Notes on the Percolation of Water.—Systems of Drainage of Cuttings and Embankments in Different Kinds of Earth and under Diverse Conditions.—The Construction of Culverts, Pipe-drains, Trenches, Ditches, and Catchwater Drains.

With respect to the percolation and drainage of water in cuttings and embankments, in cuttings the chief consideration is to gently extract and conduct the water so as to avoid any accumulation or localised flow beneath the original surface of the ground in order to prevent the surface water eroding the slopes or collecting or forming a course, saturating the ground outside them, and to ensure that the earth is not more charged with water than when in its normal condition; for, as soon as the state of absorption has reached that of dissolving or separating the particles, however fine, aqueous action is likely to produce slips, and a flow of water or vibration will supply the disturbing force necessary to commence a movement. In embankments one of the chief precautions is to obviate any flow of water upon or at a few feet beneath the land upon which an embankment has to be deposited, as it will disturb the feet of the slopes and the base, reduce the adhesion to and the friction of the tipped earth upon the ground and form a sliding surface.

An accumulation of water upon the formation must be prevented, and, as in the majority of cases a railway or road is not level, any collection of water at the commencement of an incline or at a change of gradient should be provided against, and especially any localisation of flow down the slopes from the formation; the main point being to keep a cutting or embankment in a uniform state so that settlement is equable. By carefully watching the effect of heavy rain upon the slopes and the formation, the places where water amasses can be traced, and means used to restore the surface to the same condition as the other portions of a cutting or embankment. As water is the chief cause of slips, the friction and cohesion of earth being impaired and, perhaps, destroyed by it, it is obvious that at the time percolation, which varies greatly with the seasons, is at its maximum, i.e. when the earth becomes in a damp or wet state, slips are to be most expected, and particularly soon after the commencement of the wet season. It is known that upon the thawing of a heavy fall of snow, and of quickly succeeding and separate falls of snow, percolation is great; also after heavy and continuous rains, especially if the strata dip towards a river, and in the case of springs whose yield depends more upon percolation than the amount of rainfall, a wet winter will cause an increased flow some time after, when the earth may become saturated.

If it could be determined at what depth in any earth in any state percolation, evaporation, and meteorological influences would cease, and also their effects, rules could be deduced for guidance in draining cuttings and embankments. The manner of the execution of ordinary cuttings and embankments is so dissimilar to that of filter-beds of waterworks, that the experiments made for such purposes are only of comparative value for the former works, for the condition as well as the character of the soil affects its permeability by water. For instance, in cuttings, with the exception of some surface disturbance during the process of excavation, the removal of vegetation or a covering, and the exposure of the ground to the atmosphere, &c., the normal state of the earth is not much altered; but in embankments different soils may be intermingled in a manner almost unknown in nature, the varieties of mixtures of earth being most numerous, and they may be in every condition of compactness, dryness, and dampness amounting almost to saturation, and in any case, therefore, percolation is temporarily or permanently increased consequent upon the earth having been disturbed and loosened.

The general principles of the percolation of water are here only briefly referred to, as they particularly concern slips in earthwork: but obviously the quantity of percolated water greatly affects the stability of a slope, for the surface water should not be guidelessly allowed to soak into or be absorbed by the ground at the top, and so proceed through and down the slopes, as then the pent-up water tends to press out the face which may be temporarily sun-dried. As in excavating cuttings the surface is bared and vegetation removed from the soil, water has easier access, and unless the ground when excavated is immediately covered as before, its normal state is not preserved. One point of considerable importance is to ascertain whether in any cutting or embankment percolation is uniform and regular; some infiltration will necessarily take place, as water will gravitate from the top to the bottom and will find the easiest course or line of least resistance, which may not be at the lowest level.

As the amount of the annual rainfall varies greatly according to the country, and, even in England, considerably in a small area, the earth will be more affected in one place than another; for instance, 48 to 50 inches is approximately the annual average rainfall upon the extreme S.S.W. coast of England, being greatest at the highest level nearest the sea and to leeward; it diminishes gradually from W. to E. to from 26 to 24 inches, the minimum of about 20 inches being in Essex. The differences of quantity must therefore be regarded; but such rainfall is as nothing compared with that of tropical lands, for the fall often continues many hours, and yet equals and perhaps exceeds an inch per hour. The heaviest fall and its usual time of appearance should be ascertained, as earthworks may have to be constructed in a peculiar district where the rainfall may be more than double that of the average wettest district, and it will often be much greater at the foot or the top of neighbouring hills than on a flat coast. Local information from reliable sources is the best guide when confirmed by general knowledge. In the tropics 100 to 200 inches in depth of rain instead of 20 to 30 inches has to be treated, and frequently half the total annual rainfall in England comes down in twenty-four hours. It is almost superfluous to name that the protective works which would be amply sufficient in one country may be useless in another, simply because of the variation in the amount of the rainfall and the capacity of different earths to resist or invite the percolation of water.

