SUPPLEMENTARY DATA

Experimental Design and Materials The dataset used in this study comprises 448 individual fragments from a representative gold-bearing ore. The experimental design comprised the following elements: Material: A representative gold ore subjected to staged comminution and screening, selected to reflect the compositional and textural diversity of the deposit. Fractionation: The ore was crushed and screened into 14 narrowly sized fractions. The nominal sizes (geometric mean of adjacent sieve apertures) ranged from 17.63 mm down to 0.12 mm. Sample set: A total of 448 individual samples, each consisting of numerous fragments, were riffle split from each of the 14 different size fractions. Each of the 14 size fractions consisted of 32 samples, weighed individually using a calibrated analytical balance (masses recorded in grams), and subjected to individual fire assay for gold content (grades recorded in g/t to two decimal places). Fire assay was performed using a conventional lead collection method with a flux mixture of litharge/soda/borax/silica at 1050–1100°C. The resulting prill was cupelled and parted in dilute nitric acid, and the gold residue was weighed or analysed by AAS/ICP with a detection limit of approximately 0.01 g/t. Duplicate assays of selected samples showed a relative standard deviation of <5%. Outputs: Each sample yielded a mass (g) and a gold grade (g/t). These paired values constitute the primary dataset from which and are computed for each sample in each size fraction. The sample masses and gold grades are summarised by fraction in Table 1. The data encompass 14 size fractions with nominal sizes: 17.63, 14.73, 10.42, 8.81, 7.41, 5.89, 4.17, 2.84, 1.65, 0.88, 0.62, 0.41, 0.19, and 0.12 mm.