NOTES ON TRACK CONSTRUCTION AND MAINTENANCE BY W. M. CAMP, Editor of the Railway and Engineering Review; Member of the American Society of Civil Engineers. More than 600 Illustrations PUBLISHED BY THE AUTHOR, AT AUBURN PARK, CHICAGO. 1903 GENERAL Copyright, 1903, by W. M. Camp. PREFACE. What I have attempted to do in the following pages is to treat the con- struction and maintenance of railroad track from the standpoints of both the trackman and the engineer. I am led to this from the belief that the thorough trackman must necessarily be able to comprehend some of the principles ' of engineering, and that a knowledge of some of the important details of track work is essential to the qualifications of a track engineer. As between these two classes of men, both being responsible parties concerned with the subject, the aim is, of course, to select from both views the elements which harmonize with what would seem to be the best practice. I think I understand the difficulty of producing literature en- tirely suitable to all readers who might find interest in a book on track. For the purposes of some it might answer sufficiently well to condense and digest the larger portion of the information into generalized statements, merely hinting now and then at explanations, or leaving such to be acquired by inference. While writings of this character may be entertaining, they usually fail to cover extensive practice, and I regard them as of little value to those readers who wish to make a thorough study of the subject. I have therefore addressed myself to the class of readers whom I thought were most in need of, and who would make the best use of, a thorough- going treatment of details, as well as of general principles. Considering that men in responsible charge of track may differ widely in learning, it is to be expected that a comprehensive treatment of the subject should involve some things beyond the grasp of the average track- man, while, in order that the book may accomplish the highest usefulness, the learned engineer must occasionally find what to him is unnecessary explanation; or he may find the use of some terms, common among track- men and necessary to make the matter understood, which to him are not in keeping with what he might consider the parlance of his profession. It is hoped, however, that what has been deemed necessary for the track- man to know or use will be found intelligible to him ; and what unnecessary explanation the engineer may find certainly cannot mislead him. There is much knowledge that is useful to the trackman which is not commonly sought by engineers, yet which, nevertheless, they ought to have. To deal with a structure so simple as track necessarily calls for many statements of details which may seem trivial to those not in touch with the work; but where ignorance of, or neglect to give proper attention to, such apparently trivial matters is commonly found, and must inevitably result in needless expenditure of large sums of money, it would certainly seem that a refer- ence to the same in a public utterance cannot be out of place and that no mistake can be made in pointing them out here. Some of the simple problems in switch work and curves have been taken up, not because they are not as clearly set forth in other books to be had, but because books on engineering which deal fully with these prob- lems are not, as a rule, to be found with trackmen and are not sought by IV PREFACE them, principally because one not conversant with such books in their entire scope is liable to mistrust his ability to pick out those parts which, he might comprehend. A few problems commonly met with are therefore included, with the rules and formulas applying thereto, for the benefit of those who are acquainted with arithmetic and the use of tables. Beyond this, some general problems which have not yet appeared in field books,, generally, have been worked out for surveyors and engineers. For the- benefit of persons seeking to familiarize themselves with the mathematics of easement curves, some of the problems involved have been demonstrated,, as in a text-book, one object in view being to show that the use of such curves, in all ordinary cases, is not so complicated with mathematics as some may have supposed. A considerable volume of descriptive matter that is used largely in, illustration of practice or of principles discussed has been arranged in the form of supplementary notes. While most of this matter is regarded as essential to a comprehensive treatment of the various subjects to which it relates, and therefore exceptional to the class of matter customarily reserved for use in an appended form, there were two reasons for the arrangement. In the first place, it serves the convenience of the general reader, who may wish to omit extended reference or numerous concrete applications; and, secondly, it gave opportunity to make use of a smaller size of type than seemed appropriate for the purpose of a general treat- ment, thereby effecting some economy in space, the need of which was suggested by the prospect that the amount of matter in view might expand the volume to an inconvenient size. The manuscript for the book, in its present form and scope, went seek- ing a publisher more than six years ago, but for a time it seemed that the chase would end unsuccessfully. By a fortunate circumstance, however, I was finally enabled to embrace the opportunity of publishing the matter- on the piecemeal plan, 'as a series of articles in the Eailway and Engineer- ing Review, with which I became identified as editor. Under this arrange- ment the publication of the matter continued weekly for about three - years. It is due to say that my position in an editorial capacity has afforded exceptional opportunities to enlarge upon my original work, the basis of which was notes and observations taken during years of practical experience as a trackman and an engineer. I have also profited by the- criticism of readers upon the matter published serially, and in the revision: of the same for its final appearance in book form I have increased by about 70 per cent the volume of matter contained in the series of articles. These- successive processes of elaboration have necessarily put much new cloth into the old garment and have greatly expanded its size, without, I trust,, bringing this feature of the work within the meaning of the parable. One constant aim has been to follow practice down to date, and give ref- erence to all new improvements which seemed likely to assume future ~ importance. One particular object in view has been to cover as widely as possible the development of labor-saving machinery. In this line there- has been much improvement within the past few years, and this phase of the subject is destined to be one of increasing interest. In acknowledging sources of information I must concede due credit to- several hundred railway officials and employees and to the many manufac- turers of track supplies, who have kindly responded to inquiries for infor- mation by interview or through correspondence. While I cannot under- PREFACE V take to refer personally to each and every one who has rendered valuable .assistance,, I wish to acknowledge particular indebtedness to Mr. D. M. 'Taylor, of the engineering department of the Wheeling & Lake Erie Rail- Toad, who has kindly favored me with a large amount of data and with a careful criticism of the matter published serially in the Railway and Engi- neering Review. Finally, I shall feel obliged to any reader who will notify me of typo- .graphical or other imperfections discovered, or who will give me the benefit of his criticism upon any statement or matter of opinion^ in which he may find interest, or send me data or records of work to compare with similar data herein contained. Having treated of many questions on which there are conflicting opinions among expert trackmen and engineers, I could not resist the temptation to now and then venture my own opinion on such matters. I am very sure, therefore, that certain opinions herein expressed are contrary to the views of some maintenance-of-way men. To .a very large extent, however, I have accorded questions of this character full discussion, presenting the views of both sides, believing that those interested in a work of this kind would appreciate the enumeration of established opinions fully as well as, and perhaps better than, the conclu- sions which any one person may have drawn from the same. It is needless to here dwell upon the advantages of discussion, for much benefit is some- times derived in the way of suggestion, even though the result may fall short of definite conclusions. In this light it is sometimes profitable to launch an opinion, notwithstanding that opposition to the same can readily be anticipated. A writer on track who would confine his remarks to matters of settled opinion would necessarily have to omit many inter- esting features of practice. W. M. CAMP. Chicago, January, 1908. CONTENTS. CHAPTER I. Track Foundation. 1, Introduction; 2, Meaning of Terms; 3, The Roadbed; 4, Ditches; 5, Cul- verts. CHAPTER II. Track Materials. 6, Rails; 7, Splices; 8, Bolts; 9, Spikes; 10, Ties; 11, Tie Plates; 12, Ballast. CHAPTER III. Track-Laying. 13, General Remarks; 14, Outfit Train; 15, Material Yard and Side-Tracks; 16, Unloading Material; 17, Organization of Forces; 18, Placing Ties; 19, Spacing Ties; 20, Supported or Suspended Joints?; 21, Rail Car; 22, Placing Rails; 23, Square or Broken Joints?; 24, Curving Rails; 25, Allowance for Expansion; 26, Splicing; 27, Spiking; 28, The Track-Laying Crew; 29, Tools for Laying Track; 30, Track-Laying Machines; 31, Highway Crossings. CHAPTER IV. Ballasting. 32, General Remarks; 33, Rail Grade Stakes; 34, Raising New Track; 35. Tamping; 36, Ballast Cars; 37, Lining; 38, Filling in and Dressing; 39, Quan- tity of Ballast Required. CHAPTER V. Curves. 40, General Principles; 41, Simple Curves; 42, Some Ways of Laying Out Curves; 43, To Find the Degree of 'Curve; 44, Action of Car Wheels on Curves; 45, Curve Elevation; 46, Reverse Curves; 47, Compound Curves; 48, Curve Monuments; 49, Rail Braces; 50, Transition Curves; 51, The Cubic Parabola; 52, Tapering Curves; 53, Searles Spiral; 54, The Holbrook Spiral. CHAPTER VI. Switching Arrangements and Appliances. 55, Turnouts; 56, Stub Switches; 57, Laying Stub-Switch Turnouts; 58, Frogs; 59, Guard Rails; 60, Switch Rods; 61, Headshoes; 62, Switch Stands; . 63, Headblocks; 64, Switch Ties; 65, Foot Guards; 66, Switch Lamps; 67, Clear- ance Posts; 68, Point Switches; 69, Laying Point-Switch Turnouts; 70, Chang- ing Stub Switch to Point Switch; 71, Three-Throw Switches; 72, The Lap Switch; 73, The Wharton Switch; 74, Derailing Switches; 75, Side-Tracks; 76, Crossovers; 77, Crossings; 78, Slip Switches; 79, "Y" Tracks; 80, Turntable and Drawbridge Joints; 81, Yard Tracks; 82, Machine Operation of Switches; 83, Interlocking Switches and Signals; 84, Switch Protection. CHAPTER VII. Track Maintenance. 85, Raising and Tamping Low Track; 86, Lowering Track; 87, Lining Old Track; 88, Tie Renewals; 89, Renewing Ballast; 90, Cutting Grass and Weeds in Track; 91, Mowing; 92, Cutting Brush; 93 Ditching; 94, Shimming; 95, Renew- ing and Relaying Rails; 96, Broken and Bent Rails; 97, Regaging; 98, Righting VIII CONTENTS Canted Rails on Curves; 99, Cutting Rails; 100, Expansion in Rails; 101, Stretching Steel; 102, Adjusting Bolts; 103, Creeping Rails; 104, Shoveling Snow; 105, Oil-Coated Ballast; 106, Laying Tie Plates; 107, Bank-Edging. CHAPTER VIII. Double- Tracking. 108, General Considerations; 109, Advantages, Etc.; 110, Comparative Cost of Construction nnd Maintenance; 111, Preparation for Double Track; 112, Construction of Double Track; 113, Danger to Workmen; 114, Sidings for Double Track. CHAPTER IX. Track Tools. 115, General Remarks; 116, Tools Required; 117, Shovels; 118, Picks; 119, Hammers; 120, Wrenches, 121, Claw Bars; 122, Pinch Bars; 123, Tamping Bars; 124, Chisels; 125, Rail Saws; 126, The Gage; 127 Level Boards; 128, Track Jacks; 129, Raising Bars; 130 Rail Tongs; 131, Rail Drills; 132, Rail Benders; 133, Hand Cars; 134, Push Cars; 135, Other Tools; 136, Use and Care of Tools; 137, Tool Houses; 138, Tool Repairs; 139, Section Houses. CHAPTER X. Work Trains. 140, General Remarks; 141, The Train; 142, The Crew; 143, Boarding Ac- commodations; 144, Ditching with Trains; 145, Distributing Ties; 146, Handling Rails; 147, Loading Logs; 148, Handling Ballast and Filling Material; 149, Wrecking; 150, Fighting Snow. CHAPTER XI. Miscellaneous . 151, Fence; 152, Cattle Guards; 153, Bridge Floors; 154, Snow Fence; 155, Snow Sheds; 156, Fire Guards; 157, Bumping Posts; 158, Sign Boards; 159, Signals; 160, Slides; 161, Washouts; 162, Change of Line; 163, Policing; 164, Repairing Telegraph Wires; 165, Disposition of Old Ties; 166, Taking up Track; 167, Purchasing and Handling Ties; 168, Tie Preservation; 169, Metal Ties; 170, Lag Screws vs. Spikes; 171, Effects of Bad Counterbalancing; 172, Longer Rails; 173, Compound Rails; 174, Rerolling Rails; 175, Rail Trimming; 176, Track Elevation and Depression; 177, Track Tanks; 178, Ash Pits; 179, Track in Tunnels; 180, Resurveys; 181, Rail Deflection; 182, Variations from Standard Gage; 183, Automatic Block Signals and Track Circuits; 184, Crossing Gates; 185, High Speed. CHAPTER XII. Organization . 186, General Remarks; 187, The Roadmaster; 188, Section Foremen; 189, Section Labor; 190, Watchmen; 191, Length of Section; 192, Floating Gangs; 193, Discipline; 194, Reports and Correspondence; 195, Track Inspection. Supplementary Notes and Tables. 1, Tile Drainage; 2, Some Details of Steel Working and Departures in Rail Design; 3, Material Yards in Track-Laying; 4, Rules on Care of Lamps, A., T. & S. F. Ry.; 5, Distributing Ties; 6, Tie Preservation in Europe; 7, Tree Plant- ing; 8, Metal Ties in Foreign Countries; 9, Locomotive Counterbalance Experi- ments; 10, Track Elevation and Depression; 11, The Training of Roadmasters; 12, Limit of Capacity of Single Track; Table V, Sines, Cosines, Tangents, Co- tangents, Versed Sines and External Secants; Table XI, Measurements for Stub- Switch Turnouts; Tables XIII and XIV, Measurements for Point-Switch Turn- outs; Table XV, Distances between Points of Frogs in Crossovers; Table XVI, Direct Distances between Frog Points on Ladder Tracks. NOTES ON TRACK, OF THE VNIVERSITY CHAPTER I, or RH^ TRACK FOUNDATION. 1. Introduction. The proper construction of railroad track arid the efficient and economical maintenance of the same involve the science of engineering. There are so many definitions of the term "engineering v that a new one will not be attempted here, for almost any of them apply. One which fits the case very well may be comprehended by saying that to properly construct or maintain track is to know how to "make a dollar go the farthest." Of the three recognized stages having to do with track in service, either construction or maintenance is a field of engineering of no less importance than that of track location. Considering the specializing tendency of the times, which has created such professions as bridge engi- neering, hydraulic engineering, sanitary engineering, and other departments included within the scope of .civil engineering, why should there not be recognized a distinct class of work known as "track engineering?" Track and roadbed represent a much larger investment than do bridges, water- works, or sewers, or more than all combined, and the problems which have to be studied and solved in relation to track are about as difficult as one will find in any line of engineering work. Track engineering begins with the reconnoissance or preliminary sur- veying and must be followed through the location and the construction of the roadbed, the building of the track proper, and continue with the main- tenance and repairs ever afterward ; for in no sense can it be excluded dur- ing the progress of any of these steps. In locating the line for a railroad track, it may often happen that a choice may be had between soils or substra- ta of different kinds, without sacrificing anything in matters pertaining to right of way, grades or curvature ; or the local conditions peculiar to one side of a valley may differ so widely from those of the opposite side, in such respects, for instance, as exposure to wind and drifting snow, slides, falling rocks, surface drainage, springs of water, stream encroachments on the roadbed, the shading of the right of way by steep hills or by forest, as to materially affect the cost of maintenance. Although the relation of such matters to the work and expense of track maintenace is apparent yet it has not always been considered during that part of the work so often regarded as preliminary in a too strict sense. If things are allowed to shift too much for themselves during construction it will usually be found that methods of work will be permitted which will result in inferior service. In these days when so much of industry is dependent upon the activities 2 . TKACK FOUNDATION of corporations, and when labor is becoming more and more divided, men in general will take less and less interest in that which they engage to do, except in what may appear to promise them more or less direct returns in higher compensation or in reputation. Obviously, then, there will be a larger demand for men whose occupation it shall be to maintain a close watch on details, with a view to turn aside all the undirected and mis- directed tendencies which might lead to extravagance, inefficiency, or what- ever in the end might operate depresaingly upon dividends, which consti- tute the ultimate aim of the projectors of railroads. Now, it does not matter by what name we choose to call this occupa- tion whether it be intelligent foremanship, good railroading or engineer- ing there is room for it; but if any system of work or management which can be applied to track supervision in a manner to make track more durable, safer, or less expensive to maintain, be not engineering, then I know of no appropriate term to apply to it. In almost all industrial lines, particu- larly those identified with mechanical or electrical engineering, it is the chief consideration of the science that questions of economy in maintenance or running expenses shall not only share equally with the attention usually given by the engineer to contsruction in general, but that they must be entertained by him particularly and studiously in coexistance with his plans of construction. Already a great deal more study is being devoted to track engineering than was the case when 60-ton locomotives and 20-ton freight cars were typical of rolling stock, and the tendencies indicate a still larger application of engineering principles to this line of work. It is not difficult to explain the situation in the past. Track is so extended over distance, when compared with other works or structures; the roadbed, the ballast and the materials of which the track is constructed are subject to such inequalities and irregularities; the track structure is so simple and deteriorates by such insensible degrees; and the wide-spread but mistaken idea that "main strength and awkardness" are as efficient in its service as intelligence and skill, has so prevailed, that, in the very nature of things, the officials not directly responsible for the condition of the track were slow to grasp the idea that track should be studied as thoroughly as other engineering structures. The simplicity of the track structure is the de- ceptive element in questions relating to maintenance economy, for ideas con- cerning the stability of track are too fequently confined merely to the question of approved qualities of rails, ties and ballast. The fact that the track structure lies upon the surface, exposed to the extreme action of the natural elements, is a very important consideration in track engineering. One of the most expedient resources available for moving people out of a rut is to make them feel the disadvantages of their position from a finan- cial point of view.. Opportunities for applying this principle to railroad track are easy to find. For the sake of illustration, let us for a moment contrast the track with some other engineering structure in use on rail- roads ga y an iron or steel bridge. Now the average bridge is considered a costly structure, and much care is taken with every detail which goes into its make-up. The foundation upon which it rests is usually built to stand ; all materials going into it are of the most substantial quality; all the pieces goinc 1 into the superstructure are not only carefully made and inspected but are carefully handled when being put together; connections or joints between pieces are made stronger than the pieces themselves; every piece in the whole system is carefully adjusted to its place, so as to bear its proper stress, and that before any load is allowed to come upon the bridge. The structure is supposed to be kept well painted; it is watched and inspected frequently ; and should there be found deflections much exceeding those cal- INTRODUCTION 3 culated upon, or any behavior tending to show the least weakness, the whole thing is counted a failure. * Such is engineering, and such is what makes weak' railroad bridges scarce and bridge accidents of seldom occurrence. All the care exercised costs, but everybody knows that it is money saved and that it is good economy. As for the track, who does not know that ten miles of average track costs more than the average bridge of several hun- dred feet length, foundation and all? Yet who does not know that when put together the work has too frequently been done with a rush and that reckless work due to this cause has been too frequently overlooked? In liow many instances has one not seen the work improperly finished, as, for instance, when ditching and such necessary work was Mt~to be com- pleted at a time when its cost must necessarily amount to much more than what it would have been in the first place ? How many have been the cases where costly materials are worn out or rendered useless 'through lack of attention, or through ignorant supervision, long before they should be? Now all this costs money and it is known to be false economy, yet it has not been so generally conceded as have like mistakes in some other lines of engineering. What then is needed? I maintain that the same strict and intelligent engineering is needed that is usually applied to some other railway affairs. It is popularly supposed by some trackmen that the term "engineering" relates to matters in which they are not concerned ; while on the other hand, to some railroad surveyor or draftsman the employment of the word in connection with trackmen's work is to disparage his occupation and its relative importance to the occupation of a trackman. Where such is the presumption both, parties have a mistaken conception of the word engineer- ing. I consider that there are many roadmasters and section foremen who have more to do with track engineering than some men commonly known as civil engineers, yet whose experience has been nothing more than survey- ing or drafting, no matter how extended their experience ftthin such