Problems in accounting for the production of nitrous oxide by electrical discharges and lightning are discussed. A model of nitrous oxide production by corona discharge in lightning appears able to quantitatively account for both storm cloud and laboratory observations of nitrous oxide enhancements if the N2O yield from collisions between N2(A3&Sgr;) and 02 molecule is, as found by Zipf, large. Global estimates of nitrous oxide production by lightning and by point discharges beneath thunderclouds are only of the order of thousands of tons of nitrous oxide per year; however, estimates of global production of N20 by electrical, high voltage, power line coronas are from 1 to 2 orders of magnitude larger than by lightning.