It's a very interesting question, but I don't think anybody can even come close to answering it right now.
Most speciation scenerios involve a large amount of genetic drift. The initial form of genetic drift would be sampling error. If a small population were to be isolated from a larger one, possibly by a geographic barrier, what are the chances that the allelic frequencies of the smaller population would be representative of the larger one? This is called "the founder effect", and could be considered "instant evolution."
This isn't even close to speciation, though. Next, random mutations would accumulate in the small population, further changing the allelic frequency of the new population. When the new population increases (a genetic bottleneck), it will have an allelic frequency dependent of the founder effect and these newer mutations. These new mutations may or may not have phenotypic effects. In order for speciation to occur, there needs to be some prezygotic or postzygotic mating barrier that would prevent inbreeding with the parent population. This can result from the random mutations, but a timeframe would be hard to pin down.
Natural selection may also play a large role in speciation. This will cause the new population to become adapted to new environmental conditions. After time, through pleiotropic effects, this may result in reproductive incompatibility with the parent population. Again, a timefrime would be hard to put down.
The situation you describe will be very different from the classical view of speciation I described above (allopatric). For instance, I'd imagine that a 100,000 person sample would include much of the genetic diversity of the parent population, essentially eliminating the "founder effect." Also, random mutations in the new populations aren't likely to effect the overall allelic frequency because the bottleneck effect is dependent on their being a small initial population. Next, humans are such a technologically innovative species that I think we would alter the environment to suit are needs before any serious effects of natural selection could take place.
This message has been edited by JustinC, 06-26-2005 02:39 AM