Further discussion opened by Marco on spectro-l list.

I used the spectrum sent by Marco in my excel spreadsheet.

The results are :

Direct computation ( simple integration) EW = 16.03

Gaussian fit : EW = 15.93 and center = 6562.81 (excellent wavelenght calibration, Marco !)

(The Gaussian fit is a simple gaussian curve, with one compound).

Less than 1% diffrence between the two measurements.

Determining the edges of the line does not pose problems

Here's the integration range is [6553,6573]

If I adopt the range used by Marco [6548,6578], the EW becomes 15.96 : only 0.7% difference.

The real difficulty is : where is the continuum ?

François Teyssier