SYNOPSIS Although pyelonephritis is a common disease, it escapes clinical detection in an undesirably high proportion of patients. The present unsatisfactory diagnostic position would be much improved by widespread screening of patients by simple yet reasonably accurate methods. Bacterial counts by the pour-plate technique and estimates of the white cell excretion per hour or day, while undoubtedly of diagnostic value, are probably unsuitable for use on a wide scale. In an attempt to find more convenient procedures a simplified stroke-plate method of bacterial counting and a simplified quantitative white cell count method were devised and applied to over 1,000 mid-stream urine samples from 398 patients. Good correlation was obtained between the simpler stroke-plate method of bacterial counting and the more time-consuming pour-plate method. The quantitative white cell procedure was a much more sensitive index of pyuria than wet-film microscopy, and comparison with the bacterial count results showed that it gave a useful indication of urinary infection. It is suggested that a quantitative bacterial count should replace non-quantitative culture methods when urinary infection is suspected and that the quantitative white cell count should be performed as a routine part of the initial clinical and laboratory assessment of all patients, followed by a bacterial count if pyuria is revealed. Experience has shown that routine urine microscopy by a precise method leads to the detection of many cases of occult urinary infection.Estimates of the incidence of pyelonephritis on the basis of necropsy evidence have ranged from 6 18 % (Brod, 1956) to 20% (Rhoads, Billings, and O'Conor, 1952). Many of these cases are undiagnosed, probably because obvious symptoms of urinary tract inflammation, such as dysuria and frequency of micturition, are often absent or slight, particularly if the infection is chronic. Thus, Kleeman, Hewitt, and Guze (1960) report that out of 629 cases of significant pyelonephritis seen at necropsy in a five-year period, only 104 (16.6%) were diagnosed clinically, although they stress that retrospective study of the case records indicated that there was suggestive information in 70% of all cases.