limits may have been. Eeally, surveying and mathematical calculations cut but a small figure in track maintenance. It is true that in some respects track location can be fairly well learned from books, drawings and office work, but how to build and maintain track to best meet diverse conditions cannot be learned between covers or in an office. The experience necessary to teach such knowledge must be had by actual contact with the work. Ac- cording to some men's ideas track engineering is largely a matter of sur- veying and the ability to select good materials, but in the following pages it is attempted to show that it also requires intelligent manipulation and an adjustment of parts involving no mean order of skilled labor. There can be no question but that some prestige is lost to the engineer- ing profession from the fact that so many men who have a general knowl- edge of engineering principles attempt to make their applications too specific, without having acquired that view of things which comes only by patient and earnest devotion to the partiuclar line of duty, with some responsibility therein ; and so it is with track engineering. There are men who have never so much as sweat a drop in any kind of service calculated to impart a knowledge of track work, or lost a moments sleep caring for the safety of track, who, nevertheless, are ever eager to propose what they think to be some track improvement; and as a rule their ideas on improve- ments amount to about as much as their services have. And so, many prac- tical railway men who, in reality, do most of the thinking and perform most of the work that is worthy to be called engineering, hold in a sort of contempt the very term which best describes the results of their own efforts. Between these two classes there has existed to no small extent an attitude 4 TRACK FOUNDATION which has tended to discourage practical men in what they ought to seek, after more than they do; and as a partial result of this there is, at least,, a misconception of terms. Almost everywhere one hears it said: "Theory is not practice,"' "Theory will not work in practice," etc. The term theory is too often used with that looseness of expression synonymous with inference, guesswork,, speculation, etc. ; whereas in its proper sense it applies only to those ideas- which have not been known to fail under any reasonable test, and for which there are, therefore, good grounds for putting them to further test wherever they can reasonably be applied. Again, what would be theory in a sense pure and simple might not be theory as applicable to every case where one might wish to make it hold ; example : One may find tables giving the outside rail on curves as much as 14 ins. elevation, for a certain degree of curve and 5 a certain speed not uncommonly made. Now, as for some specially built vehicle running upon some specially built track, this application may be theory; but it is not theory which has to do with the conditions which obtain in railroading; when applied to such it is not theory but nonsense, because no account is taken of circumstances which might be known. Some man's ideas about the application of some mathematical formula or scien- tific principle does not necessarily constitute a theory. The fact of the matter is that all competent men, commonly called "practical" men, use more theory than they may think they do; it guides- them in much or all of their work, although they may never have thought to- express it in so many words, perhaps. What men need to have in order to ac- complish the highest results in any line of work is a clear understanding of" the principles they are using. And then, too, unless men understand such principles, or the theory of their own work, they are unable to apply them- selves to such changing conditions as may arisa in any business experience of a few years. Practice without some knowledge of the principles involved is like working blindfolded, while, on the other hand, a knowledge of princi- ples without some practical ideas of applying them is useless. And then r in order to get a proper conception of the principles underlying a case it is essential that right premises be taken. Theory cannot be formed upon faulty observation, neither is it derived by defective reasoning, nor does it necessarily follow from hit or miss speculations. Correct theory (mean- ing theory correctly applied) and good practice are always in strict accord ; and where they apparently are not, an investigation will always lead to some interesting disclosure; generally showing that there was either a misconception of principles or an attempt to make a wrong application of them. Or, in every case where an idea is said to work well in practice it is needless to say that it conforms to principle; and if seemingly not, then the legitimate principles involved are not well understood. The only essential difference between correct theory and good practice, in one way of expressing it, is that practice so called, must employ a knowledge of de- tails, while theory, in distinction therefrom, may stand entirely upon a knowledge of principles which, of course, must be learned from details, although not necessarily from the details of the thing or practice to which the application is made. Hence it is that a knowledge of principles, as applying to some particular practice, sometimes precedes a knowledge of its details and sometimes vice versa. Good practice is theory rightly applied ; or theory may be called the explanation of, or the reason for, the practice. They both represent truth, when rightly comprehended; and how to com- prehend the two in their right relations and to carry them out in appli- cation is engineering. 2. Meaning of Terms. A railroad is made up essentially of: MEANING OF TERMS % 5 'three parts: the foundation, the ballast and the track. The foundation is the earth support, the upper surface of which is usually brought to an -established line known as sub-grade. In the case of an embankment or a fill the foundation is the earthwork: in a cut it is the lower limit of the -earthwork. Ballast may be considered to be some kind of material placed upon the foundation to put track in surface and afford drainage, and per- haps also to hold the track in alignment. In the case of dirt or "mud" bal- last the foundation might be considered as extending to the track, the ballast being in that case only that portion of the material which lies between the ties ; and the drainage, so far as accomplished, taking place on the sur- face. The track is the rails, with their fastenings, and the -ties. While, .as regards only the path of the wheels, the rails alone might be considered .as constituting the track, the fact that the tie serves as a means of holding the rails to proper .gage, as well as serving for a support, and that the whole is a separate, distinct structure, would make it seem that the tie ought to be considered an integral part of the track. For purposes of general description it is also more convenient not to subdivide the portions of the road further. As names for these parts several terms have grown out of practice, .some of which do not express the real meaning, and as a result there is more or less interchange in their application, and, consequent confusion. Tor instance, the term "grade" is in common usage among contractors and track-layers to denote the upper surface of the foundation, when, as all know, the same term is the universally accepted expression for rate of -ascent or descent with respect to level. The term "roadbed" is sometimes used to denote the foundation, sometimes the ballast, and sometimes the surface of contact of the ballast with the ties; that is, in the last case it is used in the same sense as when the bottom of a river is spoken of as its "bed." The term "track" is sometimes used to denote both the super- structure and the ballast. In imitation of the English, some members of the engineering profession choose to call the combined track, ballast and foundation "the permanent way," which, by the way, is something of a misnomer. There is nothing about American railroad track that is par- ticularly permanent, except, perhaps, the line establishing the location; but to maintain the track to this requires constant or continued labor; while the materials in track and ballast (and foundation, too, sometimes) require, in course of time, more or less frequently, either changing or replenishing. Even stone arch bridges, on some of our roads, have failed after less than fifty years of service. In the present connection it is instructive to enquire into the signifi- cance of the term "permanent way" as it is used in England. In that country the application of the term arises from the manner in which track is constructed. After the roadbed is completed a stretch of "temporary" track is laid with old materials, and after the ballast is hauled in this track is lifted approximately to the final grade and ballasted. The ballast is then leveled to the bottoms of the ties, the new rails and ties are strung out along each side of the track, the temporary track torn up, the ballast dressed off to a smooth surface, the new or "permanent" track is laid and thoroughly tamped and the ballast filled in and dressed off to standard section. Thus the road is extended by first laying temporary track, a section at a time, and repeating the process just descibed, using the old materials over and over. By such practice the lifting of new track through any considerable hight is avoided and trains are not permitted to run upon the same until it is put into final condition. The term "permanent way" is thus used in a comparative sense, to distinguish the road, as completed for traffic. () TRACK FOUNDATION from the temporary track laid down for the purpose of forwarding materials and placing the ballast. As applied to American roads, however, the term loses its English significance, for we do not build track in the manner described. As the term is also without application in a literal sense it is both un-American and in bad taste to speak of American railroad track as "permanent way." In view of possible misunderstanding, some of the terms used in this book are here defined as follows: Track foundation is called the roadbed, and its upper surface, sub-grade. The material between the ties, and between the ties and the sub-grade line constitutes the ballast. The rails and ties, when united, are called track. The combination of these fairee parts, the roadbed, the ballast and the track, is known as a railroad or railway; or, simply, the road. The term "rail," or "the rail," is some- times used to denote a piece of rail of standard length, say 30 ft., and some- times to denote all such pieces on one side of the track taken collectively, the same as though the rails were jointless. The term "steel" and "the steel" have the same significance. A "piece of rail" refers to a piece shorter than standard length, or shorter than 30 ft. Maintenance of way is commonly understood to mean or include the maintenance of all the fixed property of. a railroad, such as track, bridges, buildings and water supply. Where nothing is said to the contrary, the gage of all track herein con- sidered is understood to be standard, or 4 ft. 8-J ins. The word "ton" without qualification means the American ton of 2000 Ifos. Wherever locomotive weights are referred to the weight of the tender is not included. The term "ends of the ties" is commonly used to refer to that part of the ties which extends outside the rails. 3. The Roadbed. It is not the purpose to deal here with those problems of locating the roadbed which belong properly to surveying, treatment of which can be found in the many books on field engineer- ing. Neither can there.be included the discussion of a subject of such scope as the more intricate engineering problem known as the economic theory of roadbed location, where the topography of the country gives rise to such questions as compensation between distance, grades and curvature. The object is to take up only those features of the roadbed which may in some way affect the condition of the track built upon it, either during the construction of the track or after it is put to its use. Unlike almost all other foundations prepared to sustain a great weight, the roadbed of railroad track is largely of an unstable nature. Instead of seeking for a solid substratum, as is done when laying the walls of a building or when constructing a pier or abutment for a bridge, it is found expedient to take the surface of the ground pretty much as it is and, in excavation, go only so far as a predetermined grade line has been established, without reference to the nature or depth of the yielding soil ; while, when this same established grade line lies above the ground there is added to the top surface such material of the same yielding nature as lies most conveniently for movement: Such, at its best, is the roadbed. This, with a comparatively thin layer of ballast material of such quality as the available expenditure will admit of (and, often, no better than common soil itself), must not only bear up a ponderous load, but bear it intermittently, and under conditions which increase the effects to a degree not possible with the same burden imposed as a sta- tionary object. How, then, to construct, with such material, a formation extending over long distances, in a manner to endure not only the weight above but also the natural forces expending themselves around it, is the- work of building a roadbed. THE ROADBED 7 Roadbed Cross Sections. Almost every railway lias a standard cross section for roadbed, at least on paper, and among different o-oads these so-called "standards" vary considerably in form and in dimensions. So far as fills are concerned the width of the roadbed at sub-grade should afford at least a sufficient base for the ballast. Taking the depth of the ballast not to exceed 8 ins. below the bottoms of the ties, and allowing for a narrow shoulder of ballast against the ends of the ties, the ballast will overspread a strip of roadway about 14 ft. wide. The least permissible width for single-track roadbed at sub-grade is then about 14 ft., after shrinkage; and since the top width of an embankment should increase with hight it might be well to settle upon 16 ft., after shrinkage, as the^ieast permissible width of roadbed at sub-grade on embankments exceeding, say, 10 ft. in hight. A width of 14 ft. allows for no shoulder on the roadbed outside the ballast; and although good track can be maintained upon roadbed without such shoulders, the conditions are by no means ideal, and the highest efficiency of support cannot be looked for on these dimensoins. As for economy of maintenance, however, that is a matter having to do with the amount of surface repairs, which, of course, depend mostly upon the volume of the traffic. The situation with many railroads in this country is such that it would hardly be found economical to adopt such dimensions for roadbed as might, from the standpoint of efficiency alone, measure closely up to the ideal. For roads having but few trains daily it is a ques- tion whether the minimum widths of roadbed here given may be exceeded with any economy. On roads carrying heavy traffic 18 ft. is usually con- sidered the least available width of roadbed for fills, and a maintained width of 20 ft. is sometimes found to be standard. On the Southern .By. the standard permanent width of single-track roadbed on embankments, at sub-grade, is 14 ft. ; on the Great Northern Ry. it is 14 ft. on tangents and 16 ft. on curves; on the Southern Pacific (existing high embankments) and Atchison, Topeka & Santa Fe (branch lines) roads, it is .15 ft.; on the Atchison, Topeka & Santa Fe (main line), Baltimore & Ohio, New York Central & Hudson River, Northern Pacific, Southern Pacific (for construction of new lines) and Louisville & Nashville roads it is 16 ft. ; on the Chicago, Burlington & Quincy Ry. it is 17 ft. ; on the Burling- ton, Cedar Rapids & Northern and Cincinnati, New Orleans & Texas Pacific roads it is 18 ft.; on the Philadelphia & Reading Ry. it is 18 \ ft. ; on the Pennsylvania R. R. it is 19 ft. 2 ins. ; on the Illinois Central and Union Pacific roads it is 20 ft. ; w r hile on some cheaply built roads it is only 12 ft. The width of roadbed for double track exceeds the width for single track by the distance between track centers. Sirtce in excavations the roadbed must be made wide enough to afford room for proper drainage parallel with the track, the width of roadbed in cuts is considered in connection with the subject of ditches. The advantage of increasing the width of roadbed with depth of fill is that after the embankment has settled it can be built up to grade again by dumping material on top without having to replenish the slopes in order to get the necessary width of base. This principle of construction is adhered to on a number of roads. On the Chicago, Milwaukee & St. Paul Ry. the standard maintained width of roadbed, at sub-grade, on embankments is 18 ft., but on branch lines, a narrower roadbed has sometimes been constructed. In construction of new lines, however, it is the practice of this company to increase the width of roadbed at sub-grade 2 ft. for each 10 ft. in hight of the embankment. On the Michigan Central R. R. the roadbed is made 1 ft. wider for each 5 ft. increase in hight of embankment. The standard width of embankments on the Kansas City, Pittsburg & Gulf R. R. is 16 TRACK FOUNDATION ft. for fills up to 15 ft. in hight and 18 ft. for embankments higher than 15 ft. It is poor economy to make a fill so narrow that it must be widened while the work of ballasting is in progress, in order to keep the ends of the ties from overhanging the ballast, which has slidden down the bank. It is poor economy for two reasons: first, ballast, be it gravel or other material, is rather too expensive to use for widening embankments, if it must be hauled some distance; and secondly, gravel deposited upon a hard slope will slide off. The slope of an embankment cannot therefore be maintained in a manner to sustain the weight above, by depositing loose gravel upon a firmer substratum, without using a quantity of it which is entirely out of proportion to the quantity of filling required. Track- men can recall, the familiar spectacle of having seen half the ballast, which had been hauled for the purpose of ballasting and surfacing the track, lying either at the foot of the embankment slope or along its sides, to be crowded farther down by every workman stepping out of the way of a pass- ing train, thus weakening the support which the ends of the ties ought to have. Insufficient support for the retention of the ballast is the most frequent cause of center-bound track. The first cost of extra material required to make the fill of proper width at sub-grade, which such materials as will lie stably along the slope, is small in comparison with the cost of afterwards wasting a large amount of more costly material which is not so well suited for the purpose of an embankment. The roadbed should be brought to full width and completed before track-laying begins. Cuts and fills are often left to be widened after the track is laid, but it is nearly always one of those mistakes which cost. Engineers usually aim to have the cuttings make the fills, without hauling farther than can be done to advantage with teams. But in rare instances, where fills cannot be finished out from borrow pits, and material must be hauled a long distance, it is advisable sometimes to leave the widening of a cut, now and then, to be done with the work train rather than to waste it at the first. In such cases, however, it should be done as soon as a work train can be put on after the track is laid. Where some exceptional case of this kind is calculated upon and the practice does not become general for the whole line, there may be found a- saving; but in the main, where the road is rushed through with earthwork only partly completed there usually follows much waste of ballast, which is lost by sliding off the narrow shoulder, and the track is flooded in cuts for want of proper ditches. In at- tempting to widen an old embankment having hard slopes, stones and lumps dumped from the top will roll to the bottom and out of the fill ; and, when lying upon a harder surface, the quantity of material of any kind which must be used is disproportionate to the quantity which would be needed, during construction, just as in the case with gravel, above explained, because the weight from above will push the bottom of the slope out and the earth will not then stand at the same slope as when the fill is made with loose material all at one time. Work trains being always more or less hindered in their work by regular trains, it is well to do all that can be done on the roadbed before the track is laid, for it is nearly always the cheapest, the most economical, and by far the best plan. Where an embankment is deficient in width it is hardly worth while to raise the track to grade in ballasting, because track boosted upon a narrow heap of material obtained by robbing the side slopes will quickly settle, from want of lateral support to retain the ballast. It is just as well, and indeed better, in such cases, to permit the track to remain at such hight as is consistent with the width of the supporting base, even though there THE ROADBED jnay be local depressions below the established grade line. In widening out an embankment on which ballast has already been deposited, the material added to extend the shoulders, unless it be the same material as the ballast itself, should never be built up higher than the bottom of the .ballast, as to do so will obstruct the drainage. For drainage purposes the roadbed ought to be somewhat higher in the middle than at the sides. Unless such is the case the water which ..settles through the ballast will find its way into the roadbed, soften the material and cause it to settle under the pressure of the traffic and heave during freezing weather. On standard cross sections it is customary to show the roadbed crowned 3 to 6 ins. in the middle, but the scheme^ is seldom carried into practice. Such negligence is one of the worst mistakes that is made in railroad construction. In cuttings it is an easy matter to crown the roadbed at the center, without extra expense, and on the natural earth the material is usually firm enough to preserve the sloped surfaces permanently. On embankments filled from trestles and in filling up ravines by dumping -over the slopes it is impracticable to do this, but in filling an embankment by working from the bottom with teams the material can be so placed that the top of the roadbed will shed water fairly well. The standard roadbed of the Missouri, Kansas & Texas Ry., which is 16 ft. wide on embankments and 18 ft. wide in cuts, is flat for a distance of 4 ft. each side of the ^center line that is, over a width corresponding to the length of the ties -but from a line directly under the ends of the ties there is a slope ,of ^6 to 1 out over the shoulder. In curves the surface of the roadbed may be made flat and inclined to the same