Obviously percolation will vary considerably with the seasons, and a succession of wet periods or a continuous downpour will increase the quantity of infiltrating water; but the effect of a fall may not be experienced until some time after it occurs, as in districts or rivers that are fed with water from the thawing of snow upon surrounding or distant hills water reaches the lower tracts of country in hot and sunny, and therefore generally dry weather, when evaporation is the greatest, and not in winter. Again, there is generally very little rainfall over flat deserts, but an excess upon mountain ranges which may surround a desert, and especially in tropical countries experience has proved that storms and rainfall are often local and extend over a small area, one district being more liable to such a visitation than another; they are also, as usual, of irregular duration and severity.

The position of a river may also affect the percolation of water more upon one side of a valley than another, for a river seldom has its course in the centre of a vale, but is generally nearest to the steeper and higher side of a hill. The configuration of a country governs to a great extent the flow and quantity of the rainfall that sinks into the earth, as in a hilly country and in impervious soil the water is quickly discharged into an adjacent river, taking the easiest course. In a flat country and pervious soil rain percolates the earth and saturates the ground, or reappears in springs.

As by drainage the retentive power of the soil is not allowed free operation, water rapidly flows into the drains instead of being chiefly held by the earth and watercourses, and channels are sometimes created; and where the rainfall is heavy or occurs in a comparatively short time floods may be caused, although the soil when drained, and therefore in a drier state, absorbs more water than when undrained and in a damp condition, water will pass through it quicker, and the discharge is thus increased in volume and velocity.

Earth may be in a damp state, either from mere surface percolation and accumulation of water, or from springs which may never cease to flow; on the contrary, in rainless districts, from the almost perpetual daily drying power of the sun, the earth is sometimes found to be firmest and hardest upon the top, and for a few feet below it, than at greater depths. Separate masses of vegetation usually indicate damp places in a bare country.

As rain flows more quickly from non-absorbent soil, such as rock, and slowly permeable earth, as the clays, than from porous soil, the surface discharge is greater; and unless the water is guided, pools are likely to be formed and weak places created, especially if the ground dips towards a cutting. It is well to remember that a cutting being excavated upon the side of a hill or upon table-land may change the direction of the flow of the drainage waters, and an embankment may obstruct and interfere with them, and should the strata incline towards the excavation, it obviously favours a discharge of water into it.

If water permeates the soil or trickles down through fissures or veins, it will continue to do so until an impervious layer is reached, when it will be deflected and may become a current; therefore, whenever a permeable stratum overlies an impermeable, and the impermeable earth inclines, an increased flow may be expected, as also near the junction of tributary waters with streams or rivers.

In a drained district, water will not usually be encountered in large quantities until a depth is reached below the level of general drainage, i.e., about the invert of the nearest drains or sewers, especially in porous soils, but in the case of a pervious subsoil, such as sand or gravel, or if the tides rise in an adjacent river or the sea to the level of any foundations or above it, more water may percolate over the site or drainage area than has time to flow away between tides, the water will then rise, and systematic pumping becomes a necessity, unless the volume of the ingressing waters can be sufficiently reduced by the deposition of an impervious layer upon or in the river-bed or sea-shore, or by sheet-piling or other means, or the flow confined, which may be a risky operation in loose soil, depending upon its resistance to scour.

In clay soils, so far as slips are concerned, the action especially to be feared is the trickling of water down fissures which may extend to depths below the bottom in cuttings and create slimy surfaces, disconnecting masses of earth, and finally offering a ready means of movement, which swelling of the clay or vibration may complete. The experiments of Mr. Baldwin Latham, M. Inst. C.E., on the absorption and retention of water by clay soils, gave the following results:—The stiffest clays retained the greatest quantity of water. Clay soils can absorb and retain from 40 to 60 per cent. of water by weight. Marly clays hold less water than the pure clays. In the case of loamy soils, the percentage of water retained varied from 35 to 60 per cent. by weight, the mixture of sand and clay, therefore, limited the amount of water which it would naturally hold.

