The subject of BAL QSOs could be addressed either at a workshop on QSOs and AGNs, where discussion about the emission line region is emphasized, or at a workshop such as this one, where the origin of the narrow absorption lines seen in QSO spectra is an issue. In my opinion there is strong evidence that the BAL region gives rise to a component of the broad emission line region and, at the same time, that the phenomenon producing the BALs may give rise to narrow absorption line systems. There have been four other review papers dedicated to the subject of BAL QSOs (Weymann and Foltz 1983; Turnshek 1984b; Weymann, Turnshek and Christiansen 1985; Turnshek 1986). This current review is not designed to replace the previous ones, but is designed to complement them. I am aware of no serious errors in the previous reviews. BAL QSOs characteristically exhibit ultraviolet resonance line absorption due to highly ionized species typically extending from outflow velocities near 0-5000 km s-1 up to 10,000-30,000 km s-1. There are a great variety of BAL profile types, but they occur in otherwise relatively normal QSO spectra.
The first two BAL QSOs were discovered by C. R. Lynds (1967) and E. M. Burbidge (1970), but the identification of these types of QSOs did not become routine until objective prism surveys became a common way of discovering QSOs. Given the fact that BAL systems are widely accepted to be intrinsic to the QSO, there are basically two extreme interpretations for them which should be reviewed at the outset. One possibility is that the BALs may occur in only a peculiar subset of QSOs comprising about 3-10% of the QSO population. In this interpretation the BAL region covering factor must be large, implying that an observer would see BALs in a BAL QSO's spectrum from all lines-of-sight. The early model of Scargle, Caroff and Noerdlinger (1970) adopted this interpretation. The other extreme possibility is that the value of the BAL region covering factor is not near unity, but is approximately equal to the fraction of QSOs that have BALs. In this case an observer would see BALs in a QSO's spectrum only if he were looking in a preferred line-of-sight which intersected BAL clouds. This interpretation would be in agreement with the early model put forth by Lucy (1971). Of course, the consequence of this latter interpretation is that all QSOs may have BAL regions.
As I noted earlier, it is reasonable that a talk on BAL QSOs be given both at a conference discussing QSO/AGN emission line regions and at a conference on (narrow) QSO absorption lines. Therefore, for this conference I discuss both of these topics. First, I review the observations. Then I review the properties that a model for the BAL region must possess. Finally, I consider evidence which may suggest that the narrow absorption lines seen in QSO spectra are related to the BAL phenomenon. I suspect that this latter topic may be of greater interest to the audience of this workshop.