1911 Encyclopædia Britannica/Irrigation

IRRIGATION (Lat. in, and rigare, to water or wet), the artificial application of water to land in order to promote vegetation; it is therefore the converse of “drainage” (q.v.), which is the artificial withdrawal of water from lands that are over-saturated. In both cases the object is to promote vegetation.

I. General.—Where there is abundance of rainfall, and when it falls at the required season, there is in general no need for irrigation. But it often happens that, although there is sufficient rainfall to raise an inferior crop, there is not enough to raise a more valuable one.

Irrigation is an art that has been practised from very early times. Year after year fresh discoveries are made that carry back our knowledge of the early history of Egypt. It is certain that, until the cultivator availed himself of the natural overflow of the Nile to saturate the soil, Egypt must have been a desert, and it is a very small step from that to baling up the water from the river and pouring it over lands which the natural flood has not touched. The sculptures and paintings of ancient Egypt bear no trace of anything approaching scientific irrigation, but they often show the peasant baling up the water at least as early as 2000 B.C. By means of this simple plan of raising water and pouring it over the fields thousands of acres are watered every year in India, and the system has many advantages in the eyes of the peasant. Though there is great waste of labour, he can apply his labour when he likes; no permission is required from a government official; no one has to be bribed. The simplest and earliest form of water-raising machinery is the pole with a bucket suspended from one end of a crossbeam and a counterpoise at the other. In India this is known as the denkli or paecottah; in Egypt it is called the shadúf. All along the Nile banks from morning to night may be seen brown-skinned peasants working these shadúfs, tier above tier, so as to raise the water 15 or 16 ft. on to their lands. With a shadúf it is only possible to keep about 4 acres watered, so that a great number of hands are required to irrigate a large surface. Another method largely used is the shallow basket or bucket suspended to strings between two men, who thus bail up the water. A step higher than these is the rude water-wheel, with earthen pots on an endless chain running round it, worked by one or two bullocks. This is used everywhere in Egypt, where it is known as the sakya. In Northern India it is termed the harat, or Persian wheel. With one such water-wheel a pair of oxen can raise water any height up to 18 ft., and keep from 5 to 12 acres irrigated throughout an Egyptian summer. A very familiar means in India of raising water from wells in places where the spring level is as much sometimes as 100 ft. below the surface of the field is the churras, or large leather bag, suspended to a rope passing over a pulley, and raised by a pair of bullocks which go up and down a slope as long as the depth of the well. All these primitive contrivances are still in full use throughout India.

It is not improbable that Assyria and Babylon, with their splendid rivers, the Euphrates and Tigris, may have taken the idea from the Nile, and that Carthage and Phoenicia as well as Greece and Italy may have followed the same example. In spite of a certain amount of investigation, the early history of irrigation in Persia and China remains imperfectly known. In Spain irrigation may be traced directly to the Moorish occupation, and almost everywhere throughout Asia and Africa where the Moslem penetrated is to be found some knowledge of irrigation.

Reservoirs are familiar everywhere for the water-supply of towns, but as the volume necessary, even for a large town, does not go far in irrigating land, many sites which would do admirably for the former would not contain water Spain. sufficient to be worth applying to the latter purpose. In the Mediterranean provinces of Spain there are some very remarkable irrigation dams. The great masonry dam of Alicante on the river Monegre, which dates from 1579, is situated in a narrow gorge, so that while 140 ft. high, it is only 190 ft. long at the crest. The reservoir is said to contain 130 million cub. ft. of water, and to serve for the irrigation of 9000 acres, but unless it refills several times a year, it is hardly possible that so much land can be watered in any one season. The Elche reservoir, in the same province, has a similar dam 55 ft. high. In neither case is there a waste-weir, the surplus water being allowed to pour over the crest of the dam. South of Elche is the province of Murcia, watered by the river Segura, on which there is a dam 25 ft. high, said to be 800 years old, and to serve for the irrigation of 25,000 acres. The Lorca dam in the same neighbourhood irrigates 27,000 acres. In the jungles of Ceylon are to be found remains of gigantic irrigation dams, and on the neighbouring India. mainland of Southern India, throughout the provinces of Madras and Mysore, the country is covered with irrigation reservoirs, or, as they are locally termed, tanks. These vary from village ponds to lakes 14 or 15 m. long. Most of them are of old native construction, but they have been greatly improved and enlarged within the last half century. The casual traveller in southern India constantly remarks the ruins of old dams, and the impression is conveyed that at one time, before British rule prevailed, the irrigation of the country was much more perfect than it is now. That idea, however, is mistaken. An irrigation reservoir, like a human being, has a certain life. Quicker or slower, the water that fills it will wash in sand and mud, and year by year this process will go on till ultimately the whole reservoir is filled up. The embankment is raised, and raised again, but at last it is better to abandon it and make a new tank elsewhere, for it would never pay to dig out the silt by manual labour. It may safely be said that at no time in history were there more tanks in operation than at present. The ruins which are seen are the ruins of long centuries of tanks that once flourished and became silted up. But they did not all flourish at once.

In the countries now being considered, the test of an irrigation work is how it serves in a season of drought and famine. It is evident that if there is a long cessation of rain, there can be none to fill the reservoirs. In September 1877 there were very few in all southern India that were not dry. But even so, they helped to shorten the famine period; they stored up the rain after it had ceased to fall, and they caught up and husbanded the first drops when it began again.

Irrigation effected by river-fed canals naturally depends on the regimen of the rivers. Some rivers vary much in their discharge at different seasons. In some cases this variation is comparatively little. Sometimes the flood Irrigation canals. season recurs regularly at the same time of the year; sometimes it is uncertain. In some rivers the water is generally pure; in others it is highly charged with fertilizing alluvium, or, it may be, with barren silt. In countries nearly rainless, such as Egypt or Sind, there can be no cultivation without irrigation. Elsewhere the rainfall may be sufficient for ordinary crops, but not for the more valuable kinds. In ordinary years in southern India the maize and the millet, which form so large a portion of the peasants' food, can be raised without irrigation, but it is required for the more valuable rice or sugar-cane. Elsewhere in India the rainfall is usually sufficient for all the cultivation of the district, but about every eleven years comes a season of drought, during which canal water is so precious as to make it worth while to construct costly canals merely to serve as a protection against famine. When a river partakes of the nature of a torrent, dwindling to a paltry stream at one season and swelling into an enormous flood at another, it is impossible to construct a system of irrigation canals without very costly engineering works, sluices, dams, waste-weirs, &c., so as to give the engineer entire control of the water. Such may be seen on the canals of Cuttack, derived from the Mahanadi, a river of which the discharge does not exceed 400 cub. ft. per second in the dry season, and rises to 1,600,000 cub. ft. per second in the rainy season.

Very differently situated are the great canals of Lombardy, drawn from the Ticino and Adda rivers, flowing from the Maggiore and Como lakes. The severest drought never exhausts these reservoirs, and the heaviest rain can never convert these rivers into the resistless floods which they would be but for the moderating influence of the great lakes. The Ticino and Adda do not rise in floods more than 6 or 7 ft. above their ordinary level

or fall in droughts more than 4 or 5 ft. below it, and their water is at all seasons very free from silt or mud. Irrigation cannot be practised in more favourable circumstances than these. The great lakes of Central Africa, Victoria and Albert Nyanza, and the vast swamp tract of the Sudan, do for the Nile on a gigantic scale what Lakes Maggiore and Como do for the rivers Ticino and Adda. But for these great reservoirs the Nile would decrease in summer to quite an insignificant stream. India possesses no great lakes from which to draw rivers and canals, but through the plains of northern India flow rivers which are fed from the glaciers of the Himalaya; and the Ganges, the Indus, and their tributaries are thus prevented from diminishing very much in volume. The greater the heat, the more rapidly melts the ice, and the larger the quantity of water available for irrigation. The canal system of northern India is the most perfect the world has yet seen, and contains works of hydraulic engineering which can be equalled in no other country. In the deltas of southern India irrigation is only practised during the monsoon season. The Godaveri, Kistna and Kaveri all take their rise on the Western Ghats, a region where the rainfall is never known to fail in the monsoon season. Across the apex of the deltas are built great weirs (that of the Godaveri being 21/2 m. long), at the ends and centre of which is a system of sluices feeding a network of canals. For this monsoon irrigation there is always abundance of water, and so long as the canals and sluices are kept in repair, there is little trouble in distributing it over the fields. Similar in character was the ancient irrigation of Egypt practised merely during the Nile flood—a system which still prevails in part of Upper Egypt. A detailed description of it will be found below.

Where irrigation is carried on throughout the whole year, even when the supply of the river is at its lowest, the distribution of the water becomes a very delicate operation. It is generally considered sufficient in such cases if during Distribution of the water. any one crop one-third of the area that can be commanded is actually supplied with water. This encourages a rotation of crops and enables the precious liquid to be carried over a larger area than could be done otherwise. It becomes then the duty of the engineer in charge to use every effort to get its full value out of every cubic foot of water. Some crops of course require water much oftener than others, and much depends on the temperature at the time of irrigation. During the winter months in northern India magnificent wheat crops can be produced that have been watered only twice or thrice. But to keep sugar-cane, or indigo, or cotton alive in summer before the monsoon sets in in India or the Nile rises in Egypt the field should be watered every ten days or fortnight, while rice requires a constant supply of water passing over it.

Experience in these sub-tropical countries shows the absolute necessity of having, for successful irrigation, also a system of thorough drainage. It was some time before this was discovered in India, and the result has been the deterioration of much good land.

In Egypt, prior to the British occupation in 1883, no attempt had been made to take the water off the land. The first impression of a great alluvial plain is that it is absolutely flat, with no drainage at all. Closer examination, however, shows that if the prevailing slopes are not more than a few inches in the mile, yet they do exist, and scientific irrigation requires that the canals should be taken along the crests and drains along the hollows. In the diagram (fig. 1) is shown to the right of the river a system of canals branching out and afterwards rejoining one another so as to allow of no means for the water that passes off the field to escape into the sea. Hence it must either evaporate or sink into the soil. Now nearly all rivers contain some small percentage of salt, which forms a distinct ingredient in alluvial plains. The result of this drainless irrigation is an efflorescence of salt on the surface of the field. The spring level rises, so that water can be reached by digging only a few feet, and the land, soured and water-logged, relapses into barrenness. Of this description was the irrigation of Lower Egypt previous to 1883. To the left of the diagram is shown (by firm lines) a system of canals laid out scientifically, and of drains (by dotted lines) flowing between them. It is the effort of the British engineers in Egypt to remodel the surface of the fields to this type.

Further information may be found in Sir C. C. Scott-Moncrieff, Irrigation in Southern Europe (London, 1868); Moncrieff, “Lectures on Irrigation in Egypt,” Professional Papers of the Corps of Royal Engineers, vol. xix. (London, 1893); W. Willcocks, Egyptian Irrigation (2nd ed., London, 1899).