As in chalk soils fissures occur, the percolation of water and the effects of the atmosphere through the pores causes movement, and even crevices and breaks in rocks are not to be disregarded with impunity, as they are channels of disintegration.

Most earths when dry attract water, but if they are regularly irrigated, they require less moisture, depending upon the nature of the earth, and slope and relative level of the land. All soils when broken, as in embankments, absorb more water than when in an unbroken state, as in cuttings; for instance, it has been found by experiment that clayey and retentive earths will absorb about 7 per cent. more water, and light porous soils about 6 per centum. Of course, the increment varies. Even wet retentive soil, if handled, becomes considerably less impervious to moisture.

Mr. Evans, F.R.S., has proved by experiment that percolation through pure chalk is much greater than through ordinary top soil consisting of gravel, loam, and mould, both being covered with turf, and that in winter the average proportions of percolation are as about 1 for soil to 1·5 for chalk; in summer 1 for soil to 2·6 for chalk. The depth also to which chalk will allow a passage of water is some 60 per cent. more than ordinary top soil.

Rain will percolate through chalk or any open soil until it meets an impervious stratum, or to that place which is in a state of saturation, when the water must either flow away or the level of saturation must gradually rise. This causes rivulets to burst out in places after heavy rain, when the water has had time to percolate, and the rainfall has exceeded the average, but such springs will cease when the local excess has terminated. However, in cuttings it is the flow from fissures that is to be feared, and their size may indicate the quantity of water that may be expected to issue from them. It seems to be generally agreed that the supply of water from chalk is derived from rain, which percolates through innumerable fissures, and that in all rocks, whether limestone, sandstone, granite, sand, or clay, it is by means of the fissures, seams, and veins that the supply of water is obtained from rain, and springs created.

With regard to the percolation of water through sand, it may always be expected to be very considerable, and the soil may under certain conditions become water-charged. Mr. Greaves, M. Inst. C.E., has shown by experiments that the average percolation through ordinary top soil is only about one-third of that of sand, but the evaporation from a surface of ordinary soil was about four times more than from a similar surface of sand, and also the amount of percolation in ordinary top earth was small on the whole, and, perhaps, the percolation through ordinary ground would be about 25 per cent. of the rainfall, but 80 per cent. in average sand.

Experiments have also shown that the absorbent capacity of sand decreases regularly according to the fineness of the grain, and that “some sandy soils will not absorb more than 20 per cent., but sandy soil containing peat, as moorland, as much as 80 per cent., both computed by weight.”

The quantity of water absorbed by loamy soil will vary considerably according as clay or sand preponderates in the mass. Earth may become so mixed with coarse or fine sand that, when saturated, it approaches the condition of a quicksand.

The general effect of percolation has been briefly described as follows. Upon water entering the pores of an earth it displaces the air or liquid previously present, forcing the former upwards into the atmosphere, and the latter downwards.

Having briefly referred to the percolation of water in cuttings and embankments, the drainage is now considered. It must be either precautionary, i.e., to prevent a slip, or remedial, i.e., to drain a slip.

The aim of any draining operations to prevent slips in earthwork is to search for the source of water discharge, to tap and gently conduct it away and prevent it reaching, accumulating, percolating, or being confined within the slope of a cutting, which it may then reduce to a pulpy condition; its free effluxion being most important, as also the lessening of the percolation of rain and surface-waters. The drainage of cuttings or embankments may consist of wells, culverts, closed or open channels, pipes, and tile drains of every reasonable and economical form, and may be placed in various positions. To describe them and the different systems of draining is to open up a subject requiring several volumes; here the endeavour is made to indicate whether elaborate or ordinary drainage is required, or mere water-tables and surface drains, and care in the process of excavation and deposition, protection of the slopes, and in giving them sufficient inclination to prevent movement. If possible and time allows, the drainage in treacherous soils should always be commenced before the main excavation, and, in any case, simultaneously.

