Quantitative C-reactive protein (CRP) measurement has become increasingly valuable as a test for rapid diagnosis of infections in hospital medicine. CRP has not obtained the same importance in primary health care. This has, at least partly, been due to methodological difficulties, with no simple or rapid tests with quantitative results available. A new immunometric semi-quantitative assay, NycoCard CRP, has recently been developed. CRP was analysed at the local health centres by the new assay in 288 consultations where patients came because of infections. Parallel CRP values were obtained by an established reference method. The two procedures had an acceptable correlation (r = 0.85). The primary care doctors also registered the clinical information they obtained from each CRP result. CRP was helpful in indicating the presence, or absence of bacterial infection in more than half the consultations due to new infections. CRP was thought to yield more clinical information than the erythrocyte sedimentation rate in almost every case.