Zebra stripes and flies

A fun story by Ed Yong in The Atlantic looks at an experiment that put horses in zebra suits to test whether the stripes confound biting flies: “The Surprising Reason Zebras Have Stripes”.

When it comes to biting insects, zebras are doubly cursed. For one, they’re highly susceptible to a variety of fatal diseases, including trypanosomiasis, African horse sickness, and equine influenza, that are spread by horseflies and tsetse flies. They’re also very vulnerable to insect attacks: Compared with other grazers such as antelopes, the hairs on their coat are unusually short, allowing flies to more easily find blood vessels with their piercing mouthparts.
Stripes, for some reason, seem to help. In 2014, Caro and his colleagues showed that striped horses—three zebra species and the African wild ass with thin stripes on its legs—tend to live in regions with lots of horseflies. And several researchers, over the years, have shown that these flies find it hard to land on striped surfaces. No one, however, had watched the insects trying to bite actual zebras. That’s why Caro’s team went to Hill Livery.

This is a great story to illustrate how hard it can be to test evolutionary hypotheses. Even this one, which requires people to systematically watch lots of horses, zebras, and horses in an experimental setup, which seems pretty obvious as a test of the hypothesis, required a very special situation and tremendous effort to carry out.

And this works only because the effect is very strong. An effect strong enough to be of paramount importance to natural selection may only be a fitness increase of 1% or less. One percent fewer flies would not probably translate to one percent fitness difference; but then again, who knows? And ten percent fewer flies would require a lot of horses and zebras to show that the sample was different from chance.

And there are reasons to criticize even this setup. Are we really talking about large flies? Do tsetse flies have similar problems with the stripes? If horseflies are a problem solved by stripes, and stripes have no evolutionary cost, then why didn’t European and Asian horses have stripes? Do stripes have evolutionary costs?

Lots of questions still, but this kind of study is rare, and I’m happy to see it!