Results 
In the marking experiment, the main prediction of the Division of Labor Hypothesis was not supported, as we found no difference in thermal persistence between foragers that forage near the thermal maxima and those foraging near the thermal minima (Figure 1; Table 1). However, foragers demonstrated a higher thermal persistence than non-foragers, which is consistent with the Thermal Acclimation Hypothesis. Ants found in lower chambers of the nest had a lower thermal persistence than those collected from the higher chambers of the nest (Figure 2; Table 1), which could potentially be explained by either Acclimation or Division of Labor. There was not a strict division of labor between ants foraging during thermal maxima and those foraging during thermal minima, as several workers were marked at both time points (Figure 1). Nonetheless, their frequency was far lower than expected by random temporal distribution (Fisher’s Exact Test, p<0.0001). We separately tested for differences in thermal persistence among the 11 colonies in this experiment. One colony had an appreciably greater thermal persistence, and this difference was manifested in the whole model analysis (Table S1; Figure S2; Wald test X2 = 19.04, df=10, p<0.04; N=399).
In the circadian rhythm experiment, results indicated short-term colony thermal acclimation as well as circadian regulation of thermal persistence (Figure 3; Table 2). Colonies kept at temperatures above ambient demonstrated greater thermal persistence than those maintained at ambient, and colonies maintained at a cooler constant temperature demonstrated lower thermal persistence. Thermal persistence was different between the thermal maxima and the thermal minima, in all treatments, and particularly so under constant conditions. This effect of time of day was not found in field-collected foragers. The performance of workers removed from colony fragments every 12 hours over the time series did not atrophy over the duration of trials for up to four days, indicating that the results are not a spurious consequence of performance during captivity (Figure S1). As in the marking experiment, we separately tested for differences in thermal persistence among the 11 colonies in the experiment. Two colonies had an appreciably lower thermal persistence than the others (Table S2; Figure S3; Wald test X2 = 21.7=, df=10, p<0.01; N=402).