone Age is recognizable for its distinctive "tanged" points, and now extends from as early as 140,000 years ago to less than 40,000 years ago. The early end of this range is old enough to contribute to a possible dispersal of North/Northeast Africans into Eurasia. Hence Balter's story. On the other hand, if the Aterian were actually relevant to the movement of people into Eurasia, it is curious that Levantine Middle Paleolithic doesn't show clearer similarities to it.
Most interesting detail: new skeletal material from Morocco:
Last year, archaeologists excavating at the Grotte des Contrebandiers (Smuggler's Cave) on Morocco's Atlantic coast unearthed a rare prize: the skull and partial skeleton of a 7- or 8-year-old child. The fossils, dated to 108,000 years ago, appear to belong to an early member of our species, although study of them has just begun.
I think it is increasingly likely that we will have genetics out of these North African materials. The fluctuating humidity of the Sahara (a focus of the article) does complicate matters, but the technology has progressed so rapidly that a well-preserved skeleton will surely turn up some endogenous DNA.
Balter continues his story with some morphological analysis of North African materials. These don't as a group share any special similarities with Neandertals or people outside Africa -- although individual specimens do have features that show up elsewhere. The North African specimens (spanning from Jebel Irhoud at 160,000 years to Nazlet Khater at 40,000) are as you might expect really variable. They don't look particularly like recent peoples of North Africa, either -- "modern" in this context often means "not Neandertal" and masks some of the change that has taken place in the last 100,000 years. In other words, it's tough to tell a simple story of a North African ancestral population giving rise to variation outside of Africa, at least not without substantial evolutionary change in the non-Africans. It's not obvious how much of this evolution might be explained by mixture with Neandertals and other archaic Eurasians.
Anyway, this explains why many paleoanthropologists don't see the Upper Paleolithic and equivalent-aged specimens outside of Africa as particularly African-looking. What remains unexplained is how much does morphology reflect ancestry over this kind of time span?
I don't intend to answer the question, it needs more serious treatment.
The genetic results have changed quickly over the past year. We will need to apply a somewhat different frame to North Africa -- one that recognizes a deeper differentiation of human populations within Africa (pre-Aterian, certainly, even with the older dates). We also have to resolve the biogeographic relationship across the Sinai between the Nile corridor and Levant.
All of this means that Balter's story is very timely. Discovering more about the archaeology of Northeast Africa, the Arabian Peninsula and surrounding regions will obviously be crucial to understanding the rise of humans during the last 100,000 years.