II. Water Meadows.—Nowhere in England can it be said that irrigation is necessary to ordinary agriculture, but it is occasionally employed in stimulating the growth of grass and meadow herbage in what are known as water-meadows. These are in some instances of very early origin. On the Avon in Wiltshire and the Churn in Gloucestershire they may be traced back to Roman times. This irrigation is not practised in the drought of summer, but in the coldest and wettest months of the year, the water employed being warmer than the natural moisture of the soil and proving a valuable protection against frost.

Fig. 1.—Diagram showing irrigation properly combined with drainage
(to left), and laid out regardless of drainage required later (to right).

Before the systematic conversion of a tract into water-meadows can be safely determined on, care must be taken to have good drainage, natural or artificial, a sufficient supply of water, and water of good quality. It might indeed have been thought that thorough drainage would be unnecessary, but it must be noted that porous subsoils or efficient drains do not act merely by carrying away stagnant water which would otherwise cool the earth, incrust the surface, and retard plant growth. They cause the soil to perform the office of a filter. Thus the earth and the roots of grasses absorb the useful matters not only from the water that passes over it, but from that which passes through it. These fertilizing materials are found stored up in the soil ready for the use of the roots of the plants. Stagnation of water is inimical to the action of the roots, and does away with the advantageous processes of flowing and percolating currents. Some of the best water-meadows in England have but a thin soil resting on gravel and flints, this constituting a most effectual system of natural drainage. The fall of the water supply must suffice for a fairly rapid current, say 10 in. or 1 ft. in from 100 to 200 yds. If possible the water should be taken so far above the meadows as to have sufficient fall without damming up the river. If a dam be absolutely necessary, care must be taken so to build it as to secure the fields on both sides from possible inundation; and it should be constructed substantially, for the cost of repairing accidents to a weak dam is very serious.

Even were the objects of irrigation always identical, the conditions under which it is carried on are so variable as to preclude calculations of quantity. Mere making up of necessary water in droughty seasons is one thing, protection Quantity
of water. against frost is another, while the addition of soil material is a third. Amongst causes of variation in the quantity of water needed will be its quality and temperature and rate of flow, the climate, the season, the soil, the subsoil, the artificial drainage, the slope, the aspect and the crop. In actual practice the amount of water varies from 300 gallons per acre in the hour to no less than 28,000 gallons. Where water is used, as in dry and hot countries, simply as water, less is generally needed than in cold, damp and northerly climates, where the higher temperature and the action of the water as manure are of more consequence. But it is necessary to be thoroughly assured of a good supply of water before laying out a water-meadow. Except in a few places where unusual dryness of soil and climate indicate the employment of water, even in small quantity, merely to avoid the consequences of drought, irrigation works are not to be commenced upon a large area, if only a part can ever be efficiently watered. The engineer must not decide upon the plan till he has gauged at different seasons the stream which has to supply the water, and has ascertained the rain-collecting area available, and the rainfall of the district, as well as the proportion of storable to percolating and evaporating water. Reservoirs for storage, or for equalizing the flow, are rarely resorted to in England; but they are of absolute necessity in those countries in which it is just when there is least water that it is most wanted. It is by no means an injudicious plan before laying out a system of water-meadows, which is intended to be at all extensive, to prepare a small trial plot, to aid in determining a number of questions relating to the nature and quantity of the water, the porosity of the soil, &c.

The quality of the water employed for any of the purposes of irrigation is of much importance. Its dissolved and its suspended matters must both be taken into account. Clear water is usually preferable for grass land, thick for Quality of water. arable land. If it is to be used for warping, or in any way for adding to the solid material of the irrigated land, then the nature and amount of the suspended material are necessarily of more importance than the character of the dissolved substances, provided the latter are not positively injurious. For use on ordinary water-meadows, however, not only is very clear water often found to be perfectly efficient, but water having no more than a few grains of dissolved matter per gallon answers the purposes in view satisfactorily. Water from moors and peat-bogs or from gravel or ferruginous sandstone is generally of small utility so far as plant food is concerned. River water, especially that which has received town sewage, or the drainage of highly manured land, would naturally be considered most suitable for irrigation, but excellent results are obtained also with waters which are uncontaminated with manurial matters, and which contain but 8 or 10 grains per gallon of the usual dissolved constituents of spring water. Experienced English irrigators generally commend as suitable for water-meadows those streams in which fish and waterweeds abound. But the particular plants present in or near the water-supply afford further indications of quality. Water-cress, sweet flag, flowering rush, several potamogetons, water milfoil, water ranunculus, and the reedy sweet watergrass (Glyceria aquatica) rank amongst the criteria of excellence. Less favourable signs are furnished by such plants as Arundo Donax (in Germany), Cicuta virosa and Typha latifolia, which are found in stagnant and torpid waters. Water when it has been used for irrigation generally becomes of less value for the same purpose. This occurs with clear water as well as with turbid, and obviously arises mainly from the loss of plant food which occurs when water filters through or trickles over poor soil. By passing over or through rich soil the water may, however, actually be enriched, just as clear water passed through a charcoal filter which has been long used becomes impure. It has been contended that irrigation water suffers no change in composition by use, since by evaporation of a part of the pure water the dissolved matters in the remainder would be so increased as to make up for any matters removed. But it is forgotten that both the plant and the soil enjoy special powers of selective absorption, which remove and fix the better constituents of the water and leave the less valuable.

Of the few leguminous plants which are in any degree suitable for water-meadows, Lotus corniculatus major, Trifolium hybridum, and T. pratense are those which generally flourish best; T. repens is less successful. Amongst grasses Seeds for water-meadows. the highest place must be assigned to ryegrass, especially to the Italian variety, commonly called Lolium italicum. The mixture of seeds for sowing a water-meadow demands much consideration, and must be modified according to local circumstances of soil, aspect, climate and drainage. From the peculiar use which is made of the produce of an irrigated meadow, and from the conditions to which it is subjected, it is necessary to include in our mixture of seeds some that produce an early crop, some that give an abundant growth, and some that impart sweetness and good flavour, while all the kinds sown must be capable of flourishing on irrigated soil.

The following mixtures of seeds (stated in pounds per acre) have been recommended for sowing on water-meadows, Messrs Sutton of Reading, after considerable experience, regarding No. I. as the more suitable:

	I.	II.		I.	II.
Lolium perenne	8	12⁠	Festuca pratensis	0	2
Lolium italicum	0	8	Festuca loliacea	3	2
Poa trivialis	6	3	Anthoxanthum odoratum	0	1
Glyceria fluitans	6	2	Phleum pratense	4	2
Glyceria aquatica	4	1	Phalaris arundinacea	3	2
Agrostis alba	0	1	Lotus corniculatus major	3	2
Agrostis stolonifera	6	2	Trifolium hybridum	0	1
Alopecurus pratensis	0	2	Trifolium pratense	0	1
Festuca elatior	3	2

In irrigated meadows, though in a less degree than on sewaged land, the reduction of the amount or even the actual suppression of certain species of plants is occasionally well marked. Sometimes this action is exerted upon the finer grasses, Changes in irrigated herbage. but happily also upon some of the less profitable constituents of the miscellaneous herbage. Thus Ranunculus bulbosus has been observed to become quite rare after a few years’ watering of a meadow in which it had been most abundant, R. acris rather increasing by the same treatment; Plantago media was extinguished and P. lanceolata reduced 70%. Amongst the grasses which may be spared, Aira caespitosa, Briza media and Cynosurus cristatus are generally much reduced by irrigation. Useful grasses which are increased are Lolium perenne and Alopecurus pratensis, and among those of less value Avena favescens, Dactylis glomerata and Poa pratensis.

Four ways of irrigating land with water are practised in England: (1) bedwork irrigation, which is the most efficient although it is also the most costly method by which currents of water can be applied to level land; (2) Methods. catchwork irrigation, in which the same water is caught and used repeatedly; (3) subterraneous or rather upward irrigation, in which the water in the drains is sent upwards through the soil towards the surface; and (4) warping, in which the water is allowed to stand over a level field until it has deposited the mud suspended in it.

There are two things to be attended to most carefully in the construction of a water-meadow on the first or second of these plans. First, no portion of them whatever should be on a dead level, but every part should belong to one or other of a series of true inclined planes. The second point of primary importance is the size and slope of the main conductor, which brings the water from the river to the meadow. The size of this depends upon the quantity of water required, but whatever its size its bottom at its origin should be as low as the bed of the river, in order that it may carry down as much as possible of the river mud. Its course should be as straight and as near a true inclined plane as possible. The stuff taken out of the conductor should be employed in making up its banks or correcting inequalities in the meadow.

In bedwork irrigation, which is eminently applicable to level ground, the ground is thrown into beds or ridges. Here the conductor should be led along the highest end or side of the meadow in an inclined plane; should it terminate in the Bedwork. meadow, its end should be made to taper when there are no feeders, or to terminate in a feeder. The main drain to carry off the water from the meadow should next be formed. It should be cut in the lowest part of the ground at the lower end or side of the meadow. Its dimensions should be capable of carrying off the whole water used so quickly as to prevent the least stagnation, and discharge it into the river. The next process is the forming of the ground intended for a water-meadow into beds or ridges. That portion of the ground which is to be watered by one conductor should be made into beds to suit the circumstances of that conductor; that is, instead of the beds over the meadow being all reduced to one common level, they should be formed to suit the different swells in the ground, and, should any of these swells be considerable, it will be necessary to give each side of them its respective conductor. The beds should run at or nearly at right angles to the line of the conductor. The breadth of the beds is regulated by the nature of the soil and the supply of water. Tenacious soils and subsoils, with a small supply of water, require beds as narrow as 30 ft. Porous soils and a large supply of water may have beds of 40 ft. The length of the beds is regulated by the supply of water and the fall from the conductor to the main drain. If the beds fall only in one direction longitudinally, their crowns should be made in the middle; but, should they fall laterally as well as longitudinally, as is usually the case, then the crowns should be made towards the upper sides, more or less according to the lateral slope of the ground. The crowns should rise 1 ft. above the adjoining furrows. The beds thus formed should slope in an inclined plane from the conductor to the main drain, that the water may flow equably over them.

