ABSTRACT
H. E. Almer and J. Keller published an exhaustive presentation of surveillance test procedures. Surveillance testing looks for signs that one or more members of a weight set may have changed value since the latest calibration. The basic idea is to ensure the self-consistency of the weight set. Two distinct types of surveillance test are referred to as Type I and Type II tests. The object of the Type I surveillance test is to inter-compare all weights in a set using a minimum number of steps. The standard deviation (SD) used is the estimate of the SD of the balance being used for the surveillance measurements; this uncertainty is known from many measurements made previously on the balance. A surveillance chart would be similarly constructed and studied to detect values, that fall outside the surveillance limits. If any fall outside, one would deduce which of the weights in the weight set had changed significantly between successive surveillance measurement sequences.
