OFFICIAL DONATION. Water-Supply and Irrigation Paper No. 140 Series 0, Underground Waters, 43 DEPARTMENT OF THE INTERIOR UNITED STATES GEOLOGICAL SURVEY CHARLES D. WALCOTT, Director FIELD MEASUREMENTS OF THE RATE OF MOVEMENT OF UNDERGROUND WATERS BY CHARLES S. SLICHTER WASHINGTON GOVERNMENT PRINTING OFFICE 1905 Water-Supply and Irrigation Paper No. 140 Series 0, Underground Waters, 43 DEPARTMENT OF THE INTERIOR UNITED STATES GEOLOGICAL SURVEY CHARLES D. WALCOTT, Director s33~ V~ FIELD MEASUREMENTS OF THE RATE OF MOVEMENT OF UNDERGROUND WATERS BY CHARLES S. SLIGHTER WASHINGTON GOVERNMENT PRINTING OFFICE 1 9 5 2d. Set 13 NOV 1905 D.ofD, * • « ■ 4 ' " • • • - « • \ CONTENTS. Page Letter of transmittal 7 Introduction 9 Chapter I. The capacity of a sand to transmit water 10 Factors influencing flow 10 Transmission constant 10 Chapter II. Underflow meter used in measuring velocity and direction of movement of underground water 16 Types of apparatus 16 Test wells 16 Direct- reading instruments 19 Recording instruments 25 Chapter III. Laboratory experiments on the flow of water through sands and gravels 29 Objects of the experiments 29 Experiments in the horizontal tank 29 Experiments in the vertical tank 41 Accuracy of the electric method of determining the velocity of ground waters 49 Chapter IV. Measurements of the underflow at the narrows of Rio Hondo and San Gabriel River, California 50 Chapter V. Measurements of the underflow at the narrows of Mohave River near Victorville, Cal 55 Conditions at the station 55 Description of experiments 57 Quality of water % 63 Chapter VI. Measurements of the rate of the underflow on Long Island, New Yoik 65 Conditions existing at the stations 65 Influence of the rainfall upon the rate of motion of ground waters 69 Seepage waters from ponds and reservoirs 72 Influence of pumping upon the rate of motion of ground waters near some of, the Brooklyn driven-well stations 81 Conclusion 85 Chapter VII. Specific capacity of wells 86 General principles 86 Tests „ 91 Test I 91* Seepage and evaporation 92 Test II 92 3 4 0ONTBNT8. Page! Chapter VIII. The California "stovepipe" method of well construction for water Bupply 98 Mode of construction 98 Advantages of stovepipe construction 101 I st of construction 101 Well rigs 102 Yield of wells 103 Chapter 1 \. Tests of typical pumping plants 104 Test oi pumping plant of Felix Martinez, near El Paso, Tex 104 Pumping plants of J. A. Smith, near El Paso, Tex 106 Tesl of Roualt's pumping plant, near Las Cruces, N. Mex... 109 Test of lloraeo Ranch Company's well No. J, Berino, N. Mex Ill Summary of results of tests at pumping plants in valley of the Rio Grande, New Mexico and Texas 112 Determination of vacuum 113 Specific capacity 113 Cost and operating expenses 114 Fuel cost 114 Comments on the Rio ( rrande pumping plants 117 Index 121 ILLUSTRATIONS. / Page. Plate L First underflow stations near Garden, Kans. : A, Station 1, near "Point of Rock;" B, Station 2, south of island 10 IL Scale for graphic estimation of transmission constant of a sand 14 III. A, Ram for driving wells; B, Pulling casing with railroad jacks 18 IV. Electrode and perforated brass buckets used in charging wells 18 V. A, Underflow meter; B, Commutator clock 20 VL .4, Small well jetting rig; B, Recording instruments in field box. . . 26 VII 1 ^ Charts made by recording ammeter 26 VIII 1 .' Vertical tank used in laboratory experiments 42 IX'MSTarrows of Mohave River, California 56 XT Driving well at underflow station at narrows of Mohave River 58 XL Driving test wells in the gorge of Mohave River 58 XIL^, 12-inch "stovepipe" starter; B, Two lengths of stovepipe casing. 98 XIII. A, Side view of California well rig; B, Front view 100 XIV. A, Casing perforator; B, Rear view of well rig 102 XV. California well rig and artesian well flowing 5,250,000 gallons per twenty-four hours 102 Fig. 1. Pipe joint made with hydraulic coupling 17 2. Plan of arrangement of test wells 18 3. Diagram illustrating manner of w T orkingof electric underflow meter. 20 4. Ampere curves at station 5, San Gabriel River, California 22 5. Ampere curves at station 1, Long Island, New York 24 6. Ampere curves at station 10, Garden, Kans 27 7. Plan of horizontal tank 31 8. Experiment 4, horizontal tank 33 9. Experiment 6, horizontal tank 33 10. Experiment 8, horizontal tank 34 1 1 . Experiment 9, horizontal tank 34 12. Experiment 10, horizontal tank 35 13. Experiment 11, horizontal tank 36 14. Experiment 1 2, horizontal tank 37 15. Experiment 13, horizontal tank 38 16. Experiment 14, horizontal tank 38 17. Experiment 15, horizontal tank 39 18. Experiment 16, horizontal tank 39 19. Experiment 1 7, horizontal tank 40 20. Experiment 18, horizontal tank 40 21. Vertical tank 43 22. Experiment No. 1 in vertical tank; contours for 12.10 p. m 43 23. Experiment No. 1 in vertical tank; contours for 12.40 p. m 44 24. Experiment No. 1 in vertical tank; contours for 1.40 p. m 44 25. Experiment No. 1 in vertical tank: contours for 2.40 p. in 45 26. Experiment No. 1 in vertical tank; contours for 3.40 p. in 45 5 (') ILLUSTBATI0N8. Page. Fio. 27. Experiment No. 1 in vertical tank: contours for 4.40 p. m 46 28. Experiment No. 2 in vertical tank; contours for 2.55 p. m 46 29. Experiment No. 2 in vertical tank; contours for 7.25 p. m 47 30. Variation in rate of flow of ground water with variation in head 4S 31. Ampere curves at station- Nos. 1 and L'. Rio Hondo, California 52 :V2. Ampere curve at station No. 3, San Gabriel River, California 54 33. Map of gorge of Mohave River. Victorville, Cal 56 34. Cross Bection of gorge of Mohave River, Victorville, Cal 57 ."»•">. Plan of underflow station, gorge of Mohave River 57 .".ti. Ampere curve at station A, Mohave River 58 37. Ampere curve at station E, Mohave River, for depth of 8 feet 58 38. Ampere curve at station E, Mohave River, for depth of 14 feet 59 39. Ampere curve at stati« »n E, Mohave River, for depth of 20 feet 60 40. Ampere curve at station G, Mohave River, for depth of 25 feet 61 41. Ampere curve at station G, Mohave River, for depth of 30 feet 62 42. Ampere curve at station I, Mohave River, for depth of 24 feet 62 43. Ampere curve at station 2 X, Long Island 68 44. Ampere curves at station 5, Long Island 70 45. Ampere curve at station 6, Long Island 71 46. Ampere curve at station 7, Long Island 72 47. Ampere curve at station 8, Long Island 73 48. Ampere curve at station 10, Long Island 74 49. Ampere curve at station 12, Long Island 75 50. Ampere curve at station 14, Long Island 76 51. Ampere curve at station 15, Long Island 76 52. Ampere curve at station 15 X, Long Island 77 53. Ampere curve at station 13, Long Island 78 54. Ampere curve at station 16 X, Long Island 79 55. Ampere curve at station 17, Long Island 80 56. Ampere curve at station 21, Long Island 81 •~i7. Map showing underflow stations, Long Island 82 58. Map showing stations 5 and 6, Long Island 83 59. Map showing stations 2, 13, 16, 17, and Wantagh Pond, Long Island. 84 60. Apparatus for measuring rate of rise of water in wells 90 61 . Curves of rise of water in wells 94 62. Roller perforator for stovepipe wells 100 63. Plan and elevation of California well-rig derrick 102 64. Conditions at pumping plant of Felix Martinez, near El Paso, Tex .. 105 65. Conditions at pumping plant of J. A. Smith, near El Paso, Tex 107 66. Conditions at pumping plant of Theodore Roualt, near Las Cruo X. Mex 110 67. Conditions at Horaco Ranch Company's well Xo. 1, at Berino, X. Mex Ill LETTER OF TRANSMITTAL. Department of the Interior, United States Geological Survey, Hydrographic Branch, Washington, D. C, August 10, 1901},. Sir: I transmit herewith, for publication, a manuscript entitled "Field Measurements of the Rate of Movement of Underground Waters," prepared by Prof. Charles S. Slichter, professor of mathe- matics, University of Wisconsin. The paper presents an amplified exposition of the method of measur- ing the movement of underground waters wHich was devised by Pro- fessor Slichter in 1901 while working for the hydrographic branch of the United States Geological Survey and which has been already briefly described in Water-Supply Paper No. 67, 1902. Descriptions of the apparatus invented by him for the laboratory study of wells con- trolling horizontal and vertical movements of underground waters and the results of these studies are also presented. The laboratory studies are also supplemented by detailed accounts of several investigations made in the field. This paper should be interesting and valuable to engineers and geolo- gists, and the direct application of the results to the study of problems of vital interest to the users of artesian waters should be of great practical value to the general public. A very suggestive and interest- ing description of the California method of sinking " stovepipe" wells deserves the attention of drillers in unconsolidated deposits through- out our country, while the description of the carefully made tests on typical pumping plants in Texas and New Mexico should appeal to engineers and others who are interested in the problem of raising water for irrigation or other purposes. Very respectfully, F. H. Newell, Chief Engineer. Hon. Charles D. Walcott, Director United States Geological Survey. 7 FIELD MEASUREMENTS OF THE RATE OF MOVEMENT OF UNDERGROUND WATERS. By Charles S. Slighter. INTRODUCTION. The following paper describes the method and apparatus used in measuring the velocity of underground waters and gives the results of field work done with the apparatus in various parts of the United States, under authority of the hydrographic branch of the United States Geological Survey. The method used in making the measure- ments was devised by the writer after preliminary tests along the Arkansas River in western Kansas during the summer of 1901. This preliminary work indicated that it was practicable to measure the rate of flow of ground waters by the use of very simple apparatus. Several determinations of the rate of movement of the underflow of the Arkansas were made during that summer. These measurements, it is believed, constituted the first direct determinations of the rate of flow of ground water that had been made in this country. This preliminary work was done in the neighborhood of Dodge, Kans., and at one or two points near Garden, Kans. The photographs reproduced in PL I show the locations of the first successful stations, which were estab- lished near Garden, Kans. A brief description of the electrical method of measuring the velocities of underground waters resulting from this preliminary investigation was printed in the Engineering- News of February 20, 1902, and in paper No. 67 of the Water-Supply and Irrigation series of the United States Geological Survey. Since then, as the result of work carried on in the field and in the laboratory, the apparatus has been gradually improved, and its present form is described in these pages. The paper will also include some determinations of the manner and rate of flow of water into tubular wells, and descriptions of methods and simple apparatus designed to accurately estimate the capacity of such wells. 9 CHAPTER I. Ill i : CAPACITY OF A SAM) TO TRANSMIT WATER. FACTORS INFLUENCING FLOW. The general laws governing the flow of water through a mass of Band or erravel have been described by the writer in another paper." and will not be repeated in this place. It i- sufficient to state that experiments show that the flow of water in a giveo direction through a column of -and is proportional to the difference in pressure at the end- of the column, and inversely proportional to the length of the column, and i- also dependent upon another factor, called the trans- mission constant of the sand. TRANSMISSION CONSTANT. The resistance offered by sand or gravel to the flow of water which is percolating through it is very great. The water is obliged to pass through very small pores, usually capillary in character; indeed, they are much smaller in cross section than the soil particles between which they pass. If the particles of sand or gravel which make up the water- bearing medium are well rounded in form the pores are somewhat triangular in cross section and the diameter of the individual pores i- only one-fourth to one-seventh the diameter of the soil particles themselves. Thus if the individual grains of sand average 1 milli- meter in diameter the pores through which the water mast pass will average only one-fourth to one-seventh of a millimeter in diame- ter. If to a mass of nearly uniform sand particles larger particles be added the effect on the resistance to the flow of water will be one of two kind-, depending principally upon the ratio which the size of the particles added bears to the average size of grains in the original sand. If the particles added are only slightly larger than the original sand grains, the effect i- to increase the capacity of the sand to transmit water, and the more particles of this kind that are added the greater will be the increase in the capacity of the -and to transmit water. If, however, large particle- are added, the effect is the reverse. If par- ticle- seven to ten time- the diameter of the original sand grains be added, each of the new particles tend- to block the course of the , be motions .if underground waters: Water-Sap. and Irr. Paper No. <;:. l". 3. Geol. Survey. IW2. 1U U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. LOCATION OF FIRST UNDERFLOW STATIONS. AUGUST, 1901, NEAR GARDEN, KANS. A, Station 1 , near Point of Rock , ]>, Station 2, south of island. slichter.] TKANSMISSION CONSTANT. 11 water. Thus, for example, a large bowlder placed in a mass of fine sand will tend to block the passage of the water. As more and more of the large particles are added to a mass of uniform sand, the rate of flow of water through it will be decreased until the amount of the large particles equals about 30 per cent of the total mass. From this time on the adding of the large particles will increase the capacity of the whole to transmit water until, if a very large quantity of the large particles be added, so that the original mass of fine particles becomes relatively negligible, the capacity to transmit will approach that of the mass of the large particles alone. These facts have an important bear- ing upon the capacity of gravels to furnish water to wells or to trans- mit water in the underflow of a river. The presence of large particles is not necessarily to be interpreted as indicating a high transmission capacity of the material, for this is indicated only when the large par- ticles constitute a large fractional per cent of the total mass, as would be the case where the large particles, equal 40 or 50 per cent of the whole. The capacity of any sand or gravel to transmit water can be expressed by means of a single number which is called the transmission constant of the soil. This constant is defined to be the amount of water transmitted in unit time through a cylinder of the soil of unit length and unit cross section under unit difference in head at the ends. For example, if the foot and minute be the units of length and time, and if a column of sand 1 square foot in cross section and 1 foot in length will trans- mit 1 cubic foot of water a minute under a difference in head of 1 foot of water, the transmission constant is 1. The transmission constant of a soil varies very greatly with the size of the individual grains con- stituting the sand or gravel, and also depends in a marked degree upon the porosity or amount of open space in the soil. Table I, here- with, gives the transmission constants for a variety of sizes of soil grain and for a series of porosities varying from 30 to 40 per cent. This table is computed for a temperature of the water of 60° F. An auxiliary table, Table II, is one from which the transmission constants corresponding to other temperatures can readily be found. Transmission constant Jc is the quantity of water, measured in cubic feet, that is transmitted in one minute through a cylinder of the soil 1 foot in length and 1 square foot in cross section, under difference in head at the ends of 1 foot of water. The tabulated numbers express the transmission constant in cubic feet per minute. 12 KATK OF MOVEMENT OF DNDERGBOUHD WATEBS. • "0- Tabu I. — T 'i° m Table computed for temperature :np>eratures can be found by the Tab. Diame- - - Kind of soil. grain in mm. 30 per 32 per 34 per 0.000050 em. ent. 40 per cent. I 0.01 0. 00 0.000060 - - 0.000085 j .02 .03 .000131 • . 0002 .001 - . 00 . 00< - 4460 .0002 .0005 I . oo _ a .00 .000339 .0007 Silt. .04 .000527 .00 .00 " .00 58 .001145 . 001 5 .000822 .001012 .240 . 001495 . 001 7 .00212 .06 L182 1610 . 00 .001 - 7S4 .002 _ 1 .002 258 .00: .003050 .004155 fine - ad. - .00l .002? .00 " .oo. _ 4585 . 005425 .09 2 28 .004018 .0058 .006860 .10 282 .004960 s - " t MS .12 . 