CHAPTER II.

THE DESIGNING OF A CANAL.

1. =Headworks.=--In the design of head works no very precise rules can be laid down. Some general ideas can however be given as to the chief points to be attended to and some general and approximate rules stated. In every case a large scale plan of the river is of course required and also a close examination of it and study of its character. An attempt to forecast its action is then possible. Gauge readings for several years and calculations of discharges are of course necessary. If the bed of the river, in course of time, rises upstream of the weir or scours downstream of it, a large amount of protection to the bed and banks will become necessary. Some description of headworks and weirs, with a plan of the headworks of the Sirhind Canal, India, has been given in _River and Canal Engineering_, CHAPTERS IV. and X. Remarks regarding the collection of information for such works are given in CHAPTER II. of the same work. It is also explained how, by keeping the gates of the under-sluices closed, a “pond” is formed between the divide wall and the canal head so that heavy sand deposits in the pond and does not enter the canal. By closing the canal and opening the under-sluices the deposit is scoured away.

The best site for the headworks of a canal depends on the stability and general character of the bed of the river but in deciding between any two proposed sites, the question of the additional cost of the canal, if the upper site is adopted, has to be taken into account. Such cost may, in rugged country, be considerable.

In the case of Indian perennial canals, the head is often close to the hills where the river bed is of boulders and shingle and fairly stable, but it is often at a distance from the hills and in such cases a gradual rise in the bed of the river, even in the absence of a weir, is more probable than scour. Such a rise may necessitate a raising of the crest of the weir and of the bed of the canal.

In the general arrangement of a headworks a great deal depends on local conditions. Sometimes the river runs in a fairly straight and defined channel and the weir can then be run straight across it. Sometimes, as in the case of the Ganges Canal, there is a succession of islands and various short weirs are required in the different channels. At the heads of the Eastern and Western Jumna Canals, the river, on issuing from the hills, widens out (Fig. 5.) and the weir is built obliquely and not in a straight line. Its crest is higher at the east than at the west side. There are under-sluices at both sides. The upstream end and west side of the island are revetted. The old head of the Western Jumna Canal, as shown in the figure, existed long before the advent of the British, and a temporary weir, made of gabions filled with stones, was constructed across the river every year during the low water period and swept away during the floods. To have carried the weir along the line shown dotted, the head of the Western Jumna Canal being of course brought up to it, would apparently have been feasible and cheaper, but the off-take would have been in shallow water because of the curve in the river, and there would have been no current along the face of the head regulator of the canal.

The level of the floor of the under-sluices is generally about the same as that of the bed of the canal. The sill--made to exclude shingle and sand as far as possible--of the canal head regulator may be 3 feet higher and the crest of the weir 6 to 9 feet higher. The top of the weir shutters is 1 to 2·5 feet above the F.S. level of the canal which may be 5 feet or more above the bed of the canal. If the weir is provided with falling shutters the width of the waterway of the under-sluices may be about ¹⁄₁₂th of the width of the waterway of the weir alone, otherwise about ¹⁄₈th.

In nearly all cases the weir has a flat top and flat slopes both upstream and downstream. In a case where the river bed is of sand, the depth of water on the crest of the weir in floods may be 15 feet and the velocity 14 or 15 feet per second. The downstream slope of the weir may be about 1 in 15, and the upstream slope 1 in 6. Where the river bed is of boulders the velocity may be still higher. The faces of the weir are usually of hammer-dressed stone. A lock for the passage of rafts is added if necessary.

Unless the banks of the river are high, it is necessary to construct embankments to prevent the river water, when headed up by the weir during the floods, from spilling over the country with possible damage to the canal. If the river has side channels they have to be closed. The stream may also have to be trained, by means of guide banks or spurs, so as to remain in one channel and flow past the canal head and not form shoals against it. Where the river is unstable, it may shift its course so as to strike the weir obliquely and this may cause excessive heading up at one side of the weir. In such cases it is usual to divide the weir into bays or sections, each about 500 feet long, by “divide walls” running at right angles to the weir.

The free-board or height of the masonry walls and tops of embankments above H.F. Level is about 5 feet.

The span of each opening in the under-sluices is generally 20 to 35 feet. The piers may be 5 feet thick. It is usual to make each alternate pier project upstream further than the others so that long logs coming down the river during floods, broadside on, may be swung round and not be caught and held against the piers.

Figs. 6 and 6A show the headworks of the Upper Chenab Canal now under construction (CHAPTER IV.) The site is in a low flat plain, but no better site could be found. The weir consists of 8 bays of 500 feet each. The crest is 10 feet above the river bed and the falling shutters 6 feet high. The slopes are 1 in 6 and 1 in 15. The bulk of the work is rubble masonry in lime. The lower layer upstream of the crest is of puddle; upstream of the second line of wells it is rubble masonry in half sand and half lime; upstream of the lower line of wells it is of dry stone and there is an intermediate layer of rubble masonry in lime with the stones laid flat. Below the crest there is a wall of masonry 9 feet thick and on the crest there are two strips of ashlar between which the shutters lie when down. The extreme upstream and downstream portions of the bed protection are of dry stone and 4 feet thick while next to the weir are concrete blocks 2 feet thick resting on dry stone. The width of the crest is 14 feet, of the weir 140 feet, of the protection 70 feet upstream and 110 feet downstream. The guide banks have tops 40 feet wide and 18 and 14 feet above the crest of the weir in the upstream and downstream lengths respectively, the side slopes being 2 to 1 and the water slope being covered, up to H.W. level, by dry stone pitching 4 feet thick. The left guide bank runs upstream for 3,250 feet from the centre line of the canal and the right 2000 feet from the line of crest shutters. The under-sluices have 8 bays of 35 feet each and the canal head regulator 36 openings of 6·5 feet each, the large openings shown in the figure being sub-divided. The crest of the weir is no less than 10 feet above the river bed and the shutters add 6 feet to this. The floor of the under-sluices is 4 feet higher than the river bed. There is thus ample allowance for a possible rise in the river bed.

2. =The Contour Map.=--The contour map, besides showing the contours of the country to be irrigated and of a strip of country, even if not to be irrigated, which will be traversed by the main line, should show all its main features, namely:--streams, drainages, railways, roads, embankments, reservoirs, towns, villages, habitations, and the boundaries of woods and cultivated lands. It should also show the highest water levels in all streams or existing canals. A map showing as many as possible of the above features should be obtained and lines of levels run for the contours. In doing this, the points where the lines of levels cut or pass near to any of the above features or boundary lines, should be noted. It may be necessary to correct inaccuracies in the plan or to supply defects in it. The greater the trouble taken to do this the less will be the trouble experienced later on.

The heights of the contour lines will, in very flat country, have eventually to be only 1 foot apart. This will necessitate running lines of levels half-a-mile apart at the most, and preferably 2000 feet apart, the pegs in each line being about 500 feet apart. In less flat country the heights of the contour lines can be further apart than 1 foot. Whatever distance apart is decided on for them, the survey should be done once for all. On one of the Indian canals in flat country, the lines of levels were at first taken 5 miles apart, the branches roughly aligned and then further surveys made. This led to great expense and delay and the procedure has not been repeated.

In making a contour survey, a base line, as centrally situated and as long as possible, should be laid down, with side lines parallel to it near the boundaries of the tract. The cross lines at half-mile or other intervals should then be laid down. Some of them may run out beyond the side lines. Circuits of levels should be run along the base line, the side lines and the two extreme cross lines and be carefully checked. The remaining cross lines should then be levelled. All the levels having been shown on the map the contours should be sketched in. The scale of the map for a large project may be two inches to a mile. If it is likely that the survey will have to be extended, it will be easier to do this by prolonging the base line and running more cross lines, than by prolonging each of the cross lines already surveyed. This can be borne in mind when selecting the base line.

3. =Alignments and Discharges.=--On the contour map the proposed alignments of the canal, branches, distributaries, and escapes, determined after careful consideration of all matters affecting them, are shown. The tracts to be irrigated by each branch and each distributary are now marked off, the “irrigation boundaries” following approximately the valleys and lines of drainage. Any large tracts of land which cannot be irrigated are of course shown and are excluded. Forests or other lands which are not to be irrigated should be similarly dealt with, otherwise confusion is likely to arise later. The commanded area dependent on each distributary is now ascertained from the map. A certain percentage being deducted for scattered unculturable areas the culturable commanded areas are obtained. The proportion to be irrigated (in India in the kharif) having previously been decided, the number of acres to be actually irrigated by each distributary is arrived at.

The next step is to ascertain the discharges.[9] A general duty for the whole canal having been estimated by considering the actual figures for other canals the full supply of the canal at its head is arrived at. (CHAPTER I, Art. 6). In Northern India it will be the kharif duty and kharif full supply. Since some water is lost by absorption in the channels, the duty of the water on a branch is higher than that of the whole canal based on its head discharge, and the duty on a distributary is higher still. In designing a canal, an attempt has to be made to estimate the losses of water in the main canal and branches, so that the duties of the branches and distributaries may be estimated and the channels designed accordingly. On the Western Jumna Canal the figures were estimated to be as follows:--

Kharif. Rabi. Average discharge at canal head (c. ft. per sec.) 3536 2755 Duty based on the discharge (acres) 98 138 Estimated loss of water in canal and branches (c. ft. per sec.) 400 300 Average discharge at distributary head (c. ft. per sec.) 3136 2455 Duty based on the discharge (acres) 111 154

[9] In this Article and in the rest of this Chapter it is assumed that the canal is a Northern Indian one. Any modifications necessary to suit canals in other countries will readily suggest themselves.

