Although, I have to say that while Coyne is largely correct, he’s being a bit unfair. He’s addressing Olivia Judson’s recent article on “hopeful monsters”, a concept Coyne and the majority of the biological community reject. I reject it, too, but I think there are some legitimate issues that are associated with the idea that are also all too often and unfortunately discarded.
One point that Coyne handles well: there is a disconnect between the magnitude of genotypic changes and phenotypic effects — a single point mutation can cause amazing morphological changes. As Coyne points out, though, although this can happen, it’s not likely to be a major force in evolutionary change. Dramatic, single-step phenotypic effects are the kinds of things that geneticists select for, but they are also exactly the kinds of things that nature selects against. Evolution is much more likely to sidle up towards a major change by successive smaller steps, since those small changes are less likely to be accompanied by major deleterious side effects. Also, phenotypic outcomes of development should be robust to be advantageous, which typically means that there are many regulatory events cooperating to produce them — and they are therefore buffered by multiple controls.
But please, let’s not always dismiss Richard Goldschmidt when discussing “hopeful monsters”. It really wasn’t that awful an idea. Goldschmidt worked on stable variations in organisms: he studied sex differences (ever noticed that males and females have pretty much the same genes, but different phenotypes?) and metamorphosis (similarly, an organism builds two or more very different morphologies with exactly the same genome). He postulated that there could be specific, well-structured, stable nodes of patterns of gene expression — genes weren’t generally fluid, but tended to lock in to particular states. If he were writing today, he’d probably be bringing up the notion of attractors in chaos theory; the ideas are very similar. In that context, he was proposing a worthy concept that should have been taken more seriously than it was — Mayr’s hatchet job was particularly awful.
The “hopeful monster” concept was not shot down by the synthesis — it was ignored. I think it’s been dismantled by developmental biology, though; what we’ve learned is that the stable morphological types we see in a single species are not simply fortunate stable nodes in a nucleus that can be tuned in different ways, but that each are the product of many generations of slow sculpting by the processes of evolution, and that they are riddled with clumsy kluges that aren’t the outcome of some elegant global pattern switching mechanism, but of a long history of small tweaks.
Now also, Coyne is no fan of evo-devo, and he briefly voices the suggestion that the evolutionary developmental biologists are among the sources of this idea that saltational changes lead to sudden, drastic changes in body plans … but I’m just not seeing that. I am seeing work, for instance, that suggests that Hox duplications have been part of the process of producing additions to body plans, but it’s not a case of “poof, arthropods gain a metathorax in one change” — it’s been quite conventional. It’s more like “poof, arthropods gain an extra Hox gene, which initially adds redundancy and is later shaped by evolutionary processes that confer additional specializations on a segment,” quite ordinary stuff that shouldn’t be at all objectionable to Coyne.
It’s especially peculiar to pin the “hopeful monster” concept on evo-devo, when the one evo-devo expert he quotes, the biologist Sean Carroll, explicitly points out that evo-devo doesn’t support it.
Coyne is also going to be speaking at an evo-devo symposium I’ll be attending in April — I’m going to be very interested to hear what he has to say.