The CES suffers from scattered diffuse light, that contaminates the background.
This light has another
wavelength dependence than the normal spectrum.
All measurements were done using the saturated (interstellar) NaI D lines
(5890 Å and 5896 Å) seen in certain stars.
The uncertainties are internal errors estimated from the S/N of the spectrum, and may not
reflect the true accuracy of the measurements. Only the high-resolution slicer is monitored.
This information is
also available in tabular form.
At other wavelengths one must use a different approach, based on measuring equivalent widths (EW) in
a solar spectrum obtained with the CES and comparing with a "canonical" solar spectrum. Such monitoring
has recently (Jan. 2005) been started. Results are given only in tabular form.
Users aiming at measuring accurate equivalent widths should correct their spectra for scattered
light, using either the values quoted in the table, or (better) by estimating the scattered light in Solar Spectra.
We are building a database of calibrations from where the relevant data
can be downloaded.