Two hundred and thirty-two cows were assigned alternately to complete dry cow therapy (infusion in all quarters on the day of drying off) or selective therapy (infusion in all quarters if a history of mastitis, California Mastitis Test score of +2 or +3 in any quarter, or if cell counts from bucket milk samples as determined by the membrane filter-deoxyribonucleic acid procedure were above 500,000 cells/ml). A dry cow product containing 10(6) units of procaine penicillin G and 1 g of dihydrostreptomycin in a slow release base was used. A 1% iodophor teat dip was used throughout the experiment. Infections of Staphylococcus aureus, Streptococcus agalactiae, other streptococci, and gram negative rods were eliminated from 85.4% of the infected quarters with complete therapy and 88.2% of the infected quarters with selective therapy. New infections occurred in 3.1% of quarters with complete therapy and in 6.5% of the quarters with selective therapy. Incidence of mastitis following the dry period was less with complete therapy compared to selective therapy (4.6% vs. 7.8% of the quarters). Selective therapy was as effective as complete therapy in eliminating existing infections. Complete therapy would be the choice in situations where new infections in dry period are of concern.