Today I attended a kind of mini-symposium featuring five short talks from an ecology focused university department. The main topic was the use of molecular data in ecology, and I learned a lot of fascinating things especially about recovery of animal populations after bush fires and about Antarctic ice age refugia.
The discussion, however, turned very strongly towards species delimitation, and this is where things got a bit weird. It was remarkable how self-confidently some, say, colleagues specialising on the behaviour of a few selected fluffy mammals pronounce some black and white opinion on a problem that systematists have been struggling with for centuries.
Let's just say that if the situation were reversed - perhaps me being asked about fire ecology after giving a talk on plant systematics - I hope I would be able to stick to something on the lines of "well, this is what I think, but really you should ask an ecologist". I sure hope I would not simply say "based on what I observed in one plant species, fire has no effects whatsoever on the genetic diversity of animals". Some of the ecologists here had no such reservations.
At one end of the spectrum several audience members appeared to expect that now that we have genomic data we should finally figure out an objective cut-off for how much genetic difference between two individuals puts them into different species. This is one of those ideas that look superficially attractive but reveal new layers of wrongness the longer one thinks about them.
The easiest answer was provided by one of the speakers: Difference in what molecular marker(s)? Different genes or regions evolve at very different speeds. Taking the whole genome also seems a bit suspect as most of it is junk DNA, the exact amount varying by species. But again, there are layers. The speed of molecular evolution would also differ vastly between different lineages depending on their generation times, mode of reproduction, perhaps other life history traits, and the environment they find themselves in. The genetic diversity within populations is also vastly different from species to species.
But what mostly blows the idea out of the water is quite simply that there is necessarily a grey zone between being one species and having split into two species. ANY cut-off would have to be entirely arbitrary.
That brings me to the other end of the spectrum. After the talks, one of the speakers declared over snacks all of the following: (a) we should use the Biological Species Concept, (b) species are arbitrary human constructs and have no empirical reality, and (c) genetic data will always give you two clear groups, only we must be careful how we interpret that.
My first observation is that it should not really be possible to believe these things at the same time, because (a) and (c) are in direct contradiction to (b).
The second is that I consider all three to be wrong. It is clear that the BSC simply does not apply in many cases, especially asexual species and fossils.
As for the arbitrariness of species, well, it depends. (This is also the only answer to the species problem that I find useful.) I wrote above that there is a grey area, that a cut-off for "moment of speciation" is arbitrary. That being said, the same is true for many other things that we happily classify. We can probably all agree that the cut-off between child and adult is arbitrarily placed at the eighteenth birthday; it could just as well be a month earlier or three years later.
Now here is the question: Would you say that toddlers should be drafted into the army? No? Seems as if the difference is not so arbitrary after all, even if there is a gradient between the two categories. Likewise, there is nothing arbitrary about the question whether humans and horses are the same species or not.
Finally, genetic data always giving two discrete clusters? Ha, I wish.
In summary, species are complicated. I would be very skeptical of any claim that somebody has sorted it all out.