Rock and solid impermeable earth may merely require to be surface drained, but all treacherous and porous soil, deep draining; and granular soils, which usually exude water from the whole mass, demand different treatment to those earths which discharge water at particular places; but it may be most difficult to drain a mixed soil, such as sandy loam and silt. With the exception of a counterfort and drain at the foot of a slope, and an impermeable catchwater drain upon the slopes and top drains, to prevent and lessen surface percolation, the best method to adopt in earth of this description may be to sink wells at intervals to intercept the flow or percolation of any ground waters; to attempt to drain or draw out the water in the soil will end in comparative failure. To reduce the volume of the percolating waters is the object to be attained, and then evaporation, vibration, which tends to shake down water, and time may gradually convert the earth to the desired drier condition. The wells can be filled with broken stone or coarse gravel to support them, and prevent their closing.

It is an advantage to prevent the percolation of water into soil that will not readily part with it, such as the clay earths, as it may be economically impossible to drain or restore the earth to its normal condition, and should the strata be upheaved, intermixed, and of a permeable and impermeable character, a scientific application of drainage can alone succeed. When the source of the water is ascertained, it can be seen whether a complete system of drainage is necessary over the whole of a cutting or only a portion of the slopes. Dampness and the egression of water may be merely local; if so, by boring a hole and inserting a drain into a slope to tap the flow it may be cured, the surface being made dry by a layer of ashes or other absorbent material. Pipes or tile drains may be sufficient when springs exist, or the flow of water is local, and in loose soils it is especially advisable to provide openings for cleaning any covered drains. Brickwork, masonry, concrete, pipes, or other rigid drains, may not be suitable for ground likely to unequally subside, as they will probably crack or leak, and loose-jointed pipes or over-lapping tile drains may be required. In treacherous clay soils surface longitudinal and transverse drains will most probably be insufficient, and deep draining of the mass be necessary, also the ground upon which earth is deposited will require to be drained and a layer of rubble stone placed upon it, a cutting or embankment being divided into small drainage areas by deep open dry stone trenches.

Water may not sufficiently percolate into a hill, either because of its surface being covered with vegetation, or the soil being of an impervious nature. It may then flow down the slope of a cutting which will probably be bare and unprotected. To prevent slips the discharge must be carefully controlled and led away, particularly when the formation is drift soil upon rock, or the earth will be liable to saturation and degradation; also to prevent a flow of water under its seat, and upon the natural ground, should an embankment be deposited upon it.

When a slope upon the hill-side of a cutting is of considerable length and steepness, it is advisable to bench it and divide it into a series of terraces and short slopes, and to provide catchwater drains, with impermeable surfaces so as to prevent any surface water attaining a high velocity, and scouring power. All surface water upon the side of a hill should be controlled, and catchwater drains may suffice to do this.

Should an impervious stratum be upheaved so as to make a reservoir wall for water under the seat of an embankment, it is useless to surface-drain the valley side of it in sidelong ground, as it will not affect the waters that trickle down the hill, which will be dammed up to the top level of the impervious stratum, and may saturate the seat of an embankment and cause a slip. In such a case through drainage must be created from the hill to the valley, and the impervious upheaved cap must be pierced so as not to interfere with the passage of water.

In countries where there is a certain dry and rainy season, the necessary provision required for drainage must not be computed from the visible effects of the rainfall soon after the wet season has commenced, but the maximum flood may be discerned when the earth has absorbed or retains the moisture evaporated during the dry season, and becomes water-charged. In the tropics or exceptionally wet districts the only effectual method of draining may be to divide the area into small portions, as the rainfall may be so great, sudden, and continuous that unless it naturally flows into a channel, which should not be diverted or its course be materially altered, it will be impossible to control the waters. In an exceptional situation where a railway must be located in a ravine and close to a river, and the material of which an embankment is made is compact earth or the ground firm, it may be advisable to allow the overflow waters of a river, or the flow of surface water towards it, when the extreme flood level is known, to gently pass over a line of railway, it being kept at such a level as not to impede free working, and to ensure that any back water is not prevented from escaping. If not, an embankment may slip and require extensive and frequent bridges, culverts, or drainage channels in order to provide sufficient waterway, and to attain permanent stability of the embankment.

In all cases it is necessary to determine the depth to which the drains must be placed to be effectual, and their position, extent and number, and it should be remembered that the fewer there are the greater will be the velocity and discharge; also if many small drains are inserted the soil may stand, but if only one or two are made the surface may succumb to the erosive action of the flowing waters. The provision of a water-table or the mere surface drainage of a cutting, or the seat of an embankment may be of little use, as the superimposed soil may slide upon a stratum, and unless this bed is thoroughly drained rain may quickly destroy the equilibrium. Of course, in the case of a slip, the drains to be effectual must be placed below its level and down to the layer upon which movement has occurred, especially in clay and retentive and impervious earths, for instance, surface drainage in yellow clay is almost useless.

