Clutch

I don't want to bump up Evolskeptic's shameful thread, but there was something of interest there that I'd like to hear more about. Dr Rick asked,
To which pz replied:
My thoroughly inexpert understanding of these matters was this: Variation can accumulate within a (more or less) panmictic population, but none of the variations counts as an adaptation in the absence of some selection pressure that begins to favour it. Such a pressure "pares away" the elements of population not possessing the variation (now an adaptation), and now there's a new species or two.

I know this is oversimplified, but that's why I figured that speciation was the name for the whole process, but which only really becomes visible after the final conservative stage -- selection. So I would have thought, for example, that a "ring species" has not yet evinced speciation strictly speaking, only variation and the very beginnings of reproductive isolation. But let environmental changes wipe out populations in a few segments of the ring, and suddenly the new distinct species just pop out of the existing population. If that's right, then final stage of speciation is conservative.

Does this sound right? Am I misunderstand you, pz? Thanks for any info and clarification; I hope I haven't completely bollocksed things up here.

Itzpapalotl

I think natural selection is now viewed as the primary cause of speciation.

"it is increasingly evident that selection can also reduce assortative mating more simply, as a by-product of divergent adaptation. Adaptive divergence can cause reduced gene flow, which in turn allows further divergence, and so on, until speciation is achieved"

Adaptive divergence caused by natural selection is seen in many different sorts of organisms. It has also been shown that this divergence leads to reduced gene flow.

This change brings us back to an idea of evolutionary change primarily driven by natural selection that is much closer to the original ideas of Darwin.

In the same article this shift towards natural selection rather than drift as the primary cause of speciation is said to be like "We are apparently emerging from a 60-year-old blind alley, a veritable Dark Ages"

from: http://abacus.gene.ucl.ac.uk/jim/pap/malletjeb01.pdf

this and a few more papers can be found here:
http://abacus.gene.ucl.ac.uk/jim/Sp/Sp.html

pz

Well, sorta. You don't have to accumulate variation solely by drift -- disruptive selection for the extremes in a range of characters can also play a role. And selection isn't required for the final step, you could imagine some other isolating incident or catastrophe that separates the two groups. Even where you lose an intermediate population by selection, I find it hard to argue that the speciation event was caused by selection -- the surviving populations were not subject to it. That dilutes the meaning of selection to a remarkable degree.

I can think of one way selection can drive speciation. If there is disruptive selection for two varieties within a species, where a hybrid between the two varieties is selected against, the appearance of a novel feature that prevents matings between different varieties or promotes matings within a variety would be very advantageous. However, it is an open question about how often that kind of speciation occurs.

Clutch

Itzpapalotl, thanks.

pz, thanks to you too. Okay. But I'd thought you were saying something stronger: that selection is essentially conservative, while speciation (X) is not. I was trying to figure out what should go in for X there: essentially? often? typically? almost always? I don't see why (yet). Suppose you've an idealized sort of ring species of newt, such that only adjacent populations can interbreed. And suppose the portion of the ring inhabited by three adjacent populations sees an influx of a predator bird. All three populations serve as prey, but the centre one (say) is a colour that the birds see best, so it gets wiped out. Selection operated on all three populations, and effected a change that would make it look a lot more like we now had two (or more) distinct species -- especially if the ring gets "cut" elsewhere too. Does that dilute the notion of selection? Or is it just that things like geographic dispersal are thought to be much more usual ways of getting reproductive isolation?

Thanks again for your thoughts on this; I appreciate the tutorial!

Itzpapalotl

Although selection can be conservative in some cases it can drive differentiaition.

An interesting model has been developed to explain the pattern of data seen in the evolution of abalone egg/sperm binding proteins in which there is extreme variation (driven by positive darwinian selection) between very closely related species yet very little variation within species:

"In this process, intraspecific competition for fertilizations enlarges male gamete protein variation by means of evolutionary branching, which initiates sympatric speciation. In addition, avoidance of competition for fertilizations between the incipient species drives the rapid evolution of gamete recognition proteins. This mechanism can account for both strong stabilizing selection on gamete recognition proteins within species and rapid divergence between species. Moreover, it can explain the empirical finding that the rate of divergence of fertilization genes is not constant, but highest between closely related species"

In this situation there is both conservative stabilising selection within species and selection for change between closely related species occupying slightly different ecological niches. The model is explained at greater detail in the paper itself, i will post the longer version if people are interested.

G. Sander Van Doorn, Pieternella C. Luttikhuizen and Franz J. Weissing. Sexual selection at the protein level drives the extraordinary divergence of sex-related genes during sympatric speciation. Proc. R. Soc. Lond. B (2001) 268, 2155-2161