CHAPTER XVI.

=172. Application of Fitzgerald’s Results to the Croton Watershed.=—The evaporation data determined by Messrs. Fitzgerald and Kuichling are sufficient for all ordinary purposes in the North Atlantic States. In the discussion of the capacity of the Croton watershed Mr. Fitzgerald’s results will be taken, as the conditions of the Croton watershed in respect to temperature and atmosphere are affected by the proximity to the ocean, and other features of the case make it more nearly like the Metropolitan drainage area near Boston than the more elevated inland district near Rochester.

If the monthly amounts of evaporation be taken from the preceding table, and if it further be observed that a volume of water 1 square mile in area and 1 inch thick contains 17,377,536 gallons, the following table (Table XI) of amounts of evaporation from the reservoirs in the Croton watershed, including the new Croton Lake, will result, since the total area of water surface of all these reservoirs is 16.1 square miles.

TABLE XI.

Jan. 0.96 × 16.1 × 17,377,536 = 268,600,000 gallons. Feb. 1.05 × ” × ” = 293,800,000 ” Mar. 1.70 × ” × ” = 475,700,000 ” April 2.97 × ” × ” = 831,000,000 ” May 4.46 × ” × ” = 1,247,900,000 ” June 5.54 × ” × ” = 1,550,100,000 ” July 5.98 × ” × ” = 1,673,200,000 ” Aug. 5.50 × ” × ” = 1,538,900,000 ” Sept. 4.12 × ” × ” = 1,152,800,000 ” Oct. 3.16 × ” × ” = 884,200,000 ” Nov. 2.25 × ” × ” = 629,600,000 ” Dec. 1.51 × ” × ” = 422,500,000 ” ----- -------------- 39.20 Total = 10,968,300,000 ”

It will be seen from this table that the total annual evaporation from all the reservoir surfaces of the Croton watershed, as it will exist when the new Croton Lake is completed, will be nearly 11,000,000,000 gallons, enough to supply the boroughs of Bronx and Manhattan at the present rate of consumption for about forty days.

=173. The Capacity of the Croton Watershed.=—The use of the preceding figures and numbers can be well illustrated by considering the capacity of the Croton watershed in its relations to the present water needs of the boroughs of Bronx and Manhattan which that watershed is designed to supply. The total area of the Croton watershed is 360.4 square miles, of which 16.1 square miles, as has already been observed, is water surface. As a matter of fact the run-off observations from that watershed have been maintained or computed for the thirty-two-year period from 1868 to 1899, inclusive, covering the evaporation from the reservoirs and lake surfaces as they have existed during that period. The later observations, therefore, include the effects of evaporation from the more lately constructed reservoirs, but none of these data cover evaporation from the entire surface of the new Croton Lake, whose excess over that of the old reservoir is nearly one third of the total water surface of the entire shed. As a margin of safety and for the purpose of simplification, separate allowance will be made for the evaporation from all the reservoir and lake surfaces of the entire watershed as it will exist on the completion of the new Croton Lake, as a deduction from the run-off. The preceding table (Table XI) exhibits those deductions for evaporation as they will be made in the next table.

In Table IX the year 1880 yields the lowest run-off of the entire thirty-two-year period. The total precipitation was 36.92 inches, and only 34.21 per cent of it was available as run-off. The first column in Table XII gives the amount of monthly rainfall for the entire year, the sum of which aggregates 36.92 inches. Each of these monthly quantities multiplied by .3421 will give the amount of rainfall available for run-off, and the latter quantity multiplied by the number of square miles in the watershed (360.4) will show the total depth of available water concentrated upon a single square mile. If the latter quantity be multiplied by 17,378,000, the total number of gallons available for the entire month will result, from which must be subtracted the evaporation for the same month. Carrying out these operations for each month in the year, the monthly available quantities for water-supply will be found, as shown in the last column.

TABLE XII.

