Isotype- and subclass-specific indirect enzyme-linked immunosorbent assays were developed to detect parainfluenza-3 virus-specific IgG1, IgG2, IgM, and IgA responses. Sera were treated with protein G- agarose prior to testing for specific IgM and IgA to eliminate the possibility of false-positive results due to IgM-rheumatoid factor and to remove interisotypic competition due to specific IgG. IgM and IgA absorbance values were expressed as a percentage of the absorbance values of positive reference sera included on each plate (S/P%), and respective positive/negative threshold values of 15.0% and 28.0% were determined. The mean interval between experimental infection of 3 calves and initial detection of specific IgG1 and IgG2 responses was 8.0 and 9.3 days respectively, rising rapidly to an initial plateau 13.7 and 11.0 days postinfection (dpi). Reinfection of these calves at 30 dpi resulted in further rapid increases, with higher plateau values reached 13.0 (IgG1) and 13.7 (IgG2) days later. The mean interval between infection and the first positive IgM and IgA responses was 6.7 and 12.3 days, respectively. IgM S/P% values peaked at 13.0 dpi, with all 3 calves showing a secondary anamnestic response to reinfection, peaking 4.7 days later. The IgA response to initial infection was weak, with only 2 calves showing an obvious peak response at 15.0 dpi. A strong anamnestic IgA response to reinfection occurred in 2 calves, with a peak response 9.5 days later. Apparent biphasic and triphasic IgM and IgA responses were evident in some calves. Acute and convalescent serum samples from 80 calves involved in 17 outbreaks of respiratory disease were tested for specific IgM and IgA. Positive IgM results were detected in 15 outbreaks, with 71 sera from 44 calves testing positive. Although IgA-positive results were detected in the same 15 outbreaks, only 42 sera from 31 calves were positive. In a previous study, seroconversion was detected in 21 of these calves from 10 outbreaks. Thus the diagnostic potential of the assays was in the order IgM . IgA . seroconversion. The correlations between IgM and IgA, IgM and seroconversion, and IgA and seroconversion results for each calf were 73.8%, 58.8% and 62.5%, respectively. Infection with parainfluenza-3 virus (PI3V) is ubiq- uitous in dairy and beef herds in Northern Ireland, with serologic evidence of infection on 97.5% of farms. 19 In a serologic survey of outbreaks of bovine respiratory disease in Northern Ireland, 10 seroconver- sion to PI3V was detected in 13.7% of calves sampled, second in prevalence only to bovine respiratory syn- cytial virus (BRSV). However, serodiagnosis of respi- ratory disease based on seroconversion requires a min- imum of 2-3 weeks because of the requirement for both acute and convalescent serum samples. In addi- tion, the presence of maternally derived antibody (MDA) suppresses the humoral immune response, es- pecially of IgG. 9,14,18 IgM-specific assays have the po- tential to overcome both of these problems and have been shown to be superior to the detection of serocon