In the paper above-named work [ibid., vol. 30, no. 8 (this issue, Aug. 1982), pp. 2000- 2002] Millott and Nicholson use a regenerator noise figure to select an equalization function, and then check the suitability of this choice by examining the resulting eye diagram. This leads them to the conclusion that a raised cosine response with a roll-off factor α = 0.3 is the optimum choice. There are some aspects of the use of an eye diagram in this way that deserve further comment. The commentor makes it clear that he does not think that these observations invalidate the main conclusions presented in Millott and Nicholson's paper. He believes that their selection of a response with a relatively sharp roll-off, and the preferences for partial response modes, are correct. However, the commentor thinks that the minimum phase case should be considered, and that it is desirable to take account of the tolerances in the equalization functions when estimating system margins. When this is done the optimum point will be a little different, and the margins rather smaller, than those suggested by Millott and Nicholson. Comments by Millott are provided.