Traditional estimates of saturation flow rates have been based on a hypothesis that the time headways between successive vehicles entering an intersection from a traffic signal are constant after the first three to five vehicles are discharged. This model of constant flow has been challenged by recent field observations in which average headways between vehicles continue to decrease through the last queued vehicle entering an intersection, causing a marginal saturation flow rate that increases progressively with each entering vehicle. Large variations in saturation flow measurements also have cast doubt on the presumption that the base saturation flow rate of 1,900 passenger cars per hour of green time per lane (specified in the Highway Capacity Manual) is either stable or a constant value. The human factors that underlie the rate at which successive queued vehicles enter an intersection at a traffic signal are considered. A model based on driver behavior is formulated in a manner that can be calibrated without difficult field measurements of speeds or accelerations at precise points. Along with average queue lengths and average flow speeds, only average discharge headway times by queue position are needed. The model results suggest that the saturation flow rate probably is related to two key variables ordinarily ignored: downstream cruising speed and green-phase length. If the downstream cruising speed is slow and the green phase is long, then the discharge rate might become uniform and the base saturation flow rate might be adequately represented by a constant. For other situations, results indicate that flow rates progressively increase, as found in field studies.