Thank you all for your answers, very informative!
chadn737, as for the definition of species, is it not the ability to interbreed that is the key to differentiating between species? If not, then my question is misleading indeed. If yes, then maybe speciation arises over a bumpy period of decreasing compatibility, with a series of less fertile offspring along the way…
Charon, I get the idea of reproductive isolation. Is that a necessary requirement for all new species? EDIT: the Kronforst study you mention answers this question. With a "no".
I wonder if I can look at this differently - it is not that new distinct species arise, it is that all intermediary individuals (hybrids, those in the "gap") simply die out or are not numerous enough.