A total of 50 sheep originating from 15 Dutch farms with a known paratuberculosis infection in their cattle herd, but with no history of paratuberculosis infection in their sheep flock, were examined for infection with Mycobacterium avium subsp. paratuberculosis (Map). The sheep had been grazing on the same pastures as the cattle or on pastures fertilised with manure from these cows. The sheep were screened for paratuberculosis by serum biochemistry, serology and intradermal skin tests. At necropsy they were examined macroscopically, microscopically and bacteriologically for paratuberculosis. From 10 sheep, originating from eight flocks, Map could be isolated from various tissues but not from the intestinal contents, after an incubation period of 2.5-4 months. Six of these culture-positive sheep had no macroscopic signs of paratuberculosis at necropsy. Seven sheep were Map culture negative but showed macroscopic and microscopic lesions consistent with a paratuberculosis infection. Results of serology and skin tests did not correlate with the results of bacteriological culture. Serum concentrations of calcium, albumin and total protein of the infected, suspected and negative sheep were not different. These results indicate that a substantial number of the sheep examined were infected with Map. Even though this bacterium was not isolated from their faeces, the possibility that these sheep could have been shedding Map with their faeces below detection level or at a later stage of the disease cannot be eliminated. Map infected sheep should, therefore, be considered as a possible factor in the epidemiology of with Map infected cattle herds in The Netherlands. At necropsy bacteriological culture of Map should be performed on a routine basis to improve the diagnosis of paratuberculosis in sheep.