Search for a command to run...
It is proposed that broad emission lines in active galactic nuclei are formed by dense clouds in a molecular, hydromagnetic wind accelerated radiatively and centrifugally away from an accretion disk orbiting a massive black hole. These clouds are supposed to be photoionized by the UV continuum produced in the innermost radii of the accretion disk where the radiation flux is sufficient to evaporate the dust grains, that is, at a radius ~0.2L^1/2^_ion,46_ pc. If the wind is responsible for extracting most of the disk angular momentum, then the estimated ionizing flux, electron density, cloud filling factor, column density, and velocity are in accord with values for these quantities inferred from studies of the line widths, strengths, and variability. In order to reproduce typical line profiles with cusp shapes, blue asymmetry, and blueshifted peaks, it is found that the volume emissivity must scale roughly as r^-3.3^, where r is the cylindrical radius; in this case each decade of radius contributes nearly equally to the line profile. Simple models of the cloud emissivity approximate this dependence. We consider the possibility that the broad wings of the lines are formed through electron scattering by a ~10^6^ K warm intercloud medium. We find that such a phase can be thermally stable, when magnetically confined, and argue that a Thomson optical depth τ_T_~1 is automatically maintained. It is also found that emission over three decades of radius is necessary to reproduce a typical line profile if electron scattering is unimportant, but that two decades suffice when electron scattering broadens the line. Viewing the outflow from smaller angles with respect to the symmetry axis produces a more symmetrical line profile. Reverberation mapping and polarization observation may provide a test of this model.