This blog post is provided by Denise Dalbosco Dell’Aglio, Sebastián Mena, Rémi Mauxion, W. Owen McMillan and Stephen H. Montgomery and tells the #StoryBehindThePaper for the paper “Divergence in Heliconius flight behaviour is associated with local adaptation to different forest structures”, which was recently published in the Journal of Animal Ecology.

Butterflies are found everywhere in a tropical forest. Some are found very high in the canopy, others fly low around shrubs. What makes butterflies choose where to fly? Forest stratification is important in maintaining diversity, and butterfly stratification in flight height has an important role in maintaining diverse communities. Adaptive shifts in flight behaviour may be particularly important during ecological speciation, and could be a key attribute in this process in butterfly communities. This variation can increase local foraging efficiency and help to segregate closely related species, potentially increasing reproductive isolation, while exposing them to novel conditions that may favour further local adaptation.

Heliconius butterflies are known for their amazing colour patterns and mimicry, yet their flight behaviour in using different environments is not fully explored. To do so, we took advantage of a case of incomplete, or ‘incipient’ speciation in two closely related Heliconius butterflies: the parapatric pair of Ecuadorian species H. erato cyrbia and H. himera, which is an interesting case that demonstrates the transition between species across an environmental gradient. These “incipient” species diverged relatively recently, continue to hybridise in the wild, and the species boundary is in part maintained by ecological isolation and local adaptation.

To investigate flight height in these two species, our first stop was in a small village called Vilcabamba in the southern region of Ecuador, home of one of our focal species Heliconius himera. This species is endemic to a dry forest situated at 1500 m in altitude. The Incas referred to this region as “Sacred Valley” and indeed, it has its beauties and mysteries. To find H. himera we had to look for river ravines, or “quebradas”, places where the habitat was more preserved. Whenever we found a butterfly, we took measurements not only of its flight height but also of its surroundings, in particular which flower sources and host plants were available, and how they were located in the forest.

Our second stop was in the region of Piñas, more specifically in Buenaventura Reserve (Jocotoco Foundation), which is situated in the Western Andes at 600 m of altitude with tropical evergreen wet forest. In contrast to the dry bushy forest where we find H. himera, this locality has a very humid and well preserved broad-leaf forest. It was not hard to find dozens of butterfly species, especially our second species, H. e. cyrbia, which was flying around our lodge. We also found a high diversity of Passiflora species, in which we could see eggs and larvae confirming its use as a host plant.

In the field, we found a strong association between host plant height and flight height across all our sample sites, which suggests that the microhabitat favoured by host-plants may constrain adult movement. Our observations have shown that flight behaviours match the distribution of plant resources implying that the capacity for behavioural convergence within mimetic communities may be constrained by plant ecology.

After the field work, it was time to go back to the lab and use common garden experiments to test whether differences in the observed flight height reflect individual responses to resource distribution or are genetically determined shifts in foraging behaviour. This experimental work was done in the outdoor insectaries of the Smithsonian Tropical Research Institute in Gamboa, Panama. We built a “flight cage” where we could observe the butterflies flying around and feeding from artificial flowers. We set up two cameras inside the cage to help us collect the data.

In our common garden experiments, we found that the differences in flight behaviour we observed in nature are recapitulated in our experimental cages. This suggests that these behaviours are not merely plastic responses by individuals to perceived or experienced environmental cues, but likely have a genetic basis. By both measures in our experiments – obstacle avoidance and feeding height – H. e. cyrbia expressed a preference for flying higher and feeding from higher feeders. On the other hand, H. himera chose to fly lower and feed at lower positions, closely mirroring differences in behaviour in the wild, which align with the availability of flowers and host plants.

H. e. cyrbia and H. himera reflect an important intermediate stage in the speciation process and the differences in flight height we reported here are likely to reflect a behavioural adaptation to optimise foraging efficiency. We have shown that butterfly behaviour responds to ecological features of their environment, providing a partial explanation for the implied competitive advantage between the species in the middle of the wet and dry ecotone, and also may provide a clue to how these two species remain distinct through selection against hybrids and migrants. Our work highlights the potential role of habitat-dependent divergence in behaviour during the early stages of speciation. We still have more to learn about behaviour barriers between closely related species beyond mating behaviours.