Nov 062013

There’s a pretty remarkable fly photograph making the rounds of social media today, and while it originally had me going “Oooooh!”, the more I think about it, the more I feel like we’re staring at clouds.

It started when Ziya Tong tweeted a photo of a Goniurellia tridens (a fruit fly in the family Tephritidae) displaying its wings:

Do the wing patterns look a bit like ants to you? The photo has since been picked up by a number of high profile sites, like the New York Times Dot Earth Blog, Why Evolution is True and It’s Okay To Be Smart because, hey, it looks like this fly brings friends with it where ever it goes. I’m all for flies getting some positive exposure & attention, but the more I look at this fly and think about its relatives & evolutionary history, the more I think we’re all reading too much into that pattern.

This stellate (star-like) wing pattern is very common among flies in the Tephritinae, the subfamily of fruit flies this species belongs to, although the specifics differ between species and genera. Here are the wing patterns of the other Goniurellia species (found across the Mediterranean, Middle East, Africa and Asia).

Goniurellia wings Freidberg 1980

It’s easy to think that species in far away places like the Middle East and Africa have special, exotic adaptations, but in fact we have equally cool species in our own backyards. If you look closely at some of the species nearby, you can see this general wing pattern is conserved in other tephritine genera, like Trupanea and Euarestoides, just without the darker “eyespot” (technically called a bulla, a dark, blister-like bubble of the wing membrane, which can be found in other local-but-less-ant-like wing patterns, such as Euaresta bella‘s).

Trupanea actinobola

Euarestoides abstersus

Another theory that’s been circulating is that rather than ants, the wing patterns may be representations of jumping spiders. Again, there’s some precedence for this, as several Rhagoletis & Zonosemata species (which are fruit crop pests, and thus more heavily studied) have been shown to hold their wings in such a way as to give them the appearance of an 8-legged threat. But these flies put on a grand charade, using their entire bodies to fool predators into backing down, not just the tips of their wings. Additionally, the spider mimicking flies are in a totally different subfamily, and while the higher relationships of Tephritidae are still being figured out, as far as I can tell no one has ever suggested spider-mimicking wings are an ancestral state shared by both subfamilies, and would therefore argue there’s not a strong evolutionary link. That’s not to say there’s not some form of convergence occurring, but it’s important to note that the ant-winged flies are more closely related to species that look like this

Campiglossa albiceps

and this

Eutreta novaeboracensis

than they are to species that look like this

Rhagoletis cornivora

In all of these examples though, we’re looking at the wings splayed out, and not than in their natural resting position. Take a look at this Trupanea actinobola photo by Steve Marshall showing how the wings look when not in use.

Trupanea actinobola

When the two patterns are put together, the “ants” would be standing on top of the other’s feet, which isn’t exactly the most natural behaviour for an ant. It’s possible that the wing patterns together are serving as a false-head (as suggested to me by Richard Jones on Twitter), and thus tricking predators into attacking the wing tips first (which would still be bad for the fly’s ability to fly) or into attacking the false-head from “behind”, thus putting them squarely in front of the true head and easily watched and avoided by the fly. This strikes me as a pretty reasonable argument, but why might there be so much variation in wing patterns if species are using them as a false head?

Another factor to consider is what other uses wing patterns may have besides predator avoidance. Many fruit flies, especially those within the Tephritinae, have very complex mating rituals, and will perform routines that can involve any combination of leg waving, body swelling and wing waving & supinating (plus many other behaviours). Of course, if these wing patterns and flag-waving behaviours are for the female’s benefit, then it would follow that they can actually see and recognize the patterns and routines. While chemical stimuli (both from host plants and aggregating flies) are generally believed to draw individuals from long ranges into close proximity, it seems that fruit flies have very strong visual acuity, and can correctly identify conspecifics based on wing patterning when presented with a choice (Headrick & Goeden, 1994). This means that small changes in wing pattern can help diverging populations sort themselves out, even if host plants and other factors remain the same.

Returning to the viral ant-winged photo, it would appear that the fly is supinating its wings (twisting the wing 90o and pressing forwards to display patterns), which can occur both when a species is pretending they’re a threatening spider as well as in species with courtship wing displays, which doesn’t help us decide what’s going on — is there a female just off camera that’s being courted, or is the fly threatened by the giant camera lens in its face? However, if you look closely at the photo you can see that the middle and hind legs are actually curled up under the body and the fore legs are not resting in a natural position at all — they look more like a ballerina en pointe, with the last leg segments curled and with the tops of the “feet” (the tarsomeres) resting on the surface of the substrate — which leads me to believe the fly in the photo may be dead, or at least heavily compromised, and not actively displaying its wings at all!