(Jan. 3.43 × .3421 = 1.173) × 360.4 × 17,378,000 - 268,600,000 = 7,077,700,000 (Feb. 3.40 × ” = 1.163) × ” × ” - 293,800,000 = 6,989,900,000 (Mar. 3.90 × ” = 1.334) × ” × ” - 475,700,000 = 7,879,000,000 (April 3.57 × ” = 1.221) × ” × ” - 831,000,000 = 6,816,000,000 (May 1.04 × ” = .356) × ” × ” - 1,247,900,000 = 982,000,000 (June 1.40 × ” = .479) × ” × ” - 1,550,100,000 = 1,449,800,000 (July 5.86 × ” = 2.005) × ” × ” - 1,673,200,000 = 10,890,000,000 (Aug. 4.16 × ” = 1.423) × ” × ” - 1,538,900,000 = 7,373,100,000 (Sept. 2.42 × ” = .828) × ” × ” - 1,152,800,000 = 4,032,900,000 (Oct. 2.83 × ” = .968) × ” × ” - 884,200,000 = 5,178,500,000 (Nov. 2.32 × ” = .794) × ” × ” - 629,600,000 = 4,343,100,000 (Dec. 2.59 × ” = .886) × ” × ” - 422,500,000 = 5,126,300,000 ----- 36.92

The sum of the twelve monthly available quantities will give the total number of gallons per year applicable to meeting the water demands of the boroughs of Bronx and Manhattan.

=174. Necessary Storage for New York Supply to Compensate for Deficiency.=—At the present time the average daily consumption per inhabitant of those two boroughs is 115 gallons, and if the total population be taken at 2,200,000, the total daily consumption will be 2,200,000 × 115 = 253,000,000 gallons. If the latter quantity be multiplied by 30.5, the latter being taken as the average number of days in the month throughout the year, the average monthly draft of water for the two boroughs in question will be 7,716,500,000 gallons. The subtraction of the latter quantity from the monthly results in the preceding table will exhibit a deficiency which must be met by storage or a surplus available for storage. Table XIII exhibits the twelve monthly differences of that character.

TABLE XIII.

7,077,700,000 — 7,716,500,000 = - 638,800,000 6,989,900,000 — ” = - 726,600,000 7,879,000,000 — ” = + 162,500,000 6,816,000,000 — ” = - 900,500,000 982,000,000 — ” = - 6,734,500,000 1,449,800,000 — ” = - 6,266,700,000 10,890,000,000 — ” = +3,173,500,000 7,373,100,000 — ” = - 343,400,000 4,032,900,000 — ” = - 3,683,600,000 5,178,500,000 — ” = - 2,538,000,000 4,343,100,000 — ” = - 3,373,400,000 5,126,300,000 — ” = - 2,590,200,000 --------------- --------------- -27,795,700,000 +3,336,000,000 + 3,336,000,000 --------------- -24,459,700,000

It is seen from this table that the total monthly deficiencies aggregate 27,795,700,000 gallons and that there are only two months in which the run-off exceeds the consumption, the surplus for those two months being only 3,336,000,000 gallons. The total deficiency for the year is therefore 24,459,700,000 gallons. Dividing the latter quantity by the average daily draft of 253,000,000 gallons, there will result a period of 97 days, or more than one quarter of a year, during which the minimum annual rainfall would fail to supply any water to the city at all. These results show that in case of a low rainfall year, like that of 1880, the precipitation upon the Croton watershed would supply sufficient water for the boroughs of Bronx and Manhattan at the present rate of consumption for three fourths of the year only. A distressingly serious water famine would result unless the year were begun by sufficient available storage in the reservoirs of the basin at least equal to 24,459,700,000 gallons. Should such a low rainfall year or one nearly approaching it be one of a two- or three-year low rainfall cycle, such a reserve storage would be impossible and the resulting conditions would be most serious for the city. If an average year, for which the total rainfall would be about 48 inches preceded such a year of low rainfall, the conditions would be less serious. The figures would stand as follows:

Total run-off = 17,377,536 × 360.4 × 22.93 - 17,377,536 × 16.1 × 39.2 = 132,640,000,000 gallons.

Total annual consumption = 92,345,000,000 ” --------------- Available for storage = 40,295,000,000 ” Deficiency = 24,459,700,000 ” --------------- Surplus = 15,835,300,000 ”

The average year would, therefore, yield enough run-off water if stored to more than make up the deficiency of the least rainfall year by nearly 16,000,000,000 gallons. In order to secure the desired volume it would therefore be necessary to have storage capacity at least equal to 24,459,700,000 gallons; indeed, in order to meet all the exigencies of a public water-supply it would be necessary to have far more than that amount. As a matter of fact there are in the Croton watershed seven artificial reservoirs with a total storage capacity of nearly 41,000,000,000 gallons, besides a number of small ponds in addition to the new Croton Lake which with water surface at the masonry crest of the dam has a total additional storage capacity of 23,700,000,000 gallons. The storage capacity of the new Croton Lake may be increased by the use of flash-boards 4 feet high placed along its crest, so that with its water surface at grade 200 its total capacity will be increased to 26,500,000,000 gallons. After the new Croton reservoir is in use the total storage capacity of all the reservoirs and ponds in the Croton watershed will be raised to 70,245,000,000 gallons, which can be further augmented by the Jerome Park reservoir when completed by an amount equal to 1,900,000,000 gallons. This is equivalent, at the present rate of consumption, to a storage supply for 285 days for the boroughs of Manhattan and the Bronx.

