Application of network theory to mark recapture data allows insights into population structure of two Heliconius species

Abstract

By noting the spatial location of captured individuals mark-recapture studies create a collection of discrete events spread in space and time. This setup is appropriate for network modeling where the vertices (or nodes) are the points of capture and links are established whenever a recapture occurs. Applying network analytical tools, it is possible to ascertain aspects of spatial structure and generate predictions regarding the likely causes of structure in the network. We studied the spatial network of two tropical butterfly species, Heliconius erato and H. melpomene, using a mark-recapture database from a 2-year survey in an Atlantic Forest remnant in Brazil. The overall network structure of both species was similar in number of vertices and average connectivity. Heliconius erato had a smaller, more disconnected network structure, suggesting shorter traveling paths. The distribution of connectivity of both species was better adjusted by a power-law distribution. We found hubs in both species; hubs are points of high capture and their location is correlated with the location of flowering plants visited by adults. In complex systems, hub elimination can have a notable collapsing effect in network structure. Because resource hubs are important for butterfly network organization we suggest management as well as experimental tests with regards to the role of resource hotspots for population structure.

Keywords

• network theory
• functional habitat
• butterflies
• Atlantic forest
• movement ecology