Four tests (agar gel precipitin, seroneutralisation alpha, seroneutralisation beta, and haemagglutination-inhibition) were used to detect antibodies to avian infectious bronchitis from 3 weeks to 40 weeks of age in four breeder flocks which were given four different vaccination programmes against infectious bronchitis. During the rearing period to 20-22 weeks of age, birds were kept on the ground in isolation. Then, most of them were housed conventionally during the laying period except for 25 birds from each flock which were kept in the laboratory in strict isolation and challenged at 28 weeks old. During the rearing period variations in antibody titres were recorded according to the vaccination programme and tests used; these are discussed. Irrespective of vaccines and programmes used, low titres were observed after vaccination, as measured by seroneutralisation and haemagglutination-inhibition tests, except in one flock in which infection with extraneous infectious bronchitis virus may have occurred. In conventionally reared birds in the laying period, irrespective of the previous vaccinal response, a progressive increase of antibody titres was recorded, as measured by seroneutralisation and haemagglutination inhibition tests and titres reached a high level; at each time the agar gel precipitin test gave positive results. It is suggested that the occurrence of natural infection may explain these results more readily than vaccine virus spread from bird to bird. In birds kept in strict isolation from their arrival in the laboratory at 20-22 weeks of age until challenge at 28 weeks of age, antibody titres decreased. After challenge, high antibody titres were recorded as measured by seroneutralisation alpha and beta and haemagglutination-inhibition tests in all four groups and they persisted at a high level until 40 weeks of age, when the study was terminated; however, the agar gel precipitin responses varied greatly according to groups. Protection also varied between groups and correlations between antibody titres and protection are discussed. It is confirmed that the haemagglutination inhibition test appears to be a very useful test in diagnosis.