=175. No Exact Rule for Storage Capacity.=—This question of the amount of storage capacity to be provided in connection with public water-supplies is one which cannot be reduced to an exact rule. Obviously if the continuous flow afforded from any source is always greater per day than any draft that can ever be made upon it, no storage-reservoirs at all would be needed, although they might be necessary for the purpose of sedimentation. On the other hand, as in the case of New York City, if the demand upon the supply has reached its capacity or exceeded it for low rainfall years, it may be necessary to provide storage capacity sufficient to collect all the run-off of the watershed. The civil engineer must from his experience and from the data before him determine what capacity between those limits is to be secured. When the question of volume or capacity of storage is settled the mode of distribution of that volume or capacity in reservoirs is to be determined, and that affects to some extent the potability of the water. If there is a large area of shallow storage, the vegetable matter of the soil may affect the water in a number of ways. Again, it is advisable in this connection to consider certain reservoir effects as to color and contained organic matter in general.

=176. The Color of Water.=—The potability[7] of water collected from any watershed is materially affected by its color. Although iron may produce a brownish tinge, by far the greater amount of color is produced by dissolved vegetable matter. Repeated examinations of colored water have shown that discoloration is in many cases at least a measure of the vegetable matter contained in it. While this may not indicate that the water is materially unwholesome, it shows conclusively the existence of conditions which are usually productive of minute lower forms of vegetation from which both bad taste and odors are likely to arise.

[7] What is generally known as the “Michigan standard of the purity of drinking-water,” as specified by the Michigan State Laboratory of Hygiene, is here given:

“1. The total residue should not exceed 500 parts per million. “2. The inorganic residue may constitute the total residue. “3. The smaller amount of organic residue the better the water. “4. The amount of earthy bases should not exceed 200 parts per million. “5. The amount of sodium chloride should not exceed 20 parts per million (i.e., ‘chlorine’ 12.1 parts per million). “6. The amount of sulphates should not exceed 100 parts per million. “7. The organic matter in 1,000,000 parts of the water should not reduce more than 8 parts of potassium permanganate (i.e., ‘required oxygen’ 2.2 parts per million). “8. The amount of free ammonia should not exceed 0.05 part per million. “9. The amount of albuminoid ammonia should not exceed 0.15 part per million. “10. The amount of nitric acid should not exceed 3.5 parts per million (i.e., ‘N as nitrate’ .9 part per million). “11. The best water contains no nitrous acid, and any water which contains this substance in quantity sufficient to be estimated should not be regarded as a safe drinking-water. “12. The water must contain no toxicogenic germs as demonstrated by tests upon animals.

“The water must be clear and transparent, free from smell, and without either alkaline or acid taste, and not above 5 French standard of hardness.”

This standard is too high to be attained ordinarily in natural waters.

There are two periods in the year of maximum intensity of color, one occurring in June and the other in November. The former is due to the abundant drainage of peaty or other excessively vegetable soils from the spring rains. After June the sun bleaches the water to a material extent until the autumn, when the dying vegetation imparts more or less coloring to the water falling upon it. This last agency produces its maximum effect in the month of November.

There are various arbitrary scales employed by which colors may be measured and discolored waters compared. Among others, dilute solutions of platinum and cobalt are used, in which the relative proportions of those substances are varied so as to resemble closely the colors of the water. The amount of platinum used is a measure of the color, one unit of which corresponds to one part of the metal in 10,000 parts of water. Again, the depth at which a platinum wire 1 mm. (.039 inch) in diameter and 1 inch long can be seen in the water is also taken as a measure of the color, the amount of the latter being inversely as the depth. This method has found extended and satisfactory use in connection with the Metropolitan Water-supply of Boston, the Cochituate water having a degree of color represented by .25 to .30, while the Sudbury water has somewhat more than twice as much. The Cochituate water is practically colorless.

