Evidence suggesting the use of self-reports as an index of systemic exposure to cigarette smoke in selected study populations is highly inaccurate. In order to assess the use of hair analysis as a biochemical marker of cigarette smoking, we compared measurements of nicotine and cotinine in the hair and plasma of 36 volunteers whose reports of smoking were deemed to be reliable. A significant correlation was observed between the number of cigarettes smoked and hair measurements of nicotine (r = 0.48, p = 0.004) and cotinine (r = 0.57, p = 0.0008). In addition, a good correlation was found between the reported number of cigarettes smoked and plasma nicotine, plasma cotinine, and carboxyhemoglobin levels. These results suggest that hair analysis is a reliable noninvasive method of determining human exposure to cigarette smoke and is comparable to blood analysis.