Testing the ergodic assumption in the low-intensity interference experiments

It is well known that one must make an ergodic-type assumption in order to interpret the laboratory time averages in the low-intensity interference experiments as agreeing with the quantum-mechanical ensemble averages. In this work it is analysed the possibility of statistically testing this ergodic assumption. For the case of an ideal laser (i.e. a deterministic intensity), the number of photon-counting measurements in an experimental run and the number of experimental runs necessary to statistically compare the various time and ensemble averages are calculated