Experiments to estimate the feeding rates of aquatic insect larvae involved destructive sampling of larvae to determine the position of an introduced food marker as it passed through their guts. A normal linear model involving censoring (equal to complete passage) and truncation is fitted to the measurement of the passage of food. A method for maximum likelihood estimation is derived together with the second order variance-covariance equations. Methods for estimating the effect of larval size on feeding rate are examined.