Fore Eaatews cs papladeuseys ercoshybebccebat tat Bis snail) aii 3 te atts ee saan a Ser = cern i * HA Re a SSS SSS SSS Marthe Biological Laboratory Aug. 14, 1950 Received 64471 Accession No. < John Wiley and Sons, Inc, New-York City —— Given By Place, ( SS SSS SISSIES SSIS SSS ¢ bYTEcOO TOEO O A lOHM/T81N © APPLIED SEDIMENTATION COMMITTEE MEMBERS Ro.anp F. Beers Gait A. HatHaway Rosert F. Brack J. W. Jounson Cart B. Brown SHEPARD W. Lowman ArTHuR B. CLEAVES H. V. PETEeRson Frank C. FoLey Rocer RuHoapDEs Ratpu FE. Grim Kari TERZAGHI Frank C. Wuirmore, Jr. Parker D. Trask, Chairman Ayes LE 1D) SEDIMENTATION Edited by PARKER D. TRASK Supervising Geologist Division of San Francisco Bay Toll Crossings Department of Public Works State of California San Francisco, California Prepared under direction of Committee on Symposium on Sedimentation Division of Geology and Geography National Research Council Washington, D. C. New York - JOHN WILEY & SONS, Jue. London - CHAPMAN & HALL, Lid. 1950 () CopyricHT, 1950 BY JoHN Witey & Sons, INc. All Rights Reserved This book or any part thereof must not be reproduced in any form without the written permission of the publisher. PRINTED IN THE UNITED STATES OF AMERICA PREFACE This symposium considers the practical applications of sedimenta- tion. It is designed (1) to describe aspects of mutual interest to the geologist and to the engineer so that each can understand the other’s problems and thus cooperate more effectively in their work; (2) pro- vide information for the consulting geologist who may not be com- pletely familiar with specific problems; and (3) acquaint students with the many practical applications of sedimentation so that they may be more fully informed as to possibilities for a career in this field. Each chapter is a summary of a comprehensive subject. Bibliographies are given for benefit of readers who may wish to pursue topics further. Key articles are characterized by an asterisk in the lists of references. The work has been prepared under the sponsorship of the Committee on Symposium on Sedimentation of the Division of Geology and Ge- ography of the National Research Council. The Committee wishes to express its appreciation to Dr. Arthur Bevan, Chairman of the Division, and to Miss M. L. Johnson, Secretary of the Division, for their help in the preparation of the symposium; and to Mr. G. D. Meid, Business Manager of the National Academy of Sciences, and Dr. David M. Delo, Executive Secretary of the Division, for aid in arranging publication. Special credit is due Dr. Arthur B. Cleaves, Professor of Engineering Geology, Washington University, St. Louis, Missouri, for assistance in organizing chapters on engineering topics in Part 2 of the symposium. Acknowledgment is also due the University of Wisconsin Alumni Research Foundation for a grant that helped me in early work on the symposium. The practical applications of sedimentation are so large in number and so complicated in scope that the Committee felt that it would accomplish more good by inviting a group of specialists to write chapters in their own field of interest rather than by trying to prepare a hand- book itself, a handbook which in places would be only a compilation. As it has not been practicable to procure articles in all fields, the cover- age of the subject is not complete. A sufficient number of articles, however, is included to indicate the general usefulness of sedimenta- tion in practical endeavor. The geologist is constantly called upon to indicate the character of sediments at some place beneath the surface of the earth where he cannot see them either in outcrop or in samples from drill holes. He Vv vi PREFACE has to predict what they will be like by means of interpreting what he has seen at some other place. He is well aware of the variability of sediments, and on the basis of his past experience and his knowledge of the mode of origin of sediments he is the person best qualified to make predictions. A knowledge of the basic principles of sedimenta- tion and stratigraphy is essential for this purpose. The geologist, how- ever, also must understand the things the engineer wishes to do with the sediment; and the engineer should understand the general geologic problems involved and the limitations of geologic work, so that he can utilize the services of the geologist intelligently. The engineer above all should be conscious of the variability of sediments. It is hoped that this symposium will help accomplish these purposes. Three types of earth materials are considered in this symposium: (1) recent or slightly consolidated sediments; (2) ancient or maturely consolidated sediments, that is, sedimentary rocks formed in the ge- ologic past; and (3) residual soils, or soils weathered in place. Soil technically perhaps is not a sediment, but as it has many of the char- acteristics of soft sediments, particularly with respect to its effect upon engineering