i
 
 m^8 UBRARY
 
 THE 
 
 DESIGNING AND CONSTRUCTION 
 
 07 
 
 Storage Reservoirs. 
 
 BY 
 
 ARTHUR JACOB, B. A., 
 
 LATE EXECUTIVE ENGINEER FOR IRRIGATION H. M. 
 BOMBAY SERVICE. 
 
 REVISED AND EXTENDED BY 
 
 E. SHERMAN GOULD, M. Am. Soo. C. E., 
 
 CONSULTING ENGINEER TO THE SCRANTON GAS AND 
 WATER CO. 
 
 NEW YORK : 
 
 D. VAN NOSTRAND, PUBLISHER, 
 
 23 Murray and 27 Warren Street. 
 
 1888.
 
 John-'W-Oooam^^. 
 
 'ECEHBLR 27? 18^4 
 
 JPKIA'iF/LLP, MASS. - 
 
 EDITOR'S PREFACE. 
 
 In preparing a revised edition of 
 *' Jacob's Storage Reservoirs," it was con- 
 sidered best, in view of the standard 
 character of the work, to present it in 
 its original form, with all additional 
 matter separate from the text. 
 
 All of the editorial work, therefore, in 
 the present volume, appears in the shape 
 of foot-notes, and the additional pages 
 appended to the original discussion. 
 
 E. s. a 
 
 ScRANTON, Pa., January, 1888.
 
 THE DESIGNING AND CONSTRUCTION 
 
 OF 
 
 STORAGE RESERVOIRS. 
 
 Before entering upon such considera- 
 tions as affect the selection of reservoir 
 sites and their construction, a brief allu- 
 sion to some of the most ancient works 
 for impounding water may not be unin- 
 teresting. Of these the most prominent 
 examples are undoubtedly to be found 
 in Hindostan, where the magnitude and 
 antiquity of the storage works cannot fail 
 to arrest attention. These great works, 
 surpassing in their immensity what are 
 conventionally esteemed to be the won- 
 ders of the world, the production of other 
 countries and nations, took their origin 
 in the necessities of the people and the
 
 6 
 
 variableness of the climate of India, and 
 were, in fact, great public works on which 
 the welfare of the people mainly de- 
 pended. The climate of India, although 
 singularly uniform in some respects from 
 year to year, is remarkably variable as 
 regards the rain-fall ; and in order to 
 guard against the disasters of famine and 
 sickness, inevitably attendant on a scanty 
 monsoon, the native princes were wont 
 to make such provisions as large re- 
 sources and an almost unlimited power 
 enabled them, in order to obviate the diffi- 
 culty that they had to contend with. 
 
 The rain records of India for several 
 years past show that a scarcity of rain is 
 indicated by periods of about five years, 
 or that every fifth or sixth year is marked 
 by a scanty rainfall over certain districts. 
 The recurrence of these periods is, of 
 course, not very clearly marked, but still 
 it is sufficiently so to warrant, with ap- 
 proximate correctness, the jDrediction of 
 scarcity and famine ; and such deplorable 
 recurrence is, as all are aware, now reign 
 ing in India, and visiting with destruc-
 
 tion, by sickness and hunger, some 
 thousands whose sole dependence is upon 
 a fair season of rain, and the successful 
 maturing of their Httle crop of grain. 
 
 The natural expedient for guarding 
 against the recurrence of these periodical 
 calamities was evidently to be found in 
 husbanding a scanty supply of rain-water 
 for the purpose of irrigation, and this 
 the people of India appear to have un- 
 derstood. They took advantage, in cer- 
 tain districts, of every nook and ravine, 
 whether large or small, and converted 
 them into storage reservoirs by throwing 
 across banks of earth, or bunds, as they 
 are termed, producing, in certain dis- 
 tricts, such an elaborate and complete 
 system of irrigation as can only be com- 
 pared, for cost and completeness, to our 
 railway system in England. Taking four- 
 teen districts in the Madras Presidency, 
 where tank irrigation was most generally 
 relied upon, the records of the Indian 
 Government show that there are no less 
 than 43,000 irrigation reservoirs now in 
 effective operation, and as many as 10,000
 
 8 
 
 more that have fallen into disuse, making 
 a total number of 53,000 storage works. 
 The average length of embankment is 
 found to be about half a mile, the ex- 
 treme limit of the series being a dam of 
 the immense length of 30 miles. This 
 ancient reservoir, called the Poniary 
 tank, is no longer in use, the cost of 
 maintaining such a length of bank in 
 adequate repair having probably been 
 found disproportionate to the advantage 
 derived from the supply. The work, em- 
 bracing an area of storage of between 60 
 and 80 sq. miles, remains, however, as a 
 record of what the Hindoos are capable 
 of. To quote a second example, there is 
 the Veranum reservoir, now in actual 
 operation as a source of supply, and 
 yielding a net revenue of no less than 
 £11,450 per annum. The area of the 
 tank is 35 sq. miles, and the storage is 
 effected by a dam 12 miles in length. 
 In order to bring the immensity of 
 this system of storage works within 
 the reach of statistical minds, it has been 
 calculated that the embankments contain
 
 9 
 
 as much earth as would serve to encircle 
 the globe with a belt of 6 ft. in thickness. 
 To show that these are not singular ex- 
 amiDles, one other embankment of re- 
 markable size may be alluded to. This 
 embankment, of somewhat smgular con- 
 struction, was built on the island of 
 Ceylon, and bears testimony that the 
 Singalesemonarchs were not behind their 
 neighbors in public spirit or enterprise. 
 The embankment was composed of huge 
 blocks of stone strongly cemented to- 
 gether, and covered over with turf, a 
 solid barrier of 15 miles in length, 100 ft. 
 wide at base, sloping to a tojD width of 
 40 ft., and extending across the lower 
 end of a spacious valley. 
 
 Thus it will appear that the practice of 
 embanking across valleys, for the purpose 
 of retaining the surface water, has for 
 ages been in operation. There is no 
 doubt that the disposal of some of the 
 most remarkable works in India is not 
 what it might, with advantage, have 
 been ; the fact remains, however, that 
 the desired end was attained, and if the
 
 10 
 
 earthworks were disproportionately ex- 
 tensive, it was a source of satisfaction, at 
 least, for the projectors to know that they 
 cost, as a general rule little or nothing, 
 the practice in those days being to press 
 whatever labor was required, rendering 
 in return nominal wages or none at all. 
 
 The two main questions that it is pro- 
 posed to submit for consideration are, 
 first, the selection of a reservoir site ; 
 and, secondly, the leading principles to 
 be observed in the designing and con- 
 struction of storage works. 
 
 The purpose or purposes for which the 
 work may be required will, of course, af- 
 fect materially the choice of a position, as 
 well as the details of the structure itself ; 
 but certain general principles are avail- 
 able for our guidance in every case, after 
 considering which, it is proposed to dwell 
 upon such points as apply to the special 
 purposes for which reservoirs may be con- 
 structed. 
 
 The first and most essential point for 
 accurate determination by the engineer is, 
 undoubtedly, the amount of rainfall, both
 
 11 
 
 maximum and minimum, that may be ex- 
 pected in the district under examination ; 
 and, having arrived at rehable data on 
 this point, the next consideration will 
 obviously be, what amount may be made 
 available, due allowance having been 
 made for evaporation and absorption. 
 When we know that the annual depth of 
 rainfall taken all over the world varies, 
 according to the locality, between zero 
 and 338 in. or 28 ft. deep (which ex- 
 cessive amount was on one occasion 
 registered in the hill district of Western 
 India), it will be obvious how little 
 ground there will be for assumption, in 
 the examination of any district hitherto 
 unexplored, with regard to the question 
 of its rainfall. In the examination of any 
 given country, however, there are certain 
 phenomena connected with the rainfall 
 that will be found of almost invariable 
 acceptation, and may with advantage be 
 borne in mind. 
 
 The rainfall will, as a general rule, be 
 greatest in those districts that are situ- 
 ated towards the point from which tbe
 
 12 
 
 prevailing winds blow. If Great Britain 
 for instance be taken, the western dis- 
 tricts will be found the most rainy. 
 The very reverse, however, of this 
 phenomena is noticed in the neighbor- 
 hood of mountain ranges. If the wind 
 prevails from one side rather than from 
 the other, it is found that the greatest 
 rainfall is on the leeward side of the 
 range, and the probable solution of the 
 matter is, that the air, highly charged with 
 moisture, is carried up the hills by the 
 wind until it comes into a cold region 
 of the atmosphere. Condensation of 
 the watery vapor immediately takes 
 place, and the result is a fall of rain on 
 the side of the mountain range remote 
 from the prevailing wind. 
 
 To this cause may also be attributed 
 the fact that the rainfall is always greatest 
 in mountainous districts, while it by no 
 means follows that elevated plains are 
 more abundantly supplied with rain than 
 land lying nearer to the sea level. The 
 principles are remarkably exemplified in 
 the southern part of the Bombay Presi-
 
 13 
 
 dency, where the author has had occasion 
 to study the subject of rainfall. The 
 Western Ghauts run parallel to the coast, 
 rising to a height of 4,500 ft. above the 
 sea, and from the western support of the 
 great table-land of the Deccan, the mean 
 elevation of which may be taken at 3,000 
 ft. In the rainy season the southwest 
 monsoon, blowing from the sea, impinges 
 agaiust the Ghauts, and while passing on- 
 wards to the Deccan, parts with its 
 moisture to the average annual amount 
 of 254 in. On a spur of mountain that 
 runs eastward, the pluviometers are found 
 to register but 50 in.; and about 40 
 miles farther inland the rainfall is not 
 more in some places than 15 in., which is 
 considerably less than that registered in 
 the lower-lying districts of the Presi- 
 dency.* 
 
 *In regard to this matter, Mr. Fanning says in his 
 "Water Supply Engineering":— "A study of some of 
 our principal river valleys independently, reveals 
 the fact that their rainfall gradually decreases from 
 their outlets to their more elevated sources." And 
 again: " The oft-made statement, 'that rain falls most 
 abundantly on the high land,' is applicable, in the 
 United States, to subordinate water-sheds only, and in 
 rare instances."
 
 14 
 
 In civilized countries like our own, 
 much valuable information is usually- 
 available regarding the rainfall, if not ap- 
 plying actually to the district under ex- 
 amination, then probably to some neigh- 
 boring districts, enjoying the same 
 physical characteristics; but when any 
 project of great importance is in contem- 
 plation, it will not be sufficient to take 
 the returns of adjoining districts as 
 accurate information of the rainfall at the 
 exact locality fixed upon for the construc- 
 tion of the works. It will be necessary 
 to establish rain-gauges at different points 
 over the catchment basin of the valley 
 from which it is intended to obtain the 
 supply ; and daily observations of these 
 gauges must be taken for comparison with 
 a series of simultaneous observations 
 taken and recorded at the nearest station 
 at which the rainfall has been regularly 
 and carefully noted. It is evident that 
 a comparison of the several observations 
 taken over the area of water- shed with 
 those registered at the permanent station, 
 will convey a just estimate] of the amount
 
 15 
 
 of maximum and minimum rainfall that 
 may be relied upon. 
 
 The amount of rain falUng upon the 
 ground is not, however, the point to be 
 determined, though it will aid consider- 
 ably as a guide to the engineer. A con- 
 siderable quantity of all the rainfall is 
 either absorbed by the ground or evapor- 
 ated before it reaches the point at which 
 it can be made available for storage. 
 Regarding, then, first, the question of 
 absorption, it must be apparent that no 
 two districts, unless they are exactly 
 identical in soil, inclination of sui'face, 
 and under similar circumstances of culti- 
 vation, can give on examination the same 
 comparative result of rainfall and evapo- 
 ration. If one district or unit of area be 
 similar to the other in all respects but 
 the surface incUnation, that which has 
 the greatest slope will, as a rule, give the 
 largest percentage of water available for 
 storage, because of course there will be 
 less time for the rain to be absorbed. 
 Again, the degree of cultivation will 
 materially afiect the result when two 
 areas, otherwise precisely similar in
 
 16 
 
 their physical conformation, come to be 
 compared one with the other, it being 
 evident that an open and well-drained 
 soil will be more favorable to the reten- 
 tion of water falling upon it than com- 
 pact and impervious land. In every case 
 the physical features of a district will 
 each and every one of them force itself 
 on the attention, as influencing the con- 
 clusion to be arrived at. If any general 
 rule can be applied, it may be said that 
 the greater the slope of the valley, the 
 more rapidly it will throw surface water 
 off ; the more denuded the surface is of 
 soil of any kind, the less will the escape 
 of rain-water be retarded ; and the more 
 compact the rocks composing the geologi- 
 cal structure of a district, the better will 
 the circumstances be for impounding 
 water. The volcanic rocks and those of 
 the granite order will be as favorable as 
 any that could be desired ; while, on the 
 other hand, porous rocks, such as the 
 sandstones, chalk, etc., are too absorbent 
 to offer the desired conditions for storage. 
 It is not here asserted that all the water 
 absorbed by porous rocks is necessarily
 
 17 
 
 intercepted from passing away to con- 
 tribute to storage supply ; much of it 
 may be lost by evaporation and absorp- 
 tion by vegetables, but a considerable 
 portion will often be found to contribute 
 in the form of springs, if the disposition 
 of the strata be favorable. 
 
 As a further source of loss, evapora- 
 tion from the ground as well as from the 
 surface of the reservoir, must be taken 
 into consideration. The circumstances 
 attending the latter source of loss will be 
 considered further on, as this does not 
 affect the question of how much of the 
 total rainfall may be made available.* 
 
 The question how much water will be 
 evaporated at any moment from the sur- 
 face of land is one involved in consider- 
 able difficulty ; and so many disturbing 
 elements enter into the solution of the 
 problem, that its accurate determination 
 
 *It is customary, in this country, when estimating 
 the available areas of water-sheds, to deduct the 
 areas of all lakes, ponds and open bodies of water, as 
 the evaporation from them is supposed to cancel the 
 rainfall upon them.
 
 18 
 
 may be regarded as hardly possible of 
 attainroent. The hygrometric state of 
 the ground's surface, the aspect of the 
 sky, the amount of wind, and the tem- 
 perature, will all, in their degree, exercise 
 a sensible influence on the amount of 
 water tJiat the ground will give off from 
 its surface ; so that, in fact, it is doubtful 
 whether any reliable and philosophically 
 correct conclusions can be arrived at. The 
 resultant facts from such experiments as 
 have been carefully conducted afford, 
 after all, the only data for the engineer 
 to arrive at any general conclusion by; 
 and for forming a rough estimate for the 
 probable available rainfall of a district, 
 the following proportions of available 
 actual rainfall may be accepted as fur- 
 nishing general data ; but they are not 
 meant to obviate the necessity of a care- 
 ful and specific examination of the circum- 
 stances likely to affect the design of any 
 particular work : 
 
 steep surfaces of granite, gneiss, and slate 100 
 
 Moorland and liill pasture (JO to 80 
 
 Flat cultivated country 40 to 50 
 
 Chalk Oto
 
 19 
 
 In order to arrive at more specific, and 
 truly reliable results, the engineer will 
 have to make a series of accurate observa- 
 tions on the discharge of the stream or 
 streams that carry away the rainfall of a 
 district ; and by doing so, and at the same 
 time comparing the result with the 
 amount of rain registered by the gauges 
 — which should also, of course, be kept 
 with accuracy in the locality under ex- 
 amination — an approximately true esti- 
 mate of the available rainfall wdll be 
 arrived at. 
 
 If there is time in the preparation of a 
 project to make the necessary examination 
 of a district, it is evident that the results 
 will speak for themselves ; and there will 
 be no necessity to enter into abstract 
 speculations concerning the theory of the 
 influences affecting loss by either evapora- 
 tion or absorption. 
 
