:: [ --y:':' ; :: ' : -; 
 
 
 
 '■■■ 
 
 ■ 
 
 r, 
 

 
 CORNELL 
 
 UNIVERSITY 
 
 LIBRARY 
 
 GIFT OF 
 
 C. E. College 
 
 ENGINEERING 
 
^uEfimmSP efficien cy tests of hydraulic 
 
 DATE DUE 
 
 
 Up|>_p 
 
 
 
 
 mtTTT 
 
 "i iwf 
 
 
 
 
 
 
 
 
 
 [ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 . . 
 
 
 
 
 GAYLORO 
 
 
 
 PRINTED IN U.S A. 
 
p| Cornell University 
 f Library 
 
 The original of this book is in 
 the Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924004033746 
 
MISSISSIPPI 
 
 «&. 
 
 MAY 23 1904 
 
 if 
 
 Reports ok EimeiJ^CTjFESTS 
 
 HYDRAULIC DREDGES 
 
 Presented at the 89th and 90th Sessions of the Mississippi 
 .River Commission and ordered printed. 
 
 [Being pages 136 to 167 of annual report of Mississippi River Commission for 1903; 
 with addition of records of later tests.] 
 
 CONTENTS: 
 
 Resolution. of Commission authorizing tests 
 
 Report op Asst. Engineer F. B. Maltby 
 
 Report of Prof. W. B. Gregory 
 
 Additional reports of Asst. Engineer F. B. Maltby. 
 
 plates. 
 
 Plates 1 to 47 opposite page 
 
 Plate 48 opposite page.. 
 
 Plate 49 opposite j'age 
 
 PAGE. 
 
 135 
 
 135-159 
 
 160-167 
 
 169-174 
 
 158 
 
 174 
 
 174 
 
re 
 m 
 
 .-h—*—¥ 
 i o J 
 
 -^i-3cT 
 
 f- ST^C-llC 
 
 Q^t e.f. Ce.\W 
 
EFFICIENCY TESTS OF HYDRAULIC DREDGES. 
 
 Office of the Secretary Mississippi River Commission, 
 
 St. Louis, Mo., January 23, 1904. 
 
 These tests were undertaken in compliance with the following resolution of 
 the Mississippi River Commission of November 13, 1899, and conducted in ac- 
 cordance with directions from Committee on Dredges and Dredging, dated July 
 24, 1902: 
 
 "That the Secretary be authorized to make such experiments as may seem to 
 be advisable at a cost not exceeding $10,000.00 with the object of increasing the 
 efficiency of the dredges operated by the Commission. These experiments not 
 to be made until after a project and an approximate estimate of cost have been 
 submitted to the Committee on Dredges and approved by the said Committee". 
 
 Wm. B. Ladue, 
 Captain, Corps of Engineers, V. 8. A., 
 Secretary Mississippi River Commission. 
 
 report of asst. engineer f. e. maltby on efficiency tests of hydraulic 
 
 dredges. 
 
 During the dredging season of 1902 the following-described tests of the effi- 
 ciency of the engines, boilers and sand pumps on the various dredges belonging 
 to the Mississippi River Commission were made. The tests were made in obedi- 
 ence to the following instructions, prepared by the committee on dredges and 
 dredging of the Commission : 
 
 1. Such tests shall be made as may be necessary to determine the efficiency 
 of boilers, engines, and sand pump of each of the dredges. The relative efficien- 
 cy of the several types of jet pumps, with due consideration of the results re- 
 quired in economical dredging work, should also be carefully determined. 
 
 2. As a basis for determining the mechanical efficiency of engines and pumps 
 under working load, it is necessary to first determine their frictional horsepower 
 when running at normal speed without load. 
 
 3. The pump tests shall be made by pumping water with the intake submerged 
 
 135 
 
136 
 
 to the normal depth and the pump running at normal speed, and also at known 
 speeds both higher and lower than the normal to ascertain the effect of varia- 
 tions in speed. 
 
 4. In order to ascertain as far as practicable the effect of the form of suction 
 head, tests shall be made both with and without the suction head, where these 
 are so attached as to be readily removable. 
 
 5. Each test shall embrace the determination of the indicated horsepower of 
 the engines, the number of revolutions of pump per minute, the velocity of flow 
 in suction and discharge pipe, the suction and discharge pressures. 
 
 6. In addition to the pressure gauges now in use, mercury manometers should 
 be attached to suction and discharge pipes near the pump for the accurate deter- 
 mination of suction and discharge pressures. 
 
 7. The velocity of flow in suction and discharge pipes shall be carefully meas- 
 ured and their determinations should be made at several points in the cross sec- 
 tion of the discharge pipe, so as to determine whether or not the whole of the 
 discharge section is effective under normal pumping conditions. This test can, 
 however, only be made when pumping sand ; but it would not interfere materi- 
 ally with the regular field work if done when the dredges are in operation. Pitot 
 tubes are recommended for use in making velocity observations. 
 
 8. The loss of head due to friction in discharge pipe shall be determined. It 
 is also desirable to carry this investigation further, if found practicable, so as to 
 include the effect of curved sections, rough joints, etc. 
 
 9. It is desirable to measure, as far as practicable, the relative efficiency of the 
 double and single intake to ascertain whether the flow of two columns of water 
 from opposite directions and meeting at the center of the pump tends to materi- 
 ally reduce the efficiency. 
 
 10. In conducting the above required investigations other lines of inquiry will 
 doubtless be suggested, and if they promise results of value they should be fol- 
 lowed up. 
 
 11. "When the required observations have been completed, they shall be care- 
 fully studied and compared with a view to determine the most efficient type of 
 engine and pump now in use and how the best of these could be improved upon 
 in future construction. 
 
 12. The results of the above investigations shall be embodied in a report giv- 
 ing in detail the type and form of boilers, engines and pumps examined and the 
 observations made in each case, with a summary showing from the results which 
 type or combination of types is the most efficient and best for the conditions met 
 with in the Mississippi River. 
 
 13. It is intended that the investigations and experiments called for above 
 will be made at such times as the dredges are not otherwise employed, as when 
 lying at the bank waiting for suitable stage of water or at the close of the coming 
 ing dredging season before being laid up for the winter. It is therefore desir- 
 able to have such preparations made in the way of instruments and measuring 
 appliances and attachments as may be deemed necessary before °*oin£ into the 
 field. 
 
 GENERAL DESCRIPTION OF TESTS. 
 
 It was thought desirable in making these tests to obtain actual working results 
 and to ascertain what was actually being done, rather than what it might be & possible 
 to do. With this idea in mind the tests were all made with the ordinary regular 
 
MISSISSIPPI RIVER COMMISSION. 137 
 
 crews working in the ordinary manner. No alterations or preparations were made 
 on the "boilers or machinery except those necessary for attaching the measuring 
 appliances. 
 
 All the plant was in good ordinary working condition, as none of it had been 
 -used to any great extent since the ordinary overhauling made during the lay-up 
 season. The tests were made at such times as the dredges were not otherwise 
 engaged and at such times as could be spared by the superintendent of dredging 
 operations from his other duties. 
 
 In general the same observers were used throughout all the tests as far as prac- 
 ticable; all weighing scales and gauges used were carefully tested; thermometers, 
 pyrometers, and calorimeters were of a high grade, new, and were not tested: indi- 
 cator springs were calibrated. 
 
 The information and results desired, as indicated in the above instructions, may 
 be divided under the following distinct heads and will be treated separately* 
 
 1. The efficiency and comparison of the boilers. 
 
 2. The efficiency and comparison of the main pumping engines. 
 
 3. The efficiency, capacity, and comparison of the sand pumps. 
 
 4. The effect of the form of the suction heads. 
 
 5. The efficiency and comparison of the jet pumps. 
 
 6. The loss of head in the discharge pipes due to friction and other causes. 
 
 7. The determination of the effective cross section of the discharge pipe when 
 pumping sand. 
 
 BOILER TESTS. 
 
 In making tests of the boilers no especial preparations were made for the tests. 
 The boilers and settings were all in good condition. The regular firemen on each 
 dredge were used and watches changed in the usual manner. The only instruc- 
 tions given the firemen were to fire as regularly as possible and not to shut the 
 draft doors, as the engineers would take care of any excess of steam generated. 
 In nearly all cases the pumping engines and auxiliary engines were run in the 
 ordinary manner in order that the demands on the boilers might be the same as 
 in ordinary service. It is believed that the efficiencies obtained show fairly well 
 the actual results obtained in regular operation. 
 
 METHODS. 
 
 The method of starting and stopping a test was as follows: The plant was oper- 
 ated till everything was running smoothly and regularly. The deck of the fire- 
 room was cleaned of coal, ash pits cleaned, fire examined by the superintendent, 
 the height of the water in the gauge glasses noted, and the word given to start 
 the test. During the last fifteen minutes of the test the fires were closely watched 
 by the superintendent and brought into as near the same condition as when starting 
 the test as possible. 
 
 COAL. 
 
 The coal was what is known as ' ' Pittsburg lump ' ' and was taken directly from 
 barges to the scales. No tests were made of the coal. It was all, however, prac- 
 tically " air dry." It was all carefully weighed on tested scales and delivered to 
 the firemen in lots sufficient to last approximately one hour. All ashes were 
 removed and weighed dry. 
 
 FEED WATER. 
 
 Water was pumped from the river into a cylindrical measuring tank about 3 
 feet in diameter and 4- feet high. It was emptied through a 6-inch gate valve near 
 the bottom. The vertical height drawn out each time was measured by means of 
 a largo wooden float, varnished to prevent absorption of water, bearing a staff grad- 
 uated to hundredths of a foot. This staff passed through a hole in a board across 
 the top of the tank. The tank was first filled, water supply shut off, and the posi- 
 tion or the staff read; the water was then drawn off, care being taken that the sur- 
 face of the water did not go below the top of the opening leading to the discharge 
 valve, the discharge valve closed, and the staff again read. The weight of the 
 water per foot, as indicated by the staff readings, was determined by drawing water 
 from the tank at various heights, observing the amount so drawn off on the staff, 
 and weighing it on standard scales. The mean of twenty observations was used 
 in computing the weight per foot. The measuring tank discharged into a suction 
 tank supplying the boiler feed pump. The temperature of the feed water was 
 measured with a high-grade thermometer at a point in the feed pipe just before 
 entering the boilers. The height of the water in theboilers was kept constant as 
 
138 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 nearly as possible and pains taken to have it exa6tly the same at the beginning and 
 end of the test. All blow-off valves and pipes and pipes from the feed pump were 
 examined and made tight. 
 
 CALORIMETER. 
 
 The quality of the steam was determined by means of a Carpenter throttling 
 calorimeter attached to the main steam pipe as close to the boiler as possible. 
 
 FLUE GASES. 
 
 The temperature of the flue gases was determined by two horizontal graphite 
 pyrometers furnished by A. S. Aloe Company. They were not calibrated. They 
 were inserted in the stacks just above the breeching. 
 
 The temperature of the feed water and readings of the calorimeter and pyrome- 
 ters were taken every half hour, and the mean readings used in computing results. 
 
 Grate and heating surfaces were obtained from careful measurements of each 
 boiler. 
 
 A table showing data and results obtained is appended hereto. 
 
 DISCUSSION OF RESULTS. 
 
 It is believed that the results on the Gamma, Delta, Epsilon, Zeta, and Kappa 
 show good average working conditions and are fair examples of the efficiencies of 
 the boilers of their class. 
 
 On the Iota considerable trouble was experienced in operating the engines, and 
 for this reason it was not possible to consume the steam regularly. On the Flad 
 the same difficulty was experienced, though to a less extent. The comparatively 
 low efficiency of the latter is believed to be due to the firing, which was by far the 
 poorest during any of the tests. The mean efficiency of the Mississippi River 
 boilers per pound of combustible is 9.10 pounds of water evaporated, as against 
 9.34 for the water-tube boiler on the Delta. The high efficiency of the boilers on 
 the Gamma is due, it is thought, to the relative large ratio of heating surface to 
 grate surface and the high rate of combustion per square foot of grate surface. 
 The 5-flue boilers do not seem to be more efficient than the 4-flue, and, in fact, 
 as between the Ejwlon and Zeta and the Iota and Kappa the advantage seems to 
 lie the other way. For ease of cleaning and danger of burning the points of supe- 
 riority all lie with the 4-flue boilers. The tests show that the Mississippi River 
 type of boiler, while possessing the advantages of simplicity, ease in cleaning and 
 repairing, is also an economical boiler in operation. 
 
 These tests also show incidentally the importance of a feed- water heater. The 
 Gamma is not supplied with a heater, water from the river being pumped directly 
 into the boilers. It will be noted that the correction applied to the quantity of 
 water evaporated into dry steam to bring the feed water to 212 degrees is nearly 19 
 per cent of the total. A feed-water heater using the p.xhaust, which now goes to 
 waste from the auxiliary engines, as on the Delta, Iota, Kappa, and Flad, assum- 
 ing that the same temperature of feed water would be secured, would reduce this 
 to a little over 9 per cent; or, in other words, the same amount of coal would 
 evaporate nearly 10 per cent more water. 
 
 Sketches showing the characteristic features of the boilers and settings are 
 appended. 
 
 EFFICIENCY OF MAIN PUMPING- ENGINES, 
 
 It was not thought necessary to determine the efficiency of the engines on all 
 the dredges, as many of them have the same general characteristics. All are 
 engines of standard manufacture, modern, high speed, compound, and all use 
 about the same steam pressures, the most important difference being that the 
 Gamma, Delta, Iota, Kappa, and Flad are equipped with condensing engines, 
 while the engines on the Epsilon. and Zeta are noncondensing. To determine the 
 relative efficiency of these two types, tests were made to determine the amount 
 of water used per indicated horsepower hour by the condensing engines on the 
 Kappa and the noncondensing engines on the Epsilon. 
 
 METHODS. 
 
 It was arranged that all the steam generated by the .boilers should be used by 
 the engines under test alone on the Epsilon and by the engines and condenser on 
 the Ka%ypa. Steam for operating the feed pumps and auxiliaries was furnished 
 
MISSISSIPPI KIVEB COMMISSION. 139 
 
 from an outside source. Especial care was taken to prevent leaks or to collect the 
 drips from any that existed. The feed water was measured in the same manner 
 as described under tests of boilers. The quality of steam was determined by the 
 same calorimeter, attached to the steam pipe just above the cylinders. Indicators 
 were attached to all four cylinders and cards taken simultaneously. Indicator 
 cards and calorimeter readings were taken every half hour and the means were 
 used in computing results. 
 
 DESCRIPTION OF ENGINES. 
 
 The engines on the Epsilon are ''Ideal " tandem compound, a pair on each end 
 of the main shaft of the pump. The cylinders are 16 inches and 26 inches in 
 diameter by 18 inches stroke; piston valves on high-pressure cylinders and bal- 
 anced slide valves on low-pressure cylinders. There is no receiver between the 
 cylinders, and the cylinders have no steam jacket, but are well lagged with min- 
 eral wool. They have automatic fly-wheel governors. The normal speed is from 
 173 to 176 revolutions per minute, and they developed a mean indicated power of 
 731.82 horsepower and consumed 26.06 pounds of water per indicated horsepower 
 hour. They were made by A. L. Ide & Sons, of Springfield, 111. 
 
 The engines on the Kappa were built by the Bucyrus Company, of South Mil- 
 waukee, Wis. They are also tandem compound, direct connected to the pump 
 shaft. The cylinders are 15| inches and 30 inches in diameter, 24 inches stroke. 
 High-pressure valves are of the piston type, and the low pressure balanced slide 
 valves. The cylinders have no steam jackets, but are well lagged. They are 
 supplied with automatic fly-wheel governors. The normal speed is about 130 
 revolutions per minute, and they develop a mean total horsepower of 852.01 and 
 consume 21.42 pounds of water per indicated horsepower hour. This steam con- 
 sumption includes the amount used by the condenser on the Kappa. 
 
 It is thought that the difference of 4.64 pounds of steam per hour per indicated 
 horsepower in favor of the engines on the Kappa represents in a fair way the econ- 
 omy of the condensing engines over the noncondensing of this size and type. The 
 boilertests of the Epsilon show an evaporation of 7.96 pounds of water per pound 
 of coal, or to evaporate the excess of 4.64 pounds of water per indicated horse- 
 power hour on the Epsilon requires , when developing 730 horsepower, 551 pounds of 
 coal per hour, or about 13 per cent of the total amount consumed by the dredge. 
 This, at present prices, costs nearly $19 per twenty-four hour day for coal only, 
 not including the cost of handling it. 
 
 A table showing in detail the data and results of this test is appended. 
 
 THE CAPACITY AND EFFICIENCY OF THE SAND PUMP. 
 
 In the design of a hydraulic dredge one of the principal features to be consid- 
 ered is the design of the sand pump, its efficiency and capacity, and it is evident 
 from the resolution of the committee that the determination of the efficiencies 
 and capacity of the various forms of pumps in use was considered the most impor- 
 tant of the tests to be made. On the seven dredges under consideration all the 
 pumps are of different design except on the Kappa and Flad, which are identical. 
 The Epsilon and Zeta were originally of the same design, but the pump on the 
 Epsilon has been supplied with a closed or shrouded runner, and on both the 
 Epsilon and Zeta the number of blades has been reduced from 7 to 5. The details 
 of each pump will be referred to later in the discussion of results. 
 
 METHODS — THE PITOT TUBE. 
 
 In determing the capacity of the pumps one of the first points to be decided 
 upon was the method to be employed in determining the velocity of flow through 
 the discharge or suction pipes. The instructions recommended the use of Pitot 
 tubes. After a careful study of the subject the writer believes that there is no 
 other simple, inexpensive, and practical method of measuring the high velocities 
 in the 32 and 34 inch discharge pipes in use with the dredges. The publication in 
 the Transactions of the American Society of Civil Engineers for April, 1902, of a 
 paper on experiments on the flow of water in pipes, and especially the discussion 
 of the original paper, was most timely. A large portion of the original paper and 
 its discussion by members of the society is devoted to the Pitot tube, and atten- 
 tion is invited to it for a complete discussion of its merit as an accurate instru- 
 ment for measuring the velocity of flowing water. 
 
 This accuracy seemed to be fully conceded, with tubes properly constructed, 
 and it remained to design a tube that could be inserted readily into the suction or 
 
140 REPORT OF THE CHIEF OF ENGINEERS, 17. S. ARMY. 
 
 discharge pipes that would be strong enough to stand the rough usage it would 
 meet and still show accurate results. All the tubes were made in the machine 
 shops connected with the dredges, as well as the gauges with which they were con- 
 nected. Nine tubes in all were constructed, but the inherent defects of the design 
 of all but Nos. 1, 3, 8, and 9 were soon" apparent and they were discarded. 
 
 DESCRIPTION OF TUBES. 
 
 A sketch of tubes Nos. 1, 3, and 8 is appended hereto. 
 
 Tube No. 1 consists of two pieces of brass tubing, one-fourth inch inside diam- 
 eter, inclosed in a piece of pipe 32£ inches long and turned to lyV inches outside 
 diameter. One of the small brass tubes is bent at its lower end, and below the 
 end of the outside inclosing pipe, to a right angle and the plane of the opening 
 made truly parallel with the upright pipe. This forms the impact opening or 
 point. The other brass tube is brazed onto a solid brass piece, circular in section, 
 and placed at right angles to the tube, the upstream end is turned down to a sharp 
 point and its end is even with the opening in the impact point and below it; this 
 brass point has a one-sixteenth inch hole drilled on each side, connecting with 
 the interior of the upright brass tube, and forms the static or pressure point. 
 The upper ends of the brass tubes have one-fourth inch air cocks, which are con- 
 nected by pieces of one-fourth inch rubber tubing, about 4 feet long, to the gauges. 
 
 The small brass tubes are held in the outside tube by lead or cement poured 
 around them. The outside tube slides through a stuffing box, which is screwed 
 into a hole tapped in the side of the suction or discharge pipe with a 1-J-inch pipe 
 tap. The top of the tube has a handle set carefully at right angles with the plane 
 of the impact opening. This arrangement permits the point of the tube to be 
 placed at any point along the diameter of the pipe in which velocity is being 
 measured and permits "traversing," or the determination of the velocity curve 
 across the pipe. 
 
 Tube No. 3 is very similar to No. 1, except that the impact point is below the 
 static point. 
 
 Tube No. 8 has the same sized vertical outside pipe to permit of the use of the 
 same stuffing boxes. At its lower end it is joined at right angles to a 1-inch pipe 
 about 18 inches long, drawn down at the point to the size of a one-fourth inch 
 brass tube, which is brazed into it and forms the impact point. This small brass 
 tube runs inside the large pipe to the top, where it is provided with a one-fourth 
 inch cock. The horizontal pipe has three one-sixteenth inch holes drilled in its 
 side for determining the static pressure, the interior of the pipe being connected 
 through the upright pipe and one handle to a one-fourth inch air cock. Tube No. 
 9 was made as nearly like No. 8 as could be done by a skilled machinist. 
 
 In the design of the tube it was thought that the fact that the impact point, 
 when formed, of a surface of revolutio n, wi ll convert velocity head into static 
 head, in accordance with the law v=%/2gh, had been thoroughly proven in the 
 discussion by Mr. "W. M.White of the paper above referred to and some experi- 
 ments made by him. (See Journal of the Association of Engineering Societies, 
 August, 1901.) This point being accepted, it only remains to show that the static 
 or pressure point of the tube used gives the true static pressure, or, in other words, 
 that there is no suction action due to the flow of water past the static openings. 
 If this is true, the tube will have a coefficient of unity in the formula v=C\/2gh> 
 where c is a coefficient and h is the observed difference of pressure shown by the 
 static and impact points of the tube. 
 
 The most conclusive method of demonstrating this fact would be to rate the 
 tube under the same conditions under which it was used, but the difficulties of 
 rating a tube in a 32-inch pipe having a velocity of flow of from 14 to 25 feet per 
 second and under a pressure of from 18 to 30 feet of water can readily be realized. 
 
