Objectives. Report the 8-year impact on body size, physical activity, and diet of a community-based diabetes prevention program for elementary-school children in a Kanien'kehá:ka (Mohawk) community in Canada. Methods. Follow-up (1994 -1996) of subjects in the intervention and comparison community and repeat cross-sectional measurements in the intervention community alone from 1994 to 2002. Measures included triceps and subscapular skinfold thicknesses, body mass index (BMI), weekly number of 15-minute episodes of physical activity, run/walk test times, television watching, and consumption of sugared foods, fatty foods, and fruits and vegetables. Results. The longitudinal data of 1994-1996 showed some early positive effects of the program on skinfold thickness but not on BMI, physical activity, fitness, or diet. Repeat cross-sectional measures from 1994 to 2002 showed increases in skinfold thickness and BMI. Physical activity, fitness, and television watching showed favorable trends from 1994 to 1999 that were not sustained in 2002. Key high-fat and high-sugar foods consumption decreased, as did consumption of fruits and vegetables. Conclusions. Although early results showed some successes in reducing risk factors for type 2 diabetes, these benefits were not maintained over 8 years.