Should the source be known from which the water issues the drainage may be local, and if a spring be unsealed it may be necessary to insert pipes in the slope, for until the spring is tapped and guided no system of drains may be effectual.

By inserting a stand-pipe over a spring, the height to which the water will rise will approximately show the head-level of the supply. If possible, this should be ascertained, as it may happen that the water can be drained by gravitation in pipes without much excavation being required in the slopes or formation; but care should be taken that no water is allowed to settle or accumulate for the purpose of its being conducted away unless upon a protected surface.

With regard to catchwater drains upon the cess at the top of the slopes, they should be cut before the excavation is commenced; and it is important to remember that instead of their affording protection by guiding the surface waters, which would otherwise proceed towards the slopes, they may increase the percolation by localizing the water and allowing it to accumulate and find its way to the slope; and in sidelong ground it may, therefore, be necessary to protect their valley more than their uphill side, as the surface water will impinge against it; but when they are practically impervious and gently direct the surface water they are advantageous, and in permeable soil, unless they are so constructed as to be impervious, perhaps it is better to have none. They should be as reasonably distant from the top of the slope as is convenient without weakening the foundations of the posts supporting any fencing, and in order to quickly discharge the water and lessen the chance of their becoming choked by detritus or ice, they should have considerable inclination. As the adoption of even a moderate fall in the drains may erode the side ditches and cause water to percolate to the slopes and make a water-seam, it may be necessary to protect the bottom, and as the depth of the side drains in order to be effectual may be considerable, according to the character of the soil, the sides may also require to be covered and supported. In cuttings in many soils sufficient stones can be picked out to cover the surfaces of the side drains, and they can be roughly packed, the smaller stones being rammed into the interstices between the larger, which will gradually become filled; also a covering of brushwood, rammed earth, or puddled clay can be used, or other expedient which occasion may suggest. The inclination must not be steeper than the natural or protected bed can bear without the water scouring it, and yet should be sufficient to prevent any deposit or choking, and the drains should be cleared regularly, and especially in the autumn in England or before the wet season commences, and all depressions in which water can accumulate should be levelled in order to assist easy discharge. Small open drains become choked somewhat easily, and it is therefore advisable to make them according to the nature of the soil, situation, and requirements not less than 1 foot 6 inches to 3 feet in width at the top and 1 foot at bottom; they may be from 1 foot to 3 feet in depth. Care should be taken that the bends are not too abrupt or the water may make its own course. A gentle curve considerably increases the flow. The angle of a bend should be as easy as possible, and not exceed 26° or 2 to 1.

In sandy and loose soils, if unprotected, open drains may be difficult to maintain even when filled with broken stone, and covered or pipe drains be necessary, and those loosely filled with stones or faggots may not succeed; in any case no run of the sand must be allowed, and it is advisable to rapidly construct them. In peaty soils, from subsidence of the ground consequent upon draining, the drains often become choked. The depth of a drain in the formation may require to be deeper than elsewhere, as at the toe of the slope, the weakest part, the water will generally be most abundant.

Upon railways, the advantage of thorough drainage of the formation as regards the stability of the permanent way and reduction of the cost of repairs is proverbial, but here is only named in its relation to the prevention and the reparation of slips.

When field drains are intercepted in the slopes, drain pipes or timber ducts should be joined to them and be connected with the general drainage of a cutting. Draining the slopes, providing outlets for the water, and also support to the earth can be effected by means of a counterfort of permeable material at the toe of the slope, with its foundation a few feet below formation level, with open drains at right angles or obliquely to it extending from the foe to the top of the slope.^[1] These open drains and trenches can be filled with stones, gravel, hard chalk, ashes, brushwood and gravel, broken bricks, burnt clay or other suitable material; and a simple covering of picked ashes, &c., over a moist place in a clay cutting of little depth may be sufficient. The distance apart of such trenches in clay cuttings generally ranges between 10 and 33 feet; their location must be governed by the consideration that they exist in order to prevent a localization of water in the slopes: their width is usually from 2 feet to 8 feet, and most frequently 4 to 6 feet, and the depth from 1 foot 6 inches to 3 feet and upwards, below the surface of the slope, but in very wet embankments a width from 6 to 10 feet and of such a depth as the soil will allow without extensive support. If the earth has not slipped, and it should be found that the ground is wet for from 6 to 8 feet or so beneath the surface, and the depth of the trench is made about half that of the wetted soil, it is sufficient to collect the water, but if it has slipped most probably such drains will not be effectual until carried down below the seat of movement. At the foot of a slope these drains should be connected with longitudinal channels parallel to the formation to gently convey the water to the nearest outlet. The location, distance apart, depth, width, and direction of the trenches must be governed by the nature of the soil, the depth of a cutting or height of an embankment, the area of the surface to be drained, the quantity of water in it, the presence of water-seams and weak places, and by other minor considerations.

