Abstract
Various approaches to the study of random variables with possible values in the non-negative integers are discussed in accordance with the generalized Poisson law distribution. In the study of vehicular traffic, the customary assumption of random distribution of arrival times of cars has been found to be slightly unsatisfactory because: (1) very small gaps do not occur as frequently as that theory would require, and (2) there is some interference between vehicles, inducing an element of regularity into arrival times. The alternatives that have been suggested are presented. The assumption of linearity applied to an equation has proved accurate to three places. However, even this equation cannot be exactly correct.