Many models for the analysis of capacity and signal timing of signalized intersections are based on the assumption that queue discharge will quickly reach a steady maximum rate after the green interval begins. This assumption leads to a common practice of using saturation flow and lost time as parameters in modeling the requirements for lane capacity, intersection performance, and signal timing. However, recent field data indicate that the queue discharge rate may continue to increase long after green onset. For this type of queue discharge characteristic, the traditional applications of lost time and saturation flow may result in significant errors when lane capacity and modeling optimal cycle length are estimated. The extent of such errors is assessed on the basis of the queue discharge characteristics observed in 38 urban traffic lanes.