Book review: Experimental Evolution and the Nature of Biodiversity by Rees Kassen

Here’s something new for this blog: a timely book review. Rees Kassen‘s Experimental Evolution and the Nature of Biodiversity has just been published. Here’s my review.

Full disclosure: Rees is a friend, I spent a semester visiting his lab back in 2010. He was kind enough to send me a free copy of his book. I tried not to let it affect my review one way or the other, and I hope I managed to do that.

The book reviews what we’ve learned about evolutionary adaptation and diversification from experimental evolution of microbes. Connecting adaptation and diversification is an old problem, one Darwin himself famously struggled with. Rees is one of the world leaders in experimental microbial evolution, so his lab’s own work figures prominently in the book (without dominating it; the book is very far from just being a compilation of Rees’ own work). The chapters cover:

• an introduction to experimental evolution (starting with a really cool example dating back to shortly after Darwin’s death)
• the genetics of adaptation to a single environment
• divergent selection
• selection in spatially and temporally variable environments
• genomics of adaptation
• phenotypic disparity
• rate and extent of diversification
• adaptive radiation
• genetics and genomics of diversification
• the nature of biodiversity

Most of the chapters start with a stage-setting vignette to introduce and motivate interest in the topic. For instance, the chapter on adaptation to a single environment starts with the story of how Londoners sheltering in the Underground tunnels in WW II were plagued by mosquitoes that may have adapted to underground habitats, and to feeding on humans, during the 80 years the Underground had existed at that time. I really enjoyed the vignettes and found most of them effective. It’s too bad the approach kind of runs out of steam near the end of the book (there’s no vignette for the chapter on adaptive radiation, and the vignette for the following chapter was interesting but didn’t seem to me to be closely tied to the chapter topic).

Bottom line: I liked the book. Not surprisingly, perhaps, because I’m very much on Rees’ wavelength. He believes that our hypotheses should come from mathematical theory whenever possible. He thinks it’s really important to complement observational and comparative data with direct experimental tests. He believes that good data from a model system are better than no good data at all (and as his book shows, that is often a real choice we’re faced with in science). He believes that one can make useful comparisons between microcosms and other systems by keeping in mind the ways in which microcosms are different than other systems (e.g., large population sizes, adaptation based on new mutations rather than standing variation). He believes that microbial microcosms are simple enough to be tractable, but yet complex enough to be capable of surprising us, and so capable of inspiring new hypotheses as well as testing existing hypotheses. I agree on all counts.

Indeed, I wish I’d written the book myself. And I mean that almost literally, because this is kind of the evolutionary equivalent of an ecology book I proposed to write a few years ago, pulling together everything ecologists have learned from microcosm experiments. But Rees’ book is better than mine would have been, I think. One reason for that is that Rees’ book is about a fairly well-developed and unified body of theory, that’s been directly tested in a sufficient number of sufficiently-similar experiments that one can do meta-analyses on the results. I don’t know that you could say the same for my proposed book.

Those meta-analyses are the core contribution of Rees’ book, to my mind. There are about 10 meta-analyses in the book, depending on precisely how you count, many of which could’ve been standalone papers. I can only imagine how much frickin’ work it must’ve been to compile the data! If you want to know how often fitness trade-offs evolve under divergent selection (invariably), whether adaptation to a fitness peak typically involves fixation of few or many mutations (few), what the typical rate of substitution is during an adaptive walk, and much more, this book has the numbers.

The other bit of the book that really stood out for me was the extension of Fisher’s geometric model to multiple phenotypic optima, thereby converting the model into a tool for studying the consequences of divergent selection (e.g., the contrasting selection pressures imposed by two different habitats). This is a lovely idea, credited to unpublished work by G. Martin. Simple, elegant, and powerful–I can’t wait to see it further developed.

The book is a satisfying story of an ongoing, successful research program. On topics on which we have well-developed theory, microbial evolution experiments usually behave more or less as theory predicts, at least on average (there’s often a lot of variation around the average, which is something I wish Rees had discussed a bit more). The book also points out the most interesting and needed directions for future research, as a good book of this sort should do. It’ll be a gold mine for grad students looking to get up to speed on the literature and on the lookout for project ideas.

