The factors that control the distribution of heavy minerals in natural sandstones are so numerous and complex that their separate effects are difficult or even impossible to untangle. However, a theoretical approach to the problem indicates something of the relative importance of each. Factors such as differences in the density and hardness of the various minerals, differences in the original size of the various mineral grains in the source rock, the amount of abrasion that all grains have undergone during transportation, the different settling velocities of the various grains at the site of deposition, and the degree of sorting to which all grains were subjected there--all these factors seem competent to cause large variations in the relative abundance of various minerals in different samples of deposits that have been derived from exactly the same source rock and also in different size fractions of each of the samples. Certain of the factors, such as abrasion, tend to concentrate the heavier minerals in those samples that are predominantly finer grained; whereas other factors, such as sorting, tend to concentrate the heavier minerals in the finer grained portions of each sample, regardless of whether that sample is predominantly fine-grained or coarse-grained. The selective effects upon different minerals of the various influences, as deduced from theoretical considerations, appear to be confirmed by the evidence of actual sand samples. These effects seem sufficiently large to merit attention when using heavy mineral suites as a basis for correlation or for inferences about the character of source rock. Unless all the samples to be compared are of approximately the same grain size and the same degree of sorting, it is suggested that, in order to make allowances for these effects, heavy mineral separates be taken from two selected size fractions of each sample--one, representing the same actual grain size, the other, the same relative grain size in each sample.