Morbidity studies based on hospital admission rates are complicated because all admissions to a defined group of hospitals rarely come from a de fined population and all admissions from a defined population rarely go to a defined group of hospitals. As a consequence, if the starting point of the study is either the complete records of a hospital group or the complete enumeration of the population in a geo graphical area, the available data are usually precise for either the numerator or the denominator of admis sion rates but not for both. For a morbidity study the most realistic starting point is the population of a given area, but this is often served by a large number of hospitals both inside and outside its boundaries. As a result, the scrutiny of a mass of medical records is necessary in order that residents of the given area, if admitted to hospital, are included in the survey no matter where the hospital is situated, and in order that non-residents are excluded even though they are admitted to hospitals inside the geographical boundaries of the area studied. Apart from the practical difficulties of deriving reliable hospital admission rates with numerator and denominator properly related, their interpretation must be made with extreme care. Hospital admis sions depend on several factors, such as the number of available beds in various specialties and the attitude of general practitioners and their patients to hospital treatment. Such factors vary between areas; hence what may be true of one area may not be true of another or of the country in general. The Hospital In-Patient Enquiry (H.I.P.E.)?the 10 per cent, national sample survey of hospital patients organized by the General Register Office on behalf of the Ministry of Health?will enable the variation of hospital sickness between and within areas to be studied. Meanwhile the results of an analysis of all hospital admissions of persons living in a defined area may be of interest. The present paper gives a general analysis of the results. In a future paper, an attempt will be made to relate this analysis with certain social indices.