An examination of data from two continuous stage and discharge streamgages and one continuous stage-only gage on the Middle Mississippi River was made to determine stage-discharge relation changes through time and to investigate cause-and-effect mechanisms through evaluation of hydraulic geometry, channel elevation and water-surface elevation data. Data from discrete, direct measurements at the streamgages at St. Louis, Missouri, and Chester, Illinois, during the period of operation by the U.S. Geological Survey from 1933 to 2008 were examined for changes with time. Daily stage values from the streamgages at St. Louis (1861-2008) and Chester (1891-2008) and the stage-only gage at Cape Girardeau, Missouri (1896-2008), throughout the historic period of record also were examined for changes with time. Stage and discharge from measurements and stage-discharge relations at the streamgages at St. Louis and Chester indicate that stage for a given discharge has changed with time at both locations. An apparent increase in stage for a given discharge at increased flows (greater than flood stage) likely is caused by the raising of levees on the flood plains, and a decrease in stage for a given discharge at low flows (less than one-half flood stage) likely is caused by a combination of dikes in the channel that deepen the channel thalweg at the end of the dikes, and reduced sediment flux into the Middle Mississippi River. Since the 1960s at St. Louis, Missouri, the stage-discharge relations indicated no change or a decrease in stage for a given discharge for all discharges, whereas at Chester, Illinois, the stage-discharge relations indicate increasing stage for a given discharge above bankfull because of sediment infilling of the overflow channel. Top width and average velocity from measurements at a given discharge for the streamgage at St. Louis, Missouri, were relatively constant through time, with the only substantial change in top width resulting from the change in measurement location from the Municipal/MacArthur Bridge to the Poplar Street Bridge in 1968. The average bed elevation appeared to be lowering with time at both measurement locations at St. Louis. Flow in the Horse Island Chute overflow channel for the streamgage at Chester, Illinois had an effect on top width and average velocity from measurements, and this effect changed with time as the inflow channel to Horse Island Chute filled with sediment. Top width from measurements at a given discharge was consistent through time at the Chester streamgage when adjusted to remove the part of the flow through Horse Island Chute. Average velocity from measurements at a given discharge appears to be increasing with time, possibly as a result of a series of dikes built or extended in the channel immediately upstream from the Chester streamgage; however, the average bed elevation for all discharges less than bankfull at the Chester streamgage fluctuate around an average value from 1948 to 2000, and the fluctuations appear to be related to the occurrence of moderate and large floods. Daily stage and discharge values available for the streamgage at St. Louis, Missouri, from 1861 to 1932 display distinct, fixed relations that change slightly with time before operation by the U.S. Geological Survey, indicating daily discharge was obtained from the daily stage value during this timeframe. A sudden and substantial reduction of about 24 percent at the upper end of the ratings for discharge at a given stage occurred between 1932 and 1933 when the U.S. Geological Survey began operating the streamgage. This change likely is the result of the change to Price AA current meters from other, less-accurate methods used for discharge measurements before 1933. Based on modeling results for the Middle Mississippi River by the U.S. Army Corps of Engineers and the findings of this study, the accuracy of the historic record before 1933 is questionable, and needs to be examined further. The differ