The purpose of this study was to assess the test-retest stability of the Fagerstrom Tolerance Questionnaire (FTQ) in two samples: (a) paid subjects in an American laboratory; data were collected via telephone screen and subsequently via questions embedded in a written history; and (b) smokers hospitalized for depression in Paris; data were collected via a written questionnaire upon admission and again after 3 weeks of treatment for depression. Reliability data are also presented for a recently revised version of the FTQ, the Fagerstrom Test for Nicotine Dependence (FTND), and compared with FTQ data collected in a subsample of subjects in the American database who received both versions of the questionnaire. Both the FTQ (in both samples) and the FTND proved to be highly reliable. The validity of the scales, using cotinine, number of years smoked, and the "addictive" factor on the Classification of Smoking by Motives questionnaire as criterion variables, was also supported. No relationship between FTQ score and severity of depression was detected in either sample. Internal consistency was somewhat higher for the FTND than for the FTQ, replicating previous findings in the literature.