The bacterial denitrification method for isotopic analysis of nitrate using N2O generated from Pseudomonas aureofaciens may overestimate delta(15)N values by as much as 1-2parts per thousand for samples containing atmospheric nitrate because of mass-independent O-17 variations in such samples. By analyzing such samples for delta(15)N and delta(18)O using the denitrifier Pseudomonas chlororaphis, one obtains nearly correct delta(15)N values because oxygen in N2O generated by P. chlororaphis is primarily derived from H2O. The difference between the apparent delta(15)N value determined with P. aureofaciens and that determined with P. chlororaphis, assuming mass-dependent oxygen isotopic fractionation, reflects the amount of mass-independent O-17 in a nitrate sample. By interspersing nitrate isotopic reference materials having substantially different delta(18)O values with samples, one can normalize oxygen isotope ratios and determine the fractions of oxygen in N2O derived from the nitrate and from water with each denitrifier. This information can be used to improve delta(15)N values of nitrates having excess O-17. The same analyses also yield estimates of the magnitude of O-17 excess in the nitrate (expressed as Delta(17)O) that may be useful in some environmental studies. The 1-sigma uncertainties of delta(15)N, delta(18)O and Delta(17)O measurements are +/-0.2, +/-0.3 and +/-5%., respectively. Copyright (C) 2004 John Wiley Sons, Ltd.