A heterogeneity study was conducted at Erdenet Copper Mine in Erdenet, Mongolia, to determine the fundamental sampling error associated with blasthole sampling. Using a custom-designed radial sector sampler, six large blasthole samples were collected to perform the test developed by Pierre Gy to estimate the constitutional heterogeneity factor for several types of ore. The study investigated the entire sampling protocol, from mine site to analytical laboratory, and quantified the contributions to total fundamental sampling error (FSE) of the various stages of sampling and sample preparation. The total relative standard deviation (RSD) of sampling error was determined to be ±32 percent and ±49 percent for copper and molybdenum, respectively. A classical statistical analysis conducted on the data confirmed the level of RSD native to the mine. About 95 percent of the sampling error was produced from one step in the sampling protocol, that of reducing the 1-kg (2.2-lb) sample to a 50-g (0.1-lb) sample for chemical analysis. To minimize the FSE, the use of a slightly different protocol is recommended. The 1-kg sample (2.2-lb) should first be crushed to a 0.2 cm (0.08 in.) particle size. A 250-g (0.5-lb) split should then be taken and finely ground, from which a 50-g (0.1-lb) split should then be taken. These simple steps will reduce the FSE to ±7.9 percent and ±17 percent for copper and molybdenum, respectively. The modified sampling protocol effectively reduces to zero the probability of misclassification of low-grade ore.