Putting everything together, it leads me to believe we may be choosing to see ants where they don’t actually exist. Much like how we see sharks in the clouds or Jesus in our toast (a psychological phenomenon called Pareidolia), I think we’ve become so conditioned to expect ornate patterns on wings to be mimicking something else that we’re forcing objects to appear everywhere, even if there’s no evolutionary or behavioural explanation for it. It’s important that we don’t let our human-centric points of view, emotions and opinions bias our interpretation of what’s really going on. The fact that there are so many amazing and believable mimicry strategies in the world of insects is awe-inspiring, but rather than trying to fit them all in to the same box, we should instead be working to understand what crazy new idea natural selection is experimenting with now.

Neaspilota albidipennis

Neaspilota albidipennis

Freidberg A. (1980). A revision of the genus Goniurellia Hendel (Diptera: Tephritidae), Journal of the Entomological Society of Southern Africa, 43 (2) 257-274. 

Headrick D.H. & Goeden R.D. (1994). Reproductive behavior of California fruit flies and the classification and evolution of Tephritidae (Diptera) mating systems, Studia Dipterologica, 1 (2) 194-252.

Fruit fly wing photos and Trupanea actinobola live habitus from:

Jackson M.D., Marshall S.A., Hanner R. & Norrbom A.L. (2011). The Fruit Flies (Tephritidae) of Ontario, Canadian Journal of Arthropod Identification, 15 1-251. DOI:

  45 Responses to “Ants, spiders, or wishful thinking?”

Comments (28) Pingbacks (17)
  1. That is a fascinating explanation. Thanks for going into so much detail!

  2. Wonderful! Yet, this too is a hypothesis and one I would not automatically accept even if it is of contrarian mundanity. To decide the question is to ask which explanation is simpler? Your explanation might tempt one to bias in the opposite way but that is confusing mundanity with simplicity. I do not have an answer but I will go through a thought exercise.

    One can inspect the variation in the pattern across the species and apart from genus or family to decide whether it is truly random and merely triggering false matches in our pattern detectors or serves some utility. A stable pattern within a more variable family pattern is merely a trace of its evolution. The less the species variation, the more fixed in the genetics the less like shark in a cloud it is. So if it is not random why would it be selected for? It could be prey avoidance or mating or likely, merely correlated with something else that is selected for. But of course that is only one level of random, it’s possible that particular pattern evolved by chance.

    So the question really is how likely is it for events to align such that the wings are displayed to say a predator and aid in escape. If by random chance and across all encounters, the statistics are enough to offer some advantage in say adversary confusion then tuning the pattern (likely selected on primarily for non protective reasons) may well be one of some advantage to the insect genetics. It could very well be an indirect exploit of some conserved part of visual processing. You’re assuming it is only possible for us to falsely recognize images.

    I cannot say that yours is automatically the better hypothesis even though it is less interesting. Because other than the above, what strikes me unlike all other examples of false patterns is how stable and faithful it is across various lighting conditions and angles. Most false patterns require time to discern, attention to maintain and can sometimes morph or are at least exchangeable to some other near thing. So I remain uncertain.

    • Wow, could you have made your point any more pseudo-intelligent and difficult to read? It’s very badly written. It’s perfectly possible to get your opinion across without seeming like you are spewing thesaurus alternatives twice per sentence.
      “You’re assuming it is only possible for us to falsely recognize images.”… no he isn’t, he’s just accumulating sensible opposing views and pointing out that the people are too quick to accept what they have been told from a viral image without thinking about the theory.

    • So it’s basically “coincidences don’t exist” vs. “coincidences do exist” – and while coincidences certainly do exist, I tend to go with @Ende on this one. Reason being: if “wing-art” is so prevalent in this family, then certainly some variant could have experienced some selective (whether through predation, sexual selection or both) pressures due to its environment and other circumstances. I don’t think it’s too far fetched to say that such a pattern *could* have evolved to mimic some insect in such a way that it dimishes chances of predation.

  3. “may be dead, or at least heavily compromised”? There’s a freaking needle stuck through her. She’s as dead as one can be.

  4. Very well thought out and concise response, Morgan. It’s almost as if you’ve studied these organisms…

    I have a new theory though: that Trupanea actinobola looks like it’s mimicking Space Invaders. Space Invaders always rest on top of each other.

  5. I’d like to see the UV patterning on the wings (like in this paper: I’d bet the things light up when under UV, which would strongly suggest looking like a poor facsimile of an ant is not the function of the pattern. I’m going to go with sexual selection on this one.

