The purpose of this study was to evaluate Salmonella ELISA tests and a culture technique to determine Salmonella status using samples collected from commercial herds. Faecal samples from 15 finisher pigs on each of 40 swine farms were cultured for Salmonella. Sera from the same 600 pigs were analysed for the presence of antibodies to Salmonella by means of two different ELISA kits Salmotype (Salmotype Labordiagnostik, Leipzig, Germany) and IDEXX (Herdchek* Salmonella; IDEXX Laboratories, Schiphol-Rijk, Noord-Holland, The Netherlands). The Salmotype ELISA test demonstrated a moderate ability to differentiate culture-positive pigs from culture-negative pigs while IDEXX was relatively poor in classifying those pigs correctly (The area under the curves were 0.79 and 0.65 for Salmotype and IDEXX respectively). The maximum value of sensitivity plus specificity was gained at the cut-off optical density (OD) > or =25% for Salmotype (sensitivity 0.65, specificity 0.84) and at the OD > or =9% for IDEXX (sensitivity 0.59, specificity 0.69). The maximum herd sensitivity and herd specificity was 0.64 (95% CI 0.49-0.79) and 0.72 (95% CI 0.59-0.86) for the Salmotype test and 0.73 (95% CI 0.59-0.87) and 0.62 (95% CI 0.47-0.77) for the IDEXX. Culture and the ELISA tests showed fair agreement at the herd level (kappa=0.3, P<0.05). Likewise there was significant disagreement between the two ELISA tests at the pig level, but very good agreement between the two ELISA tests (kappa=0.8, P<0.0001) at the herd level. The coating antigens in Salmotype and IDEXX represent only 48% of the antigens of Salmonella isolated in our study and need to be revised based upon the Salmonella serovar distribution in Ontario.