Monte Carlo experiments were performed on each of several procedures for treating prerecord flood information in flood-frequency analyses. The performance of each method was judged by how well the 100-yr peak discharges determined by each method approximate the true value of the 100-yr peak discharge of the Pearson Type III distribution underlying the generated data. Results indicate that when sample skew is adjusted for historic information before weighting with a generalized skew coefficient, it is better to use historic record length (length of record extended to include historic information) rather than systematic record length (length of record in which annual discharge information is routinely collected) in computation of the weighting factor. Also, adjusting for historic information after weighting sample skew and generalized skew coefficients was the optimal method tested.