These are 20 inch air gap (actually just water) simulations
for modules M3 and M4b.  The plots show the effect of a
(minus) 5% change in widths.  The SRF is modelled using our
"z_srf" routine with gfrac=0.91, gexp=0.78, and lexp=2.5.

The bottom plot is the measured width from the test 1191 run.
The freqs have been shifted to adjust for the finite field of
view.

I have strong doubts that these gas cell/air tests are going
to work.  The 1266-77 gas cell tests are poor, so unless
something has changed, we should expect failure.  I have not
played sround to see if anything can be salvaged from the
gas cell tests.  And I can;t think of any other better test
to help determine if the width variation in AIRS is real,
so maybe we are stuck with the planned test.

Scott.