 It is known that tubes 1 and 3 have a coefficient less than unity, but it is be- 
 lieved that the coefficient of tubes 8 and 9 is unity. The following observations 
 are presented in support of this belief. The tube was rated in open running water 
 having a velocity of about 3£ feet per second by comparing it with floats. About 
 80 floats were run and 600 observations were made with tube No. 9. In this case 
 the tube was suspended in the river with the point about 12 inches below the sur- 
 face and facing the current. The impact point and static point were attached to 
 opposite ends of an inverted U tube of glass, from which the air could be partly- 
 exhausted in order to bring the water surface in the tubes high enough to be read 
 
 The floats were 24 inches long, weighted to stand vertical in the water and 
 three-quarters or more of their length was submerged. The time required to 
 pass over a course 40 feet long was observed, the tube being about onepthird the 
 
MISSISSIPPI EIVEB COMMISSION. 
 
 141 
 
 distance from the upper end of the course. In the following table the velocity 
 given by the floats is in each case the mean of ten observations and that given 
 for the tube is the mean of about seventy-five observations. 
 
 
 Velocity 
 by float. 
 
 Velocity 
 by tube. 
 
 No.l --.- 
 
 3. 316 
 
 3.284 
 " 3.517 
 3.419 
 3. 405 
 3.212 
 3. 362 
 3.332 
 
 3.420 
 
 No. 2 
 
 3. 279 
 
 No.3_. 
 
 3.393 
 
 No. 4 
 
 3.358 
 
 No.5 . 
 
 3.502 
 
 No. 6 
 
 3.228 
 
 No.7.--' .__ 
 
 3. 354 
 
 No. 8 
 
 3.218 
 
 
 
 Mean _ -- _. ._ 
 
 3.356 
 
 3.343 
 
 
 
 The means show a coefficient of unity within about one-third of 1 per cent. 
 
 If tubes 8 and 9 have coefficients of unity they should both give the same 
 velocity in the same pipe at the same time. 
 ' To demonstrate this they were placed in the same discharge pipe, 50 feet apart, 
 and three sets of observations made; the position of the tubes was then reversed 
 and three sets of observations again taken, with results as follows. Each set is 
 the mean of ten readings. 
 
 Velocity by 
 tube No. 8. 
 
 Velocity by 
 tube No. 9. 
 
 25. 370 
 24.160 
 25. 770 
 a 25. 360 
 25. 370 
 25.550 
 
 24.500 
 25.600 
 25.140 
 "25.190 
 24. 840 
 25.020 
 
 Mean, 25.263 
 
 25. 048 
 
 a Reversed positions. 
 
 The means agree within about eight-tenths of 1 per cent. 
 
 That the accuracy of measurements by the tubes is not affected by pressure 
 was determined as follows: 
 
 Tube No. 8 was placed in the discharge pipe on the second ponton in the pipe 
 line of the dredge Epsilon, where the average static pressure was 17.50 inches of 
 mercury; tube No. 9 was placed in the discharge pipe on ponton No. 9, 350 feet 
 from tube No. 8 and where the average pressure was 3.64 inches of mercury. 
 Simultaneous readings were taken on gauges connected with each tube and at 
 different points across the pipe. 
 
 The mean velocity determined by tube No. 8 from 170 observations was 25.321 
 feet per second. The mean determined by tube No. 9 was 22.351 feet per second, 
 a difference of one-tenth of 1 per cent. 
 
 The pressure indicated by the static openings in tube No. 9 was compared with 
 piezometers in the sides of a 32-inch pipe as follows: 
 
 Four holes were drilled and tapped for a one-fourth inch pipe on the four 
 "quarters" of the 32-inch discharge pipe on the dredge Kappa, all being in the 
 same vertical section. Care was taken to have the axis of these holes normal to 
 the surface of the pipe. Ordinary one-fourth inch air cocks were screwed into 
 the holes, their ends projecting inside the pipe slightly. These were afterwards 
 filed down carefully to present a perfectly smooth surface flush with the inside of 
 the pipe. It was thought that if there was any inaccuracy in the pressures indi- 
 cated by the piezometers it would not be probable that the errors would be the 
 same on each one, or if the pressures indicated by each piezometer were the same 
 all would be correct. Each of the piezometers was in turn connected with every 
 other one through a differential gauge and in no case could any difference in pres- 
 sure be observed. 
 
 Tube No. 9 was then inserted into the same vertical section as the piezometers 
 and with the point of the tube at the center of the pipe. The static side was in 
 turn connected with each piezometer through a differential gauge and in no case 
 
142 BEPOKT OF THE CHIEF OF ENG-INEEKS, U. S. ARMY. 
 
 could any difference of pressure be observed. The pressure given by the four 
 piezometers and the static side of the tube was exactly the same. ,-^ OVOTlf 
 
 This experiment, with some others taken with the point of the tube at ditterent 
 points along the diameter of the pipe, show, incidentally, that the Pressuie is the 
 same throughout any section across a pipe and that the Bernomlli theory is not 
 true when applied to a section. .. i„4.„i„ + -u 0+ 
 
 It is conceded that the above observations do not demonstrate absolutely that 
 the coefficient of tubes 8 and 9 is unity, but in connection with the discussion at 
 the paper before referred to, it is believed that they furnish sufficient evidence to 
 warrant us in assuming that their coefficient is unity, and the observations made 
 with these tubes have been so used in reducing results. ^ 
 
 Tubes Nos. 1 and 3 were compared with tube No. 8 by placing them m the same 
 discharge pipe 50 feet apart. Three sets of observations of ten each were taken 
 and the position of the tubes in the pipe reversed and three sets more 'taken 
 The results' show the coefficient of tube 1 to be 0.930 and of tube 3 to be 0.8915, and 
 these values have been used in reducing observations made with these tubes. 
 
 A brief description of the other tubes experimented with and abandoned may 
 be of interest. A sketch is attached, showing the points only, the inclosing pipe, 
 stuffing box and attachments for connecting with gauges being m each case 
 similar to tubes 1 and 3. 
 
 Tube No. 2 was composed of an impact point similar to tubes 1 and 6. lhe 
 pressure point was a simple vertical brass tube with the lower end open and the 
 plane of the opening parallel with the current. This form of pressure opening 
 showed the greatest amount of suction action of any tube experimented with. 
 In fact, the suction was greater than the impact force on the impact point, as was 
 evidenced by the fact that when the impact point was turned to point down- 
 stream it still showed an apparent difference of pressure, due to impact, about half 
 as great as when pointing against the current. 
 
 Tube No. 4 was tube No. 2 with the lower end of the static point plugged up 
 and with openings on the side. 
 
 Tube No. 5 had impact and pressure points similar, but pointing in opposite 
 directions. 
 
 Tube No. 6 was very similar to No. 1, except that larger tubes were used, the 
 impact point being filled with a plug having a hole one-eighth inch in diameter. 
 
 The coefficient of these tubes, except No. 6, was not obtained, as at _ the time 
 they were made we were in search of a tube having a coefficient of unity and it 
 was quite evident that these tubes did not possess that qualification. 
 
 In comparing the various tubes with each other they were at first placed in the 
 discharge pipe 4 feet apart. It was apparent that the one lowest downstream 
 was affected by the one above it to a very considerable extent. The distance apart 
 was increased to 8 feet, with still an appreciable disturbance. The distance was 
 then increased to 50 feet, and no disturbance could be detected, and the coefficients 
 of tubes Nos. 1 and 3, as compared with tubes Nos. 8 and 9, were determined at 
 this distance, as has been described. 
 
 THE GAUGES. 
 
 Sketches show the two different styles of gauges used. 
 
 The differential gauge is a smple U-shaped glass tube, one-fcurth inch inside 
 diameter, provided with a connection at each end for attaching rubber tubes from 
 the Pitot tubes. The whole was securely fastened to a board, with raised edges 
 to protect the glass from injury. A paper scale, divided into inches and tenths, 
 was fastened to the board behind the tubes, and both the scale and board were 
 varnished over to prevent injury from water. The open gauge was also a simple 
 U-shaped glass tube with one leg longer than the other and with an attachment 
 for connecting with a rubber tube at one end only. The attachments for rubber 
 tubes were made of cast iron. The holes, where they were joined to the glass 
 tubes, were counterbored about three thirty-seconds of an inch larger than the 
 glass and this space filled with Portland cement, which made a most satisfactory 
 joint. The gauges were connected to piezometers or Pitot tubes with cloth-inserted 
 rubber tubing one-fourth inch inside diameter and not over 4 or 5 feet in length. 
 Considerable patience and ingenuity were required to remove all the air from the 
 tubes and the gauges above the mercury. It was found that a small amount of 
 air in the tubes would affect the readings of the gauges quite materially, The glass 
 tubes were not calibrated, but a careful examination revealed no appreciable dif- 
 ference in their size. Mercury was used in the gauges and the specific gravity was 
 assumed to be 13.5. It is believed that the possible errors due to slight differences 
 in the bore of the tubes or the specific gravity of the mercury are extremely small 
 *=rlicn the extreme fluctuations of the gauge readings ars noted, as will appear later. 
 
MISSISSIPPI RIVER COMMISSION. 143 
 
 METHODS. 
 
 The measurements for determining the capacity and efficiency of the pumps and 
 the drop in pressure in the pipe line were usually made at the same time. The 
 latter will be referrrd to later. 
 
 It was necessary to vary the mode of procedure somewhat in the different 
 dredges, but, generally speaking, a Pitot tube was inserted in one of the suction 
 pipes and in the discharge pipe. The tube in the discharge pipe was placed as 
 far from the pump as possible, to avoid the effect of the disturbance caused by it. 
 Owing to the fact that on none of the dredges is there a piece of pipe forming the 
 suction pipe which has any considerable length between elbows or bends of various 
 sorts it was difficult to so locate the tubes in the suction pipes as to be beyond the 
 effects of these bends. For this reason the velocities obtained in the suction pipes 
 have not been used in reducing results. A piezometer was introduced in the suc- 
 tion pipe usually within 4 or 5 fe^t of the pump flange. 
 
 Piezometers were introduced into the discharge pipes at various convenient 
 points along its length.. All piezometers were one-fourth-inch air cocks, screwed 
 into holes tapped in the horizontal center of the pipes, and were connected by 
 short rubber tubes to gauges, already described. 
 
 Drawings submitted herewith show the position of the pumps and arrange- 
 ment of pipes on each dredge; also the points where piezometers and Pitot tubes 
 were attached. 
 
 Pitot tubes were attached to differential gauges. Indicators were attached to 
 each cylinder of the engines. 
 
 "When all was ready and the engines running regularly, the gauges were read 
 until ten sets of readings had been taken. Curing this time the speed of the 
 pump, the steam pressure, and readings of the regular spring gauges attached to 
 the suction and' discharge pipes were observed and indicator cards taken. The 
 time required for completing a set of observations was from five to ten minutes. 
 
 These observations were made at various speeds of the pumps. The limiting 
 speeds were fixed on the one hand by a speed so low that experience in operation 
 had shown it to be too slow for accomplishing the most work; on the other hand, 
 the engines were run as fast as it was thought expedient to run them without risk 
 of injury, and in each case faster than it would be feasible to operate them regu- 
 larly. The mercury in the gauges connected with both the piezometers and Pitot 
 tubes and on both the suction and discharge pipes fluctuated very rapidly and 
 irregularly and through an interval extending sometimes over a space of 2 or 3 
 inches. 
 
 At first glance it seemed impossible to read the gauges with any degree of accuracy , 
 but by watching them closely for several moments it was seen that there was a 
 mean position near which the mercury came to rest at short intervals. In reading, 
 each observation was made the mean of the fluctuations as nearly as could be judged, 
 and as each side of the gauge was read from five to ten times for each observation, 
 it is thought that the mean is probably correct within one-tenth of an inch. 
 
 In reducing the readings of gauges connected with piezometers a correction 
 was applied, due to the height of the mercury on the pressure side of the gauge 
 above or below the center of the pipes, or in other words, all piezometer pressures 
 are reduced to the pressure at the center of the pipe. In the few cases where the 
 impact and static pressure points of the Pitot tubes were attached to separate open 
 gauges, the readings were also reduced to give pressures at the level of the point 
 of the tube. Where, as was usually the case, the Pitot tube was attached to a dif- 
 ferential gauge there was no correction to be applied on account of the position 
 of the gauge above or below the point of the tube, as the same correction would 
 apply to each side. There was a correction to be made due to the unbalanced col- 
 umn of water in one leg of the gauge of a height equal to the difference of the 
 heights of the mercury in the two legs of the gauge, or in other words, the differ- 
 ence of readings of the height of the mercury in the two legs of the gauge, which 
 was apparently the height due to the pressure being observed was reduced by the 
 weight of an equal amount of water. 
 
 The mechanical or indicated horsepower developed by the engines was obtained 
 from indicator cards taken on all the cylinders simultaneously in the usual man- 
 ner. In obtaining the total head against which the pump worked the suction head 
 was taken from piezometer readings in the suction pipe near the pump. Readings 
 were taken in one suction pipe of a double suction pump only, it being assumed 
 that the suction in either pipe would be the same. Some experiments made on the 
 Epsilon showed this to be practically true. 
 
 In determining the delivery head it was thought that a gauge placed close to the 
 pump might not give the true pressures, owing to the violent disturbance in tne 
 flow. For this reason piezometers were introduced in the straight pipe and some 
 
144 REPORT OF THE CHIEF OF ENGINEERS, XT. S. ARMY. 
 
 distance from the pump. The platted line showing the fall in pressure in the pipe 
 line referred to later on, was extended back to the flange of the pump, and this 
 pressure corrected for any actual lift or fall in the pipe between the pump ana 
 first gauge was used for the delivery head. m ,™ 
 
 If the areas of the suction pipe and discharge pipe, where measurements were 
 made, were the same, the sum of the suction and delivery heads, as indicated by 
 piezometers, would be the total head against which the pump was working. 
 
 This statement is contrary to the methods used, in computing the work done by 
 the pumps, in previous tests made of the dredges and published in the annual 
 reports of the Mississippi River Commission, and may require some explanation. 
 In these reports the method has been, in determining total head to take tne sum 
 of the suction and discharge heads plus the velocity head in the discharge pipe. 
 
 The total energy imparted to the water is expended, outside the pump itself, as 
 follows: Overcoming the entrance head, friction in the suction and discharge 
 pipes, actual lift, and creating velocity. It is generally known that on a discharge 
 pipe flowing under pressure a piezometer will give all the head except the veloc- 
 ity head. On a suction pipe, however, a piezometer gives the total head including 
 the velocity head. This fact is quite clearly shown by Mr. M. White m the Jour- 
 nal of the Association of Engineering Societies, October, 1900, where he proposed 
 to and did measure velocities by means of piezometers in the suction pipes. 
 
 Some measurements made in the suction pipe of the Delta also showed that this 
 assumption concerning a piezometer in the suction pipe was true. If these facts 
 concerning the piezometers are accepted, it is evident the sum of the piezometer 
 readings in the suction and discharge pipes gives the total head, including the 
 velocity head, and to add the velocity head to this sum, as was done in the com- 
 putations mentioned above, would be crediting the pump with velocity head twice. 
 
 On all the dredges except the Delta the area of the suction pipes is somewhat 
 greater than that of the discharge pipe, and the difference of the heads due to 
 velocity in the suction and discharge pipes has been added to the sum of the suc- 
 tion and discharge piezometer readings. On the Delta conditions are reversed, and 
 the difference in velocity heads was subtracted from the sum of the section and 
 delivery heads. In other words, total head, H, may be expressed 
 
 H=h D + — +h. 
 
 2g 2g 
 
 Where h represents head on the discharge pipe and v D the velocity at the same 
 point, and \\ and v represent head and velocity in the suction pipe. 
 
 THE VELOCITY CURVE. 
 
 All measurements of velocity for determining capacity and efficiency were taken 
 at the center of the pipes and it was necessary to determine the relation of center 
 to mean velocity. This was done by " traversing," or measuring the, velocity at 
 2-inch intervals across the diameter of the pipe. This was done on each of the 
 pipes and at the points where velocity measurements were made. All were made 
 on a vertical diameter except on the Iota , where traverses were made on both hori- 
 zontal and vertical diameters. 
 
 As all bends in the pipes in the vicinity of the point of application of the Pitot 
 tubes which would affect the form of the velocity curve are in a vertical plane, 
 it is thought that the vertical velocity curve will give the correct mean velocity. 
 
 The results of the traverses were platted on cross-section paper, and the mean 
 velocity of each curve obtained by dividing the pipe into 10 concentric rings of 
 equal area and taking the mean of the velocities at the centers of these rings. 
 These curves are appended hereto; each platted velocity is the mean of ten or more 
 observations. 
 
 The mean value of —-=0.8721 for all the dredges, and it was at first intended 
 
 to use this value in reducing observations on all dredges. The application of this 
 value to results gave efficiencies widely different in cases where there was every 
 reason to believe that approximately the same conditions existed, and on further 
 
 study it was noted that although there is a variation of the value of ^^ from 8008 
 
 ve 
 on the Delta to 0.9711 on the Gamma, the variation between individual determi- 
 nations on any one boat seldom varies more than 2 or 3 per cent, and it was decided 
 to use, for reducing center to mean velocities on each dredge, the ratio as deter- 
 mined on that individual dredge. 
 
MISSISSIPPI RIVER COMMISSION. 
 
 145 
 
 This value does not seem to vary in proportion with the velocity, but is prob- 
 ably determined by local conditions on each dredge, the proximity of bends, etc. 
 This is especially noticeable on the Flad, where traverses made in the discharge 
 pipe inside the dredge, about 110 feet from the pump and with a straight pip© 
 from the pump, with a mean velocity of 17.44 feet per second, gives a ratio of 
 0.788; while a traverse made in the pipe on the first ponton, about 40 feet beyond 
 the short reversed bends at the stern of the dredge, with a mean velocity of 14.80 
 feet per second, gives a ratio of 0.910. These traverses were made at different 
 times and at different speeds of the pump. 
 
 RESULTS. 
 
 A table is appended hereto, showing data and results on each dredge and at 
 various speeds. 
 
 The efficiency of the pump and engine as a whole is given. Efforts were made 
 to ascertain the mechanical efficiency of the engines by obtaining the horsepower 
 required to run them without load. The results obtained were very unsatisfac- 
 tory, owing to the difficulty of obtaining satisfactory cards. None of the auto- 
 matic governors would hold the engines down to a normal speed without load, 
 and to control the speed by throttling resulted in very poor cards. 
 
 From such cards as were obtained the frictional resistance varied from about 6 
 per cent on the Gamma to about 12 per cent on the Iota. It maybe assumed, 
 then, that the engine efficiences are from 90 per cent to 92 per cent, and the 
 column of pump efficiency in the table of results has been computed on a basis 
 of 92 per cent engine efficiency. 
 
 In general, the total head against which the pump was working has increased 
 about in proportion with the revolutions. The efficiency, compared with speed, 
 however, varies somewhat in different pumps. On the Gamma the variation in 
 efficiency between speeds of from 142 to 167 revolutions is about 6£ per cent. On 
 the Epsilon the efficiencies vary very little up to a speed of 170 revolutions, when 
 they fall off rapidly with an increase of speed beyond that point. On the Zeta 
 "the efficiencies increase up to a speed of 170, then they fall off. On the Iota there 
 is very little difference in efficiency over the range in speed covered; on the Kappa 
 the efficiency falls off regularly as the speed increases from the lowest; on the 
 Flad, however, at a speed only five revolutions less than on the Kappa the effi- 
 ciency is about 7£ per cent less, increasing to about the same efficiency as the 
 Kappa at the same speed, showing that the maximum efficiency on these dredges 
 is attained at a speed of about 120 to 125 revolutions. 
 
 Comparing the maximum efficiencies with the peripheral velocity on the differ- 
 ent pxxmps, the maximum on the Gamma is at a velocity of r>0.27 feet per second; 
 on the Epsilon, 50.58 feet; Zeta, 51.48; Iota, 54.98; Kappa said Flad 45 to 48 feet. 
 Apparently the open-runner pumps require a slightly greater peripheral velocity 
 to obtain the maximum efficiency than those with closed or shrouded runners, 
 but the difference is slight. Generally speaking, it would seem that a peripheral 
 velocity of about 48 to 50 feet per second will give the best results. 
 
 It must be borne in mind that these efficiencies were obtained while pumping 
 water, and that the deductions may not be true when pumping sand. As velocity 
 is the vehicle of transportation, it would seem that within certain limits the effi- 
 ciency as a sand pump would increase with the velocity of discharge. It is hoped 
 to procure additional data as to the proportion of sand carried at different velocities. 
 
 The following table shows the characteristic conditions accompanying the 
 highest efficiencies of each pump: 
 
 Dredge. 
 
 Gamma . 
 Delta.... 
 Epsilon . 
 
 Zeta 
 
 Iota 
 
 Kappa _. 
 Flad..,.. 
 
 Mean. 
 
 Efficiency 
 of pump 
 and en- 
 gine. 
 
 Per 
 
 cent. 
 71.0 
 nU.li 
 74. 1 
 73.4 
 G9.1 
 71.8 
 74.6 
 
 Total head P ^P^ ral 
 , f + f velocity 
 
 wlter? ( feet Per 
 second). 
 
 Sup. Eng. 1903 10 
 
 28.03 
 34.87 
 
 38.68 
 .50. 15 
 48. tfl 
 38. -12 
 48.01 
 
 40. £0 
 
 53.27 
 44.71 
 
 50. 53 
 51.43 
 54. 98 
 43.98 
 
 48.01 
 
 49.30 
 
146 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 RELATION OF DISCHARGE TO LENGTH OP DISCHARGE PIPE. 
 
 Especial attention is invited to the mean velocities on the Kappa, with 240 feet 
 of floating discharge pipe, and the Flad, which had 480 feet For the range of 
 speeds covered on each boat, from 115 to 133 revolutions the > mean total head 
 Varies from 42.05 feet on the Kappa to 41.65 feet on the Flad; while the mean 
 velocity, however, is 21.13 and 16.75 feet per second, respectively. With a short 
 discharge pipe a greater proportion of the work of the pump is converted into 
 velocity, and, as before stated, as velocity is the only means of transportation, as 
 short a discharge pipe as possible is desirable. 
 