 In proportioning the size of a storage 
 reservoir to the area of the catchment 
 basin, the engineer will, of course, in the 
 first instance be guided by the require- 
 ments of the work. The object of the
 
 20 
 
 undertaking may be any one of the fol- 
 lowing : 
 
 To husband a scanty rainfall. 
 
 To check the injurious effect upon the 
 country by floods. 
 
 To add to the discharge of a stream, 
 by preventing the escape of the flood 
 waters. 
 
 The amount of storage will always be 
 part of an engineer's data in designing 
 works. It will either be his object to 
 store the whole of the water that the 
 drainage area will afford, which will be 
 the case in impounding water for irriga- 
 tion, for example ; or a certain fixed de- 
 mand, governed by the want of a town or 
 other requirements, will determine the 
 amount of the rainfall that it will be 
 necessary to retain for supply. In Eng- 
 land the demand for water supply may 
 be reckoned at from 150 to 180 days, 
 depending on the amount and the con- 
 stancy of the rainfall ; as a rule, the six 
 months' supply will be the safest to adopt. 
 The following table extracted from Mr. 
 Beardmore's work, shows the proportions
 
 21 
 
 that have been observed in designing 
 some of the best constructed reservoirs. 
 (See Table A.)* 
 
 From this it appears that the propor- 
 tion between the amount stored and the 
 total rainfall varies between one-half and 
 one-fourth. 
 
 *There are a number of discrepancies in this table, 
 but as the figures in the different columns are func- 
 tions of each other, it would be very diificult, if not 
 impossible, to say where the errors lie.
 
 22 
 
 Table A. 
 
 I 
 
 Locality. 
 
 Height 
 above Sea. 
 
 1 
 
 < 
 
 1 
 
 Bann Reservoirs, 1837-38.... 
 Greenock. 1827-28, flat moor. 
 
 Bute, 1820 
 
 Glencorse, Pentland Hills.. 
 
 Belmont, 1843, moorland 
 
 " 1844, 
 
 ft. ft. 
 
 400 to 2800 
 512 to 1000 
 200 to 350 
 734 to 1600 
 650 to 1600 
 
 sq. m. 
 
 5.15 
 7.88 
 7.80 
 6.00 
 2.81 
 
 " 1845, " 
 
 
 
 1846, " 
 
 
 
 Rivington Pike, 1847 
 
 Longendale " 
 
 Swineshaw " ........ 
 
 Turton and Entwistle, 1836 
 1837 
 
 800 to 1545 
 500 to 1800 
 600 to 18U0 
 500 to 1300 
 
 16.25 
 '"'3!i8' 
 
 Bolton Waterworks 
 
 800 to 1600 
 800 
 
 80 
 
 Ashton " 1844 
 
 .59
 
 23 
 
 Table A — Continued. 
 
 o 
 
 <D 
 
 
 
 ;_ 
 
 br-- 
 
 5 
 
 ^ 
 
 'd 
 
 cu 
 
 
 ta 
 
 
 
 
 
 
 
 o 
 
 1^ 
 
 Si 
 
 11 
 
 c3 
 
 
 ^3 
 
 O o 
 
 2 « 
 
 ?S 
 
 %< 
 
 1< =2 
 
 
 g"^ 
 
 m 
 
 
 a, 
 
 2<1 
 
 
 11 
 
 O 
 
 
 
 OJ 
 
 d 
 
 Sh O 
 
 H 
 
 P 
 
 « 
 
 K 
 
 
 Ph"" 
 
 c. ft. 
 
 C. ft. 
 
 ir. 
 
 . 
 
 c. ft. in. 
 
 
 perm. 
 
 perm. 
 
 in. 
 
 in. 
 
 mill. 
 
 
 1092.6 
 
 210.2 
 
 48.0 
 
 72.0 
 
 56.0 
 
 * 
 
 1416.6 
 
 197.7 
 
 41.0 
 
 60.0 
 
 38.0 
 
 2-5 
 
 819.0 
 
 105.0 
 
 23.9 
 
 45.4 
 
 
 
 600.0 
 
 100.0 
 
 22.3 
 
 37.0 
 
 7.66 
 
 3.10 
 
 630.4 
 
 224.3 
 
 50.7 
 
 63.4 
 
 26.8 
 
 \ 
 
 412.8 
 
 146.4 
 181.9 
 146.3 
 
 33.3 
 41.2 
 33.2 
 
 50 
 55.0 
 
 49.8 
 
 
 
 511.2 
 
 
 
 411.3 
 
 
 
 2880.0 
 
 176.7 
 
 40.0 
 49.5 
 37.0 
 41.0 
 
 55.5 
 55.5 
 49.3 
 46.2 
 
 29.6 
 
 i 
 
 
 
 .... 
 
 576.7 
 
 181.8 
 
 31.43 
 
 ^ 
 
 548.2 
 
 172.3 
 
 125.2 
 
 39.0 
 32.7 
 
 48.2 
 
 
 
 100.2 
 
 25.6 
 
 1-8 
 
 40.7 
 
 65.5 
 
 15.5 
 
 40.0 
 
 21.0 
 
 4-7
 
 24 
 
 The rule suggested by Professor Ran- 
 king " for estimating the available capa- 
 city required in a store reservoir, that 
 founded upon taking into account the 
 supply as well as the demand," is prob- 
 ably the best that can be adopted in de- 
 signing waterworks for the supply of a 
 town ; " for example, 180 days of the ex- 
 cess of the daily demand above the least 
 daily supply, as ascertained by gauging 
 and computation in the manner above 
 described." In order that a reservoir of 
 the capacity " prescribed by the preced- 
 ing rule may be efficient, it is essential 
 that the least available annual rainfall of 
 the gathering grounds should be suffici- 
 ent to supply a year's demand for water." 
 In calculating the capacity of a storage 
 reservoir, the consideration of the sur- 
 face evaporation must not be disregarded, 
 especially when the works are designed 
 for tropical or very dry climates. The 
 amount of loss will in some cases be very 
 considerable, for whatever depth of water 
 be assumed to pass away into the air, it 
 must be regarded as extending over the
 
 25 
 
 whole surface of the reservoir ; or, in fact, 
 the cubic quantity will be equal to the 
 product of the depth evaporated away 
 and the mean surface area of the reservoir 
 as the water rises or falls throughout the 
 year. Some have gone the length of as- 
 serting that the amount of evaporation 
 from the surface of large and deep bodies 
 of water is probably nothing at all, or at 
 any rate, not worthy of consideration ; 
 whilst others assume a much larger 
 amount of loss than appears to be sup- 
 ported by observation. The following 
 extract from the article " Physical Geog- 
 raphy," published by the Society for 
 Promoting Useful Knowledge, expresses 
 intelligibly the conditions that tend to 
 promote evaporation : 
 
 " Other things being equal, evaporation 
 is the more abundant the greater the 
 warmth of the air above that of the evap- 
 orating body, and least of all when their 
 temperature is the same. Neither does 
 much take place whenever the atmosphere 
 is more than 15 deg. colder than the sur- 
 face upon which it acts. Winds power-
 
 26 
 
 fully promote evaporation, because they 
 bring the air into continual as well as in- 
 to closer and more violent contact with 
 the surface acted upon, and also, in the 
 case of liquids, increase by the agitation 
 which they occasion, the number of points 
 of contact between the atmosphere and 
 the liquid. 
 
 "In the temperate zone, with a mean 
 temperature of 52 j^ deg., the annual evap- 
 oration has been found to be between 
 36 in. and 37 in. At Cumana, on the 
 coast of South America (N. lat. 10^), with 
 a mean temperature of 81.86 deg., it was 
 ascertained to be more than 100 in. in the 
 course of the year ; at Guadaloupe, in the 
 West Indies, it has been observed to 
 amount to 97 in. The degree of evapora- 
 tion very much depends upon the differ- 
 ence between the quantitj' of vapor which 
 the surrounding air is able to contain 
 when saturated and the quantity which it 
 actually contains. M. Humboldt found 
 that in the torrid zone the quantity of 
 vapor contained in the air is much nearer 
 to the point of saturation than in the
 
 27 
 
 temperate zone. The evaporation within 
 the tropics, and in hot weather in temper- 
 ate zones, is on this account less than 
 might have been supj^osed from the in- 
 crease of temperature." 
 
 Thus it appears that evaporation, under 
 highly favorable conditions, may take 
 place to the extent of 9ft. in depth — an 
 allowance that will demand careful con- 
 sideration in designing storage works. 
 In India, where, from the extreme dryness 
 of the atmosphere, the evaporation is 
 found to be considerable, the usual al- 
 lowance made by engineers for evapora- 
 tion from the surface of storage reservoirs 
 is at the rate of ^ in. of depth per diem 
 for eight months in the year. Regarding 
 the results that have been arrived at in 
 Bombay, this allowance would appear to 
 be about double what is necessary, for the 
 observations extending over five years 
 give a mean daily evaporation of less than 
 ^ in. In Bombay, however, the atmo- 
 sphere is much more humid than that ex- 
 perienced on the great tableland of the 
 Decaan ; and in Madras, where reservoirs
 
 28 
 
 are the specialty, it is probable that the 
 actual loss is not far from being a mean 
 between the two fractions. In Great 
 Britain the mean daily evaporation is 
 found to average less than the tenth of 
 an inch.* 
 
 In estimating the quantity of storage 
 water that will result from the drainage 
 of any particular district, it will be essen- 
 tial to consider carefully the geological 
 disposition of the strata characterizing 
 the locality in which it is contemplated 
 to establish the works. This, although a 
 matter that may influence the effective- 
 ness of an undertaking to the extent of 
 success or failure, will appear to thfe pure- 
 ly practical man to imply a degree of re- 
 finement that is uncalled for. There will 
 be no difficulty, however, in showing that 
 the geological conformation of a district 
 may be such as, on the one hand, to ma- 
 terially contribute to the efficiency of a 
 
 *See an interesting paper upon Evaporation, by Mr. 
 Desmond FitzGerald, M. Am. Soc. C. E., contained in 
 the Transactions of the American Society of Civil 
 Engineers, for Sept., 1886.
 
 29 
 
 storage reservoir, or on the other to prove 
 so defective that no engineering skill or 
 pecuniary outlay could remedy it. A 
 condition of geological structure, perhaps 
 the most favorable that could be imagined, 
 is that shown in Fig. 1. This diagram 
 represents a geological section taken at 
 right angles, or nearly so, to the axis of 
 the valley that it is proposed to convert to 
 the purpose of storage. This somewhat 
 peculiar structure is what is geologically 
 termed synclinal, the beds inclining away 
 from the axis of the valley, and is the re- 
 sult of an upheaving force having taken 
 place underneath the points of greatest 
 elevation. Subsequent to the upheaval 
 and consequent displacement of the 
 strata, the process of denudation has 
 taken place, cutting the upper beds, and 
 leaving the outcrop exposed, not only in- 
 side the basin, but in the adjoining 
 valleys at O and O. Now, it is evident 
 that if the highest ridges bounding the 
 valley be taken to mark the line of water- 
 shed, and therefore limiting the area of 
 the catchment basin, it is possible that
 
 r 
 
 ■■)
 
 31 
 
 the estimate of the umount of supply may 
 be found far short of what the district will 
 yield. A certain proportion of rain fall- 
 ing upon the outcrop at the points O O 
 will be absorbed by such of the strata as 
 are porous, and the water, percolating 
 through the bedding, till an impervious 
 stratum is met with, will find its way 
 down the course of the stratification, till 
 it ultimately reaches the reservoir in the 
 form of springs, and contributes more or 
 less to the maintenance of the supply. 
 The converse of this condition of things 
 will be readily understood by reference 
 to Fig. 2. It also represents a section 
 taken directly across the valley of the 
 proposed reservoir. Here the strata of 
 the earth's crust incline against each other 
 consequent upon some disturbing force 
 having taken place to elevate them, and 
 are said to be anticlinal to the axis of the 
 valley. In order to account for the for- 
 mation of a valley on the summit of the 
 ridge, that at first was thrown up, it is to 
 be understood that the upper beds suf- 
 fered fracture in the process of upheaval
 
 32
 
 33 
 
 and subsequently were exposed to de- 
 nudation. These valleys of elevation 
 are evidently not to be desired as situa- 
 tions for the establishment of storage 
 reservoirs. The area of the gathering 
 grounds will be much more limited than 
 the extent of the water-shed would ap- 
 pear to indicate ; and cannot safely be 
 relied upon to give an estimate of the 
 quantity of water that the valley will 
 afford. A certain amount of water will 
 undoubtedly pass over the surface in 
 times of heavy and continued rain, before 
 it can be absorbed; but there is no 
 doubt that of all the water absorbed by 
 the ground, by far the greater portion 
 will follow the inclination of the strata, 
 and come out as springs in the adjoining 
 valleys. 
 
 Fig. 3 shows a geological section that 
 combines in it favorable and unfavorable 
 conditions for the storage of water. On 
 one side the outcrops of the strata are 
 found to extend beyond the highest point 
 of water-shed line, whilst on the other 
 side the strata incline away, producing
 
 34 
 
 such a condition as would favor the 
 escape from the valley of the water ab- 
 sorbed. 
 
 Certain rules are in general use for 
 estimating the quantity of the total rain- 
 fall that will be lost by absorption and 
 evaporation, with a view to determining 
 the proper proportion to be observed be- 
 tween the reservoir and the area of the 
 catchment basin. Two-thirds of the 
 whole fall is sometimes taken to represent 
 the loss that may be expected from the 
 drainage of any district; in general terms, 
 one-third being assumed as the amount 
 that may actually be intercepted for util- 
 ization. Some authors leave a much 
 smaller margin, and state that fully two- 
 thirds of the total rainfall may fairly be 
 taken as available for storage. This is a 
 large discrepancy when the aj^plication 
 of the rules is taken to be general ; but 
 when the statements are applied to sepa- 
 rate districts and different countries, there 
 is nothing irreconcilable in them. Gen- 
 eral rules are undoubtedly of much value 
 if they be received with quahficalion,
 
 35
 
 and are not adopted as of absolutely uni- 
 versal application. They cannot, how- 
 ever, with safety be substituted for speci- 
 fic investigations, when so much depends 
 on starting with accurate data. 
 
 RESERVOIR SITES. 
 
 The special requirements of each partic- 
 ular case will, as a general rule, go far 
 towards determining the selection of a site 
 for the establishment of storage works. 
 Assuming, however, that there is a consid- 
 erable extent of country situated ad- 
 vantageously in relative position to the 
 locality at which it is proposed to utilize 
 the water, and that there is a choice of 
 ground, the point to be considered chiefly 
 will be the natural lie of the country. 
 To throw an embankment across a valley 
 at any point without due regard to the 
 .configuration of the ground would most 
 probably result in an expensive and ill- 
 designed scheme ; for under such circum- 
 stances the cost of the dam would bear 
 a very large proportion to the quantity 
 of water stored. It will rarely happen 
 that, in the examination of the resources
 
 37 
 
 of any particular piece of country, some 
 special features will not present them- 
 selves, favorable to the situation of stor- 
 age works. The most advantageous dis- 
 position of the ground will be when two 
 spurs of high land approach each other, 
 forming a narrow outlet for the stream, 
 and leaving a wide space above them in the 
 valley for storage. Such a configuration 
 is not uncommonly met with at the junc- 
 tion of two streams, as shown in Fig. 4. 
 This is merely a sketch from memory, by 
 the author, of a reservoir that he had de- 
 signed in India for purposes of irrigation; 
 and it will be evident that the dis^Dosition 
 of the ground was singularly favorable 
 in every respect for the construction of a 
 large storage work. The area of the 
 reservoir, as designed, was about three 
 square miles, and the maximum depth 90 
 ft., the area of the catchment basin being 
 about 60 square miles. Such favorable 
 situations for storage are of somewhat rare 
 occurrence ; for when the contour of the 
 land is what is desirable, it may be that 
 the area of water-shed is not adequate, or
 
 38
 
 39 
 
 possibly the geological condition of the 
 ground may be unfavorable, or the ma- 
 terials for the construction of a sound 
 bank are not available. In examining 
 large tracts of country in India, with a 
 view to the establishment of irrigation 
 reservoirs, the author found that more 
 reliance was to be placed on a careful ex- 
 amination of the map in the first instance, 
 than on the common plan of making per- 
 sonal explorations of the country. A 
 good map will show at a glance, especi- 
 ally if the hill-shading has been carefully 
 engraved, the points at which the supply 
 will be found sufficient to justify the un- 
 dertaking ; and will probably furnish a 
 pretty true indication of sites at which 
 embankments may be advantageously 
 constructed. 
 
