Faecal samples from 186 dairy cows representing ten commercial dairy herds with sporadic clinical paratuberculosis (group A), and from 100 dairy cows from herds without a history of paratuberculosis (group B) were cultured for Mycobacterium avium subspecies paratuberculosis (MAP). Two different decontamination methods, a NaOH/oxalic acid method and treatment with 0.75% hexadecylpyridinium chloride (HPC) were performed prior to inoculation of Loewenstein-Jensen agar slants with and without mycobactin. Cultures were incubated for 16 weeks. Acid-fast staining bacteria were identified as MAP on the basis of mycobactin dependency and by PCR-RFLP analysis of the IS 1311-insertion element of M. avium. MAP was grown from 15 out of 186 group A animals (8.1%) whereas faecal culture for MAP was consistently negative in group B. The growth rate of MAP was significantly higher (8.1% vs. 1.6%) and the contamination rate of cultures was significantly lower (17.6% vs. 21.5%) in faecal samples decontaminated with NaOH/oxalic acid than with HPC-treated faecal samples (p<0.01, McNemar's test). Atypical mycobacteria which were grown from 46.8% of NaOH/oxalic acid treated specimens were not obtained from any of the HPC-treated samples. A commercial ELISA with MAP-lipoarabinomannan as the antigen was used to detect MAP-antibodies in unabsorbed sera from all animals. The percentage of ELISA-positive cows was 16.8%. Overall agreement between antibody detection and MAP-positive faecal culture was 15.4%.