As explained in Section 4.4.2,
in the pipeline processing of the two grating AOT's (L01 and L02),
the dark current is taken either as the average of the
two dark current measurements performed respectively at the begining
and the end of each observation, or as a `fixed'
dark current that was determined in dedicated calibration observations
(see Section 5.4).
For the
majority of sources, i.e. those above about 20Jy at 60 
m, the dark
current is negligible. For faint sources however, the uncertainty in the
dark current measurement can become large relative to the source flux level
and can lead to sub-spectra being misaligned or even to a negative
sub-spectrum if the dark current determined is higher than the
on-target flux. In such cases, it is better to revisit the dark current
subtraction in LIA. The user can then chose to use the `fixed'
dark current or a dark current previously
interactively derived within LIA (for LIA see Section 8.2.3).
Processing of Fabry-Pérot data (L03 and L04) systematically uses the fixed dark current.