A line-emission ``gap'' is usually observed between the two regions, most objects showing an optical spectrum which can be fitted by line profiles corresponding to clouds belonging to one or the other line-emitting regions. This line-emission gap can be explained by the presence of dust mixed with the gas (Netzer & Laor 1993). Nevertheless, the existence of an intermediate region, both in terms of velocity and density, is expected; in such a region, the O III lines would be partially collisionaly de-excited and show substantially broadened wings (Shields 1978). This ILR should not be confused with the ILR found in QSOs by Brotherton et al. (1994), which is much smaller and denser, with a velocity dispersion of the order of 2000 kms-1 and density 1010 cm-3.
Mason et al. (1996) presented high-resolution (2 Å) optical spectroscopic observations of KUG 1031+398. The model they used to fit the data revealed a line-emitting region with lines of intermediate width (FWHM 1000 kms-1); according to Mason et al., this region would dominate the Balmer lines profile, being also a significant contributor to the O III4959, 5007 lines, with a flux ratio 5007/H = 1.4, suggesting an intermediate density.
Osterbrock (1978) thought that he had detected, in a few Seyfert 1 galaxies, faint wings to the O III lines with essentially the same widths as the Balmer lines. Crenshaw & Peterson (1986) and van Groningen & de Bruyn (1989) have found broad wings in the O III lines of a number of Seyfert 1 galaxies, implying the presence, in these objects, of an ILR with a density of a few times 106 cm-3, similar to the one reported in KUG 1031+398; however, all these objects show strong Feii emission, and the observed broad O III components could be due to an inaccurate removal of the Feii blends (Boroson & Green 1992).
In summary, although the presence in Seyfert 1 galaxies of emitting clouds with density intermediate between those of the ``broad'' and ``narrow'' components is not unexpected, no uncontroversial report of the existence of such intermediate components has ever been made to the best of our knowledge. Therefore, the claims by Mason et al. (1996) that the NLS1 KUG 1031+398 shows evidence for an ILR induced us to conduct new spectroscopic observations and modelling of its emission-line features.