OBJECTIVE - While high parity is hypothesized to be associated with insulin resistance and type 2 diabetes, few studies have examined this association in diverse racial samples or geographical areas. Our objectives were to estimate the magnitude of association between parity and diabetes and to determine if higher parity is predictive of future risk of diabetes. RESEARCH DESIGN AND METHODS - This was a population-based, prospective cohort study of 7,024 Caucasian and African-American women from the Atherosclerosis Risk in Communities study, a prospective epidemiological study of men and women aged 45-64 years, with 9 years of follow-up. Incident diabetes was defined by the 1997 American Diabetes Association diagnostic criteria. Parity was defined as the number of live births (no live births [nulliparity], one to two live births, three to four live births, and five or more live births [grandmultiparity]). Parity and risk of diabetes was estimated for 754 incident cases of diabetes with Cox proportional hazard regression models, adjusting for sociodemographic, clinical, and lifestyle factors and inflammatory markers. RESULTS - Incidence rates were highest among women with five or more live births (23/1,000 person-years [95% CI 20.3-26.7]) and lowest among women with one to two live births (11/1,000 person-years [9.6-12.5]). Adjustment indicated that much of the risk was due to sociodemographic factors and higher obesity, but after adjustment for all covariates, grandmultiparity (five or more) was still associated with a 27% increased risk for diabetes (hazard ratio 1.27 [95% CI 1.02-1.57]). CONCLUSIONS - Grandmultiparity is predictive of future risk of diabetes after adjustment for confounders.