We disagree with Mason et al. on the analysis of the emission line profile of KUG 1031+398, in the sense that we find no evidence for the presence of an ``intermediate'' component in which the forbidden lines are almost, but not completely, suppressed by collisional de-excitation. Nevertheless, we find that this object is exceptional in having a ``narrow'' line region (defined as a region where 5007/H 5) with almost the same width at half maximum as the ``broad'' line region (Balmer lines with no detectable associated forbidden lines); however, in the first case, the line-profile is Gaussian, while in the second case, it is Lorentzian.