Simultaneous measurements of NOx (NO+NO2), NO, and O3 production in a laboratory discharge show that within the uncertainties of the experiment, all of the NOx produced was NO, and no detectable enhancement of O3 after the discharge was observed. The laboratory experiments described gave an NO production rat of 5¿2¿1016 molecules joule-1 for a 105--106 joules m-1 spark. Assuming that the global dissipation of lightning energy is about 10-8 joules cm-2 s-1 (Dawson, 1980; and Hill et al., 1980), our NO production rate results in a global source of NO due to lightning of about 1.8 Mt(N)/yr, which is considerably lower than earlier estimates. This lower value for NOx production by lightning suggests that NOx emissions from anthropogenic sources, estimated to be at least 20 Mt(N)/yr, may be the dominant source of NOx to the global troposphere. Furthermore, since most of the anthropogenic sources of NOx are located in the Northern Hemisphere, this new interpretation of the relative source strengths of this species favors a highly skewed asymmetric distribution of NOx.