Twenty-three laboratories analyzed 5 replicate wine samples according to a specified version of the Ripper direct iodometric titration for sulfur dioxide. Each sample was analyzed for (A) free SO2, (B) total SO2, and (C) iodine-reactive substances other than SO2. Although variation of A with temperature and of A and B with time of analysis were anticipated, analysis of covariance showed no significant reduction in error when these variables were taken into account. Error did vary with SO2 level and wine type, red vs white. Pooled estimates of precision (within-laboratory error) in mg SO2/L wine were, for white wine: (A) 3.3, (B) 10.4, (C) 1.9; for red wine: (A) 3.8, (B) 7.3, (C) 1.9. Pooled estimates of systematic (between-laboratory) error in mgSO2/L wine were, for white wine: (A) 2.7, (B) 16.6, (C) 2.1; for red wine: (A) 4.3, (B) 15.1, (C) 3.0. Although rapid and convenient, the Ripper method is severely limited by poor precision and large systematic error. The Ripper method is not recommended for adoption by the AOAC.