 The same observations apply to the Epsilon, with 500 feet of floating pipe, and 
 the Zeta, with 1,000 feet, though as these pumps are not exactly alike the com- 
 parison can not be as rigidly made. 
 
 This point illustrates the advantages possessed by the discharge pipes belonging 
 to Ihe Iota, Kappa, and Flad, which are mounted on pontons and entirely above 
 the water surface and can be deflected to discharge on the side of the dredged cut, 
 while it is impossible to deflect the pipe lines belonging to the older dredges to 
 any great extent, owing to the fact that the pipes are partly submerged. It is 
 therefore necessary to use sufficient discharge pipe to carry the material into deep 
 water below the end of the cut. 
 
 EFFECT OF FORM OF SUCTION HEAD. 
 
 In order to ascertain what effect the suction head has on the efficiency of the 
 pump, it was removed on the Gamma and Epsilon and measurements taken. 
 Very little difference, if any, in results is noted, though on the Gamma the effi- 
 ciency seems to be somewhat higher. 
 
 On all the dredges the area of the openings in the suction heads is considerably 
 greater than the area of the suction pipes to which they are attached, and as they 
 contain no abrupt bends affecting the line of flow of the entering water it is not 
 thought that the form affects the velocity. 
 
 It is believed that a head built of such form and strength as to withstand the' 
 very severe usage to which it is subjected is much more important than one of a 
 theoretical form. It is believed, on the other hand, that the area of the openings 
 should not exceed the area of the suction pipe, in order that the same velocity 
 may be maintained throughout. 
 
 GENERAL COMPARISON OF PUMPS. 
 
 The table of results shows mean efficiencies of engines and pumps at what may 
 be called normal speeds, as follows: 
 
 Dr 
 
 Speed (rev- 
 olutions per 
 minute) . 
 
 Mean effi- 
 ciency. 
 
 Gamma 
 Delta . . 
 Epsilon 
 Zeta ... 
 
 Iota 
 
 Kappa . 
 Flacf... 
 
 142 to 167 
 122 to 124 
 
 158 to 171 
 
 159 to 171 
 165 to 168 
 120 to 133 
 120 to 131 
 
 Per cent, 
 66.6 
 56.1 
 73.2 
 71.2 
 68.0 
 69.4 
 71.8 
 
 Drawings submitted herewith show the general outlines of the pumps, with 
 their characteristic features and dimensions. 
 
 Nearly all the pumps differ in more than one particular, and it is difficult to 
 point with assurance to any one feature as being the one which affects the effi- 
 ciency to the greatest extent. 
 
 The shrouded or closed runner on the Epsilon has a slightly higher efficiency 
 than the Zeta, which has an open runner, but otherwise of the same form and 
 dimensions. The clearance on the Zeta's runners was very small, not more than 
 one-eighth to three-sixteenths of an inch. The Epsiloit's runner also had very lit- 
 tle clearance, not over one-eighth of an inch. The difference in efficiency is only 
 2 per cent, and this is lessthan the difference in efficiency of the Kappa and. Flad, 
 which are identically alike, of 2.4 per cent, so that very little additional efficiency 
 <;an be claimed on account of the shrouded runner. 
 
MISSISSIPPI RIVER COMMISSION. 147 
 
 The fact that the Epsilon' s efficiency is greater than that of the Kappa and Find, 
 while even the Zeta, with an open runner, equals them, seems to indicate that the 
 form of the vortex chamber and tapering runner blades on these pumps possess 
 some advantage. These runners also have narrower blades in proportion to their 
 diameter and also in proportion to the discharge than the Kappa B&d. Flad. 
 
 The principal characteristic of the Gamma, in distinction from the Epxilon, 
 Zeta, or Iota, is the long curved throat opening. In fact, the curve from the suc- 
 tion pipe to the tip of the runner is almost continuous. Compared with the Zeta, 
 which has an open runner of the same diameter, a peripheral speed approximately 
 the same, but with an efficiency of about 4£ per cent more, it does not appear that 
 this long throat opening possesses any advantage, and it is certainly much more 
 expensive to make and maintain. The runner on the Gamma has only four blades, 
 instead of five, as on the other pump, and this may also affect the efficiency, t.nd 
 probably does. 
 
 The Delta has the only single-suction sand pump in use on any of the dredges. 
 Its efficiency is so much lower than any of the others as to be quite marked. The 
 drawing of the pump shows the characteristics and dimensions. It will be noted 
 that, besides having the distinction of being the only single-suction pump, the 
 blades of the runners are considerably wider than the others. Which feature is 
 responsible for the low efficiency is hard to determine. It is possible that a higher 
 speed would have developed a greater efficiency, but it is noted that the peripheral 
 speed is as high as the most efficient speeds on the Kappa, although lower than 
 the average. The total head against which the pump was working is not greatly 
 different from a very efficient head on the Epsilon, though considerably below 
 the heads showing the greatest efficiency on all pumps except the Gamma. The 
 amount of water discharged per second is approximately the same as on the Iota, 
 which discharged also through the same length of discharge pipe. 
 
 Referring, then, to paragraph 9 of the instructions, we are forced to the conclu- 
 sion that the double-suction pump is the most efficient and that there are no indi- 
 cations that the flow of the two columns of water, entering from opposite sides, 
 tends to reduce the efficiency of the pump. It must be observed, however, that 
 owing to the construction of the runners these two columns of water do not meet 
 till their direction has been changed so that they are approximately parallel with 
 each other. 
 
 A table is submitted herewith showing the efficiencies of the pumps on the 
 Gamma, Delta, Epsilon, and Zeta, as determined in the capacity tests made in 
 1897 and published in the report of the Chief of Engineers for 1898. In computing 
 this table the mean of the measured quantities given in the above report have been 
 taken and reductions of work done, as determined by the total head, made as 
 explained above for the present report. 
 
 In this connection it should also be observed that the former report is in error, \» 
 inasmuch as inepmputing the weight per cubic foot of the material pumped it is; 
 assumed to weigh~B3| pounds, while in fact the mixture of sand and water weighed ; 
 nearly 75 pounds in some instances. The total head should also have been expressed " 
 in feet of material pumped instead of feet of water. However, as the weight of 
 the material pumped would be increased in the same proportion as the head would 
 be decreased, the results would remain the same. A comparison of this table 
 with the results of the later test shows the efficiency of the Gamma to be almost 
 exactly the same in each case. This is as it should be, as the pump and engines 
 on the Gamma have not been changed in any way except to keep them in repair. 
 It also indicates that the efficiency of a pump when handling sand or water is about 
 the same, and that we can assume that the results obtained in pumping water, in 
 the recent tests, can be applied to the pumps as sand dredges. The Delta shows the 
 lowest efficiency again, though about 8£ per cent higher than obtained in 1902. 
 The pump and engines are the same as originally made, and the only explanation 
 of the difference lies in the fact, as intimated above, that the speed in the later 
 tests was too low. 
 
 The runners on the Epsilon and Zeta were of the open pattern and had seven 
 blades in 1897 and were alike. Apparently the larger number of blades possesses 
 a higher efficiency than the present ones of five blades. 
 
 CONCLUSIONS. 
 
 From the foregoing we are led to submit the following conclusions: 
 
 That the most efficient speed for pumps of the size under consideration is one 
 
 giving a peripheral velocity of about 50 feet per second, and that the most efficient 
 
 head is 45 to 50 feet. 
 That the closed or shrouded runner, while being a necessity, to prevent abnormal 
 
148 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 wear of the pump casings and runner, and thus increasing the clearance and 
 decreasing the efficiency very rapidly, is also somewhat more efficient tnan an 
 open runner, even with a small clearance. 
 
 That the tapering.blades on the runner and curved vortex chambers are slightly 
 more efficient, but that the additional expense of construction and maintenance is 
 hardly warranted by the slight gain. . . 
 
 That the width of the blades on the Kappa and Flad is somewhat too great m 
 proportion to the size of the pump. . , 
 
 That the area of the openings in the suction head, the area of the suction and 
 discharge pipes should be approximately the same, in order that the velocity, wnic U 
 is the only vehicle of transportation, maybe maintained at a nearly constant rate. 
 
 EFFICIENCY OF JET PUMPS OB SAND AGITATORS. 
 
 On the dredges under consideration there are installed two distinct types of jet 
 pumps or agitators. On the Gamma, Epsilon, and Zeta there are single-suction 
 centrifugal pumps with discharge pipes of 17 or 18 inches inside diameter; they 
 deliver water to a pressure chamber just below the suction head provided with ten 
 to thirteen jet openings If to 2 inches in diameter. The pumps are operated by a 
 direct-connected vertical compound engine, condensing on the former and noncon- 
 densing on the two last-named dredges. In ordinary dredging operations in sand 
 the pressure on the discharge pipe is from 10 to 15 pounds per square inch. This 
 pressure is increased in material not readily loosened to 20 to 22 or 23 pounds, 
 which is about the limiting pressure possible. 
 
 The dredges Iota, Kappa, and Flad are equipped with compound, duplex, con- 
 densing steam pumps, with 16-inch water plungers. They deliver water to a 
 pressure chamber provided with 12 jet openings three-fourths of an inch in diam- 
 eter, and are ordinarily operated at a pressure of 40 to 60 pounds per square inch. 
 Their limit is about 80 or 90 pounds per square inch, and this pressure can not be 
 maintained for any considerable length of time. 
 
 The table shows the results of observations made. The velocity of discharge, 
 head, etc., for the centrifugal pumps was determined in the same manner as 
 described under the main sand pumps. They show a combined efficiency of engine 
 and pump of only about 41 per cent to 45 per cent, which seems quite low when 
 compared with the larger pumps. This low efficiency can probably be explained 
 by the fact that they have only a single suction, and there is therefore the friction 
 of a thrust bearing to overcome, and that there is probably a greater clearance 
 between runner and casing than on the large pumps. 
 
 In determining the efficiencies of the reciprocating pumps it was not possible 
 to measure the velocities of flow in the discharge pipe, as the gauges and connec- 
 tions with the Pitot tubes were not designed for such high pressures. 
 
 The indicated horsepower was determined by indicators in the ordinary man- 
 ner, but the discharge was computed from the displacement of the plungers and 
 no allowance made for slip or leakage in the valves. This is not a very satisfac- 
 tory determination, but it can readily be seen that even allowing for slip the 
 mechanical efficiency would be considerably higher than the centrifugal pumps. 
 It must be remembered, however, that the above, efficiencies are based on indi- 
 cated horsepower; and that the steam consumption with the reciprocating pump 
 would probably be about 33 pounds per indicated horsepower hour, while with 
 the compound engines we might expect at least as low as 24 pounds per indicated 
 horsepower hour 4 so that the efficiencies when referred to the coal pile would not 
 be so widely different. 
 
 The practical efficiency test would be the amount of work done in moving sand 
 per indicated horsepower, and it is hoped that some tests along this line can be 
 made later. 
 
 The writer believes the centrifugal jet pumps possess a very great advantage 
 over the reciprocating steam pumps from a practical operating point of view, even 
 should the latter show a greater steam economy for the same amount of work per- 
 formed, for the following reasons: 
 
 No instance is recalled in the writer's experience, as superintendent for two 
 dredging seasons, where a material was encountered by dredges equipped with 
 centrifugal pumps that they were not as well able to handle as one equipped with 
 the reciprocating pumps. It may be recalled that in 1898 the Zeta attempted 
 to dredge through blue mud at Cherokee during a capacity test, and that her 
 progress was very slow, the amount of material moved per hour being only about 
 600 yards. On the other hand, in 1902 the Kappa attempted to cut out a mud 
 lump at Silver Top Bar, and an advance of only 50 feet per hour could be made, 
 
MISSISSIPPI RIVER COMMISSION. 149 
 
 cutting less than 3 feet in depth and in a very strong current. The jets in this 
 material seemed to have no effect whatever. 
 
 The centrifugal jet pumps are extremely simple, contain no valves, and require 
 no repairs, and practically no attention. The centrifugal pump on the Gamma 
 ran for five seasons without even being taken apart for examination, and then the 
 runner and casing showed no wear, the scale not even being worn off. The total 
 repairs made consisted of a new brass bushing in the bearing of the shaft. 
 
 I think that the repairs of these pumps on the Epsilon and Zeta have not been 
 more elaborate. The care and repair of the engines connected with them is not 
 greater than that required for the steam end of the reciprocating pumps. TLe 
 latter pumps, on the other hand, are a constant source of labor and expense in care 
 and repair. It requires 120 feet of 1-inch square flax packing to pack the plungers 
 on each of the pumps on the Kappa, Flad, or Iota, and this packing must be 
 renewed at least once each season and more if a great deal of dredging is done. 
 
 The jet pumps on the Delta contain 2G0 hard-rubber valves, which required grind- 
 ing on each side three tim;>s during the season of 1902 , during which time five hun- 
 dred and thirty-two hours were consumed in dredging, or, in other words, the valve 
 surfaces required repair after a little less than four days' use. An entire set of 
 new valves will be required for the next season. This example is an extreme one, 
 though a large number of valves on the other pumps require renewal every year. 
 The pumps require constant attention and care, keeping packing tight, etc. 
 
 THE LOSS OF HEAD DUE TO FRICTION IN DISCHARGE PIPE. 
 
 An inspection of the table of results shows that the delivery head is a very large 
 •proportion of the total head against which the pumps were working, amounting, 
 on the Gamma, Delta, Epsilon, and Zeta, where discharge pipes are nearly hori- 
 zontal from the pumps, from about 55 per cent to GO per cent with 500 feet of float- 
 ing pipe, from 70 per cent to 75 per cent with 1 ,000 feet. On the Iota , Kappa , &n& 
 Flad this varies from about 60 per cent to 75 per cent, but on these dredges it 
 includes, besides the loss by friction, an absolute lift of about 5 feet at the gooseneck | 
 at the stern of the dredge. 
 
 The loss in head was determined in the following manner: Holes were drilled in 
 the suction and discharge pipes on the horizontal center and tapped for one-fourth 
 inch pipe thread, and into these holes were screwed ordinary commercial one-fourth 
 inch air cocks. The hole in these cocks at the screwed end is about one-fourth inch 
 in diameter and is reduced through the plug and at the outlet end to about three- 
 sixteenths of an inch. Short pieces of rubber tubing connected the cocks with 
 gauges, already described, which rested on top of the pipes or in other convenient 
 positions. Gauge readings were in each case corrected for the height of the gauge 
 above or below the piezometer to give the true pressures at the center of the pipe. 
 
 The positions of the piezometers along the pipe lines are shown in the accom- 
 panying sketches. 
 
 Circumstances were such that it was impossible in most cases to go inside the 
 pipes and ascertain if the ends of the cocks were exactly flush with the inside of 
 the pipe, but from the known thickness of the pipe and length of thread on the 
 shank of the cock they could be screwed in till they were approximately flush. 
 
 In connection with the comparison of the static pressure indicated by tube 9 and 
 four piezometers placed on a vertical section and filed smooth with the inside of 
 the pipe mentioned in the discussion of the Pitot tube four piezometers were 
 placed in a vertical section 5 feet downstream from the first set. These were put 
 in in the ordinary manner as described above, and each of the eight connected to 
 open gauges and readings taken at the same time. 
 
 The mean of twenty observations of gauges connected to the four piezometers 
 in the first set showed a pressure of 6.01, 6.05, 6.02. and 6.16 inches of mercury, 
 respectively, a mean of 6.06, while the pressures indicated by the piezometers in 
 the second set were 5.74, 5.35, 4.53, and 5.26 inches of mercury, respectively, a mean 
 of 5.22. Correcting this for the drop in pressure in 5 feet of pipe, which in this case 
 amounts to 0.25 inch of mercury, gives a means of 5.47, or 0.59 inch of mercury as 
 the error. This error will always be in one direction and probably very nearly 
 constant, so that relatively the indicated drop in pressure in a pipe line will not 
 be far from correct. 
 
 The effect on the determination of the total head against which the pumps work 
 would be to make it too small and the efficiency of the pumps too low. As the 
 delivery head has been obtained in each case, as already explained, by a combina- 
 tion of piezometers on the discharge pipe, it is thought that the error on this 
 account will not exceed 2 or 3 per cent. 
 
 It will be observed that the above observations on the first set of piezometers do 
 
150 BEPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 not agree as well as those mentioned in comparison with the Pitot tube where 
 each one was in turn connected with every other one through a differential gauge 
 and all gave exactly the same pressure. The explanation is probably to be found 
 in the fact thatthese latter measurements were made on open gauges and the vaii- 
 ations in pressure are due to errors in gauge readings due to extreme flu ^ ua ^ns 
 as mentioned before, and especially as efforts were made to read }F es ^^f 
 exactly the same instant, instead of, as was usually the .case reading ; the gauges 
 over a given period of time and attempting to get readings of mean fluctuations. 
 On the other hand, the differential gauge was very steady when attached to two 
 piezometers. The fluctuations hardly extended over more than one-tenth or an 
 inch, and three experienced observers were not able to detect any preponderence 
 of pressure in favor of one side over the other. ,, ,. 
 
 Diagrams are submitted herewith on which the observed pressures on the dis- 
 charge pipe at various points along its length have been platted for each dredge 
 and at various speeds. These diagrams also show in dotted lines how the delivery 
 head at the pump was obtained, as referred to under the efficiency ol the pumps. 
 
 The Gamma, Delta, and Zeta show a fairly uniform drop from the stern ot the 
 dredge along the pipe line. On the Gamma and Delta the drop from the last 
 gauge inside the dredge to the first one on the pipe line is not uniform with the 
 rest of the line owing to a change in diameter of pipe inside and outside the dredge. 
 The drop is also complicated inside the Delta by an absolute drop m the.elevation 
 of the pipe. On these dredges the pipe line was very nearly straight, and there 
 was very little leakage at the connecting joints. The drop in pressure was not 
 observed on the Epsilon, and the delivery head was obtained from observations on 
 one gauge and extending the line of drop back, as indicated on the Zeta. 
 
 The Iota, Kappa, and Flad show interesting indications of the effect of short, 
 abrupt bends on the loss of head. On each of these dredges there is a reversed 
 bend at the stern of the dredge, the effect of which is shown in the drop between 
 the last gauge on the dredge and the first one on the pontoons. This drop on the 
 Iota is from 9.5 to 12 inches of mercury. This amount may be partly accounted 
 for by an absolute lift of 5 feet, or 4.4 inches of mercury, and a drop of about 3.8 
 inches, due to the horizontal distance between gauges, leaving the loss of head due 
 to the bends of from 2.3 to 4.8 inches of mercury, or from about 10 per cent to 16 
 per cent of the total delivery head. 
 
 On the Kappa y with a shorter discharge pipe, the effect of the short bends is 
 more marked, the net loss being 7.1 inches, or 25 per cent of the total delivery head. 
 On the Flad this loss amounts to about 2.4 inches, or a little less than 10 per cent. 
 
 There is also shown the increase in the suction due to the reversed curves in the 
 suction pipe of the Delta. In this case the pipe on. each side of the curves was 
 about on the same level, as the bends are horizontal. The difference, then, of 
 about 7 inches, or more than 50 per cent of the total suction head, is nearly all due 
 to the bends. It does not seem to require any further argument as to the unde- 
 sirability of short bends in the pipes. 
 
 The upper line on the diagram of the loss of pressure on the Flad shows the 
 additional load caused by deflecting the pipes at each joint to the limit of their 
 throw — about 20° — indicating that the loss due to slight bends, all in the same 
 direction, is not great. 
 
 THE EFFECTIVE CROSS SECTION OF THE DISCHARGE PIPE AND PROPORTION OF SAND 
 
 PUMPED. 
 
 Paragraph 7 of the resolution of the committee instructed that the velocity of 
 discharge should be measured at several points in the cross section of the pipe, 
 while pumping sand, to determine the effective cross section. 
 i During the progress of the experiments already described we were not success- 
 f ful in determining the velocity of discharge while pumping sand, but were suc- 
 cessful in determining the effective cross section, and several determinations were 
 made of the proportion of sand carried. 
 
 The apparatus devised was as follows: A piece of 1-inch pipe, about 4 feet long, 
 was fitted to go through one of the stuffing boxes already described as used with 
 the Pitot tubes; the lower end of the pipe was bent to a right angle to face the 
 current; the upper end of the pipe was fitted with two elbows, so arranged as to 
 turn the opening downward; between the elbows was a gate valve for shutting 
 off the flow of water. The fittings on the top of the pipe formed a handle, by 
 means of which it could be manipulated. The stuffing box permitted the lower 
 end of the pipe to be placed at any point along the vertical diameter at the point 
 of application. 
 
MISSISSIPPI RIVER COMMISSION. 151 
 
 The lower end of the pipe being made to face the current and the valve being 
 opened; the water would flow into the lower end and out at the upper end, as the 
 vertical height of the pipe was much less than the head on the pipe at that point, 
 and it was assumed that this stream of water would carry the same proportion of 
 sand as was being carried in the discharge pipe at the point of the tube. 
 
 The water and sand issuing from the pipe were caught in iron buckets and tin cans 
 and the proportional part of sand was measured. No attempt was made to deter- 
 mine the sediment or mud in suspension, but only that portion of the material 
 which settled to the bottom of the receptacle at once or, say, within a minute. 
 These measurements were mad e at intervals of 2 inches along the diameter. Three 
 sets of traverses were made on the Delta while dredging at Peters Grossing and 
 eight on the Kappa in the chute of Presidents Island, and tables are submitted 
 herewith showing results. 
 
 The evidence that the whole cross section of the pipe is effective is quite con- 
 clusive, as in no case was there any evidence that any sand was resting on the bot- 
 tom of the pipe, as the small pipe was shoved down to the bottom in every set of 
 observations. 
 
 The indicated proportion of sand carried varies widely, and from the nature of 
 the work it must be true that the actual proportion of sand being carried varies 
 constantly. It can not be assumed that sand flows in a constant, uniform stream 
 into the suction head, and in fact we know very well that it does not do so. It 
 will therefore be necessary to take a large number of observations to determine 
 with accuracy what proportion of sand is carried. 
 
