Metabolic rate dictates life’s tempo, yet how ecological and environmental factors integrate to shape metabolic traits remains contentious. Considering metabolic traits of 114 species of ants from seven subfamily clades along a 1,500 km climatic and soil phosphorus availability gradient in Australia, we tested four hypotheses relating to variation in metabolic rate due to niche conservatism, temperature, aridity, and ecological stoichiometry. We also tested the contested hygric hypothesis, which predicts that insect ventilation patterns can be modified to reduce water loss in arid environments. Mass-independent metabolic rate was phylogenetically conserved. The ant clade Myrmecia had metabolic rates 3 to 10x higher than other species, likely related to their large eye size, a correlate of cognitive complexity. Metabolic rate was higher in ants from warm, arid sites relative to those from wet, cool sites. A weak positive interaction between soil phosphorus and body mass indicated that, at sites with low soil phosphorus, smaller ants respired at higher rates than expected based on their mass—consistent with ecological stoichiometry theory. Larger ants, regardless of clade, were more likely to exhibit discontinuous gas exchange (DGC) with increasing aridity, likely reflecting a water conservation strategy. Phylogenetic conservatism of metabolic rate and a moderate influence of environment suggest that, in addition to biophysical geometric constraints, metabolic rate has evolved to match the energetic demands required of ecological strategies to address environmental stressors. For larger insect species confronting their metabolic limits, DGC may promote resilience in a world that is becoming hotter and more arid.

Study sites and ant sampling

Ants were sampled from six locations along the east coast and inland of south-eastern Australia representing a precipitation (421 – 1283 mm Mean Annual Precipitation) and temperature (13 – 20˚C Mean Annual Temperature) gradient. Locations were each sampled over a one-week period between June 2022 and April 2023, with locations sampled during the warmer months. Within each location, two sites with contrasting soil phosphorus status were chosen within which four plots (10 x 10 m), separated by ~200 m, were established (total of eight plots per site) and soil cores taken to confirm soil phosphorus levels. We collected as many live ant species as possible from each site over three days of sampling. We trialled a minimum of 10 individuals per colony taken from 1-3 colonies per species (30 individuals per species) from each site for metabolic assays.

Metabolic assays

Ants were housed at 20˚C in plastic nest boxes and provided with water and honey soaked in cotton wool balls every two days.  Time between field collection and metabolic trials across sites ranged between 3 and 14 days. As digestion can influence metabolic rate, ants were starved but provided with water for 48 hours prior to trials. Carbon dioxide production (VCO₂) was used as a proxy for metabolic rate and was measured at a consistent temperature of 22˚C using 8 Sable Systems International (SSI) multiple animal versatile energetics (MAVEn (SSI, Las Vegas, Nevada, USA)) system each attached to a Li-Cor 7000 CO₂/H₂O infrared gas analyser (Li-Cor, Lincoln, Nebraska, USA). Activity readings of each individual ant were measured simultaneously using infrared light detectors.

Following experiments, ants were instantly frozen at -20 °C and wet mass was measured the following day. Ants were then dried at 50 °C for 48 hours and then weighed for dry mass. Data from assays was extracted using the software Expedata (SSI) and metabolic rate converted to microwatts per hour. Data were inspected and cleaned for technical errors and outliers resulting in a final dataset of 2805 individuals of 214 colonies. We found no relationship between activity during assays and metabolic rate and activity was therefore not included in downstream analyses.

Ventilation patterns

Insects are known to exhibit three forms of gas exchange: discontinuous gas exchange cycle (DGC), cyclic gas exchange and continuous gas exchange. We calculated three metrics to indicate DGC occurrence and frequency. We produced a binary (0,1) categorical value for whether DGC was being exhibited by a species. We then calculated the proportion of individuals per species for each site and plot that were conducting DGC, which ranged from 0 (no individuals exhibiting DGC) to 1 (all individuals exhibiting DGC). We then calculated ventilation frequency per hour (VF), for those individuals exhibiting a DGC pattern. Ventilation frequency was determined by first counting the number of complete closed and open phases in an individual’s VCO₂ trace. Then multiplying this number by six (i.e., six 10-minute periods in an hour) to give vent frequency in cycles per hour.

Microclimate variables

We modelled microclimate at each site to represent the thermal and hydric environment which ant species directly experience in the field. We estimated hourly temperature and relative humidity for the 15 years preceding the study and calculated mean annual microclimate temperature and vapor pressure deficit (VPD), with VPD representing the aridity gradient using NicheMap R (Kearney et al. 2020).