Abstract Interceptor trenches are commonly used to recover light non-aqueous-phase liquids (LNAPL) from the subsurface and are particularly effective in the recovery of floating hydrocarbons. Interceptor trenches function by creating a zone of high permeability within the aquifer either as an open excavation (where the formation permits) or filled with a suitable backfill (typically gravel) that retains the high permeability and effective porosity required of a collection system. Numerous texts suggest that placement of an impermeable membrane is required on the down-gradient side of the trench to prevent captured LNAPL from exiting the trench on its downgradient side. Installation of such a liner can be difficult if not impossible under certain soil conditions. Simple calculations show that the membrane may not be necessary due to the hydraulic conditions created by installing the trench. Using these calculations in the preliminary design will result in an effective trench installation and correspondingly reduce installation and operation costs. LNAPL has been shown to migrate primarily near the top of the capillary fringe in porous media. Since the high permeability and effective porosity of the trench backfill do not support a capillary fringe of similar magnitude, LNAPL rests essentially on the water table in the trench. The zone of capillary saturation on the downgradient side inhibits LNAPL migration from the trench into the native soil. Provided that recovery of the LNAPL is maintained and the thickness of LNAPL in the trench is less than the capillary fringe height in the native soil, the LNAPL will remain hydraulically trapped. To demonstrate this condition, a review of some of the methods used to calculate the capillary fringe height was performed. To verify the magnitude of the calculated capillary fringe height and its effects on LNAPL retention within a trench, a simple laboratory-scale simulation was undertaken.