Ants are important consumers in many ecosystems, but much remains unknown about how ant colonies process energy. Estimates of colony respiration are often based on measurements of isolated ants, assuming individual metabolic rate remains constant with group size. To test this assumption, I scaled metabolic rate (VCO₂) against group size (M) to quantify exponents (b) and determine whether the metabolic rate per ant was independent of (b = 1) or allometrically related to (b ≠ 1) group size. 
Results/Conclusions

VCO₂ scaled allometrically for Crematogaster lineolata (~M^≈1.20) and Aphaenogaster treatae (~M^≈0.80). For four of six species tested, VCO₂ scaled isometrically after abruptly increasing in samples with two ants. Two species of Pheidole tested have a dimorphic worker caste. Whereas minor workers of P. dentata (b = 1.04) and P. bicarinata (b = 1.07) had higher metabolic rates in larger groups, majors of both P. dentata (b = 0.79) and P bicarinata (b = 0.90) showed the opposite pattern. The energetic costs to colonies of maintaining majors may thus be offset by their diminished respiration in groups. Crowded ants are thought to have higher metabolic rates. A 10-fold reduction in respirometry chamber volume yielded allometry (b = 1.14) for groups of P. dentata minor workers. Observed scaling variation may reflect caste and species differences in worker behaviour and integration at the colony level. Further comparative studies will shed light on these mechanisms that may explain how the relatively simple behaviours of workers yield complex actions of whole colonies.