Ok, Robin, I'm convinced. That looks pretty nice! The continua match perfectly as if rectified at this very same wavelength. That is how the result should be. Now Berthold and you showed the real variability within the line peaks.
Just today I got the following meassage from Tony Moffat in Montréal. He made some quick calculations.
Assuming the length of the observable night is ~12h at Teneriffe on Jan 25 (the worst night to be observing WR140!), then the hour angle is +6h and increasing at the beginning of the night and -6h and decreasing - but the sky is getting bright! - at the end of the night. [On the best night, July 25, one has HA = -4h and +4h, resp.]
So, one will have to observe WR140 at HA = +/- 6h at this time of year. Of course, as you go off this date, things get better slowly at first. Anyway, HA = +/- 6h corresponds to airmass X = 3.0, increasing to 5.9 at HA = +/- 7h.
This means that you don't have much time to grab a spectrum on Jan 25, maybe 30 minutes or so. One way around this would be to observe more into twilight, although at this high airmass, the sky is already significantly brighter than at lower airmasses.
Bottom line: NOW is the time to make some tests (i.e. between say early January and mid February), to see how good the spectra will be, ready for next year when it really counts. Can you get your troops on it now?