Experimental infections were produced in 78 quarters of 17 cows by the infusion of small numbers of a single strain of Staphylococcus aureus. In each single experiment three quarters in a cow were infected, with the fourth left as a control. At times varying from three to 60 days after the infusion of organisms, a standard intramammary antibiotic treatment was administered on a single occasion. A cure was arbitrarily defined as the absence of the organism in foremilk, from direct plating and replated incubated milk, together with return to normal somatic cell count levels as determined by an electronic counter. With these standardized conditions the effects of a number of cow associated factors on the outcome of the therapy were determined.Forty-three of the 78 quarters (55%) were cured by the standard treatment. Significant differences in percentages of quarters cured were found to be associated with the duration of infection before therapy, the lactation age of the cow, the length of time in lactation, somatic cell count in milk at time of treatment, the location of the quarter in the udder and individual cows. No significant effects on the outcome of the standard treatment were found associated with the number of bacteria in the secretion at the time of treatment, previous infection and cure in a quarter nor the season of the year in which treatment was given. Of the 35 quarters in which infection recurred following treatment, organisms were reisolated from 12 within four days, 18 between five and nine days, four between ten and 17 days and one after 28 days. From these data it is apparent that if, as has been suggested, models such as described are to be used for efficacy trials, standardization of some parameters is essential.