. 
Report of Messrs. Humphreys and Abbot. 29 
medium-stage curves, deduced from 52 observations at Columbus 
and 20 at Vicksburg, the mean differences were 0056 and ‘0155 
respectively; the maximum velocities in the curve being 4°1958 
and 45709. In bayou Plaquemine the mean difference was ‘036 
a a maximum velocity of 6491; and in bayou La Fourche 
"003 on a maximum velocity of 3-250. In these two cases, the 
curves are deduced from the results of a single day’s observation. 
he same equation was also further tested by being applied 
to the original curves, which had been combined, as above stated, 
on the principle of ee eo The correctness of that 
principle of combination ha t been quite certain; but the 
te of this final test left no » ante of its legitimacy. The 
an discrepancy between the computed and observed veloci- 
ties amounted in only one case to so much as one per cent of 
the maximum velocity in the curve; and _in this oe the 
absolute ee hee Re a little over an inch. Usually the 
agreement was much n 
“But a st till more saieurteathe test of 0 accuracy of the law of 
parabolic velocities is furnished by a comparison of its results 
in the okt’ curve varied, in the first instance, from w two 
3 feet to nearly three; and in the second, from about one aie a 
to over two Set: The observations are recorded in the one 
case at fifteen different depths, and in the other at thirteen. The 
‘ a discrepancy, between these observations and the results of 
mputation for the same points from the parabolic equation 
founded on them, amounted, in decimals of a foot, to only 0330 
and 
0243 for the two cases "respectively. _The largest of these 
neh. 
_ mean errors is less than four-tenths of an i 
It will be noticed that in all the computations made for the 
_ subsurface velocity of the Mississippi, in its different stages, and 
for the aes one and the same equation was constantly em- 
BE pioyed 
=%a a (01 8560) ¥a,25 
stream, but as this will not be very far in error Kf taken at oak 
tenths of the velocity on the surface, it may be assumed to be 