The question of duty is one which if not carefully considered, may cause some confusion. A canal and branches, having been designed with certain assumed duties and with discharges based on certain values of N in Kutter’s co-efficient, have, let it be supposed, been constructed to a greater or less extent. When the time comes for constructing the distributaries, the engineers concerned may have different ideas, based on later experience, as regards the probable duty and the most suitable value of N. If they design the distributaries with a higher duty and a lower value of N, it is obvious that they can provide more distributaries than at first designed, or can increase their lengths. In either case they would provide for an increased commanded area. If they do not do this, they ought to adhere to the values at first proposed, thus making the channels larger than, according to their ideas, would be necessary. These larger channels will be able to do more irrigation, by an increase, not in the commanded area, but in the proportion of it which is irrigated. Any other course would result in the canal carrying more water than could presumably be used by the distributaries. Again, the question how the assumed duty was arrived at may need consideration. It may have been arrived at by taking the duty figures of some existing canal, based on discharge figures which were the result, not of observed but of calculated discharges, and if the calculations were based on a value of N which experience has proved to be wrong, a correction is obviously needed. Many mistakes of the kinds indicated above have been made, not perhaps in the case of a project which has been recently got up and is then quickly carried out in its entirety, but in one which is carried out slowly or after a long period has elapsed or in one which consists of extensions of an existing system. So great, however, is the elasticity of a channel--by which is meant its capacity for adapting itself to varied discharges, a small change in the depth of water causing a great change in the discharge--and so considerable has been the uncertainty as to the real duty to be expected, that any mistakes made have not usually resulted in any serious trouble.

It has been stated (CHAPTER I, Art. 2) that it is not desirable to let one channel tail into another. In old canals a distributary used sometimes, after running parallel to a canal, to be brought back towards it and tail into it. The advantage of this was that the distributary had not to be made very small towards the tail and that, if the demand abruptly ceased, the distributary was not likely to breach. The principle was, however, essentially bad. The lower part of the distributary was obviously too near the canal and not centrally situated as regards the irrigated strip. The portion at the extreme tail was superfluous. Again, whatever volume of water was carried through the distributary and back into the canal, was needlessly detached instead of being kept in bulk. Moreover the duty of water on such a distributary cannot be ascertained without a tail gauge and the observation of discharges at the tail. There are similar objections to one distributary tailing into another. Each should be separate and distinct.

A major distributary is one whose discharge is more than 40 c. ft. per second. It may be as much as 250 c. ft. per second. A branch, as soon as it reaches a point where its discharge becomes only 250 c. ft. per second should be considered as a major distributary. A minor distributary is one whose discharge is from 8 to 40 c. ft. per second. A minor distributary is nearly always a branch of a major distributary. There are instances of “direct minors,” i.e., minors taking off from canals or branches. Such a minor, unless its discharge is a large fraction of that of the canal which supplies it--and this can seldom be the case--is objectionable because the petty native official who has to see to the regulation of supplies can manipulate the supply easily and without detection, and the number of persons irrigating from it being small, he can make private arrangements with them. On the Sidhnai Canal there are some half-dozen distributaries each of which had one or two minors which took off close to the head of the distributary. The people who irrigated from the minors managed to get the heads shifted and taken off direct from the canal, on the ground that, the water level in the canal being higher than in the distributary, there would be better command and less silt deposit. The irrigation on all these minors ran up to a figure far in excess of what had been intended, to the detriment of lands further down the canal. The minor heads have all been retransferred to the distributaries, the difficulty as to command being got over, as it should have been at first, by constructing weirs in the distributaries. The fall in the water surface at the distributary head, i.e., the difference between the water level in the canal and that in the distributary downstream of its head but upstream of the weir, is quite trifling or even inappreciable.

In some of the older Indian canals it was the custom to place the heads of distributaries, not just above a fall but several hundred feet above it, the idea being that the distributary then received less silt. This practice has now been discontinued. There is no valid reason for following it.

The question whether, when a channel crosses a road on the skew, a skew bridge should be constructed or curves introduced into the road or channel, is one which requires some consideration. As far as possible the lines of channels should be fixed so as to cross important[10] roads on the square or with a small angle of skew. In the case of main canals or branches, the introduction of special curves is generally out of the question, but if the road is not straight something can be done by shifting the line one way or the other. In the case of “major” distributaries, curves can to some extent be introduced. In the case of “minor” distributaries it is often possible to curve the channel, with a radius of say 500 feet, so that it will cross the road at right angles. There is very little objection to a skew bridge if the angle of skew is not great. The angle of crossing having been made as near to 90° as possible, the bridge can be made skew though not necessarily so much askew as the road. Slight curves can be introduced into the road. When the road is made askew, a bridge on the square involves at least three considerable curves (Fig. 7) and the taking up of extra land. It also causes, in perpetuity most likely, a more or less inconvenient and unsightly arrangement and one which, in most countries, would not be tolerated. When the angle of skew is not great, it is best to introduce no curve at all into the road. In the case of a “village” road, which may be more or less undefined and liable to be shifted, the difficulty about land may not be great, but even in this case the angle of crossing should, if possible, be kept near to 90°, especially in the case of minors, and where curves have to be introduced into the road they should be suitable ones. Abrupt angles are not only unsightly but are unfair to the cart drivers. The crossings of village roads by the minors of a certain great modern canal have been stigmatised as “hideous.” Indian canals can afford to do work properly.

[10] In India “district” and “provincial” roads.

4. =Remarks on Distributaries.=--Before a canal system can be properly designed, it is necessary to determine certain points in connection with the working of the distributaries. A distributary is intended to irrigate a certain kharif area. Its average kharif supply is determined from the assumed kharif duty. It generally runs full in the kharif but not always. In a very dry tract such as the Montgomery district of the Punjab, the demand is so great and so steady that a distributary practically runs full through the greater part of the kharif. In such a case the canal or branch must be so designed that it can keep all distributaries full at the same time. Its F.S. discharge will be the sum of all the F.S. discharges of the distributaries plus the losses of water by absorption.

But in other cases, especially if the rainfall is considerable, a distributary does not require its full supply, either all through the kharif or for long at a time. An estimate must then be made of what it will require. It may be estimated that its requirements will be met if, during the period of greatest demand, it is closed for two days out of a fortnight and receives full supply for the remaining twelve days. In this case, since the various distributaries need not all be closed on the same days, the canal or branch can be so designed that it will carry a full supply equal (after deducting losses) to ⁶⁄₇ths of the aggregate full supplies of the distributaries. In other cases the fraction may be ³⁄₄ths. It is likely to be lower the greater the rainfall of the district. Even in the case when the distributaries run full through nearly the whole of the kharif, there will be periods when they only run with about ³⁄₄ths full supply. If full supply were run at such times, many of the outlets would discharge more water than was required, the cultivators would partly close them, and breaches in the banks of the distributary might result. Thus the water level of a distributary must always be so arranged that it will have a good “command” when it is running with about three-fourths of the full supply discharge. The water level with ³⁄₄ths full supply is generally ·5 to ·75 feet below the full supply level but it should be calculated in each case. Generally it will be correct to make the water level, when ³⁄₄ full supply is run, about 1 foot above the high ground traversed by the distributary, excluding any exceptionally high portions of small area. A more exact method is given in Art. 9. The greater the proportion of the culturable area which is to be irrigated, the less should be the area of any high land which is excluded. The F.S. levels of the distributaries at their off-takes must be settled in accordance with the foregoing remarks, and these F.S. levels must be entered on the plan. Neglect to thus fix the F.S. levels of distributaries before designing the canals has frequently led to trouble.

The head needed at a bifurcation in order to get the supply into a branch or distributary is always small unless the velocity is high. For a velocity of 3 feet per second the head required is only about ·16 ft., for 2 ft. per second ·1 ft.

On an Inundation Canal which has no weir across the river, the mean supply downstream of the regulator (which is built a few miles down the canal lest it should be damaged by the shifting of the river) is, as has been mentioned, about half the full supply. The command in such canals is not generally very good. A distributary can often obtain only mean supply and it should be designed so as to command the country when it is carrying mean supply. A detailed description of Inundation Canals in Northern India, is given in _Punjab Rivers and Works_.

Let M, F, m, f, be the mean and full supply discharges at the heads of a canal and of an average distributary on it and let the number of distributaries be n. It has been seen (Chap. I. Art. 6.) that M = ·8F about. Let k be the proportion of the supply lost by absorption in canal and branches. Then n m = (1 - k) M = ·8 (1 - k) F. If the distributaries all run with full supplies--at the time of greatest demand--for 4 days out of 5, then,

nf = 1·25 (1 - k)F

f 1·25 - = ---- = 1·56 m ·8

Since k depends on the wetted area, it is not likely to be so great for F as for M, but the above gives a general idea of the ratio of the full kharif discharge to the mean kharif discharge. On a large canal the circumstances of the distributaries will not all be similar. Some will run full for a greater proportion of their time than others. They can be divided into groups and the ratio of the full to the mean supply calculated for each group. The mean supply is, as above stated, obtained from the area to be irrigated, and the duty as estimated at the distributary head.

At one time a system was introduced of making distributaries of large size with the idea of running them for short periods. One reason given for abandoning this arrangement, was that there was a tendency to run such a distributary for too long. This reason is not very intelligible. It would be applicable to any distributary which was not intended to be run without cessation. The result would be that some other distributary would be kept short of water and this would imply extremely bad management. The chief reason against such a distributary is the greater cost of its construction. It would effect a saving of water. The ratio of the discharge to the wetted area would be high, though this would be to some extent neutralized by the greater frequency of closures, since, when water is admitted to a dry channel, the absorption is at first great. There would also be some difficulty in the distribution of the water because of the short period for which it would remain open. It will be seen (Chapter III. Art. 5), that it is desirable to open and close always at the same hour of the day. An ordinary distributary might run for 11 days out of 14. One of double the size could not conveniently be run for 5¹⁄₂ days. A distributary can always be enlarged if necessary, but if made too large it is extremely difficult to make it smaller.

