3067 leaf reflectance spectra data from all leaves of a single tree

As our objective was to gather spectral data for every leaf on a single tree, we opted for a relatively small-sized tree. At first, we stimulated the shedding of senescent leaves by gently shaking the tree trunk. Subsequently, we utilized pruning shears to gather terminal twigs (Fig. S1). Following collection, the twig samples were packed in ziplock bags along with moist absorbent paper. Subsequently, these bags were placed in a portable cooler together with ice packs to reduce water loss during transportation. Upon arrival at the laboratory, we proceeded to select all leaves from the twigs. Subsequently, after rinsing them extensively with deionized water, we meticulously removed surface moisture using lint-free paper. After the preprocessing, we immediately collected the fresh leaf reflectance spectra (350-2500 nm) using the ASD FieldSpec (ASD Inc., Boulder, CO, USA) with a leaf contact probe with the calibrated internal light source to minimize errors associated with stray light. Moreover, to improve the spectra measurement quality (i.e., to reduce the noise ratio of spectra signal), we optimized the spectrometer every 10 minutes using a calibration white referenced panel (Spectralon, Lasphere, Durham, NH, USA). We utilized specialized software (ViewSpecPro) to open and inspect each spectral curve. In instances where noticeable errors were detected, the corresponding leaves underwent re-measurement. Concurrently, we recorded the leaf order (LOR) from the top to bottom of the twig. For instance, the first leaf order (LOR1, and subsequent abbreviations from LOR1 to LOR7) corresponds to the uppermost pair of fully expanded leaves (Fig. S1). B. gymnorhiza typically displays opposite phyllotaxy, wherein a single leaf order (LOR) on a twig corresponds to two leaves under normal circumstances. During transportation, some leaves became detached, and we did not document the order of these separated leaves.