The biggest problem in selling evolution: how fast can selection create complexity?

April 26, 2009 • 7:36 am

I received a wonderful four-page typewritten letter from a gentleman in Oklahoma detailing, chapter by chapter, his reaction to my book.  He describes himself as “a nearly ideal representative of people who have doubts about evolution but who could be convinced. I’m a scientist and a farmer. I am also an evangelical Christian. My scientific training is in physics and meteorology. . and I’ve been a close observer of nature all my life. . I have taught science and mathematics classes at two colleges and a high school in Lousiana and Oklahoma.”

I won’t bore you with the stuff he liked (e.g., his favorite chapter was the one on biogeography).  What I want to discuss was what he didn’t like:  in particular, the idea that natural selection can build complexity over observed stretches of time.  Here is what he says about the chapter on natural selection (his underlines):

“Chapter 5 is where I have the most problems in accepting what you say. Let me ask a few questions. You have looked at living things in great detail, far more than have I. You have seen how amazingly complex living things are.  That complexity exists at the tiniest levels (DNA, proteins, cells) to the levels of complete organisms.  Surely  you must think to yourself, “Is it really possible that all this complexity happened by purely materialistic and naturalistic processes?”  Or, more specifically, do you ever think “Is natural selection enough to account for such complexity?”  Of course, mutations occur and, of course, natural selection occurs everywhere in the biosphere andf all the time.  But, is it enough? I just don’t see how it could be.  It is a ruthless conservative process.  But, is it a truly creative process? I found your description of the evolution of bird flight (p. 39) to be inadequate. You say “It’s not hard to envision. . .”  Yes it is!  How did a complex thing like a feather ever develop by mutations and natural selection?”

If you try to convey evolutionary biology to nonbiologists, this plaint will sound familiar.  Sure we have evidence for evolution, and we see evidence in the fossil record for adaptation (for example, in tetrapods adapting to a terrestrial environment), but how can we be sure that adaptive changes are caused by natural selection?  This inability to visualize how the slow, step-by-step accretion of adaptations could create the amazing diversity of life on Earth is why natural selection was not firmly established in the scientific community until the 1930s or so, even though evolution, gradualism, and common ancestry were part of scientific dogma in the 1870s.  It is not so much that the process of natural selection is hard to understand, or that it could be responsible for “simple” adaptations like antibiotic resistance in bacteria.  Rather, it is whether there has been enough time for such a process to create all the complex adaptations we see around us.  (I’m not talking about the ID objection about the sufficiency of selection to build adaptations without going through a maladaptive phase  —  a common but misguided creationist argument made infamous by Michael Behe).   And, of course, the gentleman also raises problems about whether natural selection can produce novelty.

Well, the complexity issue is easily dealt with.  Both Ken Miller and Richard Dawkins have written extensively and convincingly about whether selection can produce novel traits.  The answer of course is yes, as most biologists can see intuitively.  And I think it’s easy to convince people of this with the compelling examples that Miller and Dawkins have used in Only a Theory and Climbing Mount Improbable, among other places.   But I think the heart of the objection is not the ability of selection to build novel, complex traits, but the time available for it to have done so.

As evolutionists, how can we address this “time” concern?  We usually do it in two ways.  First, we say that we can see selection building adaptations over time.  The evolution of whales from terrestrial artiodactyls, for example, took about 10 million years, and we can see it happen in the fossils.  But this begs the question, since the questioners are asking whether this kind of change can be due to natural selection, which is said to be ineffective over such periods.  And of course we can only invoke the idea that we know of no process other than selection that could create such adaptive change.  That is satisfying to scientists, but perhaps not so convincing to people like the gentleman who wrote me.

Second — and most often — we evoke the huge stretches of time over which selection has had to work: millions and billions of years.  Indeed, such spans of time are not easily grasped by even the minds of evolutionary biologists.  We are accustomed (and perhaps evolved) to thinking of times on the order of decades or centuries.  So, we say, given the long eons since life began evolving, there has been plenty of time for selection to do its job.

I think a good way to meet this criticism is through mathematical modelling.  We simply make a model of the evolution of a complex trait (or better yet, several of them), basing it on reasonable estimates of selection pressures, mutation rates, etc.  Then we see how long it will take the model or the computer to construct the adaptation.  Then we extrapolate to how many such adaptations it would take to evolve a new “type” of creature, say a bird from a theropod dinosaur.  If our theory is right, we should be able to do this, and find that selection can indeed create adapations in reasonable stretches of time.

As far as I know, there has been only one attempt to do this: Nilsson and Pelger’s 1994 paper on the evolution of a complex camera eye from a flat, pigmented, light-sensitive eyespot. ( Nilsson, D.-E., and S. Pelger. 1994. A pessimistic estimate of the time required for an eye to evolve. Proc. Roy. Soc. Lond. B 256:53-58.).  This was a mathematical model that I describe in chapter 5 of WEIT, and which Richard Dawkins summarizes here. (See the animation of this process on Mark Ridley’s Evolution website here.) Nilsson and Pelger showed that, under conservative assumptions, the eyespot would become a complex camera eye in a little less than 400,000 years.  Since the earliest animals with eyes go back to about 550 million years ago, this span is enough for eyes to have evolved more than 1500 times in succession!

Apparently this model was not sufficient to convince my interlocutor, or he didn’t realize its implications.  What I would like to see, and what I think would be a great boon to furthering acceptance of evolution, is more models of the Nilsson and Pelger type.  When we tell people that there’s been sufficient time for everything to have evolved by natural selection, we need more hard models to back us up.  Granted, making such models isn’t going to tremendously advance the career of an evolutionary biologist, but I still think they’re important.


The stages of eye evolution in Nilsson and Pelger’s paper, with n equal to the number of generations that elapsed in their model.