AbstractA large subsurface pool of waste solvent product, consisting primarily of 1,1,1‐tichloroethanc and carbon tetrachloride, was encountered during investigations at an industrial site in northern New Jersey. In the 1950s the product was discharged through a settling chamber directly below the shallow water table. Eventually the product accumulated within elongated depressions of erosional surface of varved clays at depths 10 to 15 feet below grade. The host sediment, line to medium sand, was overlain by line sand and silt. The delineated area of pooled DNAPLs covered 2750 feet2, and the maximum pool thickness exceeded 3 feel.The primary recovery involved pumping product from nine wells. Each recovery well was equipped with a sump extending into the clay, which enabled the system to keep the product pumping level below the bottom of the pool. A total of 3495 gallons of solvent product was recovered over two years. Nearly half of this volume was produced by two wells placed at the lowest points of the pools. Postpumping sampling of the former pools indicated that 43 to 94 percent of the pooled solvent mass was removed during the primary recovery. Average initial product salutation within the pool was estimated at 53.2 percent of the total porosity measured at 31 percent. Average residual saturation after the primary product recovery was 3.7 percent of the total pore volume.To test the feasibility of residual product recovery, an experimental secondary recovery was undertaken. Using sheet piling, a 506 feet2lest cell was constructed inside the former DNAPL pool. The cell featured a central recovery well, six peripheral wells, and monitoring probes. I he selected sequence of secondary operations included partial dewatering. hot water injection, final dewatering, and thermally enhanced vapor extraction (TVE). During six weeks of the secondary recovery operations. 87.9 gallons of product were removed, of which 72 percent was from TVE. 25 percent from hydraulic mobilization effects. and 3 percent from dissolution of residuals. Confirmatory soil sampling showed an average reduction of residual contamination by 93.4 percent in comparison to the concentration of residuals prior to the secondary recovery. For the lest cell, a combined total solvent recovery of 99.6 percent was achieved. This high recovery exceeded DNAPL recoveries expected or achieved in other field‐scale attem