DEVELOPMENT OF PREDICTIVE MODEL FOR THE POTENTIAL ESTIMATION OF MOISTURE CONTENT OF LANSIUM DOMESTICUM SEEDS

Abstract

Lansium domesticum produces recalcitrant seeds that cannot tolerate excessive drying to allow them to be stored by conventional means without impairment. Moisture content in such seeds is crucial for handling purposes. It is commonly determined as destructive method according to ISTA after drying at 103°C for 16 hours. Non-destructive method for the determination of seed moisture content is yet to be explored. In this study, the potential of predictive model development for the estimation of seed dry weight of L. domesticum was explored. The morphological parameters and wet weight of seeds were used to develop predictive models for dry weight of seeds. The estimated seed dry weight was then used to calculate the seed moisture content. To obtain seeds with varying moisture contents, seeds freshly extracted from ripe fruits were cleaned and dried to remove the mucilage. Seeds were randomly drawn daily and their dry weight was determined according to ISTA. Two thirds of the data of each desiccation period were pooled and used to determine the relationships between the length (L), diameter (D), thickness (T), derived area, A (i.e., L×D), derived volume, V (i.e., L×D×T), wet weight (Fw), Fw/A, Fw/V and mass of seeds by using correlation analysis and scatter plot. Then, several regression models were developed for estimation of dry weight of seeds. Such candidate models were validated using the remaining dataset. This study indicated that the measured seed variables have the potential to be used for estimating the seed dry weight. When the estimated seed dry weight was used to calculate the seed moisture content, the estimated seed moisture content showed significant correlation (P≤0.05) with the observed seed moisture content with Pearson’s correlation coefficient (r) of 0.48. Thus, the development of predictive model for the estimation of seed dry weight for calculation of seed moisture content provides an important basis for future research on manipulation of seed moisture content for germplasm conservation of desiccation sensitive tropical fruit species.