Previous studies of tobacco dependence have tended not to examine adolescent samples. In a cohort of 18-year-olds, the 12-month prevalence rate of dependence based on DSM-III-R criteria was found to be 19.3%. Cotinine levels in saliva were positively related to level of cigarette consumption but not to tobacco dependence and may not be a suitable measure of dependence. Measures of behavioral and psychological aspects of smoking supported the utility of DSM-III-R criteria for tobacco dependence among young smokers. As a large number of young smokers are dependent on tobacco. the results support the development of treatment programmes for youth that take into consideration their degree of dependence.