[en] The validity of different colorimetric methods used to quantify various families of antioxidant compounds was evaluated with standard compounds. The colorimetric tests for global evaluation of flavonoids, anthocyans, and flavanols were found generally unreliable, as reactions could be different for individual compounds within a family (anthocyanins or flavonols or flavan-3-ols) and not specific to one family. In the flavonoid test, for example, flavonols reacted very well, anthocyanins did not react, and flavanons reacted only slightly. The same methods were applied also to beverages known for their antioxidant content (apple, orange, grape, and vegetable juices, ice tea, and red wine) and the data were compared with the results of HPLC analysis of specific compounds. The values obtained in a colorimetric test were generally higher than the sum of the values obtained for the corresponding individual compounds by HPLC analysis, mainly because other compounds can interfere with the colorimetric tests. For example, in wine, anthocyanin concentrations obtained by colorimetric test were 9068 1407 µ mol /100ml (mean SEM), higher than the sum of the six main anthocyanidins detected by HPLC, only 41 µmol/100 ml. The relative antioxidant capacity values determined for beverages on the basis of colorimetric tests could exceed by far the values previously measured in radical-scavenging tests (for instance, the antioxidant capacity attributable to anthocyans in wine on the basis of the colorimetric test was 50 times higher than the total antioxidant capacity measured by the ORAC assay). In conclusion, colorimetric tests for flavonoids, anthocyans, and flavanols appeared generally unreliable for estimating their content and thus the antioxidant capacity reliable to these compounds.