Environmental stress reduces species growth rates, but its impact on the structure and function of microbial communities is less clear. Using salinity stress as an example, we recently demonstrated that increasing salinity stress shifts community composition towards species with higher growth rates. As a result, the mean community growth rate is more robust to increasing stress than the growth of individual species. We showed this by propagating natural aquatic communities at multiple salinities and mapping the observed diversity onto the measured salinity performance curves of >80 bacterial isolates. We further validated these results with pairwise species competitions and in metagenomic data of natural communities sampled from estuarine environments. A minimal ecological model with mortality and stress-dependent growth rates could recapitulate the observed robustness, which suggests that these results extend to other environmental stressors. In this talk I will further highlight what I think the evolutionary implications of these ecological dynamics may be.