This working paper reviews some advantages and disadvantages of various alternative definitions and classification schemes. Certain of these alternatives were discussed by an international group of experts on the epidemiology of diabetes. It was recommended by the conferees that the following terms not be used as standard descriptors in epidemiologic reports: juvenile diabetes, juvenile-type diabetes, maturity-onset diabetes of youth (MODY), latent diabetes, subclinical diabetes, clinical diabetes, chemical diabetes, and J-type (Jamaica) diabetes. Alternative ways of describing these conditions were offered. Some, but not all, conferees favored abandonment or modification of the terms “secondary diabetes” and “pancreatic diabetes.” Two alternate approaches are described in classification. One is designed to classify all types of cases in a single scheme. In this system each case is described by indices that include etiology and special etiologic concomitants, fatness, severity, type of treatment, and, if present, special or peculiar manifestations, and morbidity. This system makes it possible to classify all cases by mutually exclusive categories. A disadvantage of this scheme is that the number of combinations of classifying indices and subindices is great. Another approach is to classify cases by syndromes, using in each class a mixture of indices (e.g., etiology, severity, fatness, etc.). The conferees failed to reach a consensus in the specifics of this approach. There was considerable disagreement on how certain terms should be defined, whether they should be used, and on the degree of need for specific definitions. These designations included the terms “insulin-dependent,” “idiopathic,” “insulin-requiring,” and “ketosis-prone.” The conferees agreed that in either of the two approaches to classification, the prime indices should be those that can be ascertained by ordinary clinical methods. Much progress could be accomplished, even in the absence of international standardized definitions and classification systems, provided that authors keep in mind the need to describe the major characteristics of their cases and define the sense in which they use terms such as insulin-dependent, ketosis-prone, obese, etc.