The potential value of electrical conductivity of milk as a screening test for subclinical mastitis was evaluated. Conductivity of foremilk and of postmilking strippings from 368 quarters of 92 cows was measured. Infection status of quarters was determined by bacteriological analysis of strict foremilk samples. Infections were classified as by primary or secondary pathogens, depending on the importance of the isolated organism as a mastitis pathogen. Somatic cells were counted on foremilk samples. Milk conductivity increased with infection. Conductivity of postmilking strippings was higher than that of foremilk in samples from quarters infected by primary pathogens. By thresholds which correctly classified at least 90% of normal quarters, accuracy of identifying primary pathogen infections by absolute conductivity was 62.8 and 96.2% with foremilk and postmilking strippings. Differential conductivity and combination of absolute and differential methods also were evaluated with the latter being the most effective. Number of quarters with elevated conductivity of postmilking strippings tended to be higher when somatic cell count was greater than 500,000/ml in both normal and infected groups. Conductivity of milk seems to hold promise as an indicator of subclinical mastitis.