Abstract
This paper discusses the detection of P V 1118,1128 and other broad absorption lines (BALs) in archival HST spectra of the low-redshift BALQSO, PG 1254+047. The P V identification is secured by excellent redshift and profile coincidences with the other BALs, such as C IV 1548,1550 and Si IV 1393,1403, and by photoionization calculations showing that other lines near this wavelength, e.g. Fe III 1123, should be much weaker than P V. The observed BAL strengths imply that either 1) there are extreme abundance ratios such as [C/H] >~ +1.0, [Si/H] >~ +1.8 and [P/C] >~ +2.2, or 2) at least some of the lines are much more optically thick than they appear. I argue that the significant presence of P V absorption indicates severe line saturation, which is disguised in the observed (moderate-strength) BALs because the absorber does not fully cover the continuum source(s) along our line(s) of sight. Computed optical depths for all UV resonance lines show that the observed BALs are consistent with solar abundances if 1) the ionization parameter is at least moderately high, log U >~ -0.6, 2) the total hydrogen column density is log N_H(cm-2) >~ 22.0, and 3) the optical depths in strong lines like C IV and O VI 1032,1038 are >~25 and >~80, respectively. These optical depths and column densities are at least an order of magnitude larger than expected from the residual intensities in the BAL troughs, but they are consistent with the large absorbing columns derived from X-ray observations of BALQSOs. The outflowing BALR, at velocities from -15,000 to -27,000 km/s in PG 1254+047, is therefore a strong candidate for the X-ray absorber in BALQSOs.