The review examines the effects of smoking on blood parameters, concentrating on those seen as responses or reactions to the insult of smoking, and finds evidence that smoking causes chronic leucocytosis, macrocytosis and raised haematocrit, raised plasma fibrinogen concentration, reduction of the serum ratio of high to low density lipoprotein (HDL/LDL ratio), and platelet changes. Each of the changes resolved on “quitting”, though for fibrinogen the evidence was indirect. Reports that elevations of the white cell count (WCC), and plasma fibrinogen and reduction of the HDL/LDL ratio each predict myocardial infarction, and that higher haematocrit increases the risk of cerebro-vascular incidents are reviewed, with reports of associations with other forms of arterial disease and COPD, and of their significance for the prognosis of established disease. After noting pathological mechanisms which implicate each of these factors, and platelets, in reactions likely to contribute to the development of atherosclerosis, infarction, arterial spasm, and/or lung damage, the author concludes that the evidence in man, backed up by experimental data, provides very strong support for the view that elevations of WCC, haematocrit, plasma fibrinogen and reduction of the HDL/LDL ratio represent (or are closely associated with) intermediate causal mechanisms through which smoking induces arterial disease and probably lung disease, and that experimental evidence indicates that platelet changes and macrocytosis also contribute. Differences in the extent of these responses to smoking could be valuable in differentiating the relative harmfulness of different types of cigarette (or of other inhaled pollutants) in terms of these diseases, and in predicting the susceptibility of individuals to these smoking-related diseases.