Small cursorial ectotherms risk overheating when foraging in the tropical forest canopy, where the surfaces of unshaded tree branches commonly exceed 50 °C. We quantified the heating and subsequent cooling rates of 11 common canopy ant species from Panama and tested the hypothesis that ant workers stop foraging at temperatures consistent with the prevention of overheating. We created hot experimental "sunflecks" on existing foraging trails of four ant species from different clades and spanning a broad range of body size, heating rate, and critical thermal maxima (CT). Different ant species exhibited very different heating rates in the lab, and these differences did not follow trends predicted by body size alone. Experiments with ant models showed that heating rates are strongly affected by color in addition to body size. Foraging workers of all species showed strong responses to heating and consistently abandoned focal sites between 36 and 44 °C. Atta colombica and Azteca trigona workers resumed foraging shortly after heat was removed, but Cephalotes atratus and Dolichoderus bispinosus workers continued to avoid the heated patch even after >5 min of cooling. Large foraging ants (C. atratus) responded slowly to developing thermal extremes, whereas small ants (A. trigona) evacuated sunflecks relatively quickly, and at lower estimated body temperatures than when revisiting previously heated patches. The results of this study provide the first field-based insight into how foraging ants respond behaviorally to the heterogeneous thermal landscape of the tropical forest canopy.

Woody debris (WD) stocks and fluxes are important components of forest carbon budgets, and yet remain understudied, particularly in tropical forests. Here we present the most comprehensive assessment of WD stocks and fluxes yet conducted in a tropical forest, including one of the first tropical estimates of suspended WD. We rely on data collected over 8 years in an old-growth moist tropical forest in Panama to quantify spatiotemporal variability and estimate minimum sample sizes for different components. Downed WD constituted the majority of total WD mass (78%), standing WD contributed a substantial minority (21%), and suspended WD was the smallest component (1%). However, when considering sections of downed WD that are elevated above the soil, the majority of WD inputs and approximately 50% of WD stocks were disconnected from the forest floor. Branchfall and liana wood accounted for 17% and 2% of downed WD, respectively. Residence times averaged 1.9 years for standing coarse WD (CWD; >20 cm diameter) and 3.6 years for downed CWD. WD stocks and inputs were highly spatially variable, such that the sampling efforts necessary to estimate true values within 10% with 95%