The profiles of the Balmer lines in NGC 5548 as reported by Dibai et al. (1968) were somewhat asymmetric, whereas those reported by Anderson (1970) are smooth and symmetric. We present profiles which are strongly asymmetric, resembling those of Dibai et al. Evidently electron scattering is not the sole principal broadening agent and we must deal with velocities ∼ 2500 km s−1 in a very small volume.

The transient nature of the P-Cygni type profiles in the Balmer lines of NGC 4151 has previously been noted (Cromwell and Weymann, 1970). These lines have since disappeared, at the resolution available to us, in a time of 3 months. A model in which frequent outbursts of shells or filaments produce transient features in the Balmer lines, while the accumulated material from past outbursts produces the relatively stable HeI λ 3889 feature, seems the most plausible.