Objective: A novel system for nutrient analysis has been developed and tested over 5 years. Its key features are a nutrient database of 600 commonly eaten foods (95% of foods eaten in 7-day surveys); a booklet identifying each food with a bar code, bar codes for gram weight and for portion sizes (small, medium, large) and a bar-code reader with dietary analysis software for PCs. In the present study the bar-code system has been evaluated by comparison with a commonly used manual entry nutrient analysis software for dietitians' use. Design: Cross-sectional. Setting: Glasgow city district. Subjects: One hundred and sixty adults aged 18-65 years old. Results: Comparing mean intakes for macro-and micronutrients, using the Bland and Altman method 1 , the bias between the two methods was small, ranging from 0.93 to 1.03. The bar-code system took significantly less professional time in data entry and nutrient analysis than the widely used manual system (29 min per 7-day diary vs. 47 min per 7-day diary, P Ͻ 0.001). Conclusions: It is suggested that the bar-code system offers greater speed with a saving of professional time needed for nutrient analysis of dietary surveys. This system is commended for maintaining accuracy while promoting economy.