BENJ\M From Excerpt ...i/nutes of the Proceedicgs of the Inolitatior of Civil Engineers. JX "VAN FOSTR/lSr:>, PUBLISHER, /? MURF.AY ANI> ^ ' WARREN STREET. 1882. THE VAN NOSTEAHD SCIENCE SEE! ISmo, Green Boards. Price 50 Cents Each. Demand. r _n n n n n n n n n i EPLACES . rong, C. E Chimneys = < . Zerah Col? 5^ NING WA CC ^ IN BRID( o >H v_/ flH y W. F. Bu Greenleaf. CC > < xC] f J ^RUCTION cc < ^ mr Jacob, , ORMS OF DQ U :e, C. E. ENGINE. "| ^ ^X ,' Vith Addit _J '- O v \ ^K W CO W UJ / ^ > wo fc ^ CO c V j, L., to whir VQ E ICIAL FU : ^^x^ L 3y John V v ^ P ilated fror ^^J s of Amerj -' dition. K CC -y ^ Allan. LJ .g 1. By Prof >' ^ i ^ $ FES. By J 1 ' & ^ To whic. z ?j JSTS. By u- - . is D u u u u u u--u u~u ^ I. J. Atkin^ 1 cle, C. E. ^ nj _ v _ u _ v 4 ^G CERT/ Prof. Geo. ; AJNU WATH:tt SUPPLY. By Prof. W. Corfield, M. A. No. 18.-SEWERAGE AND SEWAGE UTILIZATION, Prof. W. H. Corfield. No. 19.-STRENGTH OF BEAMS UNDER TRANSVER LOADS. By Prof. W. Allan. No. 20. BRIDGE AND TUNNEL CENTRES. By John McMasters, C. E. No. 21. SAFETY VALVES. By Richard H. Buel, C. E. THE VAN NOSTBAND SCIENCE SERIES. [22. -HIGH MASONRY DAMS. By John B McMaster. , 23.-THE FATIGUE OF METALS UNDER REPEATED STRAINS, with various Tables of Results of Ex- periments. From the German of Prof. Ludwig Spangenberg. With a Preface by S. H. Shreve i 24. A PRACTICAL TREATISE ON THE TEETH OF WHEELS, with the Theory of the Use of Robin- son's Odontograph. By Prof. S. W. Robinson. 25. THEORY AND CALCULATIONS OF CONTINU- OUS BRIDGES. By Mansfield Merriman, C. E. ).- PRACTICAL TREATISE ON THE PROPERTIES OF CONTINUOUS BRIDGES. By Charles Bender. ; .27. ON BOILER INCRUSTATION AND CORROSION- By F. J. Rowan. 28. -ON] TRANSMISSION OF POWER BY WIRE KOPES. By Albert W. Stahl. :9. -INJECTORS ; THEIR THEORY A$D USE. Trans- lated from the French of M. Leon Pouchet. TERRESTRIAL MAGNETISM AND THE MAGNET- / ISM OF IRON SHIPS. By Prof. Fairman Rogers. 31.-THE SANITARY CONDITION OF DWELLING HOUSES IN TOWN AND COUNTRY. By George MAKING FOR SUSPENSION BRIDGES, as exemplified in the construction of the East River Bridge. By Wilhelm Hildenbrand, C. E. 3. MECHANICS OF VENTILATION. By George W. Rafter, C. E. 14. FOUNDATIONS. By Prof. Jules Gaudard, C. E. Translated from the French. : 35. THE ANEROID BAROMETER: Its Construction and Use. Compiled by Prof. G. W. Plympton. 3d Edition 16. -MATTER AND MOTION. By J. Clerk Maxwell 37. GEOGRAPHICAL SURVEYING : its Uses, Methods and Results. By Frank De Yeaux Carpenter, i 38. MAXIMUM STRESSES IN FRAMED BRIDGES By Prof. Wm. Cain. 9.-A HANDBOOK OF THE ELECTROMAGNETIC TELEGRAPH. By A. E. Loring, a Practical Tel- egrapher. 2d Edition. [ 40. TRANSMISSION OF POWER BY COMPRESSED AIR. By Robert Zabner, M E. 41. -STRENGTH OF MATERIALS. By William Kent. 1 42. VOHSSOIR ARCHES, applied to Stone Bridges, Tun nels, Culverts and Domes. By Prof. Wra. Cain. ; 43.-WAVE AND VORTEX MOTION. By Dr. Thomas Craig, of Johns Hopkins University. OF EARTHWORK. BY BENJAMIN BAKER, M. Inst. C.E From Excerpt Minutes of the Proceedings of the Institution of Civil Engineers. REPRINTED FROM VAN NOSTRAND'S MAGAZINE. NEW YORK: D. VAN NOSTRAND, PUBLISHER, 23 MURRAY AND 27 WARREN STREET. 1881. OF THE UNIVERSITY PREFACE THE much discussed subject of the pressure of earthwork is in this essay so exhaustively treated that nothing is left to be desired. The engineer may find here a satisfactory explanation of the causes of the discrepancies between theory and practice, and of the differ- ences between different authorities. It was originally presented as a paper to the Institution of Civil Engineers, from the published minutes of which the essay as here presented was published in VAN NOSTJRAND'S MAGAZINE. Abstracts only of the discussion are presented here. OF THE UNIVERSITT The Actual Lateral Pressure of Earthwork. THE fact that a mass of earthwork tends to assume a definite slope, and that if this tendency be resisted by a wall or any other retaining structure, a lateral pressure of notable severity will be ex- erted by the earthwork on that structure, must have enforced itself upon the at- tention of constructors in the earliest ages. Many of the rudest fortresses doubtless had revetments, and of the hundreds of topes, or sacred mounds, raised in India and Afghanistan two thousand years ago, not a few afford ex- amples of surcharged retaining walls on as large a scale as those occurring in modern railway practice. Nevertheless, long as the subject has occupied the at- tention of constructors, there is proba- bly none other regarding which there ex- ists the same lack of exact experimental data, and the same apparent indifference as to supplying this want. Thousands of pieces of wood have been broken in all parts of the world to determine the transverse strength of timber, whilst the experiments that have been undertaken to ascertain the actual lateral pressure of Earthwork are hardly worth enumerating* One authority after another has simply evaded the task of experimental investi- gation, by assuming that some of the elements affecting the stability of earth- work are so uncertain in their operation as to justify their rejection, and have so relieved themselves from further trouble. It would hardly be less logical to assume that because timber is liable to become rotten and possesses no strength at all, it was therefore unnecessary to conduct experiments in that case also. As a mat- ter of fact, although these uncertain ele- ments are neglected in investigations, engineers in designing, and still more contractors in executing, works, do not neglect them, nor could they do so with- out leading to a blameworthy waste of OP THE money in some instances, and to a dis- creditable failure in others. The result of the present want of experimental data is then simply that individual judgment has to be exercised in each instance, with- out that aid from careful experimental investigation which in these times is en- joyed in almost every other branch of engineering. The mass of existent literature on the subject is both misleading and disap- pointing, for with little exception the bulk of it consists merely of arithmeti- cal changes rung upon a century-old theory, which even at the time of its in- ception was put forward but as a provi- sional approximation of the truth, pend- ing the acquirement of the necessary data. Writing some fifty years ago Pro- fessor Barlow excused his u very imper- fect sketch of the theory of revetments, . at least as relates to its practical applica- tion," on the ground that there was a " want of the proper experimental data ;" and but comparatively the other day Pro- fessor Eankine had to write in almost identical terms : " There is a mathematical theory of the combined action of fric- tion and adhesion in earth ; but for want of precise experimental data its practi- cal utility is doubtful." It is not, there- fore, for want of asking that the missing data are not forthcoming. Indeed, the present desiderata could not have been more clearly formulated than they were half a century ago by Professor Barlow in the following words : " To render the theory complete, with respect to its practical application, it is necessary to institute a course of experiments upon a large scale ; upon the force with which different soils tend to slide down when erected into the form of banks. A well-conducted set of experiments of this kind would blend into one what many writers have divided into sev- eral distinct data. %Thus some authors have considered first, what they call the natural slope of different soils, by which they mean the slope that the surface will assume when thrown loosely in a heap; very different, as they sup- 9 pose, from the slope that a bank will as- sume that has been supported, but of which that support has been removed or overthrown. This, therefore, leads to the consideration of the friction and cohesion of soils, and what is denomina- ted the slope of maximum thrust ; but, however well this may answer the pur- pose of making a display of analytical transformations, I cannot think it is at all calculated to obtain any useful prac- tical results. I should conceive that a set of experiments, made upon the abso- lute thrust of different soils, which would include or blend all these data in one general result, would be much more use- ful, as furnishing less causes of error, and rendering the dependent computa- tions much more simple and intelligible to those who are commonly interested in such deductions." A knowledge, however imperfect, of the actual lateral pressure of earthwork, as distinguished from what may be termed the " text-book " pressures, which, with hardly an exception known 10 to the author, are based upon calcula- tions that disregard the most vital ele- ments existent in fact, is of the utmost importance to the engineer and con- tractor. It affects not merely the stabil- ity of retaining walls, but the strength of tunnel linings, the timbering of shafts, headings, tunnels, deep trenches for re- taining walls, and many other works of every-day practice. The vast divergence between fact and theory has perhaps im- pressed itself with peculiar force upon the author, because, having had the privilege of being associated with Mr. Fowler, Past President of thelnst. C.E., during the whole period of the construc- tion of the " underground " system of railways, he has had the advantage of the experience gained in constructing about 9 miles of retaining walls, and, in relation to the subject of the present paper, the still more valuable experience of 34 miles of deep-timbered trenches for retaining walls, sewers, covered ways, and other structures. A timber waling is a sort of spring, rough it may be, but 11 still the deflections when taken over a sufficiently large number of walings afford an approximate indication of the pressure sustained an advantage which a retaining wall does not possess. Again, though numberless retaining walls have failed, in ninety-nine cases out of hun- dred the failures have been due to faulty foundations, and, consequently, experi- ences of this sort seldom afford any di- rect evidence as to the actual lateral pressure of earthwork. In timbered trenches, on the other hand, the element of sinking and sliding foundations does not so frequently arise to complicate the investigation. All kinds of earth were traversed by the above 34 miles of trenches, from light vegetable refuse to the semi-fluid yellow clay, which at different times has crushed in so many tunnel linings in the northern districts of the metropolis. The heights of the retaining walls ranged up to 45 feet, the depths of the timbered trenches to 54 feet, and the ground at the back of the former was in many 12 cases loaded with buildings ranging up to 80 feet in height. Possibly some of the author's observations and conclusions in connection with these and other works of a similar character may be of interest to engineers, though the information he is able to contribute, having been ob- tained chiefly in the ordinary routine of his practice and not in specially devised investigations, must necessarily form but a very imperfect contribution to the data which have been asked for so long. . The* theory underlying all the multi- tudinous published tables of required thickness for retaining walls is, that the lateral pressure exerted by a bank of earth with a horizontal top is simply that due to the wedge-shaped mass, included between the vertical back of the wall and a line bisecting the angle between the vertical and the slope of repose of the material. If this were true in practice, all such problems could be solved by merely drawing a line on the annexed diagram, in which a b d c is a square, a b g a triangle, having the sides of the 13 ratio of 1 : A/J, and a h d a parabolic curve.* --70 * For earthwork and masonry of the same weight per cubic foot the equation for stability is : Tit* h* 4 ~-=-j-taai } angle. Hence, the required thickness (t\) in terms of the height (K) will be t\ ^^ tan y 2 angle, which is repre- sented on the diagram by the line a g ; and the "equivalent fluid pressure " in terms of that of a cu- bic foot earthwork will be = tan a % angle, which is represented by the parabolic curve a h d. OF THE XJNIVERSIT1T 14 Thus, if it were required to know the lateral pressure per square foot of earth- work, having a slope of repose of 1^: 1, and the thickness of rectangular verti- cal wall which, when turning over on its outside edge would just balance that pressure, it would merely be necessary to draw the line c f at the given slope of 1 : 1 and the line c e bisecting the angle a c k, when the line e h would give the equivalent fluid pressure in terms of that of a cubic foot of the earth =28. 7 per cent., and the line e i the thickness of the rectangular wall in terms of the height =31 per cent , the weight of ma- sonry being the same as that of the earth. Common stocks in mortar and ballast backing each weigh about 100 Ibs. per cubic foot, hence, on the preceding hy- pothesis, the pressure acting on the wall would be the same as that due to a fluid weighing 28.7 Ibs. per cubic foot. If, as is usually the case, the masonry be heavier than the earthwork, the required thickness of wall would be reduced in 15 inverse proportion to the square root of the respective weights, so that should the masonry weigh 10 per cent, more than the ballast, the thickness would be about 5 per cent, less than before, or> say, 29.5 per cent, of the height. For other slopes of repose the equiva- lent fluid pressure and thickness of wall for materials of equal weight would be as follows : Ratio of horiz- ) ~ ontal to vertical f .6 .7 .8 .9 1.0 Fluid pressure. . . 5.6 7.7 10 12.4 14.8 17.2 Thickness . 136 160 .182 903 222 939 Ratio of horiz- [ ontal to vertical ) 1.1 1.2 1.3 1.4 1.5 1.6 Fluid pressure. .. 19.6 22 24.3 26.5 28.7 30.7 Thickness .... 956 M 984 997 31 39, 16 Ratio of horiz- ) ontal to vertical f Fluid pressure . . . 1.7 1.8 2 3 52 4 61 00 100 32.8 34.6 38.2 Thickness .33 .34 .357 .416 .451 .578 In the thickness tabulated above no allowance has been made for the crush- ing action on the outer edge ; in practice the batter usually given to the face of the wall more than compensates for this action if the mean thickness be that given in the table. No factor of safety is included, but according to theory the wall in each case would 'be just on the balance. Any one accustomed to deal with works of this class will, however, know that in practice walls so propor- tioned would in the majority of cases possess a large factor of safety. Doubtless many engineers will, with the author, have noticed that laborers and others not infrequently carry out unconsciously a number of valuable and 17 suggestive experiments on The Actual Lateral Pressure of Earthwork. In stacking materials, rough and-ready re- taining walls, made of loose blocks of the same material, are often run up, and as it is generally of little moment whether a slip occurs or not, the work- men do not trouble about factors of safety, but expend the least amount of labor that their every-day experience will justify, and so a tolerably close measure is obtained of the average actual press- ure of material retained. When the wood paving was recently laid in Eegent Street, the space being limited, the stacked wooden blocks in many cases had to do duty as retaining walls to hold up the broken stone ballast required for the concrete substructure. In one in- stance (Ex. 1) the author noted that a wall of pitch-pine blocks, 4 feet high and 1 foot thick, sustained the vertical facef of a bank of old macadam materials which had been broken up, screened, and tossed against this wall until the bank had at- tained a height of 3 feet 9 inches, a 18 width at the top of about 5 feet, and slopes on the farther sides deviating little from 1.2 to 1. Now, referring to the diagram and table of thickness, it will be seen that according to the ordi- nary theory the thickness of wall which would just balance the thrust of a bank 3 feet 9 inches high of material haying a slope of repose of 1.2 to 1 would be 3.75 X. 27= 1.01, or, say, 1 foot, which is the actual thickness of the given wall. But in the table the specific weight of the material in the wall and backing is assumed to be the same, whereas in the present case the weight of the pitch-pine block wall, allowing for the height being greater than that of the bank, would 4 feet only be, say, 46 Ibs. X . =3 Ibs. per cubic foot, whilst that of the broken granite bank would be, say, 168 Ibs. less 40 per cent, for interstices = 101 Ibs. per cubic foot. It follows, since the wooden wall stood, that if it had been made of materials having the same weight per cubic foot as the bank, the retaining wall OP THE 19 would not have been on the point of toppling over, as the ordinary theory would indicate, but have possessed a factor of safety of at least - , or, 4y lus. say, 2 to 1. The effective lateral press- ure of the earthwork in this instance consequently could not have exceeded a fluid pressure of - - = 10.7 Ibs. per cubic foot, instead of the 22 Ibs., which theoretically corresponds to the given slope of 1.2 to 1. Taking another case, in which the wall, instead of being lighter than the bank, was much heavier, the same con- clusion still holds good. In this instance (Ex. 2) the author found a wall of slag blocks having a batter of \ of the height, and an effective thickness of 1 foot sus- tained a bank of broken slag 10 feet high, with a surcharge of some 5 feet more. The battering wall, with a thick- ness of ^ of the height, would have the same stability as a vertical wall 0.173 thick, and the lateral pressure of the sur- 20 charged bank with the battering face would be practically the same as that of a horizontal-topped bank with a vertical face; hence, since the relatively closely- packed slag blocks constituting the wall would weigh about 40 per cent, more than the broken slag of the bank, the thickness of a vertical wall built of ma- terials of the same weight as the bank, and having the same stability as the wall under consideration would be = Vl 4 X 0.173 = 0.205 of the height. Referring to the table, the figure 0.205 will be found to apply to a slope of repose of 0.8 to 1, whereas the actual slope in the instance of this slag was 1.33 to 1. For the latter slope the thickness theoret- ically should have been 0.29, and since the stability varies as the square of the thickness, it follows that with the thick- ness indicated by theory, the wall, in- stead of being just on the balance, would have possessed a factor of safety of at 0.29'" least ^ 2, or 2 to 1, as in the last O.^Oo example. 