To test whether longer duration of treatment of asymptomatic bacteriuria in old age could improve the efficacy of cotrimoxazole therapy, three regimens were given to 75 ambulant bacteriuric residents of a retirement home, aged greater than or equal to 68 years. The groups and regimens were: A:23 subjects (160/800 mg b.i.d. orally x 3 days). B: 24 subjects (160/800 mg i.m. x 10 days); C: 28 subjects (160/800 mg b.i.d. orally x 20 days). One week, one month and five months post-therapy urines were negative in 78.3% vs 52.9% vs 42.9% of group A, in 54.2% vs 56.5% vs 50% of group B and in 57.1% vs 60.7% vs 68% of group C subjects respectively. The data indicate that: 1) the efficacy of any schedule is only moderate irrespective of the presence of antibody-coated bacteria in urine; 2) a 3-day course appears more effective at one week post-therapy; 3) at one and five months greater than or equal to 50% of the subjects were infection-free, the 20 day treatment resulting in fewer failures; 4) subjects with long-term eradication had no mobility problem, low serum creatinine and a normal urinary tract as seen by ultrasound.