problems, it is considered here. A sediment is an aggregate of solid particles that have been moved from one or more places of origin to some place of rest. Its properties depend on (1) the chemical, physical, and biologic nature of the solid constituents, (2) the character of the material that fills the pore spaces, and (3) the changes that take place in the sediment with time. due either to applied stress or to chemical reactions. The practical applications of sedimentation therefore involve both the nature of the sediment and the processes that affect them. The symposium contains thirty-five articles grouped under seven topics: (1) basic principles of sedimentation, (2) engineering problems involving strength of sediments, (3) applications of processes of sedi- mentation, (4) applications involving nature of constituents, (5) eco- nomic mineral deposits, (6) petroleum geology problems, and (7) military applications. The first topic on principles is essential for the proper understanding of the economic applications of sedimentation. The second and third topics, relating respectively to strength of sediments and the processes of sedimentation, are of special interest to engineers. The engineer must have adequate foundations for his structures. Frequently he is called upon to build roads, airports, dams, tunnels, or buildings in areas of relatively soft rocks, where he has little choice as to location. He therefore must do the best he can with what foundation material PREFACE Vii he has, and he is vitally concerned about the strength of the sediments or their ability to support the designed load. The processes of sedimentation, especially processes relating to the transport of constituent particles of sediments, have tremendous im- pact upon our national economy. Millions of dollars each year are lost by floods, beach erosion, shoaling of harbors, silting of reservoirs and canals, soil erosion, and landslides. Engineers and geologists are constantly called upon to cope with these problems. The nature of the sediments, discussed in Part 4, also interests the engineer. This topic includes chapters on aggregate for concrete, sand for foundry molds, and clay minerals. The last topic is of particular concern to the geologist as well as to the engineer, because of the materially different properties of the various clay minerals and the different effects of base exchange upon clay. However, the economic utilization of the constituent materials of sediments lies mainly outside the field of engineering. Both the min- ing and petroleum industries depend on the products of sedimentation for their livelihood. As McKelvey shows in this symposium, the total value of economic mineral materials, exclusive of coal and oil, obtained annually from sediments in the United States is almost one billion dollars. A great many kinds of raw mineral materials are used com- mercially, but space permits the discussion of only a few token ex- amples in this symposium. These chapters, presented in Part 5, should indicate to mining geologists the usefulness of knowledge of sedimen- tation in their work. More than 6,000 geologists are directly or indirectly engaged in the search for petroleum, which is found almost exclusively in sedimentary rocks. In their training and in their professional work, geologists thus become familiar with the applications of sedimentation to petroleum geology. Rather than duplicate material of more or less common knowledge to geologists, the Committee feels that it is preferable to emphasize the less well-known applications of sedimentation and to restrict chapters on petroleum geology (in Part 6) to a few topics of current interest. Geology is of material assistance to the Army and the Navy during time of war. Many of the geologic problems deal with sedimentation. Two chapters on this subject are presented in Part 7. The usefulness of sedimentation, however, is greater than these chapters indicate, as work of a confidential nature is not included. San Francisco, California Parker D. Trask February, 1950 * HT Latieh Mt ie an ibs ‘ CONTENTS PART 1. BASIC PRINCIPLES OF SEDIMENTATION 1. DYNAMICS OF SEDIMENTATION 3 Parker D. Trask 2. ORIGIN OF SOILS 4] Hans Jenny 3. THe LAwsS oF SEDIMENT TRANSPORTATION 62 H. A. Einstein and J. W. Johnson 4. GEOPHYSICAL PROBLEMS IN APPLIED SEDIMENTATION 72 Roland F. Beers 5. PRINCIPLES OF Som. MECHANICS AS VIEWED BY A GEOLOGIST 93 Clifford A. Kaye 6. SEDIMENTATION AND GROUND WATER 113 Frank C. Foley PART 2. ENGINEERING PROBLEMS INVOLVING STRENGTH OF SEDIMENTS 7. SEDIMENTATION AND HigHwAy ENGINEERING 127 Arthur B. Cleaves 8. FouNDATION PROBLEMS OF SEDIMENTARY Rocks 147 Shaler S. Philbrick 9. FouUNDATIONS FoR HigHway BRIDGES AND SEPARATION STRUCTURES ON UNCONSOLIDATED SEDIMENT - 169 C. H. Harned 10. EartH Dams 181 Thomas A. Middlebrooks 11. Grotocic Aspects oF Sort-GrouND TUNNELING 193 Karl Terzaghi 12. SEDIMENTARY GEOLOGY OF THE