Gast and Candelore have recently pointed out that the standard method of computing confidence limits in Monte Carlo calculations is invalid for iterated-source calculations. This note develops a formula for a confidence-limit correction factor for such calculations, and discusses its use.