slant as that to which the track is to be elevated. This arrangement allows of an equal depth of ballast under both sides of the track and, therefore, an equal settlement for both sides, because all kinds of ballast in new track will settle some. It also effects an economy in the use of ballast. Eoadbed which is not made to slope in this manner on curves should be widened on the outer side of the curve to provide the extra width of shoulder required for the higher and longer -slope of the ballast on that side. Such extra width is especially needed where the roadbed is narrow or of minimum allowable width. Grading. The manner in which a fill is made has much to do with the efficiency with which it will support the track. The most solidly compacted work is had" when horses are driven continually over the fill during the progress of its construction, as when hauling in carts, wagons or scrapers. Earthwork constructed by such means will usually settle but very little afterward. Shrinkage is greatest on embankments formed by dumpings from wheelbarrows, by machine graders or by casting the mate- rial by hand from side ditches. The method of depositing the material in roadbed construction, affecting so largely as it does the question of shrink- age, is a matter of no small importance; for when a fill settles, the track must be put again to surface with ballast. By providing against future settlement as far as possible there is curtailed an item of considerable expense. In the present connection, therefore, it may not be amiss to consider briefly some of the methods and means employed in roadbed con- struction. Wheelbarrows and plank runways are used where the material to be moved must be taken across a ravine, or across a track, as at a side-hill y sliding, and in attempting to remedy the matter by dumping more material upon the slidden portions of the embankment something like THE ROADBED 1$ $10,000 was expended without accomplishing the purpose sought. Investi- gation disclosed that the embankment had been formed of light clayey loam hauled from cuttings and deposited upon a substratum of heavy blue clay scraped up from borrow pits. The seat of all the trouble was found by excavating into the bank, when it was discovered that a well denned plane of cleavage existed between the two materials, coinciding with the inclined runway of the scrapers used during construction. The sliding of the embankment had taken place upon this inclined plane. The- t rouble was permanently cured by heroic treatment, at a total cost of $1200, in the following manner : During dry weather longitudinal trencher were excavated on the side slopes, near the foot of the embanldnent, on either side, and between these two trenches cross trenches 2 ft. wide and 10 ft. apart were cut through the embankment and carried down below the so-called plane of cleavage. Meanwhile the track found support upon ait improvised trestle formed by placing long stringers upon two-pile bents which had been driven when the sliding became serious, prior to the exca- vation work, resort to this means of support being taken to carry the trains. The trenches were then filled with pile heads and old ties and covered over with the clay excavated from them. This mass was then fired,, more clay being added as the burning material in the trenches settled down. The embankment smoldered away like a charcoal pit for six weeks and the bank was burned into a solid mass of brick, never to slide again. Fol- lowing the result of this experiment the same method of treatment has been applied to other embankments on this road containing clay, with uni- formly successful results. Where filling is to be done on a steep, hard slope, as on side hill, it is well to break up the cleavage plane by cutting lines of steps in the slope, to bind the new material. In most cases this may be done sufficiently well by plowing deep furrows several feet apart. To prevent the slopes of filling material from sliding it is sometimes the practice to dig a trench just inside each toe line of the embankment,, but with material of approved quality such precaution is not necessary. Large stones should not be placed in shallow fills, and when such are found above the surface, on location, within 2 ft. below sub-grade, they should be broken up, rolled into pits or rolled outside the slope stakes. Stumps should be cut off at least 2 ft. below sub-grade, but wood which rots quickly should not be permitted to remain in a fill at all, except, per- haps, in the case of an embankment formed upon a steep hillside where the trunks of all large trees which stand within the slope stakes should remain standing, to assist in retaining the earth and prevent sliding. Just before the embankment is completed or at that time the trees may be cut off at proper distance under sub-grade, or even with the slope. It is usually required that large trees shall be cut so that the tops of the stumps shall not be more than 3 ft. above the ground, and that where embankments are less than 3 ft. in hight all trees and stumps shall be cut close to the ground. The surface of ground to be excavated and where embankments less than 2 ft. in hight are to be bniilt should be grubbed free from stumps, roots and other perishable material, and all brush should be cleared away. In some cases where the materials vary considerably it is advisable to reserve the best of them for finishing and dressing the surface. In a wooded country, where grading is done by contractors, especially when sublet to numerous subcon- tractors, it is well to keep close watch to see that logs are not rolled in and covered up. One is justified in being suspicious of all filling done at night or not during regular working hours. Aside from rotting and weakening the fill years afterward, or burning, in case the ends get uncovered, a fill containing many logs, especially where several are rolled together in a heap v 14 TRACK FOUNDATION or are near the top in any shape, will' be springy or humpy and a positive hindrance to good surface for a long time. The top surface of roadbed should be graded off smoothly, filling up all pockets or depressions which would otherwise remain to collect and hold the water which sinks through the ballast, where it can freeze in in winter time and heave the , track. For this reason ruts from wagon wheels driven over the roadbed should be filled and the material compacted before the track is laid or the ballast deposited. The practice of running material trains over newly laid track before it is surfaced or ballasted also requires that the roadbed surface be made smooth, to afford an even embedment of the ties and thus avoid kinking the rails. As far as possible, embankments or other made ground should be com- pleted early, so as to have time to settle before the track is laid. Em- baiiJonents seldom settle evenly, and track laid thereon before 'settlement does take place is sure to require considerable labor to keep it in fair surface soon after the trains begin running. Where a short fill comes between two cuts or next to a bridge it should be put high enough to allow for settle- ment. If not, the fill will settle below the established grade, while the roadbed in the cuts or the track on the bridge will not, and there will then result an ugly sag. The slope of railway embankments varies from 1 to 1 for rock fills to H to 1 for ordinary earth, and easier slopes for soft material like clayey soil, when such must be used. By building up a rough dry Avail on the exterior of a rock fill it may be made to stand at a slope somewhat steeper than 1 to 1. Ordinary earth will stand for a time at a slope steeper than l. (horizontal) to 1 (vertical), but under the action of the rains, the winds and frost it will gradually wear dpwn to about 1J to 1. In excavations, solid rock which will not disintegrate by exposure will stand to a vertical face. " Firm dry earth well protected from water seepage is sometimes sloped 1 to 1 in cuttings, but under less fav- orable conditions it will not always stand at even 1J to 1. Under ordinary conditions, however, 1^ to 1 is considered safe. Where excavation is made through rock overlaid with earth, the earth slope should be cut back to leave a berm of 3 to 4 ft. from the brink of the rock slope, to retain loose material which slides or rolls down in moderate quantities. The slope of earth in such places should be made easier, if anything, than elsewhere, because rock cuts are usually made so narrow that only a relatively small amount of mate- rial sliding into the same is liable to fill the ditch and obstruct the rail. Sub-Drainage. A question which has received but little attention from American railway engineers as yet is the sub-drainage of roadbed. Although the value of sub-drainage in wet cuts is recognized, and in com- paratively few cases something has been done to put the principle into practice, but little or nothing has been done to carry the water from embankments by under drains. In selecting material for ballast the drain- age feature is considered one of the most important properties, but it should be understood that drainage as applying to ballast refers to the drainage -of water from the ties to the roadbed. Through stone ballast, for example^ water sinks to the roadbed as through a sieve, and in gravel ballast the same thing occurs after the material becomes thoroughly soaked. The sanrl. contained in gravel has, of course, some capacity for holding the water back. Such being the case the roadbed undei most of the track which is considered well ballasted must receive a good deal of water. So far as ballast is concerned it reaches its limit of settlement within a comparatively short time, and those who will investigate matters closely will find that rough surface in old track is caused largely by settlement of the roadbed or settle-- THE ROADBED 15 rnent bilow sub-grade. One very responsible cause for this condition is the seepage of water through the ballast and into the roadbed. .While much of this might be drained off on the top surface of the roadbed, if the same was properly crowned, it is known, nevertheless, that only a comparatively small amount of roadbed construction is brought to such a top surface and compacted sufficiently to hold its slope until the ballast is placed upon it; and in many cases, as already shown, it is impracticable to do this. As a rule, then, a good deal of water must find its way into the interior of the roadbed to soften the material and keep it continually settling, and, in cold weather, to freeze and heave the track. Moreover, as most double- tracking is done by building a second track beside the old ohe~it~is imprac- ticable to crown the roadbed midway between the tracks and slope it to either side, since the roadbed at sub grade under the old track, if prop- erly formed, is highest underneath the center of the track and, in any case, it is not accessible. Such being the case one has good reason to think that most of the water which finds its way through the ballast on double track percolates through the roadbed material to considerable depth. The value of tile drainage being well understood, there would seem to be no difficulty in keeping the interior of the roadbed reasonably dry by resorting to the usual methods of sub-drainage. It would certainly be worth the cost of thorough trial, on any railroad where the annual rainfall is considerable, to see if a longitudinal tile drain laid a few inches under sub-grade, with cross drains leading to the surface at frequent intervals, would not intercept the larger portion of the water which ordinarily sinks through the roadbed. On single track such a drain might be laid under the center of the track, preferably in sections draining into the cross drains at a considerable fall rather than in a continuous line laid to the grade of the track, except, perhaps on the steepest grades. On double track, where the roadbed is constructed before either track is laid, and properly crowned in the middle, a line of tile might be laid under the center of each track, but otherwise, as in the case where the two sides of the embankment or cut were formed at different times, only one drain would likely be used, and that could be placed to best advantage midway between the tracks, or about on the dividing line between the old and new embankments. On tho double-track lines of the Baltimore & Ohio R. R., where stone ballast is used, an 8-in. tile drain is laid upon the roadbed midway between the tracks -and the ballast is filled in level with the tops of the ties. The roadbed is crowned 6 ins. in the middle and the depth of ballast at this point is 12 ins. below the bottoms of the ties. Sodding and Seeding. Still another line in which the maintenance -of earthwork is open to improvement is the protection of slopes against washing or sliding, by the growth of vegetation. Barren embankment slopes are continually eroded by rains, and ditches at the bottom of bare slopes in cuts become obstructed by sediment washed down by the rains or which rolls or slides down when loosened by the thawing of the ground. On English railways the sodding of slopes is a feature of general practice, but in this country it has been considered too expensive to have succeeded to anything like extensive trial. Here and there the slopes of a cut will be found sodded, but in general practice the natural growth of weeds or grass, without any attempt at encouragement or cultivation, is all that can be found on either cuts or fills. A healthy growth of grass on slopes requires nourishment by a coating of soil. In cuts this may sometimes be obtained by stripping the top surface some distance back from the cut. The best -opportunity to obtain the material on the right of way, however, is before "the cut is excavated. Tlu i top Foil is scraped back beyond the slope stakes 16 TRACK FOUNDATION into heaps and after the excavation has been completed it can then be spread!' over the slopes that are to be sodded or seeded. On embankments the strip- pings from gravel pits, material cleaned from ditches, the bedding from stock cars, and other fertile material which must be hauled off for disposal may be utilized to encourage the growth of vegetation. Sweepings from streets are also good material for this purpose, as they contain a large per- centage of fertilizing matter and a considerable mixture of seeds of various kinds. Before sodding or seeding is begun the slopes should be dressed off rea- sonably smooth and angular shoulders and intersections at the top and toe of slopes should be rounded off to a natural contour. Sods about five- years old are the most vigorous for transplanting, and those from high, well drained ground are, from previous condition of growth, better able- to thrive on dry slopes than sods from swampy or wet localities. As a means of assisting the sod in getting started some recommend sowing a mixture of timothy seed and oats over it the first year. These will quickly spring up and form strong roots to help hold the sod in place. To strengthen the growth later on, Kentucky blue grass, white clover, perennial. rye, red fesco and red top, in the proportions of 8, 4, 9, 3 and 8, respectively,. are considered a good mixture for supplemental seeding. In the South the slopes of embankments are frequently set with tufts of Bermuda grass in rows 1J to 2 ft. apart, as referred to in 148. This grass will thrive- in sand, and in a short time it forms a thick sod entirely covering the ground. Its characteristics are described in 12, in connection with sand ballast. Where sod is placed on steep slopes it is customary to drive stakes,,. in rows, staggered, to hold it in place until the roots take firm hold. The stakes are usually driven flush with the surface of the sod and permitted to remain. Seeding is, of course, cheaper than sodding, but some time is required for the growth to form a sod. The variety of seed best suited to the 'climate and soil is perhaps best ascertained by observation of the- natural growth or of the grasses grown under cultivation in the locality.. Willows and scrub brush indigenous to the locality are also planted OB slopes to check the tendency to slide or wash away. An advantage in a growth of willows is that their great vitality permits close trimming, form- ing in time heavy stumps and strong roots to permeate the ground and hold it in place, without the presence of an excessive or troublesome growth above the surface. Borrow Pits. A matter which ought to receive more attention than it sometimes does with fills made from borrow pits, is the nearness of the pit to the foot of slope. At the foot of slope of shallow embankment* , say up to 3 ft. in hight, there should be a berm at least 4 ft. wide ; and for higher embankments the berm should be wider. The removal of earth in nearness to the foot of slope increases the hight of the embankment by the depth of the pit excavated, and if the pit is too near it of course weakens the embank- ment. It is also to be considered that with a narrow berm ties and other materials thrown off the cars will slide or roll out of reach; and besides, if the embankment must ever be widened, a "pit at the foot of slope must first be filled before an addition can be made. In improving the grades of a line it is not unusual to raise the track as" high as 4 ft., increasing the hight of the embankment that much, which means a widening at the foot of slope of about 6 ft. It would seem like good policy, therefore, to be mindful of a good factor of safety in establishing the width of berm, mak- ing it at least 8 or 10 ft., wherever practicable, and always leaving room for a double track on one side. The standard cross section of the Union? Pacific R. R. provides for a berm 6 ft. wide on one side and 18 ft. wide- THE ROADBED 17 on the other side, as the probable base for a widened embankment for a second track. The standard plans of the Kansas Gity, Pittsburg & Gulf E. E. require berms 6 ft. wide for banks 15 ft. high or less and 12 ft. wide for banks higher than 15 ft. The standard berm of the New York Cen- tral & Hudson Elver E. E. is not less than 6 ft. wide in any case, nor less than double the depth of the pit or ditch, always leaving room for double track on one side. On the above-mentioned considerations it is just as important that the berm should be guarded against depletion as that it should be established at proper width in the beginning which is to say that the berm should not be cut away to "build up" the bank.. Borrow pits, or ditches from which material is taken for shaping up banks, are frequently left by the construction forces in an unsightly con- dition, being excavated on irregular lines and to irregular depths, and with- out sloping the sides. Properly, the sides of borrow pits should be trimmed up parallel with the alignment, and the side next the track should be given a natural slope, according to the character of the material, never steeper than 1 to 1 for earth, and usually 1-| to 1 or the same as the slope of the embankment. The drainage of borrow pits close to embankments should also be looked after, and the proper time to do this is, of course, when the pit is being excavated. In most instances the pit may be so located that the excavation of the same will provide an outlet, or the excavation for an outlet to the pit may be made to supply part of the borrowed material. Wa- ter standing in a borrow pit at a higher elevation than the surface of the right of way on the opposite side of the embankment will naturally seep strongest in that direction, thus tending to soften the earthwork founda- tion. Roadbed over Marsh Land. Aside from the nature of the fill itself the ground upon which it is made is sometimes so unstable that proper sup- port for the track cannot be easily obtained. Such is frequently the case in swampy or boggy land and on quicksand. On an easily yielding surface a shallow fill will roll up ahead of a train and give way underneath it, with- out remedy. Under such circumstances it becomes a difficult matter to maintain the track in fair surface, and the rails will creep badly. There are several measures which may be taken to overcome difficulties or improve the conditions in a situation of this kind. In the first place, before attempting to fill across a swamp or bog the region should be drained as thoroughly as may be feasible, and in order to accomplish this it is some- times necessary to undertake the drainage of a large area, by extensive ex- cavation distant from the right of way. The hope of accomplishing results on this line of operations lies, of course, in the amount of fall obtainable. It is then important to look carefully to the drainage n.ear the track, usually by cutting ditches of good depth a's near the embankment as the conditions of stability will allow. The Galway .