 The results of the three sets of observations on the Delta have been platted and \ 
 the mean of the eight sets also platted. Two of the sets on the Delta were taken 
 dredging downstream; the mean amount shows 14.4 per cent sand, while the 
 mean dredging upstream is 28.4 per cent sand. This is explained by the fact that 
 in dredging downstream the only power available for moving the dredge is the 
 force of the current. The dredge can not be forced into a bank of sand, as is pos- 
 sible in dredging upstream. The mean of all observations on the Kappa is 16.9 
 per cent of sand carried. The mean platted line of all results shows that the pro- 
 portion generally increases from the top toward the bottom; though the table 
 shows, in one instance at least, a greater proportion near the top than near the 
 bottom. The proportion varies from 3 per cent to 46 per cent. 
 
 These measurements were made experimentally and near the close of the season, , 
 when very little time was available. They are not submitted as conclusive evidence, j 
 and it is hoped that more measurements can be made and that at the same time 
 the velocity can be determined and thus determine the relation of velocity to sand- 
 bearing capacity. 
 
 It will be noted that on the diagrams showing the loss of pressure on the Iota, 
 Kappa, and Flad the point of zero pressure is at the last gauge, or about 30 feet 
 from the end of the pipe, instead of at the end of the pipe, as would be expected. 
 An explanation of this fact is not attempted by the writer at present, but that it 
 was a fact on all three dredges is beyond doubt, as the observations were tested 
 in various ways. 
 
 When the piezometer was connected to the ordinary mercury gauge it showed no 
 pressure nor suction; the mercury would occasionally fluctuate from 0.05 inch 
 above zero to an equal distance below. The mercury was then poured out of the 
 gauge and it was filled with water to the height of the piezometer, and it still 
 showed no pressure and no suction, though the water in the glass would occasion- 
 ally fluctuate one-half inch above or below the zero. 
 
 The gauges were then removed and the air cock unscrewed from the hole in the 
 pipe. This hole is one-half inch in diameter; but although it was on the horizon- 
 tal center of a 32-inch pipe flowing entirely full and at a mean velocity of from 
 14 to 21 feet per second, no water came out except occasionally a few drops, which 
 seemed to be caught on the downstream edge of the hole and thrown out. This 
 observation was made on each of the three dredges mentioned above. How far from 
 the end of the pipe the condition of no pressure extended was not determined. 
 The fact is only true when the pipes were flowing entirely full at the ends. When 
 on the Iota the speed of the pump was reduced to such an extent that the water 
 filled the discharge pipe at the end to within 4 or 5 inches of the top, then the 
 water in the gauge attached to the piezometer rose to approximately the height of 
 the water in the pipe, as in an open channel. 
 
 After the capacity tests had been completed the Commission secured the services 
 of Prof. W. B. Gregory, of Tulane University, to "examine the results obtained 
 and suggest if any further measurements were desirable. J ' No additional capacity 
 tests were made, but some of the observations in connection with the Pitot tube, 
 
152 
 
 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 mentioned above, were made at his suggestion. I am also indebted to him for 
 suggestions as to the reduction of final results. 
 
 Asst. Engineer William Gerig rendered valuable assistance in making observa- 
 tions, and the final reductions and computations were nearly all done by him. 
 Very respectfully submitted. 
 
 F. B. Maltby, 
 Assistant Engineer, Superintendent Dredging Operations. 
 Capt. Wm. B. Ladue, 
 
 Corps of Engineers, U. S. A., 
 
 Secretary Mississippi River Commission, 
 
 Table of percentage of sand handled by Delta, Peters Crossing, Noi'ember, 1902. 
 
 Distance from top of pipe. 
 
 1. 
 Dredging 
 upstream. 
 
 2. 
 
 Dredging 
 down- 
 stream. 
 
 3. 
 
 Dredging 
 down- 
 stream. 
 
 4 inches 
 
 Per cent. 
 0.143 
 
 Per cent. 
 0.047 
 .047 
 .048 
 .058 
 .089 
 .094 
 .094 
 .117 
 .100 
 .125 
 .187 
 .187 
 .187 
 .187 
 
 Per cent. 
 
 125 
 
 6inches __ .. 
 
 111 
 
 8 inches ._ 
 
 .177 
 
 117 
 
 10 inches _ 
 
 080 
 
 12 inches 
 
 .226 
 
 083 
 
 
 200 
 
 
 .276 
 .267 
 .243 
 .286 
 .304 
 .321 
 .358 
 .371 
 
 .175 
 
 .176 
 .141 
 .235 
 .187 
 .141 
 .219 
 .250 
 .250 
 
 18 inches . 
 
 20 inches 
 
 22 inches 
 
 24 inches - 
 
 
 
 30 inches 
 
 32inches . 
 
 
 .437 
 
 .250 
 
 
 
 Mean 
 
 .284 
 
 .121 
 
 .166 
 
 
 Average of columns 1, 2, and 3,^4^0 per cent. 
 
 Table showing percentage of sandhandled by Kappa, Presidents Island, November 
 
 28 and 29, 1902. 
 
 Distance from top of pipe.' 1. 
 
 Top -. _ 
 
 2 inches 
 
 4inches 
 
 6 inches 
 
 Sinches 
 
 10 inches 
 
 12 inches 
 
 14 inches 
 
 16 inches 
 
 18 inches 
 
 20 inches 
 
 22 inches .. -.. 
 
 24 inches 
 
 26 inches 
 
 28 inches 
 
 30 inches 
 
 Bottom, 32 inches. 
 
 Mean. 
 
 0.031 
 
 .281 
 . 255 
 .271 
 .207 
 .247 
 .050 
 .110 
 . 073 
 .214 
 .147 
 .131 
 .177 
 .130 
 .137 
 .171 
 
 0. 067 
 .193 
 .081) 
 .... 
 .118 
 . 12'.) 
 .147 
 . 135 
 .152 
 .125 
 .119 
 .138 
 .214 
 
 . m 
 
 . 339 
 .467 
 
 .161) 
 
 0.125 
 .140 
 
 .125 
 . 155 
 . 155 
 
 .188 
 188 
 155 
 250 
 18.S 
 
 1SS 
 
 ;M) 
 
 0.125 
 .155 
 . 105 
 
 .188 
 .220 
 . 220 
 .235 
 .205 
 .220 
 .313 
 .280 
 . 250 
 .225 
 .219 
 .250 
 .250 
 
 .186 I 
 
 .216 
 
 0. 0H8 
 .125 
 .125 
 . 1.50 
 . 125 
 .188 
 .214 
 .175 
 .171 
 . 168 
 . 168 
 .188 
 .179 
 .174 
 .1(38 
 .250 
 
 0.100 
 .100 
 . 125 
 .250 
 . 125 
 . 168 
 .168 
 .188 
 .205 
 . 250 
 . 1(18 
 . 125 
 . 125 
 . 168 
 . 125 
 .188 
 
 .168 
 
 .161 
 
 0.063 
 .063 
 
 . 083 
 .125 
 .125 
 
 .(183 
 .188 
 .167 
 . 083 
 .083 
 .125 
 .125 
 . 125 
 .125 
 .125 
 .250 
 
 .121 
 
 0.030 
 .063 
 
 .188 
 .188 
 .168 
 .125 
 
 .088 
 
 . 125 
 
 .168 
 .125 
 .168 
 .125 
 .188 
 .125 
 . 125 
 .250 
 
 .HI 
 
 Mean. 
 
 0.126 
 .133 
 .140 
 .172 
 .163 
 .1-16 
 .166 
 .155 
 .167 
 .174 
 .171 
 .185 
 .181 
 . 198 
 .204 
 .238 
 
 The material was ordinary channel sand, possibly slightly finer than the aver- 
 age, pumped through 240 feet of 32-inch discharge pipe 
 
 The measurements were made at the center of the second pontoon, or about 100 
 feet from the stern of the dredge. 
 
MISSISSIPPI RIVER COMMISSION. 
 
 153 
 
 Data and results of feed-water test of engines. 
 
 Description of main pumping engines. 
 
 Dredge Epsilon. Dredge Kappa 
 
 Date of and duration of test. 
 
 f Dec. 9, 1902, 
 \ 7 hours. 
 
 Dec. 17, 1902, 
 6 hours. 
 
 Engines, pair of tandem compound direct connected 
 
 Dimensions of engines: 
 
 Bore inches. 
 
 Stroke do... 
 
 Diameter piston rods do... 
 
 Diameter tail rods do . . . 
 
 Average clearance, per cent of volume displaced by pis- 
 ton 
 
 Noncondensing. 
 
 Condensing. 
 
 H.P. 
 
 16 
 18 
 
 L.P. 
 
 26 
 18 
 3J 
 
 H.P. 
 15J- 
 24 
 
 L.P. 
 
 30 
 24 
 
 3£ 
 
 Horsepower constant 0™ : 
 
 Crank end _ 
 
 Headend __ 
 
 Mean effective pressure starboard engine: 
 
 Crank _ _. 
 
 Head 
 
 Mean effective pressure port engine: 
 
 Crank 
 
 Head_ 
 
 Mean steam pressure at throttle valve 
 
 Mean vacuum on condenser _ _ 
 
 Mean revolutions per minute 
 
 Mean indicated horsepower starboard engines 
 
 Mean indicated horsepower port engines 
 
 Mean total indicated horsepower per hour 
 
 Total weight of water consumed, corrected for leaks 
 
 Percentage of water in steam. __ 
 
 Total quantity of dry steam consumed 
 
 Weight of dry steam used per hour 
 
 Weight of dry steam per indicated horsepower per hour 
 
 0. 
 
 0. 009139 
 
 71.04 
 70.78 
 
 79.99 
 
 79.82 
 128.4 
 
 0. 023756 
 0. 023910 
 
 15.32 
 15.00 
 
 16.63 
 16.58 
 
 172.7 
 221.12 
 249. 16 
 731 
 136,370.990 
 
 3rft 
 
 133.507.199 
 
 19,072.457 
 
 26.06 
 
 124. 83 
 136. 71 
 
 0.010576 
 0. 010889 
 
 61.60 
 57.31 
 
 86.41 
 150 
 
 21.7 
 129.8 
 166.86 
 237.34 
 
 0.1 
 0.042527 
 
 15.90 
 18.53 
 
 22.74 
 24.18 
 
 189.50 
 258.31 
 
 852.01 
 111,970.602 
 
 109,507.249 
 
 18,251.208 
 
 21.42 
 
 Note. — The above amount of steam used includes, on the Kappa, that used in the condenser, 
 which is a compound duplex noncondensing jet condenser. 
 
 Results of boiler tests, 1902 
 
 
 Dredge 
 Gamma. 
 
 Dredge Delta. 
 
 Dredge 
 Epsilon. 
 
 Dredge Zeta. 
 
 Date 
 
 Oct. 6, 1902. 
 
 12 hours. 
 
 3-5 Flue Miss. 
 
 61.75 
 
 1,845.5 
 
 1-29.8 
 
 135.2 
 
 73.7 
 573 
 1.2 
 
 18,651 
 
 642 
 18,009 
 1,554 
 
 25.16 
 
 3.3 
 
 163,044 
 161,087 
 191,693 
 
 15,974 
 
 10.27 
 
 10.64 
 
 Sept. 18,1902. 
 12 hours. 
 3 Heine. 
 
 107 
 6,120 
 1-57 
 148.6 
 
 188.3 
 788.6 
 1.05 
 
 29, 718* 
 
 1,6571 
 
 28,061 
 
 ,2,476.5 
 
 23.1 
 
 4.0 
 
 246,503 
 243,915 
 261,905 
 
 21.830 
 
 8.81 
 
 9.34 
 
 Oct. 7, 1902. 
 
 12 hours. 
 
 6-5 Flue Miss. 
 
 154 
 
 3,879 
 
 1-25 
 
 124.1 
 
 183.2 
 599 
 1.2 
 
 39,220 
 
 2,999 
 36,221 
 3,268.3 
 
 21.2 
 
 4.3 
 
 293,915 
 290,388 
 312,167 
 
 26,014 
 
 7.96 
 
 8.61 
 
 Oct. 21,1902. 
 
 Duration of trial 
 
 Type of boiler __ 
 
 6-5 Flue Miss. 
 
 Grate surf ace square feet.. 
 
 Water-heating surface do 
 
 Ratio of grate to heating surface . _ . 
 
 .AveragR st.pin.Tn prRpsnr« 
 
 154 
 3,879 
 1-25 
 133.4 
 
 Average temperature of feed water 
 entering boilers 
 
 191 
 
 Temperature of stack gases 
 
 667 
 
 Percentage of moisture in steam 
 
 Total amount of coal consumed, 
 pounds _ 
 
 1.4 
 
 40,414 
 
 2,650 
 37,764 
 3,367.8 
 
 Total amount of refuse, dry, 
 pounds _ _ _ 
 
 Total amount of combustible 
 
 Coal consumed per hour 
 
 Coal consumed per hour per square 
 foot, grate surf ace 
 
 21.8 
 
 Coal consumed per hour per horse- 
 power 
 
 4 1 
 
 Total amount of water pumped, into 
 boilers and apparently evapo- 
 rated, corrected for difference of 
 level, etc 
 
 319,865 
 315,387 
 336,833 
 
 28,069 
 
 Water actually evaporated, cor- 
 rected for quality or steam 
 
 Equivalent water evaporated into 
 dry steam from an d at 212° _ - 
 
 Equivalent water evaporated into 
 dry steam from ana at 212° per 
 hour 
 
 Equivalent water evaporated into 
 dry steam from and at 212° per 
 
 8.33 
 8.92 
 
 Equivalent water evaporated into 
 dry steam from and at 212° per 
 pound of combustible 
 
154 
 
 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 Remits of boiler tests, 1902 — Continued. 
 
 
 Dredge 
 G-amma. 
 
 Dredge Delta. 
 
 Dredge 
 Epsilon. 
 
 Dredge Zeta. 
 
 Equivalent water evaporated into 
 dry steam from ano at 212° per 
 square foot of heating surface per 
 hour. 
 
 8.65 
 
 258.70 
 
 463 
 
 3.56 
 204.01 
 632.7 
 
 6.70 
 
 168.92 
 
 754 
 
 7.24 
 
 Equivalent water evaporated into 
 dry steam from and at 212° per 
 square foot of grate surface per 
 hour __ 
 
 182.26 
 
 Commercial horsepower devel- 
 oped, "based on the evaporation of 
 3U pounds of water from and at 
 212° per hour 
 
 813.5 
 
 
 1 
 
 
 Dredge Iota. 
 
 Dredge Kappa. 
 
 Dredge Flad. 
 
 Date j. __ _ 
 
 Duration of trial. 
 
 Type of boiler _ 
 
 Grate surface square feet.. 
 
 Water heating surf ace do 
 
 Katio of grate to heating surface 
 
 Average steam pressure 
 
 Average temperature of feed water entering 
 
 boilers _ 
 
 Temperature of stack gases 
 
 Percentage of moisture in steam .__ _. 
 
 Total amount of coal consumed pounds.. 
 
 Total amount of refuse, dry do 
 
 Total amount of combustible ___ 
 
 Coal consumed per hour. _ ___ 
 
 Coal consumed per hour per square foot grate 
 
 surface.- _ _ 
 
 Coal consumed per hour per horsepower 
 
 Total amount of water pumped into boilers 
 
 and apparently evaporated, corrected for 
 
 difference of level, etc 
 
 Water actually evaporated, corrected for 
 
 quality of steam 
 
 Equivalent water evaporated into dry steam 
 
 from and at 212° __ __ _. 
 
 Equivalent water evaporated into dry steam 
 
 from and at 212° per hour 
 
 Equivalent water evaporated into dry steam 
 
 from and at 212° per pound of coal 
 
 Equivalent water evaporated into dry steam 
 
 from and at 212° per pound of combustible- 
 Equivalent water evaporated into dry steam 
 
 from and at 212° per square foot of heating 
 
 surface per hour 
 
 Equivalent water evaporated into dry steam" 
 
 from and at 212° per square foot of grate 
 
 surface per hour _.. _ 
 
 Commercial horsepower developed, "based on 
 
 the evaporation of 34J pounds of water from 
 
 and at 212° per hour... 
 
 Sept. 17, 1902. 
 
 9 hours. 
 
 7-4 Flue Miss. 
 
 117. 5 
 
 3,769 
 
 1-23.9 
 
 150 
 
 168 
 
 681 
 
 0.75 
 
 21,189 
 
 1,639 
 
 19.550 
 
 2,354.3 
 
 14.9 
 
 170,325 
 
 169,048 
 
 185,108 
 
 20,568 
 
 8.73 
 
 9.47 
 
 5.46 
 
 130.59 
 
 5. 96 
 
 Oct. 18, 1902. 
 
 12 hours. 
 
 7-4 Flue Miss. 
 
 157.5 
 
 3,830 
 
 1-24.3 
 
 161 
 
 161 
 
 502 
 
 1.45 
 
 35,535 
 
 1,177 
 
 34,358 
 
 2,877.9 
 
 18.2 
 
 288,224 
 284,045 
 313,444 
 26,120 
 8.82 
 9.12 
 
 6.78 
 
 165.84 
 
 7.57 
 
 Oct. 3, 1902. 
 
 9.083 hours. 
 
 7-4 Flue Miss. 
 
 157.5 
 
 3,830 
 
 1-24.3 
 
 141 
 
 171 
 678 
 2.5 
 
 29,183 
 826 
 
 28,357 
 3,212.9 
 
 20.4 
 4.5 
 
 209 253 
 
 204,022 
 
 222,668 
 
 24,504 
 
 7.63 
 
 7.85 
 
 6.40 
 
 155.58 
 
 710.2 
 
 TESTING DREDGES. 
 
 Data and results of main-pump tests, 1897. 
 
 
 u 
 
 CD 
 ft 
 
 m . 
 d$ 
 
 .21 
 
 
 CD 
 
 6 
 
 u 
 
 1 
 
 CD 
 U 
 ft 
 
 a 
 
 a 
 
 $ 
 
 02 
 
 to 
 
 CD 
 
 O 
 
 ga 
 
 CD 
 TO © 
 
 d 5? 
 
 ■- 1 -f. 
 
 u 
 
 i— < o 
 
 O 
 
 B 
 
 Suction pipes. 
 
 Discharge pipe, 
 
 Name of 
 dredge. 
 
 a 
 Sjj 
 
 CD CD 
 
 "to 
 
 eg 
 CD 
 
 u 
 
 11 
 
 cd o 
 
 f> CD 
 
 $ s 
 
 8o 
 
 O+s 
 
 |3 s 
 
 a 
 
 ■M * 
 
 1* 
 
 CD CD 
 'to 
 
 n 
 
 l-H 
 
 i 
 
 >> 
 o d 
 
 Pi 
 
 8 o 
 
 Gamma - 
 
 Delta 
 
 168.2 
 150.8 
 181.6 
 181.4 
 
 Lbs. 
 141.6 
 157.9 
 159.4 
 146.2 
 
 25.52 
 23.33 
 
 485. 37 
 1,039.6 
 748.92 
 686.94 
 
 Ins. 
 
 24| 
 33* 
 244 
 241 
 
 6 2 682 
 6.212 
 6.414 
 6.414 
 
 Feet. 
 
 9.115 
 15. 107 
 14.566 
 14.343 
 
 1.29 
 3.55 
 3.30 
 3.20 
 
 Ins. 
 34 
 
 324 
 32 
 
 6?806 
 
 6.492 
 5.673 
 
 5.585 
 
 Feet 
 9.660 
 14.455 
 16.468 
 16. 472 
 
 1.45 
 
 Epsilon 
 
 3.25 
 
 Zeta 
 
 4.22 
 
 
 4.29 
 
MISSISSIPPI RIVER COMMISSION. 
 
 155 
 
 testing dredges — continued. 
 Data and results of main-pump tests, 1S97 — Continued. 
 
 Name of 
 dredge. 
 
 Dis- 
 charge 
 per sec- 
 ond. 
 
 Dis- 
 charge 
 per sec- 
 ond. 
 
 Heads feet of water. 
 
 Work 
 done per 
 second. 
 
 Efficiency 
 of pump 
 and en- 
 gine. 
 
 Suction. 
 
 Deliv- 
 ery. 
 
 "Velocity. 
 
 Total. 
 
 Gamma 
 
 Delta ... 
 
 Cu. ft. 
 60.907 
 93.845 
 93.424 
 91.996 
 
 Pounds. 
 3,806.69 
 5,865.31 
 5,839.00 
 5,749.75 
 
 9.38 
 18.62 
 17.2 
 11.52 
 
 37.0 
 45.54 
 36.9 
 35.1 
 
 0.16 
 
 -0.30 
 
 0.92 
 
 1.02 
 
 46.54 
 63.86 
 55.02 
 47.64 
 
 Ft. lbs. 
 177,163.35 
 374,558.70 
 321,261.78 
 273,918.09 
 
 Per cent. 
 Cfi. 4 
 
 Epsilon 
 
 77 9 
 
 Zeta 
 
 7;> 5 
 
 
 
 Note. — These efficiencies are obtained by using the mean quantities measured, as given in 
 Report of Chief Engineers, 1898, and reducing results by same methods as used in tests of 1902. -w, 
 
 Data and results of main pump tests, 1902. 
 
 Y-^M 
 
 
 Revo- 
 lutions 
 per 
 min- 
 ute. 
 
 Steam 
 pres- 
 sure. 
 
 Vac- 
 uum 
 mer- 
 cury. 
 
 Total in- 
 dicated 
 horse- 
 power. 
 
 Suction pipes. 
 
 Discharge 
 pipe. 
 
 Name of dredge. 
 
 Inside 
 diame- 
 ter. 
 
 Area. 
 
 Mean 
 veloc- 
 ity per 
 
 Veloc- 
 ity 
 head of 
 
 In- 
 side 
 diam 
 
 Area. 
 
 
 
 
 
 
 
 second. 
 
 water. 
 
 eter. 
 
