Note that for longer integration times (> 5h), the factor f is expected to
increase slowly (e.g. 
for 
= 5h). For broad weak lines,
because of baseline problems, it is advisable to multiply the observing times
by 1.6 or so, so that 2 independent sets of data can be obtained.
If we had used the PSW mode instead of the WSW one, although theoretically R
decreases, poorer baselines would probably lead to an increase of the total
observing time. The corresponding factor is difficult to figure out as it
depends on the atmospheric conditions. A factor of 
on 
would
however be recommended for weak signals.