The effect of dry-cow therapy was evaluated on the basis of the culling rate, occurrence of clinical mastitis, mean of cow milk somatic cell counts (CMSCC), and milk yield, in a trial including 608 cows. The cows were randomly divided into four groups: control group A (92 cows), group B treated with placebo (base ointment of Benestermycin (Leo) without antibiotics) (105 cows), one intramammary dose per infected quarter, group C treated with Benestermycin (Leo), one intramammary dose being infused in each infected quarter (196 cows); and group D treated with Leocillin with dihydrostreptomycin (Leo); one intramammary dose being used every second day, on four occasions per infected quarter (215 cows). The study included infected cows. If less than three of the quarters of the udder were diagnosed as having mastitis at first sampling, only infected quarters were treated. Otherwise, all quarters were treated. Multivariable analysis showed no significant effect of therapy on culling rate. The control groups (A + B) had a greater increase of cows having at least one case of clinical mastitis compared to the therapy groups (C + D), (from 0.26 to 0.57 in controls comparing to 0.38 to 0.43 in therapy groups). The difference between control and therapy groups during lactation was close to significant both before and after in the lactation after therapy (P < 0.10). The multivariable analysis showed a significant benefit of dry-cow therapy of 0.409 In unit in geometric mean CMSCC (corresponding to 125,000/ml), 200,000/ml in weighted CMSCC and 189 kg milk yield per lactation. According to these results selective dry-cow therapy for cows included in this study is recommended.