We have discovered two low-ionization BAL QSOs with strong absorption from metastable excited states of Fe II and III (``iron Lo-BALs''). There has been just a single member of this class until now, 0059-2735, with Hawaii 167 another likely example. One of the new objects, 1556+3517, is the first known radio-loud BAL QSO. The three iron Lo-BALs suggest a correlation between radio luminosity and absorption. We suggest that the iron Lo-BALs are transition objects evolving from radio-loud to radio-quiet BAL systems as the QSO emerges from the enshrouding material. The two new FIRST Survey objects and Hawaii 167 were all discovered in limited surveys at wavelengths longer than optical, suggesting that these objects escape detection in most optical surveys and may be common, perhaps equal in population to the radio-quiet Lo-BALs.
Whether considered extreme or transition, the iron Lo-BALs may hold clues to several aspects of the quasar phenomenon, starburst activity in the early Universe, and the relation of both to galaxy formation. Egami et al. (1996) emphasize that the heavy obscuration of the central QSO allows the surrounding galaxy starlight to be detected, opening a window on QSO galaxy hosts at higher redshift. If this is accurate, then 1556+3517 should be even more galaxy-like in the IR than Hawaii 167 because of its higher extinction of the QSO source. It is possible that these objects are a ``missing link'' between galaxies and quasars.
We acknowledge Jules Halpern for first recognizing the nature of 0840+3633 and for additional helpful discussions. We thank Paul Hewett for the spectrum of 0059-2735 and Michael Brotherton for many useful comments and suggestions. We acknowledge support from the NRAO, the NSF (grants AST-94-19906 and AST-94-21178), the IGPP/LLNL (DOE contract W-7405-ENG-48), the STScI, the National Geographic Society (grant NGS No. 5393-094), NATO (grant CRG 950765), and Sun Microsystems. This paper is Contribution Number 620 of the Columbia Astrophysics Laboratory.