Depositional models of braided rivers are necessary for rational interpretation of ancient deposits, and to aid in the characterization of subsurface deposits (e.g. aquifers, hydrocarbon reservoirs). A comprehensive depositional model should represent bed geometry, flow and sedimentary processes, and deposits accurately, quantitatively, and in detail. Existing depositional models of braided rivers do not meet these requirements, and there are still many misconceptions about braided rivers and their deposits that need to be expunged. Over the past decade, there have been major advances in our understanding of braided rivers, making it possible to develop new and improved depositional models. First, the use of groundpenetrating radar in combination with cores and trenches has allowed detailed description of the different scales of deposit in braided rivers that vary widely in channel size and sediment size. Second, the study of braided channel geometry and kinematics has been facilitated by the use of aerial photographs taken at short time intervals. In some cases, these photographs have been analysed using digital photogrammetry to produce digital elevation models. Third, water flow and sediment transport at the all-important flood stages have been studied using new equipment and methods (e.g. acoustic Doppler current profilers, positioning using differential global positioning systems). However, such high-flow studies are rare, which is one reason why there has not been much progress in development of realistic theoretical models for the interaction between bed topography, water flow, sediment transport, erosion and deposition. Laboratory experimental studies of braided rivers have continued to be useful for examining the controls on channel geometry and dynamics, but have not been able to generate all of the different types and scales of strata observed in natural braided river deposits.New depositional models for sand-bed and gravel-bed braided rivers are presented here based mainly on studies of natural rivers. They comprise: 1 maps showing idealized active and abandoned channels, compound bars and lobate unit bars; 2 cross-sections showing large-scale inclined strata and their internal structures, associated with migration of compound bars, unit bars and their superimposed bedforms; 3 vertical logs of typical sedimentary sequences through different parts of compound bar deposits and channel fills.Compound bars migrate laterally and downstream, associated mainly with accretion of lobate unit bars. Abandoned channels are mainly filled with unit-bar deposits. The geometry of the different scales of strataset is related to the geometry and migration of the bedform associated with deposition of the strataset. In particular, the length-to-thickness ratio of stratasets is similar to the wavelength-to-height ratio of associated bedforms. Furthermore, the wavelength and height of bedforms such as dunes and bars are related to channel depth and width. Therefore, the thickness of a particular scale of strataset ...