The prevalence of physician-diagnosed diabetes and of undiagnosed diabetes and impaired glucose tolerance (IGT) that meet National Diabetes Data Group (NDDG) and World Health Organization (WHO) criteria have been estimated for the U.S. population aged 20–74 yr from the 1976–1980 National Health and Nutrition Examination Survey. This survey included a demographic/medical history questionnaire administered in the participant's home and a detailed examination composed of a physician's exam, special clinical procedures, other tests, and collection of blood and urine specimens. Survey participants were selected from 1970 census data through a stratified multistage probability sampling scheme. Of 17,390 eligible residents aged 20–74 yr, 15,357 (88.3%) participated in the interview and are the basis for estimates of diagnosed diabetes; 11,858 (68%) participated in the exam. A half sample of 5901 examinees was selected to receive a 75-g oral glucose tolerance test (OGTT) performed in the morning after an overnight 10- to 16-h fast. Of these examinees, valid OGTT data were obtained for 3772 people without a medical history of diabetes, and these are the basis for estimates of undiagnosed diabetes and IGT. The major reasons for incomplete OGTT data were inability of participants to attend the examination center in the morning and lack of adherence to the fasting instructions. Despite the relatively low response rates, evidence is presented that data on both the interviewed sample and those receiving the OGTT, when adjusted for the 1970–1980 census characteristics by age, race, sex, income, and geographic location, are representative of the U.S. population. Extrapolation of these data to the U.S. population aged 20–74 yr indicates a total diabetes prevalence of 6.6% by NDDG criteria, or more than 8 million people with diabetes. The prevalence of undiagnosed diabetes (3.2%) was almost equal to that of previously diagnosed diabetes (3.4%). Total rates of diabetes increased with age, from 2.0% at age 20–44 yr to 17.7% at age 65–74 yr. Rates were approximately equal by sex but were greater in Blacks than in Whites. The prevalence of undiagnosed diabetes by WHO criteria (3.4%) was similar to that by NDDG criteria, but the rate of impaired glucose tolerance (11.2%) was more than twice the NDDG estimate (4.6%). Both obesity and parental history of diabetes were associated with significantly higher rates of diabetes and IGT. Fasting plasma glucose was relatively insensitive to age, but 1-h and 2-h post-75-g glucose values increased significantly with age.