Fuel consumption and dynamical variable data have been collected as a car was driven normally with low speed urban traffic. It has been shown, using regression analysis, that the single traffic variable of the 16 studied that can best explain fuel consumption per unit distance is the average trip time per unit distance. The regression results are similar for each of four different methods of sampling the speed-time history of the vehicle in traffic. Such fuel consumption data may therefore be collected by the method which is operationally most convenient. The variable, average trip time per unit distance, explained more than 71% of the variance in fuel consumed per unit distance in these experiments. An interpretation of this result is given in terms of engine-vehicle dynamics and the characteristics of urban traffic. The analysis of the data for each of four different drivers showed that the fuel consumed per unit distance was determined mainly by the average trip time per unit distance, essentially independent of the individual driver.