With respect to the material with which the trenches should be filled, a uniform substance having considerable power of absorption, and but few particles between the interstices, is to be preferred, as the trickling of water and vibration causes the smaller material to fall towards the bottom of the trench, which, therefore, may become partly choked, and free drainage be interrupted at the toe of the slope, the most vulnerable place. As sand is always present in unwashed gravel, it will gradually flow or fall to the base of the trench and will prevent equal drainage, but properly burnt clay, being a more uniform and fragmentary substance, is better than sandy gravel, as it is dry and porous; but much depends upon it being well burnt or it may weather. Ashes and chalk are excellent collectors of water and are usually of an even character, but ashes are better than chalk, as the latter material, unless very hard, is liable to become disintegrated by the action of the atmosphere, rain, and frost. The difficulty with respect to ashes is to obtain them in sufficient quantity, free from dust, and of the requisite size. When weight is required as well as drainage, burnt clay, gravel, or chalk is to be preferred to ashes. Burnt clay, although burned upon the site, is generally a more expensive material than gravel, but the amount of moisture in the clay will principally determine the quantity of fuel necessary to burn it, and therefore the cost. If shale is present, by thoroughly igniting it the expense of burning may be nearly reduced to that of lighting and turning over, as it will usually burn unaided. All trenches should drain into pipes placed below the formation or into open drains at a sufficient distance from the toe of a slope as not to deleteriously affect it, in order that the water may be controlled and gently conveyed to an outlet.

The chief objection to open drains is that all excavated trenches or inserted drains in the slopes destroy the cohesion of the earth and aid in detaching portions of the surface. If the cohesion and adhesion of the soil were the same under every condition, this would be a cogent reason against the system, but, as in earthwork, every degree of moisture from dampness to saturation may be attained, the cohesive power is a very variable quantity, apart from the effects of vibration; and also open drains undoubtedly do cause a slope to be drier, and moderate local humidity, and therefore increase the cohesion and general stability of the part drained; the only fear being that from inattention they may become choked; then they are dangerous, as they will permanently collect and retain water instead of temporarily retaining and gently guiding it. A careful consideration of the circumstances may much reduce this objection by indicating whether it is advisable to have only a few deep, or several small surface drains. Provided proper precautions are taken, experience indicates that filled-in trenches in the slopes are generally successful, and certainly are simpler and cheaper than a retaining wall at the toe of a slope. In pervious soil it may be economically impossible to drain the slopes unless they are divided into sections, and should the material in the trenches be well-packed and pressed down, it may even increase the friction between the separated portions.

In connection with the drainage of cuttings may be named that excavation should not be allowed to be cast upon the cess unless some distance from the edge of a slope, and only temporarily for purposes of ballasting and metalling; as such spoil-banks increase the load and localize the water to be drained.

A more complete system of drainage may be necessary than those previously named, consisting of a combination of wells, open, or holding filtering material; pipe-drains or filled-in trenches, wells with a pipe leading to a catchwater drain, &c., or other usual methods of land drainage. For instance, it may be discovered that water issues from a spring outside the fence or cutting; if so, in order to drain the slopes it may be necessary to sink a well to a stratum below the formation level so as to tap the spring, thereby preventing an exudation of water upon the slope; this is a better plan than drawing the water into the slope and then draining it.

When pervious earth overlies an impervious stratum, i.e., gravel or sand upon clay, rough-filled wells at intervals inside the fence extending 3 or 4 feet into the clay, with an outlet drain, may be required to prevent a flow of water upon the clay and a wetted surface upon which the gravel can slide; and it may be necessary to have a cess on the slope between the top of the clay and the bottom of the gravel or sand, with a catchwater drain upon it, particularly in a cutting in sidelong ground.