The origin of the color of water is chiefly the swamps which drain into the water-supply, or the vegetation remaining upon a new reservoir site when the surface soil has not been removed before the filling of the reservoir. The drainage of swamps should not, as a rule, be permitted to flow into a public water-supply, as it is naturally heavily charged with vegetable matter and is correspondingly discolored. This matter, like many others connected with the sanitation of potable public waters, has been most carefully investigated by the State Board of Health of Massachusetts in connection with the Boston water-supply. Its work has shown the strong advisability of diverting the drainage of large swamps from a public supply as carrying too much vegetable matter even when highly diluted by clear water conforming to desirable sanitary standards.

=177. Stripping Reservoir Sites.=—The question of stripping or cleaning reservoir sites of soil is also one which has been carefully studied by the Massachusetts State Board of Health. As a consequence large amounts of money have been expended by the city of Boston in stripping the soil from reservoir sites to the average depth in some cases of 9 inches for wooded land and 12½ inches for meadow land. This was done in the case of the Nashua River reservoir having a superficial area of 6.56 square miles at a cost of nearly $2,910,000, or about $700 per acre. It has been found that the beneficial effect of this stripping is fully secured if the black loam in which vegetation flourishes is removed.

[Illustration: Wachusetts Reservoir, showing Stripping of Soil.]

This stripping of soil is indicative of the great care taken to secure a high quality of water for the city of Boston, but it is not done in the Croton watershed of the New York supply. It cannot be doubted that the quality of the Croton supply would have been sensibly enhanced by a similar treatment of its reservoir sites. Mr. F. B. Stearns, chief engineer of the Metropolitan Water-supply of Boston, states that in some cases the effects of filling reservoirs without removing the soil and vegetable matter have “continued for twenty years or more without apparent diminution.” On the other hand, water discolored by vegetable matter becomes bleached to some extent at least by standing in reservoirs whose sites have been stripped of soil.

=178. Average Depth of Reservoirs should be as Great as Practicable.=—In the selection of reservoir locations those are preferable where the average depths will be greatest and where shallow margins are reduced to a minimum. It may sometimes be necessary to excavate marginal portions which would otherwise be shallow with a full reservoir. There should be as little water as possible of a less low-water depth than 10 or 12 feet, otherwise there may be a tendency to aquatic vegetable growth. The following table exhibits the areas, average depths, capacity, and other features of a number of prominent storage-reservoirs.

COMPARATIVE TABLE OF AREAS, DEPTHS, AND CAPACITIES OF STORAGE RESERVOIRS WITH HEIGHTS AND LENGTHS OF DAMS.

LEGEND: (A) = Area Square Miles. (B) = Average Depth, Feet. (C) = Length of Dam, Feet. (D) = Capacity, Million Gallons. ------------------------------+-----+---+--------------+-----+------- | | |Maximum Height| | | | | of Dam. | | Name and Location | | +-------+------+ | of Reservoir. | (A) |(B)| Above | Above| (C) | (D) | | |Ground.| Rock.| | ------------------------------+-----+---+-------+------+-----+------- Swift River, Mass |36.96| 53| 144 | ... |2,470|406,000 Nashua River, Mass | 6.56| 46| 129 | 158 |1,250| 63,068 Nira, near Poona, India | 7.25| 27| 100 | ... |3,000| 41,143 Tansa, Bombay, India | 5.50| 33| 127 | 131 |8,770| 37,500 Khadakvasla, Poona, India | 5.50| 32| 100 | 107 |5,080| 36,737 New Croton, N. Y. | ....| ..| 157 | 225 |1,270| 32,000 Elan and Claerwen, Birmingham,| | | | | | Eng., water-works | | | | | | (total for six reservoirs) | 2.34| 43|98-128 | ... |4,460| 20,838 All Boston water-works | | | | | | reservoirs combined | 5.82| 14|14-65 | ... |.....| 15,867 Vyrnwy, Liverpool, Eng. | 1.75| ..| 84 | 129 |1,350| 14,560 Ware River, Mass. | 1.62| 33| 71 | ... | 785| 11,190 Sodom, N. Y. | ....| ..| 72 | 89 | 500| 9,500 Reservoir No. 5, Boston | | | | | | water-works | 1.91| 19| 65 | 70 |1,865| 7,438 Titicus, N. Y. | ....| ..| 105 | 115 |.....| 7,000 Hobbs Brook, Cambridge | | | | | | water-works | 1.00| 12| 23 | ... |.....| 2,500 Cochituate, Boston water-works| 1.35| 8| .. | ... |.....| 2,160 Reservoir No. 6, Boston | | | | | | water-works | 0.29| 25| 52 | ... |1,500| 1,500 ------------------------------+-----+---+-------+------+-----+-------