 In tropical climates, where the rainfall 
 is in places very scanty, and where the 
 land is not of great value, it not unfre- 
 quently happens that such situations 
 prove available for the estabhshment of 
 large storage works as would not under 
 any circumstances be made available in
 
 40 
 
 England. These sites are to be found, 
 not at the head of a valley, but at some 
 considerable distance down the course 
 of a stream, where, the general inclination 
 of the country being slight, a low em- 
 bankment serves to store a very large 
 area of water. The apparent disadvant- 
 ages of such a site for storage are the 
 large area of land swamped and lost to 
 the cultivator and to Government, and 
 the great surface exposed to evaporation 
 under a tropical sun and the influence of 
 a dry wind. In India, the first objection 
 is one of comparatively little moment, 
 considering that in those districts where 
 irrigation is most required the value of 
 land is very trifling. From Is. to 2s. is 
 about an average rent per acre, where 
 land is under dry crops ; but when water 
 is available, the cultivators can, with 
 profit, afford to pay 30s. per acre. It is 
 therefore evident that, so far as Govern- 
 ment is concerned, there is no sacrifice 
 in the matter, but, on the contrary, an 
 unspeakable benefit is conferred on those 
 land owners who hold farms below the
 
 41 
 
 reservoir ; and an ample supply of water 
 is stored in the dryest seasons to mature 
 those crops whose failure almost inevit- 
 ably reduces the people to the verge of 
 starvation. The evaporation from these 
 lakes is, beyond question, a source of 
 very considerable loss, and one that ad- 
 mits of no possible abatement. Esti- 
 mated as above, at about half an inch 
 vertical for eight months of the year, the 
 loss frequently amounts to one-third of 
 the whole body of water stored. As a 
 set-off against this and other objections, 
 the facilities for constructing these 
 reservoirs of great extent, are consider- 
 able. In the first place, the embank- 
 ments, being very low, are rapidly and 
 cheaply constructed by native workmen ; 
 and when finished, the head of water even 
 at the deepest point is not sufficient to 
 try the work to any great extent. Fur- 
 ther, the greater the extent of the reser- 
 voir, the less inconvenience is experienced 
 from silting. The streams, owing to the 
 suddenness of the rainfall, come down 
 heavily charged with earth in suspension,
 
 42 
 
 the mass of which is deposited Hke a 
 miniature delta at the influx of the reser- 
 voir, instead of passing on and resting 
 near the embankment, as invariably 
 occurs in reservoirs of small extent. 
 The immense consumption of water 
 necessary to confer any appreciable bene- 
 fit by irrigation is of itself the strongest 
 argument in favor of these broad and 
 shallow reservoirs ; for it is not possible 
 to find in the upper part of a valley such 
 sites as would store the requisite quan- 
 tity of water without an embankment of 
 excessive dimensions ; and moreover, the 
 catchment area in such situations is not 
 usually sufficient to serve, with a scanty 
 rainfall, for the supply of a very large re- 
 servoir. It is not, of course, maintained 
 that this mode of storing water is by any 
 means applicable in England, for the cir- 
 cumstances and requirements in each case 
 are wholly dissimilar. 
 
 SUPPLY. 
 
 The reservoir site being supposed 
 everything that could be desired, as re- 
 gards the disposition of the ground, the
 
 43 
 
 supply will next engage attention as a 
 matter of course. Assuming that the 
 gathering grounds are sufficiently exten- 
 sive, it is presumed that the reservoir 
 will be constructed to contain sufficient 
 water to meet the maximum demand, 
 whatever that may be calculated at ; and 
 in order to determine with accuracy what 
 capacity the reservoir will have with dif- 
 ferent heights of embankment, it will be 
 necessary to carry out certain leveling 
 operations over the ground. The least 
 elaborate manner of proceeding will be 
 to run a series of cross-levels through 
 the valley, referring all to the same datum, 
 and by comparing these levels to ascer- 
 tain what the average depth will be for a 
 given height of bank. Having decided 
 the height of the water-level, the next 
 operation will be to contour round the 
 basin, and to survey the boundary-line. 
 In this way may be acquired sufficient 
 knowledge as to the storage capacity, 
 to justify the procedure with the work. 
 When the execution of the project has 
 been determined upon, it will be advis-
 
 44 
 
 able to make a more accurate survey of 
 the bed of the valley, and this can best 
 be done by covering the whole plan with 
 a series of contour lines at a vertical dis- 
 tance from each other of about 5 ft. This 
 kind of survey will be of lasting value to 
 the engineer, for it will enable him to 
 calculate what quantity of water the 
 reservoir will contain at each foot of 
 depth ; and, consequently, he will know, 
 from a mere inspection of the gauge in 
 the reservoir, how much water he has at 
 his disposal for service. 
 
 It has been assumed that the gathering 
 grounds are sufficient to maintain the re- 
 quisite suj^ply in the reservoir ; but it may 
 be well to pause and inquire what extent 
 of water-shed will be sufficient to furnish 
 a given supply, and what method may be 
 adopted for supplementing an insufficient 
 drainage area. It has before been re- 
 marked that the only reliable information, 
 when there is any question as to the 
 sufficiency of the rainfall or the area of 
 the catchment basin, can be derived from 
 careful gaugings of the stream or streams
 
 45 
 
 that may be depended upon to contribute 
 to the supply. If the catchment area is 
 very large as compared to the capacity of 
 the reservoir, a mere inspection of the map 
 and an exploration of the ground will gen- 
 erally be conclusive as to the sufficiency 
 of the supply for storage. Should there 
 not be such conclusive evidence on this 
 point, it must be determined by measur- 
 ing the quantity of water that absolutely 
 flows off the ground, at the same time 
 gauging the rainfall. This latter pre- 
 caution would appear unnecessary, but in 
 truth it is of great value, for it will fur- 
 nish by comparison with the rainfall 
 registers that have been kept through the 
 same year, and a series of previous years, 
 evidence as to the amount of available 
 rainfall that may be expected during 
 terms of comparative drought. If the 
 supply of a town with water be the 
 desideratum, the rule to be rigidly 
 observed is that of making a minimum 
 supply meet the maximum demand, and 
 therefore it is of the highest importance 
 to determine beyond any doubt, what the
 
 46 
 
 minimum yield of a catchment basin 
 will be. 
 
 As a mode of supplementing an insuf- 
 ficiently large drainage area, catchment 
 drains or feeders have frequently ren- 
 dered good service. These are cuts that 
 are carried outside the water-shed line 
 to arrest the surface drainage and catch 
 the contributions of small streams, and 
 conduct the water into the reservoir. 
 The greater the area enclosed between 
 catchment drains and the water- shed 
 line, the more valuable will they be as 
 aids to the supply of the reservoir. They 
 of course virtually extend the area of the 
 catchment, adding so many square miles 
 or acres to the rainfall. 
 
 DESIGNING OF WORKS. 
 
 Knowing the exact requirement of a 
 given population, or rather having fixed, 
 after every consideration, the daily con- 
 sumption of every individual that it is 
 proposed to supply, there will be no dif- 
 ficulty whatever in proportioning the 
 reservoir to the demand upon it. It is 
 sometimes necessary, however, to provide 
 
 1
 
 47 
 
 reservoirs for the purpose of preventing 
 damage to the country by floods, and in 
 this way the inconvenience and injury 
 naturally consequent upon very sudden 
 and excessive falls of rain may be to a 
 great extent obviated. The duty of the 
 reservoir will be to arrest all water in 
 excess of what the stream can carry 
 within its banks, and to dispose of this 
 excess water, so to speak, in detail, after 
 the excessive rainfall has become mod- 
 erated. A comparison of a stream's dis- 
 charge, taken at highest floods, with the 
 quantity that it can carry without over- 
 flowing its banks, will show the excess 
 that has to be retained by the reservoir ; 
 and these data can only be arrived at 
 through a carefully kept record of the 
 extent of the floods and of their duration. 
 The maximum flood in this consideration 
 will not be that which rises to the 
 greatest height for a short time, but will 
 be^the product of the excess above what 
 the-river can discharge by the length of 
 time the flood lasts ; which will, in fact, 
 be the necessary capacity of the reservoir.
 
 48 
 
 The table given on a preceding page 
 will afford an interesting study when 
 compared with the following table, ex- 
 tracted in part from the same work. The 
 first gives a comparative view of the 
 volume of water gauged and stored in 
 small hill districts, the last column indi- 
 cating the proportion of the total avail- 
 able rainfall to the amount actually 
 intercepted for storage. The following 
 table shows the ordinary summer dis- 
 charge of various rivers, streams, and 
 springs, as unaffected by immediate rain. 
 (See Table B.)* 
 
 Where the reservoir is designed to 
 check the injurious effects of floods, the 
 proportion of the storage to the rainfall 
 will, in most cases, be much smaller than 
 what would be necessary to provide for 
 the better part of a whole year's fall of 
 rain, for it is not probable that the max- 
 imum known flood can ever exceed the 
 amount that it would be necessary to 
 store for economic purposes. 
 
 ♦This table also contains several evident inaccura- 
 cies ; where the point of error seemed apparent, cor- 
 rection has been made in the present edition.
 
 49 
 
 Table B. 
 
 RlTERg. 
 
 Height 
 above Sea. 
 
 Thames at Staines, chalk, greensand 
 Oxford clay, oolites, etc 
 
 Severn at Stonebench, Silurian 
 
 Loddon (February, 1850), greensand. 
 
 Nene, at Peterborough, oolites, 
 Oxford clav and lias 
 
 Valley Hill, 
 ft. ft. 
 
 40 to 700 
 400 to 2600 
 110 to 700 
 
 10 to 600 
 
 Mimram, at Panshanger, chalk 
 
 Lee, at Lee Bridge, chalk -(Rennie, 
 
 April, 1796) 
 
 Wandle, below Carshalton, chalk. . . . 
 Medway, dryest seasons (Rennie, 
 
 1787), clay 
 
 200 to 500 
 
 30 to 600 
 70 to 3cO 
 
 Ditto, ordinary summer run (Rennie, 
 
 1787) 
 
 
 Verulam, at Bushey Hall, chalk 
 
 Gade, at Hunton Bridge, chalk 
 
 Plym, at Sheepstor, granite 
 
 Woodhead Tunnel, millstoue, grit. . . . 
 Glencorse Bum 
 
 150 to 500 
 150 to 500 
 800 to 1500 
 
 1000 
 750 to 1600 
 

 
 50 
 
 Table B. — Continued. 
 
 
 '•o 
 
 & 
 
 1 
 
 1 
 
 
 «f-i 
 
 m 
 
 a 
 
 ;3 
 
 c3 
 
 §)^. 
 
 
 M 
 
 *^S 
 
 2 
 < 
 
 1^ 
 
 'Si gj 
 
 ""i 
 
 bog 
 •2 2 
 
 ^1 
 
 ? 
 
 §dI 
 
 ^< 
 
 'z< 
 
 ^ 
 
 "6 
 
 s 
 
 
 ^^ 
 
 a 
 
 >-H 
 
 A 
 
 ?i ft 
 
 ^ 0. 
 
 '^ 
 
 s 
 
 S 
 
 ft 
 
 5 
 
 ;-i 
 
 o 
 
 
 
 
 
 
 Q 
 
 H 
 
 s 
 
 K 
 
 H 
 
 sq. 
 miles. 
 
 c. ft. 
 per. min. 
 
 C. ft. 
 
 per min. 
 
 in. 
 
 in. 
 
 3086 
 
 40.000 
 
 12 98 
 
 2.93 
 
 24.5 
 
 3900 
 
 33,111 
 
 8.49 
 
 1.98 
 
 
 221.8 
 
 3 000 
 
 13.53 
 
 3.01 
 
 '25!4 
 
 620.0 
 
 5,000 
 
 8.45 
 
 1.88 
 
 23.1 
 
 50.0 
 
 1,200 
 
 24.3 
 
 5.5 
 
 26.6 
 
 570.0 
 
 8 880 
 
 15.58 
 
 3.53 
 
 
 41.0 
 
 1,800 
 
 43.9 
 
 9.93 
 
 "24!0 
 
 481 5 
 
 2,209 
 
 4.59 
 
 1.04 
 
 
 
 481.5 
 
 2.520 
 
 5.23 
 
 1.19 
 
 
 120.8 
 
 1,800 
 
 14 9 
 
 3.37 
 
 
 69.5 
 
 2,500 
 
 36.2 
 
 8 19 
 
 
 7.6 
 
 500 
 139 
 
 71 4 
 
 15.10 
 
 '45;6 
 
 46 
 
 6*.6' 
 
 130 
 
 "21.6"' 
 
 ""--' 
 
 37.4
 
 51 
 
 PROPORTIONS OF BANK. 
 
 The proper proportion to be given to an 
 embankment for the support of water is a 
 question that appears to admit of a good 
 deal of difference of opinion, some de- 
 signers taking one view, some another, of 
 the proper theory that is to determine 
 the dimensions of a bank. Some few, 
 with whom the author cannot agree on 
 this point, maintain that a bank ought 
 to be designed with strict reference 
 to its theoretical power of resisting 
 hydrostatic pressure, or the effort of the 
 water to displace it. Kegarding the 
 question in its abstract form, it will be 
 evident that any structure intended to 
 sustain the pressure of water may be sup- 
 posed to fail in one of two ways — either, 
 in the first place, by yielding to the 
 horizontal pressure of the water and over- 
 turning, or by progressive motion, i. e., 
 sliding on its base. In considering the 
 first theory, that of resistance of over- 
 turning, the easiest method of examining 
 the question will be to take a simple ex- 
 ample of a vertical rectangular wall, and
 
 52 
 
 ascertain what power it exercises to resist 
 the pressure of water. The pressure of 
 water upon any plane surface immersed 
 is known to be equal to the area of that 
 surface, multiplied by the depth of its 
 center of gravity below the level of the 
 water. Generally speaking, the unit 
 adopted in calculation is a foot ; and the 
 
 Fig. 5. 
 
 unit of water being taken at a cubic foot, 
 weighing 62.5 lbs, the resulting product 
 from the multiplication of the three
 
 53 
 
 quantities will give the pressure in 
 pounds on the surface immersed. Let it 
 be supposed, for simplicity, that water to 
 the depth of 10 ft. has to be sustained by 
 a vertical rectangular wall, as in Fig. 5. 
 It is usual to take but 1 ft. length of the 
 wall for the calculation, though it will not 
 affect the result whether 1 ft. or 100 ft. 
 be the length assumed. We then have 
 the surface under pressure = 10 sq. ft., 
 the depth of the center of gravity = 5 ft., 
 and the weight of a cubic foot = 62.5 lbs., 
 the product of which quantities gives us 
 3,125 lbs. pressure on 1 ft. length of the 
 wall. But this pressure is not the whole 
 of the force that the wall has to resist ; 
 the leverage that it exerts must also be 
 taken into account. In the example 
 under consideration — viz., that of a verti- 
 cal plane with one of its sides coinciding 
 with the surface of the water, as in Fig. 
 5 — the whole of the pressure is so dis- 
 tributed as to be equal to a single, force 
 acting at a point one-third of the depth 
 from the bottom. Thus, the total force 
 to be resisted by the wall is 3,125 X 3.33
 
 54 
 
 = 10,406, which is the moment tending 
 to overturn the wall. 
 
