"there IS in fact a knowable, discrete threshold moment for many species nowadays ... it's the moment someone kicks off a maximum likelihood tree analysis on DNA data. Eureka/el voila!!! New species emerge!"
>>> I defer once again to my favorite Mike Sandel quote. (see signature block below)
"reproductive isolation also gets used a lot in textbooks, but seems to be ignored for a lot of fish taxonomy. For instance, a lot of the centarchids interbreed easily enough in the wild, but are considered separate species."
>>> They do, BUT the majority of individuals in natural habitats are pretty clearly one or the other species, indicating that the hybrids are not very successful over multiple generations, and selective mating with the "right" species is still the dominant trend. (This can certainly break down in a managed pond, or in natural habitats where an exotic has been introduced). When hybrids of two former species become reproductively successful and dominant, then a new species may be created by merger. There are several suspected hybrid-origin species, including Luxilus albeolus.
Also, morphology never really defined a species; it was (and still is, along with genes) a tool for distinguishing species. Defining a sexually-reproducing species is (and should always have been) based on drawing boundaries between populations across which there is relatively little long-term successful mixing. It doesn't matter how many hybrids are born if they fail to leave successful fertile grandkids. Now the weird part: Given that concept, the environment is also a factor in defining species. If long-term habitat conditions (temperature, chemistry, competitors, pathogens, food organisms, etc) change in ways that makes multi-generational hybrids more successful than the parent species, then species may merge.
Like Doug, I'm an old guy who doesn't know crap about current genetic research methods and evolutionary tree theory, so take my comments with that disclaimer.