Combined 17O and 18O measurements of tropospheric CO from two northern hemisphere (NH) sites reveal systematically enhanced 17O levels, confirming the existence of significant mass independent oxygen isotope enrichment in this important trace gas. When CO levels increase in the NH winter, the mass independent enrichment decreases proportionally. A possible source of this rare isotope effect in CO is transfer of the mass independent enrichment from O3 to the CO pool via ozonolysis of unsaturated hydrocarbons such as isoprene and terpenes. Laboratory ozonolysis experiments indeed confirm that this process is a potentially important source of mass independently enriched CO. The extra 17O enrichment found in ozonolysis-derived CO is similar to the mass independent fractionation measured on ozone. If ozonolysis is the only source of mass independent enrichment in CO, it is estimated that about 10% of all CO in middle to high latitude winter, at ground level, originates from the ozonolysis of unsaturated hydrocarbons. While the laboratory experiments prove that ozonolysis is a source of mass independent enrichment in CO, a further problem unfolds. Because tropospheric O3 is strongly enriched in 18O as well, this enrichment is also transferred to the CO inventory. When 10% of the CO inventory has the strong 18O enrichment, the measured low 18O abundance of atmospheric CO requires for compensation the existence of a strongly 18O depleted CO source, or a selective isotopic modification due to isotope fractionation. No such source has been identified to date, and thus ozonolysis may not be the only source of mass independent fractionation in atmospheric CO. Because the mass independent isotopic enrichment is by its nature indestructible by common (i.e., mass dependent) fractionation processes, it is an ideal tracer. ¿ 1998 American Geophysical Union