Smoking-behavior during pregnancy and the reliability of an interview were prospectively investigated. The tobacco-exposure was assessed a) by an intake interview b) by a smoking diary and c) by maternal thiocyanate and carboxy-hemoglobin concentrations. Of the 115 nulliparous women, 66 smoked (57.4%). Seventeen women (25.8%) quit and 39 women (59.1%) reduced smoking. Directly after intake, in the first week of the diary, significantly more cigarettes were smoked than stated during the intake interview. According to the diaries, the number of cigarettes increased during the first half of pregnancy. In the 24th week significantly more cigarettes were smoked than in the first week after intake. Also a significant increase of thiocyanate between the measurements at intake and in the 24th of amenorrhea was found. The maternal carboxy-hemoglobin levels did not significantly change during pregnancy. Investigators should be aware of changes in smoking behavior during pregnancy and the potential bias of self reported smoking behavior at the beginning of pregnancy. The tobacco-exposure should be assessed (anamnestic or biochemical) several times during pregnancy, anyway it should include the second half of pregnancy.