There is wide divergence of opinion between physicians regarding the use of serological measures for the diagnosis and treatment of erythema migrans, the hallmark of Lyme borreliosis. We studied the outcome of an enzyme immunoassay and immunoblot (Western blot) used on the sera of patients who had suffered tick bite and erythema migrans, and had been subsequently treated with various antibiotics. Ninety-nine consecutive patients presenting with erythema migrans after tick bite were prospectively recruited at the outpatient department of two Vienna City hospitals and at the consultation office for Lyme borreliosis of the Institute of Hygiene. University Vienna. Blood samples were taken before antibiotic treatment and 3 and 6 months thereafter. Blood samples from 100 blood donors served as controls. Antibodies against Borrelia burgdorferi sensu lato were determined by enzyme immunoassay (IgG and IgM EIA) and by IgG immunoblot. The latter was performed with isolates of B. alzelii (H2) B. burgdorferi sensu stricto (Le) and B. garinii (W) from Austrian patients. The 4 interpretation criteria for immunoblot results were: A (3 bands out of 8), B (2 bands out of 9), C and D (1 band out of 6). In all patients, the erythema resolved within the treatment period. No complications secondary to the borrelia infection were registered. After treatment there was no significant change in titre, nor was there a difference in the immunoblot pattern between the first, second and third serum samples. Serum antibodies to B. burgdorferi were positive by EIA in 22.9% (IgG) and 2.5% (IgM). Immunoblot results offered by borrelia species and by the interpretation criteria, ranging between 8.3% (criterion A, strain Le) and 44.2% (criterion D, strain H2). By EIA, control samples were IgG and IgM positive in 5% and 1%, respectively. Positive immunoblot results with strain H2 were found in 9%, 13%, 18%, and 20% by the criteria A through D respectively. After antibiotic treatment of erythema migrans the immunological response appears to be abrogated. Thus, serological results are not supportive for the diagnosis of erythema migrans, not will they retrospectively prove successful antibiotic treatment of borrelia infection.