ALLUVIAL VALLEY OF THE Lower MuissiIs- stippr RiveR AND Its INFLUENCE ON FouUNDATION PROBLEMS 210 W. J. Turnbull, HE. L. Krinitzsky, and S. J. Johnson PART 3. APPLICATIONS OF PROCESSES OF SEDIMENTATION 18. RELATION oF LANDSLIDES TO SEDIMENTARY FEATURES 229 D. J. Varnes 14. PERMAFROST 247 Robert F. Black 15. GEoLocy In SHORE-ConTROL PROBLEMS 276 Martin A. Mason 64471 x CONTENTS 16. SEDIMENTATION IN HARBORS 291 Joseph M. Caldwell 17. ConTRIBUTION OF MopEL ANALYSIS TO THE SOLUTION OF SHOALING PRos- LEMS 300 C. B. Patterson and H. B. Simmons 18. StREAM-CHANNEL CONTROL 319 Stafford C. Happ 19. Desris ConTROL 336 Burnham H. Dodge 20. SEDIMENTATION IN RESERVOIRS 347 Albert 8S. Fry 21. PRoBLEMS oF IRRIGATION CANALS 364 Alfred R. Golzé 22. EFFECTS OF SOIL CONSERVATION 380 Carl B. Brown 23. Tue ProBLeM oF GULLYING IN WESTERN VALLEYS 407 H. V. Peterson PART 4. APPLICATIONS INVOLVING NATURE OF CONSTITUENTS 24. INFLUENCE OF SEDIMENTATION ON CONCRETE AGGREGATE 437 Roger Rhoades 25. APPLICATION OF STUDIES OF THE COMPOSITION OF CLAYS IN THE FIELD oF CERAMICS 464 Ralph E. Grum 26. Founpry SANDS 475 H. Ries PART 5. ECONOMIC MINERAL DEPOSITS 27. THe Fietp or Economic GroLocy oF SEDIMENTARY MINERAL Deposits 485 V. HE. McKelvey 28. SEDIMENTARY IRON DEPOSITS 506 Stanley A. Tyler 29. SEDIMENTARY Rocks As Hosts For Ore Deposits 524 John S. Brown 30. GEOCHEMICAL PROSPECTING FOR ORES 537 H. E. Hawkes PART 6. PETROLEUM GEOLOGY PROBLEMS 31. SUBSURFACE TECHNIQUES 559 Damel A. Busch 32. Porosity, PERMEABILITY, AND CAPILLARY PROPERTIES OF PETROLEUM RESERVOIRS 579 Charles D. Russell and Parke A. Dickey 33. CARBONATE Porosity AND PERMEABILITY 616 W.C. Imbt CONTENTS PART 7. MILITARY APPLICATIONS 34. SEDIMENTARY MATERIALS IN Miirary GEOLOGY Frank C. Whitmore, Jr. 35. APPLICATIONS OF SEDIMENTATION TO NAVAL PROBLEMS R. Dana Russell xX] 635 656 t SC i i ‘ wi i p ‘ 1 a ea bes oo 5 i. a iy i e beanie? be er ah i et ; f i Matats t } — ~ 6 Part 1 BASIC PRINCIPLES OF SEDIMENTATION rhage a i CuaprTer 1 Parker D. Trask Supervising Geologist Division of San Francisco Bay Toll Crossings California Department of Public Works San Francisco, California Geologists and engineers need to understand sediments. They should know what they are like, why they are the way they are, and what factors cause them to change. The object of this paper is to sum- marize the principal features of sediments, emphasizing the processes that lead to their formation or that cause them to change once they have been formed. As sediments are largely a product of the environ- ment in which they are formed, special attention is given to environ- ments of deposition. This chapter, like all other chapters in this sym- posium, is a condensation of a complicated subject. A list of 100 refer- ences is included for benefit of persons desiring further information on particular topics. Sediments are aggregates of particles that come to rest in some place after having been transported laterally or vertically for some distance. When first deposited, the particles are unconsolidated or essentially unconsolidated. The geologist calls such deposits recent sediments. With time the sediments consolidate and harden into rock. Such con- solidated sediments are called ancient sediments. During this process, cementing material is deposited in the pore spaces; new minerals form or old minerals grow, thus binding particles to one another; and pres- sure of overburden compresses the sediment, causing the constituent particles to interlock with one another. When rocks weather in place they are altered physically and chemically and eventually are trans- formed into soil. As soils or weathered rocks have many of the proper- ties of recent or unconsolidated sediments, it is convenient to consider them in connection with sediments. Technically they could be called * Prepared with aid of a grant from the University of Wisconsin Alumni Re- search Foundation. 3 4 TRASK. PRINCIPLES OF SEDIMENTATION [Cx. 1 sediments, because their constituents have settled under the influence of gravity, albeit only a very small distance. For more detailed ac- counts of sedimentation see the works of Gilbert (1890), Walther (1894), Grabau (1913), Twenhofel (1932, 1939), Boswell (1933), Hatch, Rastall, and Black (1938), Trask (1939), Shrock (1948), and Pettijohn (1949). Sedimentation consists of five fundamental processes: (1) weather- ing, (2) erosion, (3) transportation, (4) deposition, and (5) diagenesis or consolidation into rock. WHATHERING Most of the constituent particles of sediments are derived from rocks or earth materials that have been more or less weathered. The original constituents of sediments, of course, can come from unweathered rock, but weathering and concomitant solution are the dominant processes by which rock or soil constituents are transformed into a state whereby they can be eroded and transported to some place of deposition. Weathering is essentially a process of