It was also, at one time, usual to make minors, when there were several on a distributary, of large capacities so that they ran in turns. The preceding remarks apply to this case. The system has been abandoned.

5. =Design of Canal and Branches.=--The apportioning of discharges to the various channels having been effected as described in Art. 2, the designing of the canal and branches is proceeded with. Rough longitudinal sections of all the lines are prepared by means of the contour map, the ground levels being shown at intervals of one foot--or whatever the vertical distance between the contours may be--and the horizontal distances obtained from the map by scaling.

On these longitudinal sections the lines proposed for the bed and F.S. levels are shown reach by reach and also the mean velocities and discharges.

The laws of silting and scouring and the principles on which channels should be designed are fully gone into in _River and Canal Engineering_. It is there explained that, for a channel of depth D, there is a certain critical velocity, V₀, which just prevents the deposit of the silt, consisting of heavy clay and fine sand, found in Indian rivers--this silt enters the canal in such immense quantities that the canal silt clearances would be impossible if much of it was deposited in the channels--that sand of grades heavier than [·1] may deposit in the head of a canal and well nigh threaten its existence, that the clear water entering the canal in winter may pick up and carry on some of the sand but that proper steps for preventing the deposit in the canal can be taken at the headworks. This last question has been referred to in Art. 1. The following additional rules for designing canals in Northern India are chiefly taken from those given by Kennedy in the explanatory notes to his Hydraulic Diagrams, which are in use in the Irrigation Branch in Northern India.

(1) Near the hills where the bed is of shingle the velocity may exceed V₀. A few other soils will stand 1·1 V₀.

(2) In ordinary channels any excess over V₀ will give much trouble lower down.

(3) In the first four or five miles of a distributary, V₀ should be allowed and gradually be reduced to ·85 V₀ at the tail, the gradient being reduced if convenient, while a minor or branch distributary should have less than V₀ at its off-take and still less at the tail. The sand is drawn off by the outlets and in the lower part of a distributary it is often non-existent.

(4) If there is efficient silt trapping at the head of the canal any figures arrived at by the preceding rules should be multiplied by ·9.

(5) In the case of a canal having its head far from the hills, the sand is finer and any figures arrived at as above may be multiplied by, perhaps, about ·75, but further experience is needed to decide this.

(6) If the soil is very poor, especially if the depth of water is more than 6 or 7 feet, the velocity should be less than V₀--say ·9 V₀--so as not to cause falling in of the banks. Depths of more than 9 or 9·5 feet should, as far as possible, be avoided for the same reason.

(7) At a bifurcation, one branch channel may have no raised sill, and, owing to its smaller depth, it may draw off no surface water and get an undue share of rolling sand. Its velocity should be greater than V₀ and that of the other branch be less than V₀.

(8) At such a bifurcation it may be necessary, during times of low supply, to head up the water in the main channel and some silt may temporarily be deposited in it. When the heading up ceases, the silt is scoured away but it mostly goes into the branch whose bed level is the lower. It is best to design such bifurcations so that the sill levels of the two branches are equal and, if possible, so that their bed levels are equal.[11] Otherwise the channel which is likely to get most silt should have the steeper gradient.

(9) Any existing well established régime should not be tampered with.

[11] Appendix A in _River and Canal Engineering_ deals with some instances of fallacies in questions concerning flow in open streams. An extract from it describing a remarkable divide wall recently constructed at the head of the Gagera branch, Lower Chenab Canal, is given in Appendix A of this book.

Experience shows that in designing Irrigation Channels in the plains of India in accordance with Kennedy’s figures, the maximum ratio of bed width to depth of water is as follows:--

Discharge, c. ft. per second 10 25 100 200 500 1,000 Ratio 3·5 4 4·5 5 6 6

The actual gradients of the canals generally range from about 1 in 8,000 for a main canal to 1 in 2,000 for the tail of a distributary, but near the head of a canal where the bed is of boulders and shingle, the gradient may be as steep as 1 in 1,000.[12] The velocity in this last case may be 5 feet per second but generally it is not more than 3 or 4 feet per second in canals and branches, and 1 to 2 feet per second in distributaries.

[12] On the Upper Jhelum Canal, 1 in 970.

In designing the channels, N, in Kutter’s co-efficient, may be taken as ·0225 or ·020, according to judgment. For new and smooth channels ·020 is generally correct. A channel generally becomes rougher by use but sometimes it becomes smoother. Cases have occurred in which N has been found to be ·016. This question is discussed in _Hydraulics_, Chap. VI.

The bed width of a canal is reduced, where a distributary takes off, in such a way that when the canal and distributary are both running full, the depth of water in the canal continues to be uniform and the flow to be uniform. When the distributary is closed there is heading up in the canal upstream of the off-take, but not enough to make any appreciable difference unless the capacity of the distributary is a large fraction of that of the canal and even then no harm is likely to result.

The preceding rules and principles being taken into consideration, the channels are designed. The bed levels, gradients and depths are so arranged as to give the velocities suited to the soil and to maintain the proper relation of depth to velocity. The bed width is arranged so as to give the proper discharge. The full supply level of the canal and branches has also to be so arranged that it shall be higher, at each distributary off-take, than the full supply level of the distributary. It is desirable to be able to give a distributary its full supply even when the canal is low. Generally the slope of the country along any line is greater than would be suitable for the bed, and “falls” are introduced. The off-take of a distributary is generally just above a fall and there is generally an ample margin between its F.S. level and that of the canal. The discharge of the canal during the greater part of the rabi may be only about half the full supply. This discharge should be estimated and the water level corresponding to it calculated and shown on the longitudinal section. If possible the levels should be so arranged that even with its least supply the water level in the canal will enable full supply to be given to a distributary. If this cannot otherwise be managed it may be necessary to construct a regulator in the canal below the head of the distributary so that, during low supplies, the water can be headed up. It has been stated in _River and Canal Engineering_, Chapter IV., that such heading up, if temporary, is not at all likely to cause silt deposit in the canal. The designing of the distributaries is not proceeded with at this stage.

Since no irrigation is usually done directly from the canal and branches, they are designed without any particular connection between the level of the water and that of the country traversed. Dangerously high embankments are of course avoided as far as possible. The bed is designed at such a level that the excavation and embankment at any place will be, as nearly as possible, equal. Land in India is cheap. When the excavation exceeds the embankment the balance is made into a spoil bank. When the excavation is less than the embankment the balance is got from borrow-pits.

The side slopes of channels in excavation are generally 1 to 1, in embankment 1¹⁄₂ to 1. The sides of channels of small or moderate size usually become about ¹⁄₂ to 1, or even vertical, by the deposit of silt on the slopes. This reduction of area is allowed for in the design i.e. the bed width is so designed that the channel will carry the required discharge, not with the side slopes as executed, but when they have become ¹⁄₂ to 1. In large canals however the sides do not always silt up but rather tend to fall in. When this is expected to occur the allowance above described is not made. Berms are left so that if any part of the sides fall in, the bank will not also fall in. The berms also allow of the channel being widened if that ever becomes necessary. Type sections are given in Figs. 8 and 9.

6. =Banks and Roads.=--Figs. 8 and 9 show the banks and spoil.

The scale is 6 feet to an inch. The depth of water, in this particular case, is 7 feet, and the bank, excluding the small raised bank, 2 feet above the water. The inside edge of the bank, where the small raised bank is shown, is kept parallel to the canal for a considerable distance. Its position is got by drawing a line, shown dotted, at, generally, 1¹⁄₂ to 1. The embanked part of the slope is actually made at 1¹⁄₂ to 1, but the excavation is at 1 to 1, so that a berm is left. The width of this berm of course varies as the depth of digging varies. If there is likely to be much falling in of the sides the berm can be made wider, the dotted line starting, not from the edge of the bed, but from a point further in. On an inundation canal in sandy soil the berm may be 20 feet wide. In figure 8, the inside slope above the berm is supposed to have silted up to a slope of 1 to 1. In cases where it is expected the whole inside slope will silt to ¹⁄₂ to 1 the dotted line, to give the edge of the bank, can be shifted towards the channel so that the berm at the ground level when the channel is excavated will be very small for the minimum depth of digging. There is no need for the inner edge of the bank to run parallel to the canal for great distances. Its position can be shifted whenever suitable and the width of the berm at ground level varied. This prevents the occupation of a needlessly great width of land. It used at one time to be not unusual to make a bank with a berm on the land side, similar to that formed by the spoil in Fig. 8, but at about the level of full supply in the canal. The principle is not a good one. Salient angles are liable to be worn away. If earth has to be added to a bank to strengthen it, the whole can be widened or the rear slope flattened. The roadway is shown 18 feet wide, which is nearly the maximum. For the drainage of rain water it has a transverse slope, away from the canal, of about 1 in 50. The small raised bank on the canal side is to give safety to wheeled vehicles. It is provided on the patrol bank[13] on main lines and places where there is much traffic or where there is plenty of width of bank to spare. When the ground level is, for a considerable distance, above the proper bank level--which is at a fixed height above the F.S. Level--so that the road and its side-drain have to be cut out, much earthwork can be saved by allowing them to be at a higher level and, in the case at least of the non-patrol road, giving the road a reduced width.

