I'm going to be careful because this might be considered somewhat controversial, but I don't have any sinister subtext going on or anything. I'm just curious.
In some long, long ago PoD, how likely is it for different and unique phenotypical human groups to develop. Groups that are distinct from what we know, but recognisable as groups.
An example that always comes to mind for me is a phenotype of having red hair, pale skin but with Asiatic epicanthic folds to the eyes. Red hair does very very rarely appear in Japan apparentally so it's not completely out of the realm of possibility.
Imagine if the Americas has been first colonised by Caucasians instead of Asiatics, after a few thousand years the local populations would be highly likely to be phenotypically distinct from Eurasian Caucasians. It is highly difficult to predict exactly how though, I imagine they would appear more akin to Iranians and north Indians than European populations.
Imagine that groups of Malays colonised the Australian coastline and merged with the Aborigine population, creating a new groups analogous to the Malagasy. These groups colonise New Zealand, where they are conquered by the Maori but too numerous to be wiped out completely. Instead a caste system forms, with Polynesians at the top and the Malay-Aborigine New Zealand natives below. Or perhaps they remain too numerous and technically advanced to be subject to Polynesian aggression in the first place.
Imagine uber-TLs but for some of the very marginalised people of our world. Imagine something akin to the BANW timeline, but the developers of the advanced navigational package are the Guanches of the Canary Islands, who then set out to colonise all the islands of the Atlantic and take over many islands in the Caribbean, creating peculiar islands populations of blond-haired Neolithic people. Something similar could be done with the Ainu of Japan, which could lead to a Hawaii inhabited by short, heavily-bearded Asians.
What started this idea was reading this
http://news.bbc.co.uk/2/hi/science/nature/7358868.stm article, talking about the near-split in human lineages. Imagine if the East African population had been wiped out, and the Khoi-San human groups then went on to spread out and colonise Africa and the world alone. The human populations of the world would still diverge, but they'd all be coming from the Khoisan mitochondrial roots. Which is probably what happening at an earlier stage in OTL anyway, but in this case it is likely that the resultant human phenotypical groups would look vastly different.