From 1954 to 2004, water levels declined in the nontidal reach of the Apalachicola River, Florida, as a result of long-term changes in stage-discharge relations. Channel widening and deepening, which occurred throughout much of the river, apparently caused the declines. The period of most rapid channel enlargement began in 1954 and occurred primarily as a gradual erosional process over two to three decades, probably in response to the combined effect of a dam located at the head of the study reach (106 miles upstream from the mouth of the river), river straightening, dredging, and other activities along the river. Widespread recovery has not occurred, but channel conditions in the last decade (1995-2004) have been relatively stable. Future channel changes, if they occur, are expected to be minor. The magnitude and extent of water-level decline attributable to channel changes was determined by comparing pre-dam stage (prior to 1954) and recent stage (1995-2004) in relation to discharge. Long-term stage data for the pre-dam period and recent period from five streamflow gaging stations were related to discharge data from a single gage just downstream from the dam, by using a procedure involving streamflow lag times. The resulting pre-dam and recent stage-discharge relations at the gaging stations were used in combination with low-flow water-surface profile data from the U.S. Army Corps of Engineers to estimate magnitude of water-level decline at closely spaced locations (every 0.1 mile) along the river. The largest water-level declines occurred at the lowest discharges and varied with location along the river. The largest water-level decline, 4.8 feet, which occurred when sediments were scoured from the streambed just downstream from the dam, has been generally known and described previously. This large decline progressively decreased downstream to a magnitude of 1 foot about 40 river miles downstream from the dam, which is the location that probably marks the downstream limit of the influence of the dam on bed scour. Downstream from that location, previously unreported water-level declines progressively increased to 3 feet at a location 68 miles downstream from the dam, probably as a result of various channel modifications conducted in that part of the river. Water-level declines in the river have substantially changed long-term hydrologic conditions in more than 200 miles of off-channel floodplain sloughs, streams, and lakes and in most of the 82,200 acres of floodplain forests in the nontidal reach of the Apalachicola River. Decreases in duration of floodplain inundation at low discharges were large in the upstream-most 10 miles of the river (20-45 percent) and throughout most of the remaining 75 miles of the nontidal reach (10-25 percent). As a consequence of this decreased inundation, the quantity and quality of floodplain habitats for fish, mussels, and other aquatic organisms have declined, and wetland forests of the floodplain are changing in response to drier conditions. Water-level decline caused by channel change is probably the most serious anthropogenic impact that has occurred so far in the Apalachicola River and floodplain. This decline has been exacerbated by long-term reductions in spring and summer flow, especially during drought periods. Although no trends in total annual flow volumes were detected, long-term decreases in discharge for April, May, July, and August were apparent, and water-level declines during drought conditions resulting from decreased discharge in those 4 months were similar in magnitude to the water-level declines caused by channel changes. The observed changes in seasonal discharge are probably caused by a combination of natural climatic changes and anthropogenic activities in the Apalachicola-Chattahoochee-Flint River Basin. Continued research is needed for geomorphic studies to assist in the design of future floodplain restoration efforts and for hydrologic studies to monitor change