A modification of the urine osmolal gap was evaluated as an estimate of urine [NH<sup>+</sup><sub>4</sub>]. We proposed that: Urine [NH<sup>-</sup><sub>4</sub>] =Urine osmolality – [2(Na<sup>+</sup> + K<sup>+</sup>) + urea + glucose]/2 Spot urine samples were collected from normal volunteers and from individuals with ketonuria; the modified urine osmolal gap as well as two other previously described estimates of urine [NH<sup>+</sup><sub>4</sub>] were compared with measured urine [NH<sup>+</sup><sub>4</sub>]. There was a significant positive linear correlation between the urine [NH<sup>+</sup><sub>4</sub>] and the modified urine osmolal gap in normal volunteers (r = 0.81; p < 0.01) and in individuals with ketonuria (r = 0.93; p < 0.001). The originally described urine osmolal gap greatly overestimated the urine [NH<sup>+</sup><sub>4</sub>] but also showed a significant correlation. The urine anion gap was not a valid estimate of urine [NH<sup>+</sup><sub>4</sub>] within the range of values measured in our subjects. The modified urine osmolal gap is an improvement over previously described estimates of urine [NH<sup>+</sup><sub>4</sub>] and can be used as a single calculation in place of the other two.