Conventional ribotyping was compared with the PCR amplification of the intergenic spacer region between 16S and 23S rRNA genes (PCR-RFLP ribotyping) when applied to the subtyping of sporadic Neisseria meningitidis strains. Thirty isolates out of a total of 36 meningococcal disease cases, reported as having occurred in a particular community over a 7-year endemic period, were analyzed by each of the methods. Only ribotyping with three restriction enzymes (EcoRI, ClaI and XhoI) gave acceptable discriminatory power for short-term epidemiological purposes. We conclude that conventional ribotyping is a suitable method for typing sporadic meningococcal strains and that it cannot be replaced by the more straightforward PCR-RFLP ribotyping method.