=179. Overturn of Contents of Reservoirs Due to Seasonal Changes of Temperature.=—It will be noticed that the average depth is less than about 20 feet in few cases only. If the water is deep, its mean temperature throughout the year will be lower than if shallow. During the warmer portion of the year the upper layers of the water are obviously of a higher temperature than the lower portions, since the latter receive much less immediate effect from the sun’s rays. As the upper portions of the water are of higher temperature, they are also lighter and hence remain at or near the top. For the same reason the water at the bottom of the reservoir remains there throughout the warm season and until the cool weather of the autumn begins. The top layers of water then continue to fall in temperature until it is lower than that of the water at the bottom, when the surface-water becomes the heaviest and sinks. It displaces subsurface water lighter than itself, the latter coming to the surface to be cooled in turn.

This operation produces a complete overturning of the entire reservoir volume as the late autumn or early winter approaches. It thus brings to the surface-water which has been lying at the bottom of the reservoir all summer in contact with what vegetable matter may have been there. The depleted oxygen of the bottom water is thus replenished with a corresponding betterment of condition. It is the great sanitary effort of nature to improve the quality of stored water entrusted to its care, and it continues until the surface is cooled to a temperature perhaps lower than that of the greatest density of water.

Another great turn-over in the water of a lake or reservoir covered with ice during the winter occurs in the spring. When the ice melts, the resulting water rises a little in temperature until it reaches possibly its greatest density at 39°.2 Fahr., and then sinks, displacing subsurface water. This goes on until all the ice is melted and until all water cooled by it, near the surface, below 39°.2 Fahr. has been raised to that temperature. The period of summer stagnation then follows.

=180. The Construction of Reservoirs.=—The natural topography and sometimes the geology of the locality determines the location of the reservoir. The first requirement obviously is tightness. If for any reason whatever, such as leaky banks or bottom, porous subsurface material, or for any other defect, the water cannot be retained in the reservoir, it is useless. Some very perplexing questions in this connection have arisen. Indeed reservoirs have been completed only to be found incapable of holding their contents. Such results are evidently not creditable to the engineers who are responsible for them, and they should be avoided.

[Illustration: YARROW RESERVOIR, LIVERPOOL WATER-SUPPLY]

[Illustration: SAN LEANDRO DAM, SAN FRANCISCO WATER-WORKS]

[Illustration: TITICUS DAM, NEW YORK WATER-SUPPLY]

In order that the bottom of the reservoir may be water-tight it must be so well supported by firm underlying material that it will not be injured by the weight of water above it, which in artificial reservoirs may reach 30 to 100 feet or more in depth. The subsurface material at the site of any proposed structure of this character must, therefore, be carefully examined so as to avoid all porous material, crevasses in rocks, or other open places where water might escape. Objectionable material may frequently be removed and replaced with that which is more suitable, and rock crevices and other open places may sometimes be filled with concrete and made satisfactory. Whatever may be the conditions existing, the finished bottom of the reservoir should be placed only on well-compacted, firm, unyielding material.

The character of the reservoir bottom will depend somewhat upon the cost of suitable material of which to construct it. If a bottom of natural earth cannot be used, a pavement of stone, brick, or concrete may be employed from 8 inches to a foot or a foot and a half in thickness. The reservoir banks must be placed upon carefully prepared foundations, sometimes with masonry core-walls. They are frequently composed of clayey and gravelly material mixed in proper proportions and called puddle, although that term is more generally applied to a mixture of clay and gravel designed to form a truly impervious wall in the centre of the reservoir embankment. Some engineers require the core-wall, as it is called, to be constructed of masonry, with the earth or gravelly material carried up each side of this wall in layers 6 to 9 inches thick, well moistened and each layer thoroughly rolled with a grooved roller, or treated in some equivalent manner in order that the whole mass may not be in strata but essentially continuous and as nearly impervious as possible. The masonry core-wall should be founded on bed-rock or its equivalent. Its thickness will depend upon the height of the embankment. If the latter is not more than 20 or 25 feet high, the core-wall need not be more than 4 to 6 feet thick, but if the embankment reaches a height of 75 feet or even 100 feet, it must be made 15 to 20 feet thick, or possibly more, at the base. Its top should be not less than 4 or 6 feet thick, imbedded in the earth and carried well above the highest surface of water in the reservoir.