soil formation; for details the reader should consult standard references on the subject such as: Baver (1948), Clarke (1924), Goldich (1938), Goldschmidt (1937), Jenny (1941), Leith and Mead (1915), Polynov (1937), or Reiche (1945). Weathering consists of two fundamental processes: (1) mechanical disintegration and (2) chemical solution. Living organisms, particu- larly microorganisms, also play an important role, but in the last analysis organic action is essentially a question of physical disruption or chemical solution of rock and earth constituents. Five fundamental factors, as Jenny has pointed out, influence weath- ering: (1) parent rock material; (2) climate, particularly the tempera- ture and rainfall; (3) physical environment in which the weathering takes place, especially topography or shape of the land surface; (4) length of time the processes operate; and (5) action of organisms. To this might possibly be added a sixth factor—the dynamics of the environment, that is, the ratio of the rate of weathering to rate of re- moval of weathered particles by erosion. If erosion is rapid, soil-form- ing processes cannot proceed very far toward maturity before the con- stituents are removed. Moreover, these principal variables that affect weathering and soils to a considerable extent are mutually inter- related. Mechanical weathering is largely a question of thermal expansion and strength of the mineral constituents. Mechanical disintegration Cu. 1] EROSION 5 is dominant in arid areas and in cold regions; chemical weathering prevails in humid, tropic areas. Steep slopes facilitate mechanical erosion; flat areas favor accumulation of water in the rocks and soil and thus help chemical weathering. Chemical weathering is essentially a question of exchange and move- ment of ions. Base (cation) exchange is a dominant process. Critical factors are (1) concentration of ions, particularly hydrogen, sodium, calcium, and magnesium; (2) oxidation-reduction potential; and (3) temperature. Basic rocks, that is, rocks rich in calcium and magne- sium and relatively low in silica, tend to be more susceptible to chemi- cal weathering than acid rocks, which are rich in silica and alkalies, but, because the hydrogen-ion concentration of the soil is an im- portant factor, this generalization is subject to exceptions. Many sedimentary rocks weather more rapidly than igneous rocks because of their greater permeability. Moisture and temperature affect the growth of microorganisms, which in turn influence chemical weather- ing. The longer the time that climate, topography, and organic activ- ity remain essentially unchanged for given conditions of runoff, the more thoroughly the rocks are weathered into soil. When weathering has proceeded to the state where soil has formed, the earth materials are divided into four distinct zones from the sur- face downward: (1) an upper leached zone relatively rich in organic matter, the A zone; (2) an underlying zone in which some of the ma- terials transported or leached from the upper zone have been deposited, the B zone; (3) a still lower zone of partially weathered rock, the C zone; and (4) essentially unweathered rock. If weathering proceeds for a long time, the soil approaches matur- ity, which results in the formation of clay minerals typical of the environment and in the production of many fine particles. The con- stituents of such soils, when transported to a place of deposition, form a sediment having a far different reaction to imposed loads than sedi- ments whose constituents are derived from areas in which the rocks have not developed a chemically mature soil. Students of soil mechan- ics might well consider this point in interpreting the strength of foun- dation materials. EROSION Both mechanical and chemical processes cause erosion. Because of the close relationship between erosion, transportation, and deposition, it 1s convenient to discuss the mechanical aspects of these subjects in sequence and then take up chemical weathering and solution. Mechanical erosion consists of two processes: (1) plucking or fore- 6 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 ing particles from their position of rest, and (2) abrading rocks by impact of moving particles. The size of the particle, the velocity of the transporting agent, and the cohesiveness of the rock or soil are factors in the plucking action. Alternate freezing and thawing and the growth of plants in cracks are agents in the forcing apart of rocks. The size, quantity, hardness, and velocity of the moving particles and of the object that is struck are factors in the abrading action. Water is the chief mechanical agent of erosion, but wind, ice, grav- ity, voleanic explosions, plants, and animals also act. Water acts in at least three different ways: (1) when running off the surface of the land before it collects in streams; (2) in streams; and (3) in bodies of essentially standing