The critical height for DNAPL entry into a partially water saturated, dead-end fracture is derived and compared to laboratory observations. Experiments conducted in an analog, parallel-plate fracture demonstrate that DNAPL accumulates above the water until the height of the DNAPL overcomes the sum of the capillary forces at the DNAPL-air interface and at the DNAPL-water interface. These experiments also show that DNAPL preferentially enters the water at locations where DNAPL has previously entered, and the entry heights for these subsequent entries are lower than the heights measured for the initial invasion. The wetting contact angle at the DNAPL-water interface becomes larger at the locations where the DNAPL has already entered the water because of residual DNAPL on the fracture walls, which results in lowering the critical entry height at those locations. The experiments also demonstrate that a DNAPL lens can remain nearly immobile above the water for a period of time before eventually redistributing itself and entering the water.2