The vehicle time headways of arriving traffic flows at signalised intersections are examined in this study. Low traffic conditions are considered, and traffic lanes that carry less than 500 vehicles/h are studied. Data are obtained from 10 signalised intersections in Turkey. The lanes are classified with respect to movements of traffic flows. Seven distribution models are taken into account. The goodness of fits of the observations with the theoretical distribution models are evaluated using chi square (chi(2)) and the Kolmogorov-Smirnov tests. The results are considered with respect to both lane type and traffic volumes. The presented results show that log normal, Weibull, gamma or Pearson V distribution models can be used instead of the exponential distribution in modelling vehicle time headways.