 
 
 
 Lbs. 
 
 Inches. 
 
 
 Inches. 
 
 Sq.feet. 
 
 Feet. 
 
 Feet. 
 
 In. 
 
 Sq.feet. 
 
 Gamma 
 
 142 
 154 
 
 138 
 147 
 
 22.0 
 22 
 
 297. 47 
 356.91 
 
 248 
 
 6.682 
 6.682 
 
 12. 982 
 13.680 
 
 2.62 
 2.90 
 
 34 
 34 
 
 6 305 
 
 
 6.305 
 
 
 160 
 
 139 
 
 22 
 
 407.93 
 
 24? 
 
 6.682 
 
 14. 395 
 
 3.22 
 
 34 
 
 6.305 
 
 
 167 
 
 145 
 
 22.5 
 
 457.61 
 
 24J 
 
 6.682 
 
 15.007 
 
 3.50 
 
 34 
 
 6.305 
 
 
 153 
 
 147 
 
 22.5 
 
 351. 74 
 
 241 
 
 6.682 
 
 13.600 
 
 2.87 
 
 34 
 
 6. 305 
 
 
 154 
 
 142 
 
 22.5 
 
 352. 94 
 
 245 
 
 6.682 
 
 13.870 
 
 2.99 
 
 34 
 
 6. 305 
 
 Delta 
 
 122 
 122 
 
 160 
 162 
 
 23 
 23 
 
 737.39 
 737. 39 
 
 33$ 
 
 33J 
 
 6.212 
 6.212 
 
 17_._452 
 16.965 
 
 4.73 
 
 4.47 
 
 34} 
 34.', 
 
 6 492 
 
 
 6.492 
 
 
 124 
 
 165 
 
 23 
 
 784.15 
 
 33f- 
 
 6.212 
 
 17.784 
 
 4.92 
 
 84* 
 
 ■ (J. 492 
 
 Epsilon 
 
 158 
 
 144 
 
 
 587.50 
 
 m 
 
 6.414 
 
 17.554 
 
 4.79 
 
 32} 
 
 5.673 
 
 
 165 
 
 139 
 
 
 680.70 
 
 24} 
 
 6.414 
 
 18.482 
 
 5.31 
 
 32} 
 
 5.673 
 
 
 166 
 
 141 
 
 
 687.92 
 
 24| 
 
 6.414 
 
 18.356 
 
 5.23 
 
 32} 
 
 5.673 
 
 
 168 
 
 123 
 
 
 712. 67 
 
 24} 
 
 6.414 
 
 18. 755 
 
 5.48 
 
 32} 
 
 5.673 
 
 
 171 
 
 140 
 
 
 765.97 
 
 24} 
 
 6.414 
 
 19.088 
 
 5. 66 
 
 32} 
 
 5.673 
 
 
 181 
 
 133 
 
 
 874.69 
 
 24} 
 
 6.414 
 
 19. 513 
 
 5.91 
 
 32} 
 
 5.673 
 
 
 181 
 
 133 
 
 
 867. 41 
 
 24} 
 
 6.414 
 
 19. 441 
 
 5.76 
 
 32} 
 
 5.673 
 
 Zeta 
 
 150 
 159 
 
 132 
 134 
 
 
 483. 85 
 571. 15 
 
 24} 
 243 
 
 6.414 
 6.414 
 
 12. 321 
 13.087 
 
 2.38 
 2.65 
 
 32 
 32 
 
 5 585 
 
 
 5.585 
 
 
 170 
 
 126 
 
 
 690. 40 
 
 24} 
 
 6.414 
 
 14.234 
 
 3.15 
 
 S3 
 
 5.585 
 
 
 171 
 
 125 
 
 
 712.06 
 
 24} 
 
 6.414 
 
 14.304 
 
 3.18 
 
 32 
 
 5.585 
 
 
 175 
 
 129 
 
 
 734.07 
 
 24} 
 
 6.414 
 
 14.347 
 
 3.20 
 
 32 
 
 5. 585 
 
 Iota 
 
 151 
 165 
 
 155 
 150 
 
 23 
 24.8 
 
 667. 69 
 817.83 
 
 24 
 24 
 
 6.283 
 6.283 
 
 16.016 
 16.622 
 
 3.99 
 4.29 
 
 2** 
 
 5.828 
 
 
 5.828 
 
 
 165 
 
 150 
 
 25 
 
 817.83 
 
 24 
 
 6.283 
 
 16.862 
 
 4.42 
 
 32H 
 
 5.828 
 
 
 165 
 
 150 
 
 25 
 
 817.83 
 
 24 
 
 6.283 
 
 16. 751 
 
 4.36 
 
 m 
 
 5.828 
 
 
 168 
 
 146 
 
 25.2 
 
 867.54 
 
 24 
 
 6.283 
 
 17.393 
 
 4.70 
 
 m 
 
 5.828 
 
 
 168 
 
 146 
 
 25.2 
 
 867. 54 
 
 24 
 
 6.283 
 
 17. 393 
 
 4.70 
 
 m 
 
 5.828 
 
 
 120 
 
 120 
 
 20 
 
 662. 03 
 
 24} 
 
 6.414 
 
 16.986 
 
 4.48 
 
 31 f 
 
 5.541 
 
 
 123 
 
 146 
 
 20.3 
 
 744. 80 
 
 24} 
 
 6.414 
 
 17. 893 
 
 4.98 
 
 31 fr 
 
 5.541 
 
 
 126 
 
 153 
 
 20 
 
 793. 49 
 
 24} 
 
 6.414 
 
 18.130 
 
 5.11 
 
 315 
 
 5.541 
 
 
 126 
 
 160 
 
 20 
 
 797. 19 
 
 24} 
 
 6.414 
 
 18. 114 
 
 5.10 
 
 HI J 
 
 5.541 
 
 
 127 
 
 165 
 
 19.2 
 
 817. 70 
 
 24} 
 
 6.414 
 
 18. 508 
 
 5.32 
 
 31 ,y 
 
 5.541 
 
 
 129 
 
 165 
 
 20 
 
 844.85 
 
 24} 
 
 6.414 
 
 18.552 
 
 5.35 
 
 HU 
 
 5.541 
 
 
 132 
 
 160 
 
 20 
 
 885.95 
 
 24} 
 
 6.414 
 
 18.808 
 
 5.50 
 
 31 1 
 
 5.541 
 
 
 133 
 
 165 
 
 20 
 
 926.34 
 
 24} 
 
 6.414 
 
 19.079 
 
 5.66 
 
 SH 
 
 5.541 
 
 Henry Flad 
 
 115 
 
 152 
 
 17 
 
 537. 35 
 
 24} 
 
 6.414 
 
 12. 869 
 
 2.57 
 
 32 
 
 5.585 
 
 
 115 
 
 139 
 
 15 
 
 521.53 
 
 24} 
 
 6.414 
 
 12.869 
 
 2.57 
 
 32 
 
 5.585 
 
 
 120 
 
 145 
 
 14.5 
 
 590.61 
 
 24} 
 
 6.414 
 
 14.586 
 
 3.31 
 
 32 
 
 5.585 
 
 
 126 
 
 123 
 
 20 
 
 634.98 
 
 24} 
 
 6.414 
 
 14. 796 
 
 3.40 
 
 32 
 
 5. 585 
 
 
 129 
 
 116 
 
 21 
 
 645 58 
 
 24} 
 
 6.414 
 
 14. 767 
 
 3.39 
 
 32 
 
 5.585 
 
 
 129 
 
 117 
 
 21 
 
 644.31 
 
 24} 
 
 6.414 
 
 14. 896 
 
 3.45 
 
 32 
 
 5.585 
 
 
 131 
 
 137 
 
 14 
 
 782.68 
 
 24} 
 
 6.414 
 
 16.665 
 
 4.32 
 
 32 
 
 5.585 
 
 
 128 
 
 137 
 
 18 
 
 701.29 
 
 24} 
 
 6.414 
 
 15. 249 
 
 3.62 
 
 32 
 
 5.585 
 
156 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 Data and results of main-pump tests, 190$— Continued. (_ 
 
 
 Discharge pipe. 
 
 
 Dis- 
 charge 
 per 
 
 Heads, feet of water. 
 
 Name of dredge. 
 
 Center 
 veloc- 
 
 Mean 
 veloc- 
 
 Veloc- 
 ity 
 head of 
 
 Dis- 
 charge 
 per 
 
 Suc- 
 
 Deliv- 
 
 Veloc- 
 
 Total. 
 
 
 ity per 
 
 ity per 
 
 second. 
 
 second. 
 
 tion. 
 
 ery. 
 
 ity. 
 
 
 
 second. 
 
 second. 
 
 water. 
 
 
 
 
 
 
 
 
 Feet. 
 
 Feet. 
 
 Feet. 
 
 Cu. feet. 
 
 Pounds. 
 
 
 
 
 
 G-amma 
 
 14.170 
 14.930 
 
 13.760 
 14.499 
 
 2.94 
 3.27 
 
 86. 757 
 91.416 
 
 5,422.31 
 5,713.50 
 
 7.09 
 7.90 
 
 12.54 
 13.90 
 
 0.32 
 .37 
 
 19.95 
 
 
 22.17 
 
 
 15. 710 
 
 15.256 
 
 3.62 
 
 96. 189 
 
 6,011.81 
 
 8.75 
 
 15.36 
 
 .40 
 
 24.51 
 
 
 16. 379 
 
 15.905 
 
 3.93 
 
 100.281 
 
 6,867.56 
 
 9.29 
 
 16.31 
 
 .43 
 
 26-03 
 
 
 14. 843 
 
 14. 414 
 
 3.23 
 
 90.880 
 
 5,680.00 
 
 8.24 
 
 13.80 
 
 .36 
 
 22.40 
 
 
 15.138 
 
 14.701 
 
 3.36 
 
 92.690 
 
 5,793.12 
 
 8.32 
 
 13.90 
 
 .37 
 
 22.59 
 
 Delta_ 
 
 20.854 
 
 20. 272 
 
 16.700 
 16.234 
 
 4.34 
 4.10 
 
 108.416 
 105. 391 
 
 6,776.00 
 6, 586. 94 
 
 18.86 
 
 19.72 
 
 15.32 
 
 15.52 
 
 - .39 
 
 - .37 
 
 33.79 
 
 
 34.87 
 
 
 21.250 
 
 17. 017 
 19.847 
 20.897 
 
 4.50 
 
 110.474 
 
 6,904.63 
 
 19.17 
 
 15.80 
 
 — .42 
 
 34.55 
 
 Epsilon 
 
 22.522 
 
 23. 715 
 
 6.13 
 6.78 
 
 112.592 
 118. 549 
 
 7,037.00 
 7,409.31 
 
 9.85 
 10.91 
 
 22.78 
 24.97 
 
 1.34 
 1.47 
 
 33.97 
 
 
 37.35 
 
 
 23.533 
 
 20.737 
 
 6.69 
 
 117.641 
 
 7,352.56 
 
 10.24 
 
 25.05 
 
 1.46 
 
 36.75 
 
 
 24.064 
 
 21.205 
 
 6.99 
 
 120. 296 
 
 7,518.50 
 
 11.50 
 
 25.65 
 
 1.53 
 
 38.68 
 
 
 24.492 
 
 21. 582 
 
 7.24 
 
 122.435 
 
 7,652.19 
 
 11.95 
 
 26.60 
 
 1.58 
 
 40.13 
 
 
 25.036 
 
 22.062 
 
 7.53 
 
 125.158 
 
 7,822.38 
 
 12.49 
 
 28.86 
 
 1.62 
 
 42.97 
 
 
 24.944 
 
 21.981 
 
 7.51 
 
 124.698 
 
 7,723.63 
 
 12.46 
 
 28.70 
 
 1.75 
 
 42.91 
 
 Zeta ... 
 
 17.443 
 18.500 
 
 14.150 
 15.007 
 
 3.08 
 3.50 
 
 79.028 
 83.814 
 
 4,939.25 
 5,238.38 
 
 5.85 
 6.47 
 
 29.81 
 33.75 
 
 .72 
 .85 
 
 36.38 
 
 
 41.07 
 
 
 20.150 
 
 16.347 
 
 4.15 
 
 91.298 
 
 5,706.12 
 
 7.72 
 
 39.60 
 
 1.00 
 
 48.32 
 
 
 20.250 
 
 16.428 
 
 4.19 
 
 91. 750 
 
 5,734.37 
 
 7.85 
 
 41.29 
 
 1.01 
 
 50.15 
 
 
 20.310 
 
 16.477 
 
 4.22 
 
 92.024 
 
 5,751.50 
 
 8.28 
 
 40.16 
 
 1.02 
 
 49.46 
 
 Iota 
 
 18.600 
 19.400 
 
 17. 181 
 17.920 
 
 4.59 
 4.99 
 
 100. 131 
 104.438 
 
 6,258.19 
 6,527.38 
 
 5.74 
 6.71 
 
 33.01 
 38.70 
 
 .60 
 .70 
 
 39.35 
 
 
 46.11 
 
 
 19. 680 
 
 18. 179 
 
 5.14 
 
 105.947 
 
 6,621.69 
 
 6.68 
 
 38.75 
 
 .72 
 
 46.15 
 
 
 19.550 
 
 18. 059 
 
 5.07 
 
 105. 248 
 
 6,578.00 
 
 6.77 
 
 38.75 
 
 .71 
 
 46.23 
 
 
 20.300 
 
 18. 751 
 
 5.47 
 
 109.281 
 
 6,830.06 
 
 6.79 
 
 40.30 
 
 .77 
 
 47.86 
 
 
 20.300 
 
 18. 751 
 
 5.47 
 
 109.281 
 
 6,830.06 
 
 7.22 
 
 40.30 
 
 .77 
 
 48.29 
 
 Kappa .. 
 
 22.436 
 23.634 
 
 19. 663 
 20. 713 
 
 6.01 
 6.67 
 
 108. 953 
 114. 771 
 
 6,8U9.56 
 7,173.19 
 
 10.70 
 11.22 
 
 26.19 
 28.07 
 
 1.53 
 1.69 
 
 38.42 
 
 
 40.98 
 
 
 23.947 
 
 20.987 
 
 6.85 
 
 116. 289 
 
 7,268.06 
 
 11.72 
 
 28.44 
 
 1.74 
 
 41.90 
 
 
 23.925 
 
 20.968 
 
 6.84 
 
 116. 184 
 
 7,261.50 
 
 11.72 
 
 28.44 
 
 1.74 
 
 41.90 
 
 
 24.446 
 
 21.424 
 
 7.13 
 
 118. 710 
 
 7,419.38 
 
 11.90 
 
 28.81 
 
 1.81 
 
 42.52 
 
 
 24.505 
 
 21. 476 
 
 7.17 
 
 118. 998 
 
 7,437.37 
 
 12.23 
 
 28.82 
 
 1.82 
 
 42.87 
 
 
 24.843 
 
 21.772 
 
 7.37 
 
 120. 639 
 
 7,539.94 
 
 12.75 
 
 29.05 
 
 1.87 
 
 43.67 
 
 
 25.200 
 
 22.085 
 
 7.55 
 
 122. 373 
 
 7,648.31 
 
 12.94 
 
 29.35 
 
 1.89 
 
 44.18 
 
 Henry Plad 
 
 18. 742 
 18. 742 
 
 14. 780 
 14.780 
 
 3.40 
 3.40 
 
 82. 546 
 82.546 
 
 5, 159. 13 
 5, 159. 13 
 
 9.19 
 9.19 
 
 26.91 
 26.91 
 
 .83 
 .83 
 
 36.93 
 
 
 36.93 
 
 
 21.243 
 
 16. 752 
 
 4.36 
 
 93.560 
 
 5,847.50 
 
 9.78 
 
 29.25 
 
 1.05 
 
 40.08 
 
 
 21.549 
 
 16. 993 
 
 4.49 
 
 94.906 
 
 5,931.63 
 
 10.78 
 
 30.32 
 
 1.09 
 
 42.19 
 
 
 21.505 
 
 16. 959 
 
 4.47 
 
 94. 716 
 
 5,919.75 
 
 11.30 
 
 29.87 
 
 1.08 
 
 42.25 
 
 
 21. 694 
 
 17. 108 
 
 4.55 
 
 95.548 
 
 5,971.75 
 
 11.30 
 
 30.24 
 
 1.10 
 
 42.64 
 
 
 24.270 
 
 19. 139 
 
 5.69 
 
 106.891 
 
 6,689.69 
 
 11.64 
 
 35.00 
 
 1.37 
 
 48.01 
 
 
 22.208 
 
 17. 513 
 
 4.77 
 
 97.810 
 
 6, 113. 13 
 
 11.14 
 
 31.95 
 
 1.15 
 
 44.24 
 
MISSISSIPPI RIVER COMMISSION. 
 
 157 
 
 Data and results of main pump tests, 1902 — Continued. 
 
 Name of dredge. 
 
 Gamma . 
 
 Delta. 
 
 Epsilon . 
 
 Zeta. 
 
 Iota , 
 
 Henry Flad_ 
 
 "Work 
 done per 
 second. 
 
 Foot lbs. 
 108,175.08 
 
 126,668.29 
 147,349.46 
 178,762.58 
 127,232.00 
 
 130,866.58 
 228,961.04 
 
 229,686.59 
 238,554.96 
 239,046.89 
 276,737.72 
 270,206.58 
 290,815.58 
 307,082.38 
 336,127.67 
 331,420.96 
 179,689.91 
 
 215,140.26 
 275,719.72 
 287,578.65 
 284,469.19 
 246,259.77 
 
 300,977.49 
 305,590.99 
 304,100.94 
 326,886.67 
 329,823.59 
 261,623.29 
 
 293,957.32 
 304,531.71 
 304,256.85 
 315, 472. 03 
 318,840.05 
 329,269.18 
 337,902.33 
 190,526.67 
 
 190,526.67 
 234,367.80 
 250,255.47 
 250,109.44 
 254,635.42 
 321, 172. 02 
 
 ! 270,444.87 
 
 Effi- 
 
 o 
 
 O 
 
 a © 
 
 
 ciency 
 
 >. • 
 
 
 t-i ft 
 
 
 of 
 pump 
 
 and 
 
 en- 
 gine. 
 
 o ft 
 
 a a 
 .2 5 
 2 ft 
 
 a 
 
 % © 
 + J pi 
 
 5 
 
 fto b 
 •r" o o 
 
 J^gh 
 
 Per ct. 
 
 P.ct. 
 
 In. 
 
 Feet. 
 
 
 66.1 
 
 71.4 
 
 m 
 
 42. 7519 
 
 0. 8379 
 
 64.5 
 
 69.7 
 
 69 
 
 46.3648 
 
 . 8145. 
 
 65.6 
 
 70.8 
 
 69 
 
 48.1712 
 
 .8242 
 
 71 
 
 76.7 
 
 69 
 
 50.2787 
 
 .8138 
 
 65.7 
 
 70.6 
 
 69 
 
 46. 0637 
 
 .8241 
 
 67.4 
 
 72.8 
 
 69 
 
 46. 3648 
 
 .8221 
 
 56.4 
 
 60.9 
 
 84 
 
 44. 7154 
 
 1.0436 
 
 56.6 
 
 61.1 
 
 84 
 
 44. 7154 
 
 1.0591 
 
 55.3 
 
 59.7 
 
 84 
 
 45. 4485 
 
 1.0372 
 
 73.9 
 
 79.8 
 
 69 
 
 47.5691 
 
 .9826 
 
 73.9 
 
 79.8 
 
 09 
 
 49.6766 
 
 .9866 
 
 71.4 
 
 77.1 
 
 69 
 
 49. 9776 
 
 .9728 
 
 74.1 
 
 80 
 
 69 
 
 50.5798 
 
 .9862 
 
 72.8 
 
 78.6 
 
 69 
 
 51.4830 
 
 .9869 
 
 69.8 
 
 75.4 
 
 69 
 
 54. 4937 
 
 .9647 
 
 69.4 
 
 74.9 
 
 69 
 
 54.4937 
 
 .9641 
 
 67.5 
 
 72.9 
 
 69 
 
 45. 1605 
 
 1.0711 
 
 68.4 
 
 73.9 
 
 69 
 
 47. 8700 
 
 1.0737 
 
 ,72.6 
 
 78.4 
 
 69 
 
 51. 1819 
 
 1.0892 
 
 73.4 
 
 79.3 
 
 69 
 
 51.4830- 
 
 1. 1032 
 
 70.4 
 
 76 
 
 69 
 
 52.6873 
 
 1.0705 
 
 67 
 
 72.4 
 
 75 
 
 49. 4147 
 
 1.0181 
 
 66.9 
 
 72.2 
 
 75 
 
 53.9963 
 
 1.0086 
 
 67.9 
 
 73.3 
 
 75 
 
 53.9963 
 
 1.0090 
 
 67.6 
 
 73 
 
 75 
 
 53. 9963 
 
 1.0099 
 
 68.5 
 
 74 
 
 75 
 
 54. 978 
 
 1.0092 
 
 69.1 
 
 74.6 
 
 75 
 
 54. 978 
 
 1.0137 
 
 71.8 
 
 77.5 
 
 84 
 
 43.9826 
 
 1.1302 
 
 71.7 
 
 77.4 
 
 84 
 
 45.0820 
 
 1.1388 
 
 69.7 
 
 75.3 
 
 84 
 
 46. 1815 
 
 1.1241 
 
 69.3 
 
 74.8 
 
 84 
 
 46.1815 
 
 1.1241 
 
 70.1 
 
 75.7 
 
 84 
 
 46.5480 
 
 1.1235 
 
 68.6 
 
 74.1 
 
 84 
 
 47. 2811 
 
 1.1106 
 
 67.5 
 
 72.9 
 
 84 
 
 48.3806 
 
 1.0954 
 
 66.3 
 
 71.6 
 
 84 
 
 48. 7472 
 
 1.0936 
 
 64.4 
 
 69.5 
 
 84 
 
 42. 1498 
 
 1.1563 
 
 66.4 
 
 71.7 
 
 84 
 
 42. 1498 
 
 1.1563 
 
 72.1 
 
 77.9 
 
 84 
 
 43. 9826 
 
 1. 1544 
 
 71.6 
 
 77.3 
 
 84 
 
 46. 1815 
 
 1.1280 
 
 70.4 
 
 76 
 
 84 
 
 47. 2811 
 
 1.1026 
 
 71.8 
 
 77.5 
 
 84 
 
 47. 2811 
 
 1.1076 
 
 74.6 
 
 80.6 
 
 84 
 
 48.0141 
 
 1. 1574 
 
 70.1 
 
 75.7 
 
 84 
 
 46,9146 
 
 1. 1370 
 
 rtf'ft 
 
 « 2 
 
 57.0 
 
 87.6 
 
 68.5 
 
 50 
 
 54 
 
 boo © 
 
 620.5 
 
 590 
 
 1,092 
 
 662 
 
 54 
 
 642 
 
 Remarks. 
 