21 Other instances of these unintentional experiments on the lateral pressure of earthwork will be found in the stacking of coal in station yards, in the rubbish banks at quarries, and in many other instances which have been investigated by the author, with the invariable result of finding that walls which, according to current theory, would be on the point of failure, really possess a considerable factor of safety. Turning now from indirect to direct experiments, specially arranged with a view to determine the lateral pressure of earthwork, those carried out at Chatham nearly forty years ago by Lieutenant Hope, K.E., may be referred to. His intention was to experiment first with fine dry sand, as free as possible from the complications introduced by cohesion, irregularities of mass and other practical conditions, and then to extend the in-* vestigation to ordinary shingle, and to clay and other soils possessed of great tenacity. Sand and shingle were, how- ever, alone experimented with. 22 The direct lateral thrust of sand weighing 91 Ibs. per cubic foot when lightly thrown together, and 98 J Ibs. when well shaken, was measured by balancing the pressure exerted on a board 1 foot square. The mean results of seven experiments (Ex. 3) was 9 Ibs. 7 oz., which is that due to a fluid weigh- ing nearly 19 Ibs. per cubic foot. As the slope of repose of the sand employed was 1.42 to 1, the theoretical fluid press- ure due to the weight of 98J Ibs. per cubic foot wou]d be 26.2 Ibs., or about 40 per cent, more than the observed 19 Ibs. per cubic foot. With gravel (Ex. 4) weighing 95J Ibs. per cubic foot, and having a slope of re- pose of 1J to 1, about the same lateral pressure was found to exist. Lieutenant Hope attempted to reconcile the differ- ence between theoretical and actual re- sults by adding to the measured force an estimated sum for friction against the sides of the apparatus, but experiments of the author's to be subsequently refer- red to, clearly prove that the difference 23 is not to be so accounted for. Indeed, the knowledge of what the pressure theo- retically should be would appear to have given Lieutenant Hope an unconscious bias in the direction of rather exaggerat- ing the experimental results. This it is extremely easy to do, as a trifling amount of vibration will alter the pressure from 10 to 50 per cent., and a comparatively innocent shake in a small model will cor- respond in its relative effects with an earthquake in real life. Experiments with colored sand in a vessel with glass sides did not uniformly confirm the usual theory that the angle of pressure of maximum thrust is half that contained between the natural slope and the back of the wall (Ex. 5). Thus the line of separation was at an angle of 24 with the vertical instead of 28. Again, with a gravel bank (Ex. 6) 10 feet high the line of separation ranged from ' 3 feet 8 inches to 5 feet 8 inches from the back of the wall, whilst as the natural slope was 1J to 1, the distance should have been 5 feet in all instances if Cou- 24 lomb's theory applied strictly to even such exceptionally favorable materials as dry sand and shingle. The really valuable portion of Lieu- tenant Hope's investigation was the series of experiments on walls built of bricks laid in wet sand. The first of these (Ex. 7) was about 20 feet long and two-and-a-half bricks, or say, 1 foot 11 inches thick. When raised to a height of 8 feet and backed with ballast, it had in- clined from the vertical about 1-J inch; at 9 feet the inclination had increased to 3J inches, and at 10 feet the wall fell for- ward in one mass. At the instant when the thrust of the ballast overcame the stability of the wall, the overhang must have been 4 inches, and the moment of stability per lineal foot certainly not more than 2,000 Ibs.x0.9 foot=l,800 foot-pounds. Hence, dividing by , is obtained 10.8 Ibs. per cubic foot as the weight of the fluid, which would have exerted a lateral pressure equal to that of the ballast piled against this 10-feet 25 wall. This is hardly more than half the pressure obtained with the 1-foot square board, and shows how desirable it is that even the most faithful experimenter should not know what to expect if a mere shake of a table will enable him to obtain the desired result. The natural slope of the ballast being 1J to 1, and the weight 95 ^ Ibs. per cubic foot, the pressure theoretically should have been 23.6 Ibs. per cubic foot instead of 10.8 Ibs ; hence a wall so proportioned as to be on the point of toppling over, accord- ing to the ordinary theory, would in this instance have had a factor of safety of rather more than 2 to 1. Another vertical wall (Ex. 8) was con- structed with the same amount of ma- terials differently disposed. At 8 feet high, after heavy rain, the 18-inch thick panel between the 27-inch deep counter- forts had bulged 1J inch; at 12 feet 10 inches the bulging had increased to 4 J inches, and the overhang at the top to 7 inches, when, after some hours' grad- ual movement, the wall fell. The moment 26 of stability at the time of failure could not have exceeded 2,600 Ibs. X 1 foot = 2,600 foot pounds, which, divided by A 8 , gives 7.4 Ibs. per cubic foot, instead of the theoretical 23.6 Ibs., as the weight of the equivalent fluid. This result is clearly not evidence that the pressure of the ballast was less in the counterforted wall than in the wall of uniform thick- ness, but that the binding of the ballast between the counterforts increased the stability of the wall by practically add- ing somewhat to its weight. A wall with a batter of ^ of the height, and with counterforts of the same thick- ness as the last (Ex. 9), was next tried, with noteworthy results. This wall, only 18 inches thick, with counterforts 3 feet 9 inches deep, measuring from the face of the wall, and 10 feet apart, was carried to a height of 21 feet 6 inches without any indications of movement, beyond a bulging about halfway up of 2^ inches at the panel, and 1J inch at the counterfort; and in Lieutenant Hope's 27 opinion it would probably have stood for years without giving way any more, al- though the mean thickness was less than ^ of the height. The calculated stabil- ity indicates that a fluid pressure of 8.5 Ibs. per cubic foot would have overturned the wall, and, correcting for the reduced thrust of the ballast due to the batter of its face, the equivalent pressure on a vertical wall would be that of a fluid weighing 10 Ibs. per cubic foot. Here, again, doubtless the binding of the gravel between the counterforts con- tributed to the stability of the wall ; but, even adopting the extreme and impossi- ble hypothesis that the ballast was as good as so much brickwork, or, in other words, that the wall was a monolithic structure of the uniform thickness of 3 feet 9 inches, its stability would barely balance the 23.6 Ibs. per cubic foot fluid pressure theoretically due to the weight and slope of repose of the back- ing. Assuming that the binding of the ballast between the counterforts in- creased the stability, as in Examples 8 and 28 9, by about 45 per cent., the fluid resist- ance would be 14.5 Ibs. per cubic foot ; and, remembering that this wall did not fall, though the bricks were only laid in sand, it is reasonable to infer that this interesting experiment confirms the pre- vious conclusion that a properly built wall in mortar or cement, just balancing the theoretical pressure, would really have had a factor of safety of 2 to 1. Other experiments of Lieutentant Hope's justify this inference, and so do the ex- periments of General Pasley, also made at Chatham many years ago. General Pasley experimented with loose dry shingle weighing 89 Ibs. per cubic foot, and having a natural slope of 1J to 1. His model retaining walls (Ex. 10) were 3 feet long, 26 inches high, of various forms and thickness, and weighed 84 Ibs. per cubic foot. The stability of each wall was tried by pulling it over by weights before and after backing it up with shingle, and the difference between the two pulls of course represented the thrust of the shingle. When the thick- 29 ness of the vertical wall was 8 inches, the stability, without shingle, was equiv- alent to a pull of 47 Ibs. applied at the top of the wall, and with shingle, the pull required to upset it was reduced to 30 Ibs. The difference of 17 Ibs. repre- sents the thrust of the shingle, and throughout the several hundreds of ex- periments this appears to have been com- prised within the limits of 16 Ibs. and 24 Ibs. The center of pressure being at J of the height of the wall, the mean thrust of 20 Ibs. at the top will be equiv- alent to 60 Ibs. at the center of pressure, and the area being 6.5 square feet, and the height 26 inches, the actual lateral pressure of the shingle, as deduced from General Pasley's experiments, is equiva- lent to that of a fluid weighing 8.5 Ibs. per cubic foot, instead of 21 Ibs. as theory would indicate. General Cunningham tested some model revetments, and his experiments led him to believe that General Pasley had overestimated the thickness required for stability. The 30 about 30 inches in height, were weighted with earth and musket bullets to the equivalent of an equal mass of masonry weighing 129 Ibs. per cubic foot. One of the models (Ex. 11) represented a wall 30 feet high, 6 feet thick at the base, vertical at the back, battering 1 in 10 on the face, with counterforts 4 feet 3 inches thick, 18 feet from center to center, and of a depth equal to the thickness of the wall or, say, 3 feet at the top and 6 feet at the base. This was backed up and surcharged with shingle weighing 104 Ibs. per cubic foot, but required a pull of 111 Ibs. to overturn it. Another model (Ex. 12) representing a wall 18 feet high, 4 feet 4 inches thick at the base, and 2 feet 8 inches thick at the coping, without counterforts, when sur- charged with shingle to a height great- er than that of the wall, required a pull of 84 Ibs. to upset it. A fluid press- ure of 19 Ibs. per cubic foot would over- come the stability of such a wall ; hence, having regard to the surcharge and to the pull, it will be found that the actual 31 lateral pressure of the shingle could not have exceeded that due to a fluid weigh- ing 8 Ibs. per cubic foot. General Burgoyne also commenced an experimental investigation of the ques- tion of retaining walls, but circumstances precluded his pursuing the subject. About half a century ago he built at Kingstown four experimental walls 20 feet long and 20 feet high, having the same mean thickness of 3 feet 4 inches, or ^ of the height, but differing otherwise. One of them (Ex. 13) was of the uniform thickness of 3 feet 4 inches, and battered -^ of the height ; another (Ex. 14) was 1 foot 4 inches thick at the top, and 5 feet 4 inches at the bottom, with a vertical back; the third (Ex. 15, Fig. 1; was of the same dimensions, with a vertical front ; and the last (Ex. 16, Fig. 2) was a plain rectangular vertical wall 3 feet 4 inches thick. The masonry consisted simply of rough granite blocks laid dry, and the filling was of loose earth filled in at random, without ramming or other precautions, during a very wet winter. 32 No. 1 wall stood perfectly, as might have been expected from the behavior of Lieu- tenant Hope's experimental wall of near- ly the same height and batter. No. 2 wall also stood well, coming over only Fig.1 Fig-2 abont 2 inches at the top. A fluid press- ure of 22.5 Ibs. per cubic foot would be required to overcome the stability of this dry masonry wall weighing 142 Ibs. per cubic foot. Earthwork of the class de- 33 scribed, consolidated during continuous rain, would not weigh less than 112 Ibs. per cubic foot, nor have a slope of re- pose less than 1 to 1. Referring to the table, the theoretical pressure of such earthwork would be 28.67x1.12 = 32 Ibs. per cubic foot, or nearly one-half greater than the wall could resist. No. 3 and No. 4 walls both fell when the filling had attained a height of 17 feet. The former came over 10 inches at the*top, was greatly convex on the face, overhanging 5 inches in the first 5 feet of its height and rending it in every direc- tion, when finally it burst out at 5 feet 6 inches from the base, and about two- thirds of the upper portion of the wall descended vertically until it reached and crushed into the ground (Fig. 1). The vertical wall tilted over gradually to 18 inches and then broke across, as it were, at about ^ of its height and fell forward (Fig. 2). So long as the wall remained vertical the calculated stability would in- dicate it to be equal to sustain the press- ure of a fluid weighing 20.4 Ibs. per 34 cubic foot, but the overhang of 18 inches and the bulging which occurred would reduce the stability exactly one-half, so that a fluid pressure of 10.2 Ibs. would really have sufficed to effect the final overthrow. The character of the failure both of No. 3 and No. 4 walls clearly in- dicates that if the walls had been in mortar or cement, as usual, the overhang would not have been a fraction of that occurring with the dry stone walling, and the failure would not have faken place. Since, as already stated, the theo- retical thrust of the earthwork would be 3;21bs. per cubic foot, it is hardly unfair to conclude that a wail in mortar and pro- portional to that pressure would not have come over and would have enjoyed a fac- tor of safety of at least 2 to 1. Colonel Michon carried out in 1863 an interesting experiment (Ex. 17) on a 40 feet high retaining wall of a peculiar type (Figs. 3 and 4), which, perhaps, may be best described as a very thin w&ll with numerous battering buttresses turned upside down. The face wall, battering 35 1 in 20, was only 1 foot 8 inches thick, and the buttresses, spaced about 5 feet apart from center to center, were also 1 Fig.4 foot 8 inches thick by 2 feet 4 inches deep at the base and 9 feet 2 inches at the top. The work was hurriedly con- 36 s true ted during continuous rains with any stones that came to hand, and with very bad lime. When the filling had attained a height of 29 feet the wall bulged a trifle, but no further movement was noticed, though the filling, when carried up to the top of the coping, was allowed some weeks to settle in the rain. Earth was then piled above the level of the coping to a height of be- tween 3 and 4 feet, when the wall fell. The fall was preceded by a general dis- location of the masonry at the base, a bulging at about one-third of the height, and a slight movement of the top to- wards the bank. The lower portion of the wall fell outwards, the upper part dropped vertically (as in General Bur- goyne's wall, Fig. 1), and a considerable number of the counterforts went for- ward with the slip and even maintained their vertical position. This failure arose from a flexure of the thin wall at the center of pressure of the earthwork, and would not have oc- curred had the masonry been in cement 37 instead of in weak unset lime. No direct data therefor are afforded for an exact estimate of the actual lateral thrust of the heavy wet filling on this lofty wall. Nevertheless, as the weight of the ma- sonry was only 18,000 Ibs. per lineal foot, and the center of gravity of the same from the toe but 6 feet 6 inches, it follows that the wall, even if monolithic, would be overturned with the pressure of a fluid weighing 11 Ibs. per cubic foot. How far the sodden earthwork between the counterforts contributed to the sta- bility of the wall is open to question, but it could hardly account for the differ- ence between the 11 Ibs. or less stability and the 32 Ibs. due, according to the or- dinary theory, to the weight and slope of the backing. If dirt were as good as masonry, General Burgoyne's wall with the battering back (Fig. 1) would have been more stable than the vertical wall (Fig. 2) in the ratio of the squares of their respective bases, or, say, as 2^ tol, whereas, these walls proved to be of equal stability, both falling with 17 feet 38 of filling. Colonel Michon, by assuming dirt to be as good as masonry, and a wall 40 feet high and 1 foot 8 inches thick of unselected stones and unset mortar to be as good as a monolith, succeeds in recon- ciling the behavior of his wall with the ordinary theory of the stability of earth- work ; but in the author's experience the conditions assumed are not approached in practice. The stability of this lofty wall battering only -fa of the height on the face, and averaging hardly more than T *2 of the height in thickness, is, never- theless, one of the most remarkable and interesting facts connected with the sub- ject of the present paper. To show how invariably an experi- mentalist is driven to the same conclu- sion as to the excess in the theoretical estimate of the pressure of earthwork, the " toy " experiments of Mr. Casimer Constable with little wooden bricks and peas for filling may be usually referred to. The peas took a slope of 1.9 to 1, and weighed twice as much per cubic foot as the wooden retaining wall. By 39 the table the thickness of wall, which would just balance the lateral pressure would be .35\/2 = .49 height. By ex- periment, a wall ( Ex, 18) having a thick- ness of .40 height moved over slightly, but took some amount of jarring to bring it down. Since the stability varies as the square of the thickness, the calcula- ted wall would be 50 per cent, more sta- ble than the actual wall, without consid- ering the question of jarring. If the slope of the peas had been measured also, after jarring, it would probably have been found to be nearer 2.9 to 1 than 1.9 to 1, and the calculated required thickness would have been correspondingly in- creased. The influence of even a slight amount of vibration is well illustrated by the difference between the co-efficient of friction of stones on one another in mo- tion and repose. Granite blocks, which will start on nothing flatter than 1.4 to 1, will continue in motion on an incline of 2.2. to 1, and, for similar reasons, earth- work will assume a flatter slope and ex- 40 ert a greater lateral pressure under vi- bration than when at rest. This fact has long received practical recognition from engineers ; indeed, attention was called to it by Mr. Charles Hutton Gregory, C. M.G., Past-President Iiist. C.E., in a Pa- per on slips in earthwork, read before the Institution in 1844, when the Presi- dent and others gave instances of slips in railway cuttings caused by vibration; The general results of the preceding and other independent experiments on retaining walls tending to throw a doubt on the accuracy of Lieutenant Hope's measurement of the direct lateral thrust of ballast and sand on a board 1 foot square, the Author considered it advisa- ble to repeat those experiments. Care was taken to eliminate all disturbing causes tending to vitiate the results. The pressure board was held by a string at its center of pressure, and was perfectly free to move in every direction, which, of course, a retaining wall having a greater hold on the ground than stability to re- sist overturning has not. In every in- 41 stance the filling was poured into the bo;< and allowed to assume its natural slope towards the pressure board, and the lat- ter was rotated and thumped to keep the ballast alive before the reaction was meas- ured. In order to avoid all chance of the bias which the knowledge what to expect might have given him, as it did Lieuten arit Hope, the author had the experiment 8 made by others who were ignorant even of the object of them, whilst he himself purposely experimented with an appara- tus the dimensions of which he did not know, and consequently could form no estimate of the weight which would be required in the scale With clean dry ballast having a natu- ral slope of 1 J to 1, it will be remembered Lieutenant Hope obtained a lateral press- ure on 1 square foot of 9 Ibs. 7 oz. With well-washed wet ballast of the same kind the author found the natural slope to be 1 to 1, and he decided therefore to use the ballast wet, because, possessing greater fluidity, it would give more uni- form results than dry ballast, and also 42 impose greater lateral pressure. In a, large number of independent experiments the results were uniformly as follows (Ex. 19) : With 6 Ibs. in the scale the board moved forward about ^ inch, but continued