[13] A canal has an unmetalled driving road--called the “patrol road” or “inspection road” on one bank. This road is reserved for the use of officials. Otherwise, it would soon be cut up and worn away, and the cost of repairs would be excessive. The patrol road should be on that bank which is, in the morning (the time when inspections are usually made) in the shade of trees planted on the landward side. Trees are not usually planted near the water edge as they are sometimes blown down. In Northern India the canals generally flow in a southerly direction, so that the left bank is best for the patrol bank. On the other bank there is a bridle road which is open to the public. Near a rest house--unless there is a bridge actually at the place--the patrol road should be on the same bank as the rest house. It can if necessary cross at the first bridge. Frequently there is also on one or both sides of the canal a “boundary road,” which is open to the public, along the toe of the outer slope. Along a distributary there may be a boundary road on one side. It is generally the only road which can take wheeled traffic, and in this case it should be reserved for officials unless money is provided to keep it always in repair. Officials have to be on tour for weeks or months at a time, and in all weathers. Their baggage carts also have to precede and follow them. Anything which facilitates their touring about and seeing things for themselves is, in India, most desirable. At a watercourse crossing the boundary road along a distributary should be taken by a curved incline up on to the bank and down again. Thus not only is the cost of a culvert saved, but any touring official who is driving obtains a view of the channel which he cannot get from the boundary road.

In shallow digging, the plan of setting back the banks (Fig. 10) and letting silt deposit as shown by the dotted lines, is one which should be followed much oftener than it is. It not only gives eventually a very strong bank, but it enables the borrow pits, from which the earth for the banks is got, to be dug inside the banks. Outside borrow pits, besides being a source of expense, owing to compensation having to be paid to those in whose land they are dug, cause great areas of hollows which are not only unsightly, but are often full of stagnant water and are thus a fruitful source of mosquitoes and malaria. Insufficient attention has hitherto been paid to this matter.

In designing each reach of a canal or branch, type cross sections should be drawn out for several different depths of digging, _e.g._, one for very shallow digging, _i.e._, where the bed is little, if at all, above the ground level, one for deep digging where the ground is higher than the water level, and one for the “balancing depth,” where the area of the channel excavation is equal to the earth required for the banks. In calculating the earthwork the sectional area of the digging or of the embankment is taken, whichever is the greater.

The proper width and height of bank for any channel depends partly on the maximum depth of water in the channel, and partly on the discharge. Given a depth of water of say 8 feet, a breach will obviously be more disastrous with a great volume of water than with a small volume. The following statement gives some figures suitable to the rather light and friable soils of Northern India, but the question is largely one of judgment. Generally a low and rather wide bank is preferable to a higher and narrower one. If a road, with or without the small raised bank next the canal, is required, special widths can, of course, be arranged for. A 14-foot bank is required for a driving road.

------------+--------------+---------------+--------------- Top Width of|Height of Bank| Greatest | Greatest Bank. | above F. S. | Admissible | Admissible | | Discharge. |Depth of Water. ------------+--------------+---------------+--------------- Feet. | Feet. |C. ft. per sec.| Feet. 20 | 2 | 12,000 | 12 18 | 2 | 8,000 | 12 16 | 2 | 5,000 | 11 14 | 2 | 3,000 | 10 16 | 1·5 | 2,000 | 9 14 | 1·5 | 1,500 | 9 12 | 1·5 | 1,200 | 8 10 | 1·5 | 1,000 | 7 9 | 1·5 | 700 | 6 8 | 1·5 | 500 | 5·5 7 | 1·5 | 400 | 5 6 | 1 | 300 | 4·5 5 | 1 | 200 | 4 4 | 1 | 100 | 3·5 3 | 1 | 50 | 3 ------------+--------------+---------------+---------------

The spoil in Fig. 8 is shown at a different level from the bank proper, as it should be to give a neat straight edge to the bank. The width of the spoil may vary every chain. In Fig. 9 the spoil is raised to avoid taking up too much land. The spoil presents the best appearance when its height is kept uniform for as long a length as possible, the width varying according to necessity, When the height has to be altered, the change should be made by means of a short ramp. When the spoil is higher than the road, gaps in it are left at intervals so that rain water can pass away. When the spoil is heavy for a very short length it can, in order to avoid a short and unsightly heap, which would result from the adoption of the section shown in Fig. 9, be placed as in Fig. 8, some of it being led askew.

The small channel shown outside the bank in Fig. 8 is a watercourse for enabling trees to be grown. It has, of course, to be graded, and it may be in cutting or in embankment. If any silt clearances of the canal are likely to be necessary, the watercourse must be set back to allow room for the spoil. Such spoil, if sandy, is to a large extent washed down or blown away and does not accumulate to anything like the extent that would be expected.[14] Moreover the spoil can extend onto the watercourse when the trees have grown big, and no longer need watering. Outside the watercourse is shown the boundary road and the land boundary pillar. The small channel in Fig. 9 is a drain for rain water. It can be used as a plantation watercourse if the water is lifted.

[14] This fact has been quoted (_The Pioneer Mail_, “Silt,” 8th March, 1913) as showing that the silt supposed to be cleared is not really cleared. This may be the case to some extent, but shortage of spoil is little proof of it.

Where there is no spoil, some extra land, perhaps 20 feet on either bank, is usually taken up for getting earth from for repairs.

7. =Trial Lines.=--The proposed lines of channel, determined as explained in Art. 5 should next be laid down on the ground. A line should consist of a number of straight portions. The curves should not be put in. Trial pits should be dug at intervals. Some defects in the line may at once become apparent because the contour map, owing chiefly to the lines of levels having been taken a considerable distance apart, is not perfect. A line may pass through a patch of very high or very low ground or too near to some building or other object with which it is desirable not to interfere. Alteration may be desirable at a drainage crossing or at the off-take of a branch. The lines should be corrected where necessary. Sometimes the corrections may be very considerable. Allowance can be made for the alterations which will occur when the curves are laid out. Where there is doubt as to which line is the best, trial pits may be dug to obtain further information regarding the soil.

The line should now be levelled, careful checks being made, a longitudinal section of it prepared and the proposed bed, bank and F.S. level shown. The ground levels ascertained by levelling the line, are certain to disagree, to some extent, with the contour lines. The latter were got only by inference from the levels of points in the survey lines, and they should be corrected in accordance with the fresh levels now available. If the line does not seem to be the best that can be got, a fresh line can be marked on the plan and the above procedure repeated.

8. =Final Line and Estimate.=--As soon as the best line seems to have been found, a large scale plan of the country along its course should be made by taking bearings or off-sets from points in it to the various objects and noting where the line cuts them. On this plan will be shown the exact alignment, the curves being put in and the straight portions slightly shifted where necessary so that the line may pass at a proper distance from any buildings or other objects. But before this procedure is carried out, or while it is being carried out, the estimate for the work can be prepared from the longitudinal section already taken. Such a section is of course amply sufficient for a “project estimate,” in which only approximate figures are given, and it is quite near enough for any estimate. In the case of small works which have often to be executed with great promptitude, lamentable delays have occurred owing to the engineer deferring the preparation of his estimate till he had got the line exactly fixed. Moreover there is a chance of the labour being thrown away in case the sanctioning authority directs any change in the alignment to be made.

In the case of a large scheme, a project estimate is prepared. In this the earthwork and the area of land to be occupied are calculated pretty accurately. Designs and estimates are also prepared for the headworks and for the chief regulators. For works of which there are to be many of one type--bridges, falls, distributary heads and small drainage syphons--the cost is arrived at from lump sum figures, one drawing of each kind being submitted as a type. The distributaries are approximately estimated at mileage rates. In the case of a small scheme everything is estimated in detail except perhaps the distributaries or some of them.

9. =Design of a Distributary.=--A distributary is a canal in miniature and, like a canal, it may have branches. It has masonry bridges, falls and drainage syphons. It has, as already mentioned, a masonry regulator at its head. At the off-take of any branch or distributary there is a regulator in the head of the branch. If the branch takes off a large proportion of the water there is a double regulator. A distributary gives off watercourses as a canal gives off distributaries. The watercourses belong to the people and not to Government and they are cleared and maintained by the people. Each watercourse has a masonry head known as an “outlet” (Fig. 11). The outlet is the point where the water passes from the hands of Government officials to those of the cultivators. The outlet is of masonry and its opening is not adjustable but is fixed in such a way that its discharge, when the distributary is full, bears, as nearly as can be arranged, the same ratio to the F.S. discharge of the distributary as the area intended to be irrigated by the watercourse bears to that intended to be irrigated by the distributary.

The floor of the outlet is level with the bed of the distributary. It thus draws off rolling sand which might otherwise accumulate in the distributary. Small outlets are made of earthenware pipes, about ·4 feet in diameter, laid in concrete. Two pipes, or three, may be laid side by side. If more than three would be required, a masonry opening is adopted. The discharge through an outlet, is generally 2 to 5 c. feet per second per square foot of outlet area, and the head ·1 to ·5 feet.

For the tract of country allotted to any distributary, a contour map is prepared on a fairly large scale, say 4 inches to a mile. On the map the line is laid down and a rough longitudinal section, showing the ground level, is prepared as in the case of a canal.

It has already been stated (Art. 4) that a distributary is so designed that its water level, when three-fourths of the full supply is run, shall be well above the level of most of the ground along its course. In other words it should have a good command. A good rule is to allow a fall of ·5 feet from the level of the water in the distributary to that in the watercourse, a slope of 1 in 4,000 for the water flowing along the watercourse, and a fall of ·3 feet for the water at the tail of the watercourse to the level of the ground. This last level is, like the other ground levels, taken from the contour map. This procedure, in short, consists in making the water level of the watercourse at its head govern that of the distributary, just as the water level in the distributary at its head was made to govern that in the canal.