 500-foot pon- 
 toons. 
 
 Suction head re- 
 moved. 
 Do. 
 
 500-foot pon- 
 toons. 
 
 Do. 
 
 1,000 -foot pon- 
 toons 
 
 500-foot pon 
 toons. 
 
 240-foot pon- 
 toons. 
 
 480-foot pon- 
 toons. 
 
 Pipes deflected 
 ateachi'ointto 
 full throw. 
 
 Note.— Efficiency of pumps alone are based on assumed efficiency of the engines of 92 per 
 cent. 
 
 Data and results of jet pump tests, 1902. 
 
 CENTRIFUGAL PUMPS. 
 
 
 © 
 ft 
 
 M 
 11 
 
 > 
 
 © 
 
 i 
 
 © 
 
 -i 
 
 ft 
 
 a 
 s 
 
 -4-1 
 
 CO 
 © 
 
 St 
 
 pa 
 
 o 
 
 > 
 
 © . 
 +- 1 pi 
 efi © 
 
 *& O 
 
 s a 
 
 •it © 
 CO 
 
 13 o 
 
 Heads feet of water. 
 
 u 
 o 
 -u 
 © 
 
 S3 a 
 
 S 
 "53 i 
 
 B 
 
 p— i 
 
 
 Name of dredge. 
 
 pi 
 
 S 
 
 Q 
 
 «} 
 
 0.92 
 
 1.10 
 
 1.40 
 
 .06 
 
 © 
 > 
 
 Q 
 
 >> 
 
 "8 
 ■o 
 "© 
 > 
 
 3 
 
 o 
 
 © 
 ft 
 'ft 
 
 «W 
 
 O 
 o3 
 Oi 
 
 r-4 
 
 Gamma 
 
 144 
 150 
 194 
 212 
 
 Lbs. 
 138 
 148 
 146 
 140 
 
 22.0 
 20.0 
 20.5 
 
 72.50 
 82.40 
 140.04 
 53.61 
 
 33.30 
 35.84 
 46.00 
 23.73 
 
 0.45 
 .48 
 
 .70 
 .38 
 
 34.67 
 37.42 
 48.10 
 24,17 
 
 Inches. 
 16^ 
 16J 
 1«| 
 17J 
 
 SQ./f. 
 
 1.553 
 
 Epsilon 
 
 1.553 
 1.553 
 
 1.718 
 
 
 
158 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 Data and results of jet pump tests, 1902 — Continued. 
 CENTRIFUGAL PUMPS— Continued. 
 
 Name of dredge. 
 
 Observed 
 center ve- 
 locity per 
 second. 
 
 o * 
 f-=l <J1 
 
 > © 
 
 
 rCl © 
 
 .2© 
 0^ 
 
 01 
 
 M <a 
 
 O QJ 
 
 © _ © 
 
 «H Pi© 
 
 o 
 
 11 
 
 3 
 
 h » 
 © >i . 
 
 ft 
 
 
 Gamma 
 
 Feet. 
 5.580 
 5.790 
 7.000 
 5.591 
 
 Feet. 
 5.368 
 5.570 
 6.734 
 4.926 
 
 Cu. ft. 
 8,337 
 8,650 
 
 10,458 
 8,463 
 
 Lbs. 
 521. 06 
 540.63 
 653.63 
 
 538.94 
 
 F*. Z&s 
 18,065.15 
 20,330.37 
 31,439.60 
 
 12,784.48 
 
 Pr. ct. 
 45.3 
 44.6 
 40.8 
 43.4 
 
 Ins. 
 69 
 69 
 69 
 49 
 
 Feet. 
 43. 3541 
 45.1605 
 58. 4076 
 45.3259 
 
 1.089 
 
 Epsilon 
 
 1.086 
 .952 
 .870 
 
 
 
 RECIPROCATING PUMPS. 
 
 Name of dredge. 
 
 Iotaei... 
 Kappa _ 
 
 Flad. 
 
 Revo- 
 
 
 Vac- 
 
 lutions 
 
 Steam 
 
 uum 
 
 per 
 
 pres- 
 
 inches, 
 
 min- 
 
 sure. 
 
 mer- 
 
 ute. 
 
 
 cury. 
 
 
 Lbs. 
 
 
 18 
 
 155 
 
 29 
 
 20 
 
 155 
 
 29 
 
 25* 
 
 161 
 
 18f 
 
 26 
 
 158 
 
 18* 
 
 29 
 
 452 
 
 18|- 
 
 40 
 
 139 
 
 184- 
 
 44 
 
 160 
 
 19 
 
 45 
 
 150 
 
 19 
 
 46 
 
 160 
 
 19 
 
 27 
 
 135 
 
 21* 
 
 27 
 
 135 
 
 21* 
 
 33 
 
 136 
 
 21* 
 
 34 
 
 138 
 
 12 
 
 38 
 
 135 
 
 2U 
 
 38 
 
 138 
 
 12 
 
 Total in- 
 dicated 
 horse 
 power. 
 
 41.24 
 
 54.23 
 
 32.48 
 
 35.53 
 
 61.48 
 
 102. 10 
 
 147. 19 
 
 156. 50 
 
 160. 50 
 
 42.64 
 
 48.66 
 
 80.92 
 
 94.94 
 
 107. 40 
 
 115. 46 
 
 Deliv- 
 ery 
 head. 
 
 Lbs. 
 55 
 
 ■■■: 
 
 20 
 
 Delivery 
 head 
 
 feet of 
 water. 
 
 126. 
 154. 
 
 46. 
 
 52. 
 108. 
 142. 
 184. 
 195. 
 211. 
 
 92. 
 
 :>•.*. 
 
 161. 
 161. 
 
 184. 
 184. 
 
 Dis- 
 tance 
 from 
 deliv- 
 ery 
 gauge 
 
 to 
 water. 
 
 Total 
 
 head 
 feet of 
 water. 
 
 134.72 
 162.37 
 54.08 
 60.99 
 116.29 
 150.85 
 192. 32 
 203.84 
 219.97 
 100.16 
 100,16 
 169.28 
 169.28 
 192.32 
 192.32 
 
 Name of dredge. 
 
 Diame- 
 ter of 
 plung- 
 er. 
 
 Iota 
 
 Kappa . 
 
 Flad. 
 
 Dis- 
 place- 
 ment 
 of both 
 plung- 
 ers per 
 stroke. 
 
 Inches. 
 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 16 
 
 "Water 
 
 dis- 
 placed 
 permin 
 
 ute. 
 
 Sq.ft. 
 5. 510 
 5.510 
 5.510 
 5.510 
 5.510 
 5. 510 
 5.510 
 5.510 
 5.510 
 5.510 
 5. 510 
 5. 510 
 5.510 
 5.510 
 5.510 
 
 Cu.feet 
 
 123. 975 
 137. 750 
 175.631 
 179. 075 
 199. 737 
 275. 500 
 303. 050 
 309. 938 
 316. 825 
 
 185. 976 
 185. 976 
 227. 304 
 234. 192 
 261.744 
 261. 744 
 
 Water 
 displaced 
 per min- 
 ute. 
 
 Pounds. 
 7, 748. 44 
 8,609.38 
 10,976.94 
 11, 192. 19 
 12,483.56 
 17,218.75 
 18,940.62 
 19,371.12 
 19,801.56 
 11,623.50 
 11,623.50 
 14,206.50 
 14,637.00 
 16,359.00 
 16,359.00 
 
 "Work done 
 per minute. 
 
 Ft. pounds. 
 1,043,869.84 
 1,397,905.03 
 593, 632. 92 
 682,611.67 
 1,451,713.19 
 2,597,448.44 
 3,642,660.04 
 3,948,609.10 
 4,355,749.15 
 1,164,209.76 
 1,164,209.76 
 2,404,876.32 
 2,477,751.36 
 3,146,162.88 
 3, 146, 162. 88 
 
 Work 
 done per 
 second. 
 
 Ft. lbs. 
 17,397.83 
 23,298.42 
 
 9,893.88 
 11,376.86 
 24,195.22 
 43, 290. 81 
 60,711.00 
 65,810.15 
 72,595.97 
 19, 403. 50 
 19,403.50 
 40,081.27 
 41,295.86 
 52,436.05 
 52, 436. 05 
 
 Efficien- 
 cy pump 
 and en- 
 gine. 
 
 Pet- 
 
 cent. 
 76.7 
 
 77.9 
 55.4 
 58.2 
 71.5 
 77.1 
 75. 
 76. 5 
 82.2 
 82.7 
 72.5 
 90.0 
 79.1 
 88.8 
 82.6 
 
 « Work done computed from displacement of plungers; no allowance for slip made. 
 
MISSISSIPPI KIVER COMMISSION. 
 
 159 
 
 LETTER OF ASST. ENGINEER F. B. MALTBY. 
 
 Memphis, Tenn., June 17, 1903. 
 
 Captain: As the Commission at its coming meeting may be interested in know- 
 ing the results of my recent tests with a 7-bladed runner, made on the Zeta, I 
 have the honor to inclose herewith a table of results in the same form as that sub- 
 mitted with my report on testing. 
 
 As far as possible the same conditions were obtained in both tests. The same 
 instruments for measuring were used and the same amount of discharge pipe was 
 attached. 
 
 The mean efficiency with the 5-bladed runner at a speed of 170 to 175 revolutions 
 is 72.1 per cent, while the mean efficiency with the 7-bladed runner at a speed of 
 168 to 175 revolutions is 73.1 per cent — no very great difference. 
 
 The diameter at the outside tip of the blades and at the inside shoulder was the 
 same in each case. The amount of clearance, less than one-eighth inch, was also 
 about the same. 
 
 The total amount of work done at the same speed is about 1 per cent less with 
 the 7-bladed runner than with the 5, though with an efficiency of about 1 per cent 
 higher. 
 
 Respectfully submitted. 
 
 F. B. Maltby, Assistant Engineer. 
 
 Capt. Wm. B. Ladue, 
 
 Corps of Engineers, U. S. A., Secretary Mississippi River Commission, 
 
 Data and results— Dredge Zeta with 7-blade runner, 
 
 [May 22, 1903. 1,000 feet of discharge pipe. J^=0.8099 used. ] 
 
 Vc 
 
 
 
 
 Steam 
 pressure. 
 
 Total in- 
 
 
 Suction pipes (2) 
 
 
 
 Discharge pipe. 
 
 
 Revolu- 
 tions per 
 minute. 
 
 dicated 
 horse- 
 power. 
 
 Inside 
 diame- 
 ter. 
 
 Area. 
 
 Mean 
 velocity 
 
 per 
 second. 
 
 Veloc- 
 ity 
 head of 
 water. 
 
 Inside 
 diame- 
 ter. 
 
 Area. 
 
 Center 
 velocity 
 
 per 
 second. 
 
 11.25A-.- 
 
 j Pounds. 
 180 135 
 180 135 
 180 135 
 168 | 125 
 170 I 133 
 162 132 
 161 132 
 180 | 137 
 175 j 128 
 
 
 Inches. 
 
 m 
 
 24£ 
 
 m 
 
 24} 
 24} 
 24} 
 24} 
 24* 
 24} 
 
 Sq.feet. 
 6.414 
 6.414 
 6.414 
 6.414 
 6.414 
 6.414 
 6.414 
 6.414 
 6.414 
 
 Feet. 
 
 Feet. 
 
 Inches. 
 32 
 32 
 32 
 32 
 32 
 32 
 32 
 32 
 32 
 
 Sq.feet. 
 5.585 
 5.585 
 5.585 
 5.585 
 5. 585 
 5.585 
 5.585 
 5. 585 
 5.585 
 
 Feet. 
 
 11. 38 A... 
 11.55 A... 
 
 1.15P... 
 
 1.22P... 
 
 1.39P... 
 
 764.52 
 770. 15 
 639.97 
 665.83 
 574.26 
 569.33 
 754.83 
 703.53 
 
 15.235 
 15. 643 
 14. 769 
 14. 854 
 
 3.61 
 3.80 
 3.39 
 3.43 
 
 21. 603 
 22.182 
 20.942 
 21.063 
 
 1.45P... 
 1.57P... 
 2.06P... 
 
 14.017 
 15.294 
 15.093 
 
 3.05 
 3.64 
 3.54 
 
 19.875 
 21.687 
 21.402 
 
 
 Discharge 
 pipe. 
 
 Dis- 
 charge 
 
 per 
 second. 
 
 Dis- 
 charge 
 
 per 
 second. 
 
 Heads, feet of water. 
 
 Work 
 done per 
 second. 
 
 Effi- 
 ciency 
 
 of 
 pump 
 and 
 en- 
 gine. 
 
 Di- 
 ame- 
 ter of 
 run- 
 ner. 
 
 In. 
 69 
 69 
 69 
 69 
 69 
 
 10 
 
 Pe- 
 
 
 Mean 
 veloc- 
 ity per 
 second. 
 
 Ve- 
 loc- 
 ity 
 head 
 of 
 wa- 
 ter. 
 
 Suc- 
 tion. 
 
 Deliv- 
 ery. 
 
 Ve- 
 loc- 
 ity. 
 
 Total. 
 
 riph- 
 eral 
 
 veloc- 
 ity per 
 sec- 
 ond. 
 
 11.25 A 
 
 Feet. 
 
 Feet. 
 
 Cub. ft. 
 
 Pounds. 
 
 
 
 
 
 Foot-lbs. 
 
 Pr. ct. 
 
 Feet. 
 54 192 
 
 11.38 A.. . 
 11.55 A... 
 
 1.15P... 
 
 1.22P... 
 
 1 39 P 
 
 17. 496 
 17. 965 
 16.961 
 17. ('59 
 
 4.76 
 5.02 
 4.47 
 4. 52 
 
 97.715 
 100. 334 
 94. 727 
 95. 275 
 
 6, 107. 19 
 6,270.87 
 5,920.44 
 5,954.69 
 
 6.90 
 7.13 
 6.13 
 6.28 
 
 40.73 
 41.83 
 37.35 
 38.22 
 
 1.15 
 1.22 
 1.08 
 1.09 
 
 48.78 
 50.18 
 44. 56 
 45.59 
 
 297, 908. 1 
 314,672.4 
 263,814 g 
 271,474.3 
 
 70. 85 
 74.2? 
 75. 01 
 74.14 
 
 54. 192 
 54. 192 
 50. 579 
 51. 1819 
 
 48 7728 
 
 1.45P..- 
 1.57P... 
 2.06P... 
 
 16. 1)97 1 4.03 
 17.5' ; 4 i 4 80 
 17.333 j 4.67 
 
 89. 902 
 98.095 
 96.805 
 
 5, HIS. 88 
 6, 130. 94 
 6,050.31 
 
 5.73 
 G K6 
 G. 55 
 
 34. 46 
 41 . 03 
 39.17 
 
 0.98 
 1.16 
 1.13 
 
 41.17 
 
 49. 05 
 40.86 
 
 231,329.3 
 
 30(1.721.2 
 283,517.5 
 
 73. *7 
 70. 79 
 73. 27 
 
 
 39 
 19 
 
 J9 
 
 47. 4115 
 54. 192 
 52.566 
 
160 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 Appendix 1 G. 
 report of prop. "w. b. gregory on efficiency tests of hydraulic dredges. 
 
 Tulane University of Louisiana, 
 
 New Orleans, La. , May 18, 1903. 
 
 Captain: In accordance with instructions conveyed in a letter, under date of 
 February 24, 1903, from Capt. G. P. Howell, I hereby beg leave to submit the 
 following report: 
 
 The general outline of dredge tests required by the committee on dredges and 
 dredging, under resolution of Commission, is given in the report of Mr. F. B. 
 Maltby, United States assistant engineer, superintendent of dredging operations. 
 In the main they embody the suggestions conveyed by the writer in a letter to 
 Maj. B. M. Harrod, under date of April 15, 1902, but contain many other instruc- 
 tions as to details of work. 
 
 The writer can not refrain from expressing his admiration for the thorough, 
 capable, and painstaking way in which these instructions were carried out, the 
 tests made and reported. 
 
 Boiler and engine tests were not taken up in the letter, as it was understood 
 that the tests of main pumps were considered of the greatest importance, and as 
 methods of conducting tests of boilers and engines were well known, standard 
 rules for such tests having been recommended by the American Society of 
 Mechanical Engineers. 
 
 Early in November, Capt. G. P. Howell, Corps of Engineers, U. S. Army, then 
 secretary of the Mississippi River Commission, and Major Harrod called on the' 
 writer at Tulane University and the former engaged his services as consulting 
 engineer in these tests, informing him that many tests had already been made 
 and asking him to visit the dredging fleet at its winter quarters near Memphis to 
 examine data, confer with Mr. Maltby as to the desirability of further tests, and 
 offer such suggestions as circumstances might demand. 
 
 According to instructions, writer left New Orleans on the evening of November 
 19, 1902, arriving at Memphis the next day. He returned to New Orleans 
 November 26. 
 
 During this first visit much data were examined and ways were suggested to 
 obtain constants for the various Pitot tubes and to establish what was believed to 
 be a fact, namely, that tubes Nos. 8 and 9 had a constant equal to unity in the 
 equation v—C\/2gh. 
 
 Again, on December 6, writer left New Orleans for Memphis, remaining at the 
 fleet until December 23, when he returned home. 
 
 During the second visit data obtained since first visit were examined, and, as 
 they were not conclusive, further tests were conducted, in which the writer par- 
 ticipated. These tests established the fact that the formula v~\/%g\i was correct 
 for tubes Nos. 8 and 9, and included tests of engines of dredges Epsilon, lota, and 
 Kappa. Results of tests will be discussed in the order given in the general 
 description of tests by Mr. Maltby. 
 
 boilers. 
 
 These tests were undertaken to give data for ordinary, everyday running. This 
 they do very well, but they do not contain all the information necessary for a 
 satisfactory analysis of results, and so only the most general conclusions can be 
 drawn. 
 
 Two kinds of boilers were tested, viz, the water-tube and a kind of horizontal 
 fire-tubular boiler, which is known as the Mississippi Eiver boiler. One peculiar- 
 ity of the furnace in the latter type of boiler is the small distance between top of 
 grate bars and bottom of boiler shells, which is much less than in stationary 
 boiler practice. In the furnace of the Gamma, for instance, the distance from 
 grate to bottom of boilers is 17 inches at the front, while at rear end of grate it is 
 only 15 inches. In stationary boilers 20 to 24 inches, or even more, is considered 
 good practice, the idea being to allow space enough for perfect combustion before 
 the hot gases come in contact with the comparatively cool surfaces of boilers. 
 
 The flame bed is also high behind the bridge wall. It appears that the gases in 
 passing backward along the surf aces of those boilers are compelled to pass between 
 and around the sides, which may account for the high efficiency of heating 
 surfaces. 
 
MISSISSIPPI RIVER COMMISSION. 161 
 
 According to Kent, for a fuel completely burned in the furnace and a given 
 steam pressure, the evaporation per pound of combustible depends on several 
 quantities, viz: The heating value of the coal; the difference of temperature between 
 boiler and air; the weight of flue gas per pound of combustible, which in turn 
 depends on the force of draft and the condition of the fire; rate of driving and 
 radiation. . It also depends on the manner in which gases pass through furnace, 
 dead spaces, eddies, etc. 
 
 The amount of draft is one of the most important factors. The draft in all these 
 tests was " good,' 1 but it was not measured. No analysis was made of the flue 
 gases, so we do not know how much air, above the quantity needed to supply oxy- 
 gen, was furnished, or if any. 
 
 The moisture in the coal was not determined; as coal was kept in open barges 
 the amount of moisture contained is difficult to estimate. 
 
 Attention is called to these facts merely to .show that a scientific analysis of 
 results is impossible. However, one point is obvious: it is that in these tests the 
 capacity of the Mississippi River boiler is far superior to that of the water-tube 
 boiler on the basis of square feet of heating surface. 
 
 For each pound of water evaporated per hour from and at 212° the average heat- 
 ing surface of all tested was 0.148 square foot for the former and 0.28 for the lat- 
 ter. While the average heating surface per horsepower was 5.11 square feet for 
 the fire-tu.be boilers, it was 9.69 square feet for the water-tube. 
 
 The average equivalent evaporation per pound of combustible from and at 212° 
 for the former was 9.1, while for the latter it was 9.32. For the boilers of the 
 .Gamma it was 10.64. 
 
 Great capacity is obtained, especially in the Mississippi River boilers, at a sacri- 
 fice of economy, although comparison shows the economy of the fire tube to be 
 nearly equal to, in the average case, and in one case to greatly exceed, that of the 
 water-tube type. 
 
 Mr. Maltby has pointed out the gain to be had from the use of a feed-water 
 heater; the point is well taken. 
 
 MAIN ENGINES. 
 
 In order to study the results of the tests of the engines of the Epsilon and Kappa 
 more carefully, a set of cards from each test has been selected and combined. 
 Separate cards from each test have been combined for the starboard and port 
 engines, using mean volumes of high and low pressures, as shown by average of 
 crank and head ends in each case. The particular set of cards to combine were 
 chosen because of their close agreement in mean effective pi'essures and in num- 
 ber of revolutions per minute with the average of all cards taken during the test. 
 
