AFTER infection with a group-A streptococcus, antibodies to a wide variety of extracellular and somatic antigens can be detected in the patient's serum. One of the factors that may influence the magnitude of these responses is the serotype of the infecting strain. It is well established that patients with streptococcal pyoderma have considerably lower titres of anti-streptolysin 0 (ASO) than do patients with streptococcal sore throat, but that titres of anti-deoxyribonuclease B (anti-DNAase B) are often very high in cases of skin infection (Anthony, Perlman and Wannamaker, 1967;Dillon and Reeves, 1969). It is not known to what extent the site of infection is responsible for these differences, but the serotypes that cause pyoderma are in general quite different from those that commonly cause disease of the respiratory tract (see Wannamaker, 1970).Antibody against M protein confers type-specific protection against reinfection with streptococci of the same serotype, and Denny, Perry and Wannamaker (1957) and Siegal, Johnson and Stollerman (1961) found that at least 70% of patients had type-specific antibody to the infecting strain after recovery from streptococcal sore throat. However, the outbreaks that these workers investigated were all caused by M types that did not produce opacity factor (OF), i.e., " OF-negative M-types " (Maxted and Widdowson, 1972). Information on the antibody response in man to infection with OF-positive serotypes is lacking, but vaccines of these types are generally poorly antigenic in rabbits (Gooder, 1961 ;Top and Wannamaker, 1968;Widdowson, Maxted and Grant, 1970).We have shown (Widdowson, Maxted and Pinney, 1971a) that titres of antibody against streptococcal M-associated protein (MAP) are almost invariably high in patients with rheumatic fever, but much less often raised in patients with acute streptococcal glomerulonephritis and in sporadic cases of sore throat. Later, however, we investigated an outbreak of rheumatic fever due to M-type 5 in which a considerable proportion of the patients with uncomplicated sore throat also developed high titres of anti-MAP (Widdowson et al., 1974). We have therefore studied the antibody responses of patients in a number of outbreaks due to known streptococcal types; the sera were examined for ASO, anti-DNAase B, anti-MAP, and M antibody to the supposed infecting serotype. We also examined the sera of several other groups of patients suffering from the non-suppurative sequelae of streptococcal infection of the throat or of the skin, and the sera of groups of supposedly normal persons.