& Clifden Ey., in Ireland, runs through long stretches of bog land and the matter of drainage and types of roadbed construction have been closely studied. On this road the standard arrange- ment for drainage consists of two longitudinal ditches 22 ft. apart on each side of the track. The inner ditch on each side is 4 ft. wide at the top and 3 ft. deep, and is cut at a distance of 6 ft. from the toe of the em- bankment. The outer ditch is 6 ft. wide at the top and 5 ft. deep. The -side slopes of these ditches is 1 in 3 and the two are connected by cross drains every 100 ft. In filling over boggy land it will usually pay better to haul the filling material some distance than to use the top surface of the swamp, as it is seldom fit for constructing embankments. The best results are obtained by leaving the top surface nndc-r the roadbed undisturbed, as it usually 18 TRACK FOUNDATION consists of matted vegetable matter which will carry a considerable weight without breaking. To secure a proper distribution of the weight the filling material for the embankment is sometimes deposited upon a corduro}' foun- dation, but more frequently upon a brush mattress. In making embank-, ments over marshy land in Holland it is quite commonly the practice to first lay a mattress of willow boughs, 2 to 4 ft. thick, extending from toe to toe of the embankment slopes, before the earth filling is begun. On the Galway and Clifden Ky. a layer of brushwood and poles 3 ft. thick ha? been used as a foundation for embankments on bog land, with satisfactory results ; but the most successful construction in this line is reported to have been obtained by building the embankment with turf, without brushwood,. and then covering it with an 18-in. layer of stiff, marly clay, which hardens upon exposure to the air and becomes water proof. A source of trouble to- be guarded against with embankments made by filling over a peat bog is fire. In extremely dry weather peat, especially in a bog which has been drained out, will sometimes take fire and burn over large areas, smoldering away for weeks. It is something of a task to extinguish such a fire, as it will burn on in spite of light rains, and if not stopped in some way will burn right under an embankment and let it down. About the only way to fight it is to dig a trench across its path. Where the yielding material in marshy land is shallow or extends but a few feet below the surface, the stability of an embankment constructed thereon improves with increase in hight; but if the yielding material ex- tends to considerable depth any increase to the hight of an embankment only makes matters worse. In the latter case the weight of the superim- posed embankment causes it to sink into the mass of yielding material^ the displacement of which usually takes an upward course, bulging the ground on either side of the embankment. Not infrequently a large mound will be formed at either side of an embankment which has settled in this manner, sometimes overthrowing right-of-way fence and teiegraph poles or carrying them out of line. A remedy sometimes applied is to drive a row of piles a few feet apart along the foot of each slope of the embank- ment. As a better means of preventing settlement it has been proposed to construct a sort of pier as a foundation for the embankment, by driving a row of piles outside each line of slope stakes and placing a line of timber inside each row of piles as backing for a row of sheet piling. It is proposed to then tie the two rows of -piling together with rods and in this way prevent the mucky material underneath from getting away. The practicability of this scheme would, of course, depend a good deal upon the depth of the marsh. On ground of this nature, however, it is sometimes better not to fill at all but to build the track on piling. Where the track must run but a few feet above the surface of a marsh, and piling, for some special reason, is impracticable, a substructure which will bear up evenly under the track may be made in the following manner : Grade off the top surface just enough to get a fair bearing, and upon it lay closely, thick, wide cross ties 12 or 15 ft. long. Upon these place deep stringers of good length and lay them to break joints. Use about as many as would be required to support the track on piling. Upon the stringers place sawed cross ties and drift-bolt them, and lay the track about as it would be laid on a bridge floor. Each bottom tie should be made to lie in as good a bed in the mud as may be, without special reference to the other ties each side of it. The stringers can be evenly supported by spiking shims of proper thickness to the tops of such ties as are not touched by the stringers after they are put' to place and leveled up. In this manner the stringers transmit the weight over a larger surface than a shallow earth THE ROADBED 1 fill can and, when properly made, will maintain track in quite even sur- face. The stringers should be braced to a portion of the ties underneath them to keep the track in alignment. The structure is simply a track built with a double course of cross ties with stringers between the two courses. It is to be expected that there will be more spring in such a structure than in track supported on piles. Before roadbed construction through marshy land is undertaken sound- ings should be made by a portable pile driver, or other means, to ascertain the conditions underneath. If the material underneath is found to be ex- tremely soft or readily yielding to a considerable depth, it^is_svell worth while to consider the alternative of either abandoning the location or resorting to pile construction. The sinking of embankments into marshy ground is not of unusual occurrence. It has frequently happened that fills made through swamps or bog land or over strata of quicksands have dis- placed the underlying materials and entirely disappeared. Difficulties of this kind have sometimes been met by continuing to deposit filling material until equilibrium was obtained between it and the material which it dis- placed, but in numerous instances large sums of money have been expended in this way only to abandon the work in the end. In cases of this kind the most serious trouble has not usually developed until after the roadbed was subjected to the weight and jar of passing trains, but proper examina- tion beforehand, at nominal expense, might have disclosed tne exact nature of the conditions. The perplexity of dealing with such conditions has, in a large number of instances, convinced maintenance of way men that the most satisfactory practice lay in the permanent use of piling and in fore- going any attempt at earthwork. It will be interesting to give here the particulars in a case or two of the kind under consideration. In the summer of 1897 the Chicago, Indianapolis & Louisville ("Mo- non") Ey. made extensive changes in the location of its line in the vicinity of Cedar Lake, Ind. A stretch of track about 300 ft. long, on the 'old location, crossed the eastern side of a marshy basin, generally circular in shape and about 1,000 ft. across. In eliminating a curve the track at this point was relocated farther over, so as to cross about 800 ft. of the marsh, on a line running 125 to 200 ft. distant from the old location. Filling material was obtained from a heavy cutting to the south and work had progressed on the construction of an embankment, 'with teams, for a period exceeding a month, during which time a fill about 7 ft. in hight had been extended nearly half way across the marsh. No unusual settlement had ap- parently taken place, when, upon, taking up the work one morning, it was found that 150 ft. of the fill had sunk about 16 ft. and disappeared under a depth of 7 ft. of water. This, the first subsidence, is shown in Fig. 1, the bottom picture being a nearer view looking down into the depression. The track appearing in the background of this view was the old main track, about 125 ft. distant, which had carired traffic for years without extraordi- nary settlement. The work of filling was next taken up from the north side of the basin, the material being hauled out on 'cars which were unloaded from a temporary trestle. After considerable progress had been made this embankment also disappeared, carrying the trestle and track with it so sud- denly that a train of cars which was being unloaded at the time was barely saved from sinking with the track. This part of -the fill also sank beneath the water and no trace of the track was afterward discovered. The sink- ing of the embankment caused the surrounding surface to bulge upward, and large cracks 6 or 7 ft. deep were opened up along the lines where the top stratum was sharply broken off and deflected downward by the sinking of the earthwork. These cracks revealed a bed of peat extending the whole "20 TRACK FOUNDATION -depth and below. The surface of the water which covered the sunken em- bankment was 2 or 3 ft. below the general surface of the marsh. Moisture was absent on the top surface, as may be surmised from the fact that at the time the photographs were taken the peaty surface was afire and was burning over a considerable area. The hazy appearance in the background of the illustrations is due to smoke arising from this burning peat. There was no water in sight at any point on the marsh and the land was enclosed and used for pasture. Judging of the nature of the deeper earth structure from the foregoing suggestions of instability it was decided to abandon the project of filling and carry the track across on piling. Something of the nature of the underground sup- port may be inferred from the fact that the piles had to be driven to depths varying from 75 to 150 ft. in order to secure bearing of sufficient supporting power. The piles used were of oak in 25-ft. lengths, spliced by cutting off squarely and joining end to end, using a drift bolt, and two iron straps bolted through and through at each joint. The pile driver which appears in the illustrations had progressed with the work of bridging across about one-half of the affected territory. A most remarkable and strange circumstance associated with the sink- ing of this earthwork was that immediately following the first subsidence "the water covering the sunken embankment was thickly populated with Fig. 1. Sunken Embankment, C. I. & L. Ry. THE ROADBED Fig. 2. Sink Hole, Grand Rapids, Holland & Lake Michigan Ry. fish and frogs. The fish were of all sizes, from minnows to fish 6 or 8 ins. in length, and there was no peculiarity of eye structure such as is commonly found with subterranean varieties. There was apparently no grounds for ex- plaining the presence of these aquatic inhabitants on the theory of an under- ground passage, for there was no near body of water on the same level containing fish. The surface of the water in Cedar Lake, one mile dis- tant, is 28 ft. below the water level at the point of subsidence, as deter- mined by the railroad surveys, and the track between the two points is on a grade of 26 ft. per mile. The only satisfactory explanation is that at one time an open body of water existed in this basin, and that within a com- paratively recent period it was gradually overgrown with a thick bed. of peat moss, the surface of which, over the whole area, is now found in the- decomposed state, or in the form of peat. The looseness of structure in places would admit sufficient air to maintain the life of fish descended from the earlier inhabitants of the open water, and enable frogs, by burrowing, to exist on water which did not appear at the upper surface. There are- lakes in this country where growths of this kind may now be seen in progress. Beds of moss of astonishing depth have become extended from the shores and are slowly spreading over the surface of deep water. In some cases these floating beds are so firm and their buoyancy so great that one might readily mistake them for the real shore of the lake. This fact explains why railway embankments sometimes sink so suddenly. So long as the floating crust can bear up the material filled upon it, the settlement may not be excessive, but when the conditions of support or the increase of load become such that the crust breaks through, the sinking of the em- bankment then takes place suddenly. A case similar to the foregoing, but more nearly in line with ordinary experience, occurred some miles west of Chicago, on the Chicago, Burlington & Quincy Ky., in 1890. In making a fill for an additional track beside - an old embankment the earthwork settled 1 to 3 ft. per day. By persist- ently filling, however, until the track had settled altogether about 40 ft. and 95,000 cu. yds. of gravel and dirt had been dumped into an embank- ment 1000 ft. long, the settlement finally ceased. The earth on each side- 22 TRACK FOUNDATION of the embankment was crowded outward and rolled upward in a ridge 10 ft. high, and the telegraph poles were moved 12 ft. farther from the track. It is probable that the settlement did not cease until the new em- bankment found bottom on bed rock or upon a hard stratum. In filling for an embankment on similar ground at Barclay, Ontario, on the Can- adian Pacific Ry., it was estimated, from the amount of material required, that the depth of the "muskeg" must have been 200 ft. Figure 2 shows a sink hole on the Grand Rapids, Holland and Lake Michigan Ry., near the -city limits of Holland, Mich. Only a short time after the tracks had been laid the entire roadbed sank completely out of sight in the marsh, only a pond of water remaining visible at the surface. The distance across this sink hole was 700 ft., and an attempt was made to fill it with sand and gravel, but seemingly there was no bottom, and this method had to be abandoned. Support for the track was finally provided by building a pile trestle with 30-ft. piles spliced together. Instances of extraordinary settlement or sinking of roadbed in marshes have been numerous. An account of several occurrences of this kind in Pennsylvania, New Jersey and New York may be found in the Railway Review for Nov. 21, 1891. The usual remedy is either a pile trestle or filling material in sufficient quantity to "strike bottom." A plan followed on the Detroit & Milwaukee R. R. in crossing a sink hole 2600 ft. long and 60 ft. deep, 6 miles east of Grand Rapids, Mich., was to make a slab raft 10 ft. thick and fill on top of this with sand. There was some difficulty from the tipping of the raft in sinking, but in the end the work was suc- cessful. 4. Ditches. By proper drainage of the roadbed and track much expense which otherwise might be incurred in keeping the track to surface can be saved. Water must be kept out of, or drained away from, the bal- last as much as possible, to keep it from softening and settling or to pre- vent it from freezing and heaving in winter; the roadbed must, for the same reasons, be protected. The proper form, depth and size of a ditch and its distance from the ties depend a good deal upon surrounding con- ditions. First of all, most of the ditches that are needed are located in or around cuts. A very important ditch at a cut, if it be a side- hill cut or wherever the general surface of the ground slopes toward the cut, is the surface ditch, along the upper side, the purpose of which is to intercept surface water and divert it from the cut. The size of a surface ditch must depend upon the amount of land to be drained. Ditches of this kind must sometimes be made as deep as 5 or 6 ft., and correspondingly broad, in order to afford sufficient capacity for the large amount of water which comes in torrents during hard rain storms. The ditch should stand a good distance back from the top of slope, say 10 or 15 ft., and the dirt taken out of it should be heaped up on the side toward the cut. Where such ditches are too close there is danger that the seepage of water from the bottom of the ditch will soften the earth and cause slides. Spoil banks or material wasted in the excavation of cuts should be a good distance clear of the slope stakes, say at least 15 ft. In matters of this kind it is well to have in view a safe margin to cover possible improvements, such as the widening of cuts for more ditch room or for laying a second track or to obtain easier slopes. The surface or "top" ditch should be run each way from the lateral watershed to the nearest culvert or other opening under the track; or, at all events, the ditch should be so diverted that the water discharged will not find its way into the roadbed. Where the slopes of a cut are springy it is sometimes the practice to cut diagonal ditches down the slopes, at easy grades, to prevent DITCHES 23 excessive ^vash. Where springs of considerable size gush out of the slope the best way to take care of them is to lay lines of drain tile to conduct the water directly or diagonally down the slope and into the ditch or under drain. If the slope of the hillside is a long one it is well to have a second surface ditch farther up the hill, to catch the larger part of the flow during heavy storms. By properly ditching the region above the track much water which would not appear upon the surface directly above the cut, but which, nevertheless, would find its way out at the face of the cut, can be turned aside. Cases of this kind are on record where cuts, which beforehand had been wet and bothersome, have, by surface ditching, been made nearly or quite dry. Drainage Conditions as Affecting Land-Slides. Roadbed in side- hill cuttings through clay sometimes gives way and slides down the hill, carrying the track along ; likewise slides from above become troublesome when water gets between the strata underneath the surface. In cases of this kind it is well to keep water out of the cut, as far as possible, and make search to see if by draining out some pond or swamp at a higher level the water can be prevented from soaking through the ground toward the cut. Clay, if kept dry, is tough and will sustain pressure very well, but when wet it becomes plastic and will slide down the slightest grade if not confined in some way. Much of the region surrounding the Puget sound is of clay formation and has presented special difficulties to many kinds of engineering work requiring good foundations ; all the more so, too, because of the large and long continued rainfall in that region. Railroad ditch- ing properly done means the draining of a large area, sometimes. A good illustration of the influence of seepage on earth movements is to be seen on the Canadian Pacific Ry., along the Thompson river, about 200 miles east of Vancouver, B. C. Owing to the irrigation of terrace lands some remarkably large land-slides were developed in this locality before the railroad was built. One of the slides was about -J mile wide and mile long, back from the river, and covered on area of 155 acres. This enormous mass of earth dropped vertically, in one movement, to a depth exceeding 400 ft, at the back edge, and the lower portion was pushed ahead until it came to rest against a steep bluff on the opposite side of the river, damming the river to a hight of 160 ft. and forming a lake 12 miles long. As soon as the river rose above this dam the water cut its way through and the loose material was all swept away. About a mile distant another slide occurred having a width of 1880 ft., a length back from the river of 1575 ft. and covering an area of 66 acres. Within a distance of 6 miles there were four other large slides, across all of which the road had to be constructed, following closely the contour of the river bank at an elevation of 50 to 80 ft. above low-water level. The material composing these slides consists of soil overlying strata of sandy loam and clear sand, below which the material is stratified gravel and boulders. The whole rests upon a stratum of clay silt, arid it was upon this material that the sliding took place. The slide at each point occurred between three and six years after irrigation began. The largest slide above referred to was hastened by the bursting of a reservoir two miles distant in the hills, which poured a flood of water upon fields that had already become well soaked. As all the arable land at this point was caried down with the slide the irrigation was stopped and in the course of a few years the water drained out so completely that movement ceased and no trouble with the track has occurred. At the other slide, however, there has been con- tinued application of large quantities of irrigation water upon the culti- vated fields above the slide and in consequence the track has been continu- 24: TRACK FOUNDATION ally pushed toward the river, sometimes at the rate of 8 ft. in one night., the roadbed sinking at the same time as much as 4 ft. As the material has- been forced forward the river has washed it away, and from time to time a new roadbed has been built further back and the track moved over to it. This portion of the road has had to be carefully watched, and in order to maintain a safe passage across the line dividing the stable from the moving material the track has had to be continually shifted. Track Ditches. The office of a ditch near the track, sometimes called the "track ditch," is to drain off the water which falls upon the- track and that which runs toward it from the side. As the roadbed is subject to seepage from the water collected the drainage conditions naturally improve with increase of distance between the ditch and the track. Eeference is limited to comparatively near distances, of course,. and, standing in some relation to depth of cut, there is a limit of exca- vation to be calculated upon, at which the interest on extra capital invested to save repairs will balance with the saving so made. Beyond this limit: engineers are not supposed to go, unless the extra material excavated can be disposed of to advantage. An important study in problems of track engineering is to find the proper balance between efficiency and economy. Thus, under certain circumstances, it might pay better to leave some cuts at a minimum allowable width and, with the saving in expense so effected,, widen others out to exceed standard dimensions, possibly, than to make- all of the cuts the same width merely for the sake of appearance or of con- forming to some adopted standard. To put upon paper the form and dimensions of a ditch, called a "standard" ditch, and to suppose that it will meet all the conditions economically, may be landscape gardening but it may or may not be good engineering. The idea of standardizing is a valuable one if it goes far enough to provide a standard for each of the- se veral conditions requiring different methods of treatment. Forms of Ditches. In a general way ditches take two forms, ac- cording as the roadbed is shouldered or not. Where the roadbed is not shouldered the ditch is formed by sloping the roadbed at the sides to- meet the toe of the slope of the cut. At the ditch it is usual to increase- somewhat the general side slope of the roadbed due to the crowned center, so as to gain depth for the ditch. Ditches of this form are most com- monly found in narrow" cuts, where there is not room to shoulder the- roadbed and cut a ditch beyond the same. On some roads, however, this form of ditch is preferred for any width of cut, the advantage claimed being that such is the natural form,; and that if a shoulder is inter- posed between the track and the ditch it will eventually become rounded off or worn down to a common slope from the track to the back side of the ditch. The following are some of the roads on which this form of ditch is- standard, the width of single-track roadbed, from toe to toe of slope, and the depth of ballast under the bottom of the tie being given in each case: Southern Pacific, roadbed 16 ft. (18 ft. in regions where rainfall is heavy) ballast 8 ins.; Erie, roadbed 18 ft. 8J ins., ballast 12 ins.; Atchi- son, Topeka & Santa Fe, roadbed 26 ft., ballast 10 ins.; Cincinnati, New Orleans & Texas Pacific and Baltimore & Ohio roads, roadbed 18 ft., ballast 12 ins.; Pennsylvania (light-traffic lines), roadbed 19 ft. 2' ins., ballast 8 ins.; Philadelphia & Beading (in dry cuts), roadbed 18^ ft., ballast 8 ins. ; Southern By., roadbed 18 ft., ballast 6 ins. ; Mis- souri, Kansas & Texas, roadbed 18 ft., ballast 6 ins. On this road a middle strip of the roadbed 8 ft. wide is flat, the slope into the ditch ('6 to 1) starting under the ends of the ties. On each of these roads except the Southern Pacific the ballast is either shouldered out beyond the ends- DITCHES 25 of the ties or at least filled in against the ends. On that road broken rock ballast is dressed in this way but gravel ballast is not. The other of the two forms of ditches here considered is 'made by con- structing a roadbed of ordinary width, at sub-grade, as on embankment, and then excavating the ditch beyond the shoulder. The idea in construct- ing a ditch in this manner is to remove the water as far as possible from the ballast.' Ditches of this form are either V-shaped or trough-shaped, the latter having a flat bottom. As between two ditches with the same side slopes the flat-bottomed one will, of course, carry more water for the same depth than the one that is V-shaped'. Some prefer the V-shaped ditch, however, because it takes up less room and so ^conrrerrtrates the flow of water that it is more likely to keep itself clear, especially where the flow is small. On different roads the width of single-track roadbed with V-shaped ditches varies from about 19 ft. to 28 ft. from toe to toe of slope, although in some cases it is