Another instance of easily observable speciation occurs quite frequently in plants, called polyploidy. In this case, when an offspring is produced, their chromosomes do not separate during meiosis, and they receive double the genetic material from their parents. I.e., if the parents had 14 chromosomes, the offspring has 28. Therefore, when the offspring produces gametes, the games have 14 chromosomes, but the parents generation's gametes have 7, so they the offspring of that breeding would have 21 chromosomes and be triploid. Triploid organisms more often than not are not viable. So this would be a case for speciation within one generation. Granted, it doesn't involve natural selection per se, but because the polyploid individual cannot breed with it's parents' generation, it is reproductively isolated, and natural selection and genetic drift can act upon the new species to make it different from the parent species.
Another species to look at is orcas, especially the ones living off of the coasts of British Columbia and Washington. There are 3 "types" of orcas: Transient, Resident, and Off shore. Even though their ranges overlap some, they maintain themselves as very separate groups and almost never interbreed. I did a project in undergrad for a population genetics course where I went through the literature to analyze the pop. gen. data, and it's pretty evident that this different "types" of orcas are quite possibly on the road to becoming different species. This is also a very interesting case, because the populations are not separated by physical barriers, but rather by cultural barriers. The different groups eat different food, use different sounds for communication, and have different pod structures which is why they tend not to mix.
We have many intuitions in our life and the point is that many of these intuitions are wrong. The question is, are we going to test those intuitions?
-Dan Ariely