The aim of this study was to investigate whether, and how, extreme environmental influence, such as a war situation, can affect the development of children's aggressive and prosocial behaviour. A combined longitudinal–cross-sectional design was used in the study. Subjects were preschool children whose behaviour was rated on the Children's Aggressive and Prosocial Behaviour Rating Scale (Žužul, Keresteš and Vlahović-štetić, 1990) by their teachers. At the beginning of 1991 (before the war in Croatia started) the behaviour of two groups of children (one five years old and one six years old) was assessed. One year later, the younger group was assessed again. Another group of five-year-old children was also included and rated at the same time. Data were analysed according to age, sex and the time of measurements. The analyses of longitudinal data and cross-sectional comparisons support the conclusion that the war had a strong impact on the development of prosocial behaviour. In this period, children's prosocial behaviour had increased, while aggressive behaviour had not changed. Results are discussed in terms of developmental changes and the effects of war on children's behaviour.