This paper describes an improvement of the procedure, proposed previously (Spectrochim. Acta, Part B, 48 (1993) 1517), to provide information about multiplicative and spectral interferences at lines selected for the analysis of samples. The new version of the procedure only requires intensity scans across an analyte line when three solutions are aspirated: a standard solution, the sample solution and the sample spiked with the analyte standard solution. The third solution can be omitted if multiplicative interference is not of interest. Like the initial method, the present procedure does not require or assume information about the sample composition nor is the preparation of solutions of suspected interferents required. Figures of merit, namely the structured background factor, and the overall spectral interference have been devised and are used with the true detection limit at a line for line evaluation and selection. As the nonanalyte concomitant contribution to the measured spectrum is estimated, automatic background correction is possible. The proposed method not only allows selection of the spectral line, but can give a good estimate of the analyte concentration prior to a full quantitative analysis.