Our objective was to study the relations between the risk of lung cancer and dietary factors in New Caledonia, in the South Pacific. A population-based case-control study of respiratory cancers was conducted between 1993 and 1995 in New Caledonia; we analyzed data for 134 incident lung cancer cases (109 men and 25 women) and 295 controls (227 men and 68 women) who had completed a food frequency questionnaire. Odds ratios (ORs) associated with food and nutrient intakes were calculated with unconditional logistic regression, with adjustment for smoking, age, and ethnicity. Analyses were performed separately for men and women. Among men, no significant associations were observed with any foods, including vegetables and fruits. Nonetheless, high consumption of dark green leafy vegetables (highest vs. lowest tertile of intake) was associated with decreased risk (OR = 0.5, 95% confidence interval = 0.2-1.2, P for trend = 0.12), particularly among Melanesians (OR = 0.4, 95% confidence interval = 0.1-1.0, P for trend = 0.07). A similar protective effect was also suggested for high consumption of poultry (P for trend = 0.06) and fresh fish (P for trend = 0.08). No significant association was found with nutrients. Among women, the analyses concerned few subjects and were not informative. This study suggests that high consumption of dark green leafy vegetables may reduce the risk of lung cancer among men in this population.