Conditions for speciation
I don't think that you *always* need thousands of years, geographical separation, and a long series of small mutations to begin to obtain a new species. I think you need much less than that: you probably need a single mutation (in the right place) that affects the reproductive system in such a way that the individuals who carry two copies of the new gene will able to produce a fertile descendance when mating among themselves but won't produce a fertile descendance when mating with the bearers of the original gene. If luck is on its side, this new gene may begin to spread slowly over the population. *Then* a long series of little mutations will start to accumulate in the new species, and the distance between the two species will be increasing generation after generation.
I can't see geographical separation as being a *necessary* condition for speciation because (I think) it's a rare event while speciation is a relatively frequent event. I think the way geographical separation leads to speciation is different.
Actually I have a problem figuring out how geographical separation can lead to speciation by itself, and so I would be grateful if someone explained this in genetic terms. On the other hand, I think that when speciation does occur within a geographically isolated population of a given species, it may happen that the new species will be more successful than the parent species and eventually replace it completely in their "island", while being unable to migrate to the rest of the world. If this is what happens, the results are the same as though the separation itself had caused the speciation, but I think the probabilities for this kind of event are much higher.
Well, this is just my current opinion. I may be overlooking something.