Eighty-two periodontally involved patients were treated in a split mouth design such that one quadrant received coronal scaling (CS), root planing (RP), modified Widman surgery (MW), and flap with osseous resection surgery (FO). The therapy was performed in three phases: Phase I: the teeth previously designated to receive RP, MW, and FO were thoroughly root planed and the teeth designated to receive CS were scaled with no subgingival instrumentation, plaque control was initiated and reinforced for the entire mouth; Phase II: the designated teeth received MW or FO surgery; and Phase III: maintenance therapy every three months. The CS teeth received coronal scaling and polishing during maintenance appointments, while RP, MW, and FO teeth received supragingival instrumentation, subgingival instrumentation and polishing. Clinical measurements were taken initially, four weeks post-Phase I, 10 weeks post-Phase II, and after each of two years of maintenance care. All therapy modalities resulted in a decrease of mean probing depth with the FO producing the greatest decrease followed by MW, RP, and CS. The deeper the initial probing depth, the greater was the mean reduction of probing depth. FO created a loss of mean probing attachment in the 1 to 4 mm category. RP and MW produced the greatest gain of mean probing attachment in the 5 to 6 mm category. RP, MW, and FO produced similar gains in the greater than or equal to 7 mm category. FO created the most gingival recession followed by MW, RP, and CS.