 The clearances of the engines of the Kappa were obtained from the cards, 
 assuming hyperbolic expansion. The values obtained were mean clearance, high- 
 pressure cylinder, 11.6 per cent; low-pressure cylinder, 5.6 per cent. 
 
 It was impossible to use this method on the cards from the engine of the Epsi- 
 lon as they were taken at an average of 172.7 revolutions per minute and showed 
 many irregularities, which were absent from those of the Kappa, which were 
 taken at an average of 129.8 revolutions per minute. The mean clearance of the 
 high-pressure cylinders of the Epsilon were assumed to be 11 per cent and the low 
 pressure 6 per cent. 
 
 From known clearances of other engines of the same make and approximately 
 of the same size, it is believed that these values are not far from the truth. 
 
 A glance at these combined cards is sufficient to convince one of the necessary 
 loss in the noncondensing engine. There is a large " drop " in pressure at high 
 pressure release and a low ratio of expansion. 
 
 The indicator springs were calibrated by the writer and found to be within the 
 limit of probable error of observation. 
 
 The consumption of steam per horsepower hour, 26.06 for the noncondensing 
 and 21.42 for the condensing engine, showed a balance of 4.64 pounds in favor of 
 the condensing engine. It is necessary to add about 21.7 per cent to the water 
 rate of the condensing engine to obtain that of the noncondensing. 
 
 The results correspond to carefully conducted tests made elsewhere, and are also 
 such as theory would lead us to expect when all the various losses are considered. 
 
 As the greater part of all the steam generated in the boilers is used in the main 
 engines, the coal bill can evidently be greatly reduced by using feed- water heaters 
 and condensing engines. 
 
 Sup. Eng. 1903 11 
 
162 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 MAIN PUMPS. 
 
 The various problems presented by the main pumps are of the greatest interest. 
 
 METHODS OP MEASURING DISCHARGE VELOCITIES. 
 
 The "method used to measure the velocity in discharge pipe is believed to be 
 practically the only method that could be successfully employed in this case. 
 Other methods were suggested. The Venturi meter is known to be an accurate 
 instrument, capable of giving results within 3 per cent of the truth, and possi- 
 bly closer average results; but unless an abnormally large meter could be used, 
 the loss of head at the throat, for velocities such as were to be measured, would be 
 excessive, and the pressure head necessary to overcome this loss would give abnor- 
 mal conditions for the pumps. Its weight would necessarily be great, and after 
 placing it on a suitable barge it would have to be attached successively to pipes of 
 different diameters and at varying heights from the surface of the water, as the 
 several pumps were tested. 
 
 A weir might have been constructed, but the uncertain stages of the river made 
 it doubtful how long it would have been of service. 
 
 A measuring barge might have been used in which water could be pumped for 
 a known length of time. 
 
 The Pitot-tube method is believed to be as accurate as any of the ways sug- 
 gested above, and has the great advantage of ease of application without materi- 
 ally changing the normal condition of discharge. 
 
 The Pitot tubes Nos. 8 and 9, when calibrated in running water by means of 
 floats, showed an average constant of unity in the formula u=cv Bgh within one- 
 third of 1 per cent. 
 
 These two tubes under very different static pressures showed average velocity 
 readings which differed by only one-tenth of 1 per cent. The difference in static 
 pressure amounted to about 17.7 feet of water. One tube was used where the 
 static pressure was nearly five times as great as that at the section where the other 
 tube was used; yet, as already stated, results agreed within one-tenth of 1 per cent. 
 
 As these results were obtained from 170 observations the agreement is not acci- 
 dental. This experiment shows that the results are not affected by a difference of 
 static pressure. Other data are given in the report of Mr. Malt by illustrating this 
 pomi still further. 
 
 It was also shown that the pressure obtained by means of a piezometer at the 
 outside of a straight pipe in which water is flowing is precisely the pressure of 
 the entire cross section of the pipe, corrected, of course, by height of point con- 
 sidered; and that it agreed exactly with the static side of a properly constructed 
 Pitot tube at that section. That in a straight pipe a Pitot tube can be used with 
 impact opening only, and static pressure obtained by means of a carefuliy inserted 
 piezometer at outside of pipe, is thus clearly shown. 
 
 From the above facts it is evident that the behavior of a correctly designed Pitot 
 tube, when used in an open canal or in a discharge pipe under pressure, is prac- 
 tically the same. There was no suction action on the static side of the tube, and 
 the impact reading was not affected by great difference of static pressure. It there- 
 fore follows that the constant of tubes Nos. 8 and 9 must be unity. 
 
 In making a traverse of a pipe considerable time was required; as much as an 
 hour or even more was sometimes used, consequently, irregularities appear, due to 
 the practical impossibility of keeping conditions of speed of engines and pumps 
 absolutely constant, as well as to other causes. 
 
 The pressure throughout the section of a straight pipe in which water is flow- 
 ing, corrected for difference of height of the points examined, was shown to be a 
 constant quantity, as already stated. 
 
 A great mass of data on this point was collected. The velocity parallel to axis 
 of pipe was found to be very much greater at the center than at the surface. A 
 platting of many results of the traverses in the discharge pipe of the Epsilon 
 shows the mean-velocity curve to be an ellipse, and that the velocity of the entire 
 cross section can be represented by an ellipsoid of revolution. This agrees with 
 results found, at much lower velocities, by Messrs. Williams, Hubbell, and Fen- 
 kell in their Detroit experiments. 
 
 As pointed out by Mr. Maltby, there was some variation in the traverses in the 
 general discharge pipes due to local irregularities, and so the ratio of center to 
 mean velocity was not a constant quantity. 
 
MISSISSIPPI KIVEE COMMISSION. 163 
 
 In some cases the center velocity was over 25 feet per second, while that near 
 the pipe was as low as 18 feet per second, parallel to the axis of the pipe. 
 
 Bernoulli's theorem states that the sum of static and velocity heads in any two 
 sections of the same pipe are equal when corrected for the loss of head due to fric- 
 tion between the two sections. 
 
 If it is understood to mean that the sum of the velocity and static heads at all 
 points in a given cross section are equal, it can only be true when the absolute 
 velocity of the various particles of water is considered and not the velocity par- 
 allel to axis of pipe. The particles of water in the outer part of the moving mass 
 encounter friction with the walls of the pipe and are retarded; complicated cross 
 currents are undoubtedly set up. Particles in the outer part of the moving stream 
 exert an influence on their neighbors, and this extends to the center of the pipe, 
 where the disturbances are least. The velocities of particles relative to each other 
 are thus greatest at the outside and least at the center. If it were possible to 
 measure these absolute' velocities and so to find the velocity head corresponding to 
 them.it is believed that Bernouilli's theorem would be true for all points in a 
 cross section. 
 
 METHOD OF COMPUTING- TOTAL WORK OF PUMPS. 
 
 The tests of 1897, made while pumping sand, have been reworked by Mr. Maltby 
 and form a part of his report. These tests as originally published are incorrect, 
 as they credit pumps with the sum of head shown by discharge pressure minus 
 the negative head shown by suction pressure, which includes the velocity head in 
 suction pipe and the computed velocity head in discharge pipe. Now, if suction 
 and discharge pipes were of the same cross section, the method used in computing 
 tests as reported gives twice the velocity head, and is therefore in error by the 
 amount of the velocity head. As the suction and discharge pipes are usually 
 unequal in sectional area a correction must be made , due to the difference of velocity 
 head at these two sections. This point will be taken up again later. 
 
 Another point is misleading in the report of tests of 1897, although the actual 
 error involved is very small. The various heads are always stated in feet of water. 
 As a matter of fact a mixture of sand and water was pumped, and the heads ought 
 to be measured in feet of liquid pumped — that is, the mixture. 
 
 In some cases instead of 62. 5 pound s per cubic foot the mixture weighed between 
 70 and 75 pounds per cubic foot. 
 
 In any case the work done by a pump is equal to QxWxH, where — 
 
 Q=volume pumped per unit of time. 
 W= weight per unit volume of liquid pumped. 
 
 H=total h.ea.d=h d — h s -\-h' -\-h" . 
 
 7i,2=discharge head, obtained by observing the pressure p,z, at some point in the 
 
 discharge pipe, then h=^.. 
 W 
 
 ^=^=suction head, obtained at some point in suction pipe by observing the 
 
 pressure^. It includes the velocity head, as will be shown. 
 ft'=difference between velocity heads in discharge and suction pipes. 
 ft"=difference in level of the two points where discharge and suction pressures are 
 measured. 
 
 Total work=QxWX(^-^+?i'+/t'')=QxWxr|--|+;i'+fe''\ 
 
 It will be seen that nearly all the total head is included in the first two terms of 
 the quantity within the parentheses and that the error arising from computing 
 results when pumping sand due to using the weight of a unit volume of water 
 instead of the weight of unit volume of the mixture only affects the last two terms. 
 In the case of h' it is small enough to neglect and probably so in the case of h". 
 
 It is to be noted that h s ia always negative in these tests, but both h' and h" may 
 be positive, negative, or zero, depending on conditions in individual cases. Very 
 little error will therefore be made by using W as the weight per unit volume of 
 water instead of the weight of mixture pumped, and the method of comprting 
 results is greatly simplified. 
 
 It will now be shown that the suction pressure includes the velocity head in 
 suction pipe where pressure is measured. 
 
164 REPORT OF THE CHIEF OF ENGINEERS, U. S. ARMY. 
 
 
 
 
 If 
 
 G 
 
 
 
 
 
 h 
 
 
 A 
 
 . 1 
 
 w 
 
 _ B _. 
 
 f 
 
 
 
 V* 
 
 
 
 
 C 
 
 *« 
 
 1 5 
 
 "t 
 
 £. 
 
 
 
 m 
 
 n 
 
 
 
 K 
 
 . Referring to figure, let two tanks K and L be connected by a pipe of uniform 
 cross section in which water flows with a constant velocity v when the level of the 
 water in the tank K is at ^4. Suppose a suitable supply provided for tank K or 
 that the water level is always kept at A\ also an outflow provided for tank L, so 
 that water may be discharged into L and a desired height always maintained. 
 
 Now, neglecting losses due to friction in pipe at entrance, etc., the "hydraulic 
 gradient" would take the position of line CE, and this would be the level of the 
 water in the tank L to maintain continuous flow at a velocity of v feet per second. 
 v 2 
 
 AC=BE= 
 
 H 
 
 Let two piezometers be placed in pipe at m and n; the level in 
 
 these piezometers would be s, t. 
 
 Suppose at section X we cut pipe and introduce a pump, which will always main- 
 tain the uniform velocity v in the pipe and also deliver the water into tank Lata 
 level AB. The level of water in discharge tank will then just equal that in suction 
 tank, and the work of the pump will clearly be to produce the velocity head. 
 
 At the left of section X the conditions are unchanged from those of the first case 
 taken, for the constant velocity v is maintained. 
 
 The height of the water in piezometer m will therefore be at s. On the right of 
 section X the level of water in tank L is raised to B, and therefore the water in 
 piezometer n will stand at w. 
 
 The difference between the readings of the two piezometers will thus be seen to 
 include the velocity head; in fact, in this case would be just equal to the velocity 
 head. 
 
 Again, if water is pumped into tank L and the level in tank is raised to G, while 
 the same velocity v is maintained in suction and discharge pipes, the level of water 
 in piezometer n would be at F, while in the piezometer m it would still stand at s; 
 
 the difference between these two readings would plainly be EG=£ -h IT and is seen 
 
 2g 
 
 to include the velocity head. 
 
 The simplest possible case has been taken to prove this point; the proof still 
 holds, however, if we consider all losses in the pipe. 
 
 It will be seen that if section at m differs in size from that at n, the hydraulic 
 gradient will not be a straight line; but in any case the velocity head which is 
 included is that at m, i. e., the velocity head corresponding to the velocity in suc- 
 tion pipe. The velocity head in discharge pipe will not be shown by piezometer, 
 but may be computed. If it is more than that in suction pipe the difference must 
 be added to suction and discharge heads; that is, h 1 is positive; if less, h' is nega- 
 tive: in the equation H=h d —h g -\-h' -j-h" . 
 
MISSISSIPPI RIVER COMMISSION. 165 
 
 The correction for difference of level /i", at which discharge and suction pres- 
 sures are obtained, has been already referred to. 
 
 Then, knowing discharge head, suction head, the difference (if any) in velocity 
 heads in discharge and suction pipes, the correction due to difference of level, and 
 the proper sign for each and all of these quantities, we are able to compute the 
 total head H. 
 
 From the mean velocity in discharge pipe the quantity pumped is known, and 
 we thus have all quantities necessary to determine the total energy given by pump 
 to water as accounted for between a point in suction pipe and another point in 
 discharge pipe. One possible error arises from the assumption of the mechanical 
 efficiency of all the engines as 92 per cent. 
 
 The cards taken on the Gamma showed the friction to be about 6 per cent, while 
 on the Iota it was about 12 per cent. Any engineer who has tried to obtain the 
 friction cards from an engine realizes how unsatisfactory the results usually are. 
 In this case they were so much so that it was decided to use a constant quantity 
 for all. It is not improbable that a small error was introduced here, which, if it 
 could be known, would slightly change results. 
 
 EFFICIENCIES OF PUMPS. 
 
 When an attempt is made to find the reason of high efficiency in one case or low 
 efficiency in the other of the main pumps the greatest difficulties are encountered. 
 
 Hydraulic problems are very complex and must be attacked largely from the 
 experimental side. An example will illustrate this. Only three years ago it was 
 maintained by'some engineers of the highest standing in America, in discussing 
 thePitot tube, that velocity head and static head were not mutually convertible. 
 The admirable researches of Mr. W. M. White settled this point. 
 
 In a centrifugal pump with many blades, and where the path of the water is 
 therefore well circumscribed, it would be possible to take a series of tests, r.nd, 
 knowing the quantity of water pumped, angular velocity of impeller, shape of 
 blades and of pump casing, to analyze those results with more or less accuracy 
 and locate losses with some degree of certainty. When, however, the number of 
 blades en impellers is only four or five the mean path of the water through them 
 becomes very uncertain, and thus the problem of locating lo&ses proportionately 
 harder. The problem depends on many conditions, even in pumps of the same 
 form and size, the head pumped against, the quantity pumped, and thj velocity 
 of rotation of impellers, all of which are interdependent and may vary widely. 
 
 Here, then, is a problem in which theory can only suggest probabilities and 
 experiment must decide. 
 
 In each of the main pumps the writer has attempted to follow the path of the 
 water as best he could from the suction pipe into and along runner at the periph- 
 ery and through vortex chambers to discharge pipe. The results are not satis- 
 factory for the reasons already stated; however, they suggested a few points which 
 will be mentioned. 
 
 Radial flow to impeller was assumed; its mean value was obtained from the 
 dimensions of pump and knowing the quantity of water pumped. This velocity, 
 together with the known velocity of rotation at inner ends of runners, should 
 determine the angle at which the water enters the blades and show the amount 
 of shock, if any, provided that the assumption of radial flow is correct. 
 
 Asamatter of fact we know that this assumption is not absolutely correct, as 
 shown by the wear at the intake of the pump casing. 
 
 The amount of this whirl that is set up before water enters blades of impellers 
 was not measured and is not known. It is probably less on those pumps provided 
 with plates in suction elbows to prevent such whirl. However, this assumption 
 of radial flow offers a worlnng hypothesis. 
 
 On this basis the Delta shows by far the greatest loss in shock at intake, while 
 the Gamma shows the least, the Epsilon being next to the Gamma. The runner 
 of the Gamma has only 4 blades, which makes the results possibly less reliable than 
 in the Epsilon and Delta with 5 blades. 
 
 Any sudden change of velocities in the water along its path through the pump 
 is accompanied by losses due to shock and to eddies. The velocities throughout 
 the pump of the Epsilon are far more uniform than for any of the other cases. 
 Among the other pumps, possibly the Gamma has more uniform velocities 
 throughout the pump than any other. However, there are sudden changes in all. 
 
 It is known to be a fact that where a small number of blades are used on 
 impellers cobblestones have been retained for some time behind runners instead 
 of being thrown out radially. This, of course, could only be due to eddies which 
 
\ 
 
 166 REPORT OF THE CHIEF OF ENGINEERS, U. 8. ARMY. 
 
 would be largely overcome by increasing the number of blades, but the increase 
 of surface would cause greater surface-friction losses. 
 
 In another class of hydraulic machines, viz, turbines, it has been found desir- 
 able to use a great many blades. The question of the proper number of blades 
 for a centrifugal pump can best be settled by experiment. The writer does not 
 know of any such experiments having been made. 
 
 The efficiencies obtained in every case, except that of the Delta, are very satis- 
 factory and fairly uniform. Undoubtedly they are not absolute efficiencies, but 
 it is believed that they are as close to the truth as it is possible to arrive by care- 
 ful experimentation. 
 
 It is a noteworthy fact that the Gamma gave almost identical results for the 
 efficiency tests of 1897, made while pumping sand, and the tests of 1902, made 
 while pumping water, when efficiencies are corrected in the former. 
 
 The pump was the same in both tests, having been merely kept in repair since 
 the first tests. However, the "total head" for the two sets of tests was quite 
 different. 
 
 The Epsilon and Zeta showed much higher efficiencies in 1897, when they had 
 impellers with 7 blades and the angles of intake of these blades different from 
 that on the 5-blade impellers now in use. The total head and the revolutions per 
 minute all vary, and it is therefore impossible to draw any definite conclusions 
 from a comparison of the results of these two sets of tests. However, the great 
 difference between the revolutions of the Delta may account for the low efficiency 
 of the pump in tjie test of 1902, at least partially. In both cases the Delta shows 
 tli3 lowest efficiency. 
 
 Highest average efficiency was attained with a pump having a vortex chamber 
 circular in cross section, while the lowest was attained by a pump having a rec- 
 tangular 1 cross section for its vortex chamber. The highest individual efficiency 
 was attained on the Flad. Both the Kappa and the Flad, with their square 
 casings, showed good average results. 
 
 The writer has failed to discover any obstacle worth considering to the use of a 
 double suction. Certainly from the mechanical standpoint the perfect balance 
 resulting makes it highly desirable. 
 
 The single-suction pump showed the lowest efficiency. This, however, does not 
 prove anything, as proportions of casings, shapes of impellers, total head, etc., 
 varied widely in the different pumps. 
 
 JET PUMPS. 
 
 A comparison between the compound-condensing, direct-acting steam pumps 
 and the centrifugal pumps driven by ineans of compound-condensing engines 
 should include the relative consumption of steam in the two types of engines. 
 The former, in the experience of the writer, requires about 33 pounds of steam 
 per horsepower hour, while the latter requires between 22 and 25 pounds. 
 
 The probable gain in steam consumption by using the high-speed compound- 
 condensing engine, in average cases, will not be far from 25 or 30 per cent. 
 
 The greater steam economy of the engine of the centrifugal pump will thus 
 partially offset the superior efficiency of the reciprocating pump. The slip and 
 leakage in the reciprocating pumps may also be quite an appreciable quantity. 
 
 LOSS OF HEAD IN DISCHARGE PIPE. 
 
 The loss of head due to the double elbow at the stern of dredges Kappa, Iota, 
 and Flad varies somewhat. This point has been discussed by Mr. Maltby. 
 
 The discharge pipes have countersunk rivets and are polished smooth and bright 
 by the sand. The only disturbing influences in the flow of the water with a 
 straight discharge pipe are due to the lapping of the various sections where they 
 are riveted and to the flexible rubber joints connecting the different lengths of 
 pipe. 
 
 These pipes are made up of sections, each being inserted into the succeeding sec- 
 tion, thus making the mean area of pipe at the center of the section, where all 
 these piezometer pressures were measured. 
 
 In case discharge pipes are not straight , there is loss due to the curvature of these 
 flexible joints. The drop along the discharge pipe of the Flad shows the latter 
 loss to be small even in the extreme case. 
 
 Mr. Maltby has given the data in the case of the four air cocks placed on the 
 " quarters ' * of a section of pipe. It is to be remembered that these air cocks were 
 purposely placed to show the greatest possible error, and that it is extremely 
 
MISSISSIPPI RIVER COMMISSION. 
 
 167 
 
 improbable that any such variation as they showed ever occurred in tiie data 
 obtained in the tests. 
 
 As there is very little data on losses in pipes of this size, and at high velocities 
 obtained, the results are of peculiar interest and value. 
 
 Using the formula B r = £ n v or /= f , we may compute from the data 
 • D 2g ±lhv 
 
 given the value of the coefficient of friction /, for the pipes, including the flexible 
 rubber connections. 
 
 The values of / as again given in the above formula, and obtained in most cases 
 from mean values, are shown in the table below: 
 
 Name of dredge. 
 
 Diameter 
 of dis- 
 charge 
 pipe. 
 
 /. 
 
 Gamma 
 
 Inches. 
 34 
 34J 
 33 
 
 m 
 
 33 
 
 0.00445 
 
 Delta 
 
 .00547 
 
 Zeta - - 
 
 .00569 
 
 Iota 
 
 .00455 
 
 
 .00413 
 
 Flad 
 
 .00556 
 
 
 
 
 
 
 .00503 
 
 
 
 The fact that zero pressure was found in the side of the discharge pipes of the 
 Iota, Kappa, and Flad is accounted for by the fact that the water was flowing at 
 a constant velocity in an open pipe, the only force retarding it being the skin fric- 
 tion and viscosity of water. The pressure will be zero when we go back from the 
 discharge end of pipe to a section at which the force necessary to overcome this 
 friction was just balanced by the head due to the open end of the pipe. The piezom- 
 eter was at the center of the pipe, so the pressure ought to be neutral at such a 
 
 distance as will make Bj-=—. Results show this distance to be approximately 30 
 feet. 
 
 AMOUNT OF SAND IN MIXTURE PUMPED. 
 
 It is interesting to find the sand so thoroughly mixed with the water; the results 
 indicate that there was slightly more on the lower side than on the top side of pipe, 
 but that the mixture was fairly uniform. 
 
