Over the past week or so, there has been an absolutely amazing image that has made the rounds on the internet of a fly (Goniurellia tridens)with markings on its wings reminiscent to many viewers of ants.
Goniurellia tridens is a 3-in-1 insect [photo: Peter Roosenschoon] pic.twitter.com/i8ThAOkrvN
— Ziya Tong (@ziyatong) November 4, 2013
As described in these blog posts and articles (Anna Zacharias, Jerry Coyne, Morgan Jackson, Andrew Revkin, Joe Hanson, also here and here) about it, the assumption is that these images are used by this fly to mimic the ant (or more likely a spider- more on this below), to act to ward off potential predators. However, there has been relatively little discussion about the context in which it uses it (but see Morgan Jackson's post), and demonstration of its adaptive utility. As pointed out by many evolutionary biologists, and discussed in detail by Gould and Lewontin in one of the most famous papers in evolutionary biology (The Spandrels of San Marco and the Panglossian paradigm: A critique of the Adaptationist programme), it is easy to make a "just so" adaptive story, but as scientists we need to perform critical experiments demonstrating the adaptive utility of this picture on the wing.
As numerous commenters on the blogs and on twitter have pointed out this fly is part of the family of true fruit flies (Tephritidae), that include several that are known to startle jumping spider (causing them to do a short retreat). This retreat is likely because the flies have evolved to mimic aggressive behaviours of the spiders themselves. This work was initially described over 25 years ago in a pair of papers in Science (One by Erik Greene, Larry Orsak and Douglas Whitman. The other paper by Monica Mather and Bernard Roitberg). These papers beautifully demonstrate the adaptive utility of markings on the wing combined with a rowing action of the wings that could achieve this mimicry. Neither the markings on the wings nor the rowing behaviour alone were sufficient to induce the aversion behaviour in the spiders (the spiders retreat). Indeed those of us who took biology courses in University in the early to mid 1990's probably remember this example being taught to us. What's more is that it seems to be fairly wide spread among species in this family of flies (each research group used a different species of Tephritid fly and spider.) Another paper (Oren Hasson 1995) tested about 18 different species of jumping spiders with the medfly (also a Tephritid), and showed that most spiders responded with the retreat as well. This suggests that this adaptive wing morphology and behaviour combination is probably pretty ancient.
Here I want to show you a video of a picture-winged fly, with a jumping spider. This fly is from a totally different family of flies (the picture-winged flies Ulidiidae (formerly Otitidae)) than the ones discussed above (Tephritids), but apparently does the same thing to startle jumping spiders (as a way of escaping being eaten) as the true fruit flies.
A few years ago, when I was hosting a lab bbq in my backyard, we were lucky enough to get to watch the intricate little behavioural "routine" between a fly and a jumping spider (in this case the bold jumping spider, Phiddipus audax). The spider approached the fly, got into its attack posture, and then the fly did its "wing rowing" display, the spider "retreated" (took a short jump back), and the fly took off, successfully evading getting eaten. Not too shabby, plus how often do you get to watch this for real!
Two years ago I got to watch this happen again, and this time I happened to have some collecting vials. So I collected the flies! I then put the fly in a small dish with a jumping spider (the zebra spider Salticus scenicus) so I could get some simple video of it. Here it is in all its grainy, low quality glory.
Given that I am not a great entomologist, I sent a picture of the fly off to a colleague (Jim Parsons, our collection manager in the MSU Entomology department), and he pointed out to me that this was not a true fruit fly (Tephritid) at all, but a picture-winged fly (from the family Ulidiidae). This particular fly is called Delphinia picta.
This was clearly really exciting, as it shows the potential for a whole other group of flies demonstrating a similar set of anti-predation behaviours. While both of these families belong to the same super-family, their last common ancestor lived probably 75 million years ago (give or take several million years). Is this an example of two different groups of animals independently adapting the same way (convergence) to a similar selective pressure (not getting eaten)? Or is it an adaptation that has survived for millions of years across many species? Finally the possibility exists that some aspects of the behaviours and wing spots allow this to evolve as an anti-predator adaptation over and over again (parallelism)? Whatever it is, it suggests that something even deeper and cooler has happened in evolution, and it will be great to figure this out (hint to new graduate students seeking projects!). As my colleague Rich Lenski mentioned to me (when I showed him this video), it also makes one think carefully about the appropriate "null hypothesis" regarding putative adaptations!
In my lab, one of the things we study is the fly Drosophila melanogaster, and how it evolves in response to potential predators, including jumping spiders. Drosophila is the little fly that you used in high school or university biology. Many call it a fruit fly, even though it isn't (pomace fly and vinegar fly are both used as its common names). For Drosophila we have never observed this kind of behaviours at all. However Drosophila does display a pretty wide range of behaviours, and we are writing up a paper about it right now. For a taste of some of it, check out my graduate students poster over on figshare describing some of the behaviours.
Let me know if you want more, and maybe I can post some additional video. However, to whet your appetite here is another related video that we posted a while ago to youtube (of flies with a mantid). The action starts at about 2:30 into the video.