personalities. In this experimental study we tested whether naturally occurring predation causes personalitydependent mortality of a marine isopod (Idotea balthica), which could maintain personality variation in nature. Moreover, as isopods are known to have sex-differences in behaviour, we were interested in whether personalitydependent predation was sex-specific. We also hypothesised that predation pressure among personality types could vary according to habitat type, as it has been shown in correlative studies that habitat may influence personality variation. We used natural predator (European perch Perca fluviatilis) of I. balthica and studied relative mortality of males and females with a different personality types in laboratory settings with two different habitats. We found that survival in males was lower than in females for high active individuals. Moreover, survival under predation was linked to body size differently in females and males. This, however, depended on personality class as larger size was advantageous for low-active males and middle- and high-active females. Conversely, smaller size was advantageous for low-active females and middle-active males. Size did not affect survival in high-active males. Our results suggest that predation can encourage life-history differences between sexes leading to different optimal life-history strategies and also maintains consistent activity for both sexes.