Countryside Biogeography of Tropical Butterflies

Although most conservation efforts focus on preserving biodiversity in relatively pristine ecosystems, we investigated possible conservation opportunities in human-dominated landscapes. We evaluated butterfly diversity in a tropical countryside that was converted about four decades ago from continuous forest to a mosaic of coffee farms, pasture, and forest fragments. We compared the butterfly fauna in coffee plantations with that in a forest remnant, the Las Cruces Reserve (227 ha). We used coffee plantation sites located “near” ( 2.5 km) and “far” ( 6 km) from the large forest remnant to test the effects of distance from the remnant on butterfly diversity. We also tested the effects of immediately adjacent habitat by selecting coffee plantation sites that were either contiguous with “small” (2.5–9.5 ha) forest fragments (coffee/forest) or lacking adjacent forest (coffee). Both coffee/forest and coffee habitats near the Las Cruces Reserve differed from those far from the reserve in species composition but not in species richness. Overall, coffee/forest habitats had significantly higher mean species richness and higher mean abundance of species than coffee and reserve sites. Further, butterflies with narrow geographic ranges were less likely to be found in coffee plantations than were those with larger geographic ranges. Area of forest cover within a radius of 50 to 100 m of the sampling site was significantly correlated with species richness of frugivorous butterflies during the dry season but was not correlated with richness of frugivorous butterflies in the wet season or of nonfrugivorous butterflies in either season. Nonetheless, species richness of frugivorous and nonfrugivorous butterflies was positively correlated overall; thus, frugivorous butterflies may be good indicators of the status of the entire butterfly community in a region. Our work suggests that small, isolated forest fragments may help retain butterfly diversity in the tropical countryside and increase the conservation value of agricultural landscapes. Relatively large tracts of forest remain important, however, because they maintain rare and endemic species. Biogeografía de Paisaje de Mariposas Tropicales Resumen: Aunque la mayoría de los esfuerzos de conservación se enfocan en la preservación de la biodiversidad en ecosistemas relativamente prístinos, investigamos las posibles oportunidades de conservación en paisajes dominados por humanos. Evaluamos la diversidad de mariposas en un paisaje tropical convertido, hace aproximadamente cuatro décadas, de bosque continuo a un mosaico de ranchos cafetaleros, pastizales y fragmentos de bosque. Comparamos la fauna de mariposas en plantaciones de café con la de un remanente de bosque, en la Reserva Las Cruces (227 ha). Utilizamos plantaciones de café localizadas “cerca” ( 2.5 km) y “lejos” ( 6 km) del remanente extenso de bosque para probar los efectos de la distancia del remanente sobre la diversidad de mariposas. También examinamos los efectos del hábitat inmediatamente adyacente seleccionando plantaciones de café contiguos a fragmentos de bosque (café/bosque) “pequeños” (2.5-9.5 ha) o sin bosque adyacente (café). Ambos hábitats café/bosque y café cerca de la Reserva Las Cruces difirieron de aquellos lejos de la reserva en la composición de especies pero no en la riqueza de especies. En general, los hábitats café/bosque tenían una riqueza de especies promedio significativamente mayor y una mayor abundancia promedio de especies que los sitios de café y de reserva. Más aún, en plantaciones de café fue menos probable encontrar mariposas con rangos geográficos angostos que mariposas con rangos geográficos mayores. La superficie de cobertura forestal a un radio de 50 a 100 m del sitio de muestreo se correlacionó significativamente con la riqueza de especies de mariposas frugívoras durante la época de sequía pero no se corre* Current address: Department of Biological Sciences, 371 Serra Mall, Stanford University, Stanford, CA 94305–5020, U.S.A., email [email protected] Paper submitted June 29, 2001; revised manuscript accepted February 5, 2002.