The 1st of a series of morbidity surveys on a sample basis planned for the Arsenal study area of Pittsburgh is discussed. The survey design, method of selecting the sample, and field procedures are described in detail. In addition, the accuracy of the findings for a single characteristic[long dash]persons ill during the month prior to interview[long dash]is discussed with respect to both sampling and non-sampling errors. Particular reference in the discussion of the latter is made to the errors of nonresponse, of respondents, and of enumerators.