DISTRIBUTION AND VOLUME OF WATER REQUIRED Distribution.—Where irrigation is carried on throughout the whole year its proper distribution becomes a most important operation. It is generally considered sufficient if during any one season one-third of the area commanded is actually supplied with water. This encourages that rotation of crops which is a necessity of good farming, and also the lying fallow of some part of the soil, which also appears to be a necessity in hot and arid coun tries during a portion of each year. Some crops, of course, re quire water much oftener than others, and much depends on the temperature at the time of irrigation. During the winter months in Northern India magnificent wheat crops can be produced which have to be watered only twice or thrice. But to keep sugar cane, indigo, or cotton alive in summer before the monsoon sets in in India, or the Nile rises in Egypt, the fields should be watered every is to 20 days, while rice requires an almost constant supply of water.
tion for the purpose of proper division in each particular case is essential to ensure the maximum area being fed from any par ticular source.
Measurement.—Measurements of the volume of water flow ing in a river or canar were formerly obtained by noting the speed of objects floating with the stream. In the last 3o years current meters replaced floats and give more regular results, but it was not definitely known, just as in the case of floats, whether these results indicated actual volumes or not. This doubt has now been set at rest by the calibration of water passing through the sluices of the Aswan dam in Egypt soon after its construction in 1902. This calibration was itself definitely proved correct by the full scale measurements of the volume passing in a given time by one sluice into a tank built below the dam about 1905; thus the amount of opening of the sluices indicates the actual volume of water passing in the river, at any level, to within an error of %. It was subsequently found that current meter discharges, repeated a few times to obtain a mean result, and taken simul taneously with the measurements at the dam, gave very similar results to the latter; thus the accuracy of current meters also became proved. It has further been shown that the data obtained from carefully prepared small scale models can be relied upon as correct to within a very small percentage of error. This in dicates a very great advance on the previous state of such knowledge.