January 7, 1904] 
NATURE 
227 
years been engaged in ascertaining the value of the ee 
rivers as water suppliers and in furnishing information 
upon which to base estimates of the available supply. | 
The want of this information has frequently led to 
the most disastrous mistakes in the construction of | 
hydraulic works. From ignorance of the hydro- | 
graphic condition of the drainage basin of the stream, | 
and of the region in which the stream is located, | 
engineers have | in many cases been misled by the only 
information available, that of the ‘‘ oldest inhabitant, ’ 
which may be trustworthy as to the highest level to 
which the water has reached in floods, but is frequently 
very misleading as to the low water conditions of the 
river. Amongst other instances recorded is one 
where, after an expenditure of 32,o00/. in hydraulic 
works by a town where it was expected the water | 
from a neighbouring river would be capable of develop- | 
ing 14,009 horse power and cause it to become a manu- | 
facturing centre, it was found that the estimate had | 
been based upon a miscalculation as to what the river | 
could yield to the extent of 500 per cent. 
Even where statistics as to the rainfall are avail- 
able, these may be very misleading so far as the yield 
of the drainage area is concerned. unless checked by 
stream measurements. An instance is quoted where 
a calculation of the minimum yield of a river in one 
of the States was made independently by five ex- 
perienced engineers, the quantity varying from 0.20 
to 0-40 cubic foot per second per square mile. 
The various forms of floats used to determine the 
velocity of streams are discussed in the report. For 
reconnaissance works surface floats have been found 
most convenient and approximately trustworthy. Rod 
floats consisting of cylindrical tubes or wooden rods 
2 to 3 inches in diameter, weighted at the bottom, are 
considered as more trustworthy than double floats 
having the subsurface float connected to the upper one 
by a siik cord. 
The current meter, on the whole, has been found to 
be best adapted to the general measurements made in 
the United States Survey. One method of using this 
Fic. 1.—Current Meter in Use, Suspended from a Bridge. 
in a wide channel is shown in the illustration (Fig. 1). 
In the second illustration (Fig. 2) the measurements 
are shown as being taken from a cable having 200 
feet span placed across the stream, and supported on 
the right bank by timber shears 25 feet high, and on 
the left side anchored to a large buried oak. 
NO. 1784, VOL. 69] 
| 0.13 and 0.7 
As the result of a great number of observations it 
has been found by the United States surveyors that 
the mean velocity of a stream is generally found to 
be at six-tenths of the depth of the water measured 
from the surface for wide shallow rivers, which figure 
should be increased to two-thirds in the case of canals 
and flumes or narrow natural channels. The velocity 
generally increases from the surface downward to 
5. 2,—Cable and Car used to Measure Discharge of River. 
about one-tenth of the depth, and then decreases to 
the bottom, where it reaches the minimum. 
Where more than one observation was made upon 
the channel, the ratio between the surface and mean 
velocities in a stream was found, on the average of a 
number of experiments in different rivers, to be 0.88 
of the mean of the surface velocities taken in the 
vertical in which the floats were run. Where only one 
surface float was used in the centre of the river, the 
coefficient was on an average found to be 0.80. The 
chances of error are greater where only one float is 
used. For shallow depths of from 3 to 8 feet the co- 
efficient for the mean velocity varied from 0.92 to 0.82. 
For large deep rivers, such the Mississippi, 
Humphrey’s and Abbot’s observations gave a co- 
efficient of 0.98. 
Measurements are also recorded of the flow of water 
under ice. The observations were made by cutting 
holes large enough to admit a current meter. In an 
ice-covered channel a decided drag occurs at the 
surface as well as at the bottom. Two points of mean 
velocity were found to exist in the vertical at about 
3 of the depth, the maximum being at 0-35 
of the depth. The best result was obtained by hold- 
as 
| ing the current meter at two-thirds of the depth and 
applying a coefficient of 0-95 to the observed velocity 
at that point. 
For providing uniformity in the reports of the various 
_ observers as to the quality of the water, the following 
standard of turbidity was used for field observations. 
The figure 100 was taken to represent a water con- 
taining too parts of silica per million, in such a state 
of fineness that a bright platinum wire one millimetre 
in diameter can just ‘be visible when the centre of the 
wire is 100 millimetres below the surface of the water, 
the eye of the observer being 1.20 metres above the 
| wire, the observation being made in the middle of the 
day, in the open air, but not in sunlight. 
For taking observations a graduated rod with a 
platinum wire projecting from it at right angles was 
