Lazarus (see Plasma Physics, vol.25, p.1271, 1983) discusses the merits of analysing Thompson scattering data by finding the vest fit to a straight line on a log-linear plot of Thomson scattered intensity versus the square of the wavelength shift of the scattered photons from line centre. He claims that there is an inherent error in linearizing the regression which is due to 'unintended weighting given to the low intensity data points'. It is the purpose of the Comment to show that the error is not due to the linearization, but that this 'unintended weighting' arises from incorrect transformation of the error bars of the data points, and an analytic expression is derived which transforms the error bars without distorting the weight of the individual points.