Should the soil be silty sand, or be charged with water, consequent upon the formation of the country, it may be impossible to drain a cutting without a complete system of wells, catchwater drains and pipes, and even then it may be difficult to separate the water from the earth. In building a drain-shaft it should be remembered that it may not only be subject to a compressive strain, but also to transverse strain and flexure from different pressures of the earth at various depths, especially when the soil is not the same throughout, and unequally damp.

When “boils” occur in sand-cuttings, perhaps the cheapest expedient is to place a shaft over the boil, weight the bottom sufficiently to prevent a movement of the sand, but to allow the water to escape, and make a discharge outlet after having ascertained its head level: vide Chapter XII. On no account should a spring be stopped, as such action will result in its diversion to some other place; but the water flowing from it should be guided and discharged. Weighting may arrest a slip in any sandy soil, also clay or any impervious material placed upon the sand, or sinking a well outside a cutting to a depth of some 5 feet below the bottom may effect a remedy by abstracting the surplus water, but care must be taken not to disturb the sand.

Slips in embankments frequently occur from the percolation of water through the formation to the slopes, and so to the toe, the lower portions become disintegrated by moisture and the effects of weather, and cause the upper parts to slide or move. To lessen percolation and to prevent an accumulation of water upon the formation, it is usual for its centre to be raised a few inches above the level of the top of the slope. This is undoubtedly a good practice, as it also tends to drain the ballast, but it may be nullified in time if the entire width of the formation be not covered with an impermeable layer or with ballast; for when an unprotected space remains between the toe of the slope of the ballast and the top of the slope of an embankment, water is liable to percolate through the cess and cause a slope to become wet and unstable; particularly so if the ballast is broken rock and has side walls instead of slopes, as then a depression will probably be made by the platelayers or signalmen walking upon the cess. In all cases where the material is treacherous or likely to slip, it is advisable to cover the top of embankments of considerable height throughout their width with ballast or some impervious soil, provided the permanent way is also properly drained. This is the simplest precaution to take respecting the preservation of the formation level or summit of an embankment. All grass, dirt, and refuse should be regularly removed from it and anything that obstructs free drainage. The nature of the ballast also affects the evenness of the surface of the formation, as if it consists of broken rock, the equal and regular packing of the sleepers is not so easily effected as with gravel ballast; the sleepers are frequently not uniformly supported throughout their length, the pressure upon the formation is localized, depressions are formed and water collected, and slips and subsidences in soils of a treacherous nature may be induced from this cause, as the equilibrium is soon disturbed. The formation should be so drained and constructed that water cannot percolate to or cause the surface to become soft and work up into or through the ballast, or a state of unsettlement will be produced by water soaking through the ballast to an embankment, and so saturating part of it and forcing out the upper portion of a slope. In all close granular ballast cross channels should be made to lead away the surface water. Transverse open tile-drains may be required leading to an impervious channel. Water has been known to percolate through a considerable depth of ballast when added to restore a sunken embankment, even through as much as 7 to 10 feet when two falling gradients induced a flow of the surface waters to one place. In certain situations it may be necessary should an embankment be of clay or treacherous soil when wet, to cover the formation with an impervious stratum to prevent percolation to the embankment, and to thoroughly and separately drain the ballast placed upon it.

At the base of an embankment a ditch should be cut upon the higher side, or both sides, as near as convenient to the fence. When in addition to the interception of any surface waters by an embankment the ground is very retentive of moisture, it may be necessary to drain the seat; with this object trenches can be excavated at intervals at right angles or obliquely to the centre line of the embankment, and be filled with some hard filtering substance, such as stone or gravel, so as to effect and control the discharge of the waters. Should this be too expensive a method to adopt, and always provided the surface waters are prevented from flowing or trickling into the base of an embankment, the ground might be excavated so as to equally incline downwards towards the centre, the level at that point being 1 foot to 2 feet below the toe of the slope on each side, according to the width of the base of the embankment; a small trench being cut in the centre and filled with stones, and covered at the top with brushwood or hurdles or other provision necessary to ensure it being permanently an effectual water channel, with occasional or other drains to lead the water to the nearest culvert or side ditch. When it is found that water passes over the surface of firm soil upon which an embankment is deposited, the water must be intercepted and led away; and should an embankment be of retentive earth, in order to tap the water that has flowed and percolated into it, and to restore the earth to its normal condition and prevent slips and subsidences, it may be necessary to sink shafts to a depth of a few feet below the seat of an embankment until the mass is drained.