 If the pipe were inserted in discharge pipe at such a point that the static pres- 
 sure were great enough to give the water in sampling pipe the same velocity as 
 the water surrounding this pipe has on its way to the end of discharge pipe — in 
 other words, at such a point that it is equally easy for water to escape from 
 sampling and discharge pipes — there could be no doubt about obtaining an average 
 sample. 
 Respectfully submitted. 
 
 W. B. Gregory, Consulting Engineer. 
 Capt. Wm. B. Ladue, 
 
 Corps of Engineers, U. S. Army, 
 
 Secretary Mississippi River Commission, 
 
169 
 
 Additional reports op Asst. Engineer F. B. Maltby. 
 
 Memphis, Tenn., September 1 8 ', 1903. 
 Captain : 
 
 I have the honor to submit the following report of the results of tests made 
 of the efficiency of the main centrifugal pump of the dredge Zeta with runners 
 having different numbers of blades. This report to be supplemental to my re- 
 port of May 7, 1903. 
 
 The tests described were made in obedience to the instructions from the 
 Committee on Dredges contained in your letter of May 8, 1903, as follows : 
 
 "1. Install old 7-bladed runner on the Zeta and make tests with the same 
 under the conditions when the runner now in use (5-bladed) was tested. 
 
 "2. Alter the pattern of the Epsilon*s and Zeta's runner, if practicable, so 
 as to cast a runner with three blades, and test this in the same manner." 
 
 It was found impracticable to comply with the instructions contained in the 
 second paragraph in so far as altering the old pattern and a new pattern was 
 made, from which a runner having 3 blades was cast. 
 
 A *drawing is submitted herewith, showing the outline and dimensions of 
 the pump casing and of the runners with 3, 5 and 7 blades, and attention is in- 
 vited to the following characteristics : 
 
 The pump has a double suction and discbarges horizontally at the bottom; 
 it has a vortex chamber of a spiral form in elevation and with an approximate 
 elliptical section. The sides of the casing between the vortex chamber and the 
 short curve leading to the suction openings are straight. 
 
 The runners all have an outside diameter of 69 inches at the tips and a di- 
 ameter of 30V2 inches at the inside shoulder of the blades. The degree of cur- 
 vature of the face of the blades is the same in each case. The angle which the 
 face of the blades makes with a tangent to the circle at the inside shoulder of the 
 blades is the same on the 3 and 7-bladed runners, but is somewhat sharper than 
 on the one having 5 blades. It will be noticed that the inside shoulder of the 
 blades is considerably nearer the shaft than the straight portion of the casing, 
 owing to the worn condition of the curved throat casting, but whatever effect 
 this may have on the efficiency of the pump would be relatively the same with 
 each runner.' 
 
 The amount of clearance between the edges of the blades and the casing 
 was adjusted in each case by the use of edge plates, fitted to the shape of the cas- 
 ing and bolted to the face of the blades. On the runner with 5 blades, which is 
 the one in use pumping sand at present, these edge plates are of cast iron and 
 have a thickness on the edge of about two inches. These edge plates give the 
 face of the blade as a whole a trough shaped section, as shown. The edge plates 
 used on the 3 and 7-bladed runner were plain steel plates, only about V± inch in 
 thickness, and the face of the runner is flat. The clearance in each case was 
 from Vic inch to Vs inch. 
 
 methods. 
 
 The instruments used and methods employed in making observations were 
 the same as described in my previous report as referred to above. Conditions 
 of operation were as nearly the same during the tests of each runner as it was 
 practicable to make them. One thousand feet of discharge pipe was used and the 
 same instruments were applied in the same positions in each instance. 
 
 * Not printed; see Plate 36 for details of pump and 5-bladed runner. 
 
170 
 
 RESULTS. 
 
 A table of data and results is submitted herewith. Each result is the mean 
 of about twenty observations. 
 
 In the mean results for each runner it will be observed that certain figures 
 have been omitted. This was done, not because of any lack of confidence in the 
 figures, but because they covered speeds widely different from those observed 
 with the other runners. 
 
 The mean efficiency of engine and pump with 3 blades is 71.6 per cent ; with 
 5 blades, 72.1 per cent; and with 7 blades, 73.1 per cent; a total difference of 
 only V/z per cent. The amount of work done in each case varies considerably, 
 being 231,475 foot-pounds per second with 3 blades, 282,589 foot-pounds per sec- 
 second with 5 blades, and 288,518 foot-pounds per second with 7 blades ; or the 
 5-bladed runner did 22 per cent more work than the one with 3 blades. The one 
 having 7 blades did 24 per cent more work than the one having 3 blades, but 
 only 2 per cent more than the runner with 5 blades. 
 
 CONCLUSIONS. 
 
 From the above results it appears that, considering mechanical efficiency 
 alone, it makes very little difference, within reasonable limits, how many blades 
 there are on a runner, though the greater number has a slightly higher mean. 
 
 The capacity of the plant, however, is very much more with the higher 
 number of blades, though the difference between 5 and 7 is small. 
 
 Apparently the plant ran more smoothly and with less vibration with the 
 7-bladed runner than with the 3 or 5-bladed one, and the gages fluctuated very 
 much less, showing that the work being performed, or the output of the pump, 
 was more uniform. 
 
 Very respectfully, 
 
 Your obedient servant. 
 
 F B. Maltby, 
 Assistant Engineer. 
 Capt. Wm. B. Ladue, 
 
 Corps of Engineers, JJ. 8. A. } 
 St. Louis, Mo. 
 
171 
 
 N 
 
 3 
 
 <3 
 
 •09S .lOCl 'Ijf 
 -9A |i:.Mi|ih.M,| 
 
 *inao 
 aad idomd 
 jo ^ouaiouja; 
 
 '5U83 jrad JauiS 
 -us pan sduirid 
 jo ^ou3ioiya 
 
 ■oae jod -scli '%£ 
 ■aaop ^.ioja 
 
 "Woi 
 
 "A:nooi3A. 
 
 •a\i9aii9CI 
 
 •uoipng 
 
 7^ 
 
 "2 c 
 
 i *■< 
 
 t? 
 --.A 
 
 m-HTjioo 
 
 Ol ITS to r— — ' OS 
 cq to to tOio ec 
 m i- c; w m S 
 to o: o — ■* to 
 cacoio-* — o 
 CM cm cm oi cm oi 
 
 O Or- O OO 
 
 -t-I— t— t— t~- 
 
 ic o c: o m 
 
 CO t- CO CO YD 
 
 IQ "*H CO -# "* 
 
 CO-**- t-CD 
 to — r- ic -* 
 aiooi — m 
 i--hi— uoco 
 — < cm c-i ot w 
 
 a) a- 
 
 « a 
 
 OS »o 
 
 c? cm os i-i — > cm to 
 
 OS 5: lO CO — OS tC 
 
 — rH r- .-I -+ rl IO 
 
 I— CO X CO CO i-- 1- 
 
 lO CO 7-; ^ *- OS £- 
 CO CM O i-> CO t— GM 
 
 o'^'io^Kom 
 t— t- 1— t- 1> i— *- 
 
 CO CM -* -H OS i-. - 
 
 OI>r1(--(M 1 I — 
 CS tO 70 ■* CC ' J- 'O 
 
 r»^»i-' - - - 
 
 - c; 00 
 hosco 
 
 10 to i- 1- oo 
 
 3 COCO «>« :c to 
 
 OS — < — "M 1-COIQ 
 
 to* i> cc to ic to co 
 
 ■03S jad 'sqi 
 •oSjBqoeia 
 
 ■30S I9d '1J Uf) 
 
 ■aSj^qosici 
 
 '^IBAiJO'U 
 "PB9H 
 
 ApO{9A. 
 
 •oas J9d *ij; 
 
 *A!}I0019A 
 
 UB9l\[ 
 
 09« .i9d *q,j[ 
 
 i?]00I9A 
 * J9^U90 
 
 'PB3I1 
 
 a"iiooi9A. 
 
 ■098 Jad "U 
 *X)T00T3A 
 
 nB9M 
 
 •J H *I 
 
 IB! ox 
 
 ■ecri 
 a^neaajd ulnars 
 
 •aintnui I9d 
 enoii}iitOA9*a 
 
 o o to O CO o 
 
 ^cooor-N 
 os — i— -* f— :M 
 
 O 'JO OS © -# CJ 
 •M -* (N -*■ CO lO 
 
 3QOOS to-* >o 
 
 CO CO -* CO CO CO 
 
 1O1DCDHC0O 
 
 i^r-ooco-j ss 
 
 IO to to O lO -f 
 
 r- -»■ cm ic to im 
 
 os coco tr co-* 
 
 OS OS O* OS CO CO 
 
 CO 
 CO 
 OS 
 
 o< co eo r— os -& 
 OS OS CO os to tc 
 
 OJ CM CO CM Ol O* 1 
 
 OS 
 CM 
 
 OCM "* r-. tO CM 
 
 CO COn- CO CO CO 
 
 CO CO CO CO lO I 
 OS Ol I- l-l— 
 
 -i" IO tO tO iO 
 
 co^xot 
 oi ■— m >o oi 
 
 O M Ol t- O 
 
 CO CO- ^* '■d* "^ 
 
 ot-r-coN 
 
 tO © "«*• O I L- 
 r-OW-*-* 
 
 -^m toco to 
 
 CO © o o o 
 
 -* O IO IO - 
 ^ lO ■— CM CO 
 
 CM CM CO COCO 
 
 CM tDCO O-T 
 CO O Ol CO CO 
 
 CM CO -tji -^ji ■* 
 
 t~o O-* SCO 
 
 o i- rM in — cs i~ 
 
 .-I CM 3S as ^ —t = 
 CO "-O lO o io to ^ 
 
 io -^ r- m oi -o io 
 
 — CO CM |-~- O OS CCi 
 
 r^ co i.— oi os o co 
 
 ^H lO -f -* ^ -* '* 
 
 CO IO rH OS 1 - -+ CO 
 
 OS to to lO OS » CO 
 ■* OS OS O O IO CO 
 
 I - t— CD I— to I— t- 
 
 CO Ol ^1 CO lO tr~ CM 
 
 o co-* to t-ac o 
 
 tC —1 OS O CO tC "* 
 
 CO CO CO CO CO CO CO 
 
 lO CO OS -#l— ■>* S. 
 CO -^H to IO i— 1 OS OS 
 
 cm to i— co o cm c: 
 
 iCiO^HI^l'SIO 
 
 -* o OS io os ^* eo 
 tc t^ CO co to >o © 
 
 q cj 
 
 s 
 
 D hC 
 
 § a 
 
 e3 T3 
 
 a 3 
 
 « 3 
 
 a ^ 
 
 a s , 
 
 
 d fl © 
 
 S7S 
 
 ^ r^ 
 
 ^? 
 
 CM 
 
 to 
 
 Q> 
 
 p- 
 ft 
 
172 
 
 Memphis, Tenn., October SO, 1903. 
 Captain : 
 
 I have the honor to enclose herewith a copy of the results of the boiler test 
 of the dredge Beta, made October 23, 1903. It will be noted, in comparison with 
 the tests submitted with my report on tests of dredges, that the efficiency is 
 slightly less than on the Delta. The difference is very small and I think, there- 
 fore, that the tests of the same boilers on the two dredges represent very well 
 the actual efficiencies of this class of boilers. 
 
 I do not see how a better test than we got on the Beta could be devised. 
 Everything moved very smoothly and uniformly duririg the whole duration of the 
 test. The amount of water pumped per minute varied hardly any for hours at a 
 time. The coal went in with almost equal regularity. 
 Very respectfully, 
 
 Your obedient servant. 
 
 F. B. Maltby, 
 
 Assistant Engineer. 
 
 Capt. Wm. B. Ladtje, 
 
 Corps of Engineers , U. S. A. 
 St. Louis, Mo. 
 
 Results of boiler tests. 
 Dredge Beta. 
 
 Bate of trial, October 23, 1903. Duration of trial, 9 hours. 
 Type of boiler, 2-Heine water tube. 
 
 Grate surface, 109 square feet. Water heating surface, 5,436 square feet. 
 Proportion of grate to heating surface, 1—50. 
 Proportion of air space in grate, 
 Average pressures, temperatures, etc. 
 
 Steam pressure in boiler by corrected gage, J 73. 3 165.3 lbs. 
 
 Temperature of feed water entering boiler 1G6.1 degrees. 
 
 Temperature of gases escaping from boiler 637 degrees. 
 
 Draft gage, inches of water ^ 
 
 Percentage of moisture in steam 21% 
 
 Fuel. 
 
 Total amount of coal consumed 23,066 lbs. 
 
 Total amount of refuse, dry 2,645 lbs 9.4%. 
 
 Total amount of combustible 25, 421 lbs . 
 
 Coal consumed per hour per square foot grate surface -8f ff lbs. 
 
 Coal consumed per hour per square foot heating surface 
 
 Water. 
 
 Total water pumped into boiler and apparently evaporated, corrected for dif- 
 ference of level in boiler, leakage and steam pressure 216,163.5 lbs. 
 
 Water actually evaporated, corrected for quality of e Learn.. 211,523 lbs. 
 
 Equivalent water evaporated into dry steam from and at 212 degrees 232,464 lbs. 
 
 Equivalent water evaporated into dry steam from and at 212 degrees, per 
 
 hour 25,S291bs. 
 
 Equivalent water evaporated into dry steam from and at 212 degrees, per lb. 
 
 of coal 8.28 Ibs^ 
 
 Equivalent water evaporated into dry steam from and at 212 degrees, per lb. 
 
 of combustible 9.15 lbs. 
 
 Equivalent water evaporated into dry steam from and at 212 degrees, per 
 
 square foot heating surface per hour 4.7 lbs. 
 
 Equivalent water evaporated into dry steam from and at 212 degrees, per 
 
 square foot grate surface per hour 236.8 lbs. 
 
 Commercial horse-power developed, based on evaporation of 34J lbs. of wa- 
 ter from and at 212 degrees per hour 748 H. P. 
 
173 
 
 Memphis, Tenn., November 3 7 1903. 
 Captain : 
 
 I have the honor to enclose herewith a blue print of " Table Showing Re- 
 sults of Tests of Main Pumps". 
 
 This table has been revised to include the Beta and the Iota with the 
 shrouded runner. It will be noted that the quantities for the Iota with the open 
 runner have also been changed slightly from the original table, owing to an er- 
 ror in computation. 
 
 The Beta's efficiency at ordinary speeds of operation is only about 55 per 
 cent, not quite as good as the Delta. I know of no reason for this except that 
 it is due to the type of pump. The clearance between the runner and the casing 
 is slightly more than on some of the other pumps, but I do not believe that it is 
 enough more to account for all the differences in efficiency. 
 
 The table also shows the additional efficiency, and especially the capacity, 
 obtained by putting the shrouded runner on the Iota. The efficiency of the en- 
 gine and pump at 168 revolutions per minute is about 76 per cent, which is the 
 highest efficiency we have yet obtained and is over 10 per cent greater than that 
 obtained with the open runner. The average amount of work done is 341,500 
 foot-pounds per second, as against 310,800 with the open runner at the same 
 speed, an increase of nearly 10 per cent. 
 
 In this connection I beg to invite your attention to the table, and especially 
 in reference to the capacity of the pumps at different speeds. A diagram is 
 enclosed, showing graphically the work done by each dredge in thousands of 
 foot-pounds per second by each pump at varying peripheral velocities. In every 
 case it will be noted that the capacity of the pump increases to a very marked 
 degree to the maximum speed, with no indication that the speed for maximum 
 capacity has been reached on any pump. The efficiencies at varying speeds are 
 also plotted on the same sheet and show very little difference at different speeds. 
 Generally speaking, the efficiencies are slightly lower at the higher speeds, ex- 
 cept on the Beta, where the efficiency seems to increase sligh*!y with the speed. 
 This is true, also, on the Iota'. 
 
 In the whole series of experiments, leaving out only the Beta and Delta, 
 whose efficiencies are so much lower than the others as to be very marked, the 
 whole range of speed from 42 to 55 feet per second, or 30 per cent increase, has 
 been followed by an extreme range of efficiency of from about 63 per cent to 75 
 per cent, while the range in capacities is over 200 per cent of the lowest. Evi- 
 dently, then, regarding mechanical efficiency alone, speed, within reasonable 
 limits, makes very little difference, but to obtain the maximum output from the 
 plant the pumps should be run at as high a speed as is practicable. 
 
 Very respectfully, 
 
 Your obedient servant, 
 
 F. B. Maltby, 
 
 Assistant Engineer. 
 Capt. Wm, B. Ladue, 
 
 Corps of Engineers, U. S. A., 
 St. Louis, Mo. 
 
174 
 
 4 
 
 
 = 
 - 
 
 s 
 ■^ 
 
 
 500 ft. 
 pontoons. 
 
 rtinner. 
 
 o a 
 u a 
 
 ■« P 
 
 
 
 1*2 
 I'M 
 
 p. 
 
 H^OffiSlh' ... 
 CO CO OS OS OS CO ... 
 rH O C OO — . - 
 © © C O © © ... 
 
 ^ ,_J r-1* r-I rH — 
 
 ■oas jad *jj[ 
 ■X^iooy 
 -8 a. i^iaqduaj 
 
 t- CO CO CO 
 
 "ph co coco qo oo 
 
 — o: OS c: r— r- 
 
 -sf OS CI OS OS OS 
 
 cs" GO CC CO -* ■» 
 
 or i-t- 
 
 co en os 
 
 auaa.tod idinnd 
 jo Aorcajonja 
 
 oi Ol CO CO -# -H 
 
 r- 1— c- 1 - 1 - 1- 
 
 !OCf 
 
 —'cooi 
 
 £-00 00 
 
 •3«30 .iad iauiS 
 -us pcre sduind 
 jo Aouaioroa 
 
 31 Ct-CQO)-* 
 
 Ol O* CO CO -** iO 
 
 •j t-, so ■■£■ -JD iD 
 
 OS--*i-i 
 
 iC as iD 
 CD L-l- 
 
 oas Jad -sqj •$& 
 ■auop 5[.iOA4 
 
 toco© ooio -* 
 
 -f — oo io — io 
 to <m co a -^ ■-£ 
 co r-co r~r-© 
 
 O i-H CD -H OS CM 
 
 KCOXTOH 
 
 (N oi cm oi co eo 
 
 ©-- 1— 
 -* osm 
 
 (MOD — 
 OH — 
 CO OS CM 
 O! IO 1- 
 CO -f CO 
 OICO CO 
 
 "T^OX 
 
 "i^i00T8A 
 
 ■ aigai^C 
 
 ■noijong 
 
 5 CO CO CO Iffl IO 
 
 ,0J CO CO CO CO CO 
 
 OS o o c o - 
 
 oas jad ■sq'j 
 •aSjBqostQ 
 
 oas Jtad - u -no 
 ■aSjBqosia 
 
 ■J9!JBM JO'U 
 
 ■psaq 
 
 £| 10013 a 
 
 co r ■ x. c o 
 
 io in oi i-co so 
 
 IN <a -A X 00 
 COCO CD as CO CD 
 
 — OOt-CC — — ( 
 COCO -H-*U CO X 
 r- M* CS CM CM Oi 
 
 iK •* io <n c o 
 
 03S jced 'ij 
 
 'j£;pO[3A 
 
 uboh 
 
 oas jad "iji 
 A^iooiaA 
 jajaof) 
 
 •pBaq 
 
 ^TOOiaA 
 
 ■oas Jad -^ 
 •^iooi8A 
 
 UB8RI 
 
 *J *H I 
 
 T«lox 
 
 r~ i— oo oo co 'jo 
 
 © co o © o o 
 OOCOiOCO 
 CO tH CD iO CO TO 
 
 CO "^ "^ ^ *^ *^ 
 
 ID .01 CM — CO CO 
 — Ol CD lO OS OS 
 © CD CO I - CO CO 
 
 as as cd cd r- 1- 
 
 CD CO CO 'J3 IO « 
 
 ■jCjnojaui 
 
 saqoui'uinnaBA 
 
 ajnseaJd urea^S 
 
 ■Sjjnupm jad 
 euoi^nioAaa 
 
 CO -3->rJ* 
 OlOl CM 
 
 oc co — 
 
 IO I -CD 
 CO r-t- 
 
 ior- as" 
 
 IOCN CD 
 WhX 
 
 -* CO i.O 
 
 ss coo 
 
 OS OI o 
 t-X>G0 
 
 a) 
 
 V 
 
 >. 
 
 
 
 -. O I - CD OJ C . 
 CO00-«CC;CO-f^t(CO0C 
 
 oi— 'lOolcracDXCioos 
 
 ■l-XCOCO — COOSrH 
 
 as co t- oo o o io »« 
 
 • it; c ir: ic ffl co io o 
 
 
 oa — oicNcoioini> 
 
 OJH*CO"*iOiO^O 
 i(C IO IO ID »C IO IO IO 
 
 OOC-OOOlOiO- 
 
 0ilOOH~-*r-|flf 
 
 CO t~^ id — J CO IO* IO -*" CO 
 
 -mcoi— cr f* -^ oi eo ■* 
 eocoifc — eoricscDO* 
 
 OI CO Ol (N CO CO to 00 ^* 
 
 r-iocot-ooceoo — 
 
 0JCMOJOICCOICOO1O1 
 
 Ol OS CO CM CO - 
 f CiftSCT- 
 
 o o o o" - 
 
 oj— icocsoooscooieo 
 
 COCO = COCOCS=:COCD 
 
 cooiioajooco — coo 
 i— ir- 1^ r>- 1- I— cc cd i- 
 
 CDiOe0CDCDCDt-»O>fS 
 
 »iO»oa3cDcoeD'*io 
 
 ■ io os oo o r^ i-- >o -* 
 
 ■l-HHMB'.'tlHI- 
 
 n oi o co oi oi co r— 
 
 CD Q3 
 
 .2 5 
 
 ft e3 
 A 3 
 
 0> 
 
 m 
 
 s 
 
 
 cd 
 
 CI) 
 
 73 
 
 a 
 
 
 
 TJ 
 
 
 
 4) 
 
 a 
 
 J3 
 
 1 
 
 H 
 
48