This Short Communication corrects Table 1 and Figs. 2 and 3 that were published in a recent article by the same authors, in this journal. The article discussed response surface methodology (RSM), which searches for the input combination maximizing the output of a real or simulated system. RSM uses steepest ascent (SA), which is scale-dependent. The article derived scale-independent `adapted' SA (ASA). The two search directions were explored in Monte Carlo experiments. Unfortunately, the canonical and the non-canonical cases were mixed up. This Communication still shows that-in general-ASA gives a better search direction than SA. (c) 2004 Elsevier B.V. All rights reserved.