The calculated flux distribution in the 300 to 400 nm region only fits the observed distribution closely for both the Sun and Arcturus when the complete ATLAS9 line list is used. However, a good fit to the detailed shape of the observed high resolution spectrum in the Sun may only be obtained by using a sub-set of this line list that only includes lines with the highest quality atomic data. Moreover, for Arcturus, even the sub-set of high quality lines can only provide a close fit to the observed high resolution spectrum if we include, ad hoc, an extra amount of continuous absorption opacity. We suggest that the complete ATLAS9 line list contains many lines that either do not exist, or have oscillator strengths that are too strong, and that the inclusion of a hitherto undiscovered source of continuous absorption opacity is necessary to fit the violet flux in late-type stars.