SUMMARYThe suitability of serological surveys of roe deer (Capreolus capreolus) in determining the spread of tick-borne encephalitis virus (TBEV) was tested in a south German area with a low risk of TBEV infection to humans. Sera obtained from 192 hunted roe were screened by an haemagglutination-inhibition test (HAI) and in an ELISA developed in our laboratory. Those found positive were tested in a neutralization test (NT). Fifty (26-0%) sera reacted positive by ELISA and 43 (86-0%) of these were confirmed by HAI or NT. Forty-seven (24-5 %) samples were positive by HAI, 44 (93-6 %) of which were also positive in NT or ELISA. Only insignificant increase of the antibody prevalence with age (P = 0 17 for HAI antibodies) suggests that most infections occur at an early age in scattered natural foci. The antibody prevalence in females was lower than in males (OR = 0-63; P = 0-02 for HAI antibodies). In determining the distribution of seropositive roe we increased the sample size to 235 sera. No antibodies were detected in 56 (23-8 %) sera collected in the eastern third of the county. The areas of high antibody prevalence in roe match those in which humans have been infected. We conclude that serosurveys of roe deer are useful in marking out areas in which humans face the risk of infection, provided that an adequate number of sera, preferably from males, is available.