A contentious paper suggesting that butterflies and caterpillars descended from different ancestors has been rebutted in the same journal in which the original, controversial research appeared.

In August, retired biologist Donald Williamson of the University of Liverpool in England posited in an online Proceedings of the National Academy of Sciences USA (PNAS) paper that the metamorphosis between caterpillars and butterflies stems from a past cross-breeding between butterfly ancestors and those of velvet worms. "I reject the Darwinian assumption that larvae and their adults evolved from a single common ancestor," Williamson wrote.

The paper drew a great deal of fire, both for its assertions (which one developmental biologist said were better suited to the "National Enquirer than the National Academy") and for its backdoor acceptance for publication. Williamson's study arrived in PNAS as a "communicated submission," in which a member of the National Academy of Sciences submits the paper on the author's behalf and handpicks its peer reviewers. A high-placed advocate, then, can essentially knock down for an ally some of the tallest hurdles in peer-reviewed publishing. In Williamson's case, academy member Lynn Margulis of the University of Massachusetts Amherst had ushered the research into the journal.

In August, when Williamson's paper was published, Margulis told Scientific American that she needed "6 or 7" peer reviews to secure the "2 or 3" positive responses needed to present the work for publication. That statement set off a cascade of criticism of PNAS's two-tiered submission process and, according to Nature News, led the journal's editor in chief to write to Margulis demanding "a satisfactory explanation for [her] apparent selective communication of reviews." Her reply, obtained by Nature News, explained that three researchers had declined for scheduling reasons or lack of expertise, and two were omitted from the official PNAS submission because they lacked formal credentials.

Now PNAS has published a rebuttal to the hybridization paper. In direct response to Williamson's paper, "Caterpillars evolved from onychophorans by hybridogenesis," biologists Michael Hart of Simon Fraser University and Richard Grosberg of the University of California, Davis, published a study entitled "Caterpillars did not evolve from onychophorans by hybridogenesis." (The latter paper was not a communicated submission.)

Hart and Grosberg call Williamson's claim "astonishing and unfounded," asserting that data in the scientific literature show no genetic basis for his theory or its implications. For instance, they write, "all of the available phylogenetic tests strongly reject" Williamson's hypothesis that insects with caterpillar stages would contain a package of genes from the velvet worm.

Further, Hart and Grosberg add, Williamson's prediction that metamorphosing insects—the recipients of the hybridization genes—would have larger genomes than the donor velvet worms "is easily rejected": one velvet worm species, in fact, has a larger genome than is known for almost any other insect. The genome size data, they write, "are not merely inconsistent with Williamson's hypothesis but directly contradict its simplest predictions."

It remains to be seen whether the appearance of the new paper appeases some of the harshest critics of Williamson and PNAS—evolutionary biologist Jerry Coyne of the University of Chicago pondered in his blog whether the controversial work was the worst paper of the year and bemoaned "bigwigs" pushing "substandard work into publication."

As for the backdoor path to publication, PNAS announced in September that the communicated-submission system will be terminated next year but maintained to Nature News that the Williamson uproar had nothing to do with the decision.

Illustration of butterfly metamorphosis: Wikimedia Commons