water, namely lakes and the ocean. Erosion is intimately associated with the ability of the water to transport ma- terial. If the water is fully loaded and is transporting all the material of a given size it can transport, relatively few particles of that size can be picked up; or, to be more precise, the net erosion of such particles is small. However, when the water is underloaded, erosion is effective. The erosive power, that is, the ability of water to pick up particles, depends upon the energy relationships of the moving water, particularly the turbulence. Turbulence is discussed in the next section. Water moving over relatively steep surfaces of land before it reaches streams acts In some ways as a thin, flat sheet (Paige, 1912). The thickness and continuity of the sheet depend upon the average slope of the land, the amount of rain that has fallen, the rate of infiltration of water into the soil, and the interference of grass, plants, and other obstacles with its free movement. If the obstacles to downward flow of the water are numerous, relatively little material is eroded, as, for example, in humid regions where plants are plentiful or in areas of highly pervious soil which absorbs the water readily. Sheet erosion is best developed in arid or semi-arid regions, and it results in a topog- raphy that comprises relatively steep slopes above and more gentle slopes below, where the water spreads out over a fan-shaped area be- fore collecting in a stream or some body of standing water. In humid areas the fan-shaped areas usually are less well developed. The steep slopes above seem to be related to the rate of weathering of the rocks and the permeability of the soil. The flatter slopes below represent an equilibrium profile determined by the average load of sediment that the water is carrying. If the water is underloaded it will cut; if overloaded it will deposit. The same fundamental prin- ciples of transport and erosion apply to this water moving in a sheet as to water flowing in streams or in the ocean, but the generally shal- Cu. 1] EROSION i low depth with respect to width results in a different profile of erosion. The principal erosive effect in streams is scour of the banks and the bottom. Erosion in the sea and in lakes is due principally to waves and currents, and to a minor extent to chemical solution. The abra- sion of rocks along the shore or on the bottom in shallow water is caused chiefly by the impact of sand and rock particles carried by mov- ing water. Loose or slightly consolidated sediment may be scoured by wave action or by ocean currents. Wind erosion is similar to water erosion. All kinds of earth materials are worn away by impact of particles blown against them by the wind. Constituent particles of loosely consolidated sediments are also plucked as wind blows over them, provided that the wind is strong enough and the sediments are sufficiently dry (Bagnold, 1941). Ice flows plastically, much like a viscous body of tar. The essential factors affecting its movement are: the mass of ice; the gradient, par- ticularly the surface configuration over which it flows; and the distri- bution of temperature throughout the ice (Thwaites, 1941). The rate of movement of ice and hence the rate of erosion are accelerated by increased gradient of land and addition of snow to the upper parts of the mass of moving ice. As the plasticity of most solids increases with increase in temperature, it could be argued that ice would flow faster if its average temperature was relatively near the freezing point of water. Ice acts as an agent of erosion in the following ways: Rock frag- ments frozen in the ice abrade the rock or soil over which the ice flows. Ice plucks rock fragments as it passes over them. It also pushes masses of rock and earth ahead of it as it moves. Rocks are disrupted by alternate expansion and contraction as the temperature of the ice varies or as the ice melts and refreezes in cracks or crevices. Ice is reported (Blackwelder, 1940) to increase materially in hardness as the temperature decreases, and therefore ice alone, without the aid of rock fragments, can scour soft sediments and rocks. Voleanic explosions can disrupt rock fragments as the explosive gases and masses of lava and rock are foreed upward by the explosion, but the erosion of material in this way is of minor import, as active voleanoes are relatively few in number. Where slopes are steep or smooth, gravity can be an agent of erosion; but generally it is only a secondary agent, because the stability of the rock fragments must first be upset by some other agency so that frag- ments can become unbalanced and thus slide or roll down the slope. Living matter is an important factor in erosion, but its principal effect is indirect, because the living material influences the erosive 8 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 action of some other agent. Roots, however, are