Abstract
The principles for designing a calibration scheme for a set of mass standards are discussed. On these principles, a particular scheme has been chosen as being convenient and reasonably efficient; it is pointed out that although it has not been possible to decide whether the selected scheme is optimum, it can be said that it is more efficient than most. In an actual calibration based on this scheme, the accuracy decreased from 1.5 in 107, to which the Materials Research Laboratories 1 kg standard is known, to 1.6 in 108 for the 1 g mass; such a decrease in accuracy is shown to be a property of all such schemes.