To study the incidence of upper respiratory and digestive tract cancer in Denmark in relation to changes in tobacco and alcohol consumption, the incidence trends are compared to the figures for average annual consumption during the period 1943-1982. The comparison is made using relative risk estimates available from previous investigations. From the available data on alcohol and tobacco consumption, an attempt is also made to predict the future incidence trends of these cancers. Since no data on the distribution of smoking versus drinking by sex and age are available from Denmark over the study period, two hypotheses are used. It is shown that the predicted incidence is strongly dependent on the distribution of smoking and drinking, and that the existence of a group with heavy consumption of tobacco and alcohol could be responsible for very high incidences of these cancers in the near future in Denmark.