The beds are watered by “feeders,” that is, channels gradually tapering to the lower extremities, and their crowns cut down, wherever these are placed. The depth of the feeders depends on their width, and the width on their length. A bed 200 yds. in length requires a feeder of 20 in. in width at its junction with the conductor, and it should taper gradually to the extremity, which should be 1 ft. in width. The taper retards the motion of the water, which constantly decreases by overflow as it proceeds, whilst it continues to fill the feeder to the brim. The water overflowing from the feeders down the sides of the beds is received into small drains formed in the furrows between the beds. These small drains discharge themselves into the main drain, and are in every respect the reverse of the feeders. The depth of the small drain at the junction is made about as great as that of the main drain, and it gradually lessens towards the taper to 6 in. in tenacious and to less in porous soils. The depth of the feeders is the same in relation to the conductor. For the more equal distribution of the water over the surface of the beds from the conductor and feeders, small masses, such as stones or solid portions of earth or turf fastened with pins, are placed in them, in order to retard the momentum which the water may have acquired. These “stops,” as they are termed, are generally placed at regular intervals, or rather they should be left where any inequality of the current is observed. Heaps of stones answer very well for stops in the conductor, particularly immediately below the points of junction with the feeders. The small or main drains require no stops. The descent of the water in the feeders will no doubt necessarily increase in rapidity, but the inclination of the beds and the tapering of the feeders should be so adjusted as to counteract the increasing rapidity. The distribution of the water over the whole meadow is regulated by the sluices, which should be placed at the origin of every conductor. By means of these sluices any portion of the meadow that is desired can be watered, whilst the rest remains dry; and alternate watering must be adopted when there is a scarcity of water. All the sluices should be substantially built at first with stones and mortar, to prevent the leakage of water; for, should water from a leak be permitted to find its way into the meadow, that portion of it will stagnate and produce coarse grasses. In a well-formed water-meadow it is as necessary to keep it perfectly dry at one time as it is to place it under water at another. A small sluice placed in the side of the conductor opposite to the meadow, and at the upper end of it, will drain away the leakage that may have escaped from the head sluice.

To obtain a complete water-meadow, the ground will often require to be broken up and remodelled. This will no doubt be attended with cost; but it should be considered that the first cost is the least, and remodelling the only way of having a complete water-meadow which will continue for years to give satisfaction. To effect a remodelling when the ground is in stubble, let it be ploughed up, harrowed, and cleaned as in a summer fallow, the levelling-box employed when required, the stuff from the conductors and main drains spread abroad, and the beds ploughed into shape—all operations that can be performed at little expense. The meadow should be ready by August for sowing with one of the mixtures of grass-seeds already given. But though this plan is ultimately better, it is attended with the one great disadvantage that the soft ground cannot be irrigated for two or three years after it is sown with grass-seeds. This can only be avoided where the ground is covered with old turf which will bear to be lifted. On ground in that state a water-meadow may be most perfectly formed. Let the turf be taken off with the spade, and laid carefully aside for relaying. Let the stript ground then be neatly formed with the spade and barrow, into beds varying in breadth and shape according to the nature of the soil and the dip of the ground—the feeders from the conductor and the small drains to the main drain being formed at the same time. Then let the turf be laid down again and beaten firm, when the meadow will be complete at once, and ready for irrigation. This is the most beautiful and most expeditious method of making a complete water-meadow where the ground is not naturally sufficiently level to begin with.

The water should be let on, and trial made of the work, whenever it is finished, and the motion of the water regulated by the introduction of a stop in the conductors and feeders where a change in the motion of the current is observed, beginning at the upper end of the meadow. Should the work be finished as directed by August, a good crop of hay may be reaped in the succeeding summer. There are few pieces of land where the natural descent of the ground will not admit of the water being collected a second time, and applied to the irrigation of a second and lower meadow. In such a case the main drain of a watered meadow may form the conductor of the one to be watered, or a new conductor may be formed by a prolongation of the main drain; but either expedient is only advisable where water is scarce. Where it is plentiful, it is better to supply the second meadow directly from the river, or by a continuation of the first main conductor.

In the ordinary catch work water-meadow, the water is used over and over again. On the steep sides of valleys the plan is easily and cheaply carried out, and where the whole course of the water is not long the peculiar properties which give it Catchwork. value, though lessened, are not exhausted when it reaches that part of the meadow which it irrigates last. The design of any piece of catchwork will vary with local conditions, but generally it may be stated that it consists in putting each conduit save the first to the double use of a feeder or distributor and of a drain or collector.

In upward or subterranean irrigation the water used rises upward through the soil, and is that which under ordinary circumstances would be carried off by the drains. The system has received considerable development in Germany, where the Upward or subter-ranean. elaborate method invented by Petersen is recommended by many agricultural authorities. In this system the well-fitting earthenware drain-pipes are furnished at intervals with vertical shafts terminating at the surface of the ground in movable caps. Beneath each cap, and near the upper end of the shaft, are a number of vertical slits through which the drainage water which rises passes out into the conduit or trench from which the irrigating streams originate. In the vertical shaft there is first of all a grating which intercepts solid matters, and then, lower down, a central valve which can be opened and closed at pleasure from the top of the shaft. In the ordinary English system of upward or drainage irrigation, ditches are dug all round the field. They act the part of conductors when the land is to be flooded, and of main drains when it is to be laid dry. The water flows from the ditches as conductors into built conduits formed at right angles to them in parallel lines through the fields; it rises upwards in them as high as the surface of the ground, and again subsides through the soil and the conduits into the ditches as main drains, and thence it passes at a lower level either into a stream or other suitable outfall. The ditches may be filled in one or other of several different ways. The water may be drainage-water from lands at a higher level; or it may be water from a neighbouring river; or it may be drainage-water accumulated from a farm and pumped up to the necessary level. But it may also be the drainage-water of the field itself. In this case the mouths of the underground main pipe-drains are stopped up, and the water in them and the secondary drains thus caused to stand back until it has risen sufficiently near the surface. Of course it is necessary to build the mouths of such main drains of very solid masonry, and to construct efficient sluices for the retention of the water in the drains. Irrigation of the kind now under discussion may be practised wherever a command of water can be secured, but the ground must be level. It has been successfully employed in recently drained morasses, which are apt to become too dry in summer. It is suitable for stiffish soils where the subsoil is fairly open, but is less successful in sand. The water used may be turbid or clear, and it acts, not only for moistening the soil, but as manure. For if, as is commonly the case, the water employed be drainage-water from cultivated lands, it is sure to contain a considerable quantity of nitrates, which, not being subject to retention by the soil, would otherwise escape. These coming into contact with the roots of plants during their season of active growth, are utilized as direct nourishment for the vegetation. It is necessary in upward or subterranean irrigation to send the water on and to take it off very gently, in order to avoid the displacement and loss of the finer particles of the soil which a forcible current would cause.

In warping the suspended solid matters are of importance, not merely for any value they may have as manure, but also as a material addition to the ground to be irrigated. The warping which is practised in England is almost exclusively confined to the overflowing of level ground within tide mark, and is conducted Warping. mostly within the districts commanded by estuaries or tidal rivers. The best notion of the process of warping may be gained by sailing up the Trent from the Humber to Gainsborough. Here the banks of the river were constructed centuries ago to protect the land within them from the encroachments of the tide. A great tract of country was thus laid comparatively dry. But while the wisdom of one age thus succeeded in restricting within bounds the tidal water of the river, it was left to the greater wisdom of a succeeding age to improve upon this arrangement by admitting these muddy waters to lay a fresh coat of rich silt on the exhausted soils. The process began more than a century ago, but has become a system in recent times. Large sluices of stone, with strong doors, to be shut when it is wished to exclude the tide, may be seen on both banks of the river, and from these great conduits are carried miles inward through the flat country to the point previously prepared by embankment over which the muddy waters are allowed to spread. These main conduits, being very costly, are constructed for the warping of large adjoining districts, and openings are made at such points as are then undergoing the operation. The mud is deposited and the waters return with the falling tide to the bed of the river. Spring-tides are preferred, and so great is the quantity of mud in these rivers that from 10 to 15 acres have been known to be covered with silt from 1 to 3 ft. in thickness during one spring of ten or twelve tides. Peat-moss of the most sterile character has been by this process covered with soil of the greatest fertility, and swamps which used to be resorted to for leeches are now, by the effects of warping, converted into firm and fertile fields. The art is now so well understood that, by careful attention to the currents, the expert warp farmer can temper his soil as he pleases. When the tide is first admitted the heavier particles, which are pure sand, are first deposited; the second deposit is a mixture of sand and fine mud, which, from its friable texture, forms the most valuable soil; while lastly the pure mud subsides, containing the finest particles of all, and forms a rich but very tenacious soil. The great effort, therefore, of the warp farmer is to get the second or mixed deposit as equally over the whole surface as he can and to prevent the deposit of the last. This he does by keeping the water in constant motion, as the last deposit can only take place when the water is suffered to be still. Three years may be said to be spent in the process, one year warping, one year drying and consolidating, and one year growing the first crop, which is generally seed-hoed in by hand, as the mud at this time is too soft to admit of horse labour.

The immediate effect, which is highly beneficial, is the deposition of silt from the tide. To ensure this deposition, it is necessary to surround the field to be warped with a strong embankment, in order to retain the water as the tide recedes. The water is admitted by valved sluices, which open as the tide flows into the field and shut by the pressure of the confined water when the tide recedes. These sluices are placed on as low a level as possible to permit the most turbid water at the bottom of the tide to pass through a channel in the base of the embankment. The silt deposited after warping is exceedingly rich and capable of carrying any species of crop. It may be admitted in so small a quantity as only to act as a manure to arable soil, or in such a large quantity as to form a new soil. This latter acquisition is the principal object of warping, and it excites astonishment to witness how soon a new soil may be formed. From June to September a soil of 3 ft. in depth may be formed under the favourable circumstances of a very dry season and long drought. In winter and in floods warping ceases to be beneficial. In ordinary circumstances on the Trent and Humber a soil from 6 to 16 in. in depth may be obtained and inequalities of 3 ft. filled up. But every tide generally leaves only 1/8 in. of silt, and the field which has only one sluice can only be warped every other tide. The silt, as deposited in each tide, does not mix into a uniform mass, but remains in distinct layers. The water should be made to run completely off and the ditches should become dry before the influx of the next tide, otherwise the silt will not incrust and the tide not have the same effect. Warp soil is of surpassing fertility. The expense of forming canals, embankments and sluices for warping land is from £10 to £20 an acre. A sluice of 6 ft. in height and 8 ft. wide will warp from 60 to 80 acres, according to the distance of the field from the river. The embankments may be from 3 to 7 ft. in height, as the field may stand in regard to the level of the highest tides. After the new land has been left for a year or two in seeds and clover, it produces great crops of wheat and potatoes.

Warping is practised only in Lincolnshire and Yorkshire, on the estuary of the Humber, and in the neighbourhood of the rivers which flow into it—the Trent, the Ouse and the Don. The silt and mud brought down by these rivers is rich in clay and organic matter, and sometimes when dry contains as much as 1% of nitrogen.

Constant care is required if a water-meadow is to yield quite satisfactory results. The earliness of the feed, its quantity and its quality will all depend in very great measure upon the proper management of the irrigation. The points which require constant attention are—the Management and advantages. perfect freedom of all carriers, feeders and drains from every kind of obstruction, however minute; the state and amount of water in the river or stream, whether it be sufficient to irrigate the whole area properly or only a part of it; the length of time the water should be allowed to remain on the meadow at different periods of the season; the regulation of the depth of the water, its quantity and its rate of flow, in accordance with the temperature and the condition of the herbage; the proper times for the commencing and ending of pasturing and of shutting up for hay; the mechanical condition of the surface of the ground; the cutting out of any very large and coarse plants, as docks; and the improvement of the physical and chemical conditions of the soil by additions to it of sand, silt, loam, chalk, &c.