The enlarged contour map of the distributary area shows, among other things, the boundaries of the lands belonging to each village. Generally a watercourse supplies water to only one village. When, however, a village is far from the distributary, its watercourse has to pass for a long distance through other villages and it would be wasteful of water to have two separate watercourses. In such cases one watercourse may serve two villages or more. When a village is near to the distributary and its land extends for a long distance parallel to the distributary, it may have several watercourses for itself alone. A watercourse can generally be most conveniently dug along the boundary line of two villages, or there may be some other line which the people particularly desire.[15] Subject to, or modified by, these considerations a watercourse is designed to run on high ground like a distributary.

[15] They also frequently wish the “chak”--the area irrigated by a watercourse--so arranged that two men who are “enemies” shall not be included in the same “chak.” This condition can be complied with only up to a certain point. Arrangements may be modified but not in such a way as to upset the proper rules.

The great object is to reduce the total length of channels, _i.e._, minors and watercourses. No watercourse can be allowed to run alongside of or near to another. It may run alongside a canal or distributary when really necessary to gain command but not otherwise. The longer the watercourse the larger the chak. The discharge of an outlet may be anything up to 4 or 5 c. feet per second. This limits the size of a chak. If a chak is too big it can be split up or a minor can be designed. Very small chaks are to be avoided, but it is difficult to fix a minimum size. The irrigation boundary of the distributary, as fixed in the project, is shown on the map but in practice it will not be exactly followed. For various reasons the boundaries of a chak may run somewhat outside it or stop short of it.

Where a distributary gives off a minor and there is a double regulator, watercourses should, as far as possible, be taken off from one or other of the branch channels and not from upstream of the double regulator. Otherwise, irregularities are likely to occur, both of the regulators being partially closed at the same time--a thing which is never necessary in legitimate distribution of the supply--in order to head up the water and increase the discharges of the outlets.

A watercourse nearly always gives off branches and generally a system of turns is arranged by the farmers among themselves, each branch in turn taking the whole discharge of the watercourse for a day or part of a day, the other branches being closed by small dams of earth. To irrigate a field alongside the watercourse a gap is cut in its bank. For fields further away, smaller channels run off from the watercourses at numerous points. Several gaps and several field channels may be in flow at one time, and there is a dam in the watercourse below the lowest one.

Occasionally, on an old canal, one watercourse crosses another, the lands irrigated being at different levels, but such crossings do not often occur in systems of watercourses laid out according to modern methods. They are, however, quite legitimate.

The lines of the main watercourses are sketched on the map, their irrigation boundaries shown on it, and F.S. discharges allotted to them according to the areas which are to be dependent on them. In order that this may conveniently be done the “full supply duty” or “full supply factor” for the distributary is obtained. It bears the same ratio to the ordinary duty that the mean supply bears to the full supply. The total of the F.S. discharges of all the watercourses should, with an allowance for loss by absorption in the distributary, be the same as the F.S. discharge of the distributary. If the results are very discrepant it shows that the sizes of the outlets need revision. Possibly they may all be too large.

In “colonization” schemes where a canal is constructed to irrigate waste lands--which are the property of Government and which are divided into square blocks and given out to colonists--Government has complete control of the watercourse system, and can arrange it exactly as desired, but in other cases landowners often strenuously oppose the passage of watercourses through their lands. Compulsory procedure according to legal methods is tedious, but the practical rule is not to let anyone have water until any watercourses which are to pass through his land have been not only agreed to but constructed.

In ordinary cases Government possesses no power as to the precise line on which a watercourse is dug. It fixes the site of the outlet and assigns certain land to it, and sketches out the line of the watercourse. If the people choose to alter the line they can do so, but great alterations in the main watercourses are not generally feasible.

The positions of the outlets[16] having been settled after discussion with the cultivators, a table is prepared showing the chainage of the outlets, the probable head or difference between the F.S. level of the distributary and of the watercourse, and the F.S. discharge. From this the sizes of the outlets are calculated and shown in another column. If the length of the outlet barrel is not more than 5 or 6 times the diameter--in the case of a barrel whose cross section is not round or square, the mean diameter--the discharge can be calculated as for a “short tube,” but if longer the formula for flow in pipes should be used, allowance being, of course, made for the head lost at the entrance. The outlets generally consist at first of wooden “shoots” or long tubes, rectangular in cross section. This is because, after they have been tested by a year or two years’ working, the sizes nearly always require adjustment and the cultivators often wish to have the site shifted.

[16] The positions can be slightly altered by the Engineers for any sufficient reason.

The uncertainty as to the proper size of an outlet is due to several causes. If the command is very good there may be a clear fall from the outlet into the watercourse. In this case the discharge depends only on the depth of water in the distributary, and is known pretty accurately. But ordinarily the outlet is submerged, and its discharge depends on the difference between the water levels in the distributary and in the watercourse. The latter level is not fixed. The cultivators can lower it, to an extent which depends chiefly on the distance of the fields from the distributary, by deepening or widening the watercourse. In this way the discharge of the watercourse is increased except when a dam is temporarily made in it for the purpose of irrigating any comparatively high land. This uncertainty as to the discharge can in some cases be got over by building a cistern (Fig. 11). This has the same effect as raising the level of the barrel, the real outlet being no longer submerged, and the discharge depending on the depth of the crest of the overfall below the water in the distributary. But such cisterns add greatly to the cost of an outlet, and they can only be adopted when there is good command. A great cause of uncertainty as to the proper size of an outlet is the variability of the duty of the water on the watercourse. The soil may be clayey or sandy, the watercourse may be short or long, the crops grown may be ordinary ones or may be chiefly rice, which requires three or four times as much water as most other crops, and the cultivators may be careful or the opposite. Again, the people may, if the outlet gives a plentiful supply, often keep it closed, but there is no record of such closures nor would the people admit that they occur. These causes may all operate in one direction--on a whole distributary this cannot happen to the same extent--and thus enormous differences in duty may occur. There is no way of arriving at the proper size for an outlet except trial. Observations of the discharges of the outlets are of very limited use. The discharge may vary according to the particular fields being irrigated. Observations of discharges will be useful in cases where the people complain, or when the discharge is obviously much greater or much less than intended and will in such cases enable temporary adjustments to be made, but by placing a dam in a watercourse and turning the water on to a high field near its head the people can make it appear that the discharge is only a fraction of what it should be.

On any distributary there are generally some watercourses which have a poor command, the head at the outlet being, say, ·1 ft. or even less. Probably the irrigation is a good deal less than it should be. In such cases the rules may be set aside and a liberal size of outlet given. The size may be 2 or 3 times the calculated size. There is no harm in this. The irrigation cannot increase much. Similar cases frequently occur on inundation canals especially near the heads of canals or distributaries.

The construction of masonry outlets on a distributary is not usually a final settlement of the matter. Further adjustments become necessary. This matter will be dealt with in CHAPTER III.

On the older canals little or insufficient attention was given to the question of the sizes of outlets. The sizes were far too great and, as long as all the outlets in a distributary remained open, water could not reach the tail. The distributary used to be divided into two or three reaches and the outlets in the upstream reaches used to be closed periodically. The closures had to be effected through the agency of native subordinates and the system gave rise to corruption on a colossal scale. The tail villages never obtained anything like their proper share of water. The upper villages were over-watered and the soil was often water-logged and damaged. Moreover, even if all concerned had the best intentions, it was impossible to stop all leakage in the closed outlets, except by making earthen dams in the watercourses, and great waste of water resulted from this.

The water level of the distributary with ³⁄₄ full supply, designed so as to be at least ·5 ft. above the water level in the watercourse heads--or to be 1 foot above high ground if this simpler plan is adopted--is drawn on the rough longitudinal section and also the line of F.S., falls being introduced where desirable and the gradients, F.S. depths of water and widths of channels being arranged, just as in the case of a canal, so as to give the required discharges, velocities suited to the soil and a suitable ratio of depth to velocity. The bed width of a distributary decreases in whole numbers of feet. The decrease occurs at outlets but not at every outlet. As the channel becomes smaller its velocity becomes less and this necessitates, according to the laws of silting and scour, a reduced depth of water. The height and width of the banks in the tail portion of a distributary should be made rather greater than elsewhere--regard being had to the depth and volume of the water--so that breaches may not occur when the demand abruptly slackens. The longitudinal section of a distributary should have horizontal lines for showing the following:

1. Datum |5. Draw-off | 9. Bank width |13. Depth of | | | digging 2. Bed gradient |6. F.S. discharge|10. Height of |14. Bed level | | bank | 3. Village |7. Velocity |11. F.S. depth |15. Ground | | | level[17] 4. Land width |8. V₀ |12. Bed width |16. Chainage[18] -----------------+-----------------+-----------------+----------------

[17] Called “Natural Surface” in India.

[18] Called “Reduced Distance” in India.

A specimen of a longitudinal section is shown in Fig. 12. It shows only a few of the above items. In practice all would be shown, large sheets of paper being used with all the lines and titles printed on them.

When a distributary is constructed the side slopes are made 1 to 1 in excavation and 1¹⁄₂ to 1 in embankment. The sides usually silt up till they are ¹⁄₂ to 1 or even vertical. The silting up to ¹⁄₂ to 1 is, as in the case of a canal, allowed for in the designing. The berms are left so that, if any part of the side falls in, the bank will not also fall in. They also allow of widening of the channel. The remarks made in Art. 6 regarding the design of banks, apply to distributaries, especially large ones.

On a distributary there is seldom much spoil. Where there is no spoil, a strip of land, outside the bank and 10 feet wide, can be taken up on either bank from which to obtain earth for repairs. On a minor the width of the strip is sometimes only 5 feet.

When a distributary passes through land which is irrigated from wells, it frequently cuts through the small watercourses which run from the well to the fields. In such cases, either a syphon or a supplementary well is provided at Government cost. If several watercourses, all from the same well, are cut through, it is generally possible to combine them for the purpose of the crossing. The wishes of the cultivators in this matter are met as far as possible.

