Stigmatic balloon spectra of the Sun in the vicinity of 2800 Å were obtained on September 22, 1968 and April 30, 1969. We compare the observed profiles of the H and K lines of ionized magnesium with the computed profiles of Athay and Skumanich and of Dumont. The discrepancy between observed and computed profiles of the Mgii lines is considerable. It is shown that this is a general fact for all chromospheric resonance lines of abundant elements. A brief review is given of different interpretations of this discrepancy.