The Kolmogorov?Smirnov (K-S) test for goodness of fit was developed by Kolmogorov in 1933 [1] and Smirnov in 1939 [2]. Its procedures are suitable for testing the goodness of fit of a data set for most probability distributions regardless of sample size [3-5]. These procedures, modified for the exponential distribution by Lilliefors [5] and Durbin [6], are yet to be incorporated in most undergraduate texts on probability or statistics. The modified K-S test measures how well a given set of observations fits some cumulative exponential distribution function F(x). In this article, we discuss (1) the definitions of the distributions F(x) and F*n(x), (2) Lilliefors's and Durbin's modifications of the K?S test for goodness of fit for the exponential distribution, and (3) the construction of a simple spreadsheet that facilitates the computations associated with testing whether the inter-occurrence times in days of fatal accidents suffered by scheduled large and small planes in the USA (from 1983 to 1998) are exponentially distributed. We also apply the K?S test for goodness of fit to analyse inter-occurrence times in days of fatal accidents suffered by scheduled planes in Britain (from 1946 to 1963) that were previously analysed by Pyke [7, p. 433] using the method of spacing. Taken together, results of the tests raise questions about a forgetfulness property for the inter-occurrence times of fatal accidents suffered by large planes.