Convergence falsifies evolution, according to Cornelius Hunter

Xerus princeps

Xerus princeps, the mountain ground squirrel.

Before I saw the light and switched to studying Volvox, I studied squirrels. With apologies to Iris Vander Pluym, squirrels are cool. If you grew up in the squirrel-deprived eastern U.S., you might not realize that there are over a hundred species. Chipmunks are squirrels. Marmots are fat squirrels. Prairie dogs are adorable squirrels.

Most of my squirrel work, and some of my Volvox work, has focused on understanding the evolutionary relationships among species. This fits in the subfield of evolutionary biology known as phylogenetics. Phylogenetics results are often visualized as trees and published in journals like CladisticsMolecular Phylogenetics and Evolution, and Systematic Biology. Phylogenetics is a vast subfield, with a huge number of papers devoted to developing methods that are theoretically and empirically sound.

Cornelius Hunter understands none of this. In a recent post over at Evolution News and Science Today (which used to be Evolution News and Views…when did that change?), he argues against the whole idea of common descent, the very foundation of phylogenetics. Dr. Hunter argues that convergence, similarities among distantly related species, falsifies evolution. The nature of his arguments shows pretty conclusively that he doesn’t understand the basic logic of what he’s criticizing.

He uses as an example the convergence, in terms of traits related to gliding, between flying squirrels (a placental mammal) and sugar gliders (a marsupial also known as flying phalangers). These are, of course, about as distantly related as two mammals can be, but they both have flaps of skin called patagia that stretch between their front and back legs, allowing them to glide.

The marsupial flying phalanger and placental flying squirrel, for example, have distinctive similarities, including their coats that extend from the wrist to the ankle giving them the ability to glide long distances. But evolutionists must believe that these distinctive similarities evolved separately and independently because one is a marsupial and the other is a placental, and those two groups must have divided much earlier in evolutionary history.

So far, so good. Independent evolution of the gliding flaps is exactly what we ‘evolutionists’ believe.

Simply put, evolution’s random mutations must have duplicated dozens of designs in these two groups.

It is kind of like lightning striking twice, but for evolutionists — who already have accepted the idea that squirrels, and all other species for that matter, arose by chance mutations — it’s not difficult to believe.

Oooh, he went off the rails there. This is such a common tactic of the cdesign proponentsists that I should devote a whole post to it, but in the interest of time, here’s the short version. No evolutionary biologist thinks that gliding evolved randomly (or that squirrels ‘arose by chance mutations’), so this is a straw man. Yes, mutations are random, but selection is not. Mutation is the ultimate source of genetic variation, but selection determines which variants increase in frequency and which decrease. Thus the nonrandom process of selection, not the random process of mutation, drives the evolution of similar traits in ecologically similar species. This has been pointed out to members of the Discovery Institute so many times that the failure to understand it has to be willful. Dr. Hunter is misrepresenting sound evolutionary logic to make it sound silly. If he had good arguments against the real version, he wouldn’t need to misrepresent it.

What is often not understood, however, by evolutionists or their critics, is that convergence poses a completely different theoretical problem. Simply put, a fundamental evidence and motivation for evolution is the pattern of similarities and differences between the different species. According to this theory, the species fall into an evolutionary pattern with great precision. Species on the same branch in the evolutionary tree of life share a close relationship via common descent. Therefore, they share similarities with each other much more consistently than with species on other branches.

Again, not too bad. The wording is odd: I’m not sure what it means to ‘fall into an evolutionary pattern with great precision’, and being ‘on the same branch of the evolutionary tree’ means exactly the same thing as ‘shar[ing] a close relationship via common descent’. But the last bit is right: closely related species, those that share a recent common ancestor, ‘share similarities with each other much more consistently’ than with more distant relatives.

This is a very specific pattern, and it can be used to predict differences and similarities between species given a knowledge of where they are in the evolutionary tree.

Convergence violates this pattern. Convergence reveals striking similarities across different branches.

There is more in this vein (including a lovely quote mine), but the gist of all of it is the same: evolution predicts that distantly related species should be more different from each other than closely related species, and convergence doesn’t fit this expectation. But here’s the thing: organisms have lots of traits. Distantly related species ARE more different from each other than closely related species in most of those traits. This is why we don’t infer evolutionary relationships based on a single trait. If we inferred relationships based on powered flight, for example, we’d conclude that birds, bats, pterodactyls, and some insects are all very closely related.

Let’s use Dr. Hunter’s own example: sugar gliders and flying squirrels. Yes, they have a set of traits in common, namely those related to gliding. Natural selection often finds similar solutions to similar problems. But sugar gliders and flying squirrels have many, many more traits that fit the expectations of common descent. Both are mammals, and they share the traits shared by mammals: both lactate, have fur, three middle-ear bones, a neocortex, and are warm-blooded. However, one is a marsupial and one a placental mammal, and as expected they have significant differences in their reproductive systems, structure of the skull and brain, dental formulas, length of gestation, and other traits. Differences, that is, that they share with other marsupials and placentals, respectively. DNA sequence data, both nuclear and mitochondrial, are also in line with the inference that the ancestors of sugar gliders and those of flying squirrels diverged soon after the origin of mammals.

Dr. Hunter was right when he said that closely related species “…share similarities with each other much more consistently than with species on other branches.” And that is true even for his own example. Sugar gliders share many more traits with other marsupials, and flying squirrels with other placental mammals, than they do with each other. There is no expectation that distantly related species won’t share any derived traits (traits not present in their common ancestor, that is). That’s why we have a term for shared derived traits: synapomorphy*.

Synapomorphies Homoplasies* are to be expected when distantly related species occupy similar habitats and are subject to similar selective pressures, as is the case for sugar gliders and flying squirrels. They don’t pose any particular challenge to evolution. It isn’t especially surprising that small, forest-dwelling animals would evolve a fast way to get from one tree to another without coming down to the ground. It is especially surprising that Cornelius Hunter, who has a Ph.D. in Biophysics and Computational Biology, doesn’t understand this.

*EDIT: well, that’s embarrassing. It’s true that a synapomorphy is a shared derived trait (as in the second to last paragraph), but it is a derived trait that is shared by a clade, in other words not due to convergence. Derived traits that are shared due to convergence are called homoplasies. Clearly no one on my master’s committee reads this blog, because they would have smacked me down, hard, for such a mistake!


  1. says

    evolution predicts that distantly related species should be more different from each other than closely related species, and convergence doesn’t fit this expectation.

    Well, yeah, if the environments are different enough to affect survival and reproduction differently. If the environments are more or less the same then wouldn’t we expect convergence? (I’m not up on the concept but it seems straightforward)

    Cornelius Hunter, who has a Ph.D. in Biophysics and Computational Biology, doesn’t understand this.

    The great thing about science is we’ve learned enough to be able to sub-specialize and even sub-fields of physics or biology are incredibly rich in information. Which means that you can also become a bit of a mental barnacle, if you just pay attention to what’s in front of you and nothing else. I’m pretty sure that he’d have been able to figure all this out with a bit of thinking and a couple scrolls of google.


Leave a Reply