Sera from 775 adult milking cattle on 13 farms in Friesland and from 359 of the same cattle taken during the following year, were examined for evidence of infection with M johnei by the complement fixation, fluorescent antibody, haemagglutination, haemagglutination-lysis and immunodiffusion tests. Eighteen animals, not clinically affected but which had serologically positive tests were slaughtered and M johnei was isolated in cultures from 16. The fluorescent antibody test was as sensitive as the complement fixation test. The other tests were relatively insensitive and would only be of value as confirmatory tests in the presence of reactions to the first two tests. It is concluded that the use of the fluorescent antibody and the complement fixation test together was of value in detecting sub-clinical Mycobacterium johnei infection.