ABSTRACT
The data are a special form of time series. A wide variety of aspects may be explored and it is necessary therefore to be guided to some extent by the practical object in mind. There are many ways of checking agreement with exponential form. There is a very simple representation of the data, based on exponential distributions that would make interpretation vivid. Fittings are essentially straightforward and comparison of the maximized log likelihoods allows various approximate significance tests to be made. Both aircraft show a high proportion of small values but otherwise suggest irregular variation, rather than any systematic departure from an exponential distribution. There is no suggestion that ß^ is related to the mean and the authors do not have other information about the individual aircraft on which to base an explanation. It is possible that the variation is in some sense unreal, a fairly extreme random fluctuation, conceivably amplified by any positive correlation between intervals within aircraft.
