In a double-blind prospective study of 200 sequential urine specimens the sediment count of leukocytes in the centrifuged urine (white blood cells per high power field) was compared to a chamber count of leukocytes in uncentrifuged urine (white blood cells per microliter.). There was good correlation (coefficient of correlation 0.783, sensitivity 91.9 per cent, specificity 97.6 per cent and efficiency 96.6 per cent) between the more precise chamber count and the more commonly performed sediment count if the methodology of the sediment count was standardized. In a double-blind prospective study the results of the sediment count for leukocytes and erythrocytes were compared to the leukocyte esterase and hemoglobin dipstick results of urine specimens from 1,346 adults who underwent multiphasic screening. The dipsticks were found to be sensitive to physiologic limits for leukocytes and erythrocytes, with only 0.9 per cent false negative results for each. Formed elements in the urine not detectable by dipstick, such as casts and crystals, were present in 3 per cent of the specimens. Among patients who had significant pyuria, hematuria or formed elements not detectable by dipstick chemical urinalysis, no significant pathological condition was detected upon retrospective review. Because the chemical dipstick is not quantitative and because the sensitivity of the dipsticks resulted in many false positive findings compared to the sediment count (red and white blood cells 16.4 and 13.2 per cent, respectively) a protocol is offered in which results of screening urine specimens that are positive on dipstick culture would be confirmed by a properly performed microscopic urinalysis. This protocol as applied to an adult screening population would be an accurate, cost-effective method of urine testing.