Line emission is one of the defining properties of quasars. It has also been one of the main lines of research over the last three decades. The information content of the spectrum is -trivially- richer in the lines than in the continuum. Unfortunately, the information gained from the emission line does not allow researchers to gain much understanding on the mechanisms at the origin of the radiation. This must be obtained out of the relatively information poor continuum. What the information provided by the lines allows us to do is to describe the gas surrounding the energy source. This includes a description of the physical state of the gas (temperature, density, ionization level), of its kinematics through the width of lines and of its geometrical arrangement (filling factor) through the equivalent width of the lines. A general review of these inferences can be found in [Peterson 1997] and [Netzer 1990]. A detailed fit to all line features in 3C 273 can be found in [Wills Netzer & Wills 1985].
The classical picture is that the lines are formed in a set of photoionized "clouds" in rapid movement around the central black hole. The smoothness of the lines implies that the number of clouds must be large. How large is however not known yet (see Dietrich et al. in preparation). The continuum at the origin of the photoionization is normally associated with the central source.
Whereas one would expect that the study of the line intensity ratios should be able to provide a description of the ionizing continuum (since the lines come from elements that have different ionizing potentials) particularly in the unobservable part of the spectrum between 912Å and ~ 0.1keV few concrete results have emerged [Binette et al. 1988], [Krolik et al. 1988].
[Wills Netzer & Wills 1985] claim that the photoionization models they use represent well all the line features of the quasars they describe with the notable exception of the FeII line blends observed in the optical and UV parts of the spectrum. 3C 273 is no exception to this "FeII problem". Expected values for the intensity ratios of FeII lines to Ly is of 0.3-0.5 [Netzer 1990] whereas the observed ratio (corrected for a reddening AV = 0.16 corresponding to the galactic reddening E(B-V) = 0.05 as used in [Ulrich et al. 1988]) is slightly larger than 1 [Wills Netzer & Wills 1985]. Clearly, assuming a more important intrinsic reddening will tilt the FeII lines to Ly ratio to smaller values, lessening the problem; there is, however no reason to assume a large intrinsic reddening (see above). This problem is as of yet unsolved and may point to additional energy sources in the broad line clouds (e.g. mechanical heating) and/or to a more complex structure of the broad line region than envisaged in most photoionization calculations [Collin-Souffrin & Lasota 1988].
Whatever the details of the fits and the agreement of the line intensities with various photoionization models, a very important point is that the heavy element abundances are large, in some instances larger than solar. This indicates that the gas surrounding quasars, and in particular 3C 273, has been going through one or several generations of star formation and explosion before being found in the very inner regions of the galaxies hosting the quasar.