to retain the ballast ; with 7 Ibs. very slight movement occurred ; with 8 Ibs., no movement at all ; and with 10 Ibs., under extreme vibration, the board moved forward about as much as it did with 6 Ibs. without vibration. The gen- eral opinion of the different experiment alists was, that the fair value of the lat- eral pressure of this wet ballast was 7 Ibs., because when that weight was in the scale pad a slight jolt was sufficient to let the ballast down by the run to a slope of 1 to 1. The board being 1 foot square, this of course is equivalent to the press- ure of a fluid weighing 14 Ibs. per cubic foot, instead of the 19 Ibs. obtained by Lieutenant Hope, and the 26 Ibs. indica- ted by theory. With the same ballast unwashed and mixed with slightly loamy pit sand (Ex. 20) the natural slope was 1 to 1, without 43 vibration, and 1 to 1 with a moderate amount of vibration. A weight of 3 Ibs. was as effective in retaining this ballast as 6 Ibs. in -the former instance ; 4 Ibs, held it under a moderate amount of vi- bration, but 3 Ibs. failed to hold the board under very little. Practically speaking, the lateral thrust was about half that with the clean wet ballast, and considerably less than half that theoreti- cally due to the slope of repose of the loamy ballast. In harbor works both walls and back- ing are frequently completely immersed, and, so far as gravity is concerned, stone blocks and rubble become then trans- formed into coal. The author, there- fore, experimented (Ex. 21) with some coal having a peculiarly "greasy" sur- face, and offering the advantage of ex- ceptional fluidity. With 3 Ibs. the board moved forward about 1 inch, but no more until a slight jar was applied, when it fell; with 4 Ibs. a moderate amount of vibration also generally caused failure ; with 5 Ibs. the board usually moved for- 44 ward gradually without making a rush so long as a tolerably considerable amount of vibration was maintained. When a slip occurred the slope was invariably If to 1. The coal proved to be more sensi- tive to vibration than the wet ballast, and still more so than the unwashed ballast. Aweight of 4 Ibs. with the coal appeared to be equivalent to 7 Ibs. with the washed and 3i Ibs. with the unwashed ballast- The weight of the coal being one-half that of the wet ballast, and the respective slopes of repose being If and 1^ to 1, the lateral thrusts would theoretically be as 16.8 : 28.7, which is practically the exper iinental result of 4 to 7. X The author having occasion -to design a solid pier 42 feet in height from the bottom of the harbor to the surface of the quay, where a soft bottom of great thickness and small consistency precluded the use of concrete block or other retain- ing wall, adopted an arrangement in which an iron grid of rolled joist, with a backing of large blocks of rubble, was substituted for a wall. It was necessary, 45 therefore, to know the lateral thrust of large blocks of stone in such a structure, and mistrusting theoretical deductions, the author made direct experiments on a model to a scale of 1 inch to the foot. In this instance the individual st< : ies were intended to be fairly uniform in size, and of lateral dimensions not less than 2 1 (T of the height of the wall, so that the conditions differed considerably from those assumed in theoretical investiga- tions. A number of billiard balls exactly superimposed in a tightly fitting box would exert no thrust though their slope of repose might be as flat as 3 to 1 ; and it is not quite clear how nearly or re- motely large boulders in an iron cage approximate to that condition. The stones used in the experiment were waterworn pieces of schistose rock hav- ing a "greasy" surface and a slope of repose of 1 to 1. This inclination was found by the author to obtain in natural slopes of all heights, from the pile of metalling by the roadside to the hills themselves. He ascended one slope of 46 14 to 1, over 500 feet in height, and found the balance was so nearly main- tained that a footstep at times would set many tons of stones in motion; and a few winters ago a couple of stones of the respective weights of 18 tons and 22 tons, descended this slope and acquired sufficient momentum to carry them across a road at the foot of the slope and on to the middle of the lawn in front of an ad- joining shooting lodge. The conditions of the material were thus favorable, as in the instance of the coal, for obtaining uniform results and a maximum lateral thrust. In order to exclude all possible influence from side friction, the length of the box was made four times the height of the wall. AJS the result of ten experiments (Ex. 22) it was found that a weight in the scale cor- responding to the pressure of a fluid weighing 10.2 Ibs. per cubic foot sufficed to retain the rubble, though the face planking moved forward slightly as the last few shovelfuls were thrown against it. The weight of the stone filling was 47 98 Ibs. per cubic foot or practically the same as that of the wet ballast last referred to; so the 10.2 Ibs., or say under slight' vibration 11 Ibs., in the present instance compares with the 14 Ibs. of the previous instance, and the dif- ference is a measure of the influence of large-sized, smooth-faced boulders as compared with ordinary ballast. With coal of the same size (Ex. 23) the equivalent weight of fluid was 6 Ibs. per cubic foot, which confirms the pre- ceding result when regard is had to the respective weights and slopes of repose of the two materials. The experiments collectively proved that a wall, which according to the ordinary theory would be on the point of being overturned by the thrust of a bank of big bould- ers, would in fact have a factor of safety of nearly 2 to 1. Having thus briefly reviewed some direct experiments on the actual lateral thrust of earthwork, the author proposes to revert to the consideration of indirect experiments, dealing first with a few of 48 those arising on the 34 miles of deep- timbered trenches and other works of the u underground" railway. In tunneling very valuable evidence was afforded of the direction of the line of least resistance in a mass of earth- work. From the coming down of the crown bars, the changing of props, the crushing of timber, the compression of green brickwork and other causes, a settlement of from 6 to 8 inches usually occurred overhead, with a general draw of the ground towards the working end of the tunnel and the formation of fis- sures,, attaining a maximum size where the line of least resistance cut the surface. Even when the settlement was slight, fissures were invariably observed in ad- vance of the working end, and in con- tinuous lines running parallel with the tunnel. The slope of these fissures was so uniformly at the angle of J to 1, meas- uring from the bottom of the excavation (Ex. 24, Fig. 5), that the resident engi- neer professed to be able to foretell with certainty where a building or fence wall, 49 standing over the tunnel, would crack most. Assuming this ^ to 1 to represent Coulomb's line of least resistance, then the corresponding natural slope of repose Fig.5 of the material would appear to be 1 to 1, which is considerably steeper than what it was in fact. There is nevertheless a closer accord than usual between theory and fact in 50 the instance of the several miles of fissures, which occurred during the con- struction of the tunnels of the Metro- politan railway. In other instances, such as the failure of ill- devised timbering, or the pushing forward of a retaining wall, by heavy clay pressing against its lower half, this accord was not always exhib- ited. In some cases no previous fissures have occurred, but a wedge of 1 to 1 has at once broken off and gone down with the timbering, whilst in others the fissure has appeared immediately at the back of the wall; indeed in one instance, for several consecutive weeks, the author was able to pass a rod, 15 feet long, be- tween the wall and the apparently unsup- ported vertical face of the ground behind it. The corresponding theoretical slope of repose would thus appear to be hori- zontal in one case and vertical in the other, which is sufficient evidence of the necessity of giving but a qualified assent . to any theoretical deduction affecting the line of least resistance in earthwork. A very fair notion of the relative in- tensity of lateral and vertical pressure in earthwork is often obtained in carrying out headings. The heading for the Camp- den Hill tunnel of the Metropolitan rail- way is a case in point (Ex. 25). The ground consisted of sand and ballast, heavily charged with water, overlying the clay through which the heading was driven, at a depth of 44 feet from the surface. After the heading had been completed some months, the clay became softened to the consistency of putty by the water which filtered through the numerous fissures, and the full weight of the ground took effect upon the set- tings. Both caps and side trees showed signs of severe stress throughout the en- tire length of the heading, and the occa- sional fractures in the roof and sides indicated that the timbers were propor- tionately of about the same strength, cr rather weakness. . The caps were of 14-inch square balks, with a clear span of 8 feet, and the sides of 10-inch square timber, with a clear span of 9 feet. Their respective powers of resistance per square 52 foot of poling boards, supported, would 14 s 10 3 therefore be as t : =3J : 1. 8 y Now if one thing is settled by experi- ence beyond all question, it is that the superficial beds of London Clay, sodden, as in the present case, with water, will not take a less slope of repose than 3 to 1. The average weight of the wet ground over the heading being about 1 cwt. per cubic foot, the theoretical lateral press- ure on the side trees, at a mean depth of 48 feet from the surface, would be (see table) = 48 X 0.52x1 cwt. 25 cwt. per square foot, and upon the caps =44x1 cwt. =44 cwt. per square foot, or 1.76 time greater. But the side trees, as has been seen, had on]y - of the strength o.5 of the caps, so the irresistible conclusion is that the actual lateral pressure of the earthwork in this instance did not exceed one-half of that indicated by theory.* It is readily shown that the full weight * See also " Zur Theorie des Erddrucks," Weyrauch Zeitschrift fiir Baukunde, vol. i., p. 192, 53 of the ground came upon the settings. Thus, assuming it to do so, the weight upon the caps would be =44 cwt. X 8 feet clear span X 3. 5 feet distance apart of the settings = 1,232 cwt., and taking the effective span at 9 feet, the breaking weight, upon the basis of Mr. Lyster's experiments on balks of similar size and 2 X 14 8/ 'X 2.03 cwt. quality, would be -=-? = 9 1,240 cwt; hence the occasional fractures of the balks are fully accounted for. In- deed, the heading would have entirely collapsed, in the course of time, had not the roof been supported by intermedi- ate props practically quadrupling its strength. In the early days of the construction of the Metropolitan railway, a definite type of timbering had not been arrived at, and some remarkably light systems were tried at times. The lightest the author remembers was the timbering of the 14-feet wide gullet at Baker Street station (Ex. 26). Here the soil cut through was made up of about 8 feet of 54 yellow clay and gravel, 7 feet of loamy sand, 7 feet of sharp sand and gravel, full of water, and 4 feet of London clay at the bottom of the gullet. The timber- ing of the lower half consisted of 9-inch by 3-inch walings, 3 feet apart from center to center, in 12-feet lengths, with f-inch poling boards at the back. With one-half the distributed breaking load, the deflection of this 3-inch deep beam at the span of 12 feet would be at least 4 inches, whilst the ultimate deflection would be measured by feet. As the walings did not bend nearly as much as 4 inches, it will be a liberal estimate to assume that the actual lateral pressure of the earthwork was equal to half the dis- tributed breaking weight of the wal- ing. Having reference to the quality of the timber, this may be estimated at 3 2 "x9"x2.6cwt. __ _ _ = 17.6 cwt. ; and since 12 the area of the poling boards supported by each waling was 36 square feet, it follows that the lateral pressure of the earthwork could not have exceeded 55 55 Ibs. per square foot. But the depth of the bottom waling below the surface was 23 feet, or, neglecting the clay, and taking only the sharp sand and ballast charged with water, the depth would still be 20 feet, and the weight of fluid corresponding to the 55 Ibs. per square foot pressure no more than 2.75 Ibs. per cubic foot. It will be remembered that the natural slope of the sand and ballast in Lieuten - ant Hope's retaining- wall experiments was about 1J to 1, and that the actual and theoretical corresponding fluid pressures were respectively 10.3 Ibs. and 23.6 Ibs. per cubic foot. In the case of the gullet, the natural slope of the ballast and sand would similarly be not less than 1J to 1, and yet the fluid pressure could not have exceeded 2.75 Ibs. This one fact, therefore, is sufficient to prove that the universal assumption of the pressure of earthwork being analogous to thut of fluid, and proportional to the depth, is one of convenience rather than truth. The explanation of the singularly small 56 lateral thrust of the ballast in the present case is to be found in the fact that the ballast was lying between, and partially held back by, the two relatively tenacious layers of loamy sand and clay. As an extreme example of the same kind of action (Ex. 27), the author may state that he once applied to a wooden box full of sand a pressure equivalent to a column of that material 1,400 feet high before the box burst. On the fluid hypothesis, the lateral pressure would have been 1,400 feet X 23.6 Ibs., or about 15 tons per square foot; but of course a few Ibs. would have burst the box, and the sand was retained by being jammed between the bottom and lid of the little deal box the equivalents of the tenacious strata in the gullet. In shafts the stress on the timbering- is far less than in a continuous trench or heading, by reason of the frictional ad- hesion and tenacity of the adjoining earth. Thus (Ex. 28) at a depth of be- tween 40 and 50 feet, 10-inch square timbers, 4 feet 6 inches apart, proved of 57 ample strength to support the sides of a 12-feet square shaft, though the same sized timbers at the reduced distance of 3 feet 6 in-ches apart, failed, as has been seen, to support the sides of a 9-feet square heading. After the experience of several rather troublesome slips, light timbering was abandoned, and a type which proved to be of ample strength to meet all the con- tingencies of heavy ground, vibration from road traffic, and the surcharge of lofty buildings, was adopted. In this type (Ex. 29) the 14-feet walings in- creased in scantling to 12 inches by 7 inches, were spaced 7 feet apart, and strutted at each end and at the center. At one-half the breaking weight the sup- porting power of the walings would be , 7 2 Xl2x3cwt. X 112 about- 65 a x7 =670 Iba. per square foot, and as the depth of the excavation was in some instances as much as 36 feet, this would correspond to a fluid pressure of 18.6 Ibs per cubic foot. With ground weighing 112 Ibs. 58 per cubic foot, and a slope of repose of 1-J to 1, the theoretical lateral pressure would be 32 Ibs. per cubic foot; and when it is remembered that this does not include any allowance for the sur- charge due to contiguous buildings, and that the stress on the timber is taken at fully half the breaking weight, it is clear that the average actual lateral pressure of the earthwork must have been less than half that indicated by theory. On the extensions of the Metropolitan railway, the same type of timbering was adopted, but the walings were generally 9 inches by 4 inches, and spaced 3 feet apart. The supporting power, upon the same basis as in the last instance, would be about 430 Ibs. per square foot. In most cases this strength proved to be sufficient, but in a few instances the walings broke, or showed such signs of distress that additional support had to be given. This was the case in some of the deep trenches along the Thames Em- bankment, where heavy wet silt was traversed. Near Whitehall Stairs (Ex. 59 30) the trenches were 40 feet deep, so the elastic strength of the timbering was only adequate to the support of a fluid pressure of 10.6 Ibs. per cubic foot, or probably but one fourth of that theo- retically due to the material ; it is there- fore no matter for surprise that the wal- ings proved unequal to their work. The stability of the timbering in more moder- ate depths was, on the other hand, con- firmatory of the general deductions drawn from previous examples as to the wide divergence between the actual and calculated thrust of earthwork. Turning now from the consideration of the temporary works of timbering to the finished and permanent structures on the underground railway, a similar varia- tion in strength will be found to obtain. The lightest retaining wall on the line is that at the Edgware Koad station yard (Ex. 31, Figs. 6 and 7). This wall is 23 feet in height from the top of the footing to the ground level, and has a maximum thickness of but 6 feet 3 inches at the base, out of which has to be deducted a 60 panel 2 feet 6 inches deep. Calculating the moment of stability at the level of the footings and round a point 3 inches back from the face of the pier which is a sufficient allowance for the crushing action on the brickwork M=4.4 feet, X 8,800 Ibs. = 38,720 foot pounds per Fig.6 lineal foot of wall. Dividing by , then 19^1bs. per cubic foot is the weight of the fluid^which would overturn this retaining wall. The ground supported is light dry sand, having a slope of repose of about 1J to 1, and consequently exert- ing a theoretical lateral thrust equivalent 61 to a 24-]b. fluid. There is practically no tenacity in the soil, as the author re- members seeing demonstrated on one occasion when a horse and cart, ap- proaching too near the top edge of the slope, broke it away and rolled together to the bottom of the 23 -feet cutting. Although theoretically deficient in stabil- ity, and subject to heavy vibration from the two minutes train service, the wall has stood perfectly without exhibiting the slightest movement. Upon the basis of the results of actual experiments, and having reference to the character of the soil and other conditions, the factor of safety would appear to be about 2 to 1. A far lower factor sufficed to secure the temporary stability of the dry areas at the station buildings previous to the erection of the arched roofs (Ex. 32). The arrangement at Sloane Square sta- tion is shown in Figs. 8 and 9. The joint stability of the front and back walls is the same as that of a solid rectangular wall having a thickness equal to (2.4 X 2.5 4- 3.8 X 5.2)* = 5.1 feet, or 62 say of the height. A fluid pressure of about 9 Ibs. per cubic foot would upset such a wall, so the factor of safety, until the arched roof abutted against the dry areas, was only that due to the few 14 -inch brick arches which tied the walls together with a certain amount of rigid ity. This result would perhaps have surprised the author more had he not previously investigated many cases of old timber wharves, in which the piles and plauking had lost more than f of their original strength from decay, and Section on line A B. yet held on against a theoretically over- powering thrust of earthwork. The relatively strongest wall on the Metropolitan railway system is at the St. John s Wood Koad station (Ex. 33), and that has given considerable trouble. Though 8 feet 6 inches thick at the base. 64 and backed up to a height of 16 feet only out of the total height of 21 feet 6 inches, and supported at the top by the thrust of the arched roof of the station, this wall moved over and forward to an extent which necessitated the immediate adoption of remedial measures. The moment of stability per lineal foot M= 73,000 foot pounds, conse- 1 na quently dividing by -^- the fluid resist- ance is 107 Ibs. per cubic foot, or allow- ing for the thrust of the arched roof of the station, considerably greater than that of a perfect fluid having the same density as the ground supported. It is not contended that such a pressure ever occurred upon the wall, although the ground is heavy yellow clay. The fail- ure arose from causes which will be re- ferred to more generally hereafter, and the case is only mentioned as a signal instance of the futility of hoping to re- duce the engineering of retaining walls to the form of a mathematical equa- tion. 65 It is a suggestive fact that, out of the 9 miles of retaining wall on the under- ground railway, the exceptionally weak wall should show no movement either during or after construction, whilst the exceptionally strong wall, though having six times the stability of the former, should fail. If an engineer has not had f some failures with retaining walls, it is | merely evidence that his practice has not | been sufficiently extensive ; for the at- tempt to guard against every contin- gency in all instances would lead to ruin- ous and unjustifiable extravagance, and be indeed as ridiculous a preceding as the making every soft clay cutting at a slope of 10 to 1* because in a few places such cuttings happen to slip down to that slope. In two instances comparatively heavy retaining walls have failed on the Metro- politan railway. During the construc- tion of the line, the wall on the west side of the Farringdon street station, (Ex. 34), failed bodily by slipping out at the toe and falling backwards on to the 67 slope of the earthwork (Fig. 10). This wall (Figs. 11 and 12) was 29 feet 3 in- ches high above the footings, and 8 feet bb to w?i..N^i -M--- ,o :. '.