00- "■- .007130 .008 - .OK - 122 .14 . 0C" .007 - .00972 .01172 .01404 ".OH - .15 . 007390 .009 - .01115 .01: :■- - . 01611 . 01910 -and. .16 .008 .OK .on a .01531 a [ . _ " a .01064 .01311 i .01 _ - 2745 _ .01315 .OH - .0198 2390 28 I 390 _" 2050 -" .031 740 448 ": .30 2960 - 44 S380 .0645 ' 35 i -' .(4960 .06 '330 .08790 •39 Med ium Bind. .40 .Ob.' .0648 :v40 .09571 . 1145 .1355 .45 . 061 -200 .1( I .1211 .1450 .1718 1 822 . 1012 . 1240 . 1495 .1780 .21: • .09940 . 1225 .1500 .1S10 .2. I _" I 60 .1182 . 14£ - .1781 .21" 2580 .3050 I .1390 .1710 _ " -.30 .3030 "- " .1610 .198 .24 .2930 510 .415S ■ r 8 e n ~- __"- ,278£ .3365 . 40 .477 sand. .80 .21 " 2590 71 -I" .4585 . 5425 .85 _ "" _ -' 18 .4: ■_- . 5175 .6125 .90 _ -60 28 .4018 1845 -00 .6860 I - -65 " .447 5400 ^60 7650 - 1.00 .3282 . 4' 6 .4960 " 980 .717 -4S0 ■ 2.00 1. 315 2.960 1.62 - 1.983 4.460 _ :90 - a 390 30 Fine grav- el. 4.00 _'70 7.940 _ " 11.45 " 5.00 8.22 10. 12 12.40 14.96 17.90 21 . 20 SI.ICHTER.] TRANSMISSION CONSTANT. 13 Table II. — Variation, with the temperature, of ike flow of waier of various temperatures through a sand, 60° F. being taken as the standard temperature. Temperature. Relative flow. " Temperature. Relative flow, a °F. Prr cent. °F. Per cent. 32 0.64 70 1.15 35 .67 75 1.23 40 .73 BO 1.30 45 .80 B5 1.39 50 .86 90 1.47 55 .93 95 1. 55 60 1.00 100 1.64 65 1.08 a "Relative flow'" means Hovr at given temperature compared with flow at 60° F. It is expressed as a percentage. It should be borne well in mind that the rate of transmission varies very greatly with the temperature of the water. For example, a change from 50 c to 60 : increases the capacity to transmit water under identical conditions by about 16 per cent, while a change from the freezing temperature to a temperature of 75- will nearly double the power of a soil to transmit water. This difference, of course, is not due to any change in the soil itself, but is due solely to the increased ease with which water Hows at high temperatures compared to the ease with which it flows at low temperatures. The transmission constant of a ^and can also be obtained by use of the diagram oriven iu PI. II. Graduated vertical lines will be found in this diagram corresponding to the diameter of the soil grains on m. The place where the ruler crosses the line q (0.0036), will give the discharge in cubic feet per minute. The diagram gives results based upon the assumed temperature of the water of 60 F." If any three of the four magnitudes /. », m are known, the remaining one can be found in a manner similar to the above. The removal of larger particles from a mixed sand may not only increase the transmission constant in the manner described above, but such removal may also increase the transmission capacity by permit- ting the remaining sand to pack in a more open manner, as would be shown by an increased porosity. Tables II and III give results which show that the removal of the larger grains from a sand does not nec- essarily decrease the transmission constant, but may even increase it. The results given in Table III were obtained by successively removing the larger particles from a mass of sand by means of standard sieves, and then determining the porosity, effective size, 6 and transmission constant for the finer material passing through the successive sieves. The gravel represented by Table III consisted of a mixture of all sizes of grains, from very fine grains to bowlders 2 feet in diameter. All pieces larger than 1 inch in diameter were discarded before the results shown in Table III were obtained. .It is interesting to note that the 93.4 per cent of the total sand passing through sieve 2 (2 meshes to the inch) did not have as large an effective size as the 74.2 per cent which passed through sieve 20 (20 meshes to the inch). Table III is derived from a beach sand. The 54.3 per cent of this sand which passed through sieve 10 has a smaller transmission constant than the 36. S per cent which passed through sieve 14. The following table shows the effect of removing, by means of standard sieves, the coarser portions of a natural Arizona gravel. The data in columns 2, 3, 4, 5, and 6 apply to that portion of original sample that passed the sieve named in column 1. a The diagram was computed and drawn for the writer by J. D. Suter, of the University of Wisconsin. f'The effective size of a sample of sand is such a number that if all grains were of that diameter the sand would have the same transmission capacity that it actually has. It is, therefore, the true mean or average size of sand grain in that sample. U. 8. GEOLO DIAMETER ( Miilirrv: "5 -4 — .* — .3 -.1 -.0! -.0 -.0 -.01 SCALE OR NOMOGRAPH FOR ESTIMATING GRAPHICALLY THE TRANSMISSION CONSTANT OF A SAND OR GRAVEL. SLIGHTER. TRANSMISSION CONSTANT. Table III. — Effect of removing coarser portions of field gravel. 15 1. 2. 3. 4. 5. 6. No. of sieve. Quantity of gravel passing. Differences of numbers in column 2. Porosity of portion passing. Effective size. Transmis- sion con- " stantat80°F. Meshes to inch. Per cent of to- tal weight. Per cent of to- tal weight. Per cent. Mm. Cubic feet per minute. 2 93.4 6.6 38.1 0.325 0. 102 8 10 No data. 83.7 40.0 38.6 .320 .282 .120 .080 9.7 12 82.6 1.1 40.3 .304 .108 14 80.3 2.3 39.5 .282 .086 16 78.0 2.3 39.7 .282 .088 18 76.4 1.6 39.6 .277 .085 20 74.2 2.2 40.6 .317 .119 30 55.9 8.3 41.0 .250 .076 40 32.2 23.7 40.6 .187 .042 The following table shows the effect of removing by means of stand- ard sieves the coarser portions of a natural beach gravel. The data in columns 2, 3, 4, 5, and 6 apply to that portion of original sample that passed sieve named in column 1. Table IV. — Effect of removing coarser portions of a beach gravel. No. of sieve. Meshes to inch. Total sample. 10 12 14 16 18 20- Quantity of gravel passing. Per cent of to- tal weight. 100.0 54.3 47.6 36.8 31.4 No data. 25.6 Differences of numbers in column 2. Percent of to- tal weight. 45.7 6.7 10.8 5.4 5.8 Porosity of portion passing. Per cent. 37.8 40.0 41.7 41.7 42.6 43.5 43.5 Effective size. Mm. 0.810 .634 .640 .603 .539 .520 .494 Transmis- sion con- stant at 72°F. Cubic feet per minute. 0.529 .390 .457 .406 .348 .348 .314 CHAPTER II. r\DEKFLOW METER USED IN MEASURING VELOCITY AND DIRECTION OF MOVEMENT OF UNDERGROUND WATERS. TYPES OF APPARATUS. The apparatus used is of two types: (1) Direct reading, or hand apparatus, requiring- the personal presence of the operator every hour for reading of instruments, and (2) recording apparatus, which requires attention but once in a day. Both forms are described in thi>> chapter. The arrangement of the test wells and manner of wiring the wells is essentially the same for both. TEST WELLS. The test wells suitable for use with the underflow meter in deter-, mining the velocity of ground waters may be common 1^-inch or 2-inch drive wells if the soil is easily penetrated and if the depths to be reached do not exceed 50 to 75 feet. For greater depths and harder soil wells of heavy construction should be used. The 1^-inch drive wells are much preferable to the li-inch wells because of the fact that l^-inch pipe is lap welded, while the li-inch is butt welded, and less capable of standing severe pounding. The drive point used with the well may be 1^-inch standard brass jacket well points, 42 to 48 inches long, with No. 60 brass gauze strainer. The well points should be threaded with 1£ inches of standard thread, somewhat more than is usually found on the trade goods. The pipe should be full weight strictly wrought-iron standard pipe, cut in lengths of 6 or 7 feet, and threaded li inches at each end. The couplings should be wrought-iron hydraulic recessed couplings, and the thread on the pipe should be cut in such a way that when properly screwed up the ends of the pipe will abut. The recessed couplings protect the pipe at its weakest point, while an ordinary coupling will leave exposed a thread or two of the pipe so that severe driving is liable to swell and ulti- mately rupture the pipe just above the coupling. Fig. 1 represents a hydraulic coupling, showing a properly made joint. The driving head should be made of rolled steel shafting and should be about 4 inches long, carrying 1^-inches standard thread and an air hole to permit the free escape of air from the well while the driving 16 SLICHTER.] TEST WELLS. 17 is in progress. A driving ram for putting clown the drive wells should be about 5+ feet long by 5i inches in diameter, made of heavy oak or other tough wood, with iron bands shrunk on the ends, and bearing two handles of hard wood at each end in order to facilitate the handling of the ram by two men. It is convenient to have these handles placed one about 1 foot from one end, and the other about 2 feet from the other end. By reversing the ram the handles are brought in a more convenient position for driving as the well goes down. PI. Ill, ^i illustrates the method of putting down drive points. If the test wells are to be sunk to a depth exceeding that to which drive points can be readily driven, open-end 2-inch pipe should be used. These wells should be made with full weight strictly wrought-iron 2-inch pipe with long threads and recessed hydraulic couplings, as described above. The pipe can either be put down without a screen, in which case a 1^-inch well point with turned coupling may be inserted through a drive shoe at the bottom of the casing after the pipe is Fig. 1.— Pipe joint made with hydraulic coupling. This joint will stand hard driving. driven into place, or an open-end brass jacket well point, 48 inches long, may be put down with the pipe. The pipe should be driven into place with a cast-iron ram varying in weight from 150 to 250 pounds, simultaneously hydraulicking a passage for pipe with water jet in three-fourths-inch wash pipe. There are many hand rigs on the market suitable for this work, or a rig can be readily constructed by any good mechanic. Such a rig is shown in PI. VI, A. A suitable pump for the hydraulic jet is a double-acting horizontal force pump with a 1- by li-inch cylinder. If the material in which the well is to be drilled is not too hard nor too full of bowlders, the writer recommends that an open-end well point be put down with the casing. This is apt to cause some difficulty in the proper working of the lvydraulic jet. by the escape of water through the screen of the well point. This difficulty can be obviated and a more powerful wash secured by admit- ting a considerable quantity of air along with the water at the suction end of the force pump. The exact amount of air to be admitted can irr 140—05 2 18 RATE OF MOVEMENT OF r\l)KK(iHOUND WATERS. [no. 140. be readily determined with a little experience. The effect of the air entering the well under high pressure is to form a powerful airlift which will throw the water and gravel out of the top of the well easing with great force. It has been the writer's experience that the best hydraulic samples are obtained with the combination hydraulic pneu- matic jet. If tin 4 hydraulic jet alone is used the coarser particles have a tendency to remain at the bottom of the well. After a test is completed the well casino- can readily he pulled by a No. 2 cast-iron pipe puller and two 5-ton railroad jacks. Sets of dies for the pipe puller to fit both H- and 2-inch pipe can be obtained at small cost. PI. Ill, B, shows the operation of the pipe puller and railroad jacks. Fig. 2.— Plan of arrangement of test wells used in determining the velocity and direction of motion of ground water: A, B, C. D are the test wells. The direction A-C is the direction of probable motion of the ground water. The dimensions given in plan a are suitable for depths up to about 25 or 30 feet, those in plan b for depths up to about 75 feet. For greater depths the distances A-B, A-C, A-D should be increased to 9 or 10 feet, and the distances B-C and C-D to 4 feet. The well A is the "salt well" or well into which the electrolyte is placed. The test wells are driven in groups, as shown in fig. 2, each group of wells constituting a single station for the measurement of the direc- tion and rate of flow of the ground water. In case the wells are not driven deeper than 25 feet, the " upstream" or " salt well" A is located, and three other wells, B, C, and D, are driven at a distance of 4 feet from A, the distance between B and C and C and D being about 2 feet. The well C is located so that the line from A to C will coincide with the probable direction of the expected ground-water movement. This direction should coincide, of course, with the local slope of the water plane, and if this is not accurately known, it should be deter- mined by means of leveling with a level. For deeper work the wells should be located farther apart, as shown in the right portion of fig. 2. For depths exceeding 75 feet, a radius of 8 or 9 feet and chords of 4 feet should be used, the general requirement being that the wells should be as close together as possible, so as to cut down to U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. Ill A. RAM USED FOR DRIVING SMALL WELLS BY HAND. B. PULLING WELL CASING WITH RAILROAD JACK AND NO. 2 PIPE PULLER. U. S. GEOLOGICAL SURVEV WATER-SUPPLY PAPER NO. 140 PL. IV ELECTRODE AND PERFORATED BRASS BUCKETS USED IN CHARGING WELLS. slighter.] DIRECT-READING INSTRUMENTS. 19 a minimum the time required for a single measurement, but not so close that important errors are Liable to be introduced by the inability to drive the wells perfectly straight and plumb. On this account the deeper the wells the farther apart they should be placed. The angles B A C and C A D should not exceed 30 . DIRECT-READI NO I NSTK I M ENT8. Electrical connection is made with the casing of each test well by means of a drilled coupling carrying a binding post. Each of the downstream wells (B, C, D) contains within the well point or screen section an electrode consisting of a nickeled brass rod three-eight hs inch in diameter by 4 feet long, insulated from the casing by wooden spools. The end of rod receives a No. 11 rubber-covered wire, to which good contact is made by a chuck clutch. An electrode is shown in PI. IV. This electrode communicates with the surface by means of a rubber-covered copper wire. PI. IV also shows two buckets of per- forated brass used in charging wells with granulated sal ammoniac: each is If by 30 inches. Fig. 3 illustrates the arrangement of electric circuits between the upstream well and one of the downstream wells. Each of the down- stream wells is connected to the upstream well in the manner shown in this figure. A view of the direct-reading underflow meter is shown in PI. V, A. Six standard dry cells are contained in the bottom of the box, their poles being connected to the six switches shown at the rear of the case. By means of these switches any number of tin 1 six cells may be thrown into the circuit in series. One side of the circuit termi- nates in eight press keys, shown at the left end of the box. The other side of the circuit passes through an ammeter, shown in the center of the box, to two three-way switches at right end of the box. Four of the binding posts at the left end of the box are connected. respectively, to the casing of well A, and to the three electrodes of wells B, C, and D, in the order named. The binding posts at the right end of the box are connected to the casings of wells B, ( J, and I). There are enough binding posts to connect two different groups of wells to the same instrument. When the three-way switch occupies Hie position shown in the photograph, pressing the first key at left end of the box will cause tin 1 ammeter to show the amount of current massing between casing of well A and casing of well B. When the next key is pressed the ammeter will indicate 1 the current between the casino- of well Band the electrode contained within it. In one instance the current is conducted between the two well casings by means of the ground water in the soil; in the second case the electric circuit i- completed by means of the water within well B. By putting the three-way switch in second position and pressing the first and third 20 BATE OF MOVEMENT OF UNDEBGROUND WATERS. [no. ho. keys iii turn, similar readings can be had for current between casings A ami C and between casing C and its internal electrode. Similarly with switch in third position, readings are taken by pressing first and fourth keys. The results may be entered in a notebook, as shown in Table V. B A C Fig. 3.