Not only termites should be alarmed

Lynn Margulis persuaded an ignorant and sceptical world that mitochondria and chloroplasts had originally been free-living microbes that became symbiotically associated to the point of interdependency with ancestral animal and plant cells. Those cells had themselves been derived from at least two distinct genomes. It seems that several, if not many, of the major evolutionary transitions have been associated with separate genomes coming together in what has eventually become a mutually dependent relationship - a process called symbiogenesis. I expected this book, which would review many known examples and analyse the consequences of symbiogenesis, to be a real treat. I was sadly disappointed.

Parts of the book are factually incorrect and therefore misleading to the lay reader. Several arguments are contradictory, while the work and views of several research scientists are misrepresented. Ernst Mayr's foreword tries to let the authors off the hook by suggesting that readers should ignore interpretations that are "clearly in conflict with modern biology". That is just not good enough for those who are reading a popular science text to find out about modern biology; they cannot know what to ignore.

According to Margulis and Dorion Sagan, speciation always occurs as a result of symbiogenesis but, as Mayr points out, there "is no indication that any of the 10,000 species of birds or the 4,500 species of mammals originated by symbiogenesis".

Throughout, contemporary evolutionary theory is found to be wanting, because it is supposed to be based on erroneous concepts such as inclusive fitness, evolutionarily stable strategies, mate competition, parental investment, reciprocal altruism or even altruism itself. Yet there is no coherent explanation for why these concepts are problematic, except that some are based on metaphors. But then a supposed foundation of their new evolution will, say Margulis and Sagan, be one of the weakest metaphors yet devised: Gaia, the earth goddess. Not only do they base their supposed theory on a metaphor (despite criticising others for using them), they also invent their own - for example, speaking of termite colonies: "When their walled and chimneyed fortresses are threatened... bells toll, sirens screech, professional relief teams mobilise." What are tolling bells meant to represent in a disturbed termite colony?

To dismiss concepts without justification is one thing, but to attack accomplished scientists for supposed ignorance is another. The claim that Richard Dawkins and John Maynard Smith (or their students) will need to learn something about "chemistry, microbiology, molecular biology, paleontology and the air" is outrageous. I have worked alongside Dawkins and Maynard Smith for decades, and their knowledge of these subjects is considerable.

This could have been such an important book because symbiogenesis probably is underrated as a factor in evolutionary change. That is categorically not to say, as Margulis and Sagan claim, that symbiogenesis is in any way anti-neo-Darwinian. The process fits straightforwardly into the neo-Darwinian perspective, but we do not know how important it has been. It needed somebody to bring the examples together and explain when and how acquiring genomes has happened.

Fascinating examples of genomes becoming mutually dependent, resulting in new life forms, are described in the book, and they seem to constitute the pearls. However, the extent to which even those examples are properly described cannot be judged without going to the original literature.

Margulis probably knows that literature better than anyone, and I should have liked to have been able to take her word for it. But on the evidence of other things in this book, I can no longer take her seriously.

Paul Harvey is head of zoology, Oxford University.