^ :.- -I . --"62 *-; 6 inches thick. The ground consisted of about 17 feet of made ground, 3 feet of loamy gravel, and 9 feet of clay. At a distance of 15 feet from the back of 68 the wall, and at a depth of 15 feet from the surface of the road, was the Fleet Sewer a badly constructed and much broken brick barrel, 10 feet 6 inches di- ameter and 3 rings thick. It was be- lieved that the leakage from the sewer induced the failure of the wall, but in reconstruction both wall and sewer were strengthened. The latter was made 4 rings thick in cement, and the former (Ex. 35) was increased in thickness to 12 feet 9 inches (Figs. 13 and 14). Origi- nally the stability w r as equal to the resist- ance of a fluid pressure of 24 Ibs. per cubic foot, and, as reconstructed, to 54 Ibs. On the opposite side of the same sta- tion yard the ground was retained by a line of vaults (Ex. 36, Figs. 15 and 16), 29 feet high above the footings, and 17 feet deep or double the original thick- ness of the wall last referred to. Al- though the resistance to overturning was greater in the proportion of 62 Ibs. to 24 Ibs. per cubic foot, the vaults some years after construction came over 15 inches 69 at the top, and slid forward considerably more. The movement when once fairly commenced was rapid and alarming, as a mass of densely inhabited houses was 70 within 20 feet of the back of the vaults. Steps were promptly taken to strengthen the work, by building intermediate piers and doubling the thickness at the back (Ex. 37, Fig. 17). This arrested the movement for a few months, when the vaults, whose stability had been thus in- creased to 93 Ibs. per cubic foot, again began to go over and slide forward. It was clear that mere weight would not insure stability, so 3 -feet square brick struts were carried at intervals from the toe of the piers across and under the railway to the retaining wall of the low- level line traversing the station yard, at a distance of about 34 feet from, and 8 feet below, the level of the footings of the vaults. The soil in the preceding instance con- sisted of about 12 feet of made ground overlying the clay, and, as in the former case, a sewer was to be found rather close to the back of the work. West- ward of the vaults, the clay encountered in the construction of the line was hard blue clay, requiring the use of a pick, 71 and portions of the temporary cuttings in the station yard, on the site where the vaults were subsequently built, stood fairly for many months at a slope of 1^- to 1. At one point, however, troublesome slips occurred, and even a 2 to 1 slope had to be piled at the toe to prevent for- ward movement. It was at this point that the vaults were subsequently found to be most dislocated. In neither of the above cases was fail- ure due to a deficient moment of stabili- ty in the wall, and therefore the fact of their failure does not in any way conflict with the results of the experiments pre- viously set forth. In each case water at the back of the wall was as usual the active agent of mischief not in thrust- ing the wall forward by hydrostatic press- ure, but in softening the clay and afford- ing a lubricant, so that the resistance was reduced to a sufficient extent to en- able the otherwise innocuous lateral pressure of the earthwork to tilt and thrust forward the wall. A costly, but conclusive, experience of 72 this softening action was obtained in the instance of the central pier to the double covered way on the District railway near Gloucester Eoad station (Ex. 38). The weight per lineal foot was 21 tons per foot run of pier, and 4 feet 9 inches was spread out by footings and concrete to a base of 10 feet; hence the pressure on the ground was 2.1 tons per square foot. In a similar construction near Alders - gate the load was 25 tons, and the width of base 8 feet 3 inches, giving the in- creased load of 2.9 tons upon the founda- tions. At the Smithfield market the author did not hesitate to place a column carrying 435 tons upon a 12-feet-square base, which is equivalent to a load of 3 tons per square foot ; and in the Euston Eoad, the side wall of the covered way has a load of 15 tons per lineal foot on a 4 -feet-wide base, which is at the rate of 3 1 tons per square foot. In all in- stances the foundation was clay, of ap- parently equal solidity, and in every in- stance but the first no settlement at all occurred. For some years no settlement 73 was observable in that case either, but ultimately, after an accidental flooding of the line, and permanent accumulation of water near the foundations, owing to the line being below the limits of nat- ural drainage, and the pumping being neglected, cracks were observed in the arches, and on examination the concrete Frg.18 and footings of the central pier were found to be fractured, as shown in Fig. 18. The load of 21 tons per lineal foot was thus imposed upon a base only 4 feet wide, and the softened clay proved unable to sustain the pressure of up- wards of 5 tons per square foot. Con- siderable difficulty was experienced in 74 checking the movement when once es- tablished. The center pier was under- pinned with brickwork in cement, but the footings, though of exceptional strength, were again sheared off, and it was found necessary to use 6-inch York landings. This failure shows the advisability of making concrete foundations of sufficient transverse strength to distribute the weight uniformly over the ground. As the result of experiment, the author is of opinion that the ultimate tensile re- sistance in a beam of good cement or lias lime concrete, is about 100 Ibs. per square inch, and in a beam of good brickwork in cement as much sometimes as 350 Ibs. per square inch.* Taking the former value (Ex. 39), a 12-inch thick concrete foundation, projecting 12 inches from the face of a wall, would break with a distributed loader =4,800 6x6 Ibs.; or, say, 2 tons per square foot. With a pressure upon the foundation of, * Vide "The Strength of Brickwork." By B. Baker. 'Engineering," vol. xiv. 75 say, 3 tons per square foot, and a factor of safety of 2, the thickness of a con- crete foundation would therefore be 3 tons X 2 ^=1.73 time the amount of 2 tons its projection beyond the face of the pier or wall, and the author would not advise a less thickness being used when the foundation rests on plastic clay. Water naturally gravitates to the foundation of a retaining wall, and a softening occurs. Owing to the lateral thrust of the earthwork, the pressure on the foundations is not uniform, and in- stead of settling uniformly, the outer edge descends fastest and the top of the wall is thrown outwards. The same softening reduces the clay to a condition in which it is easily ploughed up by the advancing wall, and the water acts as an admirable lubricant in diminishing the friction between the bottom of tlie wall and the clay on which it rests. These elements are exceedingly variable in their nature, and it is practically i 76 foretell the extent of their influence in each individual case. In tunneling, clay may be the best or the worst of materials almost self-sup- porting, or pressing with irresistible force on the crushing timbers and brick- work. It may be taken for granted that in good ground bad work will occasion- ally creep into tunneling, however close the inspection. How good a material clay can be was enforced upon the author's attention once in renewing a short length of defective tunnel lining, when on cutting down the work it was found that for some 50 feet the side wall, instead of being 2 feet 6 inches thick as intended, consisted merely of a skin of brickwork 9 in inches thick on the face, with a number of dry bats thrown in loosely behind this thin face wall to fill up the space excavated. This tunnel (Ex. 40) was loaded with a weight of 46 feet of clay over the crown, but no meas- urable settlement had taken place ten years after completion, and it was rather by sounding the side wall, than by the 77 observance of cracks, that a suspicion was raised as to its solidity. If the full weight of the ground had come upon the tunnel as it did upon the heading (Ex. 25), the pressure upon the side wall would have been 45 tons per lineal foot, or practically double the strength of the 9-inch work as determined by experi- ment. Of course the clay in this case was hard blue clay, which had not been af- fected by the action of air and moisture. As explained by the Rev. J. C. Clutter- buck, many years ago, the superficial lay- ers of the London clay are yellow, because the protoxide of iron is changed into a peroxide by the action of air and moist- ure in the disintegrated mass, and it is the yellow clay, therefore, which is the dread of the engineer. As good an ex- ample as any of the difference between the two materials was afforded forty* years ago, in the well known slip which occurred in the 75-feet-deep cutting at New Cross, when nearly 100,000 tons of yellow clay slipped forward on the hard 78 smooth surface of the shale like under- lying blue clay, and buried the entire line for a length of more than a hundred yards to a depth of 12 feet.* Owing to a misunderstanding, a sec- tion of concrete wall designed by the author to form one side of a running shed, and to retain the earthwork in a 13-feet cutting through light-made ground, was adopted also in a similar case, but where the ground was heavy wet clay, and the cutting 30 feet deep (Ex. 41). A wall 13 feet in height from formation to coping, and only 3 feet 3 inches thick at the base, had thus to sus- tain a surcharge of 17 feet. As the slope of repose was at least 1-| to 1, the lateral thrust was theoretically equiva- lent to a fluid pressure of about 70 Ibs. per cubic foot, whereas a pressure of less than one- third that intensity would have overturned the wall. The latter, never- theless, held up the ground fairly for some months, though the nature of the * Vide Minutes of Proceedings Inst. C.E.. vol. iii., p. 139. 79 soil was such that it ultimately became necessary to add strong counterforts to the wall, and to reduce the slope of the cuttings generally to 2 : 1. On the Thames Embankment heavy clay filling was in places cut through by the District railway, and in several in- stances the light side walls of the cover- ed way were thrust over a few inches at the top before the girders were bedded (Ex. 42). The side walls were eighteen feet in height from the invert to the ground level, and 5 feet 6 inches thick, with panels 5 feet 6 inches wide,by about 2 feet 9 inches deep,and piers 2 feet 6 inches wide. A fluid pressure of 16 Ibs. would overcome the stability of these walls, but, though subject to the pressure of the heavy clay filling, none of them failed. The existence of an undue pressure was, however, manifested by the thrusting for- ward of the green brickwork during the* few weeks that the walls were left unsup- ported by the girders. The retaining walls, at the approach to Euston Station, afford a good illustra- 80 tion of the impossibility of making any reasonable approximate estimate of the possible lateral thrust of yellow clay, or of stating positively that no movement will ever occur. These walls, soon after construction, were forced out in an ir- regular way at the top, bottom, or middle, but on pulling them down, the clay behind appeared to be free from fis- sures and to stand vertical. Cast-iron struts were subsequently put in between the opposite retaining walls ; and al- though General Burgoyne, who had given much attention to the subject of revetments, prophesied at the time that they would be removed in a few years, " when the ground had become consoli- dated," the struts still remain, and the walls still give signs of severe and in- creasing stress. It is not only London clay that proves so embarrassing to engineers. In a re- cent Paper* particular attention was called to the treacherous nature of some *Vide "Earthwork Slips," Minutes of Proceedings Inst. C. E,, vol. Ixiii,, p. 280 et seq. 81 boulder clay which, " although so tough and tenacious as to give the utmost dif- ficulty in excavation, after a short ex- posure became soft and pasty in the winter, often jolting down the slurry." Examples were given of formidable slips in this material, in contrast with which the author would point to the comparati- vely slow wasting of the huge boulder clay cliffs near the mouth of the Tyne, a mat- ter which he had occasion to investigate very closely in connection with the Duke of Northumberland's Jands in that dis- trict. From a comparison of surveys extending over a period of one hundred and fifty years, it appeared that the wasting of the cliff was very slow, and due solely to the wash of the waves at its base. At no time was the slope of repose of this 105 -feet-high cliff more than 1 to 1, and in places it stood for years at an average slope of less than to 1. With, his experience of North London clay, the author was startled to find people con- tentedly living in houses partially over- hanging the brow of this steep and 82 ragged cliff, but the stability of the clay was so great, and the wasting so uniform, that the fact of the outhouses being at the bottom of the 100 feet slope, and the main building at the top, did not appear in any way to disturb the equanimity of the householders. The failures of dock walls, though nu- merous and instructive, afford no direct evidence as to the actual lateral pressure of earthwork, because in practically every instance the failure is traceable to defec- tive foundations. The author cannot re- call any case in which a dock or quay wall founded on rock has overturned or moved forward, though on other founda- tions a movement to a greater or lesser extent is so much the rule that Voisin Bey, the distinguished engineer-in-chief of the Suez Canal, once stated to the author that he could name no exception to it, since he had failed to find any long line of quay wall, which on close inspec- tion proved to be perfectly straight in line and free from indications of move- ment. A brief examination of some in- 83 stances of the failures of dock walls will show how powerfully unknown practical elements affect theoretical deductions in such cases. A well-known and often cited case is that of the original Southampton dock wall, constructed now some forty years ago (Ex. 43, Fig. 19). This wall, 38 feet in height from the foundation to the coping, was built on a platform of 6 -inch planks, resting on a sandy and loamy bottom. Before the water had been let 84 into the dock, or the backing carried to the full height, the wall moved forward in some places as much as three feet, but came over hardly anything at the top. When the water was let into the dock, the filling behind becoming saturated, the pressure on a receding tide exag- gerated, and to secure stability it was found necessary to discontinue the filling at some distance below the full height of the wall, and to substitute a timber platform. The thickness of this wall at the base is 32 per cent, of the height between the buttresses, 45 per cent, at the buttresses, and a rectangular wall containing the same quantity of material would have a thickness equal to 26 per cent, of the height. Though the base is wide, the weight is light as compared with most other dock walls, and the tendency to slide forward is therefore greater. If founded on a rock bottom, a fluid press- ure of about 40 Ibs. per cubic foot would have been required to overturn the wall, but of course a fraction of this pressure 85 would suffice to make it move forward on the actual bottom. The conclusion drawn by Mr. Giles, M. Inst. C. E., the engineer of the docks, from this and other failures is, that the quality of a dock wall is of little conse- Fig.20 quence compared with the quantity, and that it ought to be sufficiently strong not only to hold any amount of any kind of backing put against it, but to carry a 86 head of water equal to its height if it were left dry on the other side.* These principles have been adhered to in the recent extension of the South- ampton docks (Ex. 44, Fig. 20). Here the wall is founded on a mass of con- crete 21 feet wide; the effective thickness at base is about 45 per cent, and the mean thickness 41 per cent, of the height. A fluid pressure of from 60 to 70 Ibs. would be required to overturn this wall if on a hard foundation, and probably as much to make it move forward, unless the bottom were of clay or of other un- favorable material. Mr. Giles has found even a heavier wall slide, when founded on a thin layer of gravel overlying clay. In the earlier wall, if the co-efficient of friction of the base on the ground were less than f , the wall would slide ratner than overturn ; but in the latter wall, without buttresses, any co-efficient ex- ceeding ^ would be sufficient to prevent sliding. * Vide Minutes of Proceedings Inst. C. E., vol. lv., p. 52. 