less than 19 ft. The toe-to-toe width of roadbed with trough-shaped ditches varies from about 22 to 28 ft., for single-track roads, although both wider and narrow'er measurements may be found, in cases. Size of Ditches. The width of roadbed in cuts and the size of ditch should be governed to some extent by the depth of the cut, because a long slope brings more water to the ditch than a short one, and conse- quently more sediment; hence the larger the ditch the less will be the trouble to keep it clear. In cuts of extensive length it may also be found advisable to increase the capacity of the ditches toward the outlet. AJ1 things considered, 18 ft. should be about the least allowable width of roadbed in cuts, and then the conditions must be favorable. A width of 18 ft. gives a space of 5 ft. each side, clear of the ties, to include he ditch; and it is about the least room that will allow for taking out ties in renewals, after assuming that the bank slopes well away from the ditch and that 'the track will be raised above sub-grade when ballasted. Under less favorable conditions, as when, for instance, the cut is high and long and much water is to be carried in the ditch, or when the cut is through wet material or material of the nature of clay, 18 ft. is entirely too narrow. As far as I am able to discover, the largest practice with trackmen, wherever the ballast does not exceed 12 ins. in depth under the ties, seems to be to maintain ditches at a distance of 7 ft. from the rail to the foot of slope at the back side of the ditch. This measurement gives a roadbed about 19 ft. wide; and right here it should be explained that the roadbed constructed or maintained by trackmen does not always measure fully up to the blue-print drawings in the chief engineer's office. In order to treat the subject comprehensively one should not fail to in- vestigate the least width of roadbed upon which good track can be main- tained, because it is upon roads of small earning capacity that questions of this kind must receive the most studious consideration. On roads with large earnings at disposal the question of the most economical width of roadbed largely disappears, and a width may be selected which is known to be sufficient to afford desirable conditions. I have therefore outlined, in a general way, what I consider to be typical conditions or situations to be met in ditching and the least width of roadbed which may apply to each case. In a dry gravel cut there is seldom need for a ditch, because ordinary rainfall does not, on such ground, run off on the surface. As, however, the material ought to be removed far enough back from the ends of the ties to make room for taking them out without too much digging during renewals, or without having to raise the rail, it is well to slope it away gradually TRACK FOUNDATION somewhat lower than the bottoms of the ties, as far as there is room. This will provide for those extraordinary rainfalls when water comes in quan- tities faster than it can soak away, and also for winter, when snow melts on frozen ground. Six inches is far enough to go below the level of the bottoms of the ties in this instance. Of course it improves appear- ances to make the ditch deeper, but as the whole foundation is porous this is a case where the requirements for ditches under almost all other condi- tions do not arise. Engraving A, Fig. 3, is an illustration of this ditch. In all other cases the ditch ought to fulfill the following requirements : it ought to carry the water away below the level of the bottom line of the ballast, because all ballast except common dirt is porous, and if the ditch be not lower than the bottom line of the ballast the latter will be soaked with water whenever there is any in the ditch; the ditch should also have at least a slight grade, which can be maintained, even on level ground, by 8 D/t Cut, Ballast 12 m.Deep. c Ditch in DryFarth Cuf< Ballast 6 ins. Deep. _ e maintained, for the reason that when the top of the roadbed becomes wet it will slide laterally, under the weight and shock of passing trains, and fill the ditch, if there be much of a slope toward the same. Mistakes are often made in cases of this kind. When the track is found to be settling and the ditch filling up, in such cuts, some trackmen will deepen the ditch accordingly, making allowance in depth for the plastic clay which they evidently think cannot be kept out. Such treatment only makes matters worse, for it weakens the roadbed by taking away its lateral support, and the material under the track will keep pushing into the ditch and the track will continue to settle. The proper thing to do is to widen out the cut to make room for a roadbed which slopes so gradually that it will not slide out a flat roadbed, comparatively speak- TRACK FOUNDATION I'/^wii^^i^ 6 'echo/? of Earth //? cu. \ \ t4-'o'~ .-". S'O -- Sec// or? of Ear//? 0s? Emfank'/vf fy 3 8 ectior? ofccmeflt/j?0 Grave/ Mast //? cu} Qecf/o/7 of cementing Grave/ Jfaf/ffsf on E/ntofiAmeflt Seer/Oft of coarse fff?d /0ose> 9'0'~ S&cf/on o f coarse- a /7 d /0cs& Grave/ fiaf/asf or? Em bank men. Sect/on of Stone a/ fast M cut -2'6'-+-~'0^ \v%zijm? k -.-9'0"- - y '' ! Serf/on of Stone Bal/ast or? I Eiittenkfmflt "^^ DITCHES 29 ing, or one which crowns just enough to drain the water into a broad, shal- low ditch next to the slope. In clay cuts ditches should never be formed by shouldering the roadbed. Eight inches below sub-grade provides depth sufficient to carry off the water, and the toe of the slope, that is the far side of the ditch should be at least 12 ft. from the center of the track. Engraving F, Fig. 4, illustrates this arrangement. The depth and nature of the ballast required is discussed in 12. One more case invites consideration, namely the ditch for dirt-ballasted track. Of course it goes without saying that such is not suitable ballast where there are springs under the track, but a ditch can be made which can drain the track of water coming from the side. . A cut ~bf"20-ft. width, for single track, will allow a shoulder about 2 ft. wide outside the tie ends, sloping away 2 ins., and a ditch about 18 ins. deeper, at an ordinary slope of 8 to 3, as illustrated by Engraving G, Fig. 4. The foregoing seven conditions cover perhaps all the general problems which arise in ditch construction, or those which call for different methods of treatment. The depth of ballast needed and depth of ditch should, after the nature of the cut is known, determine largely the width of cut, and not the width of cut the depth of ballast and depth and slope of ditch. Bearing in mind that the variety of roaabed and ditch meas- urements here presented is due to a close study of conditions, to ascertain what might be considered the least measurements applicable to conditions which obtain in common practice, it may be explained that it is not usual to find so many standards in practice with one company. It is more frequently the case that a single set of measurements, large enough to answer the requirements of the worst supposable conditions, is made stand- ard. The roadbed sections shown in Fig. 5 are an illustration of such practice, there being but one w r idth (18 ft.) for roadbed at sub-grade, both on embankments and in cuts, and but one toe-to-toe width of roadbed in cuts (28 ft.). These sections represent liberal roadbed measurements in cuts. The standards of the Illinois Central R. R., even though more liberal, are yet quite similar, the only differences being in width of roadbed at sub-grade, which is 20 ft. both for embankments and in cuts, making the width of ditch 6 ft. 3 ins. instead of 7 ft. 3 ins., as shown in the illustra- tions. The gravel ballast is shouldered out 18 ins. beyond the ends of the ties, instead of 6 ins., and stone ballast is shouldered out to 12 ins., instead of 6 ins. In excavating ditches the angular form of the standard cross section at the meeting lines of slopes (drawn that way for convenience of indicat- ing dimensions clearly), should not be followed. The various slopes form- ing the outlines of the ditch should meet by curved surfaces; as, for instance, the bottom of a V-shaped ditch should be rounded out and not brought to a sharp corner. The rounded corner is the shape which results from the forces of nature and it applies to the edges of embankment slopes and the top edges of slopes in cuts as well as to ditches. Where it can be done the ditch should be given grade sufficient to carry off the water with some rapidity; a fall of at least 4 ins. per 100 ft. is desirable, and 6 ins. per 100 ft. is a common specification. The dimen- sions of ditches shown on the standard drawings of railroads are usually sup- posed to represent the cross section of the ditch at the highest point. In case it becomes necessary to deepen the ditch to obtain a grade, as on level ground, the standard dimensions must then be exceeded. The standards of the Louisville & Nashville R. R., are different in this respect. On this road the ditches are made with a flat bottom and the width of roadbed between ditches is 16 ft. The grade of the bottom of the ditch must not be less SO TRACK FOUNDATION than 6 ins. per 100 ft., wherever such is practicable. In summit cuts the- ditches are 3 ins. deep at the summit and in level cuts they are 3 ins. deep at the middle of the cut, increasing to the standard depth o. 12 ins. toward the ends. In cuts where the grades are steeper than -J per cent the grade of the ditch is made parallel with the grade of the track and 12 ins. deep. This practice seems proper, because at a summit in a cut or at the middle of a level cut from which the ditches slope either way, there should be but little water in the ditches, and in long level cuts they must be made shallow at this point in order to avoid running too deep in obtaining the necessary grade. Deep ditches in soft material are objectionable, as already ex- plained, and in any material the deepening of ditches in deep cuts requires the removal of large quantities of material in making the slopes. Ditches should be made regular in width, and smooth, so that puddles of water will not stand in them, to afford a source for seepage. If, owing to the shape of the cut or for any other reason, a ditch cannot be made regular in width it is well to make the track side of it straight and par- allel with the rails. Where there is an embankment or fill adjoining a cut the offtake ditch or ditches should be diverted from the made ground and carried around on solid material. Neglect of such precautions has been the cause of many a bad washout during times of excessive rainfall or sudden thawing. The track ditch may be turned into the surface ditch and. the channel formed by the two combined may be run to the nearest culvert or stream. If the ditch must be carried over or near made ground the embankment may be protected against scour by paving or riprapping. Another arrangement having the same purpose in view is to conduct the water from the end of the cut through a line of large drain tile or sewer pipe laid to a good grade. Tile as large as 15 ins. in diameter has been used in cases of this kind. In long side-hill cuts it is seldom necessary to carry the water the entire length of the cut, as the ditch may be turned under the track at intervals, through box drains or culverts. In through cuts the ditches should be of the same depth on both sides of the track,. on curves as well as tangents. Tile Drains. Valuable assistance to the drainage can be obtained by laying drains of farm tile under the ditches (Engravings E and F, Fig. 4). Especially is this the case in wet cuts or in cuts where there is not room for a ditch of proper width and depth, and in cuts where there is trouble in keep- ing the ditch clear of sliding material. The utility of tile dram? or "blind ditches" has long been demonstrated by farmers, and the use of tile in rail- way work is increasing. The ditch serves to carry off water when it comes in quantities, as during storms or thawing weather, and seepage into the tile drain prevents water from standing jn the ditch at ordinary times and also drains out the roadbed to the level of the tile. Where under drains are used the ditches may be made shallower than otherwise, and in some lands of material this is a considerable advantage. To put the tile below frost, or at least below the action of hard frost, it is laid 2-J to -1 ft. deep. Water running continually in a tile drain will compromise the action of frost to some extent. In tile-draining a through cut it is usual to lay drains under both ditches. If practicable the drain should have some fall, 3 ins. per 100 ft. or -J in. per rod, being desirable. Tile is made in 1-ft. and 2-ft. lengths, but the 1-ft. lengths are preferable for the sizes up to 12 ins. in diameter. Bound tile at least 5 ins. in diameter is preferred for railroad service. It is frequently used in sizes up to 8 ins, in diameter, where the quantity of water so requires, and 6-in. tile is com- monly in use. In very long cuts it may be necessary to increase the size of the tile toward the end of the cut or to lay two lines of tile in the- DITCHES 31 same trench, but resort to either plan is unusual in practice. Glazed tile is stronger than the unglazed article for the same thickness, but if the unglazed tile is properly burned it is considered quite strong enough for practical purposes, and just as durable. Water enters a tile drain through the joints. Water under pressure will percolate through thu wall of the tile in some quantity,, but in ordinary drainage all that gets through in this manner amounts to but very little, and cuts no figure in drainage. Unglazed tile will absorb water until the pores become filled, but without some force behind it there is but little tendency to pass through. Experts in tile drainage work have special tools for~digging the trench and laying the tile. For excavating the top portion and body of the trench a long and narrow shovel or post-hole spade (Engraving B, Fig. 6) is used. The blade measures 5-| ins. wide at the step, 6 ins. at the cutting edge and is 18 or 20 ins. long. For sticky soil the skeleton ditching spade shown as Engraving A, with a blade 18 or 20 ins. long and 6^ ins. wide, is preferred. For taking out the bottom spading a round-pointed spade (Engraving D) with a blade 18 ins. to 22 ins. long, 5J ins. wide at the step and about 4^ ins. wide at the cutting edge, is used for ordinary soil, and for sticky soil the round-pointed skeleton spade shown as Engraving C, with a blade 18 or 20 ins. long and 4J ins. wide, is used. These tools enable the ex- cavation of a narrow trench but little wider than the tile, if desired, thus Fig. 6. Tools for Tile Drainage. saving something in material handled and expediting the work. To clean up the bottom of the trench for laying the tile a drain cleaner (Engraving E), consisting of a scoop with a long handle, arranged to draw toward the user, is employed. The blade is made of shovel steel, 15 ins. long and 4 to 6 ins. wide, according to the size of the tile. The handle can be adjusted to any angle convenient to the user, by raising the spring, and when the spring is in position the blade is locked against rocking. As a guide for dressing the bottom of the trench to a uniform grade it is customary to set grade stakes at intervals of aboHit 30 ft. alongside the line of the trench, using an engineer's level. A ditch line is then tightly stretched from stake to stake, to the grade for the tile, and after the trench has been nearly com- pleted in depth measurements are taken from this line for dressing up the bottom. A measuring instrument commonly used consists of a vertical staff graduated to feet and inches, with a horizontal sliding arm made fast by means of a thumb-screw. The arm carries a spirit level and is long enough (about 2 ft.) to reach the ditch line when the staff is stood in the trench. In railroad work where the track is in good surface the rail might be used as a reference for the grade of the trench. In order to have the trench in smooth, condition for laying the tile, the workmen dress it to grade without stepping on the bottom. The sections of tile are laid to place with a hook on the end of a long handle. For laying tile through quicksand a sheet iron box, open top, bottom and rear, and commonly known as a "coffin," is used. This box, and the use of the same are descibed in 1, Supplementary Notes. 32 TRACK FOUNDATION The tile should be laid to a straight line and uniform grade, with the joints fitting closely. Some use a pole of round or square timber somewhat smaller than the inside diameter of the tile, to keep the tile properly lined up while it is being tamped and covered over. Each advance section of tile is strung upon the pole as it is pulled ahead one section at a time, the rear end of the pole remaining continually within the covered tile. When laying tile on very soft ground the precaution is sometimes taken to lay a narrow board in the bottom of the trench, to prevent displacement of the tile sections. When filling in the trench with loose material it is the practice with some to cover the tiling with inverted sods, moss, slough grass, hay, straw or some such material, to exclude fine particles of filling which might be washed into the tile \through the joints. As a means of aiding seepage toward the drain some recommend filling the trench with coarse gravel or cinders, but not with loam or sand; while others of long experi- ence claim that the water will readily find its way into the drain through any material, however compact, and for filling in the trench such men prefer to use only the material excavated, without straw or other screening material. The bottom spading is considered the best material to place directly upon and surrounding the sides of the tile. Stiff blue clay is considered excellent material for covering over tile. In cuts with wet slopes, where the bank is liable to slide, it is sometimes the practice to cut ditches diagonally down the slope to ease the grade for the running water. A better plan, where there is tile sub-drainage for the ditch, is to conduct the water down the slope through covered diagonal branch drains leading into the main line of tiling. A pile of loose stones or wire netting should be placed over the outlet of a tile drain to keep out muskrats and other small animals. To lay drain tile properly requires experience, and some railway companies find it cheaper and productive of better results to employ experts who have worked among the farmers, to do this work. Twenty-three cents per rod for the work of digging the trenches (3J ft. deep) and laying the tile is a price that has been paid by the Chicago, Bur- lington & Quincy Ry. to contractors. Mr. Alexander Birss. Prairie, Wash., who was formerly engaged for a great many years as a tile-drainage con- tractor, in Iowa, has kindly favored me with some interesting information on tile drains and the work of laying them, which may be found in 1, Supplementary STotes, in the back part of this book. As a substitute for tiling a continuous bundle of poles, trimmed of their branches and placed butts and tops, is sometimes laid in the bottom of the trench and covered over. Blind ditching may also be done by partly fill- ing the trench with broken stones, preferably placing a plank in the bottom of the trench. A form of blind ditch much used by farmers in the eastern states is laid with flat field stones as a bottom paving, and on top of these, flat stones are stood edgewise leaning toward the middle of the trench from both sides, to form an inverted V-shaped opening 2 or 3 ins. wide at the bottom. Over these stones other stones are thrown in loosely and covered with the soil. The most desirable way to ditch yards is by sub-drainage with tiling or, if the area to be covered is extensive, with branch drains of tile feeding into sewer pipe mains. Catch basins at points where water is liable to collect are desirable, as they prevent the formation of puddles of water between the tracks, which get covered with ice during freezing weather. The work of switching may be considerably expedited by main- taining the footing in good condition in all kinds of weather. Ditch Paving. The paving of ditches with cobble stones is practiced to some extent. Where coarse gravel is on hand the paving material is CULVERTS 33 easily obtained and the work of laying it is not very expensive. Paved ditches retain their shape better than unpaved ones, because they are flushed by heavy rains, and if filled by sliding material or sediment washed down there is a good bed to shovel upon when cleaning out the ditch. To improve the appearance of ditches in the vicinity of stations the paving is sometimes whitewashed. Whitewash will keep the paving clear of grass, and if salt is mixed with the lime the whitewash will adhere better to the stones. Concrete paving or lining is also applied to ditches on a number of roads. Brick, cement and asphalt are materials used for paving some of the ditches on the Pennsylvania R. E. In ditches through soft material on steep grades some kind of paving is necessary- to prevent gullying in time of hard rain storms. On some roads old ties have been used to good advantage in ditches where such protection is necessary. Retaining Walls for Ditches. Various means, some of which are mentioned in 160, are resorted to for maintaining open ditches along sliding banks. A common method of securing the foot of a sliding bank is to build a thick masonry retaining wall and lay tile drains at the back side, under back filling of coarse gravel or broken stone. Such walls are usually built to a heavy batter, like to 1, and topped out with heavy coping stones. The bank is then sloped from the top of the wal], reduc- ing the general slope and lessening the tendency to slide. In order to make sure provision for drainage, weep holes through the wall with a ditch in front of it are recommended. An interesting piece of work that may properly be referred to in the present connection is a concrete slope fac- ing constructed at Chestnut street, St. Paul, Minn., to protect the tracks of the Chicago, Milwaukee & St. Paul Ry. from falling rock and other mate- rial. At this point there is a side-hill cut through soft sandstone for several hundred feet, the sandstone being overlaid with a limestone ledge and the latter surmounted by a glacial drift formation of sand, gravel, boulders, etc. The bank rests at a slope of about 1 horizontal to 2 vertical, and a cut-stone masonry retaining wall had been built along part of the distance to support the limestone ledge, which was constantly being undermined by the disintegration of the sandstone. As a means of cheap- ening the construction, a concrete facing wall, w r ith brick pilasters at intervals to support the limestone ledge, was substituted for the remainder of the distance, it being assumed that if the sandstone could be protected against rain and frost its stability would be secured. The foundation for the facing wall and pilasters was put into the sandstone 4 ft. below rail level. The facing is 256 ft. long and 56 ft. high, and was built up by depositing concrete behind a w r ooden form