The base-catalyzed aldol condensation between benzaldehyde and p-acetylbenzoic acid in water shows an inverse solvent kinetic isotope effect, k3,D2O/k3,H2O, of 1.33 ± 0.03. The reaction is definitely faster in D2O. This is interpreted to mean that the rate-limiting step in a five-step mechanism is Step 5, the final elimination of hydroxide from the enolate intermediate, not the formation of that intermediate. This is the same result and the same conclusion as from earlier studies in aqueous acetonitrile and refutes a suggestion, based on computations, that the rate-limiting step would change in water. Those computations are criticized as implying impossibly large isotope effects.