Even with the success of antibiotics, the rapid evolution of multiple antibiotic resistances constitutes an increasing health problem. We are studying the importance of genetic interactions in the evolution of antibiotic resistance. In particular the costs of multiple resistance and the rate of mutations which can compensate for resistance costs, need to be better understood in natural microbial ecosystems, which are characterised by a multitude of biotic interactions and rapid evolution. We are using modelling and in vivo experimental evolution to address this issue.