— Diagram illustrating electrical method of determining the velocity of flow of ground water. The ground water is supposed to be moving in the direction of the arrow. The upstream well is charged with an electrolyte. The gradual motion of the ground water toward the lower well and its final arrival at that well are registered by the ammeter A B is the battery and C a commutator clock which is used when A is a recording ammeter. The principles involved in the working of the apparatus are very simple. The upstream well A is charged with a strong electrolyte, such as sal ammoniac, which passes down stream with the moving ground water, rendering the ground water a good electrolytic con- ductor of electricity. If the ground water moves in the direction of one of the lower wells, B, C, D, etc., the electric current between U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. V UNDERFLOW METER, SHOWING CONNECTIONS WHEN USED AS DIRECT-READING APPARATUS. When used with recording ammeter, only two connections are made, one to each side of battery circuit; but the ammeier is left in circuit with the recording in- strument to indicate whether tne latter is working properly. B COMMUTATOR CLOCK, FOR USE WITH RECORDING AMMETER. The clock makes electrical contact at any five-minute interval. SLIGHTER.] DIRECT-READING INSTRTM KXTS. 21 A and B, A and C, or A and D will gradually rise, mounting rapidly when the electrolyte begins to touch one of the lower wells. When the electrolyte finally roaches and enters inside of one of the wells B, C, D, it forms a short circuit between the casing of the well and the internal electrode, causing an abrupt rise in the electric current. The result can be easily understood b} T consulting Table V and tig. 4. in which the current is depicted graphically. Table Y . — Field 'record of electric cur rent during underflow measurements at station 5, Rio Hondo and San Gabriel River, California, August 6 and 6, 1902. [Readings in amperes and decimals of an ampere.] Time. 8 a. m 8.15 a.m.. 8.30a. in.. 9 a. m 10 a. in 11.40a. m. 1 p. in 2 p. m 3 p. m 4 p. m 5 p. m 6 p. m 7 p. m 8 p. m 9 p. m 10.30 p. m 12 p.m... 1 a. m 6 . . . 2.30a. in.. 4.15 a. in. . 5.30a. in.. 7.45 a.m.. 8.15 a. in. . 9 a. m Well B. Well C. Well I). Casing. 0.140 Salt. .160 .168 . 180 . 102 .202 .205 .208 .210 .218 .225 .230 .240 .250 .275 .350 .420 . 510 .560 .550 .520 Elec- trode. Casing. 0. 360 .360 . 345 .340 .345 .342 .350 . 330 .330 .330 .330 .330 .340 .600 .850 1.550 2.000 2.200 2.250 2.250 2.200 0.142 Salt. .163 .170 .182 .195 .202 .204 .205 .205 .210 .210 .218 .222 .222 .225 .230 .240 . 240 .240 .230 .230 Elec- trode. Casing. 0.332 . 330 .325 .320 .340 .320 . 320 .310 .310 .310 .310 . 320 .315 .310 .310 .310 .310 .310 .310 0.150 Salt. .170 .180 .192 .202 .210 .210 .210 .210 .212 .218 .220 . 22:5 . 225 .225 .230 .230 .230 .230 . 230 . 225 Elec- trode. 0. 390 Remarks.a 1 NaCl .390 .380 . 370 .370 .360 .370 .360 .360 .360 .350 . 352 .360 . 340 .340 .340 .340 . 330 .330 1 XaCl 2X11,(1 1 XII ,('1 1 NH 4 C1 1 XII, CI 1 XaCl 1 X1I.C1 1 XaCl 1 XII, CI «The electrolyte was lowered into well A by means of a perforated bra>-s backet, 1| by 30 Inches in size. The formula "2 NH 4 C1" means that two of these buckets, full of ammonium chloride, were introduced into well A at the time indicated. Each of these buckets held -J pounds of the salt b August 6. The time that elapses from the charging of tin 4 well A to the arrival of the electrolyte at the lower well gives the time necessarv for the ground water to cover the distance between these two wells. Hence, QQ BATE OF KOVEMENT OF DNDEBGBOUBTJ) WATERS. [no. 140. if the distance between the wells be divided by this lapse of time, the result will be the velocity of the ground water. The electrolyte does not appear at one of the downstream wells with very great abruptness; its appearance there is somewhat gradual, as shown in the curves in X »A / >' * ,' 1 / 1 ,' t / / S / 4 £'' t D / 7 % / vV \ C i ..> r ,.n • C 2ft CO ► B t 7 1 0.90 AMPER E CUF VE WELL 'B • / / 5- '**% / 0.80 * a £/ !•" CTNDEROROUND WATEB8. [no.14D. PL V, B shows a commutator clock made for this purpose by the instrument maker of the College of Engineering, University of Wis- consin. The clock movement is a standard movement of fair grade, costing less than $5. It can readily be taken from the case for cleaning or oiling and quickly replaced. A mood movement with powerful springs is best for this purpose. It will be seen from the method of wiring the wells that the record will show the sum of the current between well A and well B added to the current between the casing of well B and its electrode. The removal of the connection to well A would permit the record to show tin 1 current between the casing of a downstream well and its electrode, but the connection to the upstream well involves no additional trouble and occasionally its indications are of much service, especially if the velocities are low. One of the instruments above mentioned can be placed in a common box. 10 by 22 by 36 inches, covered witli tar paper and locked up. PI. VI, B, is a view of the instruments thus arranged. The shelf con- tains the recording ammeter (shown at left of cut) and the commutator clock (shown at right of cut). The contacts on the commutator clock are arranged about five min- utes apart, so that the record made for the wells will appear on the chart as a group of lines, one for each downstream well, of length corresponding to the strength of the current. The increasing current corresponding to one of the wells will finally be indicated by the lengthening of the record lines for that well. This can be seen In- consulting the records shown in PI. VII. The record charts are printed in light-green ink and red ink is used in the recording pen. so that record lines can be distinguished when superimposed upon the lines of the chart. A special chart has been designed for this work and is furnished by the Bristol Company as chart 458. PI. VII shows two charts made by recording ammeter. In the upper the electrical current for wells B, C, and D, at station 14, Long Island, is recorded, in the order named, at 2.10, 2.15, and 2.20 p. m.. and hourly thereafter, the current remaining nearly constant at .22 to .24 ampere until 10.15 p. m., when the current for well C rises as indicated in the chart. In the lower chart the electrical current for wells B, C, and D is recorded, in the order named, at 6.30, 6.35, and 6.40 p. m., and hourly thereafter. The current for wells B and D remains constant at .25 ampere, but the current for well C rises as shown. The recording instruments in use have given perfect satisfaction and the method is a great improvement in accuracy and convenience over the direct-reading method. The highest as well as the lowest ground-water velocities yet found have been successfully measured U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. VI A. SIMPLE FORM OF SMALL WELL-JETTING RIG. The men driving with 150-pound weight stand on platform attached to drivepipe. B. RECORDING AMMETER, COMMUTATOR CLOCK, AND BATTERY BOX IN USE IN THE FIELD, ARRANGED IN A ROUGH BOX, 16 BY 22 BY 36 INCHES IN SIZE. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. VII CHARTS MADE BY RECORDING AMMETER. SLICHTER.] RECORDING INSTRUMENTS. 27 by the recording- instruments. By using one or two additional dry cells* the instrument may be made quite as sensitive as the direct- reading type. In using the recording instruments but a single dose of salt need he placed in the upstream well. If the wells are deep, it is important to use enough salt solution to make sure that the salt reaches as deep as the screen of the well point immediately after the solution is poured into the well. A gallon of solution will fill about 6 feet of full-weight wrought-iron pipe, so that 10 gallons of solution should be used if well is 60 feet deep. If the proper amount of solu- tion be not used, it will take an appreciable time for the solution to reach the bottom of the well by convection currents, and the results 0.30 y 0.20 5 12 M. 1 2 JULY 3, 1904 Fig. 6. — Ampere curves at station 10. Garden, Kans. The heavy curves represent the electric cur- rent between the casings of wells B and C and the electrodes inside of them. The dotted curves represent the electric current between wells B and C and between C and D. The electric current for the electrode of wellB rose earlier than for well C, and the electric current between wells B and C rose earlier and more abruptly than for wells C and D, indicating that the principal stream of electrolyte passed between wells B and C and nearer to well B than to well C will be vitiated to that extent. As before stated, it is preferable to introduce into the well granulated sal ammoniac contained in a suit- able bucket in case the depth of the well renders the use of a solution uncertain. In order to properly interpret results obtained in the field with the apparatus, it becomes necessary to investigate the behavior of the dis- solved electrolyte as it moves onward with the ground water after leaving the salted well. This matter could be investigated only in the laboratory, and it was sought to reproduce as nearly as possible the conditions found in the field. 28 KATE OF MOVEMENT OF I'XDERGROUND WATERS. [no. 140. Before Baiting the upstream well of any set of best wells the electric circuit should be closed between each adjacent pair of downstream wells, and tin 4 current should be measured with the direct-reading ammeter and recorded in the notebook. An occasional reading of these same circuits will prevent the electrolyte from passing between two of the downstream wells without the knowledge of the observer. This i- clearly shown by the results obtained with the set of wells represented in tig. (>. At the location of this station the direction of the How was at first not correctly estimated on account of its nearness to a river whose height was fluctuating. For that reason the down- stream wells were redriven at distances of but 20 inches from one another. The diagram gives the ampere curves for wells B and C, both of which were reached by the electrolyte, and also the curves of current between wells B and C and wells C and D. The actual direc- tion of flow can be seen from these curves to lie between B and C, and probably nearer B than C, since the curve for B rises somewhat earlier and the percentage increase in current is greater. The same fact is shown by the curves representing the current between B and C and between C and D. The main stream of electrolyte must have passed between B and C, as is show T n by the more abrupt and earlier use in the current between B and C as compared to that between C and D. CHAPTER III. LABOEATORY EXPERIMENTS OX THE FLOW OF WATER THROUGH SANDS AXD GRAVELS. OBJECTS OF THE EXPERIMENTS. During the winters of 1902-3 and 1903-4 experiments were carried on in the laboratory upon the flow of water through sands and gravels contained in tanks. The objects of these experiments were: (1) To verify the law of flow of water through sands and gravels under gra- dients similar to those found in the field; (2) to ascertain the law of distribution in a horizontal plane of the electrolyte used in the elec- trical method of determining the rate of flow of underground water; (3) to determine the influence of varying velocities upon this distribu- tion; (1) to determine, if possible, by means of apparatus approxi- mating actual field conditions, the relation between the distribution of the electroh'te and the current curve obtained by the electrical method of measuring ground-water velocities, thereb} T checking the accuracy of the method and furnishing data indicating more definitely the point on the current curve which should be selected in order to find the velocity of flow. For the laboratory work of 1902-3 the writer had the assistance of Mr. Henry C. Wolff, and the work of 1903-1 was clone by Mr. Ray Owen and H. L. McDonald. EXPERIMENTS IN THE HORIZONTAL TANK!. The apparatus used in the first experiments consisted of a horizontal wooden tank of inside dimensions 4 feet 6 inches long, 1 feet wide, and 8 inches deep. A chamber of perforated sheet brass 3 inches wide was inserted in each end of the tank, so that the dimensions of the compartment left for the gravel was 1 by 1 feet in horizontal extent. The area 1 by 1 feet was divided into squares 6 inches on a side, at the corner of each of which a small well of slotted sheet brass, one-half inch in diameter, was fixed in position. A larger well, 2 inches in diameter, of the same material was placed in position as shown in the plan, fig. 7. For the first experiments the tank was filled with about 7 inches of gravel, which we have designated as Picnic Point gravel. The effective size of this gravel, as determined by King's aspirators, was 0.93 mm. Mechanical analysis of the Picnic Point gravels will be found in Table VI. 29 30 BATE OF MOVEMENT OF UNDERGROUND WATERS. [no.140. Table VI. — Mechanical analysis by standard sieves of several gravels referred to in the text. Percent of total weight of Band passing No. of Bcreen- meshes to inch. Size of sep- aration of screen in millimeters. 0. 18 slew. Victorvil le gravel. Picnic Point gravel. Madison glacial gravel. 100 00.6 00.8 00.1 80 . 23 1.0 1.4 0.2 60 .32 L6 6.3 1.3 40 .411 4.2 38.6 4.7 30 .70 14.3 83. 9 13.1 20 .93 25. 6 98.1 27. S 16 1.30 31.4 99. 6 37. 2 14 1.40 36. 8 100 46.3 L2 1.70 47.6 100 . ho. compartment in the tank by means of a '-inch pipe, the height of the overflow being adjustable. Table VIII. — Data obtained during experiments on flow of water in horizontal tank. No. Date i periment, 190-2-3. Hydraulic gradient. Av« •: depth of water. Velocity of ground water. Velocity per unit head. Tempera- ture of water. Salt used in experiment. Ft. p< r milt . Inclu s. Ft. in r, Jus. Ft. in U hrs. °C. 1.... Dee. 8 25. 30 5. 84 15. 4 0.61 14.4 None used. 2.... Dec. 9 20. 50 5.83 14.2 .69 14.0 Do. 3.... Dec. 11 18. 90 5. 82 13.7 .63 14.0 Do. 4 Dec. 20 18.70 5.88 13.2 .70 20.5 Dry .\II 4 C1. 5.... Jan. 3 17.60 5.12 12.9 .73 20.0 Do. 6.... Jan. < 21.45 5.10 12.1 .56 18.8 Dry NaOl. 7.... Jan. 14 22. 00 5.12 13.8 .63 17.8 Con. XH 4 OH. 8-... Jan. 17 18. 54 5.09 9. 3 .50 18.8 Do. 9.... Jan. 30 18.92 5.18 11.6 .61 17.8 Dry A H 4 C1. ^ NaOH. 10.... Feb. 9 19.03 5.17 11.25 .59 20.0 Dry T \ NH 4 C1. T % NaOH. 11.... Feb. 18 21. 45 5.15 11.68 .54 18.8 Sol. NH 4 C1. 12.... Feb. 23 42. 35 5.28 21.70 .51 19.1 Do. 13.... Mar. 3 64.90 5. 21 36.00 .55 20.1 Do. 14.... Mar. 9 64.24 5.21 35.5 .55 20.3 Sol. W NH 4 C1. t V NaOH. 15.... Mar. 16 66. 55 5.25 36.4 . 55 21.7 Dry T 'V ^H 4 C1. T V XaOH. 16.... May 4 103.2 6.68 11.47 .11 22.0 XH 4 C1. 17.... May 14 105.6 6.70 11.60 .11 18.2 &NH 4 C1. yV NaOH. 18 May 23 107.8 6.69 11.90 .11 NaOH. Note.— In experiments 1-15 ''Picnic Point graver' was used, and in experiments 16, 17, 18 Madison glacial sand was used. The water used in the experiment was obtained from Lake Mendota. Madison, Wis., and before use was freed from suspended material by passing through a filter of charcoal and sand. Before passing through the gravel in the tank, one part in 500 of 40 per cent solution of forma- lin was added to the water so as to inhibit the growth of organisms. Previous experimenters on the now of water through sands and gravels experienced much difficulty on account of the progressive reduction in now of water through the sand when an experiment extended over a considerable length of time. No means had been found for avoiding this difficulty: even the use of distilled water was not entirely effective. It was difficult to explain this phenomenon except on the basis of the growth of organisms in the pores of the SLIGHTER.] EXPERIMENTS IN THE HORIZONTAL TANK. 33 DEC. 20, 1902 ° O O o nno i o n o ol WELL SALTED AT 9.30 A.M. NH 4 CI. Fig. 8.— Diagram showing the manner in which the electrolyte spread in passing downstream with the ground water, in experiment 4, in the horizontal tank. The dot at W shows the location of the salted well, and samples were taken from the sand from the small test wells represented by dots in the diagram. The areas of the circles are proportional to the strength of the electrolyte found at their centers. The area covered by the charged water at the time specified is shown by a roughly sketched outline. The velocity of the ground water in the direction of the arrows was 13.2 feet for twenty-four hours. Electrolyte used was sal ammoniac. 3: 00 P.M 4.0MM. JAN.7,1903 5.00,P.M^ ^6 6!00,P.M. Fig. 9.— Diagram showing the results of experiment 6. Representation of wells and other features as in fig. 8. The velocity of the ground water was 12.1 feet for twenty-four hours. The electrolyte used was common salt. JBR 140— Q§— § 34 RATE OF MOVEMENT OF UNDERGROUND WATERS. Lno.MOl Pig. 10.— Diagram showing the results of experiment 8. Representation of wells and other features as in fig. 8. The velocity of the ground water was 9.3 feet for twenty-four hours. The electrolyte used was concentrated ammonia water. 12.00 M. JAN. 30, 1903 V.00 P.M. Fig. 11.