A total of 4470 pregnant women were screened for bacteriuria by the dipslide method and significant growth found in 226 (5.1%). In 198 cases the urine was re-examined, in 119 by using suprapubic aspiration or catheterisation (62 (52%) samples contained bacteria) and in 79 by using midstream urine samples (26 (33%) samples contained greater than 10(8) colony forming units/1), showing the maximum prevalence of confirmed bacteriuria to be 2.6%. Overt urinary tract infection developed later in four of 80 patients with proved bacteriuria who had been given antibiotics, in one of eight untreated patients with bacteriuria, in one of 110 patients with unconfirmed bacteriuria, and in one of 226 non-bacteriuric controls. A history of urinary tract infection was given by 18% of controls and 42% of women with confirmed bacteriuria. Screening for bacteriuria and treatment with antibiotics to prevent later overt infection is expensive. Whether it is worth while and cost effective depends largely on the prevalence of bacteriuria in the local population and the proportion who develop overt infection. The screening and treatment programme reported here appeared to prevent only six cases of overt infection.