The definition of the boundary of a groundwater contaminant plume can be best viewed as a search strategy objective. To this end, a new approach to plume boundary identification based upon the goal of a stepwise reduction in the overall uncertainty in the location and geometry of the boundary can be used. Through the sequential selection of investigatory well locations based on the criteria of maximizing plume boundary uncertainty reduction and on updated field information, a least cost plume boundary investigation can be realized. The effectiveness of this approach can be investigated using a realistic hypothetical example and demonstrated via a field application.