— Diagram showing the results of experiment 9. Representation of wells and other features as in fig. 8. The velocity of the ground water was 11.6 feet for twenty-four h<>ur> The electro- lyte used was one-Tenth caustic soda and nine-tenths sal ammoniac. SI.ICHTEK. EXPERIMENTS IN THE HORIZONTAL TANK. 35 sand used in the experiments. For this reason the formalin was added to the water in the hope that if this were the correct explanation the difficulty would vanish. Several experiments were made for the especial purpose of determining the effect of the formalin in inhibit- ing the organic growth in the filter. Table VII gives the result of three such experiments. The duration of each experiment was divided into two nearly equal periods, and the average head of water as shown by the gages and the average flow of water, as determined by weigh- ing both the water admitted to the tank and the water leaving it at the 3.30 P.M. FEB. 9, 1903 4.30. P.M. 6.30 P.M. WELL SALTED AT 2.30 P.M. 3/io NAOH 8 /, NH 4 CI. 8.30 P.M. O OO o o o o O O o o OQOo o O O o o Fig. 12.— Diagram showing the results of experiment 10. Representation of wells and other features as in fig. 8. The velocity of the ground water was 11^ feet for twenty-four hours. The electro- lyte was two-tenths caustic soda and eight tenths sal ammoniac. lower end, were determined for each of the two periods into which each experiment was divided. It will be seen by consulting column 4 of the table that the flow of water per unit head during the first portion of each experiment was essentially identical per unit head to the second portion of each experiment. The slight differences in the numbers is much smaller than the unavoidable experimental error. It was concluded, therefore, that the progressive clogging of a sand filter is duo to the growth of organisms, and that the formalin added constituted an effective reined v. 36 RATE OF MOVEMENT OF (T2TOERGBOUND WATERS. [no.140. Altogether L8 experiments were carried out in this tank. In the first L5 tests Picnic Point gravel was used in the tank: during the last 3 tine glacial sand replaced the Picnic Point gravel. The glacial sand had effective size o\' grain, as determined by King's aspirator, of <».4<> mm. A mechanical analysis of the Band is given in Table VI, WELL SALTED AT 2:00 P.M 6 P.M. O O o 8 P.M. Fig. 13.— Diagram showing the results of experiment 11. Representation of wells and other features as in Bg. 8. The velocity of the ground water was 11.7 feet for twenty-four hours. The electro- lyte was sal ammoniac in concentrated solution. (p. 30), and a summary of the data obtained during the experiments is placed in Table VIII. No difficulty was experienced in maintaining very low gradients to the water plane in the tank, a slope of water of 18 feet to the mile being easily brought about by proper adjustment in the apparatus. In this way actual field conditions of the flow of water were very SLICHTER.] EXPERIMENTS IN THE HORIZONTAL TANK. 37 closely approximated, and velocities less than 10 feet a day could be maintained by the use of the low gradients. The large well marked W was designed to receive various electrolytes while the water was mov- ing through the gravel under the selected uniform head. The small one-half inch wells placed at the corners of the 6 inch squares were designed to serve as test wells from which samples of the water could be taken at stated intervals, and the exact area spread over by the electrolyte could be ascertained by chemical analyses. A series of pipettes were coupled together in such a way as to permit the taking of a sample from each row of test wells at the same time. By the use 9:15 A.M. 1 1:15 A.M. WELL SALTED AT 8:45 A.M. Fig. 14. — Diagram showing the results of experiment 12. Representation of wells and other features as in fig. 8. The velocity of the ground water was 21.7 feet for twenty-four hours. The electro- lyte was sal ammoniac in concentrated solution. of this device a complete set of samples could be taken from all the test wells in the tank in a very few minutes. The results of the experiments are best shown b} T the series of dia- grams figs. 8 to 20, in which the strength of the electrolyte found at each test well is shown by a circle of appropriate size. Among the various electrolytes tested were ammonium chloride (sal ammoniac), sodium chloride (common salt), concentrated ammonia water, and mixtures of ammonium chloride and caustic soda, or lye. One of the most remarkable conclusions from the experiments was that diffusion plays but a very small part in the spread of the electrolyte through the ground water. In none of the experiments was it found 38 RATE OF MOVEMENT OF UNDERGROUND WATERS. [no. mo. 7.30 P.M. MARCH WELL SALTED AT 8.30 P.M. ' N H4 CI Fig. 15. — Diagram showing the results of experiment 13. Representation of wells and other features as in fig. 8. The velocity of the ground water was 3fi feet for twenty-four hours. The electrolyte was sal ammoniac in concentrated solution. Note the narrow stream of electrolyte due to the high velocity. 7.00 P.M. MARCH 9, 1903 7^30 P.M. 8. 00 P.M. 8*30 P.M. 9.00 P.M. WELL SALTED AT 6.30 P.M. V.oNaOH 9 /ioNH4 CI. Fi<;. 16.— Diagram showing the results of experiment 14. -Representation of wells and other features as in fig. B. The velocity of the ground water was :;.'>: feet for twenty-four hours. The electrolyte waa one-tenth caustic soda and nine-tenths sal ammoniac in solution. SLICHTEB.] EXPERIMENTS IN THE HORIZONTAL TANK. 39 MARCH 16, 1903 6.00 P.M. WELL SALTED AT 5.30 P.M V10NH4 CI. VioNaOH 7.00P.M. / ° \ I O \ I O \ I O O O : 6.30 P.M. . \ . A ' o o o '.o O O; 8.00, P.M. Fig. 17.— Diagram showing the results of experiment 15. Representation of wells and other features as in fig. 8. The velocity of the ground water was 3(>.4 feet for twenty-four hours. The electrolyte was one-tenth caustic soda and nine-tenths sal ammoniac in dry crystals. Fig. 18. as in rig Diagram showing the results of experiment 16. Representation of wells and other features 8. The velocity of the ground water was 11.5 feel for twenty-four hours. The electrolyte was sal ammoniac. 40 BATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. 4.00 P. M 10. OOP. M MAY.14 5.00.P.M. 1903 G.00,P.M. 800P.M. WELL SALTED AT 3.00 P.M V. V10NH4 CI. JioNaOH Fig. 19.— Diagram showing the results of experiment 17. Representation of wells and other features as in fig. 8. The velocity of the ground water was 11.6 feet for twenty-four hours. The electrolyte was one-tenth caustic soda and nine-tenths sal ammoniac. 1.00.P.M. MAY 23, 1903- ^j^ WELL 4.00.P.M. SALTED AT 12.00 M. NaOH 6.00 P.M. 4 Fig. 20.— Diagram showing the results of experiment 18. Representation of wells and other features as in fig. x. The velocity of the ground water was 11 was caustic soda. 9 feet for twenty-four hours The electrolyte slichtek.] EXPERIMENTS IN THE VERTICAL TANK. 41 that the electrolyte extended more than about 3 inches upstream from the large well W. This fact can be seen by consulting the series of diagrams illustrating the distribution of electrolyte. In general, it can be seen that the electrolyte moves downstream in a pear-shaped mass, the width of the stream varying somewhat with the nature of the electrolyte used. The high velocities always gave a stream of electrolyte which was quite narrow and the low velocities gave broader streams. The solution of concentrated ammonia water gave the broadest stream. This was probably due not so much to the dif- fusion of the ammonia gas in the water as to the low coefficient of viscosity of the ammonia water. Experiments in the field had indi- cated that the mixture of sal ammoniac and caustic soda would spread in a broader stream than sal ammoniac alone. By comparing the results of experiments 14 and 15 with that of experiment 13, it will be seen that this assumption could not be verified to an} T considerable extent. In a similar wa} T , experiments 9 and 10 may be compared with experiment 8, and experiments 17 and 18 may be compared with experiment 16. It seems to be conclusively shown by these experiments, as has been already stated (pp. 22-23), that the diffusion of the dissolved salt plays a very small part in the wa} T in which the electrolyte is distributed in the moving current of ground water, but, as already stated, that the central thread of water in each capillary pore of the soil moves faster than the water in contact with the walls of the capillary pore. Like- wise the spread of the electrolyte, as shown by these experiments, is not to be explained by the diffusion of the salt, but must be explained by the continued branching and subdivision of the capillary pores around the individual grains of the sand. The stream of electrolyte issuing from the salt w T ell W will gradually broaden as it passes down- stream, because each thread of it must divide and divide again and again as it meets with each succeeding grain of soil. If diffusion had much to do with its rate of spread, it would also make itself apparent b} T causing an upstream motion to the electrolyte against the current of ground water. As before stated, in no case did the electrolyte succeed in moving upstream a distance as great as 3 inches. EXPERIMENTS IN THE VERTICAL TANK. The experiments carried on in the winter of 1903-4 had as their object, in addition to those of the previous year, the determination of the law of distribution of the electrolyte in a vertical plane. For this purpose a tank was constructed of wood, as shown in fig. 21 and PI. VIII. The inside dimensions of this tank were 4 feet high, 4 feet (5 inches long, and 8 inches wide. At each end of the tank chambers 3 inches wide were constructed of perforated brass, similar to those 42 RATE OF MOVEMENT <>F DNDERGBOUND WATERS. [no. uo. used in the horizontal tank. leaving a total length of 4 feet available for gravel. Horizontal tubes of slotted brass one-half inch in diameter extended through the side of the tank at the corners of squares 6 inches on a side, as shown by the small circles in the side elevation, fig. 21, These tubes of horizontal test wells were stuck through holes bored in the side of the tank and were supported at one end by a thumb tack soldered to the end of the tube and at the other end by the side of the tank, the tube being slightly longer than the inside width of the tank. A perforated rubber stopper containing a glass tube was placed in the hole, one end of the glass tube extending to the middle of the tank, the other end of the tube projecting outside of the rubber stopper to receive a small rubber tube, which was kept closed by means of a pinchcock. These tubes furnished ready means of drawing out samples of water from different positions in the tank. On top of the tank, in the reproduction of the photograph of the apparatus. PI. VIII. can be seen the scales carrying the tubs of gal- vanized iron from which the water was run to a regulating apparatus consisting of a needle valve and float at the upper left-hand corner of the box similar to that used in the horizontal tank. The head of water in the two end chambers of the tank was measured by two glass gages placed about one-half inch apart, communicating with the chambers by large rubber tubes. The readings of the meniscus in the glass tubes of the gages could be readily estimated to one-half hundredth of an inch. The gravel used in the experiments in the vertical tank was Madi- son glacial sand, the same as that used in experiments 16, 17, and 18 in the horizontal tank. Seven experiments were completed with this apparatus, the general results of which are tabulated in Table IX. Table IX. — Data obtained during experiments on flow of ivater through Madison glacial sand in the vertical tank. Date. 1904. Feb. 22 .Mar. 2 .Mar. 3 Mar. 5 Mar. 12 Apr. 18 V Apr. 11) Tempera- ture. 17.8 19.2 14.0 18.4 16.0 14.9 17.0 H yd raul ic gradient. Per cent. 3.10 2.08 5.41 1.00 2.10 11.91 5. 58 Feet per in il> . 164 110 286 53 112 630 295 Discharge. Pounds pi r hour. 36.25 26.53 56.00 9.94 23.36 129.66 51.68 Cubic feet per mi ll nil . 0. 00965 . 00706 .0149 . 002645 . 00622 . 034H .0138 Area of cross sec- tion. Velocity. Sijnare Jul. Feet ]>( r litem. 2.34 16.90 2.54 11.42 2.50 24. 50 2.56 4.28 2. 54 10.10 2.42 58. 80 2.50 22.70 Velocity per unit head. Feet per diem. 0.10 .10 .09 .08 .09 .09 .08 U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. VIM VERTICAL TANK USED IN LABORATORY EXPERIMENTS. The small test wells from which samples were drawn are equipped with glass gauges containing colored water, indicating distribution of pressure while right chamber of tank is kept empty. I.U IITER.] EXPERIMENTS IN THE VERTICAL TANK. 43 The electrolyte was introduced into the well marked V\ r , shown in fig. '21, and the samples from the various test wells were drawn out at stated intervals into test tubes and analyzed. The results of the Fig. 21.— Diagram showing construction and dimensions of vertical tank used in laboratory experi- ments on the flow of ground water. The small circles indicate the position of the test wells. experiments on the vertical distribution of the electrolyte are best shown b} 7 the diagrams, a series of which are given in figs. 22 to 29. Pig. 22.— Diagram showing results of vertical-tank experiment 1, February 22, 1904. Well W was salted at 11.40 a. m. with sal ammoniac; velocity of the ground water was 17.06 feet a day, head, 1J inches. The contours show the distribution of salt at 12.10 p. m. In the series of six diagrams for experiment 1 the distribution of the electrolyte is shown by the contour curves for each one-half hour 44 RATE OF MOVEMENT <>F UNDERGROUND WATERS. [no. 140. period after t ho beginning of the experiment. A single dose of 2 ounces of sal ammoniac was introduced into the well \V it L1.40a. in., Fig. 23.— Diagram showing results of vertical-tank experiment 1, February 22, 1904. Well W was salted at 11.40 a. m. with sal ammoniac; velocity of the ground water was 17. 0G feet a day: head, 1$ inches". The contours show the distribution of salt at 12.40 p. m. on February 22, 1904. As will be observed by consulting the dia- grams, the dissolved salt entered the ground water and passed to the Fig. 24.— Diagram showing results of vertical-tank experiment 1. February 22,1904. Well W was Baited at 11.40 a. m. with sal ammoniac, velocity of the ground water was 17.06 ieet a day; head, 1.} inches. The contours show the distribution of salt at 1.40 p. m right with the moving stream, at the same time moving si ightly down- ward, as shown bv the contour curves. The velocity of water through SUCHTER.] EXPERIMENTS IN THE VERTICAL TANK. 45 the gravel during this experiment was about 17 feet for twenty-four hours. The elliptical outline of the contour curves is due to the two Fig. 25. — Diagram showing results of vertical-tank experiment 1, February 22, 1904. Well W was salted at 11.40 a. m. with sal ammoniac; velocity of the ground water was 17.06 feet a day; head, 1£ inches. The contours show the distribution of salt at 2.40 p. m. components of motion, one component being the velocity of ground water to the right, and the other being the downward motion, due to Fig. 26.— Diagram showing results of vertical-tank experiment 1, February 22, 1904. Well W was salted at 11.40 a. m. witn sal ammoniac; velocity of the ground water was 17.06 ieet a day; head, 1£ inches. The contours show the distribution of salt at 3.40 p. m. the high density of the solution of sal ammoniac. It will be noticed that the elliptical contour lines have their longest dimension sloping 46 BATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. downward t<> tin* right, us they should if they represent the resultant of these two motions. It should also be noted (consult the diagrams) Fig. 27.— Diagram showing results of vertical-tank experiment 1, February 22, 1904. Well W was salted at 11.40 a. m. with sal ammoniac; velocity of the ground water was 17.06 feet a day; head, li- inches. The contours show the distribution of the salt at 4.40 p. m. A second dose of salt was placed in well \Y at 4 p. m., and the diagram represents the two masses of electrolyte passing for- ward with the ground water. Fig. 28.— Diagram showing results of vertical-tank experiment 2, March 2, 1904. Well W was salted at 2.25 p. m. The velocity of ground water was 11.42 feet a day; head, 1 inch. The salt used was sal ammoniac. Contours show the distribution of salt at 2.55 p. m. that after an interval of an hour nearly all of the electrolyte had left well W, and the water in the well had become fresh again, At 4 p. m. SI.ICHTER.] EXPERIMENTS IN THE VERTICAL TANK. 47 an additional dose of sal ammoniac was placed in well W, the effect of which is clearly shown in the contour curves for 4.40 p. m. Here two m?sses of dissolved electrolytes can be observed traveling simul- taneously through the sand. It should also be noted in these diagrams that the electrolyte does not pass upstream, or against the current of ground water more than lor 2 inches. Two sets of contour curves are also given for the second experiment, that of March 2, 1904, in which the same electrolyte was used, but the velocit} T of the ground water was reduced to 11.42 feet for twenty-four hours. The well W was salted at 2.25 p. m., contours being given for 2.25 and 7.25 p. m. It will be observed that for the lower velocit} T of ground water the electrolyte sinks to a greater depth than in the case Fig. 29.