The wavelength scale was determined using known positions of narrow emission
lines in the symbiotic star spectra listed above.
Aside from small offsets necessary to correct
for telescope pointing errors along the dispersion direction,
the results for different objects proved consistent within
measurement error. These measurements revealed deviations from a linear
wavelength scale by as much as 1.5 Å.
The data were fit with a four-term Legendre polynomial, resulting in an RMS
error of 0.236 Å. Additional terms
did not significantly reduce the overall RMS error of the fit.
Comparison of laboratory and in-flight positions of Lyman-
,
Lyman-
, and the calibration lamp lines shows agreement
to within 1 pixel (0.51 Å),
indicating no disturbance of the internal spectrograph alignment during launch.
Consistency checks of the wavelength scale were made using measurements of
spectra in which the slit was uniformly illuminated and which, therefore, do not
require telescope pointing corrections to the wavelength scale.
The extended sources chosen
for this purpose were a daytime airglow spectrum and a lunar (essentially
solar) spectrum. Although the RMS deviations about the adopted wavelength
scale are larger due to the
increased linewidth and blending, the data indicate that the mean
offset of the wavelength scale is 
Å.