 It is evident that a certain weight of 
 the wall must be opposed to this over- 
 turning force ; and as the height of the 
 wall and the length are determined 
 quantities, the thickness alone remains 
 for adjustment. But as a rectangular 
 wall in upsetting is considered to turn 
 upon a single point, F, in the Figure — 
 viz., the outer line of the wall — there will 
 be a certain amount of leverage to assist 
 the wall in resisting the pressure of the 
 water. This leverage is the horizontal 
 distance of the center of gravity of the 
 wall from the turning point F, and when 
 the structure is rectangular and vertical, 
 it is equal to half the thickness. The 
 amount of the wall's resistance will then 
 be equal to the number of cubic feet in 
 one foot of its length multiplied by the 
 weight of a single cubic foot of masonry 
 and by half the thickness of the wall. 
 Taking w = weight of a cubic foot of 
 water = 62.5 lbs. w' = weight of a cubic 
 foot of brickwork, say 112 lbs., x = thick-
 
 55 
 
 ness of the wall, and h = the height, the 
 conditions of simple stability will be ful- 
 filled when 
 
 2 6 ' 
 
 and solving for x, we get 
 
 -i^'i ^'\ 
 
 the thickness of the wall = 4 ft. 4 in.* 
 A simple example has been selected for 
 
 * It will grreatly facilitate all such calculations to 
 make, once for all, certain reductions and simplifica- 
 tions : Thus, the weisrht of a cubic foot of water 
 being 62.5 lbs., the thrust against a reservoir wall of 
 height h, sustaining a body of water level with its top 
 is, per running foot, 31.25^2, and the moment of the 
 same is 10.42^3. Also the thickness x of a rectangular 
 wall of density tt per cubic foot, in exact equili- 
 brium with this moment, is given by the equation, 
 
 ^ _ _ _. This shows that ;r varies inversely as the 
 ^/'^ square root of the density. It will be 
 found by trial that this formula gives the same result 
 as (2), the only difference being, that in the one just 
 given, all the roots have been extracted, except the 
 root of the density of the masonry. 
 
 A good authority---Mr. Fanning— gives as a proper 
 approximate thickness for a rectangular wall of 
 density = 140 lbs. per cubic foot, .r = 0.55A. It will 
 be perceived that this is about 40 »,^ additional beyond
 
 56 
 
 illustration ; but of course a rectangular 
 section of wall would not be found gen- 
 erally ajDplicable in practice, nor would it 
 be convenient to limit the dimensions of 
 a retaining wall of whatever kind to the 
 minimum that would sustain the pres- 
 sure. If this principle of calculation be 
 applied to ascertain the stability of a 
 bank of earth with long slopes of 2^ or 
 3 to 1, it can easily be shown that in 
 
 exact equilibrium. If we should make x = 0.60 h, we 
 would add 50%, and increase the moment of the wall 
 by 19%, while the section would be increased only 
 9%, for the moments increase as the squares of the 
 bases, and the sections, as the bases, only. 
 
 Such walls are not usually built with a rectangular 
 section, but the best way to design a trapezoidal wall 
 is first to determine the thickness of a rectangular 
 wall with the desired coefficient of safety, and of the 
 given height, and then transform it into one of 
 trapezoidal section, having an equal moment of 
 resistance. This may be readily done by using Vau- 
 ban's rule, which is, that all walls with vertical backs 
 and of equal resistance, have the same thickness at 
 l-9th of their height, to a very close approximation, 
 and within extended limits. Thus, suppose we 
 wished tt) transform a rectangular wall 27 feet high 
 and 15 ft. thick, into an e(iuivalent trapezoidal wall 
 with vertical back, and top thickness of 6 feet. Join 
 the outer extremity of the top with a point 3 feet above 
 the base of the outer face, and prolong the line till it 
 strikes the base prolonged. We shall then have a wall
 
 57 
 
 every case the resistance of the bank to 
 overturning is greatly in excess of the 
 horizontal leverage exercised by the 
 water sustained. 
 
 The only theory, then, in any degree 
 tenable, is that assuming a bank in yield- 
 ing to the pressure of water to slide on 
 its base. In order to conceive how this 
 can apply, it is necessary to assume the 
 embankment to be a rigid body resting, 
 for a given length of its section, on a 
 horizontal plane; and without any ad- 
 
 27 feet high, top thickness 6 feet, bottom thickness 
 16.125 feet, with a face batter of 4]4 inches to the foot. 
 The moment of resistance of this wall is slightly in ex- 
 cess of that of the assumed rectangular wall, while 
 its section is less by more than 26 y^. 
 
 Reservoir walls of considerable length— say over 
 five times their height— should be reinforced by ex- 
 terior counterforts, or buttresses, not further apart 
 than the height of the wall, and carried up to at least 
 half its height. The section of the wall should not be 
 diminished on account of these buttresses, which are 
 merely intended to give the long wall the same 
 strength that a shorter one of the same section would 
 possess. In all these calculations the adhesion of the 
 mortar is neglected. Perhaps this omission is counter- 
 balanced by the assumption that the wall is a mono- 
 lith. If we wished to take account of the adhesion of 
 the mortar, probably the best way would be to con- 
 sider its intensity in pounds per square inch, as so 
 much additional weight added to that of the wall.
 
 hesion, or a very small fraction, existing 
 between the surfaces pressed. The 
 amount of the friction, however, is just 
 the point upon which the whole matter 
 hinges, and until it has been ascertained 
 that the surfaces of earth that are care- 
 fully incorporated with one another have 
 any such thing as a coefficient of fric- 
 tion, it is idle to pursue the investiga- 
 tion by a mathematical mode of reason- 
 ing. The conditions of stability will be 
 satisfied when the horizontal component 
 of the water's pressure against the bank 
 will equal the weight of the bank, plus 
 the vertical pressure exercised by the 
 water to hold it down and multiplied by 
 the coefficient of friction ; but nothing is 
 known of this coefficient, and conse- 
 quently the equation remains incapable 
 of solution. As a matter of fact, em- 
 bankments do not slide bodily forward 
 on their base when they fail, but give way 
 from other causes than mathematical 
 reasoning can supply. Landshps, it is 
 true, to some extent support the principle 
 that maintains the sliding of embank-
 
 59 
 
 ments; but, here, the circumstances are 
 widely diiferent. LandsHps either take 
 place when a mass of earth rests upon an 
 inclined surface of rock, with an ample 
 supply of water to lubricate the surfaces 
 in contact, or else they are the result of 
 cutting or embanking earth to a higher 
 slope than the material will stand at ; the 
 infiltration of water also in this case is 
 the chief agent in producing the effect, 
 acting as a lubricant, and causing the 
 earth to assume its natural slope. In 
 each case the surface of separation is an 
 inclined plane, an element that does not 
 enter into the question of the stability of 
 embankments, by either of the modes of 
 reasoning above referred to. The prin- 
 ciples that direct the design of embank- 
 ments to retain water are not those that 
 apply to the calculation of the forces to 
 be resisted or the means to overcome 
 them, any more than breakwaters and 
 harbor walls can be designed on math- 
 ematical principles. The whole question 
 naturally turns on what slope the ma- 
 terial composing the bank will stand at-
 
 60 
 
 If earth could be got to remain at a slope 
 of 1 to 1, even though the embankment 
 had no thickness whatever at top, it 
 would be amply sufficient in weight to 
 uphold the water in a reservoir. This, 
 however, cannot be accomplished without 
 the assistance of retaining walls, which 
 would be found in most cases much more 
 expensive than the additional earth re- 
 quired to increase the slope to the angle 
 of stability ; and therefore the section is 
 so disposed that the earth shall stand 
 both inside and outside the reservoir at 
 such a slope as will be under all circum- 
 stances permanent. These slopes have 
 been determined by long practice and by 
 success and failure in pre-existing in- 
 stances — that is to say, the limits have 
 been laid down, for it is not to be as- 
 sumed that all descriptions of earth will 
 fall to exactly the same slope when ex- 
 posed to the constant action of water or 
 weather. Earth when subjected to the 
 contact of water almost invariably loses 
 a certain amount of its stability, and 
 therefore it is usual to give the inner side
 
 61 
 
 of an embankment a longer slope than 
 the outside. In most of the best exist- 
 ing examples the inside slope of the bank 
 is either 3 to 1 or 2J to 1, and it is rare 
 to meet any departure from this rule. 
 The outside slope may be designed at 
 from 2 to 1 to 3 to 1, depending upon 
 the character of the material, its power 
 of withstanding the erosive action of the 
 air, and the means used to protect the 
 surface from being washed off or from 
 crumbling away. In designing embank- 
 ments, the impermeability of the earth is 
 a matter that cannot be relied upon. 
 There are, it is true, innumerable em- 
 bankments now standing that have never 
 allowed the escape of a drop of water 
 from the reservoir, although no special 
 precaution was taken to make them water- 
 tight. Of these India abounds with ex- 
 amples, the introduction of a puddle wall 
 being in the older embankments of very 
 exceptional occurrence. The earth was 
 merely dug out close at hand, and carried 
 by the work-people in baskets on their 
 heads to where it was deposited, without
 
 6r> 
 
 any regard to the mode of disposing the 
 material. The author has had occasion 
 to construct a considerable length of 
 levee, or embankment, on this simple 
 plan for the protection of the country 
 from the flooding of a river; and 
 although, so far as he is aware, no flood 
 has yet taken place to test the work, 
 he has, from the study of existing ex- 
 amples, entire confidence in the result. 
 The earth, so far as practicable, was dis- 
 posed in layers, and before each was com- 
 pleted it was thoroughly consolidated 
 by the tread of the workmen. It is not 
 suggested that the puddle wall should be 
 dispensed with in designing embank- 
 ments, for the additional degree of safety, 
 in most instances, will more than compen- 
 sate for the extra expense it entails ; but, 
 in low embankments made of good re- 
 tentive clay, the precaution of puddling 
 is by no means a necessity. 
 
 In most of the best examples of em- 
 bankments in England, the practice 
 adopted has been to carry up the earth- 
 work in layers of 2 or 3 ft. in thickness,
 
 63
 
 64 

 
 65 
 
 disposed in the manner shown in Figs. 
 6 and 7, and at the same time to con- 
 struct in the center of the bank a wall of 
 well-puddled clay, the foundation of which 
 is carried down for whatever depth may 
 be necessary in order to reach an imper- 
 meable bed of earth or rock. It is not in 
 all situations possible to procure earth 
 exactly suitable and in sufficient quantity 
 for the construction of an embankment, 
 and consequently it is usual and advis- 
 able to dispose the best part of the ma- 
 terial — that is, the most retentive of 
 water — in juxtaposition to the puddle 
 wall, as indicated in Fig. 6.* In this ex- 
 ample, the selected material is disposed 
 equally at either side of the puddle ; but, 
 as its function is to withstand the ad- 
 mission of water, it would probably be 
 more consistent, though less in accord- 
 
 * This practice is sound, but would iu many cases 
 be very diflflcult of execution, at least without con- 
 siderable extra expense. When the borrow-pits are 
 opened, the material is generally taken as it comes. 
 The top soil, and all roots and sods should be first 
 removed, and all stones larger than are allowed by 
 the specifications are picked out and, generally, re- 
 erved for the rip-rapping of the bank.
 
 66
 
 67 
 
 ance with practice, to place all the select- 
 ed material on the inner side. The 
 practice of excavating the earth for an 
 embankment from the inside of the reser- 
 voir is one that should not be followed 
 without caution. Removing so large a 
 mass of material would, no doubt, give a 
 considerable increase of storage room 5 
 but sometimes the bed of a reservoir is 
 covered by a layer of impervious clay 
 that is of immense value, and if this be 
 cut through or removed, it is quite 
 possible that a bed of porous material 
 may be met with sufficient to allow the 
 escape of water when it comes to be ad- 
 mitted. In specifying for the dimensions 
 of the puddle wall, a sound rule for 
 adoption is, that it shall have a thickness 
 of 10 ft. at the top water-line and in- 
 crease in thickness to the surface of the 
 ground at the rate of 1 in. on each side 
 for every foot of height. Before any ex- 
 cavation is commenced, it will be essen- 
 tial to make a sufficient number of 
 borings to ascertain the nature of the soil 
 beneath the surface.
 
 68 
 
 It may here be mentioned that profes- 
 sional men are not apparently agreed as 
 to the principles to be kept in view in 
 constructing reservoir embankments ; and 
 this want of concurrence never was more 
 apparent than in the discussion that fol- 
 lowed the destruction of the Dale Dyke 
 reservoir, near Sheffield. Fig. 8 shows 
 a plan of the embankment site after the 
 catastrophe. The bank was 95 ft. high, 
 with slopes of 2^ to 1, and a top width of 
 12 ft. The puddle wall was 16 ft. in 
 width at the ground-line, and tapered to 
 4 ft. at the top of the bank. This em- 
 bankment, with the exception of the 
 puddle wall, was composed of rubble 
 stone and shale ; an additional price hav- 
 ing been given by the engineers to insure 
 the use of the former material ; which 
 proves, at any rate, that this mode of 
 construction was adopted on principle 
 and not through ignorance or mistake. 
 From the evidence given by the engineers 
 of the company, it appears that it was, 
 in their opinion, desirable that the inner 
 part of the embankment should be per-
 
 69 
 
 SheFa^
 
 70 
 
 meable to water, because earth was much 
 more Hkely to subside and slip than an 
 open and less yielding material like stone. 
 This mode of construction implies that 
 the puddle shall be fully sufficient of itself 
 to resist the passage of water, and that 
 there is no necessity to relieve it of any 
 part of the pressure against it. Of course, 
 if a bank be composed of open work, 
 every point in the face of the puddle is 
 exposed to the full and direct hydrostatic 
 pressure ; and if at any point there is the 
 smallest fissure or imperfection, the water 
 has full power against it, and will, to a 
 certainty, take advantage of such point to 
 breach the dam. The assumption, then, 
 of the constructors of this and the Agden 
 reservoir evidently was that a puddle wall 
 of some 25,000 sq. ft. of area was to be 
 constructed without an imperfection of 
 any kind, or a single weak point in the 
 whole surface. 
 
 The obvious reason for employing 
 puddle at all, in embaukments, is to 
 thoroughly close up any imperfection 
 that may occur in the earthwork ; it is in
 
 71 
 
 fact merely an accessory, and cannot be 
 relied upon of itself to secure the em- 
 bankment against destruction. If an 
 embankment be constructed of good 
 sound earthwork, j^roperly executed, it 
 is highly probable that the water may 
 never penetrate half way through to the 
 puddle wall, and probably, in the majority 
 of examples, has not done so. Earth- 
 work, however, is not always executed 
 without imperfection; some decompos- 
 able material maybe introduced, which, in 
 course of time, dissolves, leaving a fissure ; 
 one part may be at first less consolidated 
 than another, and, subsiding, lead to im- 
 perfection ; or an embankmeut, be it ever 
 so well constructed, may be burrowed 
 through by moles, rats, and other vermin. 
 It is to meet the first two of these 
 sources of imperfection that puddle is 
 used; and if, by such fissures as may 
 occur in ordinary eartli-work, water is ad- 
 mitted as far as the puddle wall, it can 
 only exercise pressure against it at a few 
 points, the puddle and earth being, iu 
 good work, so bonded and incorporated
 
 72 
 
 with each other that there is no space 
 left for the water to occupy and press 
 against the surface. Most who have read 
 the account of the disaster that oc(?urred 
 in March, 1864, at Sheffield, will recollect 
 how singularly conflicting the profes- 
 sional evidence on that occasion was. 
 Some of the first engineers were ranged 
 against each other in order to satisfy the 
 public as to whether the failure of the 
 embankment was attributable to bad en- 
 gineering or to a landslip ; and although 
 the impression finally remained on the 
 public mind that *' there was not that en- 
 gineering skill and attention to the con- 
 struction of the works that their magni- 
 tude and importance demanded," the en- 
 gineers were fairly divided in opinion as 
 to the cause of the disaster. One section 
 pronounced, without quahfication, that 
 the embankment gave way in consequence 
 of a landslip, and entirely ignored the 
 fact of the embankment being defectively 
 constructed ; whilst the other gentlemen 
 gave their verdict dead against the com- 
 pany, and their mode of constructing
 
 73 
 
 water-tight banks.* The two diagrams, 
 Nos. 6 and 7, may be taken as indicating 
 the system of constructing embankments 
 most generally approved of. The puddle^ 
 as will be observed, is carried up to the 
 natural surface of the ground without 
 any batter, and from that point slopes on 
 each side to the top of the bank; on 
 either side of the puddle is disposed, in 
 concave layers, the most sound and re- 
 tentive part of the material, and outside 
 of all comes the ordinary earthwork.! 
 
