Correlations of slope unity between the optical continuum luminosity and the Balmer line luminosities (Yee 1980; Shuder 1981) have been traditionally used to demonstrate that photoionization is the excitation mechanism of the emission lines in active galactic nuclei. This correlation can be expressed in terms of a constant and universal equivalent width of Hbeta over five decades in optical-UV continuum luminosity. It is found that the equivalent width EW(Hbeta) of narrow-line objects (Seyfert 2's, NLRG) is not systematically different from that of broad-line dominated objects (QSO's BLRG, Seyfert 1). Based on empirical evidence as well as on recent BLR models we argue that the global efficiency for reprocessing the number of ionizing photons into Hbeta line photons is equal to or less than that given by recombination case B. Assuming a canonical cloud covering factor of C(f) = 0.1 of an isotropic UV source, for any reasonable value of the power-law index alpha (L(nu) is-proportional-to nu(+alpha), we find a substantial deficit of ionizing photons in all AGN. For instance, an index of -1.4 or as flat as -0.6 leads to a deficit in ionizing photons of a factor of 25 and 4, respectively. A lower reprocessing efficiency as that characterizing the more recent BLR models brings the deficit in the range 8-15 (for alpha = -0.6). The solution envisaged to account for the apparent deficit and to explain how EW(Hbeta) can be aspect independent from Seyferts to radio galaxies and quasars is that the observed optical continuum originates from or near the line-emitting clouds (i.e., ''hazy'' nucleus) and is distinct from the primary ionizing radiation directly impinging on the clouds. Evidence for this is provided by the detection of continuum emission associated with the spatially resolved emission-line gas in radio galaxies and nearby Seyferts. The possible mechanisms envisaged to explain the spatial association of continuum emission and line emission are electron scattering, dust scattering, and in situ generation of continuum by high-velocity shocks, by thermal gas emission, or by inverse self-Compton scattering. The main observable consequences of this spatial association between the continuum and line emission are that (1) it can explain in a natural way the observed narrow range of EW(Hbeta) in AGN whether or not the primary ionizing source is anisotropic (2), it can result in similar temporal behavior of continuum and line emission (i.e., short time lags of the BLR lines) as a result of intermixed continuum and line reverberation, (3) and it explains why dust extinction of the nuclear continuum emission never exceeds that of the BLR lines (e.g., in Seyferts 1.8 and 1.9) since both emission processes being co-spatial are characterized by a similar path length to the observer.
