Effect of source-sink relationships from the branch to the tree scale on mango fruit size and quality at harvest

Mango fruit is a tropical crop characterized by heterogeneity in fruit size and quality at harvest. The effect of source-sink relationships on fruit production have been reported in many species but their interplay at the branch, scaffold and tree scales has been little investigated. This study aims at evaluating how source-sink relationships from a local (branch) to a global (tree) scale affect fruit production. Experiments were conducted on mango trees ('Cogshall') in Réunion Island. Fruit-bearing branches were randomly selected on six trees with low to high fruit load. Half of them were girdled and ratios of 10 and 100 leaves fruit‑1 were established on girdled and non-girdled branches. Pulp dry matter content, titratable acidity and total soluble solids were measured on one ripe fruit per branch. Number of fruits, yield and trunk and scaffolds cross-sectional areas were collected on these trees, and on other trees over several years. Average fruit fresh mass at the scaffold and at the tree scales was negatively related to fruit load, expressed in number of fruits per scaffold and per trunk cross-sectional area, respectively. A global negative effect of tree fruit load on average fruit fresh mass at the scaffold scale was also observed. Between-trees variability of average fruit fresh mass decreased with fruit load. At the branch scale, low leaf-to-fruit ratio and girdling induced fruit drop in the early stages of fruit development and reduced fruit growth, pulp dry matter content and total soluble solids. Results suggested partial autonomy of the scaffolds and fruit-bearing branches with respect to fruit growth. As a next step, a modeling approach integrating carbon-related processes from the branch to the tree scale should be used to further investigate to what extent source-sink relationships, and other factors such as light environment or water availability, can explain fruit production variability.