Describing the storage behavior solely via complex indices such as RPIKS, TSS/TA, and h◦m did not always clearly reflect individual quality stages after specific storage periods. For example, semi-fast-ripening and fast-ripening variants displayed acceptable TSS/TA ratios already on day 10±1, but partly at poor TSS (N3B,N4C). Therefore, fruit of similar quality irrespective of the storage time was grouped by means of PCA and CA (Fig. 3A) that were applied to an aggregate matrix comprising the mesocarp attributes recorded on days 10±1 and 18±1 for all 11 fruit lots. The identified
groups were subsequently compared with the harvest maturity clusters and subgroups (Fig. 1A). Because of apparently complete chlorophyll degradation in the course of storage, only 7 of the 9 independent mesocarp attributes of Tables 2–4 were used, in conformity with the previous study (Kienzle et al., 2011). Unlike the color differences between the fruit lots one day after harvest (cf. 3.1), those observed during storage were clearly expressed by both the Cartesian (Table 2) and the polar (Table 5) coordinates. Hence, the Cartesian system, which was favored over the polarone by Padda et al. (2011) for monitoring ripening-associated color changes, was included (Fig. 3B)