A predominantly sandy levee embankment along an 18 km portion of the Sacramento River had a history of heavy seepage during river flood stages. A seepage cutoff wall keyed into clayey layers in the foundation soils was chosen by the U.S. Army Corps of Engineers to remedy the seepage problems. A test cutoff program utilizing 4 different methods (backhoe, trencher, mix-in-place and vibrated beam) of installing a cement-bentonite cutoff wall was undertaken as a performance type of contract to evaluate the methods and results of installing cutoffs with specified minimum width, depth, strength, and maximum hydraulic conductivity. This was followed by a Phase I contract to install 3.2 km of cutoff wall (27,870 m2) with essentially the same performance type specification as used in the test program. The quality control programs involved sampling and molding specimens of cutoff wall materials and sampling of the hardened materials with drilled borings, in situ packer permeability testing, and numerous sampling methods. Test results raised questions of the representativeness and degrees of disturbance caused by the various sampling and testing methods and they indicated the cutoff walls did not always meet the required minimum unconfined compressive strength of 103 kPa or the maximum hydraulic conductivity of 1 × 10-6 cm/sec. The paper presents the test results and discusses the quality control methods, procedures and problems encountered in both the test program and Phase I contract.