The requirements for the isolation of Mycobacterium avium subsp. paratuberculosis (Map) may be related to the strain-type [sheep (S)- or cattle (C)-type] and not to the host. The objective of this paper was to estimate and compare strain- and biological sample (faeces or pooled-tissue)--specific sensitivities (Ses) of two solid culture media, Herrold's egg yolk medium (HEYM) and Lowenstein-Jensen (LJ) medium, for the isolation of Map from Greek dairy sheep and goats. From 400 faecal samples collected from sub-clinically infected sheep and goats of four flocks and from 214 pooled-tissue samples (142 from sheep and 72 from goats) collected, at the abattoir, from >1-year-old routinely slaughtered animals, with gross pathology suggestive of paratuberculosis, we isolated 34 Map strains. Of those, by the IS1311 PCR, 18 were categorized into the C-type and nine into the S-type; seven were not typed. We used a Bayesian approach to estimate the strain-specific Ses. SeHEYM-C-faecal=17% (95% credible interval: 7, 40) was higher than SeHEYM-S-faecal=2% (0.3, 11). Also, SeHEYM-C-faecal was higher than SeLJ-C-faecal=4% (1, 12). In pooled-tissue samples, the strain-specific Ses did not differ between the two media.