The procedure as regards laying out the line on the ground, digging trial pits, correcting the line and preparing the estimate are the same as for the case of a canal.

10. =Best System of Distributaries.=--Let AB (Fig. 13) represent a portion of a distributary, the irrigation boundary CD being two miles from AB. In order to irrigate a rectangular plot ACDB, the main and branch watercourses would be arranged somewhat as shown by the full and dotted lines respectively. Generally, the whole supply of the main watercourse would be sent in turn down each branch, the other branches being then dry. The average length open is AGE. The ends of the branches lie on a line drawn say 200 feet from the lines BD and DC, since it is not necessary for the watercourses to extend to the outside edges of the fields. Within the field there are small field watercourses which extend to every part of it. By describing three rectangles on AC, making AB greater than, equal to and less than AC, it can be seen that the average length of watercourse open is least--relatively to the area of the block--when AB is equal to AC, i.e., when the block served by the watercourse is square as in the figure. If AB is 4 times AC, the average length of watercourse open is increased--relatively to the area of the block--in about the ratio of 3 to 2. Moderate deviations from a square are of little consequence.

Suppose two parallel distributaries to be 4 miles apart, each of them being an average Indian one, say sixteen miles long with a gradient of one in 4,000, and side slopes of ¹⁄₂ to 1, the bed width and depth of water at the head being respectively 13·5 feet and 2·9 feet, and at the tail 3 feet and 1 foot. The discharge of the distributary, with N = ·0225, will be 72 c. ft. per second. The discharge available for the 2 mile strip along one bank will be 36 c. ft. per second. If the duty is 300 acres per c. ft. the area irrigated in this strip will be 10,800 acres, or 1,350 acres for each of the eight squares like ACDB. Each main watercourse would then have to discharge 4·5 c. ft. per second. Supposing its gradient to be 1 in 4,000 and its side slopes ¹⁄₂ to 1 and N to be ·0225, its bed width would be 3 feet and depth of water 1·45 feet. Its wet border would be 6·3 feet, and its average length 5280√2 + 5280 - 200 or 12,546 feet. Its wetted area would be 79,040 square feet, and the total wetted area of the 16 watercourses--on the two sides of the distributary--would be 1,264,640 square feet. The wetted border of the distributary itself is 19·5 feet at the head and 5 feet at the tail, average 12·25 feet, and its wetted area is 5,280 × 16 × 12·25 or 1,034,880 square feet.

If the distributaries were two miles apart, there would be twice the number of distributaries, and each square would be one square mile instead of four. Each watercourse would have to discharge 1·125 c. ft. per second. It would have a bed width of 2 ft., depth of water ·8 ft., wet border 3·8 feet, length 6,173 feet, and wetted area 23,457 feet. The total wetted area of the 64 water courses would be 1,501,248 square feet, or 18 per cent. more than before. Each distributary would discharge 36 c. ft. per second, the bed width and depth at the head being 10 feet and 2·24 feet, and at the tail 2 feet and ·75 feet. The wet border at the head and tail would be 14·5 and 3·5 feet, mean 9 feet, and the wetted area of the two distributaries would be 1,520,640 square feet or 50 per cent. more than before. Supposing that, in the case of the larger distributary considered above, the 2-mile square was considered too large, and that rectangles 1 mile wide were adopted, so that the watercourses were a mile apart, their number would be doubled and their length and size reduced. Their total wetted area would not be greatly affected, but the difference in the wetted areas of the two small distributaries as compared with the one large one, would be the same as before. In practice, of course, distributaries are not always parallel, nor are the blocks of irrigation all squares, and frequently, owing to peculiarities in the levels of the ground or the features of the country, or the boundaries of villages, it is necessary to align the watercourses in a particular manner, or to construct more than one watercourse where one would otherwise have sufficed, but the above calculations show in a general way the advantages of large watercourses and of not placing the distributaries too near together.

It is commonly said that a watercourse discharging more than 4 or 5 c. ft. per second is objectionable because the cultivators, if there are too many of them on one watercourse, cannot organize themselves in order to work it and keep it in order. This matter is much exaggerated. On the inundation canals of the Punjab a watercourse often discharges 10 c. ft. per second, and is several miles long and requires heavy clearances, but the people have no particular difficulty in managing it. Kennedy, a great authority on questions of irrigation, states that the length of a watercourse may be three miles. This, if the angle made by a watercourse with the distributary is 45°, gives rather more than two miles as the width of the strip to be irrigated.

Suppose that a distributary instead of being two miles from each side of the irrigated strip, ran along one side of it, and was four miles from the other side. If the block were square, as before, the side of a square would be 4 miles, and each watercourse would have to discharge 18 c. ft. per second, which is far too much. The blocks would have to be rectangles, each being only one mile wide measured parallel to the distributary. It has been already seen that the length of watercourse in this case is greater than when the block is square and each side is two miles. Thus centrality in the alignment of the distributary is an advantage.

A minor distributary has been defined (CHAPTER II., Art. 3) as being one discharging not more than 40 c. ft. per second, but the term has come to be used to designate a branch of a major distributary, and in that sense it will be used in this article. When the shape of the area commanded by a distributary is such that watercourses exceeding 2 miles in length would otherwise be required, one or more minors are often added. Frequently it is a question whether to let some of the watercourses be more than two miles long, or to construct a minor and thus shorten the watercourses to perhaps only one mile. Which method is best has not been definitely settled. It is known that the loss of water in watercourses is heavy, but if a minor is added the loss in it has to be considered. The loss must be high in any channel in which the ratio of wet border to sectional area is small. The minor also costs money in construction and in maintenance. On the whole the matter, as far as concerns cost and loss of water, is, perhaps, almost evenly balanced, but as regards distribution of the supply a system without minors is preferable. The off-take of a minor is generally far from the canal, i.e., in a more or less out-of-the-way place, and it is impossible to see that the regulation is properly carried out. Irregularities and corruption are sure to arise. Even if the supply is fairly distributed as between the minor and the distributary it is almost certain that the regulator, if a double one, will be manipulated for the illegal benefit of outlets in the distributary upstream of the bifurcation. There are sure to be some such outlets not very far distant. In any case each minor adds one, if not two, to the already very large number of gauges which have to be entered daily in the sub-divisional officer’s register (CHAPTER III., Art. 3), and adds also to the mileage of channel to be inspected and maintained. These considerations should, in many cases, though of course not in all, turn the scale against the construction of a minor. At one time it became usual to construct minors even when watercourses more than two miles long would not otherwise have resulted. This custom was condemned some years ago, and is not likely to be re-established. Most of the difficulties just mentioned can, in the case of a minor which is not too large, either absolutely or relatively to the main distributary downstream of the off-take, be got over by making the minor head like a watercourse outlet, building it up to the proper size, removing the regulating apparatus and abolishing the reading of the gauge, but in this case the minor is not likely to be bigger than a large watercourse. Such minors should not be constructed, and any existing ones should, after the head has been treated as above, be made over to the people and considered as watercourses.

11. =Outlets.=--The top of the head and tail walls of an outlet are level with the F.S. levels in the distributary and watercourse respectively. The steps in the head wall enable the cultivators to go down either to stop up the outlet or to remove any obstruction. The stepping is arranged so as to fall inside the side slope ultimately proposed. It is usual, in some places, to have the entrance to the “barrel” of the outlet made of cast iron. The cast iron pieces are made of various standard sizes. This to some extent prevents the “barrel” being built to a wrong size. A discrepancy between the size of the masonry barrel and that of the iron would be noticed, but if the masonry barrel is built too large the iron head does not always restrict the discharge. The action is the same as in a “diverging tube” well known in hydraulics.

For sizes up to about 50 or 60 square inches the barrel should be nearly square. For larger sizes the height should exceed the width. Up to about 100 or 120 square inches the width can be kept down to 7 or 8 inches so that an ordinary brick can be laid across to form the roof. For larger outlets the height can be from 1·5 to 3 times the width, and the roof can be made of large bricks, concrete blocks or slabs of stone or of a flat arch of brickwork or by corbelling, but in this last case there should be two complete courses above the top of the outlet. The less the width the cheaper the roof, the easier the adjustment of size and the less the tendency to silt deposit during low supplies. If pipes are used they should be laid in concrete. If cast iron head pieces are to be used there should be several sizes of one width and the widths of the masonry outlets should be made to suit these widths.

A masonry outlet is not generally built till the watercourse has been sometime in use. The exact position of the outlet should then be so fixed that the watercourse shall run out straight or with a curve and should not be crooked.

The width between parapets should be, for a driving road or one to be made into such, 10 ft. (if the bank is wider, it should be narrowed just at the outlet site) and for a non-driving road, 8 feet to 3 feet according to the ultimate width of the bank. Earth backing should be most carefully put in and rammed, otherwise a breach may occur and the outlet be destroyed.

Various attempts have been made to provide gates or shutters for outlets. The chief result has been trouble and increased cost. If grooves are made and shutters provided, the shutters are soon broken or lost by the people. Hinged flap shutters are objectionable because they are often closed by boys or by malicious persons or by neighbours who wish to increase the supply in their own outlet. The cultivator, when he wishes to reduce the supply or to close the outlet, can easily do this by obstructing the orifice with a piece of wood or an earthenware vessel or a bundle of brushwood or grass.