 As a security against the eroding action 
 of the water, and also against the inroads 
 of vermin, the most desirable, as well as 
 the most usual practice, is to pitch the 
 whole of the inner surface of an embank- 
 ment with stone, carefully laid by hand. 
 Neglect of this precaution has led to the 
 destruction of many embankments in 
 other respects securely constructed, and 
 even when ample height of bank above 
 
 * At the present day, it seems incredible that any 
 other opinion should have been entertained. 
 
 t It may be doubted if disposing of the material in 
 concave layers is good practice.
 
 74 
 
 the surface of highest water was provided. 
 In all ordinarily inclement weather the 
 disturbance of the surface of a reservoir 
 amounts to no more than a mere ripple ; 
 but when the surface is of large extent, 
 and a severe storm blowing, the waves 
 produced are such as to cause reasonable 
 apprehension, and, in fact, have, before 
 now, overtopped the bank and cut it 
 down, till the water flowed over and 
 caused the destruction of the work. In 
 most cases, it will be necessary to leave 
 about 5 ft. between the level of the high- 
 est water and the top of the embankment, 
 and never less than 3 ft. 
 
 A mode of construction not very gen- 
 erally used, but apparently consistent 
 with reason, is that shown in Fig. 7, the 
 embankment for the Biddeford Water- 
 works. It consists in covering the whole 
 of the inner face with a layer of puddle, 
 with sometimes a layer of peat outside it. 
 On some occasions it has been thought 
 de&irable to mix with the puddle a quan- 
 tity of small stones or furnace cinders, 
 by way of obstruction to vermin — a pre-
 
 75 
 
 caution that is bj no means unnecessary. 
 As an instance in point the author is re- 
 minded of a masonry dam in India that 
 had to be pointed every year regularly, 
 because the fresh-water crabs in the 
 reservoir found it convenient and pro- 
 motive of their development of shell to 
 appropriate the mortar to their personal 
 use. The joints were cleaned out as effec- 
 tually at the end of each monsoon as if the 
 work had been done to order. 
 
 The preparation of the foundation for 
 an embankment is a matter requiring 
 some care. The soil, consisting of grass, 
 roots, etc., and other matters of a decom- 
 posable nature, should be carefully re- 
 moved over the whole surface to be cov- 
 ered by the bank, and if any porous ma- 
 terial, such as sand or gravel, be present, 
 it must be removed, until a compact and 
 water-tight bed is arrived at.* The bank 
 
 * This must be taken with some reserve. It is evi- 
 dent that carryinfj out this n commendation hteraiiy, 
 would aiuouut to dijigin^ out tiie whole site of the 
 embankment to a depth tqual to tliat of the center- 
 wall louiidation. All that can be done in most cases, 
 is to remove the sods, roots, tic, from the site of the 
 eiuOankment, so ttiat a fresh earth surface is laid 
 bare, and commence the till upon that.
 
 76 
 
 must, in fact, be in contact with some 
 sound and reliable material that will not 
 admit the passage of water. 
 
 APPENDAGES OF RESERVOIRS. 
 
 Under this heading may be considered : 
 
 The whole apparatus for allowing the 
 water to escape, including the pipes, the 
 valve tower, and the culvert. 
 
 The waste sluices. 
 
 The waste weir or by- wash. 
 
 The most economical mode of discharg- 
 ing water from a reservoir is through a 
 single pipe passing either through the 
 embankment or immediately under it ; 
 but this plan cannot under any circum- 
 stances, be recommended, though it is 
 some times found in existing examples. 
 It is open to several grave objections, the 
 principal of which, perhaps, is that the 
 failure of a joint under the embankment 
 from unequal pressure, or from whatever 
 cause, will probably produce the destruc- 
 tion of the embankment, or at any rate, 
 entail a serious interruption to the sup- 
 ply, by the reservoir having to be emptied, 
 in order to repair the pipe. Buried in or
 
 77 
 
 under an embankment, a pipe is com- 
 pletely out of reach and out of view, and 
 may be in a very defective state with- 
 out its being possible to detect the im- 
 perfection. 
 
 In order to secure the satisfactory 
 working of a reservoir as a source of con- 
 stant supjDly, it is essential that the out- 
 let pipes, valves, and all other append- 
 ages for controlling and regulating the 
 escape of the water, should be accessible 
 for inspection and repair. The usual 
 mode of accomplishing this is to carry 
 the pipes out through a culvert of brick 
 or masonry of sufficient dimensions to 
 admit a man. This culvert communi- 
 cates with the valve tower, as shown in 
 Figs. 6 and 7, so that there is a complete 
 communication between the outside of 
 the reservoir and the inside. When un- 
 avoidable, the culvert is carried straight 
 under the embankment in the natural 
 ground ; but the safest and most gener- 
 ally approved mode of construction is to 
 bring the culvert round the end of the 
 embankment, where it will be out of reach
 
 78 
 
 of injury from unequal settlement ; a 
 source of no small apprehension when 
 either culvert or pipes alone are carried 
 under the bank. Where possible, it is 
 an excellent plan to run a heading 
 through the solid ground, lining it with 
 brickwork and puddling it, forming a 
 tunnel entirely independent of the em- 
 bankment. The principal objection to 
 carrying either the culvert or pipes 
 through or under the bank is their 
 liability to fracture from the enequal 
 settlement of the earthwork. It would 
 appear that their liability to damage can- 
 not with certainty be insured by any 
 reasonable depth of excavation, and is, 
 therefore, generally disapproved of by 
 the best authorities. 
 
 In the best constructions the culvert is 
 situated half way, or two-thirds up the 
 embankment, and in such case the outlet 
 pipes for drawing off the water in the 
 reservoir act as syphons when the water 
 surface has fallen below the culvert. 
 Fig. 6 shows a plan, as well as a cross 
 section, of a reservoir dam designed for
 
 79 
 
 general application by Mr. Rawlinson. 
 Here the bottom of the culvert is about 
 25 ft. above where the inner slope of the 
 embankment intersects the ground at 
 the lowest point. The syphon pipe is 
 also shown passing through the culvert ; 
 the horizontal culvert is connected with 
 a shaft inside the embankment, in which 
 are placed the valves for leading off the 
 supply from the reservoir. The valves 
 are made to be closed on the inside by 
 valve spindles and screws, and the inlet 
 pipes are closed on the outside by plugs 
 which can be applied from the top of the 
 valve tower. Thus the engineer has full 
 command of the whole of the outlet 
 works ; all the pipes and valves are easily 
 accessible and under perfect control, so 
 that the supply can at any time be 
 arrested for the repair of any derange- 
 ment that may occur, even to the removal 
 and replacement of all the pipes. The 
 inlet pipes are shown in this example, as 
 well as in Fig. 7, fixed at different heights 
 in the valve-tower, the object of which is 
 to draw the supply from the reservoir
 
 80 
 
 from points near the surface. The out- 
 let pipe, passing through or under the 
 embankment, may be connected on the 
 inside of the reservoir by a flexible 
 joint with another pipe of the same 
 diameter, to the upper end of which is at- 
 tached a float. This pipe is movable in a 
 vertical plane, being controlled from 
 lateral motion by the guide-posts. Such 
 an arrangement admits of the water be- 
 ing drawn off from the surface, -where it. 
 is least liable to be contaminated with 
 impurities. Whatever arrangement be 
 selected for drawing the supply off from 
 a reservoir, the system of carrying the 
 pipes, either with or without a culvert, 
 through or under the embankment, can- 
 not be sufficiently deprecated ; they are, 
 in such a position, beyond the reach of 
 inspection, and, moreover, are very likely 
 to induce leakage from the reservoir. It 
 is usual to puddle carefully the culvert 
 or pipes when carried under or through 
 the bank, but, even with such a pre- 
 caution, the water has under a consider- 
 able head a tendency to creep along the
 
 81 
 
 pipe, and, by soaking into the earthwork, 
 may cause any one of the many evils that 
 imperil and destroy embankments. 
 
 When embankments are not of great 
 height, an exceedingly cheap and simple 
 mode might be adopted for drawing off 
 the water. This would be by laying a 
 syphon over the embankment, as was 
 done in the case of the middle-level drain- 
 age in Cambridgeshire, which syphon 
 would at the inner side have a flexible con- 
 nection with another tube having a float 
 attached, as above described. Such an 
 arrangement would apply in princii^le to 
 heights not exceeding 30 ft., as the 
 pressure of the atmosphere would main- 
 tain no greater height. In practice, how- 
 ever, the syphons cannot be worked with 
 success at much above 20 ft., for it is 
 found that after a short time, the flow 
 becomes arrested by the collection of air 
 in the upper part of the syphon, and it 
 becomes necessary to pump the air out 
 constantly, to prevent it from interfering 
 with the flow, as it would do if not re- 
 moved. It would appear a simple mat-
 
 82 
 
 ter, where it is desirable to adopt a 
 syphon, to utihze the power of the water 
 flowing out for the purpose of getting 
 rid of the air ; it miglit easily be applied, 
 through a small wheel and suitable gear- 
 ing, to work an air-pump fixed at the 
 highest point of the syphon, making the 
 whole arrangement self-acting. The 
 arrangement could be successfully ap- 
 plied to irrigation tanks in India, where 
 the embankments are frequently less than 
 30 ft. Each leg of the syphon should be 
 provided with a valve to retain the water, 
 and when the supply was intermittent it 
 would be essential to have an opening at 
 the highest point of the syphon, and 
 some appliance, perhaps an air-pump, for 
 filling it with water in case of leakage. 
 
 To insure a constant discharge from 
 a reservoir with a constantly varying 
 head, several methods have been adopted ; 
 of these, one of the most ingenious is 
 that used at the Gorbals Waterworks, 
 near Glasgow. Fig. 9 represents a 
 transverse section through the regulator- 
 house, showing the arrangement by which
 
 83 
 
 the discharge is equalized. To the orifice 
 of the outlet pipe, O, is fitted a square- 
 hinged flip v.ilve of wood, against which 
 presses, by a friction roller, a lever, B, 
 the arms of which are bent. To the 
 upper arm is attached a chain that passes 
 over a pulley, and is connected with a 
 cast-iron cylinder or float, D, that stands 
 in the reservoir, E, of slightly larger 
 diameter. At the side of the entrance- 
 door of the building is placed another 
 cistern, G, of cast-iron, closed at top, and 
 communicating by a pipe, R R, with the 
 vertical pipe, H, which is in connection 
 with the outlet pipe, and passes up the 
 slope of the embankment, to carry 
 away any air that may accumulate 
 in the main. The cistern, G, is con- 
 nected with the reservoir, E, by a pipe, K, 
 which supplies water to float the cyl- 
 inder, D. Now, it is evident that the 
 discharge from the reservoir will be reg- 
 ulated by the position of the lever, B, 
 and this again will be controlled by the 
 height of the float, D. To regulate this 
 height the supply from the cistern, G,
 
 84
 
 85 
 
 must be self-adjusting, or be regulated 
 by the amount of water flowing away. 
 The float, N, has attached to it a spindle, 
 on which are fixed two double-beat 
 valves that work in the vertical part of 
 the pipe,K, one of which admits water from 
 the cistern, G, into the cylinder, E, and 
 the other allows the water to escape 
 from the reservoir, E. Now, if the sur- 
 face of the water upon which the float, 
 N, rests should rise above the proper 
 level, the float forces up the spindle, 
 closing the supply valve from the cistern, 
 and at the same time opening the lower 
 valve. Thus the supply is cut off and 
 the escape opened, enabling the float, D, 
 to fall. The subsidence of the float 
 closes more or less the flap valve, and 
 checks the discharge, in consequence of 
 which the surface of the water falls, and 
 with it the float, N, which consequently 
 opens the supply valve, and again admits 
 water into the cistern, E. Thus an 
 almost perfect equality between the con- 
 sumption and the supply of water is pre- 
 served. It would appear that the same
 
 86 
 
 effect could be produced by connecting 
 the lever directly with a float on the sur- 
 face of the water, but such an arrange- 
 ment would only apply when the pres- 
 sure against the flap is trifling.* 
 
 It is essential that every reservoir 
 should be provided with some means of 
 getting rid of the excess of water that 
 flows into it, and whether this provision 
 be made by a waste weir, sluices, or 
 waste pit, it is one that should not be 
 omitted. The most advantageous posi- 
 tion for a waste weir will generally be at 
 some point remote from, and entirely un- 
 connected with, the embankment, and 
 occasionally a natural depression in the 
 ground, as shown in Fig. 4, will afford 
 remarkable facilities for the construction 
 
 ♦ As a general rule, the discharge is regulated by 
 openiDg the gates to a greater or less extent, accord- 
 ing as more or less water is required. Indeed, it 
 would seem unwise to have an automatic system, 
 maintaining a constant discharge irrespective of a 
 diminishing head. As the supply of water in the 
 reservoir diminishes, it is better to let the draft from 
 it diminish also, as a warning that the reserve is being 
 drawn upon.
 
 87 
 
 of an escape. The level of the crest of 
 the waste weir with reference to the top 
 of the dam will require to be carefully 
 adjusted, the minimum difference of level 
 being 3 feet, and the maximum about 
 10 ft., depending on varying circum- 
 stances. The height of the waste weir 
 will, of course, regulate the top water 
 level in the reservoir ; and this must be 
 fixed with regard to the probability of 
 the embankment being over-topped by 
 waves. The circumstances influencing 
 the height of the waves in a reservoir are 
 the extent of the water surface, the 
 depth, and the amount of exposure to or 
 shelter from wind, all of which will vary 
 with each particular case. Under or- 
 dinary circumstances, the height of the 
 top of the embankment above the crest 
 of the waste weir should be for 
 
 an embankment 25 ft. deep, 4 ft. 
 50 ft. " 5 ft. 
 75 ft. '' 6 ft. 
 
 and for greater height of embankment 
 the difference of level may be propor-
 
 88 
 
 tionately increased. When the configura- 
 tion of the ground does not afford any 
 facilities for the construction of a waste 
 weir after the manner described, suffi- 
 cient provision for the escape of the 
 overflow is made through a waste pit. 
 This waste pit, or tower, is generally a 
 circular structure built over the outlet 
 culvert inside the reservoir, and serves 
 equally for access to the valves and for 
 the escape of the flood water. With 
 regard to the capacity of the waste weir 
 or waste pit, whichever be adopted, it 
 will be necessary to make ample provision 
 for the discharge of the sudden acces- 
 sions of flood water that reservoirs are 
 subject to, and which so seriously im- 
 peril their safety. To provide for this 
 there is an empirical rule amongst en- 
 gineers that is supposed to suffice for 
 the most urgent contingencies. It 
 states that there shall not be less than 3 
 ft. of length of overfall for every hundred 
 acres of gathering ground, but it is 
 obvious that to proportion the length of
 
 89 
 
 the waste weir to a given area of country 
 in all cases would be unreasonable.* 
 
 The discharge over the weir will not 
 depend only upon the quantity of rain 
 falling on a certain area of ground, but 
 also on the extent of the reservoir as 
 compared to the gathering ground, and 
 on the flat or precipitous character of the 
 basin. The only safe mode, then, of 
 proportioning the length of the escape 
 will be to ascertain with exactness what 
 the discharge of the stream or streams 
 flowing out of the reservoir was during 
 the greatest known flood, and then fixing 
 upon an arbitrary depth for the water to 
 flow over the weir, say 2 ft. or 3 ft., to 
 calculate what length of overfall will 
 suffice for the discharge of the excess 
 water. In India, where large waste- 
 weir accommodation is essentially neces- 
 
 * The above rule might answer up to say 3 square 
 miles of drainage area, or gathering ground. Beyond 
 this it would give excessive lengths. For large areas, 
 up to say 50 square miles, 3 ft. per square mile 
 would be nearer the mark. A rule that would fit-, 
 fairly well, all cases, would be : Length in feet = 20 
 times the square root of the number of square miles 
 of drainage area.
 