87 For comparison with the above, the section of the east quay wall of the Whitehaven dock may be next referred to (Ex. 45, Fig. 21). Having the same height as the Southampton dock wall, the thickness at the base is but 37 per / cent of the height, the mean thickness 31 per cent., and the concrete foundation 16 feet 6 inches, instead of 21 feet wide. This wall has stood perfectly, though it would fail to resist the head of water mentioned by Mr. Giles, but would be 88 overturned by a fluid weighing from 45 to 50 Ibs. per cubic foot. During con- struction, weep holes were, however, left in the walls to relieve them of hydro static pressure. Fig.22 .10.6-^' Another dock wall of the same height as the preceding ones, is that of the Avonmouth dock (Ex. 46, Fig. 22). In this instance the thickness is 42 per cent, of the height, and the concrete base 22 feet 6 inches wide, dimensions which, 89 with a good foundation, would enable the wall to stand a full hydrostatic pressure at the back. Owing to the treacherous nature of the bottom, a long length of this wall nevertheless slipped forward at one point as much as 12 feet 6 inches, and sunk 4 feet 6 inches without the latter being affected, whilst at another point, where there was no for ward move- ment, the wall came over about 1 foot 8 inches. When the failure occurred, the foundation rested on apparently stiff blue clay, but in subsequent portions the concrete was carried down through the clay to the sand.* On the east side of the dock, though the walls were founded at an average depth of no less than 9 feet below the bottom of the dock, they still moved forward in the mass some 15 feet 6 inches, and sunk 7 feet 6 inches, The filling was carefully punned in layers, with material which seems to have stood fairly at a slope of 1J or 2 to 1, so that the wall theoretically possessed an * Vide Minutes of Proceedings Inst. C. E., vol. lv., p. 15. 90 excess of strength, and yet, owing to the existence of conditions which it was im- possible for the engineer to foresee, failures occurred as described. A somewhat similar case of sliding for- ward occurred at the New South Dock, Fifl.23 : a^m ^ JLO.- * West India Docks (Ex. 47, Fig. 23). The wall is 35 feet 9 inches high from the top of the footings to the coping, and 13 feet, or 36 per cent, of the height, thick at the base. The concrete foundation is 17 feet wide, and 6 feet deep below the 91 bottom of the dock, and the fluid press- ure required for overturning would be about 45 Ibs. per cubic foot. A coefficient of friction of less than would be suffi- cient to guard against sliding under this pressure, but owing to the existence of a thin seam of soft greasy silt between the hard strata of blue clay upon which the foundations rested, several portions of the wall slid forward. The original ground level was about 15 feet below the top of the dock wall, and the excavation stood fairly as a slope of 1 to 1. Favorable material for backing did not appear to be available. The fact that the stability of a dock wall depends far more upon the founda- tion than upon the thickness or mass of the wall itself, is well illustrated by the quay wall at Carlingford (Ex. 48, Fig. 24). With a height of no less than 47 feet* 6 inches, the thickness of wall and width of foundation at the base are each but 15 feet, or less than 32 per cent, of the height, and the mean thickness is but 24 per cent. A lateral pressure of 92 half that due to a hydrostatic pressure would probably suffice to overturn this structure. In contrast with the preceding wall may be cited that of the dock basin at Marseilles (Ex. 49, Fig. 25). In both in- stances the foundation was good, and the wall rested immediately upon it with- out the interposition of any broad mass of concrete ; , but the French engineer, though the wall was but 32 feet high, made the thickness at the base no less than 16 feet 9 inches, or 52 per cent, of the height an unusually large propor- tion, which he was led to adopt in conse- quence of the stratification of the ground inclining towards the wall. Perhaps one of the boldest and most successful examples of a lightly-propor- tioned wharf wall is that built by Colo- nel Michon in 1857 on the Moselle at Toul (Ex. 50, Figs. 26 and 27). With a height of 26 feet, and a batter of 1 in 20, the thickness of the wall through the counterforts is but 3 feet 7 inches at the base, and though the filling is ordinary 93 * >f u. -4 94 material, having a slope of repose of 1-J- to 1, and the floods rise within 6 feet of the top of the coping, no movement whatever has occurred since the wall was built. As striking a contrast as could be wished to the above light construction is found in Sir John Macneill's quay wall at Grangemouth harbor (Ex. 51, Fig. 28). Both walls are of about the same height, but whilst the mean thickness of the first is only 3.7 feet, or \ of the height, that of the second, inclusive of the mass of concrete backing, is no less than 23 feet, or, say, f of the height. One of the most troublesome cases of dock- wall failures was that at the Belfast harbor* (Ex. 52, Fig. 29). This wall was founded upon round larch piles 15 feet long, 10 inches in diameter at the top, and 4 feet 6 inches apart from cen- ter to center. Symptoms of settlement became apparent soon after the filling was commenced, and some remedial * Vide Minutes of Proceedings Inst. C.E., vol. lv., p. 31. 95 96 measures were attempted. The ground, however, was hopelessly bad, the slope of repose ranging from 3 to 1 to 6 to 1, and the backing material being equally bad, the light piling was inadequate to resist the thrust. Two years after erec- tion a length of about 70 lineal yards of wall was overturned and carried forward into the middle of the dock entrance, the piles being sheared off about 6 feet be- low the bottom of the wall. The height from the top of the pile to the coping is 31 feet 6 inches, and the thickness at the base 16 feet, or half the height. On good ground, therefore, the wall would have had an ample margin for stability. A somewhat similar failure occurred in the instance of the original side walls of the lock chamber of the Victoria docks* (Ex. 53, Fig. 30). These docks were built at a time when little confidence was placed in concrete as a durable material for dock work, and consequently the walls were faced with cast-iron piling and plates, as in previous instances at Black- *lbid., vol. xviii., p. 462. 97 wall and elsewhere. The foundations were on a layer of gravel overlying the clay, but the face piling had little hold in the gravel, and the base of the wall itself was only some 30 per cent, of the height, hence, when the water was let into the dock, the hydrostatic pressure at the back of the lock wall forced it bodily forward into the lock, ploughing up the puddle in front of it, and break- ing tie bolts and tie piles as it advanced. In reconstruction a solid concrete wall 20 feet thick, and having nearly treble the stability, was carried through the gravel down to the clay. The wall of the Victoria Dock Exten- sion Works, by Mr. A. M. Eendel, M. Inst. C.E. (Ex. 54, Fig. 31), has a thick- ness of about 50 per cent, of the height at the point where the 18 -feet wide foundation meets what may be termed the body of the wall, and the wharf wall of Mr. Fowler's Millwall dock (Ex. 55, Fig. 32) has a maximum thickness of 13 feet 6 inches for a height of 28 from bottom of dock to coping, of 98 practically the same ratio. Either or these walls would be capable of resist ing the full hydrostatic pressure. An early example of a successful wall on a very bad foundation is afforded by Fig.32 Sir John Kennie's Sheerness wall (Ex. 56, Fig. 33) The subsoil consisted of loose running silt for a depth of about 50 feet, covered with soft alluvial mud, and the depth at low water was at some points as much as 30 feet. A piled plat- form about 42 feet in width, with sheet- OFTHE 99 ing piles on the river face, and 12-inch piles pitched from 3 to 4 feet apart over the whole area, and driven until a 15-cwt. monkey falling 25 feet did not move Fig.33 them more than ^ inch at a blow, was prepared, and upon this the wall, no less than 50 feet in extreme height and 32 feet in effective thickness at the base, was raised. In no case has any yielding 100 or unequal settlement taken place, ex- cept in the instance of the basin wall, the cracks in which Sir John Kennie at- tributed to other causes than a failure in the foundation. Although the voids in the masonry were designedly filled in with grouted chalk and other light mate- rial, the Sheerness river wall has per- haps a greater moment of stability than any other wall in the world. Another exceptionally heavy wall, more than a half century younger than the preceding, is that of the Chatham Dock- yard Extension (Ex. 57, Fig. 34). The height from the bottom of the dock to the coping is 39 feet, and the founda- tions are carried down to the loam gravel or chalk at a depth of 4 feet 6 inches below the bottom of the dock. The thickness of the wall is 21 feet at the base, or, say, J- of the extreme height. On a hard chalk bottom it would resist a fluid pressure of about 80 Ibs per cubic foot. Two examples of Liverpool dock walls, namely, that at the Canada half- 101 tide basin, and that at the Herculanean docks, are given in Figs. 35 and 36. The 102 former (Ex. 58) is 43 feet in extreme height, and 19 feet, or 44 per cent, wide at the base. The latter (Ex. 59) is 39 feet high, and 18 feet, or 46 per cent, wide at the footings, which rest on a marl bottom. A dock wall at Spezzia (Ex. 60) of somewhat similar propor- tions, the height being 41 feet, and the width at the bottom of foundations 23 feet, or 56 per cent, of the height is shown on Fig. 37. Walls made of large concrete blocks, resting upon a mound of rubble, have been constructed in many of the Medi- terranean ports, generally with success, but occasionally with failure, as at Smyr- na, where, owing to the great settlement, six and seven tiers of blocks had to be superimposed instead of four, as in- tended, and the quay wall had after all to be supported by a slope of rock in front extending up to within 7 feet of mean sea level, and seriously interfering with the use of the quays. The propor- tions arrived at by experience are a width of 9 meters at the top, and a thickness 103 S 104 of not less than 2 meters for the rubble mound ; a depth of 7 meters below the water line, and a thickness of 4 meters for the concrete block wall resting on the mound ; and a minimum thickness of 2.5 meters, and a height of 2.4 meters for the masonry wall coping the concrete blocks. At Marseilles (Ex. 61, Fig. 38), the top of the rubble mound is only 6 meters below the water-line, so vessels occasionally bump ; and the concrete block wall 3.4 meters, or 40 per cent, of the height, in thickness has proved rather less stable under the contingencies of working and the surcharge of build- ings and goods than is considered desir- able. Examples are not wanting, however, of walls founded on rubble mounds where the thickness holds a smaller ratio to the height than the 42 per cent., con- sidered necessary by the French engi- neers. Mr. Fowler has made concrete block walls in the Rosslare Harbor (Ex. 62) 42 per cent, of the height on the sea 105 face, and but 28 per cent, on the harbor side, but cross walls at 50-feet intervals considerably strengthen the work. The inner wharf wall of the Holyhead new harbor, again (Ex. 63, Fig. 39), is 27 feet high and 8 feet thick, a ratio of under 30 per cent., but though stable, the line of coping is somewhat wavy on plan. The original wall of the West Pier at White- haven (Ex. 64, Fig. 40), is 42 feet 6 inches high, with a thickness of 8 feet 6 inches between the buttresses, which latter are 6 feet deep by about 4 feet wide and 15 feet apart ; but the lightest of all, per- haps, is the dry masonry outer wall of the St. Katherine's breakwater, Jersey, (Ex. 65, Fig. 41), which is only 14 feet wide at the base for a total height of 50 feet, or a ratio of 28 per cent. It must not be forgotten, of course, that the three latter walls have to sup- port rubble hearting only, instead of sand and other material, having a much flatter slope of repose. Occasionally, as has been stated (Ex. 22), rubble will not stand at less than 1 to 1 ; but at Holy- 106 head and Alderney the slope of the rub- ble mound on the harbor side is only about 1J to 1. At Cherbourg it is 1 to 1, and at Leghorn the large concrete blocks are found to be stable at a slope Fig.40 Fig,4l of f to 1. By a very little care in selec- tion, the thrust of a rubble filling may be reduced to a fraction of that arising from bad material, and indeed in the ordinary run of fishing piers in the North 107 of Scotland, however great the height, the face wall of the rubble-hearted pier consists simply of stones from 3 to 4 feet in depth, laid dry to a batter of about 1 in 5. The north-east pier at Seham, again, has an inner wall 25 feet high* battering 1^ inch to the foot, and only 5 feet thick, and many similar examples are to be found at other points of the coast. The most cursory examination of cases of failure cited above will serve to justify the statement that the numerous dock- wall failures do not afford any direct evidence as to the actual lateral pressure of earthwork. Thus, remembering Gene- ral Burgoyne's battering wall, only 17 per cent, of the height in thickness, supported the heavy sodden filling at its back, no calculation is required to show that the 32 and 45 per cent. Southamp- ton Dock counterfeited wall, the 42 per cent. Avonmouth Dock wall, the 36 per cent. West India Dock wall, the 50 per cent. Belfast Harbor wall, and the 30 per cent. Victoria Dock wall, would all have 103 stood perfectly had the foundation been rock, as in the instances of General Bur- goyne's experimental walls, instead of the mud, clay, and silt which it actually was. Not only the strength, but the type of cross-section, is singularly indicative of the small influence which theory and ex- periment have exercised upon the design of dock walls. If the early theorists and experimentalists were in accord upon one point, it was upon the immense advant- age afforded by a court terforted wall. Lieutenant Hope was led by his experi- ments to conclude that if good counter- forts were introduced, the merest skin of face wall would suffice for the portion between them, and theorists of course arrived at the same conclusion, from a comparison of the moments of stability of rectangular blocks of masonry edge- wise and flatwise. Nevertheless, in only one of the preceding dock walls, and that one forty years old, are counterforts in- troduced. In practice it was found that counterforts frequently separated from 109 the body of the wall, and they were con- sequently regarded as untrustworthy. It is open to question whether this con- clusion does not require reconsideration in these days of cheap, strong, and easily-moulded Portland cement concrete. Nothing but blasting would separate the counterforts from a good concrete wall. The author has used concrete in many varieties of structures, and as long back as fifteen years built a four-story ware- house, walls and floors, entirely of con- crete, without the introduction of any iron girders. He is bound to admit, however, that by far the boldest and most thorough adaptation of the ma- terial to multifarious uses met with by him was in the instance of some farm buildings in an out-of-the-way district in Co. Kerry, Ireland. The small tenant- farmer and his laborers none of whom were receiving over 11s. a week with- . out skilled assistance of any kind, had constructed dwelling-house, cattle-sheds, and hay -barn wholly of concrete. The cattle-shed was roofed with concrete 110 arches of 15 -feet span, 1 foot rise, and 4 inches thick, springing from octagonal concrete pillars 8 inches in diameter, spaced 15 feet apart from center to center. A layer of concrete constituted the pav- ing, concrete slabs divided the stalls, the cattle fed and drank out of concrete troughs, the windows were glazed in concrete mullions, the gates hung on con- crete posts, and the farmer seemed to regret somewhat that he had not adopted concrete doors and concrete five-bar gates. Portland cement concrete being thus possessed of such great tenacity, there is no risk of counterforts separating from the body of a wall, but it by no means follows that there would be any advant- age in using them in other than excep- tional cases. In practice, as failures have shown, it is weight, with the consequent grip on the ground, rather than a high moment of stability, that is required iri a dock wall. It may be asked, with reason why a bad bottom should affect the thickness of a retaining wall, or, in other Ill words, why the foundation should not first be made good, and then a wall of ordinary thickness be built upon it. The answer, of course, is that if weight is re- quired to prevent sliding, it is just as economical to distribute the material over the general body of the wall as to confine it to the foundations. It follows, therefore, that under the stated condi- tions the adoption of a counterforted wall would lead to no economy in ma- terial, whilst it would involve additional labor in construction. A dock wall is subject to far larger contingencies than an ordinary retaining wall, and the required strength will be included only within correspondingly large limits. Hydrostatic pressure alone may more than double or halve the factor of safety in a given wall. Thus, with a well-puddled dock bottom, the subsoil water in the ground at the back . of the walls will frequently stand far below the level of the water in the dock, and the hydrostatic pressure may thus wholly neutralize the lateral thrust of the 112 earth, or even reverse it, as in the case of the inner retaining walls on the Soon- kesala canal, some of which, though 35 feet in height, are only 2 feet thick at the top and 7 feet 6 inches at the base. On the other hand, with a porous subsoil at a lock entrance, the back of the walls may be subject, on a receding tide, to the full hydrostatic pressure due to the range of that tide plus the lateral press- ure of the filling. Again, the water may stand at the same level on both sides of the wall, but may or may not get under- neath it. If the wall is founded on a rock or good clay, there is no more reason why the water should get under the wall than that it should creep through any stratum of a well-constructed ma- sonry or puddle dam, and under those ci -cum stances the presence of the water will increase the stability by diminishing the lateral thrust of the filling. With rubble filling, assuming the weight of the solid stone to be 155 Ibs. per cubic foot, and the voids to be 35 per cent., the \veight of the filling would be 100 Ibs. 113 per cubic foot in air, and 59 Ibs in water, and the lateral thrust will be that due to the latter weight. If, however, as is perhaps more fre- quently the case, the wall is founded on a porous stratum, the full hydrostatic pressure will act on the base of the wall, and reduce its stability in practical cases by about one -half. Thus, the 30 -ton concrete block walls on rubble mounds, at Marseilles and elsewhere, have the stability due to a weight of, say, 130 Ibs. per cubic foot in the air, and 66 Ibs. per cubic foot in sea water : but the rubble filling at the back of the wall, being simi- larly immersed, is also reduced in weight, and consequently thrust to a correspond- ing extent, so the factor of safety is un- affected. In walls with offsets at the back, as in Figs. 25 and 36, and water on both sides, the stability will be much increased by the hydrostatic pressure on the top of ' the offsets, should the wall rest on an impermeable foundation. It is generally 114 assumed, in theoretical investigations,* that the weight of earthwork super-im- posed vertically over the offsets should be'included in the weight of the wall in estimating the moment of stability ; but the author has found no justification in practice for this assumption. He has in- variably observed that when a retaining wall moves by settlement or otherwise, it drops away from the filling, and cavi- ties are formed. A settlement of but -fa of an inch, after the backing had become thoroughly consolidated, would suffice to relieve the offsets of all vertical pressure from the superimposed earth, and the latter cannot therefore be properly con- sidered as contributing to the moment of stability. A wall with deep offsets at the back is not a desirable form where the foundation is bad, and where, consequently, the pressure over the foundation should be as uniform as possible, so that a settle- ment may take the form of a uniform * Vide " A Manual of Civil Engineering." By W. J. M. Rankine, p. 402. 