Whatever may be the command of water, it is unwise to attempt to irrigate too large a surface at once. Even with a river supply fairly constant in level and always abundant, no attempt should be made to force on a larger volume of water than the feeders can properly distribute and the drains adequately remove, or one part of the meadow will be deluged and another stinted. When this inequality of irrigation once occurs, it is likely to increase from the consequent derangement of the feeders and drains. And one result on the herbage will be an irregularity of composition and growth, seriously detrimental to its food-value. The adjustment of the water by means of the sluices is a delicate operation when there is little water and also when there is much; in the latter case the fine earth may be washed away from some parts of the meadow; in the former case, by attempting too much with a limited water current, one may permit the languid streams to deposit their valuable suspended matters instead of carrying them forward to enrich the soil. The water is not to be allowed to remain too long on the ground at a time. The soil must get dry at stated intervals in order that the atmospheric air may come in contact with it and penetrate it. In this way as the water sinks down through the porous subsoil or into the subterranean drains oxygen enters and supplies an element which is needed, not only for the oxidation of organic matters in the earth, but also for the direct and indirect nutrition of the roots. Without this occasional drying of the soil the finer grasses and the leguminous plants will infallibly be lost; while a scum of confervae and other algae will collect upon the surface and choke the higher forms of vegetation. The water should be run off thoroughly, for a little stagnant water lying in places upon the surface does much injury. The practice of irrigating differs in different places with differences in the quality of the water, the soil, the drainage, &c. As a general rule, when the irrigating season begins in November the water may flow for a fortnight continuously, but subsequent waterings, especially after December, should be shortened gradually in duration till the first week in April, when irrigation should cease. It is necessary to be very careful in irrigating during frosty weather. For, though grass will grow even under ice, yet if ice be formed under and around the roots of the grasses the plants may be thrown out by the expansion of the water at the moment of its conversion into ice. The water should be let off on the morning of a dry day, and thus the land will be dry enough at night not to suffer from the frost; or the water may be taken off in the morning and let on again at night. In spring the newly grown and tender grass will be easily destroyed by frost if it be not protected by water, or if the ground be not made thoroughly dry.

Although in many cases it is easy to explain the reasons why water artificially applied to land brings crops or increases their yield, the theory of our ordinary water-meadow irrigation is rather obscure. For we are not dealing in these grass lands with a Theory.semi-aquatic plant like rice, nor are we supplying any lack of water in the soil, nor are we restoring the moisture which the earth cannot retain under a burning sun. We irrigate chiefly in the colder and wetter half of the year, and we “saturate” with water the soil in which are growing such plants as are perfectly content with earth not containing more than one-fifth of its weight of moisture. We must look in fact to a number of small advantages and not to any one striking beneficial process in explaining the aggregate utility of water-meadow irrigation. We attribute the usefulness of water-meadow irrigation, then, to the following causes: (1) the temperature of the water being rarely less than 10° Fahr. above freezing, the severity of frosts in winter is thus obviated, and the growth, especially of the roots of grasses, is encouraged; (2) nourishment or plant food is actually brought on to the soil, by which it is absorbed and retained, both for the immediate and for the future use of the vegetation, which also itself obtains some nutrient material directly; (3) solution and redistribution of the plant food already present in the soil occur mainly through the solvent action of the carbonic acid gas present in a dissolved state in the irrigation-water; (4) oxidation of any excess of organic matter in the soil, with consequent production of useful carbonic acid and nitrogen compounds, takes place through the dissolved oxygen in the water sent on and through the soil where the drainage is good; and (5) improvement of the grasses, and especially of the miscellaneous herbage, of the meadow is promoted through the encouragement of some at least of the better species and the extinction or reduction of mosses and of the innutritious weeds.

To the united agency of the above-named causes may safely be attributed the benefits arising from the special form of water-irrigation which is practised in England. Should it be thought that the traces of the more valuable sorts of plant food (such as compounds of nitrogen, phosphates, and potash salts) existing in ordinary brook or river water can never bring an appreciable amount of manurial matter to the soil, or exert an appreciable effect upon the vegetation, yet the quantity of water used during the season must be taken into account. If but 3000 gallons hourly trickle over and through an acre, and if we assume each gallon to contain no more than one-tenth of a grain of plant food of the three sorts just named taken together, still the total, during a season including ninety days of actual irrigation, will not be less than 9 ℔ per acre. It appears, however, that a very large share of the benefits of water-irrigation is attributable to the mere contact of abundance of moving water, of an even temperature, with the roots of the grass. The growth is less checked by early frosts; and whatever advantages to the vegetation may accrue by occasional excessive warmth in the atmosphere in the early months of the year are experienced more by the irrigated than by the ordinary meadow grasses by reason of the abundant development of roots which the water has encouraged.

III. Italian Irrigation.—The most highly developed irrigation in the world is probably that practised in the plains of Piedmont and Lombardy, where every variety of condition is to be found. The engineering works are of a very high class, and from long generations of experience the farmer knows how best to use his water. The principal river of northern Italy is the Po, which rises to the west of Piedmont and is fed not from glaciers like the Swiss torrents, but by rain and snow, so that the water has a somewhat higher temperature, a point to which much importance is attached for the valuable meadow irrigation known as marcite. This is only practised in winter when there is abundance of water available, and it much resembles the water-meadow irrigation of England. The great Cavour canal is drawn from the left bank of the Po a few miles below Turin, and it is carried right across the drainage of the country. Its full discharge is 3800 cub. ft. per second, but it is only from October to May, when the water is least required, that it carries anything like this amount. For the summer irrigation Italy depends on the glaciers of the Alps; and the great torrents of the Dora Baltea and Sesia can be counted on for a volume exceeding 6000 cub. ft. per second. Lombardy is quite as well off as Piedmont for the means of irrigation and, as already said, its canals have the advantage that being drawn from the lakes Maggiore and Como they exercise a moderating influence on the Ticino and Adda rivers, which is much wanted in the Dora Baltea. The Naviglio Grande of Lombardy is a very fine work drawn from the left bank of the Ticino and useful for navigation as well as irrigation. It discharges between 3000 and 4000 cub. ft. per second, and probably nowhere is irrigation carried on with less expense. Another canal, the Villoresi, drawn from the same bank of the Ticino farther upstream, is capable of carrying 6700 cub. ft. per second. Like the Cavour canal, the Villoresi is taken across the drainage of the country, entailing a number of very bold and costly works.

Interesting as these Italian works are, the administration and distribution of the water is hardly less so. The system is due to the ability of the great Count Cavour; what he originated in Piedmont has been also carried out in Lombardy. The Piedmontese company takes over from the government the control of all the irrigation within a triangle between the left bank of the Po and the right bank of the Sesia. It purchases from government about 1250 cub. ft. per second, and has also obtained the control of all private canals. Altogether it distributes about 2275 cub. ft. of water and irrigates about 141,000 acres, on which rice is the most important crop. The association has 14,000 members and controls nearly 10,000 m. of distributary channels. In each parish is a council composed of all landowners who irrigate. Each council sends two deputies to what may be called a water parliament. This assembly elects three small committees, and with them rests the whole management of the irrigation. An appeal may be made to the civil courts from the decision of these committees, but so popular are they that such appeals are never made. The irrigated area is divided into districts, in each of which is an overseer and a staff of watchmen to see to the opening and shutting of the modules (see Hydraulics, §§ 54 to 56) which deliver the water into the minor channels. In the November of each year it is decided how much water is to be given to each parish in the year following, and this depends largely on the number of acres of each crop proposed to be watered. In Lombardy the irrigation is conducted on similar principles. Throughout, the Italian farmer sets a very high example in the loyal way he submits to regulations which there must be sometimes a strong temptation to break. A sluice surreptitiously opened during a dark night and allowed to run for six hours may quite possibly double the value of his crop, but apparently the law is not often broken.

IV. Egypt.—The very life of Egypt depends on its irrigation, and, ancient as this irrigation is, it was never practised on a really scientific system till after the British occupation. As every one knows, the valley of the Nile outside of Character-istics of the Nile Valley and flood. the tropics is practically devoid of rainfall. Yet it was the produce of this valley that formed the chief granary of the Roman Empire. Probably nowhere in the world is there so large a population per square mile depending solely on the produce of the soil. Probably nowhere is there an agricultural population so prosperous, and so free from the risks attending seasons of drought or of flood. This wealth and prosperity are due to two very remarkable properties of the Nile. First, the regimen of the river is nearly constant. The season of its rise and its fall, and the height attained by its waters during the highest flood and at lowest Nile vary to a comparatively small extent. Year after year the Nile rises at the same period, it attains its maximum in September and begins to diminish first rapidly till about the end of December, and then more slowly and more steadily until the following June. A late rise is not more than about three weeks behind an early rise. From the lowest to the highest gauge of water-surface the rise is on an average 25.5 ft. at the First Cataract. The highest flood is 3.5 ft. above this average, and this means peril, if not disaster, in Lower Egypt. The lowest flood on record has risen only to 5.5 ft. below the average, or to 20 ft. above the mean water-surface of low Nile. Such a feeble Nile flood has occurred only four times in modern history: in 1877, when it caused widespread famine and death throughout Upper Egypt, 947,000 acres remained barren, and the land revenue lost £1,112,000; in 1899 and again in 1902 and 1907, when by the thorough remodelling of the whole system of canals since 1883 all famine and disaster were avoided and the loss of revenue was comparatively slight. In 1907, for instance, when the flood was nearly as low as in 1877, the area left unwatered was little more than 10% of the area affected in 1877.

This regularity of flow is the first exceptional excellence of the river Nile. The second is hardly less valuable, and consists in the remarkable richness of the alluvium brought down the river year after year during the flood. The object of the engineer is so to utilize this flood-water that as little as possible of the alluvium may escape into the sea, and as much as possible may be deposited on the fields. It is the possession of these two properties that imparts to the Nile a value quite unique among rivers, and gives to the farmers of the Nile Valley advantages over those of any rain-watered land in the world.

Until the 19th century irrigation in Egypt on a large scale was practised merely during the Nile flood. Along each edge of the river and following its course has been erected an earthen embankment high enough not to be Irrigation during
high Nile. topped by the highest floods. In Upper Egypt, the valley of which rarely exceeds 6 m. in width, a series of cross embankments have been constructed, abutting at the inner ends on those along the Nile, and at the outer ends on the ascending sides of the valley. The whole country has thus been divided into a series of oblongs, surrounded by embankments on three sides and by the desert slopes on the fourth. These oblong areas vary from 60,000 to 1500 or 2000 acres in extent. Throughout all Egypt the Nile is deltaic in character; that is, the slope of the country in the valley is away from the river and not towards it. It is easy, then, when the Nile is low, to cut short, deep canals in the river banks, which fill as the flood rises, and carry the precious mud-charged water into these great flats. There the water remains for a month or more, some 3 ft. deep, depositing its mud, and thence at the end of the flood the almost clear water may either be run off directly into the receding river, or cuts may be made in the cross embankments, and it may be allowed to flow from one flat to another and ultimately into the river. In November the waters have passed off; and whenever a man can walk over the mud with a pair of bullocks, it is roughly turned over with a wooden plough, or merely the branch of a tree, and the wheat or barley crop is immediately sown. So soaked is the soil after the flood, that the grain germinates, sprouts, and ripens in April, without a shower of rain or any other watering.

