The living rainbow: A fatal flaw in a classic study of sexual selection

06 Drosophila melanogater Mating

A mating pair of fruit flies. (Flickr: Image Editor)

ResearchBlogging.orgA key component of classical sexual selection theory is the idea that males maximize their evolutionary fitness—the number of children they ultimately have—by mating with lots of females, while females maximize their fitness by selecting only one or a few high-quality partners. It’s pretty clear that this model works well for some species (like ducks), but also that there are many it doesn’t fit so well. Now it looks like one of the “classic” experimental examples of sexual selection may actually fall into the latter category.

Sexual selection was first proposed by Charles Darwin, in his 1871 follow-up to The Origin of Species, The Descent of Man, and Selection in Relation to Sex; but one of the earliest experimental tests of the model wasn’t published until 1948 [PDF]. The biologist A.J. Bateman allowed small groups of fruit flies—good old Drosophila melanogaster—containing equal numbers of males and females to mate at random, then reared the resulting eggs and reconstructed the parentage of the offspring to determine (1) the number of offspring each of the male and female parent flies had produced and (2) how many parters each parent fly had had.

How did Bateman reconstruct parentage decades before the advent of modern genetic testing? He used mutations with known, visible phenotypic effects as “markers”:

The fertility of individual flies of both sexes was measured by means of dominant marker genes. Several flies of each sex were mated together in one bottle, each fly carrying a different dominant marker gene. In this way, assuming the complete viability of all the marker genes, half the progeny of each fly could be identified.

That’s a pretty clever design given the technological limitations of the time. But it also turns out to be the fatal flaw in Bateman’s experiment.

Using this approach, Bateman reconstructed parentage for his offspring flies, and found that the reproductive success of males varied more than that of females; and that males were more likely than females to finish the experiment without producing any offspring—both suggestive of stronger competition for mates among the males.

Examining the number of mates for each of the parent flies, Bateman found that the number of mates also varied more in males than in females. In fact, his data showed that males with more mates tended to have more offspring—while for females, having a larger number of partners didn’t do much to improve offpsring production.

The results Bateman described are the epitome of the sexual selection model, with males competing to inseminate as many females as possible, while females instead try to select only one or a few very fit males to father their children. But those results hang on the assumption that Bateman’s marker mutations didn’t create any bias in his parentage reconstructions—and some of the marker mutations were pretty wild: gene variants that distorted the flies’ wings or eyes, many of which were lethal when a fly carried two copies.

Drosophila melanogater Bar

One of Bateman’s marker mutations, “Bar,” results in dramatically distorted eyes. (Flickr: FlyIU)

In a paper that’s just been released online at PNAS, Patricia Gowaty, Yong-Kyu Kim, and Wyatt W. Anderson present the results of the first attempt to replicate Bateman’s experiment since he first performed it in 1948. To the best of their ability, they collected the same laboratory lines of Drosophila, with the same marker mutations, and conducted the same controlled matings.

Gowaty and her collaborators replicated Bateman’s discovery that different parental marker mutations were represented in the offspring with varying frequency—but they also discovered some substantial biases in the results. One of the most glaring was that, using Bateman’s approach to reconstruct parentage, they found that mothers were less often identifiable as parents than fathers were—which, as they note, is “a biological impossibility.”

Gowaty et al. also found that the fraction offspring carrying two mutant marker genes substantially deviated from 25%, which is the frequency expected if the marker genes truly have no impact on mating success and offspring survival.

In short, Bateman’s key assumption, “the complete viability of all the marker genes,” proved to be incorrect, and incorrect in a way that seems to have systematically biased his results so that his data can’t actually tell us anything about whether or not his fruit flies were experiencing sexual selection. Fortunately, we have methods available to us nowadays that Bateman didn’t, and I’m sure we can look forward to a new test of his hypothesis using DNA fingerprinting and behavioral observations of mating frequency in the near future.◼

Thanks to Daniel Decanini for pointing me to the Gowaty et al. paper on Twitter.


Bateman AJ (1948). Intra-sexual selection in Drosophila. Heredity, 2 (Pt. 3), 349-68 PMID: 18103134

Gowaty, P.A., Kim, Y.-K., & Anderson, W.W. (2012). No evidence of sexual selection in a repetition of Bateman’s classic study of Drosophila melanogaster. Proc. Nat. Acad. Sci. USA. : 10.1073/pnas.1207851109