@m Don't think of a species as an entity, but as a set of specimens. In species with obligate sex of at least two participants, speciation can't happen via a single mutation creating a reproductive barrier. Speciation in such cases must necessarily happen (somewhat) gradually, with parts of what used to be one population drifting apart, until they become two populations. (This doesn't refute punctuated equilibrium; in the hypothetical equilibrium-punctuation case, the gradual drifting-apart would just happen relatively fast, but still, in multiple steps, and both subpopulations would eventually stabilise.)