Sera from swine experimentally or naturally infected with Mycoplasma hyopneumoniae (the etiological agent of mycoplasmal pneumonia of swine, MPS) were tested by the indirect hemagglutination assay (IHA), the enzyme-linked immunosorbent assay (ELISA) and the complement fixation (CF) test. The IHA detected antibody at comparable times and levels to the other 2 serological tests following experimentally-induced infection. In the late antibody response (greater than or equal to 86 days post-infection), the ELISA titres were higher than either the IHA or the CF test. The IHA appeared least satisfactory when it was used to test sera from commercial swine herds. When 1000 sera were tested, the IHA was positive for only 30 (22%) of 135 sera which were positive by the ELISA and the CF test. The IHA titres were low; 20 of the 30 sera had a titre of only 10. The end-points for the IHA were difficult to read for sera of this low titre. The relationship between positive IHA results for the herd sera obtained at necropsy, and the occurrence of gross or microscopic lesions typical of MPS was poor (41 and 50% agreement, respectively). An agreement of 39% was noted between positive IHA results and the localization of mycoplasmal antigens by an indirect immunofluorescence (IIF) test. However, IHA results correlated significantly (P less than 0.05) with gross and microscopic lesions, but not with the IIF test. No significant correlation was noted between the IHA (or the other 2 serologic tests) and the cultural isolation of M. hyopneumoniae or M. flocculare. On the basis of these results, the IHA appears to have limited promise as a practical test for the diagnosis of MPS in commercial swine herds because of the low titres observed, poor correlation of the IHA and other indicators of MPS, the necessarily subjective determination of end-points, and other inherent technical limitations of the test.