Hi Jeremy,
I think this is probably fixed in the current 1.99c (pre 2.0 build).
I definitely recall working on a bug like this and ultimately making
the scaling code for isosurfaces and volume rendering identical. I
will send you a 1.99c build: please reply on this list to indicate if
the bug is still present, and then I will need a demo program and data
from you.
thanks - David.