115 sinking, and not a tilting forward of the coping by reason of the toe sinking faster than the back of the wall. A paneled wall, such as that shown on Figs. 11 to 14, though not admissible in dockwork, is on bad ground far less liable to come over than a wall with offsets at the back, and with a consequent concentration of weight at the front, where the conditions of a lateral thrust especially require that it should not be. The latter conditions also indicate the expediency of adopting raking piles, as in Fig. 33, rather than vertical piles, as in Fig. 29, where a piled foundation is un- avoidable- Thus, taking an ordinary case of dock wall, in which the factor of safety, as regards overturning, is 3, and the ratio of weight of wall to the lateral pressure of earthwork required to over- turn it is If to 1, it follows that if the foundation piles are driven at the rate of 1 to 3 + If = 1 : 5 there will be no transverse strain tending to break them off, as in the case illustrated by Fig. 29, and no tendency to plough up the 116 soft ground in front of the toe of the wall. If an engineer could tell by inspection the supporting power and frictional ad- hesion of every bit of soil laid bare, or see through 5 or 10 feet of earth into a " pot hole," or .layer of slimy silt, he might avoid many failures, and even hope to frame some useful equations for ob- taining the required thickness of a dock wall. Taking things as they are, how- ever, it is hardly worth while to use even a scale and compass in such work, for being in possession of all the informa- tion obtainable about the foundation and backing, an engineer may at once sketch as suitable a cross -section for the parti- cular case as he could hope to arrive at after any amount of mathematical inves- tigation. Something must be assumed in any event, and it is far more simple and direct to assume at once the thick ness of the wall than to derive the latter from equations based upon a number of uncertain assumptions as to the bearing power of the foundations, the resistance 117 to gliding, and other elements. This being so, it has often struck the author that the numerous published tables giving the calculated required thicknesses of retaining walls to three places of deci- mals, stand really on exactly the same scientific basis, and have the same prac- tical value, as the weather forecasts for the year in Old Moore's Almanack. In both cases a pretence is made of foretell- ing what experience has shown can often not be known until after the event. One well-known authority gives young en- gineers the choice of five hundred and forty-four different thicknesses for a simple vertical rectangular retaining wall, so that an unfortunate neophyte might not unreasonably conclude that the task before him was not to decide whether, say, a 32 -feet wall should be 20 feet thick, as in Example 60, or 9 feet, as in Example 62, but whether it should be 14 feet 6 inches or 14 feet 5J inches thick. Although dock wall failures do not afford any data as to the actual lateral 118 pressure of earthwork, a knowledge of the latter will enable much valuable in- formation to be deduced as to the bearing power of soil and other matters from such failures, and the data so obtained will be applicable to other structures beside retaining walls. Knowing the actual lateral thrust, the coefficient of friction of the base of a wall which has been pushed forward on the ground can be at once deduced, but if the theoretical as distinguished from the actual thrust were introduced into the equation, the result would be valueless. The aim of the author in the present paper has been to set forth as briefly as possible what he knows regarding the actual lateral thrust of different kinds of soil, in the hope that other engineers would do the same, and that the infor- mation asked for by Professor Barlow more than half a century ago may be at last obtained. Although the acquirement of the missing data would probably lead to no modification in the general propor- tions of retaining structures, since these 119 are based upon dearly bought experience, it is none the less desirable that it should be obtained ; for an engineer should be able to show why he believes that a given wall will stand or fall. To assume upon theoretical grounds a lateral thrust, which experiments prove to be excessive, and to compensate for this by giving no factor of safety to the wall, is not a scien- tific mode of procedure. Experience has shown that a wall of the height in thickness, and battering 1 inch or 2 inches per foot on the face, possesses sufficient stability when the backing and foundation are both favor- able. The author, however, would not seek to justify this proportion by assum- ing the slope of repose to be about 1 to 1, when it is perhaps more nearly 1^ to 1, and a factor of safety to be unnecessary, but would rather say that experiment has shown the actual lateral thrust of good filling to be equivalent to that of a fluid weighing about 10 Ibs. per cubic foot, and allowing for variations in the ground, vibration, and contingencies, a 120 factor of safety of 2, the wall should be able to sustain at least 20 Ibs. fluid press- ure, which will be the case if J of the height in thickness. It has been similarly proved by expe- rience that under no ordinary conditions of surcharge or heavy backing is it ne- cessary to make a retaining wall on a solid foundation more than double the above, or \ of the height in thickness. Within these limits the engineer must vary the strength in accordance with the conditions affecting the particular case. Outside these limits the structure ceases to be a retaining wall in the ordinary ac- ceptation of the term. A 9 -inch brick facing might secure the face of a friable chalk cutting which, if suffered to re- main exposed to the action of the weather, would crumble down to a slope of 1 to 1, and a massive bridge pier, with an "ice- breaker " cutwater, might stand firm against an avalanche, but in neither case could the structure be fairly stated to be a retaining wall. Hundreds of revetments have been 121 built by Royal Engineer officers in ac- cordance with General Fanshawe's rule of some fifty years ago, which was to make the thickness of a rectangular brick wall, retaining ordinary material, 24 per cent, of the height for a batter of ^, 25 per cent for |, 26 per cent, for , 27 per cent, for -fa, 28 per cent, for -^ 30 per cent, for .-fa , and 32 per cent, for a vertical wall. As a result of his own experience the author makes the thickness of retaining walls in ground of an average character equal to J of the height from the top of the footings, and if any material is taken out to form a face panel, three-fourths elf it are put back in the form of a pilaster. The object of the panel, as of the 1^ inch to the foot batter which he gives to the wall, is not to save material, for this involves loss of weight and grip on the ground, but to effect a better distribu tion of pressure on the foundation. It may be mentioned that the whole of thp walls on the District railway were de- signed on this basis, and JiaLJJifire has 122 not been a single instance of settlement, or of coming over or sliding forward. The author has in the present paper analyzed a few dozen experiments, and discussed as many more facts; but an en- gineer's experience is the outcome not of a few facts, but of the thousands of in. cidents which force themselves on his at- tention in carrying out work, and it is this experience, acquired in the construction of works of a somewhat special character, which has convinced the author that the laws governing the lateral pressure of earthwork are not at present satisfac- torily formulated. DISCUSSION. Mr. B. BAKEK desired to add, that his object in bringing forward the paper was not so much to present certain facts for criticism as to induce others to give the ( results of their experience, and if every one helped a little he thought a very use- ful result would be attained. Mr. W. AIRY said he had given consid- 123 erable thought and attention to the sub- ject of earthwork, and he considered the collection of examples in the paper would make it an extremely useful one for purposes of reference. The subject of earthwork was a very difficult one to deal with, and he wished to point out briefly in what this difficulty consisted. A B C D (Fig. 42) might be taken to be Fig.42 the section of some ground having a small vertical cliff at B C. There would be a tendency for the ground to break away and come down along some such line as D B. The whole problem of the stability of the ground, both as affecting the slope of the earth and the pressure against a retaining wall, depended upon the accurate determination of the line D 124 B. It was not an exceedingly difficult matter to determine this line, if the con- stants of cohesion, friction, and weight of the ground were known ; and he had himself dealt with the problem in a paper communicated to the Institution. The mechanical conditions of equi- librium were very simple ; the force tend- ing to bring the earth down was the weight of it ; the forces tending to keep it from coming down were the friction along the line D B and the cohesion of the ground along that line. All those forces acted according to well-under- stood laws, and therefore if the con- stants of weight, cohesion, and friction of any particular ground were known, it was not difficult to find out the exact position of the line D B, and therefore the pressure on the retaining wall, or the shape of the slope. The question then arose, what was the real difficulty of con- structing tables for practical use with re- gard to earthwork? Simply this, that the varieties of ground were infinite in number and very wide in range, and 125 when that was the case it was quite idle to think of constructing tables for prac- tical use. A man having a particular kind of earth to prescribe for, would not be able to ascertain by inspection what the constants of that earth were, and therefore he would not know where- abouts in a table to look ; he would have to determine the constants for himself ; and if he had to do that he had to do the whole work, and the tables were of no use to him. He thought the author had rather overlooked the enormous number of conditions of earth when he contrasted the small number of experi- ments upon earthwork with the large number of experiments made with tim; ber. A piece of oak would give very nearly the same results for strength, elasticity, and so on, whether it was grown in Kent or in Yorkshire ; and, therefore, when a few experiments had been made upon it, it was not necessary to repeat them over and over again. That was not the case with earthwork, because the conditions" were so exceed- 126 ingly variable. He exhibited a little rough machine he had used for testing earthwork and taking the cohesion of the ground. The block of wood might be taken to represent a block of raw clay taken out of a cutting. There was a common lever balance, and a couple of movable cheeks were fitted into chases cut in the sides of the clay block ; and the clay having been rammed in a box so that it could not move, weights were put in the scale until the head was torn off. After subtracting the weight of the piece that was torn off, and measuring the area of the cross section that was broken, the constant of cohesion was determined. For the constant of fric- tion he arranged a certain number of blocks of the same clay in a tray, and scraped them off smooth; then he had another block of clay with a smooth sur face which he put on it, and then tilted the tray until the loose block slid; that gave the coefficient of friction. He should like to refer to the exceedingly wide range of tenacity shown by different 127 kinds of clay. In one set of experi- ments with ordinary brick loam, that clay gave a coefficient of cohesion of 168 Ibs. per square foot, and a coefficient of friction of 1.15. With some shaley clay out of a cutting in the Midlands, he had found a coefficient of tenacity of 800 Ibs. per square foot, and a coefficient of friction of 0.36. That was a very wide range, but it was only a part of what was actually to be found in practice. Mr. L. F. YEKNON HARCOTJRT wished to say a few words on the subject, as the author had referred to two or three works with which he had been connected. The author had pointed out, from the experiments he had recorded, that the pressure upon the back of a retaining wall was a good deal less than it was theoretically supposed to be about one- half but as he allowed a factor of safe- ty of 2, it apparently came to very much the same thing. With regard to walls on a rubble mound, the author remarked that the base was in many cases small. That, he thought, was owing t3 two 128 causes ; first, that with a rubble mound for a base there was no chance of slid- ing ; and secondly, that in those cases there was a rubble filling behind, which he supposed was about as good a mate- rial for backing as could be got. The slope of the inner face of the rubble mound of the breakwater at Alderney harbor had been referred to as 1 J to 1 ; but it ought to be remembered that in that case the materials used were very large blocks of stone, and therefore the slope would be naturally steeper than under more ordinary conditions. Kef- erence had also been made in the paper to St. Katharine's breakwater, Jersey, as an example of a wall built with a very small base. The author took the whole of the height of that wall as the proper height; but it would be observed that the top of the wall had what used to be called a promenade along it, and there- fore the whole of the filling did not ap- ply to the entire height of the wall. The author stated that the base was 28 per cent, of the height of the wall, but 121) leaving out the promenade it would be 35 per cent. Of course it would be something intermediate, as there would only be the small piece of filling under the promenade to be taken into account additional, instead of what would be the filling at the back if it was filled up en- tirely to the top level. The author had referred to the West India dock wall, and stated that several portions of it had come forward. That, however, was not quite the case. It was true that two portions of the south wall came forward that two surfaces of clay at some little depth below the wall slid upon one an- other. Probably some seam of sand or silt was washed out by the water behind the wall from between two layers of clay, and in that way the two detached sur- faces of clay were free to slide upon one * another. He was quite certain of the exact position of the surfaces of rup- tmv, because he saw the two surfaces of clay after the excavation was made for re- build ing the wall, and they were as smooth as glabn. The romecly for that appeared 130 to him to be very simple, and it was cer- tainly successful in the case in point. The wall had failed, as the author had stated, not from any fault in the thick- ness or the weight, but simply owing to the sliding forward ; and instead of add- ing any further weight to the wall, the foundations were carried down to a greater depth ; but it only required 2 or 3 feet more in depth in the basin wall foundations that had to be executed af- terwards under precisely similar condi- tions. That was quite sufficient to. keep the wall in a perfect state of equilibrium without the least coming forward; and he should imagine that was decidedly better on the whole than adding to the weight of the wall. It appeared to him that practice was rather contrary to theory in giving too great a thickness to the top of the wall, and too small a thickness, comparatively speaking, to the bottom ; and that it would be better to have a wall more of the shape of the Sheerness wall a good deal lessened at the top, rather than a wall like those 131 generally adopted, which had more par- allel faces with a little additional thick- ness from the batter. He thought it would be better to make a wall narrower at the top and widening out more to- wards the bottom, and to bring the foundations of the wall well down into the ground so as to prevent any chance of sliding. In the case of the West India dock wall, besides the badness of the backing, there was a large amount of water that seemed to percolate from the Millwall docks, which were filled with water while the wall was being built, the docks not having been puddled. It was clearly shown that that had a considera- ble effect, because the north wall, though it was built in exactly the same manner, and though the water of the Export dock was really nearer, stood perfectly, as there was not the same amount of water pressure at the back, owing to the water being unable to penetrate through the silted-up bottom of the Export dock. He considered that the Institution was much indebted to the author for collect- 132 ing and comparing so many valuable facts, as, whilst descriptions of particu- lar works were very useful, it was by taking a general survey, from time to time, of the existing state of knowledge, in any special branch, that definite prog- ress in engineering science was most likely to be promoted. Mr. J. WOLFE BAEKY believed the state- ment was true, that the pressure against retaining walls did not approach to the theoretical thrust ; at the same time he was of opinion that large retaining walls gave the engineer as much anxiety as any work he ever undertook. It should be re- membered that, as a rule, the thrust which the walls had to bear came against them when the material of which they were composed was green, and unless con- tractors and others were very careful in strutting the new work, and allowing plenty of time for the material to set, there would be a condition of affairs in the early stages of the wall which would never arise after the materials were thor- oughly consolidated. He wished to point 133 out that it was for such reasons most en- gineers were now getting to realize the extreme desirability of using cement as much as possible. The early stages of engineering works were generally those in which the greatest risks were run, and if a slow-setting material were used, the strains would be exerted against it in its weakest condition, and disasters would occur such as would not happen at a later period. He agreed with the statement of the author with regard to the failure of retaining walls. No doubt, in ninety cases out of a hun- dred, the failure happened from bad foundations. The remedy in railway works was in many cases that shown in Fig. 17, which practically amounted to strutting the toe of the wall against the .opposite wall, and so preventing it slid- ing forward. That was a very simple arrangement, and resembled in its effect the strutting of timber, which was gen- erally carried out as a temporary meas- ure by a contractor, when, an invert was going to be put in. If the engineer , -:., - . --.