The paper considers the estimation of the potency of an unknown test preparation in relation to a standard in a dilution assay. The method utilizes the Mantel-Haenszel one degree of freedom chi-square approach to test consistency of the observed potency ratio with a range of alternative potency ratios. The method is nonparametric as it does not require any distributional assumptions. However, it requires the doses employed to be equally spaced logarithmically. The dose-response curve is also required to be monotone over the dose range used.