The Griess reaction is widely used to measure the cellular production of NO by detecting the supernatant levels of nitrite. Ordinarily, background levels of nitrite in the media are subtracted from the levels of nitrite produced by the cells by preparing a "blank" during the determination of the standard curve. Although this method is adequate for most experimental conditions, it cannot be used when cell supernatants are collected from multiwell dishes, particularly when low amounts of NO are produced and when long incubation periods are required to induce NO production. Our data show that a highly variable level of nitrite is found in the absence of cells in the media from wells at the edges of the 96-well plate while media from interior wells shows no detectable nitrite accumulation. The most likely source of this noncellular NO is from nitric oxides (NOx) found in the ambient air and reduction of air exchange or regulation of the gaseous environment eliminates this "border effect."