As regards temporary outlets, wooden outlets if large (unless made of seasoned wood and therefore costly) are liable to give great trouble. Water escapes round the outside or through the joints. Pipes may do well if laid in puddle but are brittle and costly if of large size. The irrigators may interfere both with wooden outlets and pipes and they are liable to be displaced or broken. A temporary outlet, if small, can be made of bricks laid in mud. The joints can be pointed with lime mortar. When the outlet is made permanent the same bricks are used again. But all kinds of temporary outlets are liable to give trouble especially in light or sandy soil. There is much to be said in favour of building masonry outlets at the first, making a barrel only, _i.e._, omitting the head and tail walls and taking the chance of having to alter the size. The alteration is not very expensive. The head and tail walls are built when the size has been finally settled. The adjustment can be made by raising or lowering the roof. This should be done over the whole length of the outlet but lowering can be done temporarily over a length of 3 feet at the tail end of the outlet. This can be done even when the distributary is in flow. A reduction over a short length at the upstream end of a barrel does not, as already remarked, necessarily reduce the discharge much.

On inundation canals the rules regarding outlets have to be modified. Great numbers of watercourses take off directly from the canals. In such cases, especially near the head of a canal, the ground to be watered is often 5 to 8 feet above the canal bed and it is wholly unsuitable to place the outlet at bed level. The cost of the tail wall would be excessive. The floor level in such cases must be at about the lowest probable cleared bed level of the watercourse, say, in order to be safe, a foot or half a foot below the usual cleared bed of the watercourse, so that water need never be prevented from entering the watercourse. The irrigators should be consulted as to the floor level and their wishes be attended to as far as possible. For lift outlets the floor should be at the bed level of the canal or distributary. If this bed is to be raised in the course of remodelling, the floor should be at the old bed level until the bed has actually been raised, unless there is a weir which raises the water. It is necessary that lift outlets should work however small the canal supply may be. In a distributary or small canal, the head wall should be built up to F.S. level but in a canal with deep water the head wall should reach up to just above the roof of the outlet and be submerged in high supplies. The stepping of the head wall should be set back if the channel is to be widened and should project into the channel if the channel is to be narrowed. The centre line of the channel near the outlet site must always be laid down and the outlet built at right angles to it and also at the correct distance from it.

Occasionally there is a wide berm, say 20 ft. or even 50 ft., between a channel and its bank. In such a case the outlet should be built to suit the bank. The long open cut is however objectionable because the people clear it and heap the spoil in Government land. Sometimes the bank, especially if it is crooked, can be shifted so as to come close to the channel at the outlet site. Sometimes the outlets on inundation canals are large. For outlets of more than 2·5 square feet in area, grooves should be provided so that the cultivators can use a gate if necessary.

12. =Masonry Works.=--The positions and descriptions of all the masonry works of a proposed canal or distributary are of course shown on the longitudinal section of the channel and from this the discharges and water levels are obtained. The principles of design to be followed[19] for bridges, weirs, falls, regulators and syphons, are discussed in _River and Canal Engineering_. It is mentioned that there is no special reason for making the waterway of a regulator exactly the same as that of the stream, and that the waterway may be such as to give the maximum velocity considered desirable, and that the foundations of a bridge should be made so deep that it will be possible to add a floor, at a lower level than the bed of the stream--with the upstream and downstream pitching sloping up to the bed--so as to increase the waterway and so save pulling down the bridge in case the discharge of the channel is increased. It remains to consider certain points affecting Irrigation Canals.

[19] So far as concerns their capacity for dealing with flowing water.

The span of a bridge, where there are no piers, is generally made as shown by the dotted lines in Figure 14, so that the mean width of waterway is the same as that of the channel. The arches, in Northern India, used at one time to be 60° as shown by the upper curved line, but in recent years arches of 90° as shown by the lower curved line, have frequently been adopted, the springing of the arch being below the F.S. level, so that the stream is somewhat contracted. The 90° arch gives a reduced thickness and height of abutment. It causes increased disturbance of the water, and this may necessitate more downstream protection. An advantage of having the springing not lower than the F.S. level is that this admits of a raising of the F.S. level in case the channel is remodelled, and this arrangement is still common on distributaries.

When a fall and bridge are combined, the bridge is placed below the fall as this gives a lower level for the roadway. The side walls of the fall are produced downstream to form those of the bridge.

The roads in India are generally unfenced and the banks of canals close to bridges, on both sides of the canal and both above and below the bridge, are generally more or less worn down by cattle, which, when being driven home in the evening and out to graze in the morning, go down to the stream to drink. In order to prevent this damage the banks are sometimes pitched, above the bridge as well as below it, but the cattle generally make a fresh “ghát” further away. The best plan is to allow a “ghát” on one bank either above or below the bridge and to protect the other three places.

In the Punjab the widths of roadways between the kerbs and parapets of bridges respectively have been fixed as follows:--

--------------+----------------+------------------- KIND OF ROAD.| NEAR TOWNS.[20]|IN THE COUNTRY.[21] --------------+------+---------+--------+---------- |Kerbs.|Parapets.| Kerbs. | Parapets. --------------+------+---------+--------+---------- Provincial | 22 | 23·5 | 16 | 17·5 District | 18 | 19·5 | 14 | 15·5 Village | 14 | 15·5 | 8·5 | 10 --------------+------+---------+--------+----------

[20] The figures show the maximum. The general width should be the same as for neighbouring bridges on the same road.

[21] The parapets should be whitewashed so as to be visible at night.

Fig. 15 shows a head regulator for a distributary. The scale is 10 feet to an inch. It has a double set of grooves for the insertion of the planks with which the regulation is effected. Only one set of grooves is ordinarily used, but when the distributary has to be closed for silt clearance and all leakage stopped, both sets of grooves can be used and earth rammed in between the two sets of planks. The floor is shown a foot lower than the bed of the distributary. This reduces the action of the water on the floor, and enables the bed of the distributary to be lowered if ever the occasion for this should arise. This is a good rule--in spite of the fact that in re-modellings the tendency is for the beds to be raised--for all regulators or bridges, a raised sill being added (in regulators) to reduce the length of the needles or the number of the planks. Such sill should, where needles are to be used, be fairly wide, especially if regulation is to be done while the masonry is somewhat new. The distributary shown has a bed width of 10 ft. The span of the two openings in the head might have been four feet each, but are actually five feet, and this enables the distributary to be increased in size at any time. The pitched portion of the channel tapers. Unless needles are used, instead of horizontal planks, spans are not usually greater than 5 or 6 feet. Longer spans would give rise to difficulties in manipulating the planks. Sometimes distributary heads are built skew, but there is seldom or never any good reason for this. A curve can always be introduced below the head to give the alignment the desired direction.[22] The small circles shown on the plan are “bumping posts.” On the left is shown a portion of the small raised bank at the edge of the road.

[22] The curve can be quite sharp (see CHAP. I., Art. 2), and can be made, if necessary, within the length of the pitching.

Figure 16 is a double regulator with needles. The scale is 30 feet to an inch. The spans are 15 feet. The roadway is on arches, but the regulating platform on steel beams. The needles are seen at the upstream sides of the regulators. They are worked from the platforms to which access is obtained through the gaps in the upstream parapets. The regulating platform should generally be only just clear of the F.S. level, and therefore lower than the roadway.

Frequently the roadway of a bridge or small regulator is carried, not on arches, but on steel beams. The railings may be of wood or of gas pipe with the ends plugged, running through angle iron posts. In the case of such a regulator the roadway is sometimes so light that camels are not allowed to cross over. This causes unnecessary hardship. Bridges are not too numerous. If the regulation is done by gates, both road and platform are carried on arches.

The regulators on inundation canals, and some on perennial canals, are not strong enough to admit of the flow of water being entirely stopped, so that the depth of water would be perhaps 10 feet upstream and nil downstream. This might cause the overturning of the piers, or the formation of streams under the floor. In such cases a maximum permissible heading up is decided on. Such orders are, in India, liable to be lost sight of in course of time, and they are, at least on inundation canals, where sudden emergencies often occur, hardly reasonable. An engine driver is not told that he must never entirely close his throttle valve. Regulators should be so designed that the water can be completely shut off.

The following remarks show the chief points in favour of needles and horizontal planks respectively.

_Advantages of Needles._ Needles can be placed or removed by one man.

Needles do not require hooks, etc., which are liable to be broken or lost.

A needle regulator requires few piers, and is therefore cheap.

Water falling over planks throws a strain on the floor.

Regulation with needles is easy and rapid. A jammed plank, especially if low down and not horizontal, may give great trouble.

_Advantages of Planks._ Floating rubbish is not liable to collect above the Regulator because the water flows over the planks.

By means of double grooves and earth filling, leakage can be quite stopped.

For large works the advantages are generally with needles, but for small works, _e.g._ distributary heads and shallow water, with planks. Needles 14 feet long are not too long for trained men. Planks are more likely than needles to arrest rolling sand, and this can be taken into consideration in designing double regulators. See number 8 of Kennedy’s rules, Article 5. When planks are used there should be two sets of grooves. Planks are very suitable for escape heads which have only occasionally to be opened, earth being filled in between the two sets of planks.

Regarding notched falls, in the case of small distributaries the notches are so narrow that they are extremely liable to be obstructed either accidentally by floating rubbish or wilfully by persons whose outlets are upstream of them. Weirs are not open to this objection, and are frequently adopted. There is not the least chance of their causing any silting worth mentioning. A simple weir if made of the proper height for the F.S. discharge, will cause a slight heading up with ³⁄₄ths of the F.S. discharge, and this unfairly benefits any outlets for a considerable distance upstream of the weir. This difficulty can be got over by making the weir as in Fig. 16A.

For cisterns below falls the usual rule for the depth is

K = H + ∛H √D

where H is the depth of water in the upstream reach, and D is the difference between the upstream and downstream water levels. Another rule for distributaries is

H + D K = ----- 3

the length of the cistern being 3 H and its width the bed width of the channel.