In Lower Egypt this system was somewhat modified, but it was the same in principle. No other was known in the Nile Valley until the country fell, early in the 19th century, under the vigorous rule of Mehemet Ali Pasha. He soon recognized that with such a climate and soil, with a teeming population, and with the markets of Europe so near they might produce in Egypt something more profitable than wheat and maize. Cotton and sugar-cane would fetch far higher prices, but they could only be grown while the Nile was low, and they required water at all seasons.

It has already been said that the rise of the Nile is about 251/2 ft., so that a canal constructed to draw water out of the river while at its lowest must be 251/2 ft. deeper than if it is intended to draw off only during the highest Irrigation during
low Nile. floods. Mehemet Ali began by deepening the canals of Lower Egypt by this amount, a gigantic and futile task; for as they had been laid out on no scientific principles, the deep channels became filled with mud during the first flood, and all the excavation had to be done over again, year after year. With a serf population even this was not impossible; but as the beds of the canals were graded to no even slope, it did not follow that if water entered the head it would flow evenly on. As the river daily fell, of course the water in the canals fell too, and since they were never dug deep enough to draw water from the very bottom of the river, they occasionally ran dry altogether in the month of June, when the river was at its lowest, and when, being the month of greatest heat, water was more than ever necessary for the cotton crop. Thus large tracts which had been sown, irrigated, weeded and nurtured for perhaps three months perished in the fourth, while all the time the precious Nile water was flowing useless to the sea. The obvious remedy was to throw a weir across each branch of the river to control the water and force it into canals taken from above it. The task of constructing this great work was committed to Mougel Bey, a French engineer of ability, who designed and The Nile Barrage. constructed the great barrage across the two branches of the Nile at the apex of the delta, about 12 m. north of Cairo (fig. 2). It was built to consist of two bridges—one over the eastern or Damietta branch of the river having 71 arches, the other, over the Rosetta branch, having 61 arches, each arch being of 5 metres or 16.4 ft. span. The building was all of stone, the floors of the arches were inverts. The height of pier from edge of flooring to spring of arch was 28.7 ft., the spring of the arch being about the surface-level of maximum flood. The arches were designed to be fitted with self-acting drop gates; but they were not a success, and were only put into place on the Rosetta branch. The gates were intended to hold up the water 4.5 metres, or 14.76 ft., and to divert it into three main canals—the Behera on the west, the Menufia in the centre and the Tewfikia on the east. The river was thus to be emptied, and to flow through a whole network of canals, watering all Lower Egypt. Each barrage was provided with locks to pass Nile boats 160 by 28 ft. in area.

Fig. 2.—Map showing the Damietta and Rosetta dams on the Nile.

Mougel’s barrage, as it may now be seen, is a very imposing and stately work. Considering his want of experience of such rivers as the Nile, and the great difficulties he had to contend with under a succession of ignorant Turkish rulers, it would be unfair to blame him because, until it fell into the hands of British engineers in 1884, the work was condemned as a hopeless failure. It took long years to complete, at a cost which can never be estimated, since much of it was done by serf labour. In 1861 it was at length said to be finished; but it was not until 1863 that the gates of the Rosetta branch were closed, and they were reopened again immediately, as a settlement of the masonry took place. The experiment was repeated year after year till 1867, when the barrage cracked right across from foundation to top. A massive coffer-dam was then erected, covering the eleven arches nearest the crack; but the work was never trusted again, nor the water-surface raised more than about 3 ft.

An essential part of the barrage project was the three canals, taking their water from just above it, as shown in fig. 2. The heads of the existing old canals, taken out of the river at intervals throughout the delta, were to be closed, and the canals themselves all put into connexion with the three high-level trunk lines taken from above the barrage. The central canal, or Menufia, was more or less finished, and, although full of defects, has done good service. The eastern canal was never dug at all until the British occupation. The western, or Behera, canal was dug, but within its first 50 m. it passes through desert, and sand drifted into it. Corvées of 20,000 men used to be forced to clear it out year after year, but at last it was abandoned. Thus the whole system broke down, the barrage was pronounced a failure, and attention was turned to watering Lower Egypt by a system of gigantic pumps, to raise the water from the river and discharge it into a system of shallow surface-canals, at an annual cost of about £250,000, while the cost of the pumps was estimated at £700,000. Negotiations were on foot for carrying out this system when the British engineers arrived in Egypt. They soon resolved that it would be very much better if the original scheme of using the barrage could be carried out, and after a careful examination of the work they were satisfied that this could be done. The barrage rests entirely on the alluvial bed of the Nile. Nothing more solid than strata of sand and mud is to be found for more than 200 ft. below the river. It was out of the question, therefore, to think of founding on solid material, and yet it was desired to have a head of water of 13 or 14 ft. upon the work. Of course, with such a pressure as this, there was likely to be percolation under the foundations and a washing-out of the soil. It had to be considered whether this percolation could best be checked by laying a solid wall across the river, going down to 50 or 60 ft. below its bed, or by spreading out the foundations above and below the bridge, so as to form one broad water-tight flooring—a system practised with eminent success by Sir Arthur Cotton in Southern India. It was decided to adopt the latter system. As originally designed, the flooring of the barrage from up-stream to downstream face was 111.50 ft. wide, the distance which had to be travelled by water percolating under the foundations. This width of flooring was doubled to 223 ft., and along the upstream face a line of sheet piling was driven 16 ft. deep. Over the old flooring was superposed 15 in. of the best rubble masonry, an ashlar floor of blocks of close-grained trachyte being laid directly under the bridge, where the action was severest. The working season lasted only from the end of November to the end of June, while the Nile was low; and the difficulty of getting in the foundations was increased, as, in the interests of irrigation and to supply the Menufia canal, water was held up every season while the work was in progress to as much as 10 ft. The work was begun in 1886, and completed in June 1890. Moreover, in the meantime the eastern, or Tewfikia, canal was dug and supplied with the necessary masonry works for a distance of 23 m., to where it fed the network of old canals. The western, or Behera, canal was thoroughly cleared out and remodelled; and thus the whole delta irrigation was supplied from above the barrage.

The outlay on the barrage between 1883 and 1891 amounted to about £460,000. The average cotton crop for the 5 years preceding 1884 amounted to 123,000 tons, for the 5 years ending 1898 it amounted to 251,200 tons. At the low rate of £40 per ton, this means an annual increase to the wealth of Lower Egypt of £5,128,000. Since 1890 the barrage has done its duty without accident, but a work of such vast importance to Lower Egypt required to be placed beyond all risk. It having been found that considerable hollow spaces existed below the foundations of some of the piers, five bore-holes from the top of the roadway were pierced vertically through each pier of both barrages, and similar holes were drilled at intervals along all the lock walls. Down these holes cement grout was injected under high pressure on the system of Mr Kinipple. The work was successfully carried out during the seasons 1896 to 1898. During the summer of 1898 the Rosetta barrage was worked under a pressure of 14 ft. But this was looked on as too near the limit of safety to be relied on, and in 1899 subsidiary weirs were started across both branches of the river a short distance below the two barrages. These were estimated to cost £530,000 altogether, and were to stand 10.8 ft. above the river’s bed, allowing the water-surface up-stream of the barrage to be raised 7.2 ft., while the pressure on that work itself would not exceed 10 ft. These weirs were satisfactorily completed in 1901.

The barrage is the greatest, but by no means the only important masonry work in Lower Egypt. Numerous regulating bridges and locks have been built to give absolute control of the water and facilities for navigation; and since 1901 a second weir has been constructed opposite Zifta, across the Damietta branch of the Nile, to improve the irrigation of the Dakhilia province.

In the earlier section of this article it is explained how necessary it is that irrigation should always be accompanied by drainage. This had been totally neglected in Egypt; but very large sums have been spent on it, and the country is now covered with a network of drains nearly as complete as that of the canals.

The ancient system of basin irrigation is still pursued in Upper Egypt, though by the end of 1907 over 320,000 feddans of land formerly under basin irrigation had been given, at a cost of over £E3,000,000, perennial irrigation. This conversion work was carried out in the Basin irrigation
of Upper Egypt.provinces situated between Cairo and Assiut, a region sometimes designated Middle Egypt. The ancient system seems simple enough; but in order really to flood the whole Nile Valley during seasons of defective as well as favourable floods, a system of regulating sluices, culverts and syphons is necessary; and for want of such a system it was found, in the feeble flood of 1888, that there was an area of 260,000 acres over which the water never flowed. This cost a loss of land revenue of about £300,000, while the loss of the whole season’s crop to the farmer was of course much greater. The attention of the British engineers was then called to this serious calamity; and fortunately for Egypt there was serving in the country Col. J. C. Ross, R.E., an officer who had devoted many years of hard work to the irrigation of the North-West Provinces of India, and who possessed quite a special knowledge as well as a glowing enthusiasm for the subject. Fortunately, too, it was possible to supply him with the necessary funds to complete and remodel the canal system. When the surface-water of a river is higher than the fields right and left, there is nothing easier than to breach the embankments and flood the fields—in fact, it may be more difficult to prevent their being flooded than to flood them—but in ordinary floods the Nile is never higher than all the bordering lands, and in years of feeble flood it is higher than none of them. To water the valley, therefore, it is necessary to construct canals having bed-slopes less than that of the river, along which the water flows until its surface is higher than that of the fields. If, for instance, the slope of the river be 4 in. per mile, and that of the canal 2 in. it is evident that at the end of a mile the water in the canal will be 2 in. higher than in the river; and if the surface of the land is 3 ft. higher than that of the river, the canal, gaining on it at 2 in. per mile, will reach the surface in 18 m., and from thence onwards will be above the adjoining fields. But to irrigate this upper 18 m., water must either be raised artificially, or supplied from another canal taking its source 18 m. farther up. This would, however, involve the country in great lengths of canal between the river and the field, and circumstances are not so unfavourable as this. Owing to the deltaic nature of the Nile Valley, the fields on the banks are 3 ft. above the flood, at 2 m. away from the banks they may not be more than 1 ft. above that level, so that the canal, gaining 2 in. per mile and receding from the river, will command the country in 6 m. The slope of the river, moreover, is taken in its winding course; and if it is 4 in. per mile, the slope of the axis of the valley parallel to which the canals may be made to flow is at least 6 in. per mile, so that a canal with a slope of 2 in. gains 4 in. per mile.