  6. I love humans. Always seeing patterns in things that aren’t there.
    ~The Eighth Doctor

  7. Thanks, Morgan! This analysis (though much based on the Headrick & Goeden, 1994 idea), is an urgent relpy for that 50-years old speculation by Pavel Marikovsky, which has become a “wandering sujet” among Russian entomologists, but only recently penetrated into the New World. My comment was that — in addition — Inula hirta, the host plant of thsi fly, IS NOT COLONIZED BY APHIDS and visited by ants!

  8. I’m no entomologist, but to me it doesn’t look like most ants. The neck is too small. It looks like a wasp or a wasp-waisted fly.

  9. That wings are an example of pareidolia, but is this way that evolution works. Mutations happens randomly and then if the organism benefits of those mutations, its offsprings with the same mutation will benefit of it and have an advantage over others. If that random image on the wings of that midge will make it more resistant to predators, then it will have an advantage over other midges and in the future there will be more of its species with an ant or spider like image on the wings. But those wings could be also ineffective in frightening predators: after all is a random image, not a carefully designed wing.

  10. Oh my, an ant-wing denialist. Keep your head down and maybe the spider and ant people will go to war and not notice. Doesn’t really look like either to me. I think they are chalcidoid wasps and the fly uses them to confuse egg parasitoids.

  11. This looks like the early stage of evolution. In many thousands of years from now, if the species is not extinct, the wing patterns will be more consistent and have greater survival advantages. If the current mild advantage of some patterns does not prevail, the pattern randomness will either continue or start to disappear and other new survival advantage mechanisms will supersede.
    PS: This is a layman with very little science background commenting.

  12. It’s funny, when I first saw it I actually thought the pattern looked more like a fly, because of the way the “head” looks with it’s stubby short “antennae”. But now I’m heavily leaning towards the Dead Dipteran hypothesis.

  13. i have yet to see an ant, but immediately upon seeing the image linked from here i could have sworn it was jumping spiders one on each wing, and i started pondering wether the fly could have been sucked up by the spiders and now they are hunting for more prey, or what the……… i couldnt imaging a fly letting spiders on his wings. So NEVER did i assume ants, then when i read the article, taken from, it said ants, and on this page mentioning ants i dont know what you all are talking about seriously. Im guessing maybe in most corners of the world you guys dont see jumping spiders too often but i have grown up with them.

    The fact that i see jumping spiders and you see ants DOES sort of lead to the theory that we see what we want, BUT im almost sure that this is a defense, and i would venture whatever preys on this fly, or atleast one of its main predators are careful to not eat spiders.

    i couldnt see the same effect for ants. it would make more sense to be eaten if you have an ant on your wing since thats just more protien to a predator more than likely.

  14. Hi Morgan, I got to this post via a comment posted by Alex Wild on FB. Very well written, and I agree with your analysis of the subject. People are so used to looking for familiar patterns everywhere including animals and plants, and the best example for this is the huge number of butterflies/bugs/spiders-with-a-face-on-their-body and monkey-face/butterfly/bird-like-flowers photos surfacing every once in a while on the web.
    I know these flies from Israel, and I immediately recognized the image you posted from Amnon Freidberg’s work on them. I think the only way to tell for sure what these wing patterns are used for is by observing the flies’ behavior in the field – how are they posing? Are they waving their wings? What other organisms are in the nearby area interacting with the flies? Someone commented above that these flies are found on plants that are not even visited by aphids or ants. This is an important observation. However, I have a different experience – I have seen the flies on the inflorescence of several thistle species (Asteraceae), and the stems were always heavily infested with aphids (usually Brachycaudus helichrysi) and tended by ants. I actually intend to visit Israel later in the spring so this should be a nice topic to look into, since these flies are relatively easy to find.

  15. Thanks for such a thought provoking and in depth article. I beg to differ,however, concerning your “Jesus in toast” theory. When I initially saw the photo I thought it was an image of two ants crawling on the wings of the fly. I tend to think of flies wings an opalescent or transparent and unmarked. I am no entymologist! (forgive misspelling?) It was only after reading the accompanying commentary from the Center for Biological Diversity (which forwarded me to this site) that I understood it to be wing markings. Thanks again for your knowledge and considerations of the subject. Personally I was as fascinated by the beautifully colorful eyes as much as the awesome wings!
    Thanks, Dasi Bhaktivedanta

  16. I agree that more testing of the hypotheses is needed. Have posted a slightly alternative view –

Leave a Reply to Dasi Bhaktivedanta Cancel reply



You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>