Principal component analysis was used to construct plots with the PC1 versus PC2, and PC1 versus PC3 axes accounting for more than 72.7% and 63.5% phenotypic variations, respectively (Fig. 4).Samples with similar values for the variables explained by the PC appeared closely related. Principal component analysis score plot provided an overview of the similarities and differences between samples. Most herbal teas that are located at the left part of the coordinates can be characterized by low content of phenolic and low antioxidant capacities (Fig. 4). Moving along PC1 from bottom to top in the graph, most of the samples with low activity were grouped together. By contrast, Yue Ji Hua (H90) and Teng Cha (H109) with the higher antioxidant activity were separated and were quite far from other herbal and green teas. Five herbal teas and all green teas were grouped together because of their similar antioxidant capacities and color parameters (Fig. 4).