A multi-informant and multi-measure short-term longitudinal study of the association between subtypes of aggression and peer victimization was conducted in an early childhood sample (M = 44.36 months; SD = 11.07; N = 120). Observational and teacher report measures demonstrated appropriate reliability and validity as well as stability across an academic year. Concurrent associations revealed that observed relational aggression was uniquely associated with teacher reported relational victimization and observed physical aggression was uniquely associated with teacher reported physical victimization. Prospective findings indicated that observed relational aggression predicted increases in teacher reported relational victimization for girls only, controlling for the variance associated with physical aggression, prosocial behavior, physical victimization, and gender. Peer rejection partially mediated the association between observed relational aggression at time 1 and teacher reported relational victimization at time 2. Ways in which these and other prospective findings extend the extant literature are discussed.