The system of having one canal overlapping another has one difficulty to contend with. Occasionally the desert cliffs and slopes come right down to the river, and it is difficult, if not impossible, to carry the higher-level canals past these obstructions. It should also be noticed that on the higher strip bordering the river it is the custom to take advantage of its nearness to raise water by pumps, or other machinery, and thereby to grow valuable crops of sugar-cane, maize or vegetables. When the river rises, these crops, which often form a very important part of the year’s produce and are termed Nabári, are still in the ground, and they require water in moderate and regulated quantities, in contradistinction to the wholesale flooding of the flats beyond. Fig. 3 will serve to explain this system of irrigation, the firm lines representing canals, the dotted lines embankments. It will be seen, beginning on the east or right bank of the river, that a high-level canal from an upper system is carried past a steep slope, where perhaps it is cut entirely out of rock, and it divides into two. The right branch waters all the desert slopes within its reach and level. The left branch passes, by a syphon aqueduct, under what is the main canal of the system, taken from the river close at hand (and therefore at a lower level). This left branch irrigates the Nabári on the high lands bordering the river. In years of very favourable flood this high-level canal would not be wanted at all; the irrigation could be done from the main canal, and with this great advantage, that the main canal water would carry with it much more fertilizing matter than would be got from the tail of the high-level canal, which left the river perhaps 25 m. up. The main canal flows freely over the flats C and D, and, if the flood is good, over B and part of A. It is carried round the next desert point, and to the north becomes the high-level canal. The masonry works required for this system are a syphon to pass the high level under the main canal near its head, bridges fitted with sluices where each canal passes under an embankment, and an escape weir at the tail of the system, just south of the desert point, to return surplus water to the river. Turning to the left bank, there is the same high-level canal from the upper system irrigating the basins K, P and L, as well as the large basin E in such years as it cannot be irrigated from the main canal. Here there are two main canals—one following the river, irrigating a series of smaller basins, and throwing out a branch to its left, the other passing under the desert slopes and supplying the basins F, G, H and S. For this system two syphons will be required near the head, regulating bridges under all the embankments, and an escape weir back into the river.

Fig. 3.—Map of the Basin System of Irrigation.

In the years following 1888 about 100 new masonry works of this kind were built in Upper Egypt, nearly 400 m. of new canal were dug, and nearly 300 m. of old canal were enlarged and deepened. The result has been, as already stated, that with a complete failure of the Nile flood the loss to the country has been trifling compared with that of 1877.

The first exception in Upper Egypt to the basin system of irrigation was due to the Khedive Ismail. The khedive, having acquired vast estates in the provinces of Assiut, Miniah, Beni-Suef and the Fayúm, resolved to grow sugar-cane on a very large scale, and with this object constructed a very important perennial canal, named the Ibrahimia, taking out of the left bank of the Nile at the town of Assiut, and flowing parallel to the river for about 200 m., with an important branch which irrigates the Fayúm. This canal was badly constructed, and by entirely blocking the drainage of the valley did a great deal of harm to the lands. Most of its defects had been remedied, but one remained. There being at its head no weir across the Nile, the water in the Ibrahimia canal used to rise and fall with that of the river, and so the supply was apt to run short during the hottest months, as was the case with the canals of Lower Egypt before the barrage was built. To supply the Ibrahimia canal at all during low Nile, it had been necessary to carry on dredging operations at an annual cost of about £12,000. This has now been rectified, in the same way as in Lower Egypt, by the Assiut Weir and Esna Barrage. construction of a weir across the Nile, intended to give complete control over the river and to raise the water-surface 8.2 ft. The Assiut weir is constructed on a design very similar to that of the barrage in Lower Egypt. It consists of a bridge of 111 arches, each 5 metres span, with piers of 2 metres thickness. In each arch are fitted two gates. There is a lock 80 metres long and 16 metres wide at the left or western end of the weir, and adjoining it are the regulating sluices of the Ibrahimia canal. The Assiut weir across the Nile is just about half a mile long. The work was begun at the end of 1898 and finished early in 1902—in time to avert over a large area the disastrous effects which would otherwise have resulted from the low Nile of that year. The money value of the crops saved by the closing of the weir was not less than £E690,000. The conversion of the lands north of Assiut from basin to perennial irrigation began immediately after the completion of the Assiut weir and was finished by the end of 1908. To render the basin lands of the Kena province independent of the flood being bad or good, another barrage was built across the Nile at Esna at a cost of £1,000,000. This work was begun in 1906 and completed in 1909.

These works, as well as that in Lower Egypt, are intended to raise the water-surface above it, and to control the distribution of its supply, but in no way to store that supply. The idea of ponding up the superfluous flood discharge of the river is not a new one, and if Herodotus is to be believed, Storage. it was a system actually pursued at a very early period of Egyptian history, when Lake Moeris in the Fayúm was filled at each Nile flood, and drawn upon as the river ran down. When British engineers first undertook the management of Egyptian irrigation many representations were made to them of the advantage of storing the Nile water; but they consistently maintained that before entering on that subject it was their duty to utilize every drop of the water at their disposal. This seemed all the more evident, as at that time financial reasons made the construction of a costly Nile dam out of the question. Every year, however, between 1890 and 1902 the supply of the Nile during May and June was actually exhausted, no water at all flowing then out into the sea. In these years, too, owing to the extension of drainage works, the irrigable area of Egypt was greatly enlarged, so that if perennial cultivation was at all to be increased, it was necessary to increase the volume of the river, and this could only be done by storing up the flood supply. The first difficulty that presented itself in carrying this out, was that during the months of highest flood the Nile is so charged with alluvial matter that to pond it up then would inevitably lead to a deposit of silt in the reservoir, which would in no great number of years fill it up. It was found, however, that the flood water was comparatively free from deposit by the middle of November, while the river was still so high that, without injuring the irrigation, water might go on being stored up until March. Accordingly, when it was determined to construct a dam, it was decided that it should be supplied with sluices large enough to discharge unchecked the whole volume of the river as it comes down until the middle of November, and then to begin the storage.

The site selected for the great Nile dam was at the head of the First Cataract above Assuan. A dyke of syenite granite here crosses the valley, so hard that the river had nowhere scoured a deep channel through it, and so it was found possible to construct the dam entirely in the open air, without the necessity of laying under-water foundations. The length of the dam is about 6400 ft.—nearly 11/4 m. The greatest head of water The
Assuan
Dam. in it is 65 ft. It is pierced by 140 under-sluices of 150 sq. ft. each, and by 40 upper-sluices, each of 75 sq. ft. These, when fully open, are capable of discharging the ordinary maximum Nile flood of 350,000 cub. ft. per second, with a velocity of 15.6 ft. per second and a head of 6.6 ft. The top width of the dam is 23 ft., the bottom width at the deepest part about 82 ft. On the left flank of the dam there is a canal, provided with four locks, each 262 by 31 ft. in area, so that navigation is possible at all seasons. The storage capacity of the reservoir is about 3,750,000 millions of cub. ft., which creates a lake extending up the Nile Valley for about 200 m. The reservoir is filled yearly by March; after that the volume reaching the reservoir from the south is passed on through the sluices. In May, or earlier when the river is late in rising, when the demand for water increases, first the upper and then the under sluices are gradually opened, so as to increase the river supply, until July, when all the gates are open, to allow of the free passage of the flood. On the 10th of December 1902 this magnificent work was completed. The engineer who designed it was Sir W. Willcocks. The contractors were Messrs John Aird & Co., the contract price being £2,000,000. The financial treaties in which the Egyptian government were bound up prevented their ever paying so large a sum as this within five years; but a company was formed in London to advance periodically the sum due to the contractors, on receipt from the government of Egypt of promissory notes to pay sixty half-yearly instalments of £78,613, beginning on the 1st of July 1903. Protective works downstream of the dam were completed in 1906 at a cost of about £E304,000. It had been at first intended to raise the dam to a height which would have involved the submergence, for some months of every year, of the Philae temples, situated on an island just upstream of the dam. Had the natives of Egypt been asked to choose between the preservation of Ptolemy’s famed temple and the benefit to be derived from a considerable additional depth of water storage, there can be no question that they would have preferred the latter; but they were not consulted, and the classical sentiment and artistic beauty of the place, skilfully pleaded by archaeologists and artists, prevailed. In 1907, however, it was decided to carry out the plan as originally proposed and raise the dam 26 ft. higher. This would increase the storage capacity 21/2 times, or to about 9,375,000 millions of cubic feet.

There is no middle course of farming in Egypt between irrigation and desert. No assessment can be levied on lands which have not been watered, and the law of Egypt requires that in order to render land liable to taxation the water during the Nile flood must have flowed naturally over it. It is not enough that it should be pumped on to the land at the expense of the landowner. The tax usually levied is from £1 to £2 per acre.

See Sir W. Willcocks, Egyptian Irrigation (2nd ed., 1899); Sir C. C. Scott-Moncrieff, Lectures on Irrigation in Egypt. Professional Papers on the Corps of Royal Engineers, vol. xix. (London, 1893); Sir W. Garstin, Report upon the Basin of the Upper Nile. Egypt No. 2 (1904).

V. India.—Allusion has already been made to the irrigation of India. The year 1878, which saw the end of a most disastrous famine, may be considered as the commencement of a new era as regards irrigation. It had at last been recognized that such famines must be expected to occur at no very long intervals of time, and that the cost of relief operations must not be met by increasing the permanent debt on the country, but by the creation of a famine relief and a famine insurance fund. For this purpose it was fixed that there should be an annual provision of Rx. 1,500,000, to be spent on: (1) relief, (2) protective works, (3) reduction of debt. Among protective works the first place was given to works of irrigation. These works were divided into three classes: (i.) productive works; (ii.) protective works; (iii.) minor works.

Productive works, as their name implies, are such as may reasonably be expected to be remunerative, and they include all the larger irrigation systems. Their capital cost is provided from loan funds, and not from the relief funds mentioned above. In the seventeen years ending 1896–1897 the capital expenditure on such works was Rx. 10,954,948, including a sum of Rx. 1,742,246 paid to the Madras Irrigation Company as the price of the Kurnool-Cuddapah canal, a work which can never be financially productive, but which nevertheless did good service in the famine of 1896–1897 by irrigating 87,226 acres. In the famine year 1877–1878 the area irrigated by productive canals was 5,171,497 acres. In the famine year 1896–1897 the area was 9,571,779 acres, including an area of 123,087 acres irrigated on the Swat river canal in the Punjab. The revenue of the year 1879–1880 was nearly 6% on the capital outlay. In 1897–1898 it was 71/2%. In the same seventeen years Rx. 2,099,253 were spent on the construction of protective irrigation works, not expected to be directly remunerative, but of great value during famine years. On four works of this class were spent Rx. 1,649,823, which in 1896–1897 irrigated 200,733 acres, a valuable return then, although in an ordinary year their gross revenue does not cover their working expenses. Minor works may be divided into those for which capital accounts have been kept and those where they have not. In the seventeen years ending 1896–1897, Rx. 827,214 were spent on the former, and during that year they yielded a return of 9.13%. In the same year the irrigation effected by minor works of all sorts showed the large area of 7,442,990 acres. Such are the general statistics of outlay, revenue and irrigated area up to the end of 1896–1897. The government might well be congratulated on having through artificial means ensured in that year of widespread drought and famine the cultivation of 27,326 sq. m., a large tract even in so large a country as India. And progress has been steadily made in subsequent years.

