From April 1987-March 1988, the first Dutch National Food Consumption Survey (DNFCS) was carried out among 2203 households comprising 5898 subjects, aged 1-85. In 1992 (January-December) the second DNFCS was conducted among 6218 subjects, aged 1-92 belonging to 2475 households. In both surveys, among other things, information on individual food intake was obtained using 2-day dietary records. Since changes over time are the most important topic of this study special attention is given to the methodological differences between the surveys. Because the food consumption methods used in both the surveys and the sampling frame were similar for both surveys there was a fair level of standardization of research methods and procedures. However, both surveys differ with regard to non-response, survey starting date, distribution of survey days over the week and the seasons, sample size, Dutch food composition data bank (NEVO) and coding arrangements. The random sample deviated from the Dutch population at large with respect to sex and age distribution. Because sex and age are major determinants of food consumption weighing factors were used in calculations not restricted to specific sex-age categories. The outcome of calculations correcting for variables other than sex and age hardly differed from those in which sex and age were the only variables adjusted for. Therefore, these additional variables were not included in the weighing factor. The results indicate that non-response bias is unlikely to have a strong effect on the survey outcome. In comparing intake figures from the two national surveys, special attention is given to those nutrients of which the mean intake differed most from the Dutch guidelines for a healthy diet or from the Dutch Recommended Dietary Allowances (RDA). The survey in 1987-88 indicated that overweight and obesity was prevalent among the Dutch population, especially among the elderly. The data of that survey suggested that protein intake may be too high for part of the population, especially the elderly. A high (saturated) fat intake was identified as a major nutritional risk in the Dutch diet. Other undesirable characteristics were a low intake of dietary fibre and a low percentage of energy derived from carbohydrates. In 1987-88 the largest difference (below 100%) between mean intake levels of micronutrients and the corresponding RDAs were observed for vitamin B6 (per gram of protein), iron (boys aged 1-3 and 13-18, women aged 1-3 and 10-49) and vitamin A (boys 10-12 years old, girls aged 7-12 and pregnant women). In 1992, the mean intake of energy was lower than 4-5 years before (9.3 vs 9.7 MJ/day), whereas no clear differences were observed regarding mean body weight and mean body mass index. There are no indications that the differences in average energy intake between both surveys can be attributed entirely or partly to differences in reporting behaviour (over- or under-reporting). The average contribution of fat (saturated and monounsaturated fatty acids) to energy intake was lower in 1992 (36.9 energy%) than it was in 1987-1988 (40.0 energy%), whereas in 1992 more energy was obtained from protein and carbohydrates (poly saccharides but not mono- and disaccharides). Both intake of dietary fibre and cholesterol were substantially lower in the second DNFCS. As to micronutrients, the mean intake of retinol equivalents was lower and the intake of thiamin and vitamin B6 was higher in 1992. The above described differences were seen in almost all sex-age groups. Differences in the results of the two surveys can be the result of four broad categories of causes, namely changes in socio-demographic characteristics, methodological differences, food choice and composition of food products. Of the differences observed a major part could be attributed to food choice changes. This indicates the importance of a regular monitoring of the food consumption. The classification into food groups was not identical for both surveys. Generally speaking, these differences will give rise to only minor apparent changes. With respect to the food groups 'fats and oils' and 'mixed dishes', however, differences between the two surveys should be interpreted with caution. Part of the differences result from the updating of the Dutch food composition data bank. Such an update reflects real changes as well as changes in the analytical method. In particular, the differences in the intake of dietary fibre and, to a lesser extent, of vitamin B6 and vitamin A are (partly) due to changes in the method used to analyse the amount in food products. In conclusion, the average Dutch diet still does not fulfil the criteria for a prudent diet and will most likely induce increased risk for some chronic diseases, whereas among some population groups a low intake was prevalent for iron and vitamin A. Marginal vitamin B6 intake is more general although the risk for marginal status is not as clear as in the first survey. Dietary intake among the Dutch population has changed substantially in the period 1987-92, whereas both positive and negative changes were observed regarding the criteria of a healthy diet.