Ephemeral gullies are major sources of sediment yet the dynamics of their development are not well understood and are poorly quantified. Subsurface lateral flow over water-restricting horizons can initiate ephemeral gully development as a result of seepage erosion and preferential flow through soil-pipes. Quantification of the hydrologic properties under which subsurface flow contributes to gully initiation and head-cut migration are needed in order to improve our prediction and control capabilities. The objective of this research was to determine the effect of subsurface flow on ephemeral gully formation during rainfall events and specifically to quantify the effects of slope and pressure head on ephemeral gully erosion. Simulated rainfall was applied at 65 mm/h for 1 hour to a 1.5 m long by 1 m wide soil bed under different slopes. The soil profile consisted of a 30 cm depth of Providence silt loam packed to a bulk density of 1.35 g cm-3 over a 5 cm thick water restricting layer packed to 1.57 g cm-3. Preferential flow through an artificial soil-pipe was simulated using a 2 cm diameter porous pipe that extended 50 cm into the soil bed from the upper end. Pipe flow was controlled under a constant head of 0 (no pipe flow), 15, and 30 cm. The lower end of the soil bed was an open-face such that gully development was not hindered by an endplate. Tensiometers with pressure transducers were inserted into the soil bed at 12 positions to monitor soil water pressures during flow events. Experiments were performed with and without rainfall, and with and without pipe flow. Rainfall and pipe flow individually did not result in mass wasting, however, their combination did produce pop-out failures. The total soil losses by sheet erosion were 2-3 times higher with rainfall and pipe flow combined than by rainfall alone. The total soil loss by ephemeral gully erosion was 5 times higher than sheet erosion as a result of pipe flow combined with rainfall.