[Groxx]

It also wouldn't surprise me if "can reproduce with" isn't a transitive property (edit: actually, I'd be very surprised if it always is. genetic stuff is never that simple). So even if you try to collapse species into that strict definition, you'll still have edge cases that don't fit.

Categories are hard.

[bigbubba]

You're right, it isn't necessarily transitive. So called 'ring species' are cases where it isn't.

[scarmig]

You don't even have to go as exotic as ring species. Climb up the evolutionary tree through each of your ancestors to some slime creature a billion years ago, and then climb down to your pet dog. Each connected node in that step can interbreed, but you can't mate with Fido.

I guess it's kind of a temporal ring species.

[thaumasiotes]

Ring species are so called because their breeding relationships form a cycle. You're only describing a chain. You would have described a ring species if you could mate with your dog.

(Also note that while it's perfectly possible to walk up from humans to an ancestral slime creature, and then back down to dogs, most of the steps along that path will be shared between humans and dogs, like a Y with a very, very long leg. The recent common ancestor of humans and dogs would have already been a mammal.)

[scarmig]

The point about the most recent common ancestor is well taken, but, at least according to my layman's level understanding of ring species, the two ends of the "ring" aren't expected to interbreed. Wikipedia supports that definition (and if it's wrong, should be updated by someone more knowledgeable than myself).

[thaumasiotes]

You seem to be right about the idea of "ring species". Papers from the past several decades all introduce the concept the same way.

[scarmig]

Yeah, though I wouldn't be surprised if your suggested "continuous" ring species exists somewhere, with gene flow possible from both sides at every point but not from one side of the ring to the other (even if brought into the same geographic proximity).

In principle you even could have some kind of genetic helix, where one end of the ring species not only meets but overlaps with the other, perhaps going so far as to do a full wraparound a number of times. But probably the two ends of the ring would occupy too similar an ecological niche (even with all the genetic divergence) to allow for much time of overlap.

[thaumasiotes]

> It also wouldn't surprise me if "can reproduce with" isn't a transitive property

"Can reproduce with" is not a transitive (or even a permanent!) property among individual humans. (For a more general transitivity allowing for the fact that it doesn't make sense to ask whether males can reproduce with males.)