Urine ammonia concentration is crucial to understanding and quantifying the kidney's response to metabolic acidosis. This test is generally not performed by clinical laboratories. The urine anion gap and osmolar gaps have been proposed as surrogate measures of urine ammonia in patients with hyperchloremic acidosis. We measured ammonium and other electrolytes in the urine of patients attending our renal disease clinic who did not have severe metabolic acidosis and compared the results with those calculated by standard formulae for the anion and osmolar gaps. We found no correlation between measured ammonium values and the anion gap and attributed this lack of agreement to the presence in urine of substantial amounts of unmeasured inorganic anions, which the formula fails to consider. There was significant correlation between measured ammonium and the osmolar gap but not good agreement between the absolute values provided by the 2 methods. Solutes including sulfate and phosphate were quantified in 24-hour urine collections and showed great variability with respect to measured chloride and estimated protein catabolism. We conclude from these studies that there is no substitute for the direct determination of urine ammonium when an accurate concentration is desired.