178 Transactions of the Academy of Science of St. Louts 
by the barometer were unsatisfactory or were lacking, elevations 
were estimated as accurately as possible from topographic maps. 
Intermediate flat surfaces caused by the erosion of an alluvial 
terrace and which represent only a very temporary stage in the 
evolution of the valley were excluded in all cases where they 
could be recognized. Every effort was made to eliminate sur- 
faces which had been built up by eolian deposition subsequent 
to the aqueous deposit. 
The study reveals that alluvial terraces are far more com- 
mon than rock terraces, although the latter are not entirely 
lacking. Where the rock terraces were found, they were reg- 
ularly covered with a thin veneer of soil of an alluvial rather 
than a residual nature. 
In the case of several tributaries, the terraces were studied 
for a number of miles along the tributary to determine whether 
or not the terrace had a slope parallel to the profile of the stream. 
Some stream terraces have slopes essentially parallel to the pro- 
file of the stream. In other cases, the top of the terrace appears 
to be essentially horizontal; while, in still other cases, the slope 
of the terrace is actually contrary to the gradient of the stream. 
A map showing the locations of streams in the Middle Mississippi 
River Region which have been studied is shown in Figure 1. 
This map includes other geographic localities referred to in 
this report. 
A correlation of the terraces from stream to stream has also 
been made. For this purpose, the elevations of the terraces near 
the mouth of each tributary, together with the elevation of such 
terraces as may be found in the trough of the master stream, 
have been tabulated. ‘This data is shown in Table 1. The data 
have been plotted, together with the longitudinal profile of the 
mean datum surface of the master streams. To determine dis- 
tances along the master stream a broken line was drawn essen- 
tially along the axis of the trough and the distances were taken 
as offsets from this line (Fig. 3). Distances are, therefore, 
slightly shorter than they would be if measured on the axis of 
the river itself; but since the river channel changes from time 
