The results of studies on cigarette smoking as a risk factor for stroke are more controversial than for cardiovascular disease. The CO-induced increase in the corpuscular elements of blood (erythrocytes), the influence on other parameters (such as RBC volume, haemoglobin, haematocrit, blood and plasma viscosity, tendency of erythrocytes and platelets to aggregate, fibrinogen level etc.), as well as the increase in catecholamine level are taken to be reversible. On the other hand, the association between cigarette smoking and probably irreversible morphological changes in the craniocervical vessels--possibly via lipid metabolism--is well documented. The following possible explanations for discrepant results in the literature are discussed: different extent of daily cigarette smoking, inhomogeneous populations, difficulties in diagnosing stroke, especially before the introduction of computed tomography and the common failure to consider other risk factors.