The reliability and validity of the 24-hour dietary recall method was tested statistically by comparing observed with recalled intakes of kilocalories and protein. Data on observed intake of 28 children, aged 10 through 12, with asthma, cystic fibrosis, or diabetes, participating in a summer camp program, were obtained by unobtrusive observation. The same subjects were interviewed within 24 hours after the observation. Intakes of kilocalories and of protein were computed for both observed and recalled data. The two sets of data were compared by paired t-test and regression analysis. A large and significant difference was found between mean recalled and mean observed intakes for both kilocalories and protein. This coupled with a low but significant coefficient for reliability limits the usefulness of this dietary assessment tool in the age group studied.