Navigating social relationships frequently rests on the ability to recognize familiar individuals using phenotypic characteristics. Across diverse taxa, animals vary in their capacities for social recognition but the ecological and social sources of selection for recognition are often unclear. In a comparative study of two closely related species of poison frogs, we identified a species difference in social recognition of territory neighbors and investigated potential sources of selection underlying this difference. In response to acoustic playbacks, male golden rocket frogs (Anomaloglossus beebei) recognized the calls of neighbors and displayed a “dear enemy effect” by responding less aggressively to neighbors’ calls than strangers’ calls. In contrast, male Kai rocket frogs (Anomaloglossus kaiei) were equally aggressive to the calls of neighbors and strangers. This species difference in behavior is associated with key differences in reproductive ecology and characteristics of territories. Golden rocket frogs defend reproductive resources in the form of bromeliads, which is expected to create a threat asymmetry between neighbors and strangers favoring decreased aggression to neighbors. In contrast, Kai rocket frogs do not defend reproductive resources. Further, compared with Kai rocket frog territories, golden rocket frog territories occur at higher densities and are defended for longer periods of time, creating a more complex social environment with more opportunities for repeated but unnecessary aggression between neighbors, which should favor the ability to recognize and exhibit less aggression towards neighbors. These results suggest that differences in reproductive ecology can drive changes in social structure that select for social recognition.

These data were collected by James Tumulty in the field in Kaieteur National Park in Guyana. Included are data on aggressive thresholds of male golden rocket frogs (Anomaloglossus beebei) and Kai rocket frogs (Anomaloglossus kaiei) in response to acoustic playback experiments. Also included are spatial data from long term monitoring of territorial males of both species as well as egg clutches, tadpoles, and reproductive resources (bromeliads). Spatial data have been processed as described in the manuscript to produce home range estimates, nearest neighbor distances, and spatial associations between territorial males and egg clutches, tadpoles, and bromeliads. Additionally, simulated spatial association data are included.

Please see README files associated with each data set.