Previous
abstract 
Next abstract
We present the results of our recent statistical investigations of broad emission line profiles (the $\lambda$1400 feature, C IV $\lambda$1549, C III] $\lambda$1909, and Mg II $\lambda$2798) in high quality spectra of intermediate and high redshift QSOs (originally observed by Sargent, Steidel, \& Boksenberg for absorption line studies). The most striking trends are found with increasing line width ($\geq 3\sigma$, several $> 8\sigma$): the intensity ratios C III]/C IV, Mg II/C III], and $\lambda$1400/C IV increase, the equivalent width of C IV decreases, the C IV profile becomes less-sharply peaked, and the peaks of C III] and C IV become increasingly blueshifted relative to the peak of Mg II.
Additionally, we find that radio-loud objects have narrower C III] lines than radio-quiet objects, and that this depends on radio properties and not luminosity as has been previously reported. Other differences in correlations involving radio properties may be attributed to an extra axisymmetric continuum component present in the radio-loud objects not present in the radio-quiet objects.
We interpret these results in terms of a (currently) empirical two-component scheme: the combination of a blueshifted ($\sim$1000$ km\ s^{-1}$) broad (FWHM $\sim$ 7000 $km\ s^{-1}$) base and a narrow (FWHM $\sim$ 2000 $km\ s^{-1}$) core entering in different ratios can describe most C IV and C III] profiles reasonably well. The line width is determined by the ratio of the two components, as are other measurements of the profile. A changing core/base ratio can account for several of the above trends simultaneously, including the correlations between the line width and the line shifts, line ratios (not all of them, however), and line shapes. We propose some mechanisms which may give rise to two such quasi-independent components.
