Quantifying organismal sensitivity to heat stress provides one means for predicting vulnerability to climate change. Birds are ideal for investigating this approach, as they display quantifiable fitness consequences associated with behavioural and physiological responses to heat stress. We used a recently developed method that examines correlations between readily-observable behaviours and air temperature (Tair) to investigate interspecific variation in avian responses to heat stress in seasonally hot, arid regions on three continents: the southwestern United States, the Kalahari Desert of southern Africa and the Gascoyne region of Western Australia. We found substantial interspecific variation in heat dissipation behaviours (wing-drooping, panting, activity-reduction, shade-seeking) across all three regions. However, pooling the data revealed that little of this interspecific variation was systematically explained by organismal traits (foraging guild, diet, drinking dependency, body mass, or activity levels) at the scale we tested. After accounting for phylogeny, we found that larger birds engaged in wing-drooping behaviour at lower Tair and had lower activity levels at high Tair compared to smaller birds, indicating an effect of body mass on heat dissipation behaviour (HDB). In the Kalahari, reliance on drinking was correlated with significantly lower Tair at which panting commenced, suggesting a key role of water acquisition in HDB in that region. Birds also tended to retreat to shade at relatively lower Tair when more active, suggesting a behavioural trade-off between activity, heat load, and microsite selection. Our results imply that the causes underlying interspecific variation in heat dissipation behaviours are complex. While the variation we observed was not systematically explained by the broad scale organismal traits we considered, we predict that the indices themselves will still reflect vulnerability to potential fitness costs of high air temperatures. Further research is needed on a species-specific basis to establish the functional significance of these indices.