1921.] Mead.—Impounding-dams for Multiple Functions. 169 
stream-flow; this is the capacity which is usually considered desireable in 
city water-supply schemes supplied from upland watersheds. A capacity 
of 6 in. or 7 in. will balance the flow of the driest year, giving a yield 
rather over half the mean stream-flow in the climatic conditions of the 
Auckland Province. With the still smaller capacities which would pro¬ 
bably be obtainable in practice, natural storage begins to play a considerable 
part, and no general results can be stated. It is instructive to note the 
considerable improvement on the minimum flow of the stream shown by 
quite small reservoirs of 1 in. or 2 in. capacity. 
The yield at zero capacity is, of course, the minimum dry-weather flow 
of the stream ; as it depends entirely on natural, storage, it will differ pro¬ 
foundly for different classes of country. Natural storage is provided by the 
porosity of the soil and substratum, forest, swamps, crops and cultivated 
areas, and occasionally lakes. The pumice area of the central portion of 
the North Island is an instance of splendid natural storage, and consequent 
uniformitv of stream-flow : while the clav hills of North Auckland are 
examples of the opposite extreme, where rains descend in short heavy floods, 
and a few weeks of drought reduce the streams to mere trickles. 
The relation between capacity and yield is easily calculated from a table 
of stream-gaugings extending over a sufficient period of time. A method 
of calculation is indicated in Appendix I. The accuracy of the result 
depends on the completeness of the information at disposal. The ideal 
would be a complete series of stream-gaugings extending over a sufficiently 
long period of years to include all possible climatic variations. If the 
period were twenty years the result should be correct within 4 per cent. 
In practice the calculation will usually have to be made from infor¬ 
mation which can be gathered in a year or two. 
Let a stream-gauge be established at the site of the proposed dam and 
read continuously by means of an automatic recorder, or else sufficiently 
often to give a true record of the stream-flow—say, daily in normal con¬ 
ditions and more frequently in floods. A complete year’s record, from one 
wet season over the dry period to the next wet season, will permit the 
determination of the yield of any reservoir up to 6 in. of capacity. This 
yield will apply to that particular year only, and must be reduced to obtain 
the probable yield of a year of minimum rainfall. If there is a rain-gauge 
station in the vicinity, or situated where the climatic conditions are reason¬ 
ably similar to those of the watershed under test, providing rainfall records 
covering twenty or thirty years, let the rainfall (R inches) during the 
twelve months of the test and the minimum recorded fall (r inches) for any 
consecutive twelve months be ascertained. Then the true perennial yield (y) 
of any reservoir may be deduced from the calculated yield (Y) by some 
such relation as 
y __ rj-_c 
Y ~~ ll~- c 
where c is a constant representing loss by evaporation, &c., and is usually 
between 14 in. and 20 in. If the watershed is compact and does not con¬ 
tain large variations of topography a tolerably accurate mean value of the 
rainfall on it may be obtained from several judiciously placed rain-gauges, 
when an alternative value for the constant c may be obtained directly 
from a comparison of the stream and rain gaugings. In any case, a con¬ 
siderable margin of possible error must be allowed in deductions based on 
the observations of a single year, unless it fortunately happened to be an 
exceptionally dry one. 
In some instances the information immediately available may com¬ 
prise the area of catchment, the mean annual rainfall, and a low summer 
