Hi Ernst,
I really appreciate your feedback. It is fascinating to see the comparison with the high resolution campaign results, there seems to be a reasonable correspondence. The integration interval that I use is determined by the points at which the Hα obviously emerges from the underlying continuum, so the width of the interval can vary slightly from spectrum to spectrum. So for example the width of the interval on my last spectrum captured on 11th Feb was 6554 - 6576Å, on some other spectra it has been maybe 1 or 2Å wider. Regarding the accuracy, I have been wondering for a while how to estimate the error bar I should asscoiate with the values in the graph. Would you have any advice on how to estimate that when the true value may be varying naturally from night to night?
Cheers
Hugh