 90 
 
 sary, while it is equally a necessity to 
 save every gallon" of water that is pos- 
 sible, it is a common practice to form a 
 temporary dam, of earth and sods, on 
 the top of the waste weir ; this serves to 
 pond up some 3 ft. or 4 ft. of water over 
 the whole surface of the reservoir, and 
 does not imperil the security of the 
 works. In times of heavy floods the 
 water rises and overtops the temporary 
 dam, and no sooner does so, than the 
 whole is carried away, and the water in 
 the reservoir quickly subsides. 
 
 In works designed for the supply of 
 towns, it is sometimes necessary to make 
 provision to arrest the entrance of flood- 
 water into the reservoir, as the streams 
 may come down charged with large 
 quantities of matter in suspension that 
 would injure the purity of the water for 
 domestic consumption. These streams 
 may be diverted and carried round the 
 margin of the tank past the dam, and 
 can be admitted into the channel of the 
 stream, or be utilized for mill power. 
 On tbe Manchester Waterwoiks, are con-
 
 91 
 
 structed across the mountain streams, 
 weirs of an ingenious design, for the pur- 
 pose of separating the flood-waters from 
 the ordinary flow. The dimensions are 
 adjusted from observations of each par- 
 ticular stream, so that the discharge up 
 to a certain amount will take place into 
 the channel for the supply of the town ; 
 but when the discharge increases, and 
 the water becomes turbid, it has suffi- 
 cient velocity to carry it over the open- 
 ing, and flows down to the compensation 
 reservoir for the supply of mill power. 
 
 In determining the dimensions of a 
 weir of this kind it is first to be ascer- 
 tained what the mean velocity of the 
 water flowing over will be for a given 
 depth of water, A, above the crest. The 
 mean velocity, v, will be 
 
 V = \x 8.024^^" = 5M^h. 
 o 
 
 If the vertical height of the crest of the 
 weir above the point to be overleaped by 
 the cascade be called x, the distance 
 across will be
 
 9^ 
 
 2 v^Jx 4 , — 
 
 Before concluding, it will be well to 
 give a brief consideration to the causes 
 tending to the failure of embankments. 
 The foregoing remarks will, in suggest- 
 ing the best mode of construction, have 
 anticipated much that might be said on 
 the subject of failures ; but there are a 
 few points, the recapitulation of which 
 the importance of the subject demands. 
 
 There are unfortunately on record, 
 accidents, if they can be so called, from 
 the bursting of embankments, that, if 
 estimated by the loss of life attending 
 them, are as appalling as anything within 
 the memory of man. Thousands of 
 human lives have been sacrificed to ig- 
 norance and false economy, as well as in 
 some instances to natural defects that it 
 would have been difficult to foresee. 
 
 The existence of springs on the site of 
 an embankment is an undoubted cause 
 for apprehension, and considerable care 
 should be taken to carry all water from
 
 93 
 
 this source away;* that it may not, as 
 it certainly will if not checked, force its 
 way between the surface of the ground 
 and the seat of the embankment. In 
 doing so there is every probability that 
 the earth of the embankment will be 
 washed out by constant trickling, till a 
 fissure is formed of sufficient dimensions 
 to render the destruction of the bank a 
 certainty, if the water from the reservoir 
 should ever penetrate so far. As a pro- 
 vision against this source of injury, all 
 springs found on the site of an embank- 
 ment should be taken up and carried 
 away in proper drains sufficiently and 
 securely puddled. Thus the water is 
 confined to a single channel, and has no 
 tendency to soak into the earthwork and 
 blow it up in endeavoring to escape. In 
 embankments of all kinds the presence 
 of water is a most serious evil, and 
 
 * Where to? All these springs, and all percolations 
 whatever, must be cut off and prevented from travers- 
 ing the embankment by a water-iig:ht puddle or 
 center wall, which is the only safeguard against this 
 dauirer, to meet which is one of the principal objects 
 of such wail. .."I'l. i la-jaiii
 
 94 
 
 one by which may be accounted for, some 
 of the most extensive land slips that are 
 on record. It is erroneous to assume 
 that when water is the active element in 
 producing disruption in an embankment 
 or mass of earth of any kind, that it only 
 acts as a lubricant between the surfaces 
 in contact. The truth is, the bulk of 
 earth is sensibly affected by the amount 
 of moisture in it, as is seen in the sub- 
 sidence of newly-formed railway banks 
 when exposed to rain. If, then, a suffi- 
 cient quantity of water find its way into 
 the center of a bank that has been put 
 together in a comparatively dry state, it 
 will rise and soak into the earth until at 
 length what was a solid mass becomes 
 semi-fluid, settles into a smaller space 
 than it before occupied, and, as a conse- 
 quence, will leave a vacuity above it. The 
 inevitable result is the subsidence of the 
 superincumbent earth ; but instead of 
 resting, as at first, on a resisting ma- 
 terial, it floats, so to speak, on the semi- 
 fluid mass underneath, and having little 
 or no friction to overcome, slips away to
 
 95 
 
 a lower angle than it before stood at. 
 Natural springs, therefore, whenever 
 they occur, must be dealt with carefully 
 and completely. Exactly similar effects 
 to those produced by natural springs 
 may result from the defective practice of 
 carrying outlet pipes through or imme- 
 diately under embankments. Be the 
 pipes ever so well puddled, there will be 
 a tendency to trickling along the line of 
 their direction, and assuredly if this 
 trickle makes its way to the center of the 
 bank it will carry mischief with it. It is 
 true that springs are occasionally found 
 issuing from the foot of an embankment, 
 without after several years causing any 
 appearances to justify apprehension. The 
 Doe-park reservoir is an example in 
 point, and though at one time fears for 
 its safety were entertained, the embank- 
 ment is still standing, and, so far as the 
 author is aware, the spr.ng is still trick- 
 ling away. An engineer of eminence was 
 called upon to report upon the state of 
 the works, and gave his opinion that, as 
 the spring came away without any earth
 
 96 
 
 in suspension, there was no mischief 
 taking place, and that the work was in a 
 safe condition. There is no doubt that 
 embankments in this condition require to 
 be narrowly watched, although the pre- 
 sumption may be that, having lasted for 
 several years, they will continue in safety. 
 
 The empirical and unscientific mode of 
 proportioning the length of waste weirs 
 has proved before now a source of danger 
 and destruction to embankments, from 
 the space afforded not being sufficient to 
 discharge the excess water without the 
 surface rising to such a height as to top 
 the embankment. To avoid risk, the 
 stream must be gauged with great care, 
 and the discharge calculated for the 
 greatest known flood ; and if with a 
 given head the length of the weir be 
 adjusted to discharge this amount, or a 
 little in excess, there will be no risk to 
 the embankment. 
 
 Regarding finally the whole subject, 
 the danger that may result from careless 
 or unscientific construction, the large 
 outlay entailed in the establishment of
 
 97 
 
 storage works, and the benefit that may 
 accrue from them whatever their purpose 
 may be, the subject cannot be under- 
 taken on merely rational grounds. Its 
 successful application will rest alone on 
 the study of the question in its scientific 
 details, and an ample practical experi- 
 ence. 
 
 DISCUSSION. 
 
 Mr. H. P. Stephenson said he entirely 
 agreed with the author as to the impro- 
 priety of carrying a pipe through the 
 embankment of a reservoir. He would 
 extend his objection to the passing of a 
 culvert through the embankment. If the 
 culvert were laid on the natural ground, 
 they would avoid the risks pointed out 
 by the author, either of the settlement 
 from the joints of the pipe, or of the 
 water creeping along between the ma- 
 terial and the pipe. He believed that 
 the true principle of construction for 
 reservoirs was the placing of a good 
 puddle dam in the center, and on each 
 side of this dam layers of earth well 
 punned in. One reason why he should
 
 98 
 
 prefer the puddle wall in the center was 
 that there was less tendency in the pud- 
 dle to slip in such a position than when 
 laid on the slope. 
 
 Mr. Albert Latham agreed with Mr. 
 Stephenson in his remarks as to the 
 pipes and culverts ; but he thought it 
 was an open question whether the puddle 
 wall should be in the center of the dam. 
 He had a strong opinion that it should 
 be on the face of the dam. 
 
 Mr. Cargill said that he believed that 
 the reason the puddle wall was not re- 
 quired in Indian embankments, referred 
 to by the author of the paper, was that 
 the earth seemed to have been thor- 
 oughly consolidated by the continual 
 trample of people upon it. That thor- 
 ough-consolidation was the great point 
 in all puddling, and it was on that ac- 
 count that specifications were generally 
 so stringent as to the thickness of the 
 layers of the puddle. As to the position 
 of the puddle wall, he could not see the 
 particular value of having it in the 
 middle of the dam, and he thought that
 
 99 
 
 a far better place for it would be the 
 face, because the object of the puddle 
 wall was to prevent the infiltration or 
 the escape of the water.* This could be 
 effected by puddling the whole slope 
 right down to the permanent strata. 
 The puddle wall was not required to 
 promote the stability of the dam. The 
 question of putting pipes or culverts 
 under the dam required more considera- 
 tion. It was alleged that the putting of 
 a naked pipe through the dam of the 
 Bradfield reservoir was one of the causes 
 of its bursting. In some very large, 
 waterworks now being constructed in 
 Dublin there were two distinct sets of 
 main pipes, and they were laid in two 
 large culverts at the bottom of the dam. 
 The culverts were large enough for a 
 man to walk upright in them. If the 
 foundation were well looked after, there 
 
 *.Asthe center wall, of puddle or masonry, is the 
 citaael of the dam, it would seem clear that it sbould 
 be protected by being placed within the embankment. 
 Placing the puddle on the outside, in the shape of a 
 face covering, would seem to invite such a disaster 
 as that mentioned a«. having occurred at Sheffield.
 
 100 
 
 would be no fear of the arch or dome of 
 the culvert giving way in consequence 
 of any inequality of pressure above it, as, 
 if proi3erly constructed, an arch would 
 stand any amount of pressure short of 
 what would crush the material. 
 
 Mr. Baldwin Latham said he could 
 not agree with Mr. Jacob that a dam 
 could not be constructed from theo- 
 retical deductions ; for unless regard was 
 paid to theoretical considerations there 
 might result either a deficiency of 
 strength or a waste of material and 
 labor. In the dam shown in the draw- 
 ings, and designed by himself, the pipe 
 did not run through, but on the outside 
 of the dam, on the solid ground. It was 
 a well received opinion among enghieers 
 that if you had a pipe or culvert running 
 through an embankment, that pipe or 
 culvert would be unsafe. He believed 
 that well made and properly tested pipes 
 were quite as safe as culverts when in 
 the solid ground. A pipe was simply a 
 small culvert made of iron instead of 
 brickwork. In cases in which there was
 
 101 
 
 a tendency for the water to creep along 
 the outside of the pipe, that might be 
 stopped by having projecting flanges on 
 the pipe. The same creeping of water 
 might take place along a culvert as along 
 a pipe. With regard to the slope of a 
 dam, the inside slope should be greater 
 than the outside slope, because the 
 greater would be the stability of the 
 dam, and the water would have less de- 
 structive effect on the dam ; he had 
 effectually prevented leakage by the use 
 of socket-pipes. The square projection 
 of the sockets was always presented to 
 the reservoir, and the pipes were laid in 
 the virgin ground. It was very bad 
 practice to lay the pipes in made ground, 
 and especially through a dam. Pipes 
 laid under a dam should be tested under 
 pressure after being laid and before 
 being covered up, so that any defective 
 joint might be discovered. In cases in 
 which he had laid pipes through dams, 
 they had been so tested, which resulted 
 in good and effective work ; but he was 
 bound to say that, if the pipes had not
 
 102 
 
 been tested in situ the result would not 
 have been satisfactory. 
 
 Mr. Scbonheyder said that Mr. Jacob 
 had said that wherever springs occurred 
 they should be well carried away. He 
 (Mr. Schouheyder) wished to know how 
 a spring was to be prevented from dif- 
 fusing through the earth. 
 
 Mr. Hendry said that he had seen pipes 
 which were laid through embankments, 
 but had never seen one that was per- 
 fectly tight. It was almost impracticable 
 to make it so, owing to the continuity of 
 the puddle being disturbed at the point 
 where the pipe passes through. 
 
 The chairman asked what was the 
 largest diameter of pipe Mr. Hendry had 
 seen used. 
 
 Mr. Hendry replied that the largest 
 was 18 in. He had heard of several 
 methods being tried, but he did not 
 think it was possible to prevent leak- 
 ing, more or less, from the leservoir 
 along the outside of the pipe. He 
 should like to be informed how it was 
 possible to connect the puddle with the
 
 103 
 
 pipe ; if the pipes be laid in the natural 
 ground below the foundation of the em- 
 bankment, then there is no fear of leak- 
 age, provided the pipes are properly laid. 
 Mr. Jacob, in replying to the discus- 
 sion, said, that in the opinions that had 
 been expressed there were but few jooints 
 of disagreement with those that he him- 
 self held. He could not agree with Mr. 
 Latham in his belief that embankments 
 could be calculated on mathematical 
 principles. In order to deal with em- 
 bankments theoretically, they must be 
 regarded as rigid masses, and be assumed 
 to rest upon a horizontal plane. It 
 could be shown mathematically that a 
 rigid body of the same specific gravity 
 as ordinary earth need not present the 
 same section as is usually given to 
 embankments, in order adequately to 
 resist the pressure of water. A right- 
 angle prism with the hypothenuse rest- 
 ing upon the plane would be quite suffi- 
 cient to resist the pressure of water, 
 even supposing the surface of the water 
 to coincide with the upper edge of the
 
 104 
 
 prism. The reason of giving long slopes 
 to an embankment is discoverable from 
 the fact that banks, when exposed to the 
 action of water, are found to waste and 
 slip away to such an angle as will with- 
 stand the action of the water. The chief 
 reason of the failure of embankments is 
 the infiltration or soaking of the water 
 from the inner side, which renders the 
 material semi-fluid and causes it to sub- 
 side into a smaller space than it origin- 
 ally occupied. The superincumbent 
 mass then sinks and allows the water 
 to overtop the embankment. The earth 
 used for making embankments in the 
 Deccan and in parts of the Madras Presi- 
 dency in India is of a most suitable 
 quality for the purpose. It is what is 
 called " black soil," being very dark in 
 color, and of a highly argillaceous charac- 
 ter. The color is, no doubt, due to the 
 presence of carbon. The clay makes 
 most excellent puddle ; but, no doubt, the 
 consolidation produced by the tread of 
 the work-people is the real secret of the 
 earth resisting the pressure of water so
 
 105 
 
 successfully as it does. In North 
 America, the levees for protecting the 
 country from flooding by the Mississippi 
 are sometimes constructed simply of sand; 
 and are found, for the most part, suffi- 
 cient for their purpose. A s regards carry- 
 ing away springs from the seat of an em- 
 bankment, there is no difficulty in ascer- 
 taining where they exist when the ground 
 is laid bare, as they are generally well- 
 defined streams. Before the earthwork 
 is commenced it is necessary to construct 
 drains of masonry, or brickwork, or to 
 lay iron piping to carry away the water 
 clear of the work.* 
 
 The chairman said that the paper of 
 Mr. Jacob was a very interesting one, 
 and the subject was one which, during 
 the last year or two, or, he might say, 
 within the last week or two, had com- 
 manded the attention of the whole body of 
 engineers. Last session a special Act of 
 Parliament was passed, that all reservoirs 
 and embankments should be constructed 
 
 * We are again compelled to ask : Where to ?
 