Some description will now be given of the chief of these irrigation works. Beginning with the Punjab, the province in which most progress has been made, the great Sutlej canal, which irrigates the country to the left of that river, was opened in 1882, and the Western Jumna canal (perhaps the oldest in India) was extended into the dry Hissar and Sirsa districts, and generally improved so as to increase by nearly 50% its area of irrigation between 1878 and 1897. Perhaps this is as much as can well be done with the water at command for the country between the Sutlej and the Jumna, and it is enough to secure it for ever from famine. The Bari Doab canal, which irrigates the Gurdaspur, Amritsar and Lahore districts, has been enlarged and extended so as to double its irrigation since it was projected in 1877–1878. The Chenab canal, the largest in India and the most profitable, was only begun in 1889. It was designed to command an area of about 21/2 million acres, and to irrigate annually rather less than half that area. This canal flows through land that in 1889 was practically desert. From the first arrangements were made for bringing colonists in from the more congested parts of India. The colonization began in 1892. Nine years later this canal watered 1,830,525 acres. The population of the immigrant colony was 792,666, consisting mainly of thriving and prosperous peasants with occupancy rights in holdings of about 28 acres each. The direct revenue of this canal in 1906 was 26% on the capital outlay. The Jhelum canal was opened on the 30th of October, 1901. It is a smaller work than the Chenab, but it is calculated to command 1,130,000 acres, of which at least half will be watered annually. A much smaller work, but one of great interest, is the Swat river canal in the Peshawar valley. It was never expected that this would be a remunerative work, but it was thought for political reasons expedient to construct it in order to induce turbulent frontier tribes to settle down into peaceful agriculture. This has had a great measure of success, and the canal itself has proved remunerative, irrigating 123,000 acres in 1896–1897. A much greater scheme than any of the above is that of the Sind Sagar canal, projected from the left bank of the Indus opposite Kalabagh, to irrigate 1,750,000 acres at a cost of Rx. 6,000,000. Another great canal scheme for the Punjab proposed to take off from the right bank of the Sutlej, and to irrigate about 600,000 acres in the Montgomery and Multan districts, at a cost of Rx. 2,500,000. These three last projects would add 2,774,000 acres to the irrigated area of the province, and as they would flow through tracts almost unpeopled, they would afford a most valuable outlet for the congested districts of northern India. In addition to these great perennial canals, much has been done since 1878 in enlarging and extending what are known as the “inundation canals” of the Punjab, which utilize the flood waters in the rivers during the monsoon season and are dry at other times. By these canals large portions of country throughout most of the Punjab are brought under cultivation, and the area thus watered has increased from about 180,000 to 500,000 acres since 1878.

It is on inundation canals such as these that the whole cultivation of Sind depends. In 1878 the area was about 1,500,000 acres; in 1896–1897 it had increased to 2,484,000 acres. This increase was not due to famine in Sind, for that rainless province depends always on the Indus, as Egypt does on the Nile, and where there is no rainfall there can be no drought. But the famine prices obtained for agricultural produce doubtless gave an impetus to cultivation. In Sind, too, there is room for much increase of irrigation. It has been proposed to construct two new canals, the Jamrao and the Shikárpur, and to improve and extend three existing canals—Nasrat, Naulakhi and Dad. The total cost of these five projects, some of which are now in progress, was estimated at Rx. 1,596,682, and the extension of irrigation at 660,563 acres.

Turning from the basin of the Indus to that of the Ganges, the commissioners appointed to report on the famine of 1896–1897 found that in the country between the Ganges and the Jumna little was left to be done beyond the completion of some distributary channels. The East India Company’s great work, the Ganges canal, constructed between 1840 and 1854 before there was a mile of railway open in India, still holds its place unsurpassed among later irrigation work for boldness of design and completeness of execution, a lasting monument to the genius of Sir Proby Cautley, an officer of the Bengal Artillery, but a born engineer. Ever since 1870 consideration has been given to projects for irrigating the fertile province of Oudh by means of a great canal to be drawn from the river Sarda. The water is there in abundance, the land is well adapted for irrigation, but as there is a considerable rainfall, it is doubtful whether the scheme would prove remunerative, and a large section of the landowners have hitherto opposed it, as likely to waterlog the country. Among the four protective works of irrigation which were said above to have irrigated 200,733 acres in 1896–1897, one of the most important is the Betwa canal, in the parched district of Bundelkhand. This canal has cost Rx. 428,086, and causes an annual loss to the state in interest and working expenses of about Rx. 20,000. It irrigated, however, in 1896–1897 an area of 87,306 acres, raising crops valued at Rx. 231,081, or half the cost of the canal, so it may be said to have justified its construction. A similar canal from the river Ken in the same district has been constructed. Proceeding farther east, we find very satisfactory progress in the irrigation of southern Behar, effected by the costly system of canals drawn from the river Sone. In 1877–1878 these canals irrigated 241,790 acres. Rapid progress was not expected here, and 792,000 acres was calculated as being the maximum area that could be covered with the water supply available. In the five years preceding 1901–1902 the average irrigated area was 463,181 acres, and during that year the area was 555,156 acres, the maximum ever attained.

The canal system of Orissa was never expected to be remunerative, since in five years out of six the local rainfall is sufficient for the rice crop. In 1878–1879 the area irrigated was 111,250 acres, and the outlay up to date was Rx. 1,750,000. In 1900–1901 the area was 203,540 acres, the highest ever attained, and the capital outlay amounted to Rx. 2,623,703. It should be mentioned in favour of these canals that although the irrigation is not of yearly value, they supply very important water communication through a province which, from its natural configuration, is not likely to be soon intersected by railways. If, moreover, such a famine were again to occur in Orissa as that of 1866–1867, there would be no doubt of the value of these fine canals.

In the Madras presidency and in Mysore irrigation has long assumed a great importance, and the engineering works of the three great deltas of the Godavari, Kistna and Cauvery, the outcome of the genius and indefatigable enthusiasm of Sir Arthur Cotton, have always been quoted as showing what a boon irrigation is to a country. In 1878 the total area of irrigation in the Madras presidency amounted to about 5,000,000 acres. The irrigation of the eight productive systems was 1,680,178 acres, and the revenue Rx. 739,778. In 1898 there were ten of these systems, with an irrigation area, as shown by the accompanying table, of 2,685,915 acres, and a revenue of Rx. 1,163,268:

Irrigation.	Area Watered.	Total Revenue.	Total  Expenditure.	Net Revenue.	Capital and Indirect Charges.	Percentage  of Net Revenue to Capital.
Major Works.	Acres.	Rx.	Rx.	Rx.	Rx.
1. Godavari Delta	779,435	328,443	68,376	260,067	1,297,807	19.15
2. Kistna Delta	520,373	254,579	74,142	180,437	1,319,166	13.18
3. Pennar Weir System	70,464	28,160	5,937	23,123	189,919	7.59
4. Sangam System	76,277	32,627	7,037	25,590	385,601	3.68
5. Kurnool Canal	47,008	15,622	12,404	3,218	2,171,740	.15
6. Barur Tank System	4,421	1,162	385	777	4,250	1.39
7. Cauvery Delta	989,808	434,346	43,464	390,882	199,458	44.87
8. Srivaikuntam System	41,668	19,349	4,680	14,669	147,192	5.45
9. Periyar Project	89,143	37,526	10,751	26,775	852,914	.27
10. Rushikulya Canal	67,318	11,454	3,678	7,776	464,423	.54
⁠Total	2,685,915	1,163,268	229,954	933,314	7,032,470	7.88
Minor Works.
23 Works for which Capital and
Revenue Accounts are kept	535,813	200,558	34,655	165,903	1,693,878	4.44
Minor Works for which such
Accounts are not kept	3,131,009	830,175	193,295	636,880	..	..
Grand Total	6,352,737	2,194,001	457,904	1,736,097	..	..

In the three great deltas, and the small southern one that depends on the Srivaikuntam weir over the river Tumbraparni, extension and improvement works have been carried on. The Sangam and Pennar systems depend on two weirs on the river Pennar in the Nellore district, the former about 18 m. above and the latter just below the town of Nellore. The former irrigates on the left, the latter on the right bank of the river. This district suffered severely in the famine of 1877–1878, and the irrigation works were started in consequence. The Barur tank system in the Salem district was also constructed after the famine of 1877–1878. As yet it has not fulfilled expectations. The Periyar scheme has for its object both the addition of new irrigation and the safeguarding of that which exists in the district of Madura, a plain watered by means of a great number of shallow tanks drawing their supply from a very uncertain river, the Vaigai. This river takes its rise on the eastern slopes of the Ghat range of mountains, and just opposite to it, on the western face of the range, is the source of the river Periyar. The rainfall on the west very much exceeds that on the east, and the Periyar used to find its way by a short torrent course to the sea, rendering no service to mankind. Its upper waters are now stemmed by a masonry dam 178 ft. high, forming a large lake, at the eastern end of which is a tunnel 5700 ft. long, piercing the watershed and discharging 1600 cub. ft. per second down the eastern side of the mountains into the river Vaigai. No bolder or more original work of irrigation has been carried out in India, and the credit of it is due to Colonel J. Pennycuick, C.S.I. The dam and tunnel were works of unusual difficulty. The country was roadless and uninhabited save by wild beasts, and fever and cholera made sad havoc of the working parties; but it was successfully accomplished. The last of those given in the table above was not expected to be remunerative, but it should prove a valuable protective against famine. The system consists of weirs over the rivers Gulleri, Mahanadi and Rushikulya in the backward province of Ganjam, south of Orissa. From these weirs flow canals altogether about 127 m. long, which, in connexion with two large reservoirs, are capable of irrigating 120,000 acres. In 1901 the works, though incomplete, already irrigated 67,318 acres.

In addition to all these great engineering systems, southern India is covered with minor works of irrigation, some drawn from springs in the sandy beds of rivers, some from the rainfall of 1/2 sq. m. ponded up in a valley. In other cases tanks are fed from neighbouring streams, and the greatest ingenuity is displayed in preventing the precious water from going to waste.

Allusion has been already made to the canals of Sind. Elsewhere in the Bombay presidency, in the Deccan and Gujarat, there are fewer facilities for irrigation than in other parts of India. The rivers are generally of uncertain volume. The cost of storage works is very great. The population is backward, and the black soil is of a nature that in ordinary years can raise fair crops of cotton, millet and maize without artificial watering. Up to the end of 1896–1897 the capital spent on the irrigation works of the Deccan and Gujarat was Rx. 2,616,959. The area irrigated that year was 262,830 acres. The most important works are the Mutha and Nira canals in the Poona district.

