From the data of an experiment of selection for increased longevity, a realized heritability of longevity was calculated. The low value of this heritability (3.4%) was very close to values observed in other experiments concerning Drosophila melanogaster wild strains. The error variance of the heritability estimator was calculated through the use of orthogonal contrasts. In this way, it was possible to calculate the confidence interval of the realized heritability. The amplitude of this interval was wide although the size of the sample was large. This shows the difficulty of estimating with precision the heritability of longevity from data obtained in selection experiments.