At “incomplete” falls, i.e., where the tail water level is above the crest, it is not unusual to construct a low-level arch, which forms a syphon. The object is to allay the surging of the surface water.

The question of skew bridges has been dealt with in Art. 3. Another question is that of the heights of bridges. Irrigation channels, especially the smaller ones, are very frequently at a high level, and bridges have ramps which are expensive to make and to maintain, and are inconvenient. The lowering of distributary bridges in such cases, so that they become syphons, or nearly so, has often been advocated and is frequently desirable. The bed should slope down to the floor and up again. The heading up can be reduced by giving ample waterway, but it will not be necessary to do this if there is head to spare. The fall in the water surface can be recognised and shown on the longitudinal section. The structure becomes one of the incomplete falls above described. The crown of the arch can, if desirable, be kept above F.S. level, so that floating rubbish will not accumulate.

The width between the parapets of a regulator can be 10 feet in the case of a driving road. It may be less, according to the width of the bank, in other cases.

The upper layer of the floor of a bridge or regulator is of brick on edge. Below this there is a layer of brick laid flat, and below this, concrete of a thickness ranging from ·5 feet to 3 feet. The thicknesses of piers range from 1·5 to 3 feet.

The bricks used for canal work in Northern India are 10 inches long, 4⁷⁄₈ inches wide, and 2³⁄₄ inches thick. The thicknesses of walls are about ·83, 1·25, 1·7, 2·1, 2·5 feet, and so on.

The slopes of ramps should be about 3 in 100 for district roads, and 5 in 100 for village roads.

Railings should be provided along the tops of high walls and top of pitching near to public roads or canal patrol roads. Bumping posts should be provided for all parapets, and should not be so placed as to seriously obstruct the roadway.

The quarters for the regulating staff should, when convenient, be in the fork between the two principal branches. They may be on the bank--with foundations on pillars carried down to ground level--but not in such a position as to obstruct the road or any road likely to be made. Rests consisting of two parallel timbers bolted to blocks of masonry reaching up a foot from the ground, should be provided for the needles or planks. The bolt head should be countersunk so as not to damage the needles and planks when they are hurriedly laid down.

When two or more works are close together they should be made to conform, and the whole site should be considered with reference to a neat and suitable arrangement of works, ramps and roadways. If an outlet is near to a minor or distributary head the parapets of the two should be in line. If two masonry works of any kind are near together it is often suitable to pitch the intervening space. If there are outlets or distributaries on opposite banks they should be exactly opposite each other. Where a road crosses a bridge or regulator, the bank should be at the same level as the road, the bank being gradually ramped back to its original level. The space in front of any quarters should have a slight slope for drainage, but otherwise be at one level and be connected with the road or bank by proper ramps. The berm or bank should be made at the exact level of the top of any pitching or side wall which adjoins it. Wing walls are frequently made too short, so that the earth at their ends forms a steep slope and is worn away, and the bank or roadway is cut into. The walls should extend to such a point that the earth at their ends cannot assume a slope steeper than the slope of the bank.

It is obvious that for every masonry work there should be a large scale site plan[23] showing all roads, ramps, and adjoining works, both existing and proposed roads being shown for some little distance from the work.

[23] It is, or was until recently, in some parts of India, the custom to omit the preparation of site plans, and to leave the fixing of the exact site of a work and the arrangement of ramps and other details to the judgment of the assistant engineer who was building it. Much unsightly work resulted. A chief engineer in the Punjab recently issued some orders on the subject.

For each kind of masonry work there is usually a type design. A few of its dimensions, which are fixed, are marked on it. The other dimensions are variable. It would be a great advantage to add to the design a tabular statement to show how these dimensions should vary under different circumstances.

13. =Pitching.= The object of pitching upstream of bridges or regulators or downstream of bridges where there may be little or no scouring action, may be partly to protect the bank from damage by cattle or wear, or to prevent sandy sides from falling in. In such cases there may be pitching of the sides only, and it may be of brick on edge laid dry and under this one brick flat resting on rammed ballast (Fig. 17). Downstream of regulators or weirs and downstream of bridges if contracted or having piers which cause a rush of water, especially if the soil is soft, the side pitching may be as above, but with the bricks over one-sixth of the area placed on end and projecting for half their length. This “roughened pitching” tends somewhat to reduce the eddying. The bed protection should be solid concrete or blocks of concrete or masonry. Immediately downstream of regulators or weirs where there is great disturbance, both side and bed pitching may consist of solid concrete or of concrete or masonry blocks (Fig. 18).

Three kinds of toe walls are shown in Figures 17, 19 and 20. The kind shown in Fig. 19 contains, for a given depth below the bed, far more masonry than the one shown in Fig. 17. It is also liable to be displaced and broken if scour occurs.

The earth should in all cases be carefully cut to the proper slope, so that no made earth has to be added. If the slope has already fallen in too much, well rammed earth should be added. The flat brick and rammed ballast can be varied as the work proceeds, more being used in soft places and less in hard.

In some parts of the Punjab, large bricks, the length, breadth, and thickness being about twice the corresponding dimensions of an ordinary brick, are made, and are extremely useful and cheap for pitching. Where the soil is sandy such bricks can be burned without cracking.

Sometimes the curtain wall which runs across the bed at the downstream end of the pitching is carried into the banks and built up so as to form a profile wall (Fig. 21). This is not very suitable, because the pitching of the sides is apt to settle and leave the profile wall standing out. It is better to lay a row of blocks on the slope. If a hole tends to form in the bed downstream of the curtain wall, blocks of masonry or concrete can be laid and left to take up their own positions (Fig. 22).

When scour of the bed or sides occurs downstream of pitching, it is sometimes said that any extension of the pitching downstream is followed by extension of the scour. This may happen if the cross section of the stream downstream of the pitched section has become greater than the pitched section. In this case there is eddying, due to abrupt enlargement of the stream where the pitching ends. The increased width and lowered bed level (not counting mere local hollows) of the stream should be adhered to in the pitching. Where the masonry of the regulator ends and the pitching begins, there will be an abrupt or tapered enlargement, but the eddies--at very low supplies there may be a fall--cannot do harm.

This principle of enlarging the pitched cross section can be followed, even in a new channel, if the soil is light and scour is feared, and for this reason the matter is mentioned in the present Chapter instead of in Chapter III. It was once the custom to splay out the sides of a channel, downstream of a regulator or weir, so as to form a sort of pool in which the eddies exhausted themselves, but this gives curved banks and requires extra land and is not a very convenient or neat arrangement. Where scour of the sides is likely to occur, or has occurred, immediately downstream of the pitching the latter may be turned in as shown in Fig. 23.

Pitching has constantly to be replaced or extended owing, generally, to failure to pitch a sufficient length or to ram well the earth under the pitching, or to use properly rammed ballast or flat brick, or to give proper bed protection, or to the use of dry brick pitching when a stronger kind is needed.

The side slopes of pitching should be 1 to 1. They can be ¹⁄₂ to 1 in rare cases, _e.g._, when there is no room for 1 to 1, or in continuation of existing ¹⁄₂ to 1 pitching. No absolute rule can be laid down as to the length to be pitched, but in a Punjab distributary it is often about 5 times the bed width.

14. =Miscellaneous Items.= On Indian canals the chainage[24] is marked at every thousand feet. Five thousand feet is called a “canal mile.” The distance marks are often cast iron slabs, fixed in a cylindrical block of brickwork about 2·1 feet in diameter and 1·5 feet high, the upper edge being rounded to a radius of ·4 feet. The wedge-shaped bricks for these blocks are specially moulded. The iron slab should project about eight inches and have about a foot embedded in the brickwork.

[24] In India, instead of the simple word “chainage” the term “reduced distance” is used. It is the distance reduced to a common starting point as levels are reduced to mean sea level. The expression is puzzling to non-professionals and new comers.

On a canal having a wide bank the distance mark is put at the outer edge of the patrol bank, earth being added, if necessary, to increase the width. On a distributary with a narrow bank the mark should be on the opposite bank not the patrol bank. To enable the miles to be easily distinguished the masonry block can be sunk only ·5 foot in the ground, the others being sunk a foot. In all cases the masonry block rests on a pillar, 1·7 feet square, of bricks laid in mud, carried down to the ground level.

Profile walls (Fig. 21, page 92) used occasionally to be built at frequent intervals along a distributary. They will not prevent scour occurring, if the stream is tending to scour, unless very close together. Such walls are of some use as showing whether the channel is altering, but they are expensive and have to be altered if, as often happens, the channel is remodelled. It is a much better plan to lay down blocks--about 1¹⁄₄ foot cubes--of masonry or concrete, along the centre line at every 500 feet, with their upper faces level with the bed. If the bed scours they may be displaced but otherwise they are useful not only for showing what silt, if any, has deposited, but for showing the centre line of the channel. Without them the centre line is liable to be altered in silt clearances or berm cuttings. To enable a block to be readily found and to be replaced in proper position if displaced, there should be two small concrete pillars exactly opposite to it and equidistant from it, one on either bank of the channel. Such blocks and pillars may with advantage be placed at quite short intervals on curves.

The rest houses for the use of officials on tour are generally at intervals of about 8 to 14 miles. There is generally a rest house near to a large regulator and frequently there is one near to a small regulator. This facilitates inspection work and discharge observations and it saves money, because the house can be looked after by one of the regulating staff. Not infrequently the house is placed just too far away from the regulator. Similarly if a rest house is near a railway station it should be within a quarter of a mile of it--always provided that this does not bring it too near to villages or huts--and not a mile or more away as is sometimes the case. It is also a mistake to place a rest house off the line of channel unless perhaps when it is on a district road which crosses the channel.