The overall objective of this study was to characterize the reservoir mechanisms that cause premature oxygen breakthrough, and develop practical tools for controlling it. We did not expect to develop a universal solution, or a solution for each specific mechanism. The focus was on those conditions most likely to be encountered in moderate-depth to deep reservoirs; candidates included gas override, high-permeability steaks, and water underlie. Mitigation measures were selected based on due consideration of the cost of implementing them and their potential application to a wide range of specific problems. An overriding consideration was practicality for field use. The total program was to extend over a 3-year period. It was divided into three phases to permit assessments of completed efforts before proceeding with dependent activities, and to simplify the allocation of funds. ? Phase I consisted primarily of characterizing the problem and selecting candidate mitigating measures suitable for simulation testing and modeling. Some of the design fabrication efforts that could be performed in parallel were also to be included. ? Phase II was to consist of completing the simulator test installation, selecting and modifying the computer model, and then performing the testing and modeling to simulate breakthrough and mitigation in moderate-depth (low-pressure) reservoirs. ? Phase III was to consist primarily of the testing and modeling associated with simulating breakthrough and mitigation in deep (high-pressure) reservoirs.