Messy Eating

Mergers at z=0 leave lots of crumbs behind (Hibbard & van Gorkom 1996)

A significant fraction of the extended HI in disks can be sent to many tens of kpc in mergers, where it can slowly "rain" back onto the remnant (Spergel & Hernquist 1992; Hibbard & Mihos 1995).

Further disk building?
Gas rich ellipticals?
QSO absorbers at high-z?

Mergers are not very efficient at scrambling energy (White 1978): tightly bound material stays tightly bound. To send large quantities of gas to large radius, you need lots of loosely bound gas in the progenitor.

This can be done through a shallower potential well or by a more extended gas distribution (Barnes 1988; Dubinski etal 1996; Mihos etal 1998; Springel & White 1998; Barnes 1999; Mihos 2000).

Are gaseous galaxy disks more extended at higher redshift?
How would that affect tidal gas properties?

Inside-out models: Gas accretes in the inner portion of the halo first; outer regions later -- disk grows with time.	galaxies at higher redshift do seem to be systematically smaller (Lowenthal etal 1997; Dickinson 2000). Better explain chemical abunances (Matteuchi, this conference). seen in disk galaxy formation models (eg Kauffmann 1996; Mo etal 1997)	High redshift mergers may be neater: more gas thrown down the well, less ejected back into extended (unbound?) tidal debris. Bad for tidal debris models for QSO absorbers Old ellipticals less able to reaccrete. (Gas poor old E's, gas rich young E's?)
Extended disk models: Gas initially more extended. (contracts with time due to angular momentum transfer / secular effects?)	better explain kinematics of DLAs (Maller etal 2000) slow buildup of disk at later times? (eg Steinmetz 2000)	High redshift mergers may be very messy. Possibly better for QSO absorber models, but not metal systems. Lots of gas to continue accreting.