-:: -:::.: -r - ..__..^ _. .. _ . ... -'.' .._.. . _ :. . ^~.. 1 -- . V '. : giuond rmQwmr Of 8UM Of cencdin It: and HWT for Ike foDnegs uid ' Wt latter pax* he slated that hi had ptelly wall accorded ittlli I3G For instance, he gave the theoretical thickness for a retaining wall in ground that naturally stood at a slope of 1^ to 1 as 31 per cent, of the height ; and in the last paragraph but one he said his habit had been to make his walls ^, or 33 per cent.; and in the Table with slopes from 1 to 1 to 4 to 1, which included all that engineers usually had to deal with, his theoretical thickness ran from 0.239 to 0.451, while in the concluding para- graphs of the paper he stated that the engineer must work between the limits of ^ the thickness and , which seemed to agree with the theoretical thickness. The general conclusion that engineers must work between J and . was differ- ent from the practice in which Mr. Lewis had been trained, and he had therefore brought a diagram (Fig. 43) of a retain- ing wall constructed according to Mr. Brunei's rules. Of course Mr. Brunei, who had to carry out very great works, modified his rules to suit the circum- stances ; but the diagram represented his standard section of wall such as was 137 constructed at Lord Hill's land in the early days of the Great Western railway, and at the Britain Ferry docks two years before his death. It would be seen that Fig.43 8cale,lC feet = 1 inch. the dimensions and peculiarities of that wall differed very much from those given in the paper. In the first place the wall had an average batter of 1 in 5, and at the top a batter of 1 in 10. Batter 138 was a point on which Mr. Brunei al- ways insisted, and Mr. Lewis was a little surprised that the author seemed to treat it with so much indifference. He was evidently aware of its value, because in the early part of the paper he mentioned a wall with a batter of 1 in 5, and a thickness of 1 foot, which he said was^ equivalent to a vertical wall of 1 foot 9 inches. Now anything that was equiva- lent to an increase of the original value of 73 per cent, was well worthy of con- sideration. Mr. Brunei's custom was to curve the face of the wall. The radius was 150 feet in the case of a 30 feet wall, or five times the height. The thickness was -J- to J- the height. The counterforts were 2 feet 6 inches thick, and placed 10 feet apart from center to center, but were omitted in good clay cuttings. In the case of docks sometimes there was a difficulty, in consequence of the neces- sity of having the top more upright, and at Britain Ferry docks the radius was reduced by nearly one-half. Mr. Brunei, too, was in ihe habit of building 139 behind what he called sailing courses and the projections in Fig. 43 were 1 foot 3 inches. In the case of embank- ments the wall was supported by earth carefully punned against it and against the sailing courses, thereby adding con- siderably to the weight that had to be overturned when pressure came from be- hind. Then his rule for thickness was i, which was below the minimum given by the author. There were a number of such walls at Paddington, Bath, Ply- mouth, Briton Ferry, 30 feet high and 5 feet thick, and generally of nearly the same thickness at the top as at the bot- tom. Another point Mr. Brunei was- particular about was that the footings were made square to the batter, and when the ground was not good consider- ably larger footings were introduced. At Briton Ferry a 2- feet lining of con- crete was employed at some places for watertightness. Concrete was not then in such general use as it was at present, Of course when exceptional ground was met with it was dealt with exceptionally. 140 At a tunnel on the Wilts, Somerset, and Weymouth railway, some heavy ground had been found ; the tunnel mouth was in a 60-feet cutting, a retaining wall 30 feet high was built, and the top was sloped back at | to 1, with a 2-feet cov- ering of masonry, and the wall was built precisely of the dimensions represented by Fig. 43 ; but as the ground was heavy, the batter, instead of being 1 in 5, was 1 in 4, and that was the only alteration. That wall was builfc in 1854, had never given any trouble, and was standing at the pres- ent moment. It seemed to him that Mr. Brunei, forty years ago, came nearer to the teaching of the experiments and of the reasoning in the paper, than the au- thor had ventured to do in his own prac- tice. Mr. J. B. EEDMAN observed that the author had undoubtedly filled a void in the literature of engineering; for, not- withstanding the great experience that most of the members of the Institution had of such catastrophes as those which had been referred to, it was only human 141 like that they had not been often record- ed by the designers of the works. Those who constituted what was now a select minority of the Institution would re- member the partial failures of Mr. Bob- ert Stephenson's retaining walls in the Euston cutting of what was then the London and Birmingham, and now the London and North-Western railway. Those partial failures were met by over- head horizontal girder struts supporting the walls, and it was rather curious that, notwithstanding all the experience that had been since gained, in a large number of instances, in metropolitan railways, the overhead girder had been, as it were, the natural sequence of what might be termed the unretaining wall. There was one circumstance which very much com- plicated the question of the direct lateral thrust of earthwork upon a retaining wall, and which rather curiously had not been mentioned by the author. It was incidentally referred to in the latter part of the paper where the author said French engineers, in designing a wall at 142 Marseilles, made the width of the base 58 per cent, of the vertical height, in consequence of the dip of the strata be- ing towards the wall. In a large num- ber of cases of the failures of retaining walls in open cuttings near London, he thought it would be found that the fail- ure was entirely on one side. Where the dip of the strata was towards the cutting, and more especially if there were lamince of clay, the superimposed strata often struck near the base of the wall ; and a retaining wall on that side not only had to support the normal lat- eral thrust of the mass of earthwork im- mediately behind, but it had also a long wedge-like piece of earth impinging against the earth at the back of the wall, so that in many cases the thrust on the wall at the one side must be something like double the amount that it was on the other ; because on the other side, the dip being away from the wall, the wall was subject only to the lateral thrust of the earthwork in its rear. The author had stated that the failures of many dock walls did not illustrate entirely the ordinary lateral thrust of earthwork ; but Mr. Redman thought thafc such cases as the failure of the walls constructed by the late Mr. G. P. Bidder, Past-Presi- dent Inst. C.E., at the Blackwall entrance to the Victoria docks, the partial failure of the same engineer's walls in the en- largement of the Surrey Docks, the simi- lar catastrophe at the Victoria dock, Hull, in the work designed by the late Mr. John Hartley, and possibly also a similar movement in the South West India Dock wall, were all clearly attributable to lat- eral thrust. It might be said that the foundation was not taken down deep enough, and consequently the wall did not resist that thrust,; but having had a somewhat extended and varied experi- ence for a great number of years, he * certainly was not prepared to indorse the dogma that a dock wall or a river wall must necessarily be so strong as to resist a head of water, or in width at the base equal to one-half the height. In the first place, the water ought not to be al- 144 lowed to come behind the wall. There were exceptional cases, perhaps, where that could hardly be avoided; but it seemed to him that laying down such a tenet was a premium for loose engineer- ing, imperfect supervision, and lavish expenditure. He had himself, in the lower reaches of the Thames, erected some of the heaviest embankment walls on the river, where the thickness was only J of the vertical height. It was true that the walls were founded on the best possible foundation Thames ballast and it was done as tide work ; and the greatest possible care was also taken to keep the backing up to the same level as the wall, and indeed rather above the wall. In fact, the great mistake in re- taining walls was the imperfect supervi- sion exercised over the backing. If the backing were put in with tolerably fair material in thin horizontal layers and brought up in that way, the lateral thrust was reduced to a very small mat- ter. The author had stated that the de- cayed timber wharves on the Thames and 145 in other neighborhoods showed that the lateral thrust must be over-estimated ; but it should be remarked that the skin might be stripped off the face of the earthwork, assuming that no water was coming against it, and it would stand, because from the length of time and consolidation of material, there was no lateral thrust. The example quoted of the breakwater at St. Katherine's, Jer- sey, appeared to be a case in point. He had nothing to do with the inception or execution of that work; but he thought the wall might be taken down and the heart of the pier would still stand. He would refer to two great Metropolitan failures which were well known, and which might be interesting in illustra- tion of this subject. One was that of Greenwich Pier and the other of the. Island Lead Works. The Greenwich Pier was constructed nearly half a cen- tury ago from the design of a local archi- tect, Mr. Martyr. It was one of the heaviest embankments on the Thames ; it had the greatest depth of water up to 146 it, and it was, being in the hollow of the reach, subject to every condition of weather. The base was formed by cast- iron piling and cast-iron sheeting be- tween, constituting a half-tide dam, and concrete was got in behind. Upon the top of the concrete there were large 6-inch York landings and a very solid, heavy brick wall. There were also outer piles, and the work was constructed in the best possible way. The case was somewhat complicated by the fact that a large amount of land-water came down and a large amount of spring water. There was a common sewer running through the heart of the work, and a large tidal reservoir for the Ship Hotel. The whole of that work, with the exception of the two returns and quoins and a small por- tion in front of the Ship Hotel, slipped into the river during the night some forty years back. The late Mr. Chad- wick, who built the Hungerford suspen- sion bridge, entered into a contract to restore the work on his own plan, acting as engineer and contractor, and he re- 147 stored the portion that had failed with timber-bearing piles and a solidly-con- structed brick wall. Shortly after the demise of Mr. Chadwick, the restored portion showed signs of failure, and Mr. Redman was called in by the Pier Direct- orate, and the matter resulted in a law- suit, and a large sum of money was ob- tained in compensation. All he did was to bleed the pier by inserting a cast- iron pipe with a self-acting flap at the eastern end, and to remove and sub- stitute with better material some part of the backing. He proposed driving land- tie piles at the back and some in front ; but on consideration with the Director- ate, it was thought that driving piles might be a ticklish operation. That was twenty years ago, and up to the present time the work had remained in the same state. It had settled somewhat at the eastern end, and there were reopened fissures in front, so that the movement had not altogether ceased. The wall of the Island Lead Works designed by the late Mr. K. Sibley, M. Inst. C.E., was the 148 pioneer of cast-iron wharfing ; and from the fact of the Limehouse cut having been deepened too close up to it, the wall failed. As the author had said, that case did not illustrate the absolute lat- eral pressure of earthwork, because this work, as long as it was not meddled with, stood satisfactorily. The leaseholders called in Mr. Redman on that occasion, and the freeholder consulted Mr. Bate- man, Past-President Inst. C.E., and the late Mr. N. Beardmore, M. Inst. C.E., Tery wisely to avoid a lawsuit con- structed a wall deeper down, to their satisfaction. Mr. W. ATKINSON agreed with Mr. Lewis's remarks with respect to the large amount of masonry or brickwork that the author had introduced in the cases of the metropolitan railways. He had been much struck with the propor- tion of J of the height for the mean thickness of a wall ; but looking at the diagrams, and taking into consideration what he had seen of the work, there was a very good explanation. It struck him 149 that on the Metropolitan railway, where property was so valuable, the batter which the late Mr. Brunei introduced of 1 in 5 would be extremely inconvenient ; either the roadway would have to be narrowed, or a great deal more property would have to be taken, than would be otherwise necessary. No doubt the author would be able to say whether that had any influence in the carrying out of the work. Then with regard to the gen- eral question of the walls and their fail- ure due to bad foundations, it struck him that the two things should be en- tirely separate ; that the foundation should be treated as a foundation, and that having been made sufficiently strong, a properly proportioned wall should be placed upon it. He rather gathered^ from the paper that the two points had been taken as a whole, and that the au- thor meant, "I have a bad foundation, and I will make the whole to stand." If that were so, it would have been better policy to have made a foundation of con- crete, and then put a wall sufficiently 150 strong. With regard to the question of theoretical calculation, there was a French formula which agreed remark- ably with what might be considered the ordinary practice. He himself had put up a good many walls, not perhaps as distinct retaining walls, but in connec- tion with bridges on 48 miles of the Mid Wales railway, and he had found practi- cally that the ^ of the height for the mean thickness stood perfectly well. In that case, it was to be borne in mind that there were two elements in addition to the theoretical calculation, namely, the projection of the footings where there was so much leverage, which was not taken into account in the calculation, and the weight of the earth resting on the projections or steppings at the back of the wall, Fig. 44. That, of course, aided the wall very materially ;* in fact, it might be called so much masonry saved. At all events, if merely the theoretical thickness of the wall was given, then, with the projections of the footings, and the weight of earth on the steppings, 151 there was a very good margin of safety; and in that way the wall was erected with J-^, or 33 per cent, less than the dimension advocated by the author, and Fig.44 Scale JG feet = 1 inch. was a good and sufficient wall. One point with regard to walls was brought to his notice when in Canada, namely, the thickening of the top to resist frost. In ordinary circumstances the practice 152 would be to put about 2 feet at the top, and then about 9 feet down a projection of 9 inches, and so on ; but in Canada, on account of the penetration of the frost, it had been found necessary to make the top of the wall much thicker than was the practice in England. Mr. H. LAW desired to add his testi- mony to the great value of the facts laid before the Institution. It was upon such facts, the result of actual experience, that the most valuable data were formed. In the early part of the paper the author had pointed out that the formula usually adopted Coulomb's did not give the results which were obtained when loosely heaped materials were placed at the back of the wall; but a little consideration would show that that formula never was intended to apply to such cases. Cou- lomb's theorem distinctly took into ac- count the adhesiveness or coherence of the ground, and then determined, de- pending upon the line on which the ground separated, what the amount of pressure would be; and the value to the 153 engineer was, that it determined what was the maximum which that pressure could be. Putting w=ihe weight of a cubic foot of the soil in Ibs., h = the height of the wall in feet, r = the limit- ing angle of resistance of the soil, s = the angle between the line at which the soil separated and the horizontal, and P = the horizontal pressure in Ibs. of the soil against the wall, then Coulomb's theorem might be thus expressed: P=:-^ . cot s. tan(s r). 