Of late, there has been a considerable volume of empirical research on freeway operations during periods of high flow. Part of the motivation for this realm of research is the attractive possibility of increasing flow and speed by preventing breakdown in congested operations. This possibility has been fueled by observations of a "capacity drop," wherein the discharge flow from the resulting queue is smaller than those observed before breakdown. The research on which this paper is based was aimed at exploring in more detail the implications of metering traffic flow at levels below that at which a breakdown occurs. From this perspective, the nature of precongestion flows is investigated empirically to determine the potential gains from such a policy. The most important contribution is an examination of the likelihood of success of such metering by establishing the probability of breakdown at various traffic flow levels. It is shown, for example, that 1-min median lane flows that are 20 percent larger than queue discharge flows have only a 10 percent probability of breakdown and that 1-min flows equal to the mean queue discharge flows have a negligible probability of breakdown. These results not only have practical significance but also are useful for improving the efficiency of freeway simulation models.