- We incorrectly claimed quantitative agreement with no free parameters with the data of Ralph et al.. In effect, a free scaling factor was included in our fit to their data, because of an improper use of the temperature-dependent conductance to normalize the differential conductance data . A portion of Fig. 1 with the correct no-free-parameters curve and the curve with the free scaling factor is presented here. Our discussion of the relevance of the two-channel Kondo interpretation to the data is not affected by this error.