built from the foot of the slope by stages and supported on bolts anchored to the standstone. These bolts were jointed about the middle of their length, and after the con- crete had hardened the outer half of the bolt was withdrawn, the hole filled with cement mortar, the concrete facing thus remaining bolted to the sandstone bluff. The average thickness of the concrete facing below the ledge is 2 ft. 5 ins., and above the ledge, 3 ft.. 2 ins. 5. Culverts. The drainage of roadbed comprises ditches and cul- verts, the purpose of the latter being to convey ditch water or small streams underneath the track or to permit the escape of rain water, melted snow or springs draining toward an embankment. To be serviceable under all con- ditions a culvert must answer the requirements of size, and be secure in foundation and end construction against washing out. Respecting the first essential, engineers when laying out culverts should exhaust every resource available for estimating the quantity of water liable to flow aloiio- the streams crossed, especially those which are found dry at times. Where 34 TRACK FOUNDATION opportunities are at hand, as in settled districts or where roads or railways traverse the country, the most satisfactory way to estimate the size of cul- vert openings is by observation of the volume of now during high water, either at the time or by high-water marks. To find the latter, observation may be made of Existing openings on the stream or by examination of the banks, preferably where the stream is contracted; or by inquiry of parties familiar with the locality. Along w T ith every party doing the preliminary surveying for a railroad there should be some man experienced in exploring or "cruising," who should scour the country surrounding for such infor- mation regarding the rainfall and the streams as will be of benefit to the company in the proper construction of its bridges and culverts. In placing culverts on an old road the determination of the sizes of the openings is less problematical, because exact records of high water should then be obtainable. In constructing railways in this country, and particularly in the West, it is extensively the practice to bridge the water courses with timber trestles, so that the construction of permanent works at the drainage openings is usually postponed until the wooden structures need renew- ing. This gives a period of some eight or ten years, during which time it is customary for both the frridge and track departments to keep record of high water at the various openings, as reported by the bridge inspectors and the section foremen. In a large proportion of the cases, therefore, there need be no uncertainty regarding the capacity of culvert openings. The con- tingency does sometimes arise, however, that the capacity of culvert openings must be determined upon where reliable information concerning the streams cannot be had. Under such a circumstance the engineer is compelled to resort to some basis for estimating the maximum rate of discharge through each opening. The investigation of such problems proceeds so largely upon mat- ters of judgment that many are disposed to regard the accepted methods of calculation as in large degree conjectural, or as processes more or less en- tangled with guesswork. While it is true that much of the data made use of in such cases are necessarily assumed, or even guessed at, it is also true that some determination is compulsory, and guessing by method is certainly preferable to guessing at random. The amount of confidence to be reposed in calculations of this kind depends, of course, upon the experience and observing capacity of the engineer in charge. Calculation of Maximum Flow. Where reliable information cannot be obtained regarding the maximum flow of the streams the size of opening or, what leads to the same end, the maximum flow through the opening, is determined by some empirical rule or method of calculation. One of the simplest rules is to base the unit of opening area upon acreage. For instance, it is commonly the practice to allow a s'liiave foot of culvert opening for some certain number of acres drained, the relation of drainage area to the unit size of opening being ascertained from experience with the topographical conditions and rainfall of the particular section of country. To give one or two illustrations of such practice, the Chicago, Eock Island & Pacific Ky. allows for drainage, in Nebraska, Kansas and eastern Colorado, as follows, a single line of cast iron pipe being referred to each case: 16-in. pipe for 20 to 40 acres; 20-in. pipe for 30 to 60 acres; 24-in. pipe for 45 to -90 acres; 30-in. pipe for 70 to 140 acres; 36-in. pipe for 110 to 220 acres; 48-in. pipe for 180 to 3GO acres. These allow- ances are based upon 14.3 to 28.6 acres per square foot of opening, or an average of about 21^ acres per square foot of opening. The wide latitude left to the discretion of the party in charge of construction enables him to take into account the variability of the topographical features, such as the slope of the ground, the state of the soil (whether cultivated or not), and CULVERTS 35 the shape of the drainage basin; as, other features being similar, water draining out of a circular valley will flow off more rapidly than from a long, narrow valley of the same area. The rules in force on the New York Central & Hudson Eiver R. E. when reliable records of the flow of water to be taken -care of at any new culvert cannot be had, are similar. For 5 acres of steep land or 10 acres of flat land, 10-in. pipe is used; 12-in. pipe for 10 acres; l()-in. pipe for 20 acres, and so on up to 36-in. pipe for 110 acres. Com- pared with rules in force on some other roads for small culverts these open- ings seem small. On the Missouri Pacific Ry. it has been the practice to allow one square foot of opening to drain four acres of steep or moun- tainous land or six acres of flat or rolling land. Another way of determining the area of culvert openings is by the use of an empirical formula, in which the factors are the drainage area and a variable coefficient to suit the conditions of the locality. The best known formula of this class, or the one which has been most extensively used in American railway practice, is the Myers formula, proposed many years ago by Mr. E. T. D. Myers, since then president of the Richmond, Fred- -ericksburg & Potomac R. R., by which Area culvert opening in sq. ft. = C X V (Drainage area in acres) The values usually given to C are : for flat or slightly rolling ground, 1 ; for hilly ground, about 1.5 ; and for mountainous and rocky ground, 4. The important respect in which this formula differs from the foreging rules is that the size of opening varies as the square root of the drainage area instead of by a straight proportion; which would make it appear that the so-called "rules" require an opening too large for the larger drainage -areas. Such is probably the case, for it is well established that the rate of flood discharge from a large area compared with the rate from a small area for the same rainfall and same duration, is not as great as the ratio of the two catchment areas. In the Talbot formula the size of opening is made to vary as the fourth root of the cube of the drainage area, thus : Area culvert opening in sq. ft. = C X 4 V (Drainage area in acres) s In this formula the coefficient C takes a value varying from f to 1 for steep and rocky ground; and -J for rolling agricultural country subject to floods at times when snow melts, where the valley is three or four times as long as it is wide; if the valley is longer in proportion to width the value of is decreased still further. In districts where snow does not accumulate, C is taken at or -J, or even less, for oblong valleys. In the case of either of these two formulas it is quite apparent that experience and good judgment are essential to a proper choice of coefficients. In any case of uncertainty in this respect the opening should be given the benefit of the doubt. The most thorough way of getting at the proper size of waterway, where authentic report concerning the maximum flow of the stream is not procurable, is by a survey of the drainage basin and the various con- ditions which affect the situation. Such work is sometimes undertaken for openings of considerable importance. On the Atchison, Topeka & Santa Fe Ry. it is the practice when constructing new culverts to send an engineer- ing party around the watershed and have a rough survey made of the drainage basin. Although there are numerous formulas which may be applied to some of the elements concerned in an investigation of this kind, I think I can do no better than follow the sense of a paper on this subject pre- sented before the Institution of Civil Engineers by Mr. George Chamier, in 1898. The elements which must be taken into account as a basis for cal- culating the maximum discharge are (1) drainage area, (2) rainfall, (3) 36 TRACK FOUNDATION amount of surface discharge and (4) the diminution in proportionate flood discharge due to area. Regarding the drainage area the form and greatest length of the catchment basin are all important, as well as the extent, for upon these features depends the time required for the flood water to reach the outlet from all parts of the drainage basin. Thus, with surrounding ridges of the same elevation, in either case, the discharge of flood water from a circular basin takes place more rapidly than from an oblong basin, for the reason that the distances traversed by the various streams are shorter and the declivities greater. The general slope of the ground over the catchment area and the outlines of the valley traversed by the main stream are also important, as affecting the velocities of the streams. The estimation of the time required for the flood water to reach the outlet from the farthest point of the basin calls, of course, for the judgment of the investigator. The velocity of the water increases as it collects into well defined channels. The time required for rain water to flow off the surface into the brooks is rather conjectural, in any case, but the rate of flow over grassy surface may be taken at -J mile per hour for moderate slopes, and 1 mile per hour for steep side-hill. Under average conditions the velocities of streams range from 2 to 4 miles per hour, but in mountain torrents and rapid rivers much higher velocities have to be considered. The velocity in any case can be ascertained approximately from the dimensions and inclina- tion of the channel, with some assumption as to the probable volume of flow at times of flood. The reliability of all such estimates depends, of course, upon the experience and judgment of the calculator. As to rainfall it is desired to know the maximum -downpour during a period corresponding to the size of the drainage area that is, for such a time as is required for the flood water to reach the outlet from the farthest extremities of the basin. The maximum rate of precipitation occurs only during short periods, of an hour, or a few hours, at most, so that, for the smaller drainage areas, for which the duration of fall to be considered is necessarily short, it would be incorrect to estimate the maximum fall in proportion to the maximum daily fall. Thus, for instance, it is not un- commonly the case that 25 per cent of the maximum daily fall is registered in an hour. In order to get at flood discharge it is, of course, essential to have some record of the rainfall for the section of country, and in order to anticipate the greatest rainfall for short periods with a reasonable degree of assurance it is necessary to have approximate data as to the diminution of the rate of fall with the duration. As to surface discharge, it is known that, owing to absorption of the soil, evaporation and percolation into subterranean passages, only a portion of the rainfall need be taken into account. The ratio of the water which flows off the surface (and finds its way into streams) to the total amount of rainfall is known as the "coefficient of surface discharge." For countries where heavy rains are liable to occur when the ground is frozen this coefficient is usually assumed at f, while for rocky mountain slopes with- out fissures, very steep ground, or paved streets the assumed value may exceed 0.80. As a general rule the coefficient of surface discharge is taken at some value between J and f . Mr. Chamier's estimates are as follows: For flat country, sandy soil or cultivated land the coeff. disch. is f aken at 0.25 to 0.35 ; for meadows and gentle declivities, absorbent ground, 0.35 to 0.45; for wooded slopes and compact or stony ground, 0.45 to 0.55; for mountainous and rocky country or non-absorbent surfaces, 0.55 to 0.65. Tt is clear, of course, that the maximum ratio of surface discharge to rain- fall does not obtain until the ground has become thoroughly saturated. CULVERTS 37 Aside from the diminution of the discharge due to the above causes there is also a diminution due to causes which act upon the water after it has collected into streams, such as evaporation, which in hot climates is great; and absoption by overflowed lands or by irrigation; or percolation through the banks. In limestone countries streams of considerable size sometimes entirely disappear into underground channels. And then the flow of some streams is impeded, and the rate of discharge diminished, by obstructions in the form of dams or accumulations of flood debris, while lakes and swamps are well known regulators of flood discharge. For average cases Mr. Chamier gives ( 4 ~\/M 3 ) -j- M as the ratio of- decrease of flood discharge due to area, where M denotes the area of the drainage basin in square miles. Observation of, and experience with, the condi- tions in particular localities would quite likely find different powers of M suitable to the various conditions obtaining. Thus it appears that in deter- mining upon the data for the solution of the problem the judgment of the investigator is called into service at every step. Having considered the various elements of the problem in some detail the formula for flood discharge at the outlet follows by the simplest logical process, being Q = A X R X C, where Q denotes the maximum discharge in cubic feet per second; A, the number of acres ; R, the average rate of greatest rainfall anticipated, in inches per hour, for such duration as will bring flood water to the out- let from the most distant point of the drainage basin ; and 0, the coefficient of surface discharge. If the drainage area exceeds 1 square mile, the formula must include the factor for diminution of discharge according to area, and it then becomes Q = A X R X G X (V ^ 3 ) + M; or, substituting for A in terms of square miles, the factor (A -f- M ) disap- pears and we have Q = 6-40 XRXC X V^ 3 One inch of rainfall per hour over a surface of 1 acre is at the rate of 1 cu. ft. of water falling per second, which is the rate of discharge, supposing all the water to flow off. Having ascertained the maximum dis- charge to be anticipated at the outlet, the area of the opening will depend, of course, upon the velocity of flow, which is frequently assumed at 10 ft. per second. In the case of moutain torrents and rapid streams, where the velocity exceeds this figure, the error is on the safe side; and if the natural velocity is less than the figure assumed the amount of head necessary to produce a velocity equivalent to the difference is but small, and if the foundation of the structure is secure against scour there need be no concern if discharge occurs under moderate pressure. Before dismissing the subject of calculating culvert openings it should be borne in mind that rainfall is not always the only source of flood water to be taken into account. In regions where snow accumulates or falls to considerable depth the highest floods may occur when the snow melts, as then the flow of water may be due to heavy rainfall and melting snow combined. Thus, for instance, in some parts of the State of Washing- ton, on the western slope of the Cascade mountains, the most troublesome freshets occur late in the fall, when heavy rains, accompained by a "Chinook" wind, fall upon heavy accumulations of October snow. To know the extent to which melting snow contributes to the volume of flood Wiitcr requires special knowledge of the climatic conditions obtaining in particular sections of country, and the matter is so important that it should never be lost sight of in fixing upon culvert areas. In building a 38 TRACK FOUNDATION road through a wooded country waterways should be made about double 1 the size found necessary for use at the time they are built, so as to- allow for the increased rate of drainage after forests are cleared away,, swamps drained, etc. Open Culverts. Where the track crosses small rapid streams which wash down large quantities of drift, and the track is close to the bed of the stream, it is sometimes necessary to construct an open culvert, in order that the opening may be accessible for cleaning out when it becomes filled or obstructed. Formerly it was much the practice to construct such culverts by merely laying two stringers across walls of masonry or heavy sills, to carry the rails. In some sections such is known as a "beam" culvert. Such openings in the track are not to be advised, as in the case of a derailed car or truck running into the same there is certainty of wreck to the train. Where an open culvert is unavoidable a standard bridge floor should be built over the opening. When it becomes necessary to clean out out such an opening it is an easy matter to remove some of the ties or spread them apart. Another occasion for shallow openings under the- track arises in irrigation districts, where the right of way is crossed by ditches or canals in which the water level is but slightly lower than the track. On the El Paso division of the Southern Pacific road trough stringers have been used for the support of the track rails over irrigation ditches. This stringer is formed by riveting two 12-in. channels (placed back to back) to a third channel of same width placed open side down- ward between them. The rail rests upon 4xl2xl2-in. creosoted blocks placed in the trough, the depth of which is such as to bring the top of rail flush with the top of the stringer. The fastening for the rail consists of clips,, with bolts passing through block and bottom channel. For 15-ft. spans the sides of the trough are formed of channels each weighing 50 Ibs. per foot, and for 12-ft. spans the channels weigh 30 Ibs. per foot; the bottom channel used in either case weighs 30 Ibs. per foot. At the ends the stringers are anchor-bolted to 12xl2-in. caps with bearing plates 1 in. thick between the two. On the Pacific system of the same road a similar method of support is employed in crossing irrigation ditches, the stringer supporting each rail in this case consisting of two pieces of T-rail 5 ft. long, spaced just far enough apart to permit the flange of the track rail to lie in the opening between their webs. The flange of the track rail fits closely under - the heads of the two stringer rails and is supported upon bolts passing through the webs of the stringer rails at intervals of 18 ins. In some places there are as many as three consecutive spans of these T-rail stringers. Wooden Box Drains. Box drains made of plank of durable lumber, like cedar, are admissible for small cross drains where the depth under the track is not sufficient for masonry or pipe culverts. Such a box may be made by spiking together four 3xl2-in. planks, standing the side planks within the edges of the bottom and cover planks. It should be made 15 or 16 ft. long for single track roadbed, or long enough to reach to the ditch line. The ends may be sloped to conform to the side of the ditch or slope of the shoulder. When placed between the ties such boxes are usually left open on top, the sides being held apart by flat strips nailed across the top edges. If there is much fall in the water directly upon leaving the box the ground should be paved with stones for a safe distance, or, if the quan- tity of water amounts to nothing more than a trickling stream, as from a permanent spring, it may be conducted away by a V-shaped trough made by nailing together two Ix6-in. boards. Where one box of this kind is not large enough to carry all the water which may come at times, and the- track is not high enough above the stream to admit of a deeper closed cul- CULVERTS 39 vert, a partitioned box may be used, spacing the partition planks, say, 12 or 15 ins. apart, the number of partitions made depending upon the width of opening desired. In this case the bottom and cover planks should be spiked on crosswise the box. All such small passageways should be at least 6 ins. below the bottoms of the ties. The use of a box drain is superior to the practice of leading small streams or springs under the track by an open ditch and placing the track upon stringers thrown across the ditch. These stringers will continually settle and give trouble. For unimportant side-tracks such a makeshift may answer well enough, using ordinary 8-ft. ties for stringers. To prevent the track ties from slewing out of their proper positions they may be drift-bolted to the stringers, or cleats may~be nailed to the stringers between the ties*, or a board may be laid outside each rail, parallel to the same, and nailed to the ties. A substitute for a box drain is sometimes made by building a walled trench about 12 ins. wide in the clear and paved on the bottom. The trench opening comes in the space between two ties, and, being open, can readily be cleaned out when clogged with mud or ice. To prevent people from stepping into such openings in the dark they should be provided with a removable plank cover. Along side-hill cuts it is a good plan to carry springs (where there are but a few of them some distance apart) directly across the track at the point where each comes out, instead of allowing two or more to flow together before leading them across. In long through cuts where the track is curved and water runs continually, it should be carried under the track from the ditch on the outside of the curve to the ditch on the inside of the curve, at frequent intervals, so that it may run against the bank ; if it was to run altogether in the ditch on the outside of the curve it might cut into the roadbed or into the ballast on the shoulder. Timber Culverts. In districts where timber has been plentiful it has been used in culverts a great deal, especially where difficulties have stood in the way of delivering permanent materials at the site of. the culvert in time for the graders. In timber countries such culverts can be quickly and cheaply built, and under certain circumstances such construc- tion is undoubtedly economical. As most kinds of wood placed in the ground will rot out in 8 or 10 years the larger number of timber culverts have been built only with the idea of temporary construction. In such cases it is intended to make ample allowance in size for reconstructing the culvert at some future time with durable materials, as of masonry or pipe, placing the new