In order to better define the specificity of Lyme disease (LD) seropositivity and seronegativity during a study of Lyme arthritis in a pediatric rheumatology population, three groups of controls are in the process of being developed. Children with other rheumatic disorders. Siblings and parents of children with Lyme arthritis. Individuals with intense environmental exposure to LD. Sera from each have been tested for antibody to Borrelia burgdorferi (Bb), by indirect immunofluorescence assay (IFA), enzyme linked immunosorbent assay (ELISA), or both. Of the 57 children with rheumatic disorders of whom 25 had JRA, 4 (7%) reached a border zone, non-specific level of 1:64 by IFA, and 53 sera were non-reactive. 7 household relatives of LD cases had non-reactive sera. The 42 intensely Bb-exposed individuals all had non-reactive sera. The fact that none of the 106 controls attained a clinically significant level of antibody is a general indication of the specificity of the serologic tests. The use of selected control groups such as these might potentially solve the ambiguities in clinical-laboratory correlations in Lyme arthritis.