 106 
 
 to tbe approval of the Board of Trade. 
 The subject of irrigation in India, which 
 was alluded to in the paper, was one of 
 vital importance. There was no question 
 that the only means we had of irrigating 
 that country in an efficient manner was 
 ,by the construction of reservoirs.
 
 107 
 
 Additional Remarks by the American 
 Editor. 
 
 The importance of storage reservoirs 
 for the purpose of equalizing the flow of 
 water furnished by streams, is so gener- 
 ally recognized that they must be con- 
 sidered as forming essential parts of all 
 well-planned systems of water supply, 
 excepting those maintained by bodies of 
 water of such large relative dimensions 
 that their minimum flow in the driest 
 seasons exceeds the maximum con- 
 sumption. 
 
 The admirable paper of Mr. Jacob, 
 which forms the first part of the present 
 volume, covers the ground embraced by 
 its title, and, together with the subse- 
 quent discussion, forms a body of infor- 
 mation of the highest interest and value. 
 The art of hydraulic engineering, how- 
 ever, is a progressive one, and at the 
 present day there is much which may be
 
 108 
 
 added to the original paper, particularly 
 from the point of view of American en- 
 gineering, which it may be fairly claimed 
 is likely to enhance its usefulness. 
 
 In planning storage reservoirs, there 
 are two questions which naturally com- 
 mand attention from the very start: 
 How much stored water does our supply 
 require ? And how much can our available 
 resources be counted upon to furnish ? 
 In answer to the first of these questions, 
 I cannot probably do better than quote 
 the following statement made by Mr. 
 Pole (Proceedings of the Institution of 
 Civil Engineers, 1884-5) as regards Eng- 
 land: "The general judgment of ex- 
 perienced practitioners appears to be, 
 that for large rainfalls, a storage of 150 
 days' supply, or even less, will sufi&ce ; 
 but in drier districts it may be necessary 
 to go as high as 200 days." This rule 
 will, I think, hold good for this country 
 also, and we may, as a general thing, con- 
 sider that a water supply which compre- 
 hends a storage equivalent to 150 days' 
 consumption, is in a good position for
 
 109 
 
 the maintenance of a constant supply at 
 all seasons in districts enjoying an aver- 
 age rainfall. It will be understood that 
 this amount of storage applies to cases 
 where it is proposed to use the whole, 
 or greater part, of the total yield of the 
 stream. 
 
 As to the second question, the answer 
 depends upon the minimum annual rain- 
 fall over the basin drained by the 
 given stream, and the amount of this 
 rainfall— deduction made for losses by 
 evaporation, percolation, etc., as well as 
 for consumption, which is available for 
 storage. In regard to this matter, Mr. 
 Jacob dwells upon the necessity of ascer- 
 taining the amount of annual rainfall, 
 and also of carefully gauging the stream, 
 in order to establish its discharge at dif- 
 ferent seasons of the year. Now, it is 
 evident that these operations, to be of any 
 real value, require a long period of time 
 for their accomplishment, and can be 
 therefore, in the great majority of cases, 
 of only pai'tial utility. Fortunately, we 
 can frequently dispense with any con-
 
 110 
 
 sideration of this point, for our stream 
 may be of such size, and may drain so 
 large an area, that we need not be at all 
 troubled as to its ability to fill our j^ro- 
 posed reservoir to the desired capacity. 
 On the other hand, however, cases often 
 present themselves when it becomes a 
 question, if we build a reservoir to con- 
 tain the required amount of water, 
 whether we can be sure of filling it every 
 year. It is not safe, as a general rule, 
 and in average locations in this country, 
 to count on more than 12 inches of avail- 
 able annual rainfall, for storage purposes, 
 and even this limit, if great interests are 
 at stake, should be approached with 
 caution. Twelve inches of rainfall will 
 furnish 27,878,400 cubic feet, or a little 
 more than 208.5 million U. S. gallons per 
 square mile of drainage area, and, for a 
 round number, 200 million gallons per 
 square mile may be considered as the maxi- 
 mum that it would be safe to count on, 
 although it is very probable that there 
 would be times every year when water 
 ran to waste over the spill- way, indicat-
 
 Ill 
 
 ing that the capacity of the reservoir was 
 not as great as the flow of the stream 
 would warrant. It might therefore be 
 wise, in cases where it was important to 
 store every available gallon, and worth 
 while to spend a good deal of money to 
 make sure of doing so, to increase the 
 relative capacity of the reservoir. 
 
 In the Ci'oton basin the average yearly 
 precipitation is almost exactly 46 inches. 
 The very careful observations made by the 
 Croton Aqueduct Department, and ex- 
 tending over a long period of years, indi- 
 cate that in this basin each square mile 
 of drainage area will furnish an average 
 water supply of at least one million 
 gallons per day. In districts of similar 
 geological character, and with equal 
 average yearly rainfalls, the same rule 
 may be considered to hold good. Under 
 such circumstances, a reservoir capacity 
 of 200 million gallons per square mile 
 would furnish storage for 200 days' 
 supply. 
 
 It will be seen that the three points 
 governing the question of storage capa--
 
 112 
 
 city are — the drainage area, the available 
 rainfall, and the daily supply which it is 
 desirable to maintain. 
 
 These preliminary questions being 
 settled, the next point coming up for 
 decision will be, the best location for the 
 dam which is to form the reservoir. The 
 approximate location can generally be 
 readily selected : The natural features 
 of the ground usually clearly indicate the 
 point near which the dam should be built. 
 In order to determine the exact position, 
 the whole probable area should be cross- 
 sectioned in 20 feet squares (with addi- 
 tional points where needed). This work 
 will not only show the precise line where 
 the longitudinal section of the dam has 
 the smallest area, which, other things 
 being equal, will be the best line, but 
 will also preserve a record of the original 
 surface of the ground, from which the 
 amount of work done at any time can be 
 readily estimated. 
 
 The height and location of the dam 
 being thus settled, the next important 
 question is, What sort of a dam shall we
 
 113 
 
 build ? It will be observed that in Mr. 
 Jacob's paper, he confines himself to the 
 consideration of earthen dams only. 
 These dams can usually be built cheaper 
 than masonry dams, and have this ad- 
 vantage, that they admit of being built 
 in places where masonry dams would be 
 unsafe. It will very seldom be found 
 advisable to build a high masonry dam 
 on anything but a foundation of solid 
 rock. When such natural foundation is 
 not found, and it is still determined to 
 build a masonry dam, the only resource 
 in order to secure safety, is to carry the 
 foundations down to a depth which will 
 very greatly increase the cost of the 
 work. 
 
 The dams spoken of by Mr. Jacob are 
 of a type very prevalent in England and 
 India : namely, earthen dams with a 
 puddle core. Such dams are rare in the 
 United States, where the puddle core is 
 commonly replaced by a masonry center 
 wall. The object in either case is, 
 primarily, to establish an impervious cut- 
 off against any possible percolation
 
 114 
 
 through or under the earthern embank- 
 ment. I think there can be no doubt as 
 to the superior efficiency of the masonry 
 wall, and I doubt if there is as great an 
 economy in the puddle, as would at first 
 appear probable. In the first place, 
 proper material for good puddle is not 
 always obtainable, and even when it is 
 found in abundance near by, its proper 
 preparation and placing are matters re- 
 quiring a good deal of careful and ex- 
 pensive manipulation. Good puddle 
 should resemble, when made and placed, 
 in character and composition, an unburnt 
 brick. When we read, as we do in Mr. 
 Jacob's paper, and elsewhere, of the great 
 precautions necessary in putting in a 
 puddle wall, and the disastrous conse- 
 quences attendant upon some appar- 
 ently trifling neglect in doing so, I think 
 most engineers intrusted with the design- 
 ing of so important a work as a large 
 storage reservoir, would hesitate before 
 risking these consequences in order to 
 effect an economy which perhaps might 
 eventually prove to be not so great as 
 was anticipated.
 
 115 
 
 There is another particular in which 
 the masonry center wall presents a great 
 superiority to the puddle core. One of 
 the weakest points of an earthen dam lies 
 in the neighborhood of the conduit used 
 for drawing off the water. A perusal of 
 Mr. Jacob's paper and the discussion 
 shows what importance is attached to 
 having this conduit form a perfectly water- 
 tight connection between the inside and 
 outside of the reservoir. It is clear that 
 the existence of solid and water-tight 
 masonry center wall, running through 
 the entire length of the dam, and firmly 
 and deeply imbedded in the banks of the 
 valley on each side, affords an excellent 
 opportunity for making all the necessary 
 connections in a satisfactory manner. 
 The masonry culvert through which the 
 water passes, and the gallery containing 
 the pipes, may be bonded in with the 
 center wall and made to form a part of it 
 in such a way that the possibility of water 
 following along the outside of these 
 structures is wholly precluded, and one 
 chief danger of destruction of the dam 
 entirely averted.
 
 116 
 
 Having now decided upon the character 
 of our proposed dam ; viz : — an earthen 
 dam with masonry center wall, or cut off, 
 it is next in order to consider its design, 
 dimensions and accessories. 
 
 As regards the latter, the spill-way 
 merits the first mention. The dimen- 
 sions of this portion of the dam must be 
 ample, and sufficient to safely pass all the 
 water which may come to it in times of 
 heaviest freshet, without the possibility 
 of its overtopping the dam. As to the 
 proper length of the spill-way, it is per- 
 haps impossible to lay down any fixed 
 rule, or to attempt to make it a given 
 function of the water shed.* If, in design- 
 ing a dam, we find any existing dams in 
 the neighborhood, upon the same stream, 
 or, failing this best indication, any rail- 
 road bridges through which the whole of 
 the stream has to pass, we may have a 
 good opportunity of ascertaining the 
 amount of water passing off in freshets, 
 and proportion our spill- way accordingly. 
 If none of these indications are to be 
 
 * See page S7.
 
 117 
 
 found, the next best course to pursue 
 (and this should be done in any event, 
 as a check), is to ascertain the dimensions 
 of the spill- way of some existing dam, 
 built elsewhere, but upon a stream of 
 about the same drainage area as the one 
 in question, and, if possible, of the same 
 character. 
 
 "When a natural spill-way cannot be 
 found, such as Mr. Jacob speaks of, and 
 which, of course, is always preferable, its 
 construction of stone masonry is always 
 a very expensive piece of work, and the 
 tendency is therefore to reduce the length, 
 and provide for a deep wave passing over 
 it. This kind of economy should not be 
 pushed too far ; it is better and safer to 
 provide a long spill-waj^ over which a 
 comparatively thin sheet of water shall 
 pass. 
 
 The dimensions of this important part 
 of our dam having been settled, the next 
 point to consider is what relative 
 position it should occupy. If the sides 
 of the valley are of rock, even if not of 
 the very soundest and most solid char-
 
 118 
 
 acter, there will be an economy in placing 
 the spill-way at one end of the dam, as 
 its height will thereby be diminished. If, 
 however, the sides are of earth, the safest 
 place for it is directly in line with the 
 stream, as it would be dangerous to dis- 
 charge a large volume of water upon the 
 unprotected hillside, and any attempt to 
 protect it against wash, by means of walls 
 and paving, will generally involve an ex- 
 pense equal, or nearly so, to the higher 
 structure. This is a point, however, 
 which is by no means to be taken for 
 granted, and in cases where great 
 economy of construction is necessary, the 
 ground should be carefully examined,and 
 comparative estimates made. 
 
 As regards the form of the spill-way, 
 the curved form adopted for the old 
 Croton dam, and the dam on the Bronx, 
 at Kensico, N. Y., is no doubt the best, 
 but its great cost in the way of cut stone 
 voussoirs will generally preclude its use 
 except for municipal work, and where 
 economy is studied, the form adopted 
 will generally be that of steps, or offsets.
 
 119 
 
 A good form of spill- way is shown in Fig. 
 72, page 382, of Tanning's " Water Sup- 
 ply Engineering." A very good and 
 massive form, which I have had occasion 
 to adopt with some modifications for a 
 spill- way about 30 feet high, is described 
 as follows : Back, vertical ; top width, 
 7 feet ; batter on face, an inch and a half 
 to the foot, for 8 feet, making a thickness 
 of 8 feet at the bottom. Thence, steps 
 ranging in height from 15 to 22 inches, 
 and following a general slope of 45 de- 
 grees. In this way, it will be perceived 
 that at any given elevation, the thickness 
 of the wall is always equal to its height 
 above such elevation. 
 
 The waste pits and sluices spoken of 
 by Mr. Jacob as substitutes for a waste 
 weir, are not, I think, to be recom- 
 mended. Nothing can be better than a 
 straight opening and clear escape for the 
 surplus water. 
 
 As regards the means adopted for 
 drawing off the water, they should be ns 
 simple as possible, with good facilities f )r 
 inspection and repair. Those menti^^f I
 
 120 
 
 by Mr. Jacob seem to be unnecessarily 
 complicated. My own opinion is strongly 
 in favor of having a cast-iron pipe or 
 pipes built into the center wall, and ex- 
 tending outside of the dam, in an arched 
 gallery, founded in the natural formation 
 sufficiently large to admit of free circula- 
 tion all about the pipes. Within this gal- 
 lery, or gate chamber, are the gates, two 
 upon each pipe, the inner one, or one 
 nearest the water, to be kept habitually 
 open, and the delivery of water regulated 
 by means of the outer one. If any accident 
 occurs to this gate, the inner one is 
 closed, and it can then be got at for re- 
 pairs. On the inside, a tower is built, 
 one side of which is formed by the center 
 wall through which the pipes run. This 
 tower, or well, is rectangular in shape, 
 and its sides are from one foot to 
 eighteen inches outside of the pipes ; 
 that is, its inside width is from 2 to 3 
 feet greater than the outside diameter of 
 tli(^ reducer of the largest pipe, so as to 
 afford a chance to lead and calk the re- 
 ducer when set in the cut ring stones
 
 121 
 
 which surround it. The length of the 
 tower may be from 10 to 15 feet. It con- 
 tains two sets of cut stone grooves, in which 
 stop-plank can be placed. By this means, 
 should it ever become desirable to get at 
 the mouth of the pipes or reducers with- 
 out first emptying the reservoir, the stop- 
 plank can be placed, and, if necessary, the 
 space between them puddled, and a water- 
 tight coffer dam is thus made, inside of 
 which work can be carried on while the 
 reservoir is full of water. From this 
 tower, a masonry culvert, also provided 
 with stop-plank tower, if thought neces- 
 sary, or an open passage, with wing 
 walls, extends through the embankment. 
 The gate chamber, tower, wing walls and 
 spill- way should be, if possible, grouped 
 together, with a view to economy of ma- 
 terials and increased strength, and an 
 effort should be made to so locate the 
 work that these important features should 
 be set upon the most favorable natural 
 foundation. 
 
