The fluorometric scopoletin-horseradish peroxidase method has been modified for field determinations of hydrogen peroxide concentrations in ground waters. Standard additions calibration compensates for background fluorescence and inconsistent stoichiometry of the fluorescence quenching reaction due to interferences by the matrix. The detection limit, defined as the blank plus three standard deviations, ranged from 3.6 to 44.6 nM. However, this limit was more an indication of the difficulty of preparing peroxide-free water than the actual limit imposed by the sensitivity of the method for the peroxide contamination introduced with the reagents. For 111 field determinations the weighted average uncorrected hydrogen peroxide concentration was 20.2 nM and the pooled standard deviation was 7.7 nM. The average of 45 field blanks was 7.8 nM with a pooled standard deviation of 5.2 nM. (Copyright (C) 1987 American Chemical Society.)