Behe 'replies' to TREE review

Well, Michael Behe has responded to my TREE review of The Edge of Evolution in a 3-part series posted on his blog. And the Discovery Institute has put up something lauding the reply.

The funny thing is how both Behe and the DI claim that I don’t address the substance of Behe’s book, all while ignoring the substance of my review, which addressed the substance of Behe’s book. All they can do is splutter that I am a biased reviewer who until recently worked at NCSE and therefore I must be wrong. (This, by the way, is how you know when you’ve got creationists where you want them.) I’ll make a list of my points and the non-replies below. It’s pretty shocking what they missed, considering my review was only 800 words or so, one of the shortest ones out there.

Here are the links to Behe’s three-part reply: Part 1 Part 2 Part 3

OK, let’s take stock. Each point below lists my scientific criticism and then the reply from Behe.

  1. Criticism: Behe claimed that the bacterial flagellum had 4 additional required regulatory proteins. But these proteins are not universally found in bacterial flagella. Behe reply: none.

  2. Criticism: Behe claimed that the eukaryotic flagellum/cilium was actually a case of “irreducible complexity squared” because a complex multiprotein system called intraflagellar transport (IFT) was required for cilium assembly. But this system is missing in malaria parasites (and Drosophila sperm, actually) which successfully build cilia anyway (they just build it in the cytoplasm which removes the need for the transport system). Behe reply: nada.

  3. Criticism: Behe’s claim that chlororquine resistance (CQR) occurs only by a rare lucky double mutation in 1 in 10^20 parasites is contradicted by research which shows that resistance alleles don’t always have the double mutant, and that in fact numerous resistance alleles with 1, 2, 3, 4, etc. mutations exist. Behe’s number is derived from an offhand guessestimate in the literature that considered only alleles that swept to high frequency in populations via natural selection, which only counts the “winning”, best resistance alleles, whereas what needs to be counted is the number of mutants which give selectable resistance, which are known to exist and to represent intermediate stages between the nonmutant gene and the “winning” resistance gene.

Behe’s Reply:

The most reliable data we have on the independent occurrence of resistance is that which surveys not just the mutations in the pfcrt resistance gene itself, but looks at surrounding DNA sequences for sequence heterogeneity. If drug resistance arose many times, easily and frequently, DNA surrounding the resistance gene would be expected to be as heterogeneous as other DNA regions in the genome. On the other hand, if resistance arose rarely with difficulty, heterogeneity would be suppressed around the resistance gene because of something called “hitchhiking” with the selected DNA. In several thorough studies, DNA heterogeneity was seen to be quite suppressed around pfcrt (6,7) (the chloroquine resistance gene), meaning that the resistance gene arose rarely and swept through a population.


  1. Volkman,S.K., et al. 2007. A genome-wide map of diversity in Plasmodium falciparum. _Nat. Genet._39:113-119.

  2. Wootton,J.C., et al. 2002. Genetic diversity and chloroquine selective sweeps in Plasmodium falciparum. Nature 418:320-323.

My surreply: all this data shows is that the advanced CQR alleles spread because of a selective sweep, which I stated in my review. All this means is that in a particular region, out of many alleles accumulating resistance mutations one at a time (and spreading relatively slowly with weak resistance), one of them got to the “best” combination first and subsequently beat out all the competition. This tells you nothing about how many times weak resistance alleles that are transitional to the “best” combination of resistance mutations originate. The data I cited, from India where this diversity of weaker resistance alleles still exists, shows that Behe’s “miraculous double mutant origin of CQR” model is wrong.

  1. Criticism: Because Behe’s 1 in 10^20 number is wrong, his estimate for the probability of evolving a protein-protein binding site (which was based on equating CQR to a binding site, since CQR mutants and binding sites both involve a combination of several important amino acids) is also wrong. Reply: none.

  2. Criticism: Behe’s decision to square the probability to get the probability of evolving 2 binding sites is wrong, both because this assumes that these 2 binding sites have to evolve at once (contradicted by the previous objections to the “irreducible complexity” argument) and because an experiment evolving binding sites managed to accidentally evolving two different binding sites to a target protein without squaring the population size. Reply: none.

  3. Criticism: Behe is wrong to think that evolution says a protein-protein binding site should evolve in any situation. Behe’s reply: evolution doesn’t predict anything about where binding sites should evolve. Surreply: Sure it does. But you have to at least have a system where protein-protein binding is an important factor, e.g. different receptors on two different host species for a virus. An evolutionary response to a small molecule like chloroquine doesn’t cut it.

  4. Criticism: Protein-protein binding sites actually aren’t hard to evolve, see antibodies, snake venom proteins, and especially the recent evolution of snake venom proteins within genera and species. Reply: none.

  5. Criticism: Behe is driven not by science but by his mistaken but obsessive-held view that evolution is “random” in a metaphysical sense meaning “purposeless.” Behe’s reply: “Wasn’t it Darwin himself, we are constantly assured, who based his theory on ‘random’ variation?” Surreply: The word “random” does not appear in the Origin of Species. Search for yourself. Darwin based his theory on natural selection, which is nonrandom. The source of variation was unknown. But thanks to Behe for proving my point – for him, evolution = randomness = purposelessness = no God or meaning of life, this is precisely why he keeps playing the same old game: set up an all-at-once chance event as if it were a good model for a gradual evolutionary process operating under the guidance of natural selection, then declare the all-at-once chance event wildly improbable, then infer ID and thereby rescue the world from the purposelessness which is somehow produced by mere description of a physical process.

By my count, Behe only bothered to give it a try on 3/8 points, only gave it a significant shot on one, and was easily shot down on all three. If anyone wonders why Behe has repeatedly failed to convince when he has informed opposition – for example, in the scientific community, or in court – now you have your answer. He gives excuses rather than answers, and when problems are pointed out, he mostly just hopes that his fans will remain ignorant of them.

[Added in edit: Filled in some stuff on the randomness point which needed a bit more. May discuss further in the comments if warranted.]