From the prokaryotic mayhem of the primordial soup emerged the multicellular eukaryotes, which occupy more space, do cleverer tasks and have bigger and more complex genomes. But what route did phenotypic evolution take in going from one to the other? Did adaptive diversification depend on having a more complex genome or was it the other way around? Mike Lynch and John Conery provide statistical evidence that the more complex genomes of multicellular eukaryotes arose passively and therefore without much adaptive purpose. They propose that only once the new genomic features were in place would they have been exploited for adaptive purposes.

The genomes of multicellular eukaryotes are not only much bigger than unicellular ones, but they also have more genes, more introns and more mobile elements. Although there are plausible advantages that these features would bring, it is also possible that more complex genomes arose because there was nothing to prevent them from arising. The 'something' that could prevent this is purifying selection, which purges undesirable variants from a population.

Purifying selection is less powerful in smaller populations, in which traits have a better chance of spreading through the population owing to random forces. The authors have calculated that as organisms get larger, on average their population size gets smaller. They show that the effective population size (Ne) — the number of individuals that actually contribute to the next generation — can vary by several orders of magnitude between the largest and smallest organisms. So, multicellular species, with their much smaller Ne — which, in turn, is probably caused by their larger cell and body sizes — would be freer to accumulate non-selected changes to their genome.

The authors tested their theoretical expectations against the characteristic features of multicellular genomes. For example, they show that multicellular species probably have more genes because they retain duplicated genes longer than do unicellular species, as mutations take longer to erode them. So, rather than one copy of the duplicate pair degenerating out of existence, both could survive by splitting between them the role of the ancestral locus (through a process known as 'subfunctionalization'). A similar sort of reasoning can be proposed to explain the emergence of a large number of introns and mobile elements.

Of course, the idea is not that all complex traits arose by chance — rather, it is that a non-adpative expansion in the genome provided the genetic raw material for selection to work on. For example, it is perfectly feasible that once a large number of introns were present, they would be put to use in alternative splicing, thereby paving the way for more adaptive evolutionary changes. As the authors point out, more directed experiments are needed to prove their model and that 'exceptional species' within each group should be good testing ground for their theories.