It’s time for a book review: Jerry Coyne’s “Why evolution is true”. My review will be in three parts, the first of which will give me the chance to display my ignorance of biology. Such ignorance should not be underestimated – many of the questions I ask below are not rhetorical. I just don’t know enough to answer them. Regardless, I tried to read the book critically.
Firstly, the book was an enjoyable read and I learned a lot. Coyne writes quite well – one reviewer complained that he doesn’t have Dawkins’ flair, but who does? He has the occasional inspired metaphor or turn of phrase …
“The story of life on Earth is written in the rocks. True, this is a story book torn and twisted, with remnants of pages scattered about, but it is there, and significant portions of it are still legible.” (pg 21)
Explanations are clear and logically laid out, and his biological and historical examples (especially the story of Robert Scott) are mostly well selected. A good example of Coyne’s clarity is in explaining to the oft-heard claim that chimps and humans are 98.5% alike:
“… recent work shows that our genetic resemblance to our evolutionary cousins [chimps] is not quite as close as we thought … to consider an analogy, if you change only one percent of the letters of this page, you will alter far more than 1 percent of the sentences … more than 80% of all the proteins shared by the two species differ by at least one amino acid.” (pg 230)
I have a few gripes, however.
Gripe 1: On page 67, he shows a photo of a human baby with a small, tail-like protrusion, while two pages later he informs us:
“… some species of whales retail vestigial pelvises and rear leg bones, but one whale in 500 is actually born with a rear leg that protrudes outside the body wall. These limbs show all degrees of refinement, with many of them clearly containing the major leg bones of terrestrial mammals – the femur, tibia, and fibula. Some even have feet and toes!” (pg 69)
What?! There are whales with legs, feet and toes?! Why not show us a photo of that?! Why hasn’t a suitable skeleton been mounted in a large truck and parked outside the creation museum? That’s the stuff I want to see.
Gripe 2: Coyne reaches for an illustration from astronomy:
“The way we have discovered how species arise resembles the way astronomers discovered how stars “evolve” over time … [We find] snapshots of the process at different evolutionary stages and [put] these snapshots together into a conceptual movie. For stars, astronomers saw dispersed clouds of matter (“star nurseries”) in galaxies. Elsewhere they saw these clouds condensing into protostars. And in other places they saw protostars becoming full stars, condensing further and then generating light as their core temperature became high enough to fuse hydrogen atoms into helium … By assembling all these stages into a logical sequence, based on what we know of their physical and chemical structure and behaviour, we’ve been able to piece together how stars form, persist, and die.” (pg 199)
This example almost works. Coyne paints a portrait of a inductive reasoning – make lots of observations and then try to see a pattern. But the physical sciences rarely work like this. We tend towards hypothetico-deductivism – start with general physical principles (e.g. the equivalence principle), deduce equations, apply them to a given scenario, make predictions, and then go make observations to see if we are correct. It isn’t true that astronomers saw protostars becoming stars as they collapsed and their cores become hot enough to sustain nuclear fusion. Stellar nurseries are obscured by dust, we can’t resolve gas motion on such small scales, and we can’t see into the centre of stars. What we see are emission lines, which quantum mechanics tells us come from certain elements and molecules. Astrophysical models then tell us that these are the signatures of cold molecular clouds, the precursors to young, hot stars.
I know what it means when a physicist makes a prediction with a model. It means a set of mathematical equations and quantities. But the predictions of evolution seem to be largely qualitative, which beings me to my third example.
Gripe 3: Coyne tells us repeatedly that evolution makes testable predictions (or retrodictions – either is fine):
“No theory of special creation, or any theory other than evolution, can explain these patterns [in the fossil record].” (pg 30)
To illustrate this point, Coyne tells us that columnar cores can trace a single fossil species through time. The figure below shows a single-celled protozoan through 8 million years, and in particular plots the number of chambers in the final whorl of the shell.
Now, to my astrophysicist eyes, this plot looks like a random walk. The canonical example is of a random walk is a drunk man leaving the pub, who is so disoriented that the direction of each step is in a random direction. All sorts of processes exhibit random walks – part of my research involves modelling radiative transfer by tracing photons as they random-walk their way out of protogalaxies.
So what exactly is the prediction of evolution? Coyne tells us that: “…evolutionary theory does not state that all species must evolve!” (pg 34). Neither does it predict that change will “proceed smoothly, or at an even pace.” (pg 32). It doesn’t tell us the direction of the change i.e. more or less chambers, because there is no independent way of knowing which way selection pressures were pushing. Thus, given the number of chambers 8 million years ago, evolution predicts that the number might increase or decrease, at any rate, smoothly or not, or not at all. And having observed what looks like a random walk – a rather generic, vanilla-flavoured statistical process – we are told that Darwinian evolution is the only possible explanation. Your first response might be to ask what alternative hypotheses there are. I’ll have none of that – frankly, I’d like a little more from a scientific theory than: “at least it’s better than intelligent design”.