In Upper Burma three productive irrigation works were planned at the opening of the century—the Mandalay, the Shwebo, and the Mon canals, of which the first was estimated to cost Rx. 323,280, and to irrigate 72,000 acres. The area estimated from the whole three projects is 262,000 acres, situated in the only part of Burma that is considered liable to famine.

In 1901, after years of disastrous drought and famine, the government of India appointed a commission to examine throughout all India what could be done by irrigation to alleviate the horrors of famine. Up to that time it had been the principle of the government not to borrow money for the execution of irrigation works unless there was a reasonable expectation that within a few years they would give a return of 4 or 5% on the capital outlay. In 1901 the government took larger views. It was found that although some irrigation works (especially in the Bombay Deccan) would never yield a direct return of 4 or 5%, still in a famine year they might be the means of producing a crop which would go far to do away with the necessity for spending enormous sums on famine relief. In the Sholapur district of Bombay, for instance, about three years’ revenue was spent on relief during the famine of 1901. An expenditure of ten years’ revenue on irrigation works might have done away for all future time with the necessity for the greater part of this outlay. The Irrigation Commission of 1901–1903 published a very exhaustive report after a careful study of every part of India. While emphatically asserting that irrigation alone could never prevent famine, they recommended an outlay of £45,000,000 spread over a period of 25 years.

See also Annual Reports Irrigation Department Local Governments of India; Reports of the Indian Famine Commissions of 1878, 1898 and 1901; Sir Hanbury Brown, Irrigation, its Principles and Practice (London, 1907).

VI. United States.—At the opening of the 20th century, during Mr Roosevelt’s presidency, the new “Conservation” policy (i.e. conservation of natural resources by federal initiative and control), to which he gave so much impetus and encouragement, brought the extension of irrigation works in the United States to the front in American statecraft (see Vrooman, Mr Roosevelt, Dynamic Geographer, 1909). Though the carrying out of this policy on a large scale was hampered by many difficulties, the subject was made definitely one of national importance.

On account of the aridity of the climate throughout the greater part of the western third of the United States, the practice of agriculture is dependent upon an artificial supply of water. On most of the country west of the 97th meridian and extending to the Pacific Ocean less than 20 in. of rain falls each year. The most notable exceptions are in the case of a narrow strip west of the Cascade Range and of some of the higher mountain masses. In ordinary years the climate is too dry for successful cultivation of the field crops, although under favourable conditions of soil and cultivation there are certain areas where cereals are grown by what is known as “dry farming.” The progress in irrigation up to the end of the 19th century was spasmodic but on the whole steady. The eleventh census of the United States, 1890, showed that 3,564,416 acres were irrigated in 1889. This included only the lands from which crops were produced. Besides this, there were probably 10 million acres under irrigation systems constructed in whole or in part. In 1899 the irrigated area in the arid states and territories was more than twice as great as in 1889, the acreage being as follows:—

Arizona		185,936
California		1,445,872
Colorado		1,611,271
Idaho		602,568
Montana		951,154
Nevada		504,168
New Mexico		203,893
Oregon		388,310
Utah		629,293
Washington		135,470
Wyoming		605,878
	Total	7,263,813

In addition to the area above given, in 1899, 273,117 acres were under irrigation in the semi-arid region, east of the states above mentioned and including portions of the states of North and South Dakota, Nebraska, Kansas, Texas and Oklahoma. The greater part of these lands was irrigated by canals or ditches built by individuals acting singly or in co-operation with their neighbours, or by corporations. The national and state governments had not built any works of reclamation excepting where the federal government, through the Indian department, had constructed irrigation ditches for Indian tribes, notably the Crow Indians of Montana. A few of the state governments, such, for example, as Colorado, had built small reservoirs or portions of canals from internal improvement funds.

The construction of irrigation canals and ditches was for the most part brought about by farmers joining to plough out or dig ditches from the rivers, descending on a gentle grade. Some of the corporations constructing works for the sale of water built structures of notable size, such, for example, as the Sweet-water and Hemet dams of southern California, the Bear river canal of Utah, and the Arizona canal, taking water from Salt river, Arizona. The cost of bringing water to the land averaged about $8 per acre where the ordinary ditches were built. The owners of extensive works were charged from $12 to $20 per acre and upwards for so-called “water rights,” or the privilege to take water from the canal, this covering cost of construction. Besides the first cost of construction, the irrigator was usually called upon to pay annually a certain amount for maintenance, which might often be worked out by labour on the canal. The cost ranged from 50 cents to $1 per acre; or, with incorporated companies, from $1.50 to $2.50 per acre and upwards. The largest expense for water rights and for annual maintenance was incurred in southern California, where the character of the crops, such as citrus fruits, and the scarcity of the water make possible expensive construction and heavy charges. The legal expense for the maintenance of water rights was often large because of the interminable suits brought during the times of water scarcity. The laws regarding water in most of the arid states were indefinite or contradictory, being based partly on the common law regarding riparian rights, and partly upon the Spanish law allowing diversion of water from natural streams. Few fundamental principles were established, except in the case of the state of Wyoming, where an official was charged with the duty of ascertaining the amount of water in the streams and apportioning this to the claimants in the order of their priority of appropriation for beneficial use.

It may be said that, up to the year 1900, irrigation progressed to such an extent that there remained few ordinary localities where water could not be easily or cheaply diverted from creeks and rivers for the cultivation of farms. The claims for the available supply from small streams, however, exceeded the water to be had in the latter part of the irrigating season. There remained large rivers and opportunities for water storage which could be brought under irrigation at considerable expense. The large canals and reservoirs built by corporations had rarely been successful from a financial standpoint, and irrigation construction during the latter part of the decade 1890–1899 was relatively small. Owing to the difficulty and expense of securing water from running streams by gravity systems, a great variety of methods were developed of pumping water by windmills, gasoline or hot-air engines, and steam. Ordinary reciprocating pumps were commonly employed, and also air lifts and similar devices for raising great quantities of water to a height of from 20 to 50 ft. For greater depths the cost was usually prohibitive. Throughout the Great Plains region, east of the Rocky Mountains, and in the broad valleys to the west, windmills were extensively used, each pumping water for from 1 to 5 acres of cultivated ground. In a few localities, notably in South Dakota, the Yakima valley of Washington, San Joaquin, and San Bernardino valleys of California, San Luis valley of Colorado, and Utah valley of Utah, water from artesian wells was also used for the irrigation of from 1 to 160 acres. The total acreage supplied by such means was probably less than 1% of that watered by gravity systems.

The development of irrigation was in part retarded by the improper or wasteful use of water. On permeable soils, especially those of the terrace lands along the valleys, the soluble salts commonly known as alkali were gradually leached out and carried by the percolating waters towards the lower lands, where, reaching the surface, the alkali was left as a glistening crust or as pools of inky blackness. Farms adjacent to the rivers were for a time increased in richness by the alkaline salts, which in diffuse form might be valuable plant foods, and then suddenly become valueless when the concentration of alkali had reached a degree beyond that which the ordinary plants would endure.

The situation as regards the further progress of irrigation on a large scale was however dominated in the early years of the 20th century by the new Conservation policy. Mr Roosevelt brought the whole subject before Congress in his message of the 3rd of December 1901, and thereby started what seemed likely to be a new sphere of Federal initiative and control. After referring to the effects of forests (see Forests and Forestry) on water-supply, he went on as follows:—

“The forests alone cannot fully regulate and conserve the waters of the arid regions. Great storage works are necessary to equalize the flow of the streams and to save the flood waters. Their construction has been conclusively shown to be an undertaking too vast for private effort. Nor can it be best accomplished by the individual states acting alone.

“Far-reaching interstate problems are involved, and the resources of single states would often be inadequate. It is properly a national function, at least in some of its features. It is as right for the National Government to make the streams and rivers of the arid regions useful by engineering works for water storage, as to make useful the rivers and harbours of the humid regions by engineering works of another kind. The storing of the floods in reservoirs at the headquarters of our rivers is but an enlargement of our present policy of river control, under which levees are built on the lower reaches of the same streams.

“The government should construct and maintain these reservoirs as it does other public works. Where their purpose is to regulate the flow of streams, the water should be turned freely into the channels in the dry season, to take the same course under the same laws as the natural flow.

“The reclamation of the unsettled arid public lands presents a different problem. Here it is not enough to regulate the flow of streams. The object of the government is to dispose of the land to settlers who will build homes upon it. To accomplish the object water must be brought within their reach.

“The reclamation and settlement of the arid lands will enrich every portion of our country, just as the settlement of the Ohio and Mississippi valleys brought prosperity to the Atlantic States. The increased demand for manufactured articles will stimulate industrial production, while wider home markets and the trade of Asia will consume the larger food supplies and effectually prevent Western competition with Eastern agriculture. Indeed, the products of irrigation will be consumed chiefly in upbuilding local centres of mining and other industries, which would otherwise not come into existence at all. Our people as a whole will profit, for successful home-making is but another name for the upbuilding of the nation.”

In 1902, by Act of Congress, a “reclamation fund” was created from moneys received from the sale of public lands; it was to be used under a “Reclamation Service” (part of the Department of the Interior) for the reclamation of arid lands. The “Truckee-Carson project” for irrigation in Nevada was immediately begun. About thirty other government projects were taken in hand under the new Reclamation Service, in some cases involving highly interesting engineering problems, as in the Uncompahgre Project in Colorado. Here the Uncompahgre and Gunnison rivers flowed parallel, about 10 m. apart, with a mountain range 2000 ft. high between them. The Uncompahgre, with only a small amount of water, flowed through a broad and fertile valley containing several hundred thousand acres of cultivable soil. The Gunnison, with far more water, flowed through a canyon with very little land. The problem was to get the water from the Gunnison over the mountain range into the Uncompahgre valley; and a tunnel, 6 m. long, was cut through, resulting in 1909 in 148,000 acres of land being irrigated and thrown open to settlers. Similarly, near Yuma in Arizona, a project was undertaken for carrying the waters of the main canal on the California side under the Colorado river by a siphon. In the report for 1907 of the Reclamation Service it was stated that it had dug 1881 m. of canals, some carrying whole rivers, like the Truckee river in Nevada and the North Platte in Wyoming, and had erected 281 large structures, including the great dams in Nevada and the Minidoka dam (80 ft. high and 650 ft. long) in Idaho. As the result of the operations eight new towns had been established, 100 m. of branch railroads constructed, and 14,000 people settled in what had been the desert.

A White House conference of governors of states was held at Washington in May 1909, which drew up a “declaration of principles” for the conservation of natural resources, recommending the appointment of a commission by each state to co-operate with one another and with the Federal government; and by the end of the year thirty-six states had appointed Conservation committees. Thus, in the first decade of the 20th century a great advance had been made in the way in which the whole problem was being viewed in America, though the very immensity of the problem of bringing the Federal power to bear on operations on so vast a scale, involving the limitation of private land speculation in important areas, still presented political difficulties of considerable magnitude.