Effects of Density on Spatial Aggregation and Habitat Associations of the Glass Frog Espadarana (Centrolene) prosoblepon
Social behaviors often are context dependent; behaviors at high density may change or disappear at lower population densities. We analyzed spatial dispersion patterns and habitat associations at high and low population densities of male Espadarana (Centrolene) prosoblepon that declined following the invasion of a pathogenic fungus, Batrachochytrium dendrobatidis (Bd). We mapped the spatial location of male frogs on four permanent 200-m stream transects over 4 pre- and 7 postdecline yr. We calculated distances between nearest neighbors and between hot spots—defined as areas of high male density—and compared habitat associations of frogs at high and low densities. Average predecline density across transects was 0.034 ± 0.02 males m−1 (mean ± SE) and dropped to 0.019 ± 0.02 males m−1 postdecline. Nearest neighbor and between cluster distances did not change significantly between pre- and post-decline yrs. Nine of 21 hot spots persisted in the same location across 12 yrs, whereas 8 hot spots disappeared and 4 new hot spots appeared postdecline. We determined male abundance was correlated with the amount of leafy vegetation and stream width at high density, but we found no associations at low densities. Overall, spatial dispersion patterns were highly conserved for over a decade, despite a new cohort of individuals, a decline in abundance, and changes in physical structure associated with natural forest dynamics. Our results suggest spatial dispersion is density independent and that male hot spots are influenced by both physical habitat and social interactions.Abstract

Parque Nacional G. D. Omar Torríjos Hererra (8°40′N, 80°37′17"W; 600–900 m) in El Copé, Coclé Province, Panama. Stream transects are indicated by dashed lines with names labeled. The dark solid line represents the main Rio Guabal drainage and its tributaries; light solid lines are labeled elevation contours (m).

Number of uniquely marked individual male E. prosoblepon from 2000 to 2012 on four transects (Cascada, Guabal, Loop, Silenciosa) at El Copé, Panama. 2008 and 2009 were excluded because of low sampling effort. Fewer males were found following 2004, the year Bd was detected.

Cluster hot spots of male E. prosoblepon representing pre- (2002) and post- (2012) decline years. Hot spots did not change regardless of frog density or changes to the physical stream habitat across years. Transects are shaded according to frog intensity (frogs/area): dark black = low intensity (0 frogs), gray–white = medium intensity (1–4 frogs), white = high intensity (>5 frogs).

Nearest neighbor male spacing and between cluster spacing averaged across predecline (A, C) and postdecline (B, D) years.
Contributor Notes