English: Intensity corrected version of this image: Image:Spectrum of blue flame.png taken on same hardware setup but using the irradiance mode of the spectrometer where the nonlinear response of the detector to the intensity of light falling on it is corrected for. The region below ~420nm is very noisy due to the lack of light at those wavelengths in the tungsten halogen lamp which is used for intensity calibration, all information below 375nm is too noisy to be useful. Spectrum taken by me using an Ocean Optics HR2000 spectrometer [1]. The flame of the torch was placed adjacent to the fiberoptic input port of the spectrometer so that attenuation of the shorter wavelengths by the fiber could be avoided. This may have introduced a slight "smearing" of the spectrum due to the lack of collimation of light entering the spectrometer but in comparing this spectrum to one taken through the fiber there is very little if any difference between the two. The data from the spectrometer used to create this image is in CSV format on the talk page.