— Diagram showing results of vertical-tank experiment 2, March 2, 1904. Well W was salted at 2.25 p. m. The velocity of ground water was 11.42 feet a day; head, 1 inch. The salt used was sal ammoniac. The contours show the distribution of salt at 7.25 p. m. A comparison of figs. 28 and 29 with 22 to 27 shows the larger vertical motion of the electrolyte, in the case of the lower velocity of experiment 2, as compared with the higher velocity prevailing during experiment 1. of the higher velocities of the first experiment. Experiments were also carried out with common lye as ■ electrolyte. This salt is very much heavier than sal ammoniac, and it was noted that it sank much faster than the solution of sal ammoniac for similar velocities of ground water. One of the most interesting experiments with the vertical tank was made for the purpose of determining the amount of diffusion of the electrolyte. For this purpose the electrolyte was introduced into the well W and the ground water was permitted to remain stationary, no water being run into or out of the tank during the eight hours covered by the experiment. The well \Y is placed exactly midway between 48 KATE OF MOVEMENT OF I' N DKRGROUND WATERS. [no. 140. columns and 1 of the small test wells, as can be seen from the dia- gram. For the purpose of the " still" experiment the uppermost test well of column 2 was removed and well \Y was placed directly over column 2. A charm 1 of salt was introduced into the well W at 9 a. m.. and samples were taken at the end of one-half hour and at the end of each hour thereafter until 5 p. in. The salt was found to drop verti- cally with a rapidity equal to the vertical component of motion noted in the experiments in which now took place. In the eight hours of the test no portion of the charge could be detected in the test wells of columns 1 or 3. This experiment showed that the electrolyte had not ( )58.9 62.6 37.2 1^/ y^6' 60.8 65.1 ,005 .01 .015 .02 .025 .03 -035 DISCHARGE IN CUBIC FEET PER MINUTE Fig. 30.— Diagram illustrating the variation in the rate of flow of ground water with the variation in head or hydraulic gradient, as observed in the experiments in the vertical tank. The figures attached to the small circles in the diagram designate the temperature, Fahrenheit, of the ground water during the experiment. The straight line represents the theoretical law of flow if the rate of flow varies directly as the head. diffused sufficiently to reach the wells of columns 1 and 3, while drop- ping a vertical distance of about 3 feet. The law of direct variation of the flow of ground waters with the head under which the flow takes place are verified by the experiments in the tank. The results are represented graphically in fig. 30. Exact agreement with this law would require all of the plotted points in this diagram to lie upon the straight line, provided the temperatures were the same. The larger departures from the straight line are not due to temperature differences, but to the high viscosity of the lye solutions used in those particular experiments^ SLIGHTER. j LABORATORY EXPERIMENTS. 49 INVESTIGATION OF THE ACCURACY OF THE ELECTRIC METHOD OF DETERMINING THE VELOCITY OF THE FLOW OF GROUND WATERS. The vertical tank offered a ready means of checking the accuracy of the method of measuring the velocity of ground waters with the electric underflow meter. For this purpose the chambers of per- forated brass at the upper and lower ends of the tank served as the upstream and downstream wells, respectively, and an electrode was sunk in the sand 2 inches from the lower partition, which answered the purpose of the electrode usually placed inside the downstream well. The apparatus was then connected in accordance with the method used in actual held work. A solution of sal ammoniac was placed in the upper chamber. The water running through the tank was weighed before it entered and after it left the apparatus, and observations were made of the electric current every fifteen minutes and sometimes oftener. Two experiments were made, one with a head of water of 2.68 inches and one with a head of 5.75 inches. From the weight of water discharged the computed velocity during the former was 23.15 feet a day, and during the latter the velocity was 58 feet a day. From the points of inflection of the two electrode curves the velocities were computed to be, respectively, 23.25 and 64.10 feet a day. This shows agreement in the case of the lower velocity within a very small fraction of 1 per cent, and in the case of the higher velocity within 10^ per cent of the actual rates. These results show that the electric method is sufficiently accurate for the purposes for which it is intended. It is very likely that if the tank in which these experiments were carried out had been wider the percentage agreement for the high velocity would be even closer than 10 per cent, for it must be remem- bered that the narrowness of the tank tended to bring the concentrated portion of the stream of electrolyte to a given downstream point more rapidly than if the tank had been wide enough to permit the electro- lyte to spread in its natural way. irr 140—05 4 CII A l'TER IV. MEASUREMENTS OF THE UNDERFLOW AT THE NARROWS OF THE RIO HONDO AND SAN GABRIEL RIVER, CALI- FORNIA. The following underflow measurements were made during the sum- mer of 1902 at the narrows of the Rio Hondo and the San Gabriel River, about 10 miles east of Los Angeles, Cal. The ultimate source of the streams referred to is found in the San Gabriel Mountains, a range which runs nearly east and west about 40 or 50 miles from the southern coast line of California. The main portion of the mountain drainage which supplies this particular stream is collected into one of the large canyons of the range, known as the San Gabriel Canyon. Like that of other streams that originate in these mountains, the water is not carried above ground much farther than the mouth of the can} T on, except in times of extreme flood. The ordinary flow of the river sinks into an enormous alluvial delta cone of gravel and mountain debris, and passes underground in a broad, gently sloping valley until it is interrupted by a line of shale hills about 10 miles south of the mountain range. This line of hills acts as a dam to the underground waters, except for a break about 2 miles in width, where the drainage of the valley escapes to the sea. This break constitutes the so-called "Narrows" of the river. In consequence of the narrow outlet a large quantity of the ground water is brought to the surface, first showing itself about 2 miles above the narrows, and increasing in volume as it enters the contracted part of the pass. At the present time the surface waters appear as two distinct streams, the Rio Hondo on the west side and the San Gabriel River on the east side of the narrows. In August, 1900, the flow of the Hondo at Old Mission bridge was 23 second-feet. The flow of the San Gabriel was somewhat larger. The Whitney electrolytic bridge indicates that the ground water and surface waters are substantially identical in character, containing 15 to 25 parts per 100,000 total solids. The measurements of the rate of underflow were made by the elec- trical method as previously described in this paper. At the time of making these measurements the recording instruments had not been perfected for field use. so that all of the work was done with the hand apparatus. The test wells used were 2-inch drive wells, with 42-inch 50 slkhter] NARROWS OF HONDO AND SAN GABRIEL, CALIFORNIA. 51 points and 18-inch well-point extensions. The wells were arranged as usual, one upstream, designed to receive the electrolyte, and the others downstream, 2 feet apart on an arc of a circle of 4-foot radius. In most instances the electrolyte moving with the ground water would show itself at but one of the downstream wells, but in one or two cases it reached two of the downstream wells, and in a few cases the first setting of the wells did not correspond to the actual direction of the motion, so that the lower wells were not touched at all by the dissolved electrolyte. Measurements were made at four stations located within the narrows of the rivers named. The first and second stations were located under a wagon bridge over Rio Hondo near the Old Mission. Two groups of wells were driven, and the location and direction of these wells with reference to the bridge are shown in fig. 31. The electrical cur- rent was observed separately for a circuit between the casing of the upstream well and the casing of each of the downstream wells, and also for the circuit between the casing of each downstream well and the brass rod electrode contained within it, a direct-reading ammeter being used. The velocity at the first station was 3.8 feet per diem. The direction of flow departed slightly from the direction of the sur- face river, being 10 degrees west of south. At the second station the electrolyte showed itself at both of the downstream wells, being stronger, however, in well F. The velocity here was 6.6 feet a da}\ Points of special interest are the several steps in which the ampere curves rise, as shown on the electrode cir- cuits for both wells E and F. These indicate different velocities of ground water in the different strata penetrated by the wells. The w T ell points and well -point extensions being covered with new bright brass gauze in these first tests, the different porous strata registered them- selves on the brass gauze by blackened bands caused by the corroding influence of the electrolyte. In the present case there were three dis- tinct zones marked off on the wells, of about 24, 20, and 8 inches each. The velocities in these strata undoubtedly differed from one another, and hence caused the steps in the ampere curve. At both stations 1 and 2 the ground water was artesian in character, rising in the wells about an inch for each additional foot increase in depth. The wells were 16 feet deep and showed about 15 inches of artesian head above the water in the flowing stream. The third station was established on the San Gabriel River just below the wagon bridge on the Whittier road. A special point of interest at this location is the fact that the river totally disappears in its gravel bed a few rods below this bridge. We hoped to secure some facts concerning the direction and velocity of the disappearing water. A double row of wells was driven across the river bed in three -VJ RATE OF MOVEMENT OF UNDERGROUND WATERS. [No.l4a groups about 33 feet apart, as shown in fig. 32. All of the wells. except a group near the left bank, pumped very poorly and were evidently in very tight material. The group of wells near the left 1.40 1.30 1.20 1.10 AMPERE CURVE WELL' B' "*\ \ \ 1 1 1 IM. 7 9 AM. 11 A.M. 1 1 >.M. 3 b 7 3 11 P.M. 1 A.M. 3 ! < ) 11 A.M. 1 ( >.M. 3 JULY 20" JULY 21- Fk.. 31.— Diagrams showing the velocity and direction of How of the bank consisted of one upstream well, E, and two downstream wells, F and G. These wells evidently penetrated two different strata of water, as a velocity of 48 feet a day was observed between E and F, 6Lichtbb.] NARROWS OF HONDO AND SAN GABRIEL, CALIFORNIA. 53 while a second result was strongly developed, indicating a velocity between E and G of 4.8 feet a day. This latter rate was due to the lower stratum of water, whose direction of motion crossed at an angle of 35 : that of the upper current of disappearing river water. Fur- ther observations at this point were rendered impossible by the break- ing of a dam some distance above the station, which completely flooded the wells after the above observations had been made. The fourth station was established in a walnut grove on Temple's ranch, near the main road from El Monte to Downey. This location is about half way between the two bluffs of the narrows. The group consisted of four wells, and a velocity was found to be rather low, 14 inches a day. The direction of flow was due south. The fifth station was on the bank of the San Gabriel River, just 0.70 AMPE RE CI RVE WELL" F" > • -- 1 0.60 * * / HI ' cc 1 . 4 Ft. 6.6 Ft. Per Da) / Casing 111 Ve '• 14.5Hrs. 0- 2 0.30 < z^^" <^**" ' i^~ / l AMPERE CURVE \ WEL L*E* r~~ m t \ e c* oi e .-_' / / / * Casi ng 8 10 P.M. 12 -JULY 22 x- 10 A.M. 12 M, 2 P.M. — JULY 23-- 6 P.M. underground water at tin- narrows of the Rio Hondo, stations 1 and 2. above the head works of the Ranchita and Los Nietos ditches. The velocity determined was 5.3 feet a day, in a direction due south, mak- ing an angle, however, of about 45° with the direction of the surface 54 RATE OK MOVEMENT OF UNDERGROUND WATERS. [no. no. stream at the same point. The direction of How and ampere curve arc shown in fig. 4. The cross section of the alluvial depositsat the narrows of the Hondo and the San Gabriel is about L0,000 feet wide and probably docs not exceed 600 feet in depth. If we assume that the porosity of the under- flow gravel is 33 per cent and that the average velocity of the ground water is 1" feet a day. the resulting estimate of the amount of water which passes underground through the narrows is 230 second-feet, or Dearly four times the How of the surface streams. This is undoubtedly a maximum estimate, as there is no indication that the average velocity i.^ as high as io feet a day. Four feet a day may he assumed as a fair minimum value of the average velocity. This would correspond to a total underflow of 92 second-feet. 1.00 ! 0.90 AMPERE CI RVE WELLS F&G ' ' i i i i i • O.80 _> VEL.=J_ FT. = 57.6 FT. PER DAY. , ' 1 y 3 hrs. C.70 i 0.60 i i o°^-~ .•i i 1 ,°# Vc- i CO 111 £0.50 G. **$ \ * i i \ / s ■*0.40 V~- ""-'' - -\ — . __Elec TRODE WELL ._'9l- / C.30 0.20 n£« C.10 12 M. 2 P.M. 4 6 8 «; JULY-29 — 10 P.M. 12 2A.M. 4 6 8 10 A.M. -X JULY 30 > Fig. 32.— Diagram of velocity and direction of flow of underground water at the narrows of the San Gabriel River: station 3. The measurements established the existence of a distinct underflow of moderate velocity through the alluvial deposits of the narrows. In low stages of the surface streams the underflow probably represents a drainage from the upper valley in excess of that discharged by the surface streams. The substantial identity of the water of the under- flow and the water of the surface streams is proved by tests with the Whitney bridge, so that we may conclude that the original mountain stream appears at the narrows as a composite river, consisting of sur- face streams bordering both the east and the west bluffs of the narrows, together with a very wide and deep bat slowly moving underflow occupying the entire major trough of the valley. CHAPTER V. MEASUREMENTS OF THE I XDERFEOW AT THE NARROWS OF THE MOHAVE RIVER NEAR VICTOR VlXIiE, CAL.. CONDITIONS AT THE STATION. The Mohave River rises on the slope of the Sierra Madre Moun- tains in San Bernardino County, Cal., its headwaters flowing from elevations of 5,000 to 8,000 feet. After following a general north- erly course the stream disappears in the Mohave Desert a short dis- tance below Rarstow, Cal. After leaving its mountainous canyon the stream gradually loses water, and for a large portion of the summer its bed is dry for a greater part of its course on the plains. At a point about 16 miles north of its source the river passes through a narrow gorge called the " Narrows" of the river. The granite uplift which forms this gorge constitutes a dam that raises the underflow to the surface, so that within an area that extends from a point a mile and a half above the gorge to a point a considerably greater dis- tance below the gorge the stream is of perennial flow. A view of the narrows of the river is shown in PL IX. This gorge is just south of the village of Victorville, Cal., a station on the Santa Fe Railway. The place had been under investigation as a possible site for a dam by the United States Geological Survey during the season of 1899 and previously. Reports on this subject will be found in the Eighteenth Annual Report, United States Geological Survey, Part IV, page 708, and the Twenty-first Annual Report, United States Geological Survey, Part IV, page 471. The permanent diy -season flow of the river at the gorge varies from about 30 to 60 second-feet. On August 15, 1902, the discharge, as measured by J. B. Lippincott, was 33 second- feet. Soundings have been made to bed rock at three different lines across the narrow part of the gorge by the United States Geological Survey. The positions of these cross sections are shown in fig. 33. The river-gage rod of the United States Geological Survey is located on the right bank of the river at the end of line 3. This line was selected as the location for the underflow measurements. . The maxi- mum depth to bed rock at this point is 46 feet, as is shown in the approximate cross section given in fig. 34. The material tilling the gorge and constituting the bed of it is a coarse angular granite debris of a size somewhat larger than buckwheat. Mechanical analysis of the gravel tilling the gorge is given in Table VI (p. 30). Determinations of 55 5C BATE OF MOVEMENT OF rXDKRl'JRorND WATERS. [no. 140. tin 1 effective size of the gravel by King's aspirator showed a mean diameter of 0.72 mm. The samples of gravel were taken from the surface material. High velocities of the underflow determined at cer- tain depths indicate that there are sonic streaks of coarser material. z t slighter] NARROWS OF THE MOHAVE NEAR VICTORVILLE, CAL. 57 DESCRIPTION OF EXPERIMENTS. The site selected for the measurement of the underflow at the upper narrows of the river was in a line extending across the river at right angles to its main coarse, as shown in the plan given in tig. 33. The water in the river at this point hardly exceeded a foot in depth, but the water spread over the greater portion of the space between the banks, necessitating the construction of a small foot bridge, shown in ''BED ROCK Fig. 34.