Convergent evolution is a thing of beauty. Whether it is wings in bats, birds and pterosaurs, or eyes in a range of organisms, evolution often seems to come up with functionally similar solutions to life’s challenges. So how predictable is evolution? This is a question that has fascinated generations of biologists and, as Losos quickly makes clear, two famous figures loom large.

Stephen Jay Gould believed evolution is unpredictable* and put forth his viewpoint in the 1990 book Wonderful Life: The Burgess Shale and the Nature of History. He coined the now-famous thought experiment of “replaying the tape of life”. Firmly rooted in a time when video came on VHS and audio on cassettes, he proposed that if you could rewind time and replay the proverbial tape of life, very different life forms would evolve. The other camp is spearheaded by names such as the British palaeontologist Simon Conway Morris, who argues that evolution is deterministic (in his 1988 book The Crucible of Creation: The Burgess Shale and the Rise of the Animals and more recently in his 2015 book The Runes of Evolution: How the Universe Became Self-Aware). There are only so many ways to make a living, and for each challenge the environment throws at living beings, there is only one or a limited number of optimal solutions, leading natural selection to repeatedly produce the same evolutionary outcomes.

For a long time, the debate basically consisted of people compiling lists of examples. Morris and others have shown numerous examples of convergent evolution, not just of traits, but even complete groups of animals evolving in similar fashion time and again (Losos’s work on comparable Anolis lizard communities popping up on different Carribean islands being one example). The other camp points out this is cherry-picking of biological examples after the fact and have drawn attention to all the biological oddities and one-offs that have evolved. New Zealand’s extraordinary fauna is a prime example of this.

To really get a grip on this matter we need to do experiments. Once biologists realised that evolution can occur really quickly, in the span of a few generations, studies were undertaken to examine natural selection in wild populations. The book is pleasantly up to date, discussing books published only months prior to this one, such as 40 Years of Evolution: Darwin’s Finches on Daphne Major Island or How to Tame a Fox (and Build a Dog): Visionary Scientists and a Siberian Tale of Jump-Started Evolution (reviewed here) and well-known work on guppies and sticklebacks. Losos is on form here, writing in an accessible style, leaving out extraneous details, and telling the stories of the researchers engaged in these studies with flair. It’s an incredibly fun section of the book and the glee with which it was written jumps off the page. The general picture after decades of field evolution experiments is that, yes, evolution is repeatable. If multiple populations experience the same environment, they tend to evolve in similar ways. However, critics argue that if you start with related populations with similar genetic material for selection to act on, it’s quite logical that these populations will evolve along similar lines, subject as they are to the same genetic limitations and options. And by design, these studies ignore the statistical outliers. Of course, populations not showing convergent evolution could also be due to environmental variations for which you cannot control. Doing studies in the wild is a messy affair.

“Losos is on form here […] the glee with which it was written jumps off the page”

Luckily, a whole cadre of scientists has taken up the challenge of studying evolution in microbes under laboratory conditions, where environmental variables can be precisely controlled. Even this section of the book, which is not Losos’s stomping ground, is very well written. Working with microbes has certain advantages for a biologist. Their quick generation times means you can observe changes over thousands of generations in the course of a few years. Microbes can be frozen, which has been used to great effect in long-term experiments to create snapshots of evolving populations, and allowing researchers to thaw these earlier populations out at a later time, literally restarting the tape as Gould suggested in his thought experiment. Here, too, studies show that evolution is repeatable, with identical populations showing identical evolutionary responses.

And yet, these same studies have also shown that this is not always the case, as a rare change in bacterial metabolism in one of these studies revealed. In this case, it took tens of thousands of bacterial generations, and the change only happened in one of twelve replicates. Further studies showed that a series of unlikely mutations had to conspire in just a certain order for this rare event to happen. But, as the saying goes, given enough time, even the statistically improbable will take place eventually. Maybe we haven’t looked long enough in our experiments?

Also, different mutations can produce functionally similar end results. These details can matter later on. Gould talked about contingency in this context. If you rewind the tape of life and make small changes at the beginning, you might get radically different outcomes. Though this hasn’t been explicitly addressed experimentally, some researchers have started with genetically dissimilar microbial populations to see if they evolve in the same way when exposed to similar environmental conditions. And, lo and behold, under these circumstances, populations don’t necessarily show convergent evolution.

So, is evolution predictable? With Improbable Destinies, Losos gives a grand tour of the arguments both camps have put forth, and the scientific research that has been done to answer this question. Along the way you get to think along with him as he ponders the different sides and lines up the available evidence. Losos’s conclusion seems perfectly reasonable: part of it depends over what time scale you look, and how comparable your starting points are. Both Gould and Morris have a point. Yes, evolution is subject to constraints, and often there is only a single optimum solution to a problem. With close relatives sharing many genetic similarities, it should not come as a surprise that they will converge on the same evolutionary solutions to life’s problems. But it doesn’t have to be like this. Starting from different points, with different genetic make-ups and evolutionary histories, evolution can also cause organisms to completely go their own way.

Improbable Destinies is a splendid piece of science writing that I can highly recommend. If the question this book poses sounds even remotely interesting, you should do yourself the pleasure of reading it.

Shortly after publication of this book, I had the pleasure of interviewing Losos by email to talk more about convergent evolution with him – you can read that here.

*Three years after writing this, my review of Contingency and Convergence taught me that this is a common misinterpretation of what Gould exactly had in mind.

Disclosure: The publisher provided a review copy of this book. The opinion expressed here is my own, however.

Improbable Destinies paperback, hardback, ebook or audiobook

Other recommended books mentioned in this review:

__________________________________________________________________