structure inside the old one without disturbing the embankment, which, during the life of the wooden culvert, should become well settled. Where durable timber, such as cedar, is obtainable, however, wooden culverts are sometimes built with a view to permanency. On the Canadian Pacific Ey. cedar timber is used in small culverts quite extensively and in the west- ern parts of Washington and Oregon, where such timber is abundant, it is largely used for railway culverts. The estimated life of such timber is 150 to 200 years, as determined by the fact that almost everywhere in the cedar forests trees may be found lying in the ground partially or wholly covered, in perfectly sound condition, with other trees growing on top of them as old as the age stated. Some consider that timber of this quality will outlast many kinds of stone, where the masonry is exposed to mois- ture and freezing. On the Southern Pacific road timber barrel culverts built of creosoted staves are used to some extent. The material for the culvert, including the portals, is cut according to plan before creosoting. Culverts of this material are built in sizes of 24 ins., 36 ins., 48 ins., 66 ins. and 72 ins. diameter, the cost, for material and labor, not including the portals, ranging from $1.06 to $4.90 per lineal foot of pipe. There 40 TRACK FOUNDATION are also some locations where timber culverts are applicable to better sat- isfaction than others, owing to the difficulty of obtaining suitable founda- tions. Such is the case on marshy ground, where, in considerable depth of peat, the cost for masonry foundations would be heavy. Furthermore, peat is said to have a preserving effect on timber, while in some cases bog water has been found to act injuriously on cement and concrete. Timber culverts of large size, or when placed under high embankments, are usually built on a flooring of 6x8-in. or 6xl2-in. timbers laid on flat, and in contact, crosswise the channel. These floor timbers usually project some distance beyond the sides of the culvert all the more so if the foun- dation is soft or yielding. The side walls or partitions (if any) are formed of 8x8-in., 10xl6-in. or 12xl2-in. timbers -laid on top of one another and drift-bolted together and to the floor. In some cases the floor timbers are gained out for the side-wall timbers about 1 in., forming a shoulder to prevent the side wall from being crowded in by earth pressure. The thick- ness of the cover timbers may vary from 6 ins., under the outer portions of the slope, where the depth of filling is shallow, to 8 ins. farther in, and to 10 or 12 ins. under the central part of the embankment, where the fill- ing is deepest. Of course the span and depth of filling has all to do with the thickness of the cover timbers, which are laid crosswise the culvert, with every fourth, fifth, or sixth piece notched 2 ins. over the side timbers to take the thrust of the side walls. The span of opening in timber culverts usually ranges from 3 to 6 ft., partitions being used if a single opening of the larger dimension does not afford sufficient area. Old bridge timbers are riot infrequently utilized in culverts of the kind here considered, and what are known as "bridge-tie" box culverts are sometimes built of new tim- ber. Such are constructed of 6x8-in. bridge ties laid on flat in the floor and walls, and edgewise in the covering. On the Tennessee Central Ry cul- verts are constructed of oak timber, the largest openings so built _being 4 ft. wide and 5 ft. high. The sub-sills are 10x1 2-in. timbers laid flat and the side walls are 12x1 2-in. timbers drift-bolted together and stepped off on the faces. The floor is laid with 2-in. oak plank and the covering of the culvert consists of 8xl2-in. timbers laid flat, or 12xl2-in. timbers, varying with the hight of the embankment. An interesting application of the scheme of building a wooden culvert to be reinforced later with more durable material inside is to be found on the Chicago, Burlington & Quincy By., where a heavy timber barrel culvert is first constructed, and after the embankment has settled the barrel is lined with brick. These culverts, which are made as large as 6 and 8 ft. in diameter, are built of staves 10 or 12 ins. thick, according to the size of the structure, and 8 ins. wide at the outer circumference. The staves are drift-bolted together and formed over heavy rings made of old rails, spaced 10 ft. apart. These rings remain in the culvert, and to prevent distortion of the barrel where the pressure of overlying material is excessive it is propped with heavy posts, and cross bolts are placed to prevent bulging at the side. After the embankment has ceased to settle the barrel is lined with a single ring of brick placed edgewise and faced with a layer of cement mor- tar, and parapet walls of stone masonry are built. Stone Box Culverts. Masonry culverts of suitable weathering stone, if properly built, are very durable, and in localities where such material is obtainable within convenient distance it is frequently selected for per- manent work. The side walls of stone box culverts are usually laid with rubble stone, and preferably in cement mortar, so as to provide for discharge under head. Water discharging under head through a dry stone box will be forced behind the walls, gradually carry out the back filling and CULVERTS 41 eventually cause a washout. Sand in embankments will also find its Avay through the openings in such culverts, leaving cavities which will cause the roadbed to settle or lead to a washout should the culvert become surcharged. Right in this connection attention should be called to the importance of filling over and around the culvert with material which will become compact and form a barrier against filtration through the bank. If the space about a culvert is filled in with loose stones and the water becomes dammed, part of the flow will take place outside the culvert opening, and where the loose stones meet the earth filling the water will cut a hole for itself and cause a washout. Some remarks contained in a letter from Col. E. T. D. Myers to a committee of the Association of Rail- way Superintendents of Bridges and Buildings,' in 1897, bearing on this matter, are to the point. He says, in part : "I am persuaded that it is rather in the superior construction, the infin- ite painstaking to insure the safety of a culvert when it ceases to be a mere covered channel and becomes a pipe discharging under pressure. When this takes place the ordinary culvert is too apt to fail to do its duty. Between the hastily constructed dry stone box and the thoroughly-built concrete, brick, or stone culvert there is room for a legion of catastrophes. . . .' . I am of the opinion that it is more often the crude method of construction than the underestimation of the area of the waterway that gives us trouble on the railroads. When a railway embankment is called upon to act as a dam, as it may be in great floods, it should possess the qualities of a dam, and the outlet from the piled-up waters above it should possess the same integrity as the drainage culvert of a reservoir. Its foundation should be as secure, its masonry as impervious, the embankment immediately surrounding it as free of voids, the inlets and outlets as carefully protected from abrasion." Whatever the size of opening required the culvert should, if the depth of filling will admit, be made high enough to permit a man to walk through it say 4 ft., if possible, although a less hight will answer. On some roads the minimum size of masonry box culverts is limited to convenient propor- tions, as on the Northern Pacific Ry., where the smallest opening allowed is 9 sq. ft. clear of all obstructions, the hight of the opening never being less than the width. In other cases it is considered that nothing is saved in making stone box culverts smaller than 3 ft. square, for streams however small. As to width, the natural channel of sluggish streams may be con- tracted to some extent, where ordinary conditions prevail, but in building over rapid streams or ravines it is not safe to encroach upon the natural width of the stream as indicated by the channel which it has cut for itself. The hight of the culvert floor relatively to the surrounding surface is important. In the case of a well defined stream it is of course necessary to go at least as deep as the bed of the stream, in order to secure a suit- able foundation, but in any case the culvert should be low enough to drain low-lying land without backing the water, particularly where the land is under cultivation. But unless the land falls away immediately from the outlet the culvert floor should not be lower than this. Submerged culverts are unsafe, as, sooner or later, they are almost sure to silt up and become reduced in effective area if not completely obstructed. If a channel be cut from the outlet deep enough to drain the culvert floor the culvert may be placed lower than the situation would otherwise permit. The foundation work of culverts signifies all that the term implies in connection with other structures of permanent character. On a solid substratum, such as rock, hardpan, gravel or firm clay, the matter is easily decided upon, as then it is only necessary to prepare a smooth bed for the footing courses. On yielding ground some additional means of support 42 TKACK FOUNDATION should be resorted to. This may consist of a timber platform or grillage,. to distribute the weight of the walls over more surface than would he-- possible with the footing courses ; or it may consist of a bed of concrete ; or a brick or concrete invert; or, if the ground is yielding to an unusual degree, it may consist of piling overlaid with a timber platform or bed of concrete. The use of timber is not advisable unless the foundation is- continually submerged. For light walls the timber foundation may consist of a simple flooring of timbers laid in contact, crosswise the direction of the walls, but for heavy work a framework of crossed courses is usually required. Old bridge timbers, ties, floor beams or stringers still in fairly sound condition answer for such work just as well as new timber. On firm ground concrete is much used in footing courses, and on yielding- ground it is used in beds or in the form of an invert. Except in rock or hardpan the excavation for culvert foundations should extend at least & ft. below the surface. It is a common fault with culverts that the side walls do not extend deeply enough into the ground, the result of which is- that in freezing climates the frost heaves them up at the ends. The original and most common covering for stone box culverts is large flat stones. In practice the thickness of cover stones is independent of the hight of embankment, being about 12 ins. for openings of 3-ft. span, 15 ins. for 4-ft. spans and 2 ft. for 6-ft. spans, which is about the widest clear opening under stone covering in general practice, although in excep- tional cases there are openings as wide as 8, and even 10, ft. covered in. this manner. It is usually required that cover stones shall have a bear- ing upon the side walls of at least 12 to 15 ins. and be laid to close joints^ which should be filled and spread over with cement mortar, to form a tight covering, lest the filling material might become softened and ooze out through the openings should the culvert discharge under head. Late years old rails, laid across the opening and covered with concrete,, have been much used for culvert covering, as on most roads the material is- conveniently available and such a covering gives the maximum permissible headroom under a shallow embankment. Another advantage is that in. case of an unstable foundation the rail top adjusts itself better to settlement of the walls than is the case with stone covering or arches. The rail top is proportioned somewhat roughly to the load or amount of filling- material supported. This may be done in one way by spacing the rails, lay- ing them close together, with the flanges touching, under the central portion of the embankment, and spreading them apart under the sloping parts of the- embankment. For heavy loads the rails may be laid in a double course by inverting the top rails so that their heads hang downward between the rails of the lower course, which stand workwise. Over long-span openings it is customary to reinforce the rails with two or more I-beams of good strength laid among the rails directly under the track, old bridge girders or floor beams being suitable for such purpose. Openings up to 12 ft. and sometimes 14 ft. clear span are covered in this manner, while rails alone, laid in a single course, are considered sufficiently strong for openings up to 8 ft. span. For spans longer than 12 ft. two openings may be used with a common pier between them. On the Portland & Eumford Falls Ey. four- pairs of rails riveted together base to base, to form girders, are placed under each track rail to strengthen the covering for openings of 10-ft. span. Kails under the shallower parts of the embankment are usually spaced 6 to 1 ins. apart, according to the span of the opening, and the intervening spaces are planked or set with paving brick placed side by side upon the flanges- of the rails (endwise between the webs) to form a bottom for the concrete before it has set. The rails usually extend 12 to 18 ins. over the culvert CULVERTS 43' walls, according to the span. To protect the rails from rust they are usually given a coating of hot coal tar and filled in between the heads and over the- tops with cement mortar, which is then covered with concrete. On the Chi- cago, Burlington & Quincy Ey. rail tops for culverts are laid in 6-ft. sec- tions, separate from the walls, so that if it should become necessary to- remove the same in places it can be done without badly breaking up the culvert. The concrete covering for rail-top culverts is mad 3 4 to 1.8 ins. thick above the tops of the rails, according to the culvert span and hight of fill, and to provide for drainage the top surface is sloped either way from the center of the embankment. This increase in thickness of cover from the ends of the culvert toward the middle also accords with the increase of pressure from the embankment and provides some margin against a dished' top in case of excessive settlement under the central portion of the embank- ment. Weep holes are somtimes left to permit the escape of such water as may collect. Where the grade of a culvert is very steep, as on a rock slope, the covering of the culvert, instead of being laid parallel with the floor, may be stepped, so as to better retain the filling on top. At the ends of" the culvert an I-beam, channel iron, stone or concrete parapet is placed upon the rails to retain the foot of the embankment; slope. This parapet is backed up by the top step or top course of the wing wall, the first stone of which is doweled to the wall underneath to resist the pressure of the em- bank against the parapet. On the New York Central & Hudson Eiver R. R. rail tops are used on concrete or masonry culvert walls for spans of 4 to 14 ft., the larger open- ings that are covered in this way being under the higher banks. Old rails in weights of 60 to 100 Ibs. per yd., according to the length of span, are used, being first thoroughly cleaned and then painted with a coat of red lead, and oil and a second coat of bridge paint. The rails are set workwise, close together, and under each track rail the cover is reinforced with six inverted rails with the heads matched in between those of the bottom laj^er. The spaces between the rails are then filled with concrete made of finely broken stone or gravel, deposited in a layer 1^ ins. higher than the tops of " the rails at the center and -J in. higher at the sides. The concrete layer- is covered with a J-in coating of American straight run coal tar pitch. The edges of the covering are finished with a concrete curb 2 ft. wide and 1 ft. thick, and in shallow banks the space above the culvert, as far as sub- grade, is filled in with gravel. For spans of 16 and 21 ft. the Cleveland, Cincinnati, Chicago & St. Louis Ry. has used a covering that is known as "concrete girder" con- struction. It consists of two layers of old rails embedded as a reinforce- ment to a thick bed of concrete, one layer of rails being near the bottom and the other near the top. For a clear span of 16 ft. the covering is 2 ft., thick, and the two layers of rails are molded in 2 ins. from the bottom and top, the top layer standing workwise and the bottom layer inverted. Under the tracks the rails in each layer are spaced 9 ins. centers, but between this and the parapets the spacing increases to 11| ins. and 17 ins. To give the rails holding power in the concrete J-in. dowels 8 ins. long were inserted in holes drilled in the rails 12 ins. apart. In the construc- tion of a covering for a double-track structure, 70 rails of 60-lb. section, 20J ft. long, were used. For a double-track structure of 21 ft. clear span, a composite steel and concrete covering 2-J ft. thick is used. Under the tracks the rails are spaced at 9 ins. centers, as in the design for the 16-ft. span, but outside the tracks the spacing increases to 11^ ins., 17 ins., and 18 ins., centers. The rails in this case are 25 ft. long. These concrete- steel coverings, which have been applied to culvert or bridge openings on 44 TRACK FOUNDATION the St. Louis division, were constructed at the side of the track and rolled into place afterwards. These side walls are of concrete. Under deep embankments the walls of culverts may be proportioned to the load, with some economy of masonry, the thickness being gradually increased from the ends, where the load vanishes, to a maximum under the central portion, where the load is greatest. Means must also be provided to prevent the walls of culverts from being forced in by the pressure of the earth filling. On rock foundation the bottom of the wall may be secured by doweling, and in the case of other foundations the paving acts as a strut to prevent crowding of the walls at the bottom. It is cus- tomary, however, to build cross walls or concrete struts between the foot- ings, at intervals, to resist the side thrust against the walls. To secure the walls at the top it is usual to abut the rail top or stone covering against -a shoulder on the top face of each wall. In building culverts where the ground freezes to good depth in winter some attention should be directed to the conditions influenced by the qual- ity of filling material over the top of the culvert. If this filling is shallow it will freeze deeper or harder than other parts of the embankment, being exposed both top and bottom. In such cases the filling directly over the culvert should be made with broken stone, coarse gravel, slag or other material which does not heave when frozen. The range of depth to which the necessity for such filling material applies depends, of course, upon the severity of the winters, or upon the thickness of material that will be frozen entirely through, from top to bottom. Then, too, when not deeply covered up the culvert itself is liable to be disturbed by the action of the frost, if the filling material is retentive of moisture. Owing to the action of frost and the jarring effect of trains it is desirable, at least with large culverts, to have a good depth of filling over the top. It is sometimes necessary, however, to make it as shallow as 1J or 2 ft. Where the available headroom under the track is not sufficient for a single opening of the required area and of desired proportions the requirements may usually be fulfilled in the construction of a wider culvert of less hight, by partition- ing. In such event the movement of ice and flood trash and the matter of protecting the culvert against the same may have to be taken into account. End Construction, Paving etc. Culverts without end walls should extend from toe to toe of the embankment slopes. It improves the general appearance of things to have the end construction conform to the slope of the embankment. In the case of box culverts, of either timber or masonry, or arch culverts of short span, this is usually done by stepping the side walls beyond the parapet, which is placed about where the top of the culvert meets the embankment slope. With masonry culverts the walls are not usually stepped lower than 3 ft. above the floor, at the* end. The stepped portion of the wall should be coped with stones of good size, each step being formed by a single stone (Fig. 7 C), block rubble or roughly dressed stones being used for rubble masonry. Such end construction leaves the walls in convenient shape for the extension of the culvert should occa- sion arise in double-tracking the road or in the construction of side-track. In rare instances the sloping of stone culvert walls beyond the parapet is -done by laying the coping stones to the slope of the wall (Fig. 11) instead -of stepping them, while with concrete walls the sloped coping is found more frequently than the stepped coping. To increase the capacity of a -culvert built with straight side walls from end to end, as presently con- sidered, the face of the wall is sometimes splayed by gradually decreasing the thickness of that portion of the wall lying between the parapet and the CULVERTS 45 end of the culvert, the wall remaining straight on the back side. Where drift is bothersome stepped culvert walls which run straight beyond the parapet are not considered as safe as walls carried to full bight all the way to the end, for the reason that, should the opening become clogged the full hight at the end of the wall the water may still pour into the culvert through the open top between the parapet and the end of the wall. Culverts built to carry streams which overflow their banks at times -should be provided with end walls to protect the embankment from scour by the currents which converge to the opening. End walls for culverts are of two kinds: head walls, which stand at a right angle to the axis of the culvert (Fig. 12) ; and wing walls, which stand at an oblique -angle to said axis 15 to 45 deg., but usually about 30 deg. (Fig. 8). In either case the end wall usually serves to retain the embankment, as it usually stands on or starts from the meeting line between the top of the culvert and the embankment slope (or some little distance back of such a meeting line, de- pending upon the hight of the parapet). The coping of the wing wall is usually stepped to conform to the embankment slope. All coping stones should extend the full width of the wall. The earth which slopes past or around the flanks of a head wall is usually retained and protected by hand-placed riprap. To prevent undermining, head and wing walls should be carried down to good depth. For large culverts such walls usually stand at a batter. The junction between a wing wall and the body of a culvert should 'be on line with the face of the side wall; that is, at the point where the side wall of the culvert joins the wing wall there should be no re-entrant angle or projection of the side wall into the splayed opening between the wing walls, such as occurs in Fig. 8. The projecting corners or shoulders impede