The optimization step was carried out using a two-level full factorial design, and the effect of the three factors on the high and low level was evaluated: type of solvent (methanol or ethanol), percentage of the solvent in water (100% or 60%) and range of extraction times (1.5 or 3.0 min). Minimum and maximum levels of each factor (Table 1) were chosen according to flavonoid extraction data in the literature [7,12], and also considering preliminary experiments using the passion fruit pulp. The relative area of the peak isoorientin (area of isoorientin peak at λ=330 nm/total area of flavonoids at λ=330 nm, Fig. 2) was used as a dependent variable, expressed in percentage.