-Cross section of the narrows of the Mohave River at which the underflow was measured This line of cross section corresponds to line 3 in the preceding diagram. The rectangles inclosing the figures represent the position and depth at which the underflow was measured The figures inclosed in the rectangles represent the velocity in feet per day at that point in the cross section. PI. X. Another view, PL XI, illustrates the method used in putting down the test wells. The double row of test wells, A, B, C, D, E, F, G, II, I. were driven across the river at this point as located on the plan shown in fig. 35. The gorge at this place is only 120 feet wide, and it was at first thought unnecessary to drive more than a single downstream well for each measurement, as it was believed that the PIG. 35.-Thia plan shows the position of the various test wells used in the underflow investigation, the principal line of test wells corresponds to line 3 of fig. 33. underflow must move through the gorge in a very direct course. It was determined, however, after first salting the upstream wells, that the ground waters must be moving through the gorge at this point at a slightly different direction from that of the surface waters, as abso- lutely no results could be obtained from the wells as first arranged Accordingly, directional wells M, X, Y, Z, U, K were driven as 58 RATE OF MOVEMENT OF UNDERGROUND WATERS. >•■ no. 1.00 .90 .80 .70 - .-■: - ui a. < M T M .*o J .90 •: ■2 2 3 4 5 6 AMPERE CURVE FOR WELL Y. DEPTH 8 FEET. VELOCITY: 52.4 FEET PER 24 HOURS 36.0 " " " Fig. 37.— Diagram showing velocity of underflow at narro. ave River, station E. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. X • I V *•*: DRIVING WELL G AT UNDERFLOW STATION AT NARROWS OF MOHAVE RIVER. Showing shallow character of river at this place during dry season. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER NO. 140 PL. XI DRIVING TEST WELLS AT U N DERFLOW STATION AT NARROWS OF MOHAVE RIVER. The wooden ram used weighed about 60 pounds. 9LICHTEB.] NARROWS OF THE MOHAVE NEAR V10TORVILLE, CAL. 59 shown in fig. 35. In order to determine whether a variation in the direction was the cause of the failure to measure the underflow, some preliminary experiments were carried out at the position lettered E, test wells X and Y being driven 18 inches from the well F. After these wells were put in place, the well E was salted. This test turned out to be unsatisfactory also, but a slight rise in the electric current in well Y was noticed, which encouraged the belief that the direction 1.30 1.20 .90 AMPERE CURVE FOR WELL Y. DEPTH 14 FEET. VELOCITY: 55 FEET PER 24 HOURS. Fig. 38.— Diagram showing the velocity of underflow at narrows of Mohave River, station E. of flow had been missed in the first test. It was surmised that the ground water might be moving so rapidly as not to give time for the electrolyte to spread sufficiently, and hence was able to pass between two downstream wells 18 inches apart. It was also possible that the direction of the flow was to the left of the new well Y. In order to provide for this contingency, the well X was pulled and redriven in the position indicated by the letter Z, 18 inches from Y. 60 RATE OF IfOVEMENT OF rXDEKciROrND WATEBS. [no.140. If it were true that the motion of the ground water of the underflow might be so great that the electrolyte used in well E would not have time to spread sufficiently so as to be certain of coming in contact with one of tN\o downstream wells is inches apart, it would be necessary to modify the method of salting E so as to cause, if possible, a wider path to be traversed by the electrolyte. This was finally accomplished by introducing into the well E, along with the dry sal ammoniac, about 1(> per cent caustic soda, or common soda lye. This, of course, brought about a reaction between the two salts in solution, ammonia CO III £T LOO III a. .2' • J»M 9 10 11 12 1 2 3 * 5 6 «-A.M. AUGUST 18THt^ P.M. AUGUST 1 8TH: — > AMPERE CURVE FOR WELL F. DEPTH 20 FEET. VELOCITY: 24 FEET PER 24 HOURS Fig. 39.— Diagram showing the velocity of underflow at narrows of Mohave River, station E. gas being liberated. It was believed that the liberation of the gas would cause the mixed chemicals to spread more rapidly in the ground waters. This later seemed to be the case, for, carrying out the experi- ment in this way. it was found that the electrolyte reached well Y, where it showed itself very strongly and just grazed well F. The electrode curve for well Y is shown in tig. 40. which indicated a velocity of ground water of about 52.4 feet for twenty-four hours. This velocity was several times greater than any previously deter- mined, which accounted foi the difficulties encountered in interpreting the first failure and in getting the direction of flow. After this expe- slichter.1 NARROWS OF THE MOHAVE NEAR VICTORVILLE, CAL. 61 rience all of the upstream wells were salted with the same mixture of sal ammoniac and caustic soda, and no further difficulties were expe- rienced in getting the velocity and direction of flow. At the various stations, A, E, G, I, determinations of velocity were made at various depths, as is shown by numbers inclosed in rectangular lines at appro- priate points in the cross section, fig. 3i. The pipe casing of well D 2.80 2.40 2.20 2.00 1.80 CO Id F UNDERGROUND WATERS. [no. 140. the well points were driven a few feet below the point at which the deepest measurements were made the points were completely embedded 1.00 CO UJ cr UJ ^.90 < fu 9 10 11 12 ^ 2 3 4 5 P.M.-AUGUST 18TH-.-- AMPERE CURVE FOR WELL H.- DEPTH 30 FEET VELOCITY: 11.7 FEET PER 24 HOURS Fig. 41. — Diagram showing the velocity of underflow at the narrows of Mohave River, station G. in the silted gravel and no water could be drawn from the wells with an ordinary pump. It was therefore very plain that the underflow of .80 .70 .*0 .10 .«0 .80 / / f / 7 i 1 1 1 1 1 1 2 2 l - \ 5 1 -A.M. AUGUST 18TH: -P.M. AUGUST 18TH- AMPERE CURVE FOR WELL K. DEPTH 24 FEET. VELOCITY: 9.1 FEET PER 24 HOURS Fi«i. 42.— Diagram showing the velocity of underflow at the narrows of Mohave River, station I. the gorge was confined to a depth not exceeding about 30 feet. All of the velocities determined in the clear gravel of the gorge ran very slighter] NARROWS OF THE MOHAVE NEAR VICTORVILLE, CAL. 63 high, three of them exceeding a velocity of 50 feet for twent} T -four hours. Taking Schuyler's figures for the area of the cross section of the gorge, 4,160 square feet, and assuming a mean velocity of ground water in the entire section of 50 feet for twenty-four hours, and estimating the porosity of the gravel at 33^ per cent, the total under- flow in the gorge will be found to he less than 1 second-foot. This must be understood to be a maximum estimate. The underflow probably does not exceed 300,000 gallons for twent}-four hours. The gradient of the water plane at and above the gorge is almost exactly 20 feet to the mile. The reproductions of photographs shown herewith illustrate the method of driving the wells used and show the small footbridge and the test wells that were put in place during- the investigation. The tent appearing at the left of the cut contained the instruments from which wires were led to the various test wells (PI. IX). QUALITY OF THE WATER. The quality of the water in the surface stream and in the underflow of the Mohave River at the narrows was determined by tests with the Whitney electrolytic bridge, and the amount of chlorine in the water was determined by titration. The water is remarkable as a desert water for its unusual softness, being very much softer than the usual water found in southern California, as at Los Angeles and neighboring points. This softness is undoubtedly due to the insoluble character of the granitic deposit through which the water flows after leaving its mountain source. It should be remembered that both the surface water in the stream and the underground waters at the location of the narrows of the Mohave River has been flowing in a ground-water stream for 10 or 12 miles of its course. As all of this water reaches the narrows by passing underground for a considerable distance, the results of the few tests made, given herewith in Table X, are of con- siderable interest. 64 RATE OF MOVEMENT OF UNDERGROUND WATERS. ■•HO. Table X. — Quality of the water in tfo dls -it the narrows of tin Mohan /i'' iiy increasing somewhat the number of stations in nX : 1.20 1.00 .80 u QT.60 UJ Q. < .40 .20 6 A.M. 8 AUG. 21 VELOCITY 6 FEET PER DAY. Fig. 43.— Diagram showing velocity and direction of the flow of underground water at Wantagh pumping station (station 2X). Velocity, G feet a day. S.40°E. This velocity was determined while pumps were drawing water from the wells of the driven well plant at a rate of 4,3,000 gallons per twenty-four hours. No velocity detected when not pumping. the area already covered and comparing- with results from drainage areas in Suffolk County, a comparative study of underground drain- age systems would result which ought to have much value in planning sources of supply for Brooklyn. The details of the measurements are given in the reports on indi- vidual stations contained in Table XL The locations of the stations are shown in tig 57. and the curves of electrical current for the vari- ous stations are given in tigs. 5 and 43 to 56. SLICHTER.J LONG ISLAND, NEW YORK. Table XI. — Underflow measurements on Long Island. 69 Number of station. Velocity ofground water; feet a day. Direction. Date, 1903. Depth of wells i t i feet, Kind of point. No. of text figure. 1 5. 5 S. 10° E . . . June 21 22 Perforated pipe. 5 •> 2 ... June 24.... 22 Do. 2 X . . . 6 S. 40° E . . . Aug. 21 22 Do. 43 3 2 June 26 22 Do. 4 2 June 27 22 Do. 5 6.4 S. 8° W . . . June 29.... 22 Common point. ■ 5X1... 5.4 S. 8° W . . . July 3,4... 22 Do. 44 5 V . . . 8.0 S. 22° E . . . Aug. 19 22 Do. _ 6 5.0 S.8°W... July 1, 2... 34 Do. 45 7 2.6 S July 5, 6... July9,10,ll. 20 Do. 46 8 0.0 s 21.6 Open-end point. 8 3.1 X. 34° W . . Julv 14, 15, 16, 17. 21.6 Do. 47 10 2.6 S. 37° E... Julv 17, 18, 19, 20. 28 Common point, 48 11 0.0 ' July 27- Aug. 8. 22 Do. 12 1.07 S. 3° E.... July 27- Aug. 1. 27 Open-end point. 49 13.. 96. S Aug. 3, 4... Aug. 3,4... 16 Common point. Do. | 53 13 6.9 S 16 14 9.3 S Aug. 5, 6 17 Do. 50 15 1.53 S Aug. 6, 7, 8, 9,10. 42 Open-end point. 51 15 X... 6.00 S. 15° W .. Aug. '17, 18, 19. 62.5 Do. 52 16 0. S. 30° E . . . Aug. 10, 11. 16 Common point. 16 X... 77 S. 60° E . . . Aug. 13, 14. 16 Do. } ■ M 16 X... 11.6 S. 60° E . . . Aug. 13, 14. 16 Do. 17 10.6 S. 30° W . . Aug. 12, 13. 20 Do. 55 18 1 S Aug. 15-21. 62 Open-end point, 21 21.3 S. 50° E . . . Aug. 18, 19. 16.5 Common point. 56 22 5.6(?) S. 30° E . . . Aug. 20, 21. 16 Do. INFLUENCE OF THE RAINFALL UPON THE RATE OF MOTION OF GROUND WATERS. An excellent opportunity was presented at one of the stations for noting- the influence of a heavy rain upon the velocity of ground waters. At station 5, Agawam pumping station (see tigs. 58 and 44), the upstream well, A. was salted at 9.45 a. in.. June 29, 1903. Between To BATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. 9 a. m. and 1 p. m. nearly 3 inches of rain fell, so t hut the heavy precipitation coincided with the early part of the ground-water meas- urements. The velocity found was *>.4 feet a day. On July 3 the experiment was repeated, there being no rain in the intervening time. The velocity found in the second trial was 5.3 feet a day. The change in velocity was undoubtedly due to the enormous rainfall during the first experiment. Part of the high velocity during the rain storm may be attributed to the low barometer accompanying the storm, but part JUNE29 10 A.M. 12 JULY 3 12 M. 2 AUG. 19 4 P.M. 6 8 10 12 2 4 6 8 10 12 2 10 12 2 4 6 8 10 12 2 4 2 4 6 8 10 12 2 4 6 8 Fig. 44.— Diagram showing three determinations of velocity and direction of flow of underground water at Agawam pumping station (station 5). Normal velocity of ground water, 5.4 feet a day. S. 8° W. Velocity during heavy rain. G.4 feet a day, S. 22° E. Velocity while pumping from the lines of driven wells, 8 feet a day, S. 22° E. of it should be assigned to the increased head of ground-water pressure and to increased load carried by the soil, caused by the heavy rainfall upon the receiving area. As is shown in another place a ground waters move very much as electricity is conducted in a good conductor,- the most striking qualit}^ being an almost complete absence of true inertia in ground-water motions. The motion of a ma'ss of ground water, even for the highest a Nineteenth Ann. Kept. U. S. Geol. Survey, pt. 2, 1899, p. 331. SLICHTEK.] LONG ISLAND, NEW YORK. 71 velocities, is so slow that the resistance to an accelerating ''orce repre- sented by the inertia of the ground water is almost nothing- when compared with the component of the retarding force due to the capillary resistance in the small pores of the sand or gravel. Actual computation will show that in a uniform sand of diameter of grain oi one-half millimeter, the ground water will reach within 1 per cent of its final maximum velocity due to a sudden application of pressure, or head, in approximately thirty seconds of time. This surprising result of the theory of ground-water motions receives a very striking verification in the increase in velocity noted during the rainstorm as described above. / / / / y Elec trode _-^V // Case 1 c/>1. u Ld 11 -JULY 1- -JULY 2- VELOCITY 5 FEET PER DAY. Fk;. 45. — Diagram showing velocity and direction of flow of underground water at Agawam pumping station I station 6). This station is located near station 5, but wells were 12 feet deeper. These results have important bearings on our knowledge of ground- water phenomena in the neighborhood of a well. They indicate that tin 1 velocitvof the ground waters in the neighborhood of a well reaches a maximum value soon after pumping is commenced. The gradual formation of the cone of depression near the well shows that there must be a progressive augmentation of the initial velocity of the ground waters toward the well. Nevertheless, the rate of depression of the water table is so slow that the ground-water motion established soon after the pumping has begun is substantially the same as its value after prolonged pumping. These remarks have their most 72 RATE OF MOVEMENT <>F UNDERGROUND WATERS, [no. 14a important bearing upon the phenomena of the mutual interference of wells. The interference of one we'll with the supply of a neighboring well is thus seen to conic into existence almost instantaneously and need not wait for tin* establishment of a cone of depression of large area. The phenomenon of tin' cone of depression has much to do with the permanent supply of the well, hut has slight bearing upon the proper spacing of the wells or the percentage of interference of one well with another. Center line of road. 1 P.M. 5 * JULY 5- — JULY 6— - VELOCITY 2.6 FEET PER DAY. JULY 7- Fu}. 46. — Diagram showing velocity and direction of flow <>f underground water at Bast Meadow Brook arid Babylon Road (station 7). This station is a short distance above Aga warn Pond, and the velocity is reduced by the flat water plane due to the presence of the pond. Velocity. 