With regard to culverts, a settlement or slip of an embankment over a culvert may unequally strain, fracture, and displace portions of it, and therefore interrupt the flow of the drainage waters, which may then reach the seat of an embankment and cause it to be in a dangerous condition. They are usually necessarily placed at the deepest point of an embankment, and consequently the most difficult to make repairs. In such a situation they should be built sufficiently large to allow of the easy passage of a man, in order that due inspection may be made, and be constructed of materials of a durable character. Should a naturally firm bank exist on one or both sides of a stream, it should be stripped of all plant growth and decaying matter, and be preserved in order to form a natural wall to relieve a culvert from side pressure, but firm or hard material must be inserted between the back of the wall and the face of the stream bank so as to support the wall against the pressure it receives from the arch. This leads to a consideration of the best form of culvert. In clay soils, and those which exert pressure from expansion, especially if the culvert is surrounded by clay earth, the circular is generally considered to be the best form, and this has been proved to be so in tunnels in similar soil, with splayed wing-walls to assist and guide the flow and help to keep a clear entrance. In granular soils, such as dry sand or gravel, the earth acts differently and more in accordance with the angle of repose theory of pressure, therefore the strain upon the arch would be the greatest, and its thrust must be counterbalanced at the sides, the strain upon the invert being probably very little and due to the tendency to an overturning movement of either of the straight walls. However, in culverts as in tunnels, it is impossible to say the exact amount or direction of the strain, although it may be approximately computed.

Should it be necessary to erect a culvert upon soft ground, as much of it as practicable should be excavated and a concrete foundation be placed thereon; it is also advisable to allow an extra length to that required by the calculated slope.

A culvert should have an invert unless upon a hard rock bed, and care should be taken that there shall be no leakage at the springing of the inverted arch or under it, or beneath the sidewalls at the level of the surface of a flat stone drain or below it.

The banks of a stream or watercourse should be inspected occasionally, especially on the up-stream side of a culvert, in order to note whether they are stable or crumbling away, as then the course of the stream may be widened or diverted, and so erode the toe of an embankment and cause a slip of earth. Splayed wing-walls are a protection against such a danger besides aiding the flow of water through them, and they also lessen the chance of a damming back of the water and undermining, as they increase the discharge; for instance, a splay of 53° has been found to increase the flow about 25 per centum. The surface of the material of which a culvert is constructed should be as smooth as practicable, so as to reduce the friction of water flowing past it, which, in the case of unplaned timber, cast and wrought ironwork, ashlar masonry, brickwork, and concrete, is about the same, but is some 30 per cent. more when the surface consists of rubble masonry.

When an embankment crosses a narrow valley, in which no watercourse exists, instead of a culvert, a bed of loose stones has been placed upon the ground under the whole area of the seat of an embankment at its greatest depth, a proper fall and bed being given to it.

The toe of the slope on the up-stream side should be protected from any wash, and the stone layer be carried a few feet beyond the foot of the slope on the lower side.

The following general principles it is well to remember in designing culverts.

When a culvert is of uniform section, which is almost invariably the case, it should have the same inclination throughout.

Avoid, or ease, all bends as much as possible.

Have splayed wing-walls.

Make provision against undermining by scour or percolation of water.

Have smooth and even surfaces so as to reduce the friction and increase the discharge.

Have an arched invert to a culvert, and a flat bed stone to all small surface drains, with complete connection to the side walls.

In some countries having a severe climate, or in high mountainous districts where the soil is rock and a heavy discharge of flood-waters occurs, instead of placing a culvert and gathering the waters at or about the level of the toe of an embankment, in deep hillsides or ravines an unlined tunnel is made under the embankment in the rock, thus avoiding a masonry or brickwork structure, which could only be set in the summer months, and preventing the waters touching the seat of an embankment and promoting a slip.

When the level of a rivulet allows, and the waters are simply surface discharge, the system has been used of making two open channels, one upon each side of a steep valley, thus retaining the waters, allowing the adoption of two short span open culverts and two channels instead of one large culvert at the deepest place, and saving expense, the original bed of watercourse being filled with the surplus excavation. The embankment consisting of broken rock or hard granular soil, any little percolation of water along the old bed will not deleteriously affect it, but will find a passage.