 As regards the center wall of masonry, 
 it is, of course, impossible to lay down
 
 122 
 
 hard and fast rules for its proportions, 
 any more than for the other parts of the 
 dam. But certain general principles 
 may be established, to do which it is first 
 necessary to get a clear idea of the 
 precise function of the center wall, and 
 of the part it is to play in the dam. As 
 we have seen, it is intended, primarily, to 
 afford a water-tight cut-off, to arrest 
 any percolations which may reach it, by 
 trickling through the bank. Indeed, we 
 may consider the center wall as constitut- 
 ing the dam proper, for it is to the center 
 wall that we finally look for the retention 
 of the water within the reservoir. Re- 
 garded in this way, the earthen embank- 
 ments on each side are only provisions 
 for keeping the center wall from being 
 thrown down. In point of fact, however, 
 we do expect more than this from the 
 embankments. We expect the inner em- 
 bankment to be very nearly impervious, 
 and of itself to be almost, if not quite, 
 suflScient for the retention of the water. 
 In any event, we expect it to be a power- 
 ful auxiliary to the center wall, by keep-
 
 123 
 
 ing the deepest water well back from it, 
 and increasing the distance that a given 
 drop of water would be obliged to travel 
 in order to pass underneath the wall. In 
 this way, the embankment is equivalent 
 to a deepening of the foundation of 
 the wall. Moreover, water reaching 
 the center wall by traversing the em- 
 bankment, comes to it in the form 
 of a percolation, modified by capillary 
 action. The object of the exterior bank 
 is mainly to keep the wall from being 
 thrown out, but it serves an excellent 
 purpose also in smotheiing down any 
 slight percolations issuing from under- 
 neath the wall, and still further increases 
 the distance that a given drop of water 
 must travel in order to pass freely from 
 the inside to the outside of the reservoir. 
 As to the proper height to give to the 
 center wall, although it is highly probable 
 that in many cases it is not necessary to 
 carry it up to the full height of the sur- 
 face of the water, when the reservoir is 
 full, we cannot say that an earthen dam is 
 absolutelv safe unless the wall is carried
 
 124 
 
 at least as high as the Up of the spill- 
 way. It is still better, in very high dams, 
 to raise it as high as the extreme eleva- 
 tion of flood freshets. As to its thick- 
 ness, economy will prompt a reduction to 
 the narrowest limits, but it must not be 
 forgotten that, although it depends on the 
 banks for support, rather than to its own 
 moment of inertia, yet, owing to the fact 
 that some movement may take place in 
 the banks themselves, it is liable to be 
 subjected to unbalanced pressures, and 
 should therefore be, to some extent, self- 
 supporting. If we start with a top thick- 
 ness of 5 feet, and increase the thickness, 
 by off-sets, to the extent of 2 feet addi- 
 tional, every 10 feet as we go down, we 
 shall have dimensions which, in a great 
 majority of cases, will be abundant to 
 afford the required strength. 
 
 A still more important question in re- 
 gard to the center wall, is the depth to 
 which its foundations should be carried. 
 For this there is absolutely no rule, 
 and the question must be decided accord- 
 ing to the engineer's best judgment for
 
 125 
 
 each particular case. Of course, if a rock 
 foundatiou is encountered, the problem is 
 greatly simplified, for we have only to 
 remove the loose and disintegrated sur- 
 face, until we come to clean, live rock, 
 and place our bottom course upon that. 
 The uncertainty occurs when our test 
 pits reveal the existence of coarse and 
 permeable strata, when our only resource 
 is to carry our wall well down. In some 
 cases such walls have been put down to 
 such a depth that in places the distance 
 below ground is greater than the height 
 above, and frequently sheet pihng has 
 been resorted to. This, however, it is 
 best to avoid, if possible. 
 
 In general, it may be said that, the 
 finer the material, the better it is adapted 
 to serve as a footing for the center wall. 
 Clay, fine gravel and sand, and quick- 
 sand when found at a sufficient depth to 
 prevent its being forced up at the sides, 
 are all excellent materials to build such a 
 wall upon. Where a compact, water- 
 tight material is found, it is needless to 
 carry the footing courses to any great
 
 126 
 
 depth below the surface of the ground, 
 but when loose cobbles, coarse gravel and 
 shingle, or any other ground affording 
 an easy passage for water, is encountered, 
 the only safety is in depth of foundation. 
 It now remains to say a few words 
 respecting the earthen embankments. 
 On page 85 Mr. Jacob gives a little table 
 of the proper heights for the top of em- 
 bankments above the crest of the spill- 
 way, or waste weir. It is very important 
 that this height should be amply suffi- 
 cient to prevent the embankment being 
 overtopped during a freshet. Of course, 
 the length of spill-way is a factor in this 
 matter, but the top of the embankment 
 should be kept well up, and I should be 
 disposed to add a foot, or even two, to 
 the figures given by Mr. Jacob, making 
 them respectively 5, 7 and 8 feet, and 
 I would change the maximum from 10 to 
 ] 2 or 14 feet. The top width may vary, 
 according to the height of the dam, from 
 5 to 20 or more feet. Perhaps a fair ap- 
 proximate rule would be, to give the top 
 of the embankment such a width that,
 
 127 
 
 taken in connection with the height above 
 the crest, or Hp of spill-way, and the 
 slopes, would give at the level of the lip, 
 a thickness at least equal to the greatest 
 depth of water in the reservoir. 
 
 For the slopes, much depends upon 
 the material used for embankments. If 
 a good quality of earth is to be found, 
 such as a sandy loam, containing a good 
 proportion of clay and few stones, great 
 thickness of embankment is not necessary. 
 If the quality of the stuff is inferior, it 
 must be made up for in quantity. Gener- 
 ally slopes of from 2^ to 3^ to 1 for the 
 inside, and 2J to 1 for the outside are 
 found sufficient. An excellent profile is 
 obtained, with good material, by starting 
 from the top with a slope of 2^ to 1 
 on both sides, and at about half-way 
 down introducing a level berme on the 
 inside, from which a slope of 3 or 3^ to 
 1 is continued down to the foot of the 
 embankment, the outside slope being 
 kept at the same slope of 2^ through- 
 out. 
 
 The manner in which the embankment
 
 128 
 
 is formed, is a matter of great import- 
 ance. It is necessary to put it in in such 
 a manner as to minimize the subsequent 
 settlement. It must not, therefore, be 
 carried along like a railroad embank- 
 ment, but brought up in horizontal 
 layers from the bottom. Generally 
 speaking, if the material be brought on 
 in carts and wagons, and kept constantly 
 moist by sprinkling, the travel of the 
 vehicles and animals will be suflScient to 
 secure the necessary degree of compact- 
 ness. The face of the interior slope 
 should be well pitched, or rip-rapped, as 
 recommended by Mr. Jacob, and the 
 exterior slope sodded, or sown to grass. 
 To return to the center wall, in order 
 to say a few words as to the manner in 
 which it, and in general all the masonry 
 work of the dam, should be executed. 
 The first essential of all hydraulic ma- 
 sonry is, that it shall be as nearly as 
 possible water-tight. In order to effect 
 this, there must be no vacancies what- 
 ever between any two contiguous stones, 
 but the wall must be so perfectly laid
 
 129 
 
 that whatever is not stone, is mortar, and 
 compact mortar. In order to secure 
 perfect work, the stones used must be 
 clean and bright, with sharp quarry- 
 faces. All soiled and dirty stones should 
 be, if possible, discarded, and if not, they 
 should be thoroughly cleaned, by brooms 
 or brushes, or by washing, before using, 
 for the mortar will not properly adhere 
 if there be a loose coating of any foreign 
 substance upon the surface of the stone. 
 If necessary, they should be wet before 
 laying, but frequently, particularly when 
 Portland cement is used, the mortar 
 throws off so much water that the chief 
 difficulty is to keep the work sufficiently 
 dry. When a piece of wall is to be 
 worked upon, it should be carefully 
 cleared of all loose and dry mortar; in- 
 deed, it is often necessary to keep men 
 employed constantly in sweeping off the 
 wall in order not to delay the masons in 
 cleaning it up. This point should always 
 be insisted upon, and in summer there 
 should be a number of large watering- 
 pots on the work, to sprinkle the surface
 
 130 
 
 after it has been swept off. Great care 
 should also be taken to have the mortar 
 properly mixed — a point too often neg- 
 lected. The sand and cement should be 
 thoroughly mixed together when dry> 
 by repeated turnings, until the mass has 
 become perfectly homogeneous, and of 
 an uniform color. 
 
 It is generally a provision of the speci- 
 fications, that stones must not exceed a 
 certain limit in size — perhaps 8 or 10 
 cubic feet — it being supposed that small 
 stones will be more easily bedded than 
 large ones. I think that this is at least 
 doubtful, and that stones measuring a 
 cubic yard, or even a yard and a half, can 
 be properly bedded with very little addi- 
 tional trouble. The gain in rapidity of 
 execution, when larger stones are used, is 
 very great, as a derrick can swing a large 
 stone as quickly as a small one, and there- 
 fore many more yards can be laid in a 
 day. Large stones also bed themselves 
 more perfectly, from the greater force 
 with which they compress the mortar 
 under them, and drive out the super-
 
 131 
 
 flaous water. Of course, stones should 
 not be allowed of such dimensions as to 
 run through the wall from front to rear, 
 and it is always well to reserve the right 
 to discard stones above, say, 10 cubic 
 feet, by a clause to that effect, in the 
 specij&cations. 
 
 After a stone has been set, it should 
 be tested with bars, so as to see if it is 
 entirely clear of everything, and swims 
 on its bed, without rocking. No two 
 stones, however small, should be allowed 
 to touch each other without the inter- 
 vention of compact mortar. When there 
 is any doubt as to whether a stone is 
 properly bedded, it should be raised, so 
 as to see if the mortar under it has taken 
 a good impression. It is frequently 
 necessary to raise a stone a second and 
 even a third time, before it can be pro- 
 nounced to be satisfactorily set, particu- 
 larly with inexperienced masons. The 
 masons should, however, soon become 
 accustomed to properly spreading the 
 beds, so that a majority of the stones 
 can be set by first intention. All work
 
 132 
 
 should be done with cranes or derricks, 
 so that stones can be dropped in their 
 places, and readily lifted again, when 
 required. All spalls and small stones 
 should be settled in their beds by a blow 
 with the hammer. This scrupulous care 
 in laying the masonry does not consume 
 as much time as might be anticipated. 
 There should be no trouble, with good 
 and shapely stones, in laying from 30 to 
 35 cubic yards per day, per derrick 
 (steam), with say six masons and a suffi- 
 cient number of helpers, and yet take all 
 the precautions — and many more — above 
 indicated. Of course, to effect this, the 
 work must be well systematized, and the 
 derricks constantly fed with material, so 
 that the masons are never kept waiting. 
 In putting down foundations, care 
 must be taken to provide abundant 
 pumping power. All pumping sumpts 
 should be put down well below the 
 bottom of the trenches they are designed 
 to drain, otherwise these latter will never 
 be kept properly dry. The sumpts 
 should also, if possible, be placed entirely
 
 133 
 
 outside the lines of the masonry, so as to 
 be out of the way, and not draw the ma- 
 terial from under work already laid. 
 
 The management of the water is always 
 one of the greatest difficulties connected 
 with this class of work. The foundations 
 extend entirely across the valley of the 
 stream, and provision must be made for 
 the passage of the water while the work 
 is in progress. Very frequently, partic- 
 ularly in the case of small streams, the 
 difficulty can be met by first building 
 the gate chamber and setting the pipes 
 and gates, and then turning the stream 
 through them, thus getting rid of it at 
 once. In larger streams this is not 
 always practicable, as we may not be 
 sure that the pipes will carry all the 
 water in case of a freshet. In such 
 cases, we may be obliged to omit setting 
 the pipes till after the completion of the 
 work, allowing the water to pass through 
 the gate chamber and gallery in which 
 they are subsequently to be set. When 
 this course is pursued, the opening in 
 the center wall must be so planned and
 
 184 
 
 built that it can be quickly and perfectly 
 closed, when the time comes, by cut-stone 
 pieces carefully dressed to fit. When 
 the dam is otherwise completed, the stop- 
 plank can be placed, and the water held 
 back while the pipes and gates are set. 
 In cases where the stream is so large 
 that even these means are not sufficient, 
 the problem becomes exceedingly diffi- 
 cult, and demands very careful special 
 study. 
 
 Mr. Jacob dwells much upon the 
 danger caused by the existence of springs 
 underneath the embankment, and recom- 
 mends that they be "taken up and carried 
 away in proper drains sufficiently and 
 securely puddled." Now, unfortunately, 
 this is easier said than done. The site 
 of an embankment is frequently covered 
 by a multitude of small springs, the 
 sources of which it is impossible to 
 locate, and to which they cannot be 
 traced. Besides, where should we carry 
 them to? Probably the best way of 
 dealing with such springs, is to get the 
 site of the inner embankment well
 
 135 
 
 stripped down to the live earth, and then 
 commence placing the embankment next 
 to the center wall (not, of course, carry- 
 ing it up so high as to create a dump, 
 such as we have already seen should be 
 avoided,) and then advance it out from 
 the wall, toward the toe of the slope, in 
 the hopes of smothering down the 
 springs, and driving them, to some ex- 
 tent at least, back from under the em- 
 bankment. Indeed, it is difficult to see 
 what else can be done, and it may be 
 fairly doubted if the existence of such 
 springs is fraught with such serious 
 menace to the work as Mr. Jacob men- 
 tions. Very large springs might occa- 
 sion trouble, if the weight and compact- 
 ness of the bank do not force them back, 
 and, if possible, such should be traced to 
 their source, and diverted within the 
 reservoir. Unfortunately, these large 
 veins of water are frequently fed by a 
 number of separate springs, which finally 
 run together, and if the flow be stopped 
 in one point, it breaks out in another. 
 Large springs, such as have been just
 
 136 
 
 ni( ntioned, often give great trouble in 
 closing the gap in the foundation of the 
 center wall, which has been left for them 
 to flow through temporarily, the diffi- 
 culty being of the same class, though less 
 in degree, as that occasioned by the main 
 stream itself. They may sometimes be 
 passed over from side to side, until a 
 point is reached, above which they do not 
 rise in any considerable volume. Some- 
 times they are best handled by reducing 
 the width of the gap progressively, until 
 it becomes a very narrow passage, and 
 then closing it boldly, working in the 
 water with mortar made of neat cement, 
 without any admixture of sand. There 
 can be no rule laid dowQ for the treat- 
 ment of such cases, which must be 
 handled as best we may, the only thing 
 necessary and sufficient being to get the 
 wall so built that the water does not, 
 finally, wash out the mortar and run 
 through it. If we fail in accomplishing 
 this in one way, there is nothing for it 
 but to take out the defective part and 
 try some other means, never leaving the
 
 137 
 
 job till the essential object aimed at has 
 been accomplished. 
 
 In conclusion, it may be said that the 
 building of dams and reservoirs belongs 
 to one of the most important and re- 
 sponsible classes of work undertaken by 
 the engineer, and no pains should be 
 spared in their construction to ensure 
 satisfactory results. It should be borne 
 in mind that the first essential in a dam 
 is stability; the second impermeability. 
 At the present day, with all past experi- 
 ence as a guide, there should be no 
 uncertainty as to the realization of the 
 first of these essentials ; the perfect ful- 
 fillment of the second cannot always be 
 so surely counted upon, because it de- 
 pends not only upon the quality of our 
 work — which is a controllable factor — 
 but also upon the character of the ground 
 itself on which the work is placed, which 
 may present disadvantageous features 
 that we cannot wholly overcome. But 
 by care and the exercise of sound judg- 
 ment these adverse conditions can gener- 
 ally be modified to the extent of securing 
 satisfactory, if not perfect results.
 
 UCSB LIBRARY
 
 FACIUTV 
 
 'b 000 007 947 5