Appetite measurements were performed in 109 Beninese children aged 18–30 months to develop a tool for appetite evaluation in young children in nutritional intervention programmes. Two test foods were identified as appropriate for these children: a maize porridge (aklui) and rice (riz-au-gras). Ad libitum intakes of the foods served after an overnight fast according to a standardized offering procedure were measured on 3 days. The children's habitual intakes were measured during 3 consecutive days not overlapping with the days when the test meals were provided. Energy intake from the test foods was comparable to breakfast energy intake which was 0?8–1?0 MJ, representing 21 f total daily energy intake. Energy intake from aklui was significantly correlated with daily intake (r=0?41,n =38, p<0?05) and with energy intake from breakfast (r=0?52,p <0?01). Correlations concerning riz-au-gras were less pronounced and non-significant. Reproducibility (as coefficient of variation) of the appetite test as calculated from the triplicate measurements was 40 or aklui and 25 or riz-au-gras. This reproducibility is better than that of the habitual breakfast intake (43–45Ž In conclusion, the appetite test used in our studies can be considered as an appropriate tool in appetite evaluations.