2 Now in the case of a fluid, r, or the limiting angle of resistance, vanished, and consequently the result was that the co-tangent of s into the tangent of s became equal to unity, and _rf " 2 ' When the ground was sufficiently co- herent to stand vertically, then the angle of separation being 90 the co-tangent of s became nothing, and the pressure became nothing. When the line of 154 separation coincided with the limiting angle of resistance or r, that was to say, when there was a mass of earth suffi- ciently coherent not to break of itself, and lying upon a bed which happened to be at the limiting angle of resistance, the tendency of the earth to slide was exact- ly overcome by its friction, and r being equal to s, the tangent vanished, and P again became nothing. Now, between tho^e two values there was a certain angle at which, if the ground separated, it would produce the maximum pressure, and that was given by Coulomb's theorem, which proved that when the line of sep- aration bisected the angle made by the limiting angle of resistance with the vertical, then cots=tan(s r), and and the maximum pressure was obtained. The great value of the formula was to show, with a given weight of earth and a given limiting angle of resistance, what the maximum pressure was. It could 155 not exceed the value expressed by making s half the angle between the limiting angle of resistance and the verticaJ. This formula could not be applied in the case of loose materials, as sand and gravel, because it was impossible for such ma- terials to stand at any other than than their limiting angle of resistance ; and under such circumstances there would be upon the wall only a comparatively small pressure, due to the unbalanced weight which remained from the efforts of the sand and the gravel to roll down upon itself. He wished to direct attention to one or two interesting exemplifications of excessive pressure which were met with in the works for the Thames tunnel. The Rotherhithe shaft, 50 feet in diameter, was built upon the surface and sunk by excavating beneath. That operation was* successful until a depth of 40 feet was reached, and then, although the exterior surface had been made perfectly smooth by being rendered, it became earth- bound, and notwithstanding the earth w as excavated to a depth of 2 feet round 156 the whole margin, and 50,000 bricks were placed upon the top as a load, making the total weight 1,100 tons, and water was allowed to rise inside, the shaft refused to sink any farther. Now, taking the weight of the ground at 120 Ibs. per cubic foot, which was about what it was on the average, and taking the coefficient of friction at 67, it would be found that a limiting angle of resistance of about 31 15', and a line of fracture of about 27 30', would show, by Coulomb's theorem, that the shaft would be bound, and therefore the practical result was quite in accordance with the pressure given by the formula. The author had mentioned a case of some heavy clay which had a pressure equivalent to a fluid pressure of 107 Ibs., and if that clay was taken as having a limiting angle of resistance of about 5 or 1 in 10, and the weight was assumed to be 130 Ibs. per cubic foot which clay of that descrip- tion might very well have the formula would give 107 Ibs. for the fluid pressure. He therefore thought these circumstances 157 fully showed that where ground was co- herent and adhesive, Coulomb's theorem applied. In the progress of the Thames tunnel there had been some remarkable cases of excessive pressure, where of course the weight of the water was super- added to that of the ground. He knew many instances of poling boards, 3 feet in length, 6 inches wide, and 3 inches thick, supported by two poling screws bearing against cast iron plates, being split lengthwise by the pressure of the earth against the outer surface. Mr. E. A. BEENAYS said the inconsis- tencies alluded to in the paper tended to make it still more interesting than it otherwise would have been. There were few engineers who had carried out works, but were conscious of inconsist- encies in their own practice and theories* The author had quoted M. Voisin Bey, the distinguished French engineer, as saying that he had rarely seen a long wall straight, and Mr. Bernays' expe- rience fully confirmed that view. When it was straight the chances were there 158 was a superabundance of material to keep it so. If it was run fine, as the calcula- tions advised, the chances were 50 to 1 against having a straight wall. With re- gard to Mr. Brunei's section of wall, no doubt if it had a good foundation it was very strong for the material in it. It not only had a rising abutment to bring the pressure down upon the foundation, but it had counterforts, which added greatly to the strength .of the wall r although of late they had gone out of fashion. He considered it was nearer 10 feet at the base than 5 feet, as, if the counterforts were 10 feet apart, the wall was, practically, a solid wall. If made of concrete instead of brickwork, it would probably be found better to make it solid at once. The batter added consider- ably to the strength, but it was not without practical disadvantages. The greater the batter the greater the disad- vantage. The tendency of the batter was to throw the side of a vessel farther away from the wall than need be, and to entail cranes with longer jibs, as well as the use 159 of much larger fenders. Iron ships were now all covered with anti-fouling com- position, which might easily be scraped off. With all its disadvantages he would rather have a smaller batter for practical purposes when ships were to lie along- side the wall. He had seen a wall of this section in Woolwich Dockyard (built, he believed, by Sir John Bennie, Past- President Inst. C.E.), partially pulled down and refaced by the late Mr. James Walker, Past-President Inst. C.E., for the purpose of deepening the dock. It was about 30, feet deep, and was increased to about 38 feet by putting a thin wall in front of it. In pulling clown such walls he had always found that the backing in settling hung upon the set off, and he had seen holes under the backing large enough for a man to creep in. He would not say that they were objectionable in other respects, but he preferred a battered back to a retaining wall to square sets off. The author had alluded to a wall that he was building, and had character- ized it as 4 1>. VAN NOSTRAND'S PUBLICATIONS. MacCORD. A PRACTICAL TREATISE ON THE SLIDE VALVE, BY ECCENTRICS examining by methods the action of the Eccentric upon the Slide Valve, ami explaining the practical processes of laying out the move- ments, adapting the valve for its various duties in the steam-engine. By C. W. Mac Cord, A. M., Professor of Mechanical Drawing, Stevens' Institute of Technol- ogy, Hoboken, N. J. Illustrated. 4to, cloth $3 00 PORTER. A TREATISE ON THE RICHARDS' STEAM-ENGINE INDICATOR, and the Devel- opment and Application of Force in the Steam-Engine. By Charles T. Porter. Third edition, revised and enlarged. Il- lustrated. 8vo, cloth, . . 3 50 McCULLOCH A TREATISE ON THE M^IANI- CAL THEORY OF HEAT, AND ITS APPLICA- TIONS TO THE STEAM-ENGINE. By Prof. R. S. McCulloch, of the Washington and Lee University, Lexington. Va. 8vo, cloth, . 3 50 VAN 6UREN. INVESTIGATIONS OF FORMU- LAS for the Strength of the Iron parts of Steam Machinery. By J. D. Van Buren, Jr., C. E. Illustrated. 8vo, cloth, . 2 00 STUART. Ho iv TO BECOME A SUCCESSFUL EN- GINEER. Being Hints to Youths intending to adopt the Profession. By Bernard Stuart, Engineer Sixth edition 18mo, boards, ..... 50 SHIELDS. NOTES ON ENGINEERING CONSTRUC- TION. Embracing Discussions of the Prin- ciples involved, and Descriptions of the Material employed in Tunneling, Bridging Canal and Road Building, etc., etc. By J. E. Shields, C. E. 12mo. cloth, . . . 1 50 D. VAN NOSTRAND'S PUBLICATIONS. jVEYRAUCH. STRENGTH AND CALCULATION OF DIMENSIONS OF IRON AND STEEL CON- STRUCTIONS. Translated from the German of J. J. Weyrauch, Ph. D., with four fold- ing Plates. 12nio, cloth, . . . $1 00 STUART. THE NAVAL DRY DOCKS OF THE UNITED STATES. By Charles B. Stuart, Engineer in Chief, U. S. Navy. Twenty- tour engravings on steel. Fourth edition. 4to, cloth, . . . . . 6 00 COLLINS. THE PRIVATE BOOK OF USEFUL AL- LOYS, and Memoranda for Goldsmiths, Jewellers, etc. By James E. Collins. 18nio, flexible cloth, 50 TUNNER. A TREATISE ON ROLL-TURNING FOR THE MANUFACTURE OF IRON. By Peter Tun- ner Translated by John B. Pearse. With numerous wood-cuts, 8vo, and a folio Atlas of 10 lithographed plates of Rolls, Measurements, &c. Cloth, . . 10 00 GRUNER. THE MANUFACTURE OF STEEL. By M. L. G rimer. Translated from the French, by Lenox Smith, A.M., E.M. ; with an Appendix on the Bessemer Pro- cess in the United States, by the transla- tor. Illustrated by lithographed drawings and wood-cuts. 8vo, cloth 3 50 BARBA. THE USE OF STEEL IN CONSTRUCTION. Methods of Working, Applying, and Test- ing Plates and Bars. By J. Burba. Trans- lated from the French, with a Preface by A.L.Holley,P.B. Illustrated. 12mo, cloth. 1 50 SHOCK. STEAM BOILERS ; THEIR DESIGN. CON- STRUCTION, AND MANAGEMENT. Hy William A. Shock, Engineer-in- Chief, U.S.N.; Chief of Bureau of Steam Engineering, U.S. N". 480 pages. Illustrated with 150 woodcuts and 80 full-paste plates (30 double). 4to, illustrated, half morocco, ] 5 CO 6 D. VAN NOSTIIAND'S PUBLICATIONS. WARD. STEAM FOR THK MILLION. A Popular Treatise on Steam and its Application to the Useful Arts, especially to Navigation. By J. H. Ward, Couiinauder U. S. Navy. 8vo. cloth. . . . . . $1 00 CLARK. A MANUAL OF RULES, TABLES AND DATA FOR MECHANICAL ENGINEERS. Rased on the most recent investigations. By Dan. Kinnear Clark. Illustrated with numerous diagrams. 1012 pages. 8vo. Clot h. $7 50; half morocco, . . . .1000 JOYNSON. THE METALS USED IN CONSTRUC- TION : Iron, Steel, Bessemer Metals, etc, By F. H. Joynson. Illustrated. 12mo, cloth, 75 DODD. DICTIONARY OF MANUFACTURES, MIN- ING, MACHINERY, AND THE INDUSTRIAL ARTS. By George Dodd. I2mo, cloth, 1 50 PRESCOTT FIRST BOOK ix QUALITATIVE CHE- MISTRY. By Albert K PiescoU, Professor of Organic ard Applied Chemistry in the University of Michigan. 12mo, cloth, . 1 50 HLATTNER. MANUAL OF QUALITATIVE AND QUANTITATIVE ANALYSIS WITH THE BLOW- PIPE. From the last German edition. Re- vised and enlarged. By Prof. Th. Richter, o the Royal Saxon Mining Academy. Translated by Professor H. B.Cornwall. With eighty-seven wood-cuts and lithogra- , phic plate. Third edition, revised. 568pp. 8vo, cloth, . . . , . . . . 5 00 PLYMPTON. THE BLOW-PIPE : A Guide to its Use in the Determination of Salts and Minerals. Compiled from various sources, by George W. Plympton, C. E., A. M., Pro- fessor of Physical Science in the Polytech- nic Institute, Brooklyn, N. Y. 12ino, c/.oth 1 50 7 D. VAN NOSTRAND'S PUBLICATIONS. JANNETTAZ. A GUIDE TO THE DETERMINATION OF ROCKS ; being an Introduction to Lith- ology. By Edward Jannettaz, Docteur des Sciences. Translated from the French by G. W. Plyuiptou, Professor of Physical Science at Brooklyn Polytechnic Institute. 12mo, cloth, $1 50 MOTT. A PRACTICAL TREATISE ON CHEMISTRY (Qualitative and Quantitative Analysis), Stoichioinetry, Blowpipe Analysis, Min- eralogy, Assaying, Pharmaceutical Prepa- rations Human Secretions, Specific Gravi- ties, Weights and Measures, etc., etc., etc. By Henry A, Mott, Jr., E. M., Ph. D. 650 pp. 8vo, cloth, 6 00 PYNCHON. INTRODUCTION TO CHEMICAL PHY- SICS ; Designed for the Use of Academies, Colleges, and High Schools. Illustrated N with numerous engravings, and containing copious experiments, with directions for preparing them. By Thomas Ruggles Pyii- chon, D. D., M. A., President of Trinity Col- lege, Hartford. New edition, revised and enlarged. Crown 8 vo, cloth, . . 3 00 PRESCOTT. CHEMICAL EXAMINATION OF ALCO- HOLIC LIQUORS. A Manual of the Constit- uents of the Distilled Spirits and Ferment- ed Liquors of Commerce, and their Quali- tative ami Quantitative Determinations. By Alb. B. Prescott, Prof, of Chemistry, University of Michigan. 12mo, cloth, . 1 50 ELIOT AND STORER. A COMPENDIOUS MANUAL OF QUALITATIVE CHEMICAL ANALYSIS. By Charles W. Eliot and Frank H. Storer. Re- vised, with the co operation of the Authors, by William Riploy Nichols, Professor of Chemistry in the Massachusetts Institute of Technology. New edition, revised. II- lustrated. I2mo, cloth, . . . . 1 50 8 D. VAN NOSTRAND'S PUBLICATIONS. NAQUET. LEGAL CHEMISTRY. A Guide to the Detection of Poisons, Falsification of Writ- ings, Adulteration of Alimentary and Phar- maceutical Substances ; Analysis of Ashes, and Examination of Hair, Coins, Fire-arms and Stains, as Applied to Chemical Juris- prudence. For the Use of Chemists, Phy- sicians, Lawyers, Pharmacists, and Ex- perts. Translated, with additions, includ- ing a List of Books and Memoirs on Toxi- cology, etc., from the French of A. Naquet, by J. P. Battershall, Ph. D. ; with a Preface by C. F. Chandler, Ph. D., M. D., LL. D. Illustrated. 12mo, cloth, . . . . $2 00 PRESCOTT. OUTLINES OP PROXIMATE ORGANIC ANALYSIS for the Identification, Separa- tion, and Quantitative Determination of the more commonly occurring Organic Compounds. By Albert B. Prescott, Pro- fessor of Chemistry, University of Michi- gan. 12mo, cloth, . . . 1 75 WUGLAS AND PRESCOTT. QUALITATIVE CHEM- ICAL ANALYSIS. A Guide in the Practical Study of Chemistry, and in the work of Analysis. By S. H. Douglas and A. B. Prescott; Professors in the University of Michigan. Third edition, revised. 8vo, cloth, 3 50 RAMMELSBERG. GUIDE TO A COURSE OF QUANTITATIVE CHEMICAL ANALYSIS, ESPE- CIALLY OP MINERALS AND FURNACE PRO- DUCTS. Illustrated by Examples. By C. F. Rammelsberg. Translated by J. Tow- ler, M. D. 8vo, cloth, 2 25 BEILSTEIN. AN INTRODUCTION TO QUALITATIVE CHEMICAL ANALYSIS. By F. Beilstein. Third edition. Translated by I. J. Osbun. I2mo. cloth, 75 POPE. A Hand-book for Electricians and Oper- ators. By Frank L. Pope. Ninth edition. Revised and enlarged, and fully illustrat- ed. 8vo, cloth, ' 2 OO D. VAN NOSTRAND'S PUBLICATIONS. SABINE. HISTORY AND PROGRESS OF THE ELEC- TRIC TELEGRAPH, with Descriptions of some of the Apparatus. By Robert Sabine, C.E. Second edition. 12mo, cloth, . . $1 25 DAVIS AND RAE. HAND BOOK OF ELECTRICAL DIAGRAMS AND CONNECTIONS. By Charles H. Davis and Frank B. Rae. Illustrated with 32 full-page illustrations. Second edi- tion. Oblong 8vo, cloth extra, . . . 2 00 HASKINS. THE GALVANOMETER, AND ITS USES. A Manual for Electricians and Students. By C. H. Haskins. Illustrated. Pocket form, morocco, , 1 50 LARRABEE. CIPHER AND SECRET LETTER AND TKLEGRAPAIC CODE, with Hogg's Improve- ments. By C. S. Larrabee. 18mo, flexi- ble cloth, . . . . ... . . 1 00 GILLMORE PRACTICAL TREATISE ON LIMES, HYDRAULIC CEMENT, AND MORTARS. By Q. A. Gillmore, Lt.-Col. U. S. Engineers, Brevet Major-General U. S. Army. Fifth edition, revised and enlarged. 8vo, cloth, 4 00 GILLMORE. COIGNET BETON AND OTHER ARTIFI- CIAL STONE. By Q. A. Gillmore, Lt. Col. U. S. Engineers, Brevet Major-General U. S. Army. Nine plates, views, etc. 8vo, cloth, 2 50 GILLMORE. A PRACTICAL TREATISE ON THE CONSTRUCTION OF ROADS, STREETS, AND PAVEMENTS. By Q. A. Gillmore, Lt.-Col. U. S. Engineers, Brevet Major-General U, S. Army. Seventy illustrations. 12mo, clo., 200 GILLMORE. REPORT ON STRENGTH OF THE BUILD- ING STONES IN THE UNITED STATES, etc. 8vo, cloth, I 00 HOLLEY. AMERICAN AND EUROPEAN RAILWAY PRACTICE, in the Economical Generation of Steam. By Alexander L. Holley. B. P. Wiih 77 lithographed plates. Folio, cloth, 12 00 10 D. VAN NOSTRAND'S PUBLICATIONS. HAMILTON. USEFUL INFORMATION FOR RAIL- WAY MEN. Compiled by W. G. Hamilton, Engineer. Seventh edition, revised and en- larged. 577 pages. Pocket form, morocco, gilt, , . $2 00 STUART. TFIE CIVIL AND MILITARY ENGINEERS OF AMERICA. By General Charles B. Stuart, Author of "Naval Dry Docks of the United States," etc., etc. With nine finely-executed Portraits on steel, of emi- nent Engineers, and illustrated by En- gravings of some of the most important and original works constructed in Ameri- ca. 8vo, cloth, 5 00 ERNST. A MANUAL OF PRACTICAL MILITARY ENGINEERING. Prepared for the use of the Cadets of the U. 8. Military Academy, and for Engineer Troops. By Capt. O. H. Ernst, Corps of Engineers, Instructor in Practical Military Engineering, U. S. Military Academy. 193 wood-cuts and 3 lithographed plates. I2mo, cloth, . 5 00 SIMMS. A TREATISE ON THE PRINCIPLES AND PRACTICE OF LEVELLING, showing its ap- plication to purposes of Railway Engineer- ing and the Construction of Roads, etc. By Frederick W. Simms, C. E. From the fifth London edition, revised and correct- ed, with the addition of Mr. Law's Prac- tical Examples for Setting-out Railway Curves. Illustrated with three lithograph- ic plates, and numerous wood-cuts. 8vo, cloth, 2 50 JEFFERS. NAUTICAL SURVEYING. By William N. Jeffers, Captain U. S. Navy. Illustrat- ed wiW 9 copperplates, and 31 wood-cut illustrations. 8vo, cloth 5 00 THE PLANE TABLE. ITS USES IN TOPOGRAPHI- CAL SURVEYING. From the papers of the U. S. Coast Survey. 8vo, cloth, . 2 00 D. VAN NOSTRAND'S PUBLICATIONS. ELLIOT. EUROPEAN LIGHT-HOUSE SYSTEMS. Being a Report of a Tour of Inspection made in 1873. By Major George H. Elliot, U. 8. Engineers 51 engravings and 21 wood-cuts. 8vo, cloth, $5 00 SWEET. SPECIAL REPORT ON COAL. By S. H. Sweet. With Maps. 8vo, cloth, . 3 00 COLBURN. GAS WORKS OF LONDON. ByZerah Colburn. 12mo, boards 60 WALKER. NOTES ON SCREW PROPULSION, its Rise and History. By Capt. W. H. Walker, U.S. Navy. 8vo, cloth, .... 75 POOR. METHOD OF PREPARING THE LINES AND DRAUGHTING VESSELS PROPELLED BY SAIL OR STEAM, including a Chapter on Laying- offou the Mould-loft Floor. By Samuel M. Pook, Naval Constructor. Illustrated. HVO, cloth, ....... 5 00 SAELTZER. TREATISE ON ACOUSTICS in connec- tion with Ventilation. By Alexander Saeltzer. 12mo, cloth, 1 00 ASSIE. A HAND-HOOK FOR THE USE OF CON- TACTORS, Builders, Architects, Engineers, Timber Merchants, etc., with information for drawing up Designs and Estimates. 250111 stratious. 8vo, cloth, . 1 50 SCHUMANN. A MANUAL OF HEATING AND VEN- TILATION IN ITS PRACTICAL APPLICATION for the use of Engineers and Architects, embracing a series of Tables and Formulae for 'iiiaeusions of heating, flow and return Pipes for <*team and hot water boilers, flues, etc . e*/< Bv F Schumann, (1. E.. U. S. Treasurv Department 12ino. Illustrated Full roan, , . 1 50 TONER. DICTIONARY OF ELEVATIONS AND CLIMATIC REGISTER OF T|[E UNITED STATES. By J. M. Toner, M D. 8To, Paper, $3.00; cloth. . . . 375 THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL PINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO SO CENTS ON THE FOURTH DAY AND TO $1.OO ON THE SEVENTH DAY OVERDUE. 1939 RBC*D UP- YA 01463 THE VAN NOSTRAND SCIENCE SERIES. 0> 72.-TOPOGRAPICAL SURVEYING. By Geo. J. Specht, Prof. A. S. Hardy, John B. McMaster and H. F Walling. o. 73.-SYMBOLIC ALGELRA ; or The Algebra of Algebraic Numbers. By P* of. W. Cain. o. 74. TESTING MACHINES; their History, Construction and Use . By Arthur V . Abbott . o. 75.-RECENT PROGRESS IN DYNAMO- ELECTRIC MA- CHINES. Being a Supplement to Dynamo-Elec- tric Machinery. By Prof. 8ilvanusP. Thompson, o. 76- MODERN REPRODUCTIVE GRAPHIC PRO- CESSES. By Lt. Jas. S. Pettit, U. S. A. 3. 77. -STADIA SURVEYING. The Theory of Stadia Measurements . By Arthur Winslow . >. 78.- THE STEAM ENGINE INDICATOR, and its Use. By W. B. LeVan. ). 79.-THE FIGURE OF THE EARTH. By Frank C Roberts, C. E. ). 80.-HEALTHY FOUNDATIONS FOR HOUSES, ^v Glenn Brown. -Y >. 81. WATER METERS : Comparative Tests of Ar , Delivery, etc. Distinctive Features of t v ington, Kennedy, Siemens and Hesse ' Ross E. Browne. . 82 THE PRESERVATION OF TIMBF of Antiseptics. By Samuel Bags' . 83.-MECHANICAL INTEGRATORS. S. H. Shaw, C. E. >. 84.-FLOW OF WATER IN OPEN CHA* JJfc CONDUITS, SEWERS, &c. ; X P. J. Flynn, C. E. .. 85.-THE LUMINIFEROUS MTV'' son Wood . . 86. HANDBOOK OF MINF and Description of ?' States. F rof. ? 87.TREATIS" ^ ^^^ STRU , E AR^ . 88. BB/ .ias for . 89.- jture, Prop- kisser, U.S. A. ' 90-\ the Gyroscope. , 91. LEVELING : Barometric, Trigonometric and Spirit. By Prof. I. O Baker. . 92. PETROLEUM : Its Production and Use. ;, 93, SANITARY DRAINAGE of Buildings. ,. 94. THE TREATMENT OF SEWAGE. By Tidy. . 95. PLATE GIRDER CONSTRUCTION. By Hiroi. TTI. BASIS OF LIFE. By Prof. T. H. HUXLEY, LL.B. I cents. ^ V. bCIi'-.-NTIFiO Ar ? E3SSES: 1. Cn ^ Jfc thods arul Tzndznc,' ;f nysical Invest ; (ia -it, .*,. 2. On Haze, and I* '. .t, i,<. Scientific the of the Imagi- ''.on. By*^ , OHN TYNDAI-L, F.R.S. 12^o, 7 pp. Pap-r Co\t.^ ^''rice 25 cents. .Blex. Cloth. 50 ots.' NO. VI.N// - 4L SELECTION AS APPLIED TO 'lAN. ."By / LFRED RuSSEi.^ WALLACE. This prtmpl. 1 ':. ti.rttp v . ) of the Development of Human Races under the la f 01 *!eciirjn ; (2 )*"" jimits of IN at- v \i i al Selection as applied , ; nan. 54pp. Price 25 otnty. ^^ , IL 81 171 RUM AN/ LYSIS. Threo lec- , t ^ V,V"PT.>IS, ^oe, Huggins, and Lc kye. . ^ine- lyiJ.*^.,,, f.d. bo pp. Paper Covsrb. _'"' ^ 25 cen^. Is'O. ' ' -THE SUN. A sketch the present state of sH^ti^c .:voinion as regards this > >u/, "with an '>-i t}.~ or rscenr ^iscoveric ., nnd methods of ; YOUNG, Ph.]cJ ., of jOa-t- nthCoJ tr Cov 1* ..:er CoverH. ' >r , sfH-.ite. ^L r^os'''. a. tliv^ ^a.bject of mag 1 - Price 25 cents. Fle?> . .\ A.. .TYSTEFJ^^ C/J x^^; VOICE AND jf."AR. .13^ ^'^-v, 0. !N. OOD, Cul'imbia ^olle^e, N* - ".''. One 01 ,"\-3 most interesting lee tares on S.^T . 1 " -"*- ^ ^'iv^rinal discoveries, b?'* n iani txneri- in^uts i>. . tly iilus. 38 pp. l- t .^r Cov. ? 25 cts.