the flow of water at the entrance between the side walls, thereby diminishing the discharging capacity of the culvert, other condi- tions remaining the same, and they also form lodging places . for drift material. A plank or stick of timber floating against such a projection will swing around with the current and, if it be longer than the width of the culvert opening, will meet the opposite wall and become lodged across the channel and held there with great force by the pressure of the current. Danger of obstruction in this manner is greater with small culverts than with large ones, but such construction is inadvisable in any case, because the extra corners increase the cost of the masonry. One way of avoiding this projection is to build the side walls of the culvert to a batter, as in Figs. 8 A and 12 A. To avoid such projections with battered wing walls and plumb side walls it is necessary to set the wing wall inward far enough to bring it flush with the face of the side wall at the top of the culvert or springing line of the arch, and then form to a vertical face, on line with the side wall, that portion of the wing wall corner which would otherwise project into the waterway. For rapid streams of considerable volume wing walls are preferable to a head wall, on the up-stream end, as they increase the discharging capacity of the opening and facilitate the passage of drift material. The intermediate walls of partitioned culverts should be pointed or formed into a cutwater on the up-stream end to split the current and increase the flow at the entrance. On the down-stream end flared walls are not usually necessary. The floor of a culvert should, if practicable, be laid to a grade, a fall of at least 3 or 4 ins. per rod being desirable. The grade actually required is something more than the average fall of the stream in the vicinity of the culvert. If this is exceeded the increased velocity of the stream through the culvert will keep it clear of sediment and increase the discharging 46 TKACK FOUNDATION capacity. In deeply covered culverts the grade should be increased over' the lower third of the length so as to drain the middle portion in case the cul- vert sags, which is quite liable to happen, owing to the preponderance of earth pressure over the middle portion. This arrangement, in effect, gives the culvert a camber, with down grade all the way. If the culvert stands level it should be cambered a few inches to allow for sagging. Should set- tlement at the middle not take place to the extent anticipated, the camber that remains will do no harm, as it will only run the water to the ends in case the culvert goes dry, and possibly cause the culvert to silt up a little at the upper end, which is not so objectionable as to have the culvert sag and silt up nearly its whole length, and deepest at the middle. The sagging of culverts sometimes stretches the walls apart, but where the walls are cambered this is not liable to happen unless the settlement exceeds the amount of cambering put in. In order to prevent parting of the walls in this manner, in culverts where unusual settlement is -anticipated, resort is sometimes had to means for tying the walls together longitudinally. For this purpose old bridge rods, old chain, or old rails with drift bolts through the bolt holes, have been embedded in the concrete footings for the walls. Culvert bottoms are sometimes paved with flagstones, but more fre- quently with rubble stones set on edge crosswise the culvert. Such paving -should be at least 12 ins. deep and, for security, the interstices should be grouted with cement mortar. A bed of concrete 6 to 12 ins. thick is also -commonly used for culvert paving. Where timber grillage is used, as in submerged foundations (Fig. 8), it usually answers for the culvert floor without paving. Where stone paving is used it should be laid between the side walls, and not as a foundation for them, as in the latter case the walls are easily undermined if the paving becomes washed out. A very secure way of paving a culvert, sometimes resorted to where the current is strong, is to build masonry cross walls into the side walls, 6 to 10. ft. apart through- out the length of the culvert, These cross walls extend from the founda- tion up to the floor level and the paving is set in between them. At the outlet of a culvert, particularly where the water falls away rapidly, there should be a paved apron or spillway, extending some distance down stream, to prevent undermining of the foundation; and if the current is strong the bed of the stream for some distance approaching the up-stream end should also be paved. The end of culvert paving which extends beyond the opening should be protected against undermining by deeply set curb- stones. The paving at the outlet is usually the more important, as, in the absence of the same, the outflow of water is liable to scour out a hole and then begin to eddy back under the pavement or masonry, causing sec- tion after section to fall in, until the culvert is finally washed out or par- tially destroyed. To fortify against trouble in case the paving should l>e washed out the end wall or foundation at this end of the culvert should be run down to good depth or to solid bottom. To prevent culverts from being clogged with ice or driftwood a fender of small piles or old rails is sometimes driven across the stream a little distance above the culvert. The piles are usually spaced 12 to 18 ins. apart, in a semicircular row -extending up stream. By making the fender V-shaped up stream it will -carry the drift to the shores and maintain a clear channel in the middle of the stream. If old rails are used the driftwood can be burned during dry weather without destroying the fender. Pipe Culverts. For small culverts vitrified clay pipe and cast iron pipe are in extensive service. In shallow embankments, where there is not sufficient hight to build a masonry culvert, pipe is commonly used, while CULVERTS 47 on some roads, vitrified pipe is used under banks of considerable hight and cast iron pipe under banks of almost any hight. The chief considerations in the use of these materials are cheapness and the rapidity with which they can be laid. In filling open culverts the use of pipe affords a con- venient means of maintaining an opening, and iron pipe is extensively used inside old wooden culverts that are about to fail. Vitrified clay culvert pipe is used in sizes up to 36 ins. (and even 48 ins.) diameter, although many roads limit the maximum size to 24 ins. diameter and some roads have established 18 ins. diameter as the largest size. The quality used in railroad culverts is known as "extra thick" or "double strength"' pipe, being 25 to 33 per cent thicker than the clay pipe commonly used in sewers. The standard thickness of shell for railroad culvert pipe is one twelfth of the inside diameter of the pipe, while for sewer pipe the standard thickness is one fifteenth or one sixteenth of the inside diameter. The most common length of section is 2J ft., net, for all sizes, but 2-ft. and 3-ft. lengths are made. Experience teaches that culvert pipe, both vitrified and iron, but par- ticularly vitrified, must be very carefully laid if good results are to be expected. The pipe must be thoroughly bedded and given a uniform bear- ing. In firm earth the trench should be rounded out to fit the lower half oi the pipe, as well as may be, and the earth (or preferably sand) should be well tamped about the pipe up to the center line. The precaution should also be taken to excavate suitable depressions for the pipe sockets, so that no por- tion of the under surface of the pipe will have to sustain more than its due proportion of pressure. The too common way of laying the^ bottom of the pipe on solid bearing, with only a narrow segment of the bottom wall of the pipe supported, and then filling in loose earth carelessly about the pipe, causes almost the entire pressure to fall upon a comparatively small por- tion of the pipe surface, thus operating to crush the pipe. The same results follow from the use of boards or timber laid in the trench to .sup- port the pipe. It is usually recommended that on soft ground the bedding ior the pipe should be rammed, but the reliability of such work is question- able. On soft ground a platform of timber, or cross timbers rounded out to fit the bottom of the pipe, or a bed of broken stone, or a pile bent under or just ahead of each socket, or a brick or stone pier under each joint, is sometimes used for the immediate support, but is not generally approved of. A timber platform covered with 2 ft. of sand for the embedment of the pipe is a better arrangement, but a bed of concrete caried up half the diameter of the pipe is still better, and probably the best foundation for either vitrified or iron pipe laid on yielding ground. Without such a foundation the advisibility of using vitrified pipe on ground of the char- acter referred to is questionable. The pipe should be laid with the sockets up stream and the joints should be filled with cement mortar; on iome roads both cement mortar and oakum are used. If the joints are not made tight roots may enter and choke the pipe, or when discharging under head the pressure of water may wash out cavities around the joints, leaving the pipe unevenly supported and liable to break. The cementing of the joints also strengthens the pipe at the spigot end, which, being thin and not fitting the socket snugly, obtains a bearing on the bottom only, and is liable to spring under the pressure and crush if not firmly packed about. Some manufacturers of vitrified clay pipe groove or corrugate the inside of the socket and outside of the spigot end circumferentially, to give the cement a better bond in the joint. After a joint has been made the cement should be wiped clean from the inside of the pipe, else if allowed to harden in place it may remain a permanent obstruction. Defective pieces of pipe, 48 TRACK FOUNDATION if at all suitable for use, should be laid next the ends of the culvert, where the pressure is light and where they can be easily removed in case of failure. As the ground under the center of a high embankment is bound to- settle, the pipe should be cambered, if anything more than in the case of masonry culverts, because the foundation of a pipe culvert is seldom carried much below the surface. As previously explained, the best way to do this is to lay the center sections somewhat higher than a plane grade from one end to the other, or lay the upper half level and then drop off from the center to the lower end. The amount of camber to be used in any case is a matter of judgment, as it depends upon the hight of the bank and the bearing ~ properties of the top strata. Two to four per cent of the hight of the fill is sometimes allowed. The cambering of pipe culverts produces the desired effect of forcing the joints together, in case of settlement at the middle, whereas the sagging of pipe laid to an even grade stretches it apart, tending to disjoint the sections. The water which stands in a sagged pipe will cause the pipe to silt up to the same depth and reduce the area of opening by that much, and if water freezes in a virtified pipe standing more than half full the expansion of the ice. will burst the pipe. This consideration makes it undesirable to use vitrified pipe on low-lying ground or wherever water is liable to stand in the opening. Owing to the heavy expense of replacing vitrified clay pipe in case it becomes crushed, the use of the same under high embankments does hot meet with general approval, even among roads where such pipe is ex- tensively used under different conditions. Most roads limit the use of it to embankments not higher than 20 ft., and many roads to banks of still less hight, as is the case with the Pennsylvania and Chicago, Milwaukee & St. Paul roads, where 7 ft. is the maximum fill in which vitrified pipe is used. On the St. Louis Southwestern By. the limiting hights of embankment for the use of vitrified pipe are 4 ft. and 18 ft. On the Nashville, Chattanooga & St. Louis Ry. the limiting fill for 18-in. vitrified clay pipe of double strength is 2o ft. high; for 24-in. vitrified pipe it is 15 ft. high. This leads up to the question, frequently raised, as to the load sustained by pipes, arches and other structures which stand under earth filling. Although earth pressure on unit areas is much of an uncertainty, the only- reasonable assumption is that, with settled embankments, the load bear- ing upon any interior area is equivalent to the weight of a prism of the material having a base equal to the given area and length equal to the hight of material above said area. Before the embankment becomes set- tled the load may be more or less than this, according to the amount of friction set up between masses of settling material and the manner of distribution of the forces arising from such friction. Of these we can know nothing. Owing to the fact that a tunnel through an embankment will stand for some time without supports under the roof it has bem erroneously supposed by some that the weight upon a culvert ceases to in- crease after a certain hight of fill is reached, or, in other words, that the load upon a culvert is not proportional to the hight of fill above it. A little reflection will dispel the fallacy. According to the tunnel hypothesis there- should be no load at all upon the top of a culvert under a high fill. As a matter of fact we have no reason to suppose that the arching or beam -effect of the earth, as displayed over the roof of a tunnel, existed before the material was removed from the space occupied by the tunnel. If it did we might then suppose that some portions of the interior of an embank- ment stand under no pressure. But this cannot be, for once pressure is removed from the material at any point there is a gradual movement of CULVERTS 49 material toward the point relieved, as evinced by the bulging of the floor in deep coal mines and in tunnels through soft material. And then, if we were to assume that the distribution of earth pressure takes place through beams or arches of material,, why is it not just as reasonable to suppose that the culvert will stand under the end of span of such beam or arch as that it should stand between the end supports? A simple illustration may serve to clarify this matter. In a pile of "boards laid in crossed 'courses there is a beam over every board in the interior of the pile, and a large percentage of the boards in any course may be removed (if not consecutively) without perceptibly affecting^ the sta- bility of the pile; yet if one attempts to pull boards from the interior of the pile he will find them harder to pull as he proceeds towards the bot- tom, showing that the beam covering any board is not called into action as such until the board is removed. So far as my reading has extended those who hold to the claims here disputed have not propounded any argu- ment to show that the case is different with earth in an embankment. An impropped tunnel through earth stands more firmly under a high embank- ment than a low one because the earth at the bottom of the high embank- ment is more solidly compacted, and therefore better constituted to hold together in a large mass. The pressure due to live load (the trains) is transmitted through earth filling in diverging lines and approaches uni- formity with depth. It therefore acts with greatest concentration through shallow embankments and the crushing effect upon culvert pipe increases with nearness to the track. According to one rule, followed to some extent for vitrified pipe, 1^ times the diameter of the pipe is taken as the mini- mum allowable depth of fill over the pipe, while on some roads 3 ft., and on other roads 4 ft., is considered the least allowable depth for pipe of any diameter. Where the action of frost must be taken into account either j bb bb CULVERT o c S rt 'c"" O ' t~.S u rt c 1 P.O C ^ s ** v -SU*^ S ~ S*"" rt & * CS * g *ji! Sf, ,22 < o* '3 at < e 2 o 5 2 * 1 5 sfo u S 3d SJ g Clay pipe; i-ig in. 2-18 1.77 3-54 464-7 929 4 '8 06 $ 08 16 M? $ .305 $777 IA 8a $ 20 $ 2 28 $ 10 25 1-24 3 T 4 958 3 86 3 45 2-24 6.28 1916 6 I 72 9 43 3 3 30 4.91 1679 3 I 58 06 12 i yy I 76 ?6o 12 78 16 48 1.77 Iron pipe: 18 in. 464.7 Iron. 150 Ibs. $ I 24 $ 03 $ 18 $i 45 $ .82 $777 $ 20 $2 28 $ 10 25 24 3-H 958-3 208 i 72 06 24 2 02 .643 9 43 24 3 03 12 70 30 4.91 1679-3 300 2 4 8 07 36 285 58 12 78 40 3 3 16 48 36 7.07 2651.0 417 3 44 07 36 3 87 547 H 79 52 428 J 9 59 4 3 60 12-57 19.63 54i6.o 9444.0 750 1250 6 19 10 31 09 IO 48 60 6 76 II 01 20 80 36 37 I 22 2 60 6 72 14 40 28 74 7? 28.27 14816.0 1500 12 38 12 72 13 22 .468 3880 3 80 16 02 58 62 Brick bbl.: 36 in. 7.07 2651.0 132 brick $ 8 5 $ 09 $ 85 $i 79 $ -253 $22 II $ 66 $ro oo $ 32 77 48 Brick arch: 12-57 54i6.o 286 I 8 3 ii i 30 3 24 .258 45 92 i 54 20 56 68 02 36 in. 7.07 2651.0 250 I 60 10 i 33 3 03 .428 22 II 66 IO OO 32 77 48 54i6.o 372 238 16 I 8q 4 43 352 42 68 I 54 18 16 6238 60 19-63 9444.0 560 358 22 2 49 6 29 .320 62 81 2 20 28 50 93 5i 72 28.27 14816.0 840 32 3 30 9 oo .318 73 65 4 20 33 15 III OO Stone box: Mason 'y 2x4 ft 8.00 1684.0 i. 33yds $ 3 36 $ II *252 $599 $ -75 $56 45 $ I 92 $42 43 $ 100 80 3x4 12.00 3903.0 1.463 369 13 2 77 6 59 55 60 98 2 20 45 59 108 77 4x5 20.OO 7920.0 1.77 4 45 18 3 34 7 97 7888 4 oo 57 93 140 8 1 4x6 24.00 10145.0 2.065 5 25 23 3 81 9 29 .388 108 36 6 oo 79 05 193 4i DATA USED IN COMPUTING ABOVE COSTS. CLAY PIPE. IRON PIPE. BRICK CULVERTS. STONE CULVERTS. 18 in. Pipe $1.07 per ) 2^ ft. joint/ i8inatiSooft>s{P t e y t 2 Brick 2^x4x8 at $4.50 perM. per cu yd. Stone $2.00 24 " " $2.15 per f 2V 2 ft joint f 24 " 2500 " ' Brick perc ft. 16.4 allowing % in. joint Cement ^bbl .45 Building Parapets $3 ) per M. brick f 30 " 3600 " Mortar, i part Cement 2 parts sand Sand i bbl. 07 Coping Stone 27 and \ JIG per Hn. ft. j 36 " 5000 " " Cement .81 bbls. per cu yard, or: Labor 1.98 Apron Stone 25C per ) liu. foot) 48 " 9000 " " Cement 1.83 bbls per M b-ick Total $4.50 60 ' 15000 " " Coping Stone 3ic per lin. foot 72 " 18000 " " Apron Stone $1.50 per All at $16.50 per 2000 Ibs cubic yard NOTE. All prices F. O. B. Nashville, February. 1898. 58 TRACK FOUNDATION sary to increase the bight of these culverts for any purpose other than that .of increasing the area of the culvert opening, as, for instance, to allow for the passage of cattle, straight side walls 2 ft. high are put in between the upper and lower semicircles of the culvert and the haunch walls are carried up in proportion. Wherever the nature of the founda- tion will permit, brick arches with straight side walls and a compara- tively flat invert are built. A particular culvert of this class (Fig. 70) under a 60 ft. embankment has a semicircular arch of 7 ft. span, straight side walls 3-J ft. high, and an invert of 7 ft. radius. The foundation consists of five rows of piles driven about 3 ft. apart in each row, around the tops of which is placed a bed of concrete 2 ft. deep. Haunch walls are carried "up to a hight of 7 ft. 9 ins. above the foundation and then sloped off rapidly to meet the top of the arch on tangent. The width over the haunch walls at the foundation is 14 ft. and the batter on the back sides of these walls about 1 in 10, making the top width 12 ft. 5 ins. The plain barrel culverts are used only on hard, firm ground, where the natural surface is above the axis of the culvert and where the hight of the earth above the culvert does not exceed 30 ft. Haunch- wall circular culverts are used wherever the brick barrel will not answer, Fig. 7C. Brick Arch Culvert, N., C. & St. L. Ry. and an extra ring of brick is added to the arch for every 15 ft. of fill over 20 ft. high. On solid rock bottom semicircular arches with straight side walls are used. The minimum grade of the culvert floor is -J per cent. In the haunch-wall circular culverts the barrel is not made as strong as in the plain barrel culverts, as in the former only two rings of brick are used up to and including culverts of 5 ft. diameter. The culvert of 6 ft. diameter of this class has 3 rings of brick. It will be noticed that the culvert shown in Fig. 70 has vertical wing walls and that there are no re-entrant angles at the end of the arched opening. Arch Culverts. Well constructed stone arches are considered the highest class of masonry for culverts and bridges, and on many of the best built roads such is the standard construction. Particularly is this the case on the Pennsylvania R. R., where many fine examples of heavy arch construction are to be found. It cannot be expected to here go into the subject of arch construction and masonry specifications com- prehensively, but some ruling principles may be considered. In railway work semicircular and segmental arches predominate, with an increas- ing preference for the segmental reach. This for the reasons that the seg- mental arch permits of a wider opening where the depth of embankment is a CULVERTS 59 limiting feature (Fig. 7D, for example), and the amount of sheeting in the semicircular arch is the greater and therefore the more expensive, particularly in cut-stone work. First-class work calls, of course, for dressed sheeting stones, but roughly dressed and rubble stones are very commonly used in arches of short span, having quarry-faced stones with chisel draught edge lines for ring stones. The Canadian Pacific Ey. builds rubble masonry arches as large as 60 ft. span, the only cut stone used being in the ring courses or those which show at the ends of the arch. This kind of work laid in Portland cement mortar has cost -about $6 per cubic yard, and, owing to the long distance -over which cement must be hauled, is considered cheaper than concrete. For abut- ments and wing walls rock-faced ashlar masonry is quite frequently found in high-class work, while range work and broken ashlar are very common. In the smaller culverts rubble masonry throughout is quite general. Figure 8, showing the plans of a stone arch culvert of 20-ft. span located near Watervliet, Mich., on the Chicago & West Micigan division of the Pere Marquette R. R., represents a good example of durable con- Fig. 7 D. Flat Arch Culvert Construction. struct ion for openings of this size. The arch is 29 ft. long, and the total length of the structure, from end to end of wing walls, is 76^ ft. The arch is segmental, with moderate rise to span, the arc or central angle being 139 deg. 58 min., the rise 7 ft. and the radius at the intrados 10 ft. 7j ins. The arch sheeting is 2 ft thick and the spandrel walls 2J ft. thick and 3 ft. llf ins. high at the crown. The filling over the arch -crown, or the distance from the crown to the base of rail is' 9 ft. 4 ins. The i\ hutment walls of the arch are 9 ft. 9 ins. high to the springing line. The excavation was carreid 4J ft. below the surface of the water and the foundation consists of timbers placed at 3 ft. centers and over- laid with two crossed courses of 3-in. plank. The wing walls are 24| ft. Jonu', and open out at an angle of 30 deg. with the center line of the arch. The footing course of the wing wall where it joins the arch abut- ment is 8 ft. wide, and at the base of the battered portion the wall is