The thickness of the clay puddle-wall employed as the central core of the reservoir embankment is usually made much thicker than that of masonry. As a rough rule it may be made twice as thick as the masonry core at the deepest point and not less than about 6 feet at the top. The thickness of the puddle core is sometimes varied to meet the requirements of the natural material in which it is embedded at different depths.

Frequently, when embankments are under about 20 feet high, the core-walls may be omitted, excavation having been made at the base of the embankment down to rock or other impervious material, and if the entire bank is carried up with well-selected and puddled material.

The interior slopes of reservoir embankments are usually covered with roughly dressed stone pavement 12 to 18 inches thick, laid upon a broken stone foundation 8 to 12 inches thick, for a protection against the wash of waves, the pavement in any case being placed upon the bank slope after having been thoroughly and firmly compacted. The sloping and bottom pavements, of whatever material they may be composed, should be made continuous with each other so as to offer no escape for the water. In some cases where it has been found difficult to make the interior surfaces of reservoirs water-tight, asphalt or other similar water-tight layers have been used with excellent results.

The care necessary to be exercised in the construction of storage or other reservoirs when earth dams or embankments are used can better be appreciated when it is realized that almost all such banks, even when properly provided with masonry or clay-puddle core-walls, are saturated with water, even on the down-stream side, at least throughout their lower portions. A board of engineers appointed by the commissioners of the Croton Aqueduct in the summer of 1901 made a large number of examinations in the earth embankments in the Croton watershed, and found that with scarcely an exception those embankments were saturated throughout the lower portions of their masses, although in every case a masonry core-wall had been built. The results of those investigations showed that the water had percolated through the earth portion of the embankments and even through the core-walls, which had been carried down to bed-rock. This induced saturation, more or less, of the material on the down-stream slopes of the embankments. When material is thus filled with water, unless it is suitably selected, it is apt to become soft and unstable, so that any superincumbent weight resting upon it might produce failure. The fact that such embankments may become saturated with water fixes limits to their heights, since the surface of saturation in the interior of the bank has generally a flatter slope than that of the exterior surface. The height of the embankment therefore should be such that the exterior slope cannot cut into the saturated material at its foot, at least to any great extent. From what precedes it is evident that the height of an earth embankment will depend largely upon the slope of the exterior surface. This slope is made 1 vertical to 2, 2½, or 3 horizontal. The more gradual slope is sometimes preferable. It is advisable also to introduce terraces and to encourage the growth of sod so as to protect the surface from wash. The inner paved slope may be as steep as 1 vertical to 1½ or 2 horizontal.

[Illustration: BOG BROOK DAM NO. 1.—RESERVOIR 1.]

[Illustration: TITICUS DAM.—RESERVOIR M.]

[Illustration: AMAWALK DAM.—RESERVOIR A.

Earth Dams in Croton Watershed, showing Slopes of Saturation.]

=181. Gate-houses, and Pipe-lines in Embankments.=—It is necessary to construct the requisite pipe-lines and conduits leading from the storage-reservoirs to the points of consumption, and sometimes such lines bring the water to the reservoir. Wherever such pipes-line or conduits either enter or leave a reservoir gates and valves must be provided so as properly to control the admission and outflow of the water. These gate-houses, as they are called, because they contain the gates or valves and such other appurtenances or details as are requisite for operation and maintenance, are usually built of substantial masonry. They are the special outward features of every reservoir construction, and their architecture should be characteristic and suitable to the functions which they perform. Where the pipes are carried through embankments it is necessary to use special precautions to prevent the water from flowing along their exterior surfaces. Many reservoirs have been constructed under defective design in this respect, and their embankments have failed. Frequently small masonry walls are built around the pipes and imbedded in the bank, so as to form stops for any initial streams of water that might find their way along the pipe. In short, every care and resource known to the civil engineer must be employed in reservoir construction to make its bottom and its banks proof against leakage and to secure permanence and stability in every feature.