direct agents of ero- sion when they grow in cracks and force a part of the rock to fall away. The most profound effect of living material, however, is the action of vegetation in inhibiting the flow of rainwash. Man, because of his various uses of land and the resulting increase in soil erosion, is also an indirect agent. (See Chapters 22 and 23.) TRANSPORTATION The same six agents of erosion, namely, water, wind, ice, voleanic explosions, gravity, and biologic activity, also control transportation but in a varying degree. With water, the principal factors that affect transport are: turbulence; ratio of settling velocity to lateral motion of water (currents); shape, size, density, and quantity of particles; and movement along the bottom by saltation (jumping), rolling, or undermining. (Rubey, 1938; R. D. Russell, 1939; Rittenhouse, 1944). Turbulence is non-linear motion of masses of water as the water moves. (See Chapter 3.) Turbulence is characterized by vertical and horizontal eddies which tend to keep particles in suspension. The principal factors that affect transport by turbulence are the quantity, velocity, and temperature of the water that is moving, the load, and the shape or roughness of the surface over which the body of water moves. Thus gradient of stream bed, volume of water, and presence of objects that impede flow are major factors. The quantity and shape of ma- terial already in suspension also influence the amount of additional material that can be transported, because turbulence is related to energy and a given discharge moving over a given slope can transport in suspension only a given amount of material of a given size. Debris also is transported along the bottom of a stream. Turbulence may lift particles for a short distance above the bottom, but it may not be strong enough to maintain them in suspension. Consequently they fall to the bottom, only to be lifted again at a later time when turbulence once more reaches a sufficient magnitude. Thus particles jump or roll along the bed in pulses. This process is called saltation. In addition many particles are round and act like ball bearings, causing heavier rock fragments that rest upon them to glide downstream. This process is particularly noticeable in the alluvial fans of the desert. In the sea or in a lake many detrital particles are carried a long distance after they enter the sea or lake. Even though turbulence may not be strong enough to maintain the particles in suspension, the mass of water in which the particles he may move. The particles thus are transported laterally while they settle downward through the mass Cu. 1] WATER 9 of moving water. Transportation of this type is a statistical phenome- non. For given size distribution of sediment and for given velocity and depth of water, the average size of particles that are transported decreases progressively with distance of transport, but the range in size of particles found in a mass of water under any given conditions varies about a mean grain size, and the average deviation in size of particle from that mean grain size varies according to the conditions of trans- port. The subject of transport in the sea is still not well understood, but it is of such vital importance that it is now being attacked actively by oceanographic institutions. Transport of sediment by wind follows essentially the same laws as does transport by water, except that, because of its low density com- pared with water, air is a less effective agent of transport (Bagnold, 1941; Loess symposium, 1945). However, particles are also blown through the air or moved along the ground by the same principle of saltation that applies to water. Ice transports material either within or on top of the ice or by shoving masses of earth and rock ahead of it as it moves. Transporta- tion by volcanism is a question of the force of the explosion and the direction and strength of the prevailing wind. Fine ash can be ear- ried many miles, as the Katmai explosion in Alaska indicated. The distance that objects will be transported under the influence of gravity is governed by the angle and smoothness of the slope along which the material moves, the coefficient of friction, and the amount (mass) of material that moves. Organisms are minor agents of transportation. Even man moves relatively little material compared with nature. DEPOSITION The most important factor in sedimentation is deposition. Once the sediments have accumulated in their final resting place, their general nature is fairly well formulated. Subsequent changes, for most sedi- ments at least, alter their characteristics relatively little. The proc- esses of deposition are complex and the products are manifold. The same six agents that affect erosion and transportation likewise control deposition. WATER The processes of deposition in water are essentially a question of energy, place, and time. If the energy available to move