Samples of foremilk were collected from 261 clinically normal glands of 150 ewes, and tested using the California mastitis test (CMT). Further samples were collected from 195 of these glands for determination of automated somatic cell counts (SCC), and from 60 of these glands for bacteriological assessment. The sensitivity and specificity of CMT for detecting samples with SCC above different threshold levels and for CMT and SCC in determining bacteriological status were evaluated using two-graph receiver operating characteristics (TG-ROC). Milk samples were obtained subsequently from ten CMT positive, and five CMT negative first- and second-lactation ewes. Samples were cultured using a variety of media, incubation temperatures and atmospheric conditions, immediately after collection, and 1 week after storage at 4°C and −21°C. Results suggested that CMT is best used as a diagnostic test for ovine subclinical mastitis (SCM) with a cut-off of 3 (distinct gel formation), and that automated SCC thresholds of >1200×103 cells/ml are appropriate, especially where low prevalences are expected (e.g. <5%). Additionally, this study showed that routine bacteriological methods were appropriate for isolation of most species of pathogen responsible for ovine SCM, but storage of samples prior to culture, either at 4°C or −21°C, was detrimental to the isolation of several of these organisms.