There are some weak points, though they’re far outnumbered by the strong points. There’s perhaps a bit too much repetition, with some of the same concepts and examples reintroduced in two or three places. But then, a reader who was completely new to this material might appreciate the repetition. The chapters on diversification (the second half of the book) in general weren’t quite as strong as the chapters on adaptation, probably because Rees had less material to review. So there’s less meta-analysis and more qualitative discussion of isolated examples. And I had several quibbles with the chapter on spatial and temporal variation in selection. Rees’ explanation of why geometric rather than arithmetic mean absolute fitness is of interest in temporally-varying environments isn’t as precise as I’d have liked. I wish this chapter had been clearer up front (rather than partway through) about the difference between selection that merely varies in direction in space or time, and selection that can actually stably maintain genetic variation. But I admit that’s a personal hangup of mine. I also would’ve liked to see this chapter compare spatially- or temporally-varying selection to selection in non-varying environments with the same average conditions as the varying environments, since otherwise you’re confounding the effects of environmental variance with effects of average environmental conditions. Apparently most theory or experimentation on this topic doesn’t make that comparison (unless I misunderstood something?) But that’s another personal hangup of mine. Finally, throughout the book I found myself wanting more comparison of the results of microbial evolution experiments with results from other systems. Rees’ comparative remarks often are quite brief. A few more comparative remarks might have helped to “sell” skeptical readers on the value of the experimental evolution approach. But then again, I suspect that for many topics there’s just no comparable data from other systems, and so probably there’s not much that could be said by way of comparison.

The writing is solid–simple, straightforward, and clear. The writing in several of the vignettes is really nice. In his understated way, Rees is a fine storyteller. I found myself wishing (greedily, I know) that Rees had adopted the voice of the vignettes throughout the book. The book is not at all technical, and so is quite accessible. There are hardly any equations and jargon is kept to a minimum (including genomics and sequencing jargon, thank god–I freely admit I find that stuff impenetrable). The figures are greyscale, which is sometimes ok, but sometimes you have to squint to distinguish different shades of grey. The cover art is cool, it recalls the multiple optima extension of Fisher’s geometric model that’s one of the highlights of the book.

Anyone who does experimental evolution needs a copy of this book. Who else will want to read it? In particular, why should an ecologist want to read it? I can think of a few reasons:

• You’re broad-minded and you want to know something about how evolutionary biologists who are interested in ecology think about biodiversity. That caveat “interested in ecology” is important. Rees doesn’t take selection coefficients and population sizes as god-given. Rather, much of the book is about how the ecology of the system (and ongoing evolution) sets those parameters, thereby affecting the future course of evolution. For instance, he talks about how predators change the selection pressures to which prey are subject, while also reducing prey population sizes and so reducing the expected supply rate of beneficial mutations. And if you’re the sort of ecologist who sees genetics and genomics as far removed from anything you could possibly be interested in, well, I think you might be pleasantly surprised if you read this book.
• You just want to understand evolution better. In particular, I think many ecologists would be surprised by, and learn a lot from, Rees’ emphasis on fitness and its evolution. For instance the idea that fitness trade-offs (i.e. high relative fitness in one environment is associated with low relative fitness in a different environment) are a result of natural selection rather than a constraint on natural selection (see this old post). And how fitness trade-offs can emerge even if the same traits are favored in all environments (just to differing degrees in different environments).
• You’re into eco-evolutionary dynamics. Rees doesn’t use that term, and doesn’t talk a lot about coevolution (though he does talk about it a bit), but if you’re serious about the “evolution” bit of “eco-evolutionary”, you’ll want this book.
• You buy the suggestion of Mark Vellend and others that community ecology can learn a lot from (asexual) population and evolutionary genetics. Right now, community ecologists who believe this are focusing on neutral models and elaborations thereof. I think the first community ecologist who starts translating other sorts of evolutionary genetic models into community ecology terms could (deservedly) make a splash. For instance, I bet the multiple-optima version of Fisher’s geometric model could be used to make novel predictions about community assembly and structure in spatially heterogeneous environments.

The book is softcover and it’s not expensive, so if you’re interested you should definitely buy a copy.