constituents that are in the act of being transported decreases, some of the particles can no longer be transported and they come to rest. The environment 10 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 in which the particles are being transported profoundly affects the types of sediment that are deposited, primarily because of the effect of the configuration of the basin of deposition upon the movement of the transporting agent, be it water, air, or ice. Diastrophism, or move- ment of the crust of the earth, is a critical factor, for diastrophism not only influences the shape of the environment of deposition, but it also affects the configuration of the land which supplies the debris that is deposited. Time is a factor because the rate of movement of particles determines whether the particles will be picked up, moved, or de- posited. Furthermore, sediments, once they have been deposited, change in the course of time (diagenesis). The three factors—energy, place, and time—are intimately inter- related. The principal source of energy comes from streams, currents, wind, voleanic explosions, or from material that has acquired potential energy by having been placed in a higher position than it once was, by evaporation, convection, turbulence, or diastrophism. ‘The effect of environment is treated in a later section. The influence of time, particularly the rate of change, is complicated. The principal effect of time, especially with respect to deposition by water, is the rate of change in velocity, ds/dt. To this also should be added the change in direction and the rate of change in direction. In other words, deposi- tion is strongly influenced by turbulence when the interval of time is short and by changes in average velocity when the time is not short. Turbulence affects the size of particle that is laid down, that is, the size distribution of the sediments (grading curve). If the average ve- locity of the water is fairly constant and fluctuations in turbulence are not too great, well-sorted sediments are deposited. Particles small enough to stay in suspension for a significant interval of time are trans- ported away from the locus of deposition, and the particles that settle to the bottom are comparatively well graded in size. Consequently a relatively large proportion of the particles deviate in size slightly from the average particle deposited. The energy relations that govern transport are too complicated to discuss here, but, even though the average deviation in velocity of water is proportionately the same with respect to the average velocity of the water, the degree of sorting of the constituents (uniformity coefficient of the engineers) is not the same. Slowly moving water seemingly leads to poorer sorting than fast-moving water. Silts are rarely as well sorted as sands. The size distribution of the constituent particles, because of its de- pendence upon the motion of the water in which the sediments accumu- late, therefore is an index of the mode of deposition. If one process is operating, the size distribution is more likely to be symmetrical than Cu. 1] WATER 11 if more than one process is operating. For example, if the locus of dep- osition receives debris from two or more sources the material de- posited will have one or more size distributions deposited upon another, with the result that the size distribution of the sediments becomes skewed or has two or more modes. Likewise, if an area is subject to variations in average velocity, a similar result could arise. Also, as discussed below, sediments deposited upon submerged ridges commonly are disturbed or riled by moving currents, which cause the fine par- ticles to be carried away and deposited elsewhere. The sediments on the ridges are relatively well sorted, but those deposited in the lee (if the word lee can be used with respect to moving water) are likely to be skewed because they receive the normal supply of detritus for that particular distance from land as well as debris winnowed away from the nearby ridge. If the velocity of the water changes so that it moves at a different velocity or in a different direction for any appreciable length of time, or if the supply of debris changes in character, the resulting size distri- bution of the sediments is different. In this way laminae or strata are formed. The development of layers is one of the principal character- istics of sediments (Andrée, 1916; Barrell, 1917; Bucher, 1919; Antevs, 1922; Rubey, 1930; Weller, 19830; McKee, 1939; Payne, 1942; Alling, 1945). Sediments ordinarily are deposited in nearly horizontal layers. However, if the surface upon which they are laid down is uneven, the individual layers tend to conform to that uneven slope. With con- tinued deposition, the inequalities in slope become less, and eventually the angle at which the layers are deposited corresponds to the equi- librium position for the prevailing load and velocity of