Abstract
Localised patterns of resource use can be constrained by multiple factors. Comparison of resource use at multiple locations with differing resource availability can allow fundamental specialists to be distinguished from species that simply feed predominantly on prey types that are locally abundant. This study investigates geographic variation in the feeding ecology of coral-feeding butterflyfishes to examine whether patterns of resource use and levels of dietary specialisation vary among distinct locations, corresponding with changes in resource availability. Our specific aims were to investigate whether the dietary niche breadth of four butterflyfishes varies among five geographically separated locations and assess whether each species utilises similar resources in each location. Resource availability and dietary composition of four butterflyfishes were quantified at three sites across each of five geographic locations throughout the Pacific. Niche breadth, niche overlap, and resource selection functions were calculated for each species at each site and compared among locations. Availability of dietary resources varied significantly among locations and sites. Chaetodon vagabundus, C. citrinellus and C. lunulatus had low levels of dietary specialisation and used different resources in each location. Chaetodon trifascialis had high levels of dietary specialisation and used the same few resources in each location. Our results indicate that relative levels of dietary specialisation among different butterflyfishes do hold at larger spatial scales, however, geographical variation in the dietary composition of all butterflyfishes indicates that prey availability has a fundamental influence on dietary composition. Highly specialised species such as C. trifascialis will be highly vulnerable to coral loss as they appear to be largely inflexible in their dietary composition. However, the increased feeding plasticity observed here for C. trifascialis suggests this species may have a greater capacity to respond to coral loss than previously assumed.