=182. High Masonry Dams.=—The greatest depths of water impounded in reservoirs are found usually where it is necessary to construct a high dam across the course of a river, as at the new Croton dam. In such cases it is not uncommon to require a dam over 75 to 100 feet high above the original bed of the river, which is usually constructed of masonry with foundations carried down to bed-rock in order to secure suitable stability and prevent flow or leakage beneath the structure. It is necessary to secure that result not only along the foundation-bed of the dam, but around its ends, and special care is taken in those portions of the work.

The new Croton dam is the highest masonry structure of its class yet built. The crest of its masonry overflow-weir is 149 feet above the original river-bed, with the extreme top of the masonry work of the remaining portion of the dam carried 14 feet higher. A depth of earth and rock excavation of 131 feet below the river-bed was necessary in order to secure a suitable foundation on bed-rock. The total maximum height, therefore, of the new Croton dam, from the lowest foundation-point to the extreme top, is 294 feet, and the depth of water at the up-stream face of the dam will be 136 feet when the overflow is just beginning, or 140 feet if 4 feet additional head be secured by the use of flash-boards. In the prosecution of this class of work it is necessary not only to reach bed-rock, but to remove all soft portions of it down to sound hard material, to clean out all crevices and fissures of sensible size, refilling them with hydraulic cement mortar or concrete, and to shape the exposed rock surfaces so as to make them at least approximately normal to the resultant loads upon them, to secure a complete and as nearly as possible water-tight bond with the superimposed masonry. If any streams or other small watercourses should be encountered, they must either be stopped or led off where they will not affect the work, or, as is sometimes done, the water issuing from them may be carried safely through the masonry mass in small pipes. The object is to keep as much water out of the foundation-bed as possible, so as to eliminate upward pressure underneath the dam caused by the head of water in the subsequently full reservoir. It is a question how much dependence can be placed upon the exclusion of water from the foundation-bed. In the best class of work undoubtedly the bond can be good enough to exclude more or less water, but it is probably only safe and prudent so to design the dam as to be stable even though water be not fully excluded.

[Illustration: Cross-section of New Croton Dam.]

The stability of the masonry dam must be secured both for the reservoir full and empty. With a full reservoir the horizontal pressure of water on the up-stream face tends to overturn the dam down-stream. When the water is entirely withdrawn the pressure under the up-stream edge of the foundation becomes much greater, so that safety and stability under both extreme conditions must be assured. There are a number of systems of computation to which engineers resort in order to secure a design which shall certainly be stable under all conditions. That which is commonly employed in this country is based upon two fundamental propositions, under one of which the pressure at any point in the entire masonry mass must not exceed a certain safe amount per square foot, while the other is of a more technical character, requiring that the centre of pressure shall, in every horizontal plane of the dam, approach nowhere nearer than one third the horizontal thickness of the masonry to one edge of it. A further condition is also prescribed which prevents any portion of the dam from slipping or sliding over that below it. As a matter of fact when the first two conditions are assured the third is usually fulfilled concurrently. Obviously there will be great advantage accruing to a dam if the entire mass of masonry is essentially monolithic. In order that that may be the case either concrete or rubble is usually employed for the great mass of the masonry structure, the exterior surfaces frequently being composed of a shell of cut-stone, so as to provide a neat and tasteful finish. This exterior skin or layer of cut masonry need not average more than 1½ to 2½ feet thick.

The pressures prescribed for safety in the construction of masonry dams vary from about 16,000 to 28,000 or 30,000 pounds per square foot. Sometimes, as in the masonry dams found in the Croton watershed, limits of 16,000 to 20,000 pounds per square foot are prescribed for the upper portions of the dams and a gradually increasing pressure up to 30,000 pounds per square foot in passing downward to the foundation-bed. There are reasons of a purely technical character why the prescribed safe working pressure must be taken less on the down-stream or front side of the dam than on the up-stream or rear face.

The section of a masonry dam designed under the conditions outlined will secure stability through the weight of the structure alone, hence it is called a gravity section. In some cases the rock bed and sides of a ravine in which the stream must be dammed will permit a curved structure to be built, the curvature being so placed as to be convex up-stream or against the water pressure. In such a case the dam really becomes a horizontal arch and, if the curvature is sufficiently sharp, it may be designed as an arch horizontally pressed. The cross-section then has much less thickness (and hence less area) than if designed on a straight line so as to produce a gravity section. A number of such dams have been built, and one very remarkable example of its kind is the Bear Valley dam in California; it was built as a part of the irrigation system.

[Illustration: Foundation Masonry of New Croton Dam.]