There is a lack of information concerning the diet of vegetarian children. The present study compared the dietary intake of fifty vegetarian children, aged 7–11 years, with fifty matched omnivores. Three 3 d food records were completed by each child at intervals of 6 months. The day after completing the record each child was interviewed to clarify food items and assess portion sizes. Food records were analysed using Microdiet (University of Salford). Finger-prick cholesterol and haemoglobin measurements were taken from a subsample of the group. Only one child's family was a member of the Vegetarian Society and almost one-third of vegetarian children had omnivorous parents (seventeen of fifty subjects). The energy intake (MJ) of the vegetarians was significantly lower than that of the omnivores, 7·6 (SD 1·05) and 8·0 (SD 1·36) respectively; there were no significant differences in Fe or fat intakes. For the vegetarians polyunsatnrated:saturated fat ratio (P:S 0·7 (SD 0·04)) and NSP intake (13·8 (SD 0·7) g/d) were significantly higher than those of the omnivores (P:S 0·5 (SD 0·02), NSP 10·3 (SD 0·4) g/d). There was no significant difference in cholesterol measurements (mmol/I) between the two groups: vegetarian 3·5 (SD 0·12), omnivores 3·7 (SD 0·15). The haemoglobin level (g/I) of the vegetarians (11·8 (SD 0·2)) was significantly below that of the omnivores (12·4 (SD 0·2)); 47·5% of the vegetarian children fell below the third percentile of the Dallman reference curves (Dallman & Siimes, 1979). The intake of the vegetarians more closely resembled current recommendations (Department of Health, 1991), although they need to be as aware as omnivores of the need to reduce fat intake. The haemoglobin levels of vegetarian children suggest that they need dietary advice to ensure optimal absorption of Fe.