2.6 feet a day. south. SEEPAGE WATERS FROM PONDS AND RESERVOIRS. The work on Long Island afforded some unusually good opportuni- ties of determining the rate of seepage below tin 1 impounding dams of some of the storage ponds which the Brooklyn Water Works has established north of the conduit line referred to in the opening pages of this chapter. The batteries of driven wells, which have been placed a few hundred feet south of nearly all of these ponds, were not used during the summer of 1908. as the heavy rains furnished a sufficient quantity of surface water, and the auxiliary supply from the wells was not drawn upon, a- usual, during July and AuguSt. Station ~> is below the Agawam Pond and somewhat within the line of seepage from the pond, as can be seen by consulting fig. 58. The normal velocity of SLICHTER.] LONG ISLAND, NEW YORK. 73 ground water at this station is 5.3 feet a day. At station 7, just north of the pond, the velocity was 2.6 feet a day. It seems clear that the natural velocity at these points, if the influence of the dam and pond were removed, would be about -1 feet a day. The velocity at station 6, located but a few feet from station 5, was 5 feet a day at a depth of 34: feet, as compared with 5.3 feet a day at a depth of 22 feet. The dam has the effect of making the water table nearly level in the immediate neighborhood of the pond, and also of greatly augmenting the slope of the water table for a short distance below the pond. The lower velocity above the pond and the higher velocity below the pond correspond Pipe line .00 .80 60 .40 .20 4 P.M. 8 12 ■*-i JULY 14^ — 8 12 —JULY 15- 8 12 4 — JULY 16 VELOCITY 3.1 FEET PER DAY Pig. 47. — Diagram showing velocity and direction of flow of underground water near Merrick pump- ing station (station 8). The ground water at this point slopes in a northerly direction toward the brick conduit north of the Long Island Railroad. The velocity found was 3.1 feet a day. N. 34° \Y. 'The northerly flow at this point is undoubtedly due to seepage into the conduit. with these facts. When there was no flow over the waste weir of the dam. the flow of the small stream which rises below the dam was meas- ured at the bridge marked A in tig. 58. On- July 10 this flow was 1.2 second-feet, practically all of which represented seepage water from the reservoir. A flow of 1.2 second-feet or 103,000 cubic feet a day represents an amount of water flowing through a bed of sand 30 feet deep and 1,000 feet wide, at a velocity of 1 foot a day. the porosity of the 74 RATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. sand supposedly being equal to one-third. The normal velocity of the ground water is augmented, as shown by the measurement quoted above, by somewhat more than 1 foot a day. The width of the lower end of this pond, or the length of the earthen dam. is about 1,400 feet, SO, basing the estimate on this minimum length and on a minimum depth o\' :*><> feet, an augmented velocity of 1 foot a day would give a minimum estimate of the seepage from the dam of 1.(5 second-feet. Since L.2 second feet are known to actually come to the surface to feed the stream below the dam, it is evident that this estimate of seep- age is a minimum. It seems evident that a considerable volume of ROAD. .70 .60 .50 7° E. seepage water could be recovered by extending the line of driven wells of the Agawam pumping station to the east of the present termi- nus, a distance of 600 or 700 feet, without serious lowering of the water plane. A test well was driven in the lower south end of Agawam Poni to a depth of 10 feet to determine the pressure gradient of ground water beneath the surface of the pond. The water in this test well stood about 1 foot lower than the water in the pond itself, showing a slope of the water plane or a hydraulic gradient of 7 feet to a mile. SLICHTER.] LONG ISLAND, NEW YORK. 75 The gradient of the water plane below the dam — that is, between the dam and station 5 — was 17 feet to the mile, so that the velocities to be compared are: Station 7 above pond; gradient, 7 feet per mile; velocity, 2.6 feet a day. Station 5 below pond; gradient, 17 feet per mile; velocity, 5 feet a day. These results check very favorably, especially when it be considered that the gradient above or north of station 7 was probably 10 or 12 feet per mile, which would make the effective gradient at this station somewhat greater than 7 feet per mile. Grand Avenue si 2.00 1.60 to 111 ff 1.00 a. 2 < .80 .40 .20 12 M. 12 M. JULY 27-* 28— 12 M. -29- 12 M. -* 30— 12 M. -31- 12 M. -AUG.1- 12 M. — 2— ^- 12 M. — 3— VELOCITY 1.07 FEET PER DAY Fi<;. 49. — Diagram showing velocity and direction of the flow of underground water at Grand Avenue and Newbridge Brook (station 12). Velocity, 1.07 feet a day, S. 3° E. This is the lowest velocity determined on Long Island. Very striking results were obtained below the dam at the Wantagh Pond, where measurements were undertaken to determine the rate of seepage. The dam of Wantagh Pond runs parallel to the right of way of the Long Island Railroad, about 75 feet north of the road, and has an extreme length of 500 or 600 feet. About 150 feet south of the railroad, downstream from the reservoir, the city of Brooklyn began in 1903 the construction of an infiltration gallery, consisting of a line of 36-inch double-strength tile, laid at a depth of 16 feet below the 76 BATE OF MOVEMENT OF FNDFKO K<>F N I) WATERS. [NO. 110. CenteMineoijoad. ft .80 ce Ul a. < -40 Mjl 1 P.M. 3 5 7 9 11 1 3 5 < AUG-5 x AUGr-6- VELOCITY 8.6 FEET PER DAY. Fig. 50. — Diagram showing velocity and direction of flow of underground water at BeUevue road (station 14). Velocity. 8.6 feet a day. south. 1.40 1.20 1.00 to UJ .80 m O. 2 .60 < .40 .20 -Aug. 6-x- 12 M. Aug. 7- 12 M. -Aug. 8 12 M. -Aug. 9--- 12 M. Aug. 10 VELOCITY 1.53 FEET PER DAY. Fio. 51.— Diagram showing velocity and direction of the How of underground water at BeUevue road (station 16). Velocity, 1.53 feet a day. south. SLIGHTER.] LONG ISLAND, NEW YORK. 77 water plane. It is proposed to extend this gallery for a mile east and west from the Wantagh pumping station. Stations 13, 16, and 17 were established for the purpose of measuring the normal ground-water velocities at the depth (16 feet) of the proposed gallery. Two of these stations are immediately south of the pond and in the apparent direct line of seepage, while station 17 is located slightly east of the edge of the pond and, as seems evident from fig. 59, just on the edge of the main influence of seepage from the ponds. The seepage velocities at stations 13 and 16 turned out to be enormous, the velocity at station 13 being 96 feet a day, south, while at station 16 the velocity was 77 feet a day, the direction being about 30° east of south, the deflection being toward the neighboring stream, as shown in fig. 59. These .50 .40 g .30 tc l±J | .20 < .10 - y^ .0 2 P.M. 6 10 < — AUQ.17— 2 6 10 2 AUG. 18 VELOCITY 6 10 2 AUG. 19- FEET PER DAY. Fk;. 52.— Diagram showing velocity rnd direction of the flow of underground water at Bellevue road (station 15 X). Velocity, 6 feet a day, S. 15° W. This station is the same as station 15, but measurement of velocity was made below a stratum of clay or bog material at a depth of 62.5 feet, 20 feet deeper tban tbe measurement shown in fig. 51. velocities are the highest the writer has determined. They may be regarded as record-making rates for the horizontal motion of ground waters. Both measurements were made with the recording instru- ments, and by consulting the curves in figs. 53 and 54 it wnll be noted that eaeh curve has two maximum points, which must correspond to the velocities in two distinct layers of gravel. The secondary veloc- ity for station 13 was 7.4 feet a day and for station 16, 11.3 feet a day. A very striking verification of the fact that the high movements here found were due to the escape of water from the pond will be noted when the temperatures of the waters in the wells of these stations are compared with the temperatures of the water in the pond and the water 78 RATE OF MOVEMENT OK UNDERGROUND WATERS. [no. 140. in wells outside of the influence of seepage from the pond. Practically all well water taken from wells on Long Island have temperatures Lying between 58 F. and 60 F. In the present case the temperature o\' water drawn from H. A. Russell's well. 22 feet deep, just west of Wantagh Pond, was 59 F. on August 8, 1903. The temperature of water from well I) of station IT, just cast and slightly below the pond, was 61.2° F. on August 11. L903. This well was 20 feet deep, the bottom being at the same depth as the wells of stations 13 and 16. The Gate house ^ Gate house Conduit 3 1.20 CO* 111 CE 100 Hi a. < .80 2 6 P.M. 8 10 < AUG-3- — AUG.-^ VELOCITY :(J) 96 FEET PER DAY 5(2) 6.9 FEET PER DAY. Fig. 53. — Diagram showing velocity and direction of the flow of underground water south of Wan- tagh Pond (station 13). Two velocities are shown, two different depths. The high velocity, 96 feet a day, is the highest yet determined. Seepage from the pond accounts for the high velocities, 96 and 6.9 feet a day, south. Ammeter chart for this station is shown in fig. 12. temperature of water in the pond varies more or less, especially the temperature of the surface layer. The temperature of the pond water on August 8, a cloudy day, was 7'2.o z F.,and on Jul}' 30, a sunny day. it was s,) F. The temperature of water from the wells of station 13 was 65.8° F. on July 30, and from the wells of station 16 on August 8 was 69.5° F. These high temperatures at stations 13 and 16 show that a large portion of the moving ground water must come directly from the pond, and the rate of motion is so great that the ground SLICHTER.] LONG ISLAND, NEW YORK. 79 water has not time to be reduced to the normal temperature of the ground. The velocity at station 17 was L0.6 feet a day in a direction 30 west of south. The temperature of the water was 61.5° F. The ground water at this point is probably not entirely free from the seepage water from the pond. The direction of flow, the velocity, and the temperature of the water all indicate, however, that a considerable w.iy TAGH POSD 0)1. 00 UJ cc UJ a. .80 2 VELOCITY: (7) 77 FEET PER DAY"; (2) 11.6 FEET PER DAY. Fig. 54.— Diagram showing velocity and direction of flow of underground water at Wantagh Pond (station 16 X). This station is near station 13, and the curve shows two distinct velocities in dif- ferent strata. Velocities, 77 a:ad 11.6 feet a day, S. 60° E. The stream jnst east of the station seems to deflect the direction of flow toward .tself. part of the water is the natural underflow, which at this point is diverted toward the lowland occupied by the streams below the pond. There can be no doubt but that the proposed infiltration gallery will intercept a large amount of seepage water from the pond, which at the present time runs entirely to waste. The amount of seepage in the first 16 feet in depth is probably somewhat less than 3 second-feet per 1,000 feet of length of cross section, or about 2,000,000 gallons per twenty-four hours. 80 BATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. At station 21, Located just above Wantagh Pond, the velocity at a depth of 17 feet was 21.3 feet a day in a direction 60 east of south. This station is near the west bank of the main brook that feeds the pond, and the greater portion of the ground-water at this point percolates into the bed of the stream. The true underflow at this point can be found by taking the southerly component of this velocity, which gives L0.6 feet a day. The temperature of the ground water at this point was 58 F. WANTAGH rOSD Z.2U 2.00 1.80 1.60 1.40 tO W 1.20 CL UJ O. 2 1.00 < .80 .60 .40 .20 « 10 A.M. 12 4 6 8 10 12 2 4 6 -AUG.-12 * AUG.-13- VELOCITY 10.6 FEET PER DAY. Fi .. 55.— Diagram showing velocity and direction of the flow of underground water at Wantagh Pond (station 17). Velocity, 10.6 feet a day. S. 30° W. The increase of underflow rate at the Wantagh Pond from 10.(3 feet a day above the pond to 96 and 77 feet a day below the pond, as com- pared with velocities above and below Agawam Pond, 2.6 and 5.3 feet a day, respectively, is easily understood when the material constitut- ing the bottom of the ponds is inspected. The material at Agawam is good, the soil being tine and compact, while at Wantagh the bottom of the pond is very sandy, in some places having a closer resemblance to a filter bed than to a puddled floor. SLICHTEB.] LONG ISLAND, NEW YORK. 81 INFLUENCE OF PUMPING UPON THE RATE OF MOTION OF (JKOUND WATERS NEAR SOME OF THE BROOKLYN DRIVEN -WELL STATIONS. Through the courtesy of Mr. I. de Verona an excellent opportunity was furnished the writer of making- some observations on the influence of pumping upon the normal rate of motion of ground waters in the neighborhood of some of the Brooklyn driven-well stations. For this special purpose the pumping stations at Agawam and Wantagh, which had been idle since December, 1902, were started up for two days each in August, 1903. Agawam was operated continuously from 7 a. m., Fifth telephone pole south of grist mill 2.40 2.20 CO UJ 0:1-20 UJ Q. *! 1.00 Elec trode / Casin g / / / / / ' / ' l^J7 ' M 2 ' 4 6 8 10 12 2 < AUG. 18 * » VELOCITY 21.3 FEET PER DAY. Fig. 56. — Diagram showing velocity and direction of the Aoav of underground water (station 21). Velocity. 21.3 feet a day, S. 50° E. August 19, to 7 a. m., August ^0. At the Agawam station observa- tions were made at station 5 bv means of the recording instrument. Well A was charged at 1 p. m.. August V.K or after nine hours of con- tinuous pumping. After this length of time it was expected that the maximum rate of flow of ground water would be established, although, of course, the cone of depression near the wells would still be chang- ing quite rapidly. Station 5 is 3<> feet north of the intersection of the chief suction mains communicating with the line of driven wells and 12 feet east of irk 140—05 6 82 RATE OF M<>Y i:\llN i OF UNDERGROUND WATERS I 140. the centra] discharge main (see fig. 58). The depth of the test we3s is 22 feet, while the depth of the 30 wells of the Agawam -tat ion system varies from 30 to L05 feet. The rate of pumping during the forty-eight-hour tesl was eery uni form, at an average rate of 2,250,000 gallons per twenty-four hours. The vacuum at the pump was maintained at l ; 4 inches, while that at the first well east of the engine-house was 23.2 inches. The charge of the centrifugal pump was dropped from 4 p. m. to 4.40 p m Fig. 57. — Map showing location of overflow stations, at which determinations of the rate of flow of underground water were made on Long Island. The Brooklyn driven-well pumping stations are located on the south side of the railroad and are named, from east to west, Massapequa, Wantagh, Matowa, Merrick, and Agawam. August 19, during which time the vacuum fell to 7 inches. This was the only interruption during the test. The velocity determined at station 5 during the test was 8 ft et a day in a direction S. 22 K. The normal velocity 'dt this station is 5.4 feet a day. S. 8 W., so that the influence of the pumping was to increase the velocity by 2.6 feet a day. or an increase of about 50 per cent (fig. 44). The actual velocity found and the percentage of increase are both very moderate, and indicate that the pumping station is not making an unreasonable draft upon the ground-water supply at this point. SLIGHTER.] LONG ISLAND, NEW YORK. 83 The 30 wells of the Agawam supply station have screens each 10 feet long, or, altogether about 730 square feet of screen. The maximum velocity of the ground water as it enters these screens must be 1,230 feet a day, since the actual pumpage was 2,250,000 gallons or 300,000 cubic feet per twenty-four hours. The mean velocity in the area (10 by 1,500 feet cross section) immediately drawn upon b} T the wells was about 30 feet a day. The reduction of this rate to 2.7 feet a day rep- resents a ratio of reduction of 11, which could be taken care of by a depth of 110 feet in the water-bearing gravels, without going outside of the 1,500 foot east and west line of the driven wells. To put this in another way: the daily pumpage of 300,000 cubic feet of water could be supplied b\ r the normal rate of motion of the ground Fig. 58.