Discharges of canals and branches at their heads or at the boundaries of divisions, are observed by the Subdivisional Officer about once a month. Discharges of distributaries are observed about once a month, usually by zilladars. They are also to some extent observed by the Subdivisional Officer, but much is left to his discretion. Delta is worked out for each distributary month by month, and also, of course, for each crop. Thus a general duty “at distributary heads” can be obtained, and may be used in new projects[26] instead of the duty at the canal head, allowance being made for the water lost by absorption in the canal and branches. [26] See CHAP. IV., Art. 2. It cannot be said that these important figures are obtained as carefully as they could be. If the Subdivisional Officer personally observed the discharge at each distributary head, even every other month, the reliability of the results would be much increased. In addition to this the discharges of canals and branches at the boundaries of subdivisions should be observed and the results compared with the distributary discharges, so as to show the loss by absorption. At first grave discrepancies among the results would be found, but they would be reduced as the causes of error became known. For the method of investigating the causes of discrepant discharges see _River and Canal Engineering_, CHAP. III., Art. 5. A specimen of a Subdivisional Officer’s gauge register is given in table I. The zilladar keeps a similar register. The columns headed G contain the gauge readings, those headed D the discharges. Until some years ago there were no columns for discharges. The daily discharges of the canal and of the branches at their heads--and at intermediate points if they were at the boundaries of divisions--were entered in the Executive Engineer’s office and the duty was worked out at the end of each crop. The zilladar merely indented for a certain gauge reading at the distributary head, and the Subdivisional Officer could tell pretty nearly what gauge reading he required in the canal at the beginning of his subdivision. Since the year 1900 or thereabouts, the zilladars have been required to learn a good deal about discharges. They have to know how to observe the discharge of a distributary, and to learn how the discharge of an outlet varies with the head or difference between the upstream and downstream water levels. They are supposed to indent for certain discharges, and not merely for certain gauge readings. All this knowledge is useful to the zilladars and tends to increase their efficiency, but a practice of constantly thinking in discharges instead of in gauge readings is unnecessary. If the channels were of all sorts of sizes matters would be different. Actually the size of a channel is apportioned to its work, and the proportion of its full supply which it is carrying at any moment is easily grasped by means of gauge readings alone. TABLE I--GAUGE AND DISCHARGE REGISTER.