Fireflies (which in this case are beetles rather than flies) light up to attract mates. But this is a conspicuous signals, and a conspicuous signal has its downsides: eavesdroppers can pick up on those signals, and even imitate them to lure in prey. Moosman and colleagues suspected this flashing light had yet a third effect: to be a warning sign.
This particular firefly, Photinus pyralis, is thought to have chemical compounds that make it distasteful to come predators. Another firefly, Photuris, eats these smaller Photinus, and gain the same defensive chemical compounds from their prey.
In that case, is it possible that predators might come to recognize their flash as a warning colouration, like the bright colours of poison dart frogs or some venomous snakes? This paper ran several tests of this hypothesis.
First, the authors confirmed that the three species of insect-eating bats they were examining and fireflies overlapped in both space and time: fireflies were signaling at times and places where bats were flying.
Second, they examined a lot of bat poo for traces of firefly remains, and found none. They found plenty of other insects, including others that were firefly sized, suggesting that bats avoided these particular insects.
Third, they gave captive bats food pellets containing portions of fireflies. Bats rejected food significantly more often if it contained traces of fireflies.
At this point, all this is pretty strong evidence that bats don’t like to eat these bugs. But is it because the bats recognize the flashing light signal? The authors tested this capturing several wild bats, and exposing them to artificial lures that flashed... or not. If the lights were a warning sign in the bat’s mind, you’d predict more attacks on the non-flashing lures than the flashing lures.
The behavioural results don’t strongly support the idea. Of the three bats species, only one, Eptesicus fuscus (shown), preferentially attacked the non-flashing lures, and the only the larger, Photuris-sized lures. The authors do point out that E. fuscus is the species that overlaps with the fireflies the most, and thus may be the bat that has the most to gain by recognizing a flash as a danger sign.
Another potentially problematic aspect of their experimental design was that they presented the bats not with a natural, intermittent rate set of flashes, but with a super-firefly, continuous set of flashes. This might be okay, because greater stimuli often generate greater responses (think of the goose that will retrieve an enormous egg over a proper sized one), but it’s also possible that the unnatural stimulus is getting an unnatural response.
Given that only one of the six combinations of three bat species and two sizes of lures showed evidence of flashes being warning signs, this sentence in the conclusion paints the situation with far too wide a brush:
In conclusion, bioluminescence of adult fireflies should indeed be considered in the context of a warning signal against bats(.)
Although it is often important to show that something can happen, that this happens in only one out of six cases tested raises real questions about whether this does happen with regularity. The authors risk overreaching just a little past the data they have.
Moosman, Jr, P, Cratsley C, Lehto S, Thomas H. 2009. Do courtship flashes of fireflies (Coleoptera: Lampyridae) serve as aposematic signals to insectivorous bats? Animal Behaviour DOI: 10.1016/j.anbehav.2009.07.028