Metoprolol and a number of related amino alcohols and similar analytes have been chromatographed on aminopropyl (APS) and ethylpyridine (EPS) silica columns. The mobile phase was carbon dioxide with methanol as modifier and no amine additive was present. Optimal isocratic conditions for the selectivity were evaluated based on experiments using design of experiments. A central composite circumscribed model for each column was used. Factors were column temperature, back-pressure and % (v/v) of modifier. The responses were retention and selectivity versus metoprolol. The % of modifier mainly controlled the retention on both columns but pressure and temperature could also be important for optimizing the selectivity between the amino alcohols. The compounds could be divided into four and five groups on both columns, with respect to the selectivity. Furthermore, on the aminopropyl silica the analytes were more spread out whereas on the ethylpyridine silica, due to its aromaticity, retention and selectivity were closer. For optimal conditions the column temperature and back-pressure should be high and the modifier concentration low. A comparison of the selectivity using optimized conditions show a few switches of retention order between the two columns. On aminopropyl silica an aldehyde failed to be eluted owing to Schiff-base formation. Peak symmetry and column efficiency were briefly studied for some structurally close analogues. This revealed some activity from the columns that affected analytes that had less protected amino groups, a methyl group instead of isopropyl. The tailing was more marked with the ethylpyridine column even with the more bulky alkyl substituents. Plate number N was a better measure than the asymmetry factor since some analyte peaks broadened without serious deterioration of symmetry compared to homologues.