Out Damned Spot! Or, How They’d Be Loving If They Weren’t Fighting

Hey, that's not an anole! Rubyspot Damselfly (Hetaerina americana) Copyright Steve Collins

Hey, that’s not an anole! American Rubyspot Damselfly (Hetaerina americana). Copyright Steve Collins

My colleagues and I recently published a paper documenting character displacement in Anolis carolinensis following the invasion of A. sagrei into Florida. The former moved up into the trees and evolved larger toepads. We did a lot of work in that paper to show with a high degree of certainty that the interaction between the two species is what led to character displacement in A. carolinensis. However, an open question remains as to exactly what kind of interaction, or interactions, they share. Most likely, the two species are competing for food (i.e. exploitative competition). They may also be interacting indirectly through a shared predator or parasite (i.e., apparent competition), and they are  known to eat each other’s hatchlings (i.e., intraguild predation).

Today, I’d like to explore another possible interaction in depth: perhaps the two species have diverged to lessen aggressive interspecific interactions for space and territory (i.e., interference competition). For more, let’s turn to the anoles of the Odonata world (provocative statement, I know!): rubyspot damselflies (Hetaerina spp).

In a recent issue of the Proceedings of the Royal Society B, Jonathan Drury and Greg Grether investigated the role of aggressive (or agonistic) interactions in driving divergence between two species of rubyspot damselflies.

Previous work [1,2,3] in Grether’s group had shown that male competitor recognition in rubyspot damselflies depends on hindwing coloration, and that cross-species recognition and male wing coloration diverges between species living in the same area. This suggests that aggressive interactions between males of different species have driven divergence in wing color to reduce the frequency of energy-intensive, aggressive interactions between species. This divergence is consistent with a type of character displacement called Agonistic Character Displacement (ACD), which is the divergence between species in some sort of species recognition trait to lessen the negative effects of aggressive encounters.

However, another type of character displacement, Reproductive Character Displacement (RCD) is also consistent with these previous findings. RCD is divergence, usually in some sort of mate recognition trait, between two species. By diverging in such a trait (think anole dewlaps), males and females of different species are less likely to spend precious time courting or mating in wasteful, failed cross-species reproductive efforts.

By this point, you, the astute reader, may have noticed that both ACD and RCD predict changes in signaling traits–the former species recognition traits, and the latter mate recognition traits.

Whenever the same trait functions as a signal for both species and mate recognition, and that does happen often, telling apart the action of these two distinct processes (i.e., selection to reduce wasted aggressive effort versus selection to reduce wasted reproductive effort ) can be very difficult*.

Drury and Grether designed a very nice test for successfully discerning between these two hypotheses.

On two different riparian transects in Texas, the authors captured wild, mature H. americana (the American rubyspot) males and assigned them to one of three treatments: (i) an unmanipulated control, (ii) a clear control to which clear ink was painted on the hindwings, and (iii) a blackened treatment, to which black ink was added to the hindwings (Figure 1). The blackened treatment made H. americana males look like males of H. titia (the smoky rubyspot). All damselflies (of both species and sexes) caught in the study were given unique markings for quick field identification.

Screen Shot 2014-11-05 at 1.52.30 PMThe authors then observed behavior along their transects to observe distances between individuals, and the ID and prevalence of mating and fighting individuals. These observational efforts, involving teams of 5 people and 9-hour surveys daily across several weeks, generated an impressive data set. The researchers marked 146 H. americana males, observed 444 unique mating events, and saw 1207 fights involving at least one H. americana male. They also re-sighted over 75% of marked males at an average of 12.1 times per day. This was a lot of work!

Drury and Grether found two major results. First, there was no effect of wing blackening on mating success in H. americana males (regardless of whether those males held territories or not) (Figure 2). This suggests strongly that  hindwing coloration, the major wing color difference between the two species, is not used for mate recognition by H. americana females. Thus, divergence in this trait is not likely a result of RCD as females have not diverged in their mating preferences along with wing color**.

Screen Shot 2014-11-05 at 2.18.51 PM

What about ACD? The authors found that male H. americana with blackened wings were much more likely to (i) fight with males of H. titia, (ii) have those fights escalate, (iii) hold territories for shorter periods of time, and (iv) disappear from study transects (Figure 3). In other words, H. americana males painted such that they looked like H. titia males suffered much more from interspecific aggressive interactions.

Screen Shot 2014-11-05 at 2.20.31 PM

Thus, these data provide strong evidence that ACD has been responsible for divergence in hindwing color. At the time of secondary contact between these two species, it is likely that natural selection  favored those H. americana that differed in hindwing color, as they would fight less and survive longer. Thus the two species diverged in hindwing color whenever they were in sympatry. With this study, the authors described an elegant experimental approach to tease apart ACD from RCD, and have further documented one of the best known cases for ACD.

Rubyspot damselflies in a mating tandem. Note the paint on the abdomen, added by the experimenters for field identification. Copyright Andrew Chao.

To bring it back to the start of this post: what roles did ACD and RCD have in the character displacement we observed in A. carolinensis, responding to the invasive A. sagrei? The two anole species differed markedly in the color of their dewlap (a signalling trait) long before secondary contact. This suggests that aggressive interactions between the two species are not influencing signalling trait evolution.

However, one trait that did change with the invasion of A. sagrei was perch height in A. carolinensis. It is possible that perch height shifts in A. carolinensis happened to lessen aggressive interactions between the two species as they built their territories. In the field, we rarely saw aggressive interactions between the two species, but we need to experimentally investigate this possibility further.

With respect to RCD, males of A. sagrei and A. carolinensis nearly exclusively ignore heterospecific females in staged encounters, and hybrids have never successfully been reported from the wild. That, combined with the fact that the perch height shift occurred within months of A. sagrei’s arrival , suggests that perch height shift in this system is not the result of RCD. It seems that the two species were already quite reproductively isolated before secondary contact.


CITATION: J.P. Drury and G.F. Grether. 2014. Interspecific aggression, no interspecific mating, drives character displacement in the wing coloration of male rubyspot damselflies (Hetaerina). Proceedings of the Royal Society B 281. DOI: 10/1098/rspb.2014.1737


* In fact, it can be impossible to support ACD if the same signalling trait is used for both mate and species recognition. In another paper from Grether’s group, they modeled the interaction of RCD and ACD and found that whenever the a trait is used for within species mate recognition, it cannot evolve ACD to lessen aggressive cross-species interactions, even if that were favored by selection. This is because divergent selection to reduce aggressive interactions will always be less strong than stabilizing selection that maintains the signal for  mating recognition. In other words, to conclude that ACD is responsible for an observed pattern of character displacement, the potential for RCD needs to be ruled out.

** The authors do note that other mating recognition traits, such as appendages used for clasping during mating, could diverge because of RCD in this system.

About Yoel Stuart

I am interested in whether, how, and why ecology shapes evolution (and evolution shapes ecology) through time, with an emphasis on microevolutionary pattern and process, adaptation, and field experiments. I completed my Ph.D. on Anolis lizards in the Department of Organismic and Evolutionary Biology at Harvard University. I am currently a post-doctoral researcher at the University of Texas, Austin studying threespine stickleback. They're not anoles, but they're cool too.

One thought on “Out Damned Spot! Or, How They’d Be Loving If They Weren’t Fighting

Leave a Reply

Your email address will not be published. Required fields are marked *

Optionally add an image (JPEG only)