The Weinberg et als.' paper that I cited documents the following direct evidence of speciation observed under controlled conditions: In 1964, Dr. D. J. Reish removed 5 or 6 polychaetes (Nereis acuminata) from Los Angeles/Long Beach Harbor, and grew his sample to a size of thousands. In 1986, four pairs from this group were brought to Woods Hole Oceanographic Institution; the population at Woods Hole thus had gone through two bottlenecks, which are supposed to help drive evolution through genetic drift. In 1977-1978, two new cultures of N. acuminata were gathered from nearby Long Beach and Newport Beach, and grown under the same conditions as the Woods Hole sample. The three populations were later crossed, and it was found that the only crosses that would not produce viable offspring were the crosses involving the Woods Hole culture and the two new cultures. This signifies nothing less than speciation, and all in the laboratory as well.
There are literally hundreds of other observed instances of speciation taking place in real-time. See http://www.talkorigins.org/faqs/faq-speciation.html
"This signifies nothing less than speciation, and all in the laboratory as well. "
Sorry, but I find it absurd to simply accept the "it depends on the meaning of the word species" as a justifiable argument.
Do it with a horsefly and you might have my attention.