The divergent incidence patterns of gastric cardia and distal stomach cancers suggest different etiologies. Although obesity has recently been linked to cardia cancer in Western populations, its association with distal stomach cancer remains unclear. This study examined the relation of anthropometric measurements to risk by subsites of stomach cancer in a Chinese population. We identified 1124 population-based cases of stomach cancer, ages 20-69 years, newly diagnosed between December 1988 and November 1989 in Shanghai, China. Controls (n = 1451) were randomly selected from permanent Shanghai residents and frequency-matched to cases by age and sex. Information on demographic characteristics, height and weight, diet, smoking, and other exposures was obtained by trained interviewers in person. The body mass index (BMI) was calculated as weight in kilograms divided by height in square meters and categorized into quartiles based on the distribution among controls. Odds ratios and 95% confidence intervals were estimated using logistic regression models, simultaneously adjusting for age, education, income, cigarette smoking (men only), alcohol drinking (men only), intake of total calories, and chronic gastric diseases. For gastric cardia cancer, the odds ratios among men were 1.4, 1.5, and 3.0 in the second, third, and fourth quartiles of usual BMI (P for trend, < 0.01). Among women, elevated risks also were associated with excess weight, but the gradient in risk was not smooth. Risk patterns for usual body weight, maximum BMI, and minimum BMI were similar to those found for usual BMI. For distal stomach cancer, no association with usual BMI was observed among men, but a slightly elevated risk was seen among women. Our observations in China support recent findings in Western populations that obesity contributes to the risk of gastric cardia cancer, especially among men.