During recent years very wide interest has been shown in the relation between smoking habits and health, and the literature has been well summarized in the publication Smoking and Health Now by the Royal College of Physicians (1971). For the most part emphasis has been given to the effect of smoking habits on the shortening of life and it has been estimated that in the United Kingdom two smokers out of every five are likely to die before reaching the age of 65 years, as opposed to only one out of five non-smokers. However, death is often the final stage of a protracted disease process and morbidity may then be an even more sensitive indicator of impairment of health than mortality. Although the effect of smoking on morbidity has been studied in a wide variety of circumstances the approach most commonly adopted has been to carry out a special survey of a specific population involving a standard examination or enquiry. For example, studies of coal miners in the United Kingdom (Ashford, Brown, Morgan, and Rae, 1968; Ashford, Morgan, Rae, and Sowden, 1970) have shown that smoking is associated with impairment of ventilatory function, with an increase of respiratory symptoms, and with absence from work due to chest illness. In contrast, the present study is concerned with the use made of the health services by a large community over a substantial period of time. The Exeter Community Health Research Project has been described by Ashford and Pearson (1970). This project involved a study of morbidity in the City of Exeter during the calendar year November 1966-October 1967. For a population consisting at any one time of some 70,000 persons defined by the National Health Service general practice lists of about three-quarters of the general practitioners then working in the city, a record was made of each general practice surgery attendance, general practice home visit, hospital outpatient attendance, and hospital inpatient stay. In order to obtain informa tion about the personal characteristics of the survey population a private census was undertaken. This census included a record of individual smoking habits and it was thus possible to examine the relation between smoking and the use made of the local health services in that section of the popu lation for which both sources of information are available.