It is a bit difficult to tell from the scaled stretched image posted but with the configuration (35mm fl lens, 500 l/mm grating, 11.5mm wide sensor, rescaled to 640 pixels, the linear dispersion will be ~10.3A/pixel in the posted image (small angle approximation) The weak lines at the right hand edge are therefore likely to be the second order image of the group at the left hand edge. With this information we can estimate the location of the zero order and produce a wavelength calibrated spectrum.The calculated calibration is also in agreement with the quoted approx 3600-6500A response
(The small angle approximation will not hold precisely at the dispersion angles from a 500l/mm grating but the method could be finessed using angular dispersion in place of linear dispersion to produce an accurate wavelength calibration)
Cheers
Robin