Two cases of meningitis due to Streptococcus suis in humans are reported here. A butcher and an abattoir worker were referred to a health centre in Castellón (Spain) with fever and symptoms of meningitis. After adequate treatment, a slight hipoacusia persisted as sequelae in both cases. Colonies of S. suis group R, serotype 2 and phenotype MRP+EF+ were isolated from cerebroespinal fluid. Epidemiological studies showed that both workers had in common the handling of pork meat of slaughtered healthy pigs from three closed farms. A study of the tonsils from apparently healthy, slaughtered pigs was carried out. A total of 234 tonsillar samples were obtained and 81 strains of S. suis were isolated from them. Serotype 2 appeared to be the most frequent (50.6%), and the analysis for phenotype showed a high percentage of tonsillar strains with the phenotype MRP+EF+ (35.9%). The humans and 28 tonsillar swine strains showed a similar profile (S. suis group R, serotype 2 and phenotype MRP+EF+). A total of 26 of the swine isolates were analysed by ribotyping using EcoRI. The human strains showed the same six-band hybridization pattern that shared five bands with the pattern most frequently shown by most of the tonsillar N. suis group R, serotype 2 and phenotype MRP+EF+ strains, differing only in the lightest, faintest band which was slightly less anodical in human (> or = 1.8 kb) than in swine (approximately 1.8 kb). From these results, both groups of strains, humans and porcine, showed differences; how can these differences in the pattern of ribotyping be explained if they should have the same origin? Is it possible that they have undergone an adaptation to the new host or perhaps the modification is due to other unknown causes? Further studies in this area are required in order to answer these questions.