Introduction For nonsolution,orally administered dosage forms,the USP Performance test can rely on either disintegration (General Chapter Disintegration ) or dissolution (General Chapter Dissolution ). The dissolution procedure is an important test both to establish bioequivalence and also for quality control. The procedure includes a complex sample preparation step in a dissolution apparatus. General Chapter specifies acceptance ranges for the dissolution vessel and variables of operation (2). During its multilaboratory studies of new Lot P Prednisone Reference Standard Tablets (Prednisone RS Tablets) used in periodic performance testing (PVT),USP noted substantial interlaboratory variability in results. This has led USP to conduct a comprehensive set of studies to examine potential assembly deviations that contribute to this variability. This study is one of a series of reports of these investigations (3–6). Based on these studies,USP may change statements in and allied documents to assist manufacturers in improving the USP Performance test when the dissolution procedure is chosen. We designed this study to screen variables for Apparatus 2 (Paddle), with plans for a similar study of Apparatus 1 (basket) later on. Because this was a screening study, we did not examine exhaustively the effects of each variable but rather attempted to identify which variables were the most essential to control in order to achieve accurate and reproducible dissolution results. Previous research has focused on assembly variables such as dissolved gas, vessel centering, vibration, vessel tilt, paddle height, rotation speed, base plate levelness, shaft wobble, temperature, vessel shape, belt tension, and shaft verticality (3, 7–12). Building on these earlier studies, we attempted a more systematic approach that included an assessment of combination effects from two or more concurrently perturbed variables. In this study, nine variables were evaluated using Abstract The specifications and acceptable ranges for nine critical dissolution apparatus variables were examined during a perturbation study of USP dissolution Apparatus 2 using USP Prednisone Reference Standard (RS) Tablets. A Design of Experiment (1) approach was used to screen the nine variables to determine the contribution of each,alone and in combination, to mean percent dissolved and standard deviation results. We observed a wide range of dissolution results, including several values that fell outside of current acceptance limits,even though the variables were kept within currently acceptable ranges. When we analyzed mean percent dissolved results,we found three variables that were statistically significant: level of deaeration,vessel type,and rotation speed. When we examined standard deviation results,we found that five variables or combinations of variables were statistically significant:vessel type, level of deaeration,paddle height,paddle height–vessel type interaction,and paddle height–level of deaeration interaction. We also found that the other variables examined—temperature,shaft wobble,vessel centering,vessel tilt,and base plate levelness—were not statistically significant within the ranges explored in this study. Acceptance ranges for several assembly variables may need to be more stringent or more precisely defined in order to decrease interand intralaboratory variability (reproducibility and repeatability) in dissolution testing.