The experimental conditions that affect equilibrium solubility values measured by the classical saturation shake-flask method have been examined, using hydrochlorothiazide as a model compound. Modifications in temperature, sedimentation time, composition of aqueous buffer and the technique of separation of solid and liquid phases were all found to influence the equilibrium solubility results strongly. However, variations in the amount of solid excess and stirring time were found to have less influence. In the light of these observations, a new, shorter protocol has been developed for measurements of equilibrium solubility, together with recommendations for good analytical practice. The equilibrium solubilities of five other drugs were measured to verify the new protocol.