Butterfly dispersal across Amazonia and its implication for biogeography

Submitted by editor on 27 November 2014. Get the paper!
Marking an individual Morpho achilles (Nymphalidae, Morphini) in Ecuadorian Amazon rainforest. Photo by P. J. DeVries.


By Carla Penz, Phil DeVries, Jarle Tufto and Russ Lande

Spatial movement plays a fundamental role in ecological and evolutionary processes as it promotes range expansion and exposes species to novel biotic and abiotic factors. The appearance of geographical barriers such as rivers or mountains can interrupt the spread of individuals across the environment, or fragment a previously continuous range (vicariance). Therefore, both dispersal and vicariance can lead to local adaptation and evolutionary diversification.  Life history parameters like longevity and body size vary among species, and such characteristics may influence dispersal ability and the efficacy of geographical barriers.  Highly diverse biomes such as the Amazon forest are of particular interest for biogeography, the investigation of species area occupancy in evolutionary time. 

Bia actorion (Nymphalidae, Brassolini) in Ecuadorean Amazon rainforest understory.  Photo by P.J. DeVries.


In this study we used empirical movement estimates of Amazonian butterflies to address two basic questions relevant to ecology and biogeography: (1) How fast could two butterfly species move across Amazonia? and (2) Does the speed of dispersal calculated for our focal species have consequences for interpreting historical biogeography?  We employed life history data, dispersal rates estimated from a mark-release-recapture study, and reaction-diffusion or integro-differential equations to calculate an annual speed for the wave of dispersal across suitable unoccupied habitat for the butterflies Morpho achilles and Bia actorion, both which range across the Amazonian biome. Our techniques estimated that M. achilles could disperse across the Amazon in approximately 1061 years, and B. actorion would take 2587 years to reach it’s current geographical range.  These estimates reflect the differences in daily dispersal speeds between the two butterflies.  Most importantly, however, we found that both species can potentially spread across Amazonia three to four orders of magnitude faster than their estimated ages (1.5 Mya for M. achilles and 39 Mya for Bia).  Morpho achilles almost certainly traversed large rivers during its range expansion, given that the Amazon river became a major eastward drainage system approximately 11.8–5 Mya, and its current width varies between 1.6 and 10 km in the dry season.  In contrast, divergence between localities north and south of the Amazon River suggests that Bia butterflies expanded their range in the Amazonian forest prior to the full development of the Amazon basin, and that wide rivers constitute a barrier to their dispersal. 

The Isthmus of Panama was formed approximately 5 Mya, and it is notable that Central American and Mexican butterfly faunas are composed largely of South American genera. We estimated that members of the butterfly genera Morpho, Colobura, Panacea, Historis and Nessaea could have expanded their range from northern Colombia to central Mexico in approximately 1319–3171 years.  Depending on species life history characteristics and age, dispersal likely played a major role in shaping present day species ranges in the Amazon, Central America and Mexico. 

Our work represents the first empirical study showing how fast Neotropical butterfly species could colonize suitable habitats.  Estimates of dispersal such as those calculated here represent a step toward placing dispersal-based hypotheses of diversification within a testable framework.