ATM Active mirror figure correction?

[Lawrence Sayre <lsayre@stratos.net>]

I own a 3/4" thick plate glass Coulter mirror (13.1" dia.) which has a
smooth figure, but is under corrected.  This is quite evident via the
star test method, and was confirmed by Steve Swayze.

Is it possible to glue a small "swivel level" pad onto the dead center
of the back side of this mirror (using an appropriate glue, such as the
glue used to adhere the rear view mirror to a windshield), and then pass
the swivel level bolt through the back side of the cell (where it would
be as accessible as the three collimation bolts), and then torque a nut
on the swivel level bolt to literally pull this thin plate glass mirror
into a good parabolic curve?  Or would this induce loads of stress
related astigmatism, or worse, shatter the mirror?  The mirror cost less
than the fix (re-figuring), so I'm willing to attempt this if it sounds
possible.

PS: If this is possible, what glue should I use?

Lawrence Sayre