highly significantP < 0.0001
As shown in Table 5 , the model was highly significant ( P < 0.0001), and the lack of fit was not significant ( P = 0.2434 > 0.05), indicating the model was able to predict the content of saponins under any combination of values of the variables.