I’m just not sure what this example is suppose to show – most of Coyne’s other examples are right on the money. Does it show that species change through time? No one denies that – not even creationists (more on that soon). It doesn’t show a particularly dramatic change – 8 million years for a protozoa that reproduces on a timescale of 1 day is equivalent to about 50 billion years of human evolution. In that time, evolution has taken a feature that already existed and reduced the number by a third. (Perhaps there is a statistical bias here – any lineage that underwent more significant evolutionary change would no longer be considered the same species, and thus wouldn’t be included in the plot). Does the plot show that evolution predicts monotonic change? Only if the selection pressure is also monotonic, and we have no independent information on that. If the line wandered left and then wandered back to the right, we wouldn’t conclude that this contradicted the theory of evolution. The plot just shows the data – there is no “prediction” line to compare the data to, as you would see in a physics paper (like this: 400 sigma!). The plot only seems to show that we can study fossils through time with columnar cores, which demonstrates the methods of paleontology rather than the prediction of Darwinian evolution.
I’ll finish with a few more questions and confusions. I’ll repeat my warning from the beginning – these aren’t rhetorical. This is my ignorance on parade. Feel free to provide answers in the comments.
1. “[Darwin] was not the first to think of evolution … But [he] was the first to use data from nature to convince people that evolution was true”. (pg 3) Really? What about 1844’s “Vestiges of the Natural History of Creation”?
2. “This principle of superposition was first devised in the seventeenth century by the Danish polymath Nacolaud Steno, who became an archbishop and was beatified by Pope John Paul II in 1988 – surely the only case of a future saint making an important scientific contribution.” (pg 24) Firstly, this is pointless needling of the very group that this book should be aimed at – Christian creationists. And secondly, Coyne gives no references, and I have my doubts as to his expertise in the scholarship of Christian saints through the ages. Three minutes on Wikipedia revealed that Leonhard Euler is included on the Lutheran calendar of saints. (Laplace: “Read Euler. Read Euler. He is the master of us all”.) Still, I’m not a historian either.
3. “… accidental colonists were able to form many species because oceanic islands offer lots of empty habitats that lack competitors and predators.” (pg 116) But on page 201 we’re told that oceanic islands don’t facilitate speciation because there are too few chances for geographic isolation to produce non-interbreeding populations. I thought that evolution would be slow in the absence of predators and competitors, because the weak would be able to survive and keep their genes in the genepool.
4. I’ve often wondered evolution’s ability to mould multiple traits simultaneously. Coyne gives the example of the selection of fur colour on mice (pg 126), which is basically the same moral as the peppered moth. Here we have a scenario where (apparently) the most important factor deciding whether an organism will reproduce or not is fur colour. Does this mean that other, potentially useful mutations are ignored? Suppose that, of 10 progeny, one has a mutation that improves its eyelids, so that it can clean its eyes faster in dusty conditions. All things being equal, it has a survival advantage. But all things aren’t equal: the animal faces many predators, and only its browner sibling survived and reproduced.
One might have thought that, if a certain beneficial mutation can be expected to occur once in 1000 generations, then on average we will have to wait 1000 generations for the mutation to enter the genepool. But there appears is an extra “wastage factor” – the fraction of times that a beneficial mutation arises in an individual that, for other reasons, fails to reproduce.
In fact, given a strong enough selection pressure for one trait (or set of traits), we could end up going backwards. If harmful mutations are more common that beneficial ones (I seem to remember Richard Dawkins saying this in one of his books), then it is more likely that the “brownest mouse” will pick up a mutation that adversely affects its other traits – e.g. a compromised immune system – but this disadvantage is swamped by the brown-fur-advantage.
To make a mouse, evolution has a lot of work to do. There are a lot of traits/genes to be selected and others to be eliminated. Does each feature have to wait for the appropriate selection pressure before any real evolutionary progress is made? Are potential advances in “secondary features” often lost? Can these secondary features get a look in at all? Does a dominant selection pressure along one direction of genome space retard progress in other directions? Are these questions quantifiable? Just how good is the Darwinian mechanism?
5. What exactly what is an evolutionary “just-so story” and what separates it from informed speculation / prediction? In the final chapter, Coyne says:
“There is an increasing (and disturbing) tendency of psychologists, biologists, and philosophers to Darwinise every aspect of human behavior, turning its study into a scientific parlour game. But imaginative reconstructions of how things might have evolved are not science; they are stories. Stephen Jay Gould satirized them as “Just-so Stories,” after Kipling”. (pg 248)
And yet, he seems to be perfectly willing to imaginatively reconstruct how flight evolved:
“… we can make some guesses about how natural selection fashioned modern birds … then followed a feathery covering, probably for insulation …” (pg 49)
I’d have thought that the problem with “just-so stories” is that, in spite of their convenience, they are known to be false. What am I missing here? Is it that a just-so-story currently lacks any evidential support? Or is it untestable? Or is it the existence of a multitude of competing just-so-stories, and no way to decide between them? Or, slightly more cynically, is it OK to speculate about birds but not about human behavior? Stay tuned …