During a period of 9 years (December, 1980, through November, 1989), 407 Group A streptococcal strains were isolated from 294 children with acute rheumatic fever and 303 of their family contacts, 234 children with acute post-streptococcal glomerulonephritis and 242 of their family contacts and 219 children with uncomplicated Group A streptococcal pharyngitis. Of the 407 straisn 216 (53%) were M and/or serum opacity factor typable, 143 (35%) were only T typable and 48 (12%) were nontypable. Throughout the period of study the M12 and M1 were the most prevalent strains; however, important changes among the prevalent strains were observed. Although the study started in 1980 the serotypes M18, M81, M3, M15 and M58 made their first appearance 7 to 9 years later. These findings show the value of long term studies in detecting the changes in the prevalence of streptococcal strains in the community. M18 was isolated from three children with nephritis but not from children with rheumatic fever; this association has not been reported before. M12 was isolated from 26% of the nephritic children and their families vs. 7% from the rheumatic children and their families (P < 0.05) vs. 17% from children with uncomplicated streptococcal pharyngitis. M49 was isolated from 7% of the nephritic children and their families vs. none from rheumatic children and their families vs. 1.4% from children with uncomplicated streptococcal pharyngitis. These findings support the concept of nephritogenicity of some streptococcal strains. the serotypes M9, M33, M5, and M29 were exclusively isolated from rheumatic children and their family contacts, whereas the serotypes M49, M4, M81 and M18 were exclusively isolated from nephritic children and their family contacts. These findings are in favor of the concept of rheumatogenicity. Our data are compared with those of similar recent studies.