— Map showing location of stations 5 and 6 with reference to Agawam pumping station and East Meadow Brook Pond. The surface stream was gaged at the bridge marked A. The normal direction of ground-water motion at station 5 was S. 8° W. During a heavy rain, and also when the pumps were drawing water from the lines of driven wells, the direction of flow changed to S. 22° E., as shown by the arrows drawn from station 5. water at this point (5.4 feet a day) through a cross section of 510,000 square feet, or, say, 100 feet deep by 1 mile wide. To supply this amount of water, if removed from the ground on each of the three hundred and sixty- five days in a year would utilize 1 foot of rain- fall on 12 square miles of catchment area. These amounts are not excessive. The rate of removal of ground water at the Agawam station must therefore be regarded as exceedingly moderate. The observations at Wantagh pumping station were made on August 21 and 22. The pumping dt this station began at 7 a. m., August 21, and continued forty-eight hours at the uniform rate of 1,366,000 gallons per twenty-four hours. The water at this station is drawn S4 RATE OF MOVEMENT OF UNDERGROUND WATERS. I No. 110. from 48 driven wells, arranged on three lines of suction mains as shown in 6g. 59. The easl and west expanse of the two chief lines of wells is about L,500 feet. The wells of this station are of two different types, shallow wells of depth of about 24 feet and deeper wells, extending below an impervious bed to depths of from 60 to 112 feet. These lat- ter wells have an artesian head of 3 or 4 feet, and when the pumping plant is idle the water from the deep wells flows into the suction main and into the shallow wells, whence the water escapes into the sands and gravels of the upper zone of flow. An attempt was made on dune 24 to measure the rate of motion of the ground water at station 2, situ- ated 17 feet west of the chief discharge pipe and 300 feet north of the intersection of the main suction pipes from the driven wells, as shown in fig. 59. The attempted measurement was a failure, it not being FIG. 59. — Map showing location of stations 2, 13. 16, and 17 near Wantagh pumping station and Wan- tagh Pond. The arrows indicate the directfon of flow of ground water. The flow at station 2 was observed while pumps were drawing water from the three lines of driven wells. know^n at that time that the discharge from the numerous artesian wells wbs entering the surface layers of gravels and hence interfering with the normal flow in these gravels. The ground water at station 2 was, on account of this situation, either entirelv stationary or moving slightly toward the north. On August 21 well A of station 2 was charged at 6 p. m., or after eleven hours of continuous pumping from the driven wells. The velocity of the ground waters observed was at the rate of 6 feet a day in a direction 40° east of south. As this station is distant only 300 feet from the lines of driven wells, it is evident that the with- drawal of 4, 366, 000 gallons, or 582.000 cubic feet, per twenty-four hours has not an excessive influence on the normal rate of motion of the ground waters. The results at Wantagh compare very well witli slichter.] LONG ISLAND, NEW YORK. 85 the results at Agawam, and indicate that the driven-well plants have not exhausted the possibilities of ground-water developments. CONCLUSION. The very evident conclusion from observations on Long Island is that large amounts of ground water can still be obtained along the south shore of the island, especially if deep wells of large diameter can be successfully bored. The writer has already called attention to the possibility of constructing 12-inch wells of the California or "stovepipe" type in the unconsolidated material which extends from the surface to considerable depths on Long Island. Such wells, sev- eral hundred feet in depth, with perforations opposite the best water- bearing material, would utilize a large part of the underflow which now escapes to the sea. The practicability and success of such wells in this locality seems very probable, but the only way to arrive at an entirely satisfactory conclusion is to actually construct a test well. CHAPTER VII. TIIK SPECIFIC CAPACITY OF WEMjS. ( ; KNE B AL PRINC1 PLES. The amount of water discharged or obtained from a tubular well is a quantity which is as rigidly dependent upon certain definite and measurable factors as the total horsepower of a steam engine is depend- ent upon the elements in its design and the pressure of steam furnished to the engine. Very few persons realize, however, the closeness and intimacy of the dependence of the yield of a well upon the various causes represented by the character of the water-bearing material in which it is constructed and the size and shape of the well itself. In fact, the available published data containing the results of actual tests of the capacity of w^ells are usually incomplete in some important par- ticular, so that no laws or general principles are discernible even where they exist. With every well, no matter what its size or method of construction, there can be associated a perfectly definite quantity which expresses the capacity of that well to furnish water. In order to add definite ness to well construction and well data, such a quantity should be applied to every well whose capacity is measured. It can conven- iently be designated by the term " specific capacity." By "specific capacity" of a well is meant the amount of water furnished under a standard unit head, or the amount of water furnished under unit lower- ing of the surface of the water in the well by pumping. This number can be made definite by agreeing upon the unit of measure of quantity of water and on the unit in which the head is to be measured. If the unit of yield be the "second-foot," or cubic foot of water per second of time, and if the hydraulic head be measured in feet of water, then the specific capacity of an} T well is found by dividing the number of second-feet by the hydraulic head. For example, if an artesian well flows 2 second-feet, and if the static head in the well when the water is not permitted to flow is equivalent to a head of 20 feet of water, then the specific capacity of the well is 2 divided by 20, or 0.1 second- foot. We describe the specific capacit} 7 by saying that the specific capacity of the well is 0.1 second-foot. Likewise, if we desire to speak of the specific capacity of a common tubular well which is not artesian in character, we can proceed in a similar way. For example, if the well yields 2 second-feet w T hen the water in the well is lowered 20 86 blichteb.] SPECIFIC CAPACITY OF WELLS. 87 feet below its normal position, the specific capacity is found by divid- ing 2 second-feet by 20, giving a specific capacity of 0.1 second-foot. For the purpose of expressing the capacity of wells, the second-foot will be found to be a large unit of capacity, so that it will often be convenient to express the yield in gallons per minute, rather than in second -feet. One second-foot is equivalent to about 450 gallons per minute, so that the specific capacity of the wells above given might be stated as ''45 gallons per minute." Another convenient unit of measure for the capacity of the well is the miner's inch, the California miner's inch being one-fiftieth of a second-foot, and hence of a very convenient size for the measurement of well capacity. However, the different values of the miner's inch prevalent in various sections of the country make this unit of measure undesirable for general use. The importance of accurate knowledge of the specific capacit} T of the wells of a locality can not be overestimated. To the owner of a well it is very important that he know whether or not his well is better or poorer than neighboring wells, and whether the difference is due to a diversity in pumping machiner} r or to a difference in the well itself. To one who contemplates the construction of a well it is of the first importance that he know how much water he may expect to obtain, and in what manner it can best be obtained. In spite of striking examples of irregularity, it is usualh^ true that the same water-bearing material is very uniform in a given locality, and by properly designing a well one should be able to estimate in advance of construction the capacity of a well with a very small per cent of error. However, tests on existing wells and all data concerning them will have to be obtained and recorded with much greater accuracy and completeness than heretofore if this desirable result is to be realized. It does not count against the above statements concerning the ability^ to determine in advance the probable yield of a well, to find that neighboring wells, similarly constructed, jaeld very different amounts of water, or that water can not be obtained a short distance from a good well. Such a discovery always causes considerable comment, while the numerous cases in which ground water is found at very uniform depths and in nearly identical material call forth no comment whatever. The amount of water yielded by a common open well or by a non- flowing tubular well is dependent first of all upon the degree of fineness of the material in the various strata from which the water is obtained. The size of the soil grains not only controls the rate at which water can be transmitted to the well under a given head, but it also determines the proportion of contained water which the soil will freely part with. The fine-grained soils retain a considerable proportion of the water of saturation as capillary water even after free means of drainage are established, so that tine-grained material will not only deliver water SS RATE OF MOVEMENT OF UNDERGROUND WATERS. [no. 140. slowly, Imt will furnish only a small total amount. Some quicksand is so fine that the water can be pulled away from the fine grains with tin 1 greatest difficulty, while silt with a diameter of grain of about one one-thousandth of an inch (not at all an unusual size) will part with its water very slowly even when it is placed on a piece of blotting paper. The above factor in the specific capacity of the well can be expressed by means of the transmission constant, /\ of the material furnishing the water. Other things being equal, the yield of the well will vary directly with the transmission constant. Another cause effecting the yield of the well is the thickness of the water-bearing stratum. If the transmission constants of all water- bearing strata are the same, the amount of water available is directly dependent upon the thickness of strata penetrated, provided, of course, that only such material is counted as is in contact with a suitable well screen or strainer. An important factor in determining the yield o;f a well is the diam- eter of the well. By the diameter is meant the diameter of the well where it penetrates the water-bearing- stratum. The diameter of the well is a factor which determines the rate at which the water must move in the water-bearing material as it enters the well. A well Inning a large casing- will permit a given amount of water to enter under a low velocity, and hence with little friction in the pores of the water-bearing medium. The dependence of the } 7 ield upon the diam- eter of the well is not expressible in a very simple way. In fine mate- rial, the dependence of the yield upon the diameter of the well is very much less than is commonly supposed. Only in material that is veiy coarse is it usual that any great advantage is obtained by using casing as large as 16 to 24 inches in diameter. The friction of the water as it flows upward in the casing of a well, a factor which is often very large in the case of an artesian w T ell, is usually small or negligible in common tubular wells from which the water is pumped with a suction pipe much smaller than the diameter of the well itself. This statement must not be understood to imply that the amount of water discharged by the pump is not influenced by the size of the suction and discharge pipes. What is meant is that, with a given lowering of the water in the well the yield of water will not be dependent upon the friction in the casing to the upward-moving water, while of course the amount of power applied to the pump will be greatly influenced by the size of the suction and discharge pipe and upon the manner in which these pipes are installed. Finally, the specific capacity, if the well be not too shallow, varies directly as the distance the surface of the water in the well is lowered by pumping. Thus, if the water in a well is lowered 2 feet below the natural level by pumping from it at the rate of 20 gallons a minute, the same well may be expected to yield approximately 40 gallons a SLICHTEK.] SPECIFIC CAPACITY OF WELLS. 89 minute if the water is lowered -± feet below the natural level. For shallow wells the yield will not increase in this direct ratio, but will be considerably less on account of the decrease in percolating- surface due to the lowering of the water plane in the neighborhood of the well. Besides the advantages just mentioned, tubular wells, owing to their greater depth, are much more likely to strike a vein of coarse mate- rial, a small stratum of which may be expected to furnish much more water than a considerable depth of line material. This accounts for the well-known superiority of deep tubular wells over common dug wells. If a well be cased through the water-bearing medium, the character of the screen or perforations in the casing will of course influence the yield of the well. If a screen is clogged, or if the perforations are not ample, the capacity of the well will be cut down because of this imperfect casing. All of the factors named above influence the yield of a flowing arte- sian well, except that in place of the distance the water is lowered by pumping we must substitute the static head at the point of discharge of the flowing water. By the static head is meant the pressure when the well is closed at the point at which the flow is measured. This static head is conveniently expressed in terms of feet of water. For example, instead of giving the static head in pounds per square inch we can state it in feet of water. The flow of water from the porous medium into the well will vary directly as the static head, but the total yield of the well will not vary in this simple way on account of the frictional resistance which the water suffers in flowing through the casing and drill hole of the well. This last component of the specific capacity while usually small in a well that is pumped is often of the very first importance in the case of a flowing* well. To the friction in the casing and discharge pipe should be added the influence of all turns and bends and reductions in size and the like. This factor is often a very large one in the determination of the amount of water yielded by an artesian well. The resistance due to friction increases very greatly with a decrease in the size of pipe and also with an increase in the length of the pipe, and is materially influenced by the curves and variations' in size of the pipe and by the rivets and joints in the well casing or discharge pipes. The friction in pipes does not vary directly with the hydraulic head, but approximately as the square root of the head at which the flow takes place. As stated before, complete data concerning tubular wells are very difficult to obtain. Complete data concerning an artesian well should consist of the following: First, exact dimensions of all casing and sizes of the bore hole, including, of course, total depth; second, the static head of the well measured at a point a known distance above the sur- face of the ground; third, an accurate measurement of the amount of 90 BATE OF MOVEMENT OF UNDERGROUND WATERS. [No. 140. Wires to battery and ammeter -^ \Wires to similar floats below water yielded by the well when freely flowing under the measured static head; fourth, the thickness of the various water-bearing strata furnishing water to the well. The data for common tubular wells should include the following facts: First, the diameter of the well cas- ing; second, the depth to water: third, the depth of the well; fourth. the length of screen or perforations in the well; fifth, the character of the perforations; sixth, the amount of water obtained from the well under continuous pumping; seventh, the amount that the water in the well is lowered below its normal level dur- ing such pumping. From these facts the specific capacity of the well can he computed and many important facts can be determined. Additional data of considerable importance would he the following: Eighth, the distance the water i< raised by the pumps; ninth, the cost or expense of pumping. Complete estimate of the specific capacity of a tubular well can he made if the following data can be obtained: First, the amount that the water in the 4 well is lowered below the undisturbed water plane: second, the rate at which the water rises in the well after pump- ing cease-: third, diameter of casing and of suction pipe. The determina- tion of the rate at which the water rises in the well casing requires some special appliances, a stop watch being usually a necessity. Additional appa- ratus for this purpose has been con- structed and is shown in tig.