The evidence that at least some mergers of gas-rich disk galaxies can make elliptical-like remnants is very strong (e.g. Schweizer 1998). Whether these merger remnants are true ellipticals or anomalous in some manner is still a matter of debate (see van den Marel & Zurek, these proceedings). H I observations addressed one important aspect of this question: do mergers get rid of the atomic gas of the progenitors? It has often been stated that H I will be ejected into the tidal features, but in fact at least as much (and likely much more) outer gas should be sent into the inner regions as is found in the tidal tails (see Fig. 15 of Toomre & Toomre 1972). It was therefore reassuring to find that progressively more advanced merging systems have less and less atomic gas in the bodies of the remnants (Hibbard & van Gorkom 1996).
It was not clear how most of the original atomic gas was removed from the inner regions, or how they remain largely H I free in light of the H I which continues to fall back from the tidal regions. Recent observations have shed some light on this subject, by showing that two processes - galactic superwinds and ionization by continued star formation - can have a strong effect on the observability of tidal H I (Hibbard, Vacca & Yun 1999). Superwinds are likely to be important in helping the most gas-rich systems get rid of much of their cold gas reservoirs, but the wind phase is short lived, and would not explain the continued removal of returning tidal H I. Simple calculations suggest that the UV flux from on-going starformation is sufficient to ionize diffuse H I in the tidal regions (see also Bland-Hawthorn & Maloney 1999).
Photoionization is an attractive mechanize for explaining tidal features which are gas rich in the outer radii, but gas poor at smaller radii (e.g., the northern tail of The Antenna, Fig. 1; Arp 105 Duc et al. 1997; NGC 7252 Hibbard et al. 1994). The dynamics of tail formation require that the gas-rich outer radii of the progenitor disks extend all the way back into the remnant (see Fig. 2 of Toomre & Toomre 1972). The geometry of a preliminary numerical fit to the NGC 40389/9 data (Hibbard, van der Hulst & Barnes, in preparation) suggest that the northern tail has an unobstructed sightline to the numerous starforming regions in the disk of NGC 4038, while the southern tail does not, explaining why it remains gas rich along its entire length. This process may explain how merger remnants remain gas poor in the presence of the continued return of tidal H I. Such an on-going process is required if remnants are to evolve into normal ellipticals in terms of their atomic gas content.