California Mastitis Test scores and bacteriological analysis of quarter foremilk samples from 133 cows of three herds were used to determine which quarters or cows to treat in a selective dry cow therapy program. Effect of score, week of the test, and type and number of intramammary infections were examined. Score was independent of the type of intramammary pathogen infection. When all positive reactions were considered, approximately 80% of the infections by major pathogens were diagnosed on a single milk sample taken at either 8 or 4 wk before drying off. At 8 and 4 wk, 13 and 23% of samples from uninfected quarters or quarters infected with minor pathogens had positive reactions. Selective treatment of all cows that had one or more positive quarters led to treatment of twice as many quarters than if only positive quarters were treated. The most discriminating, simplest, and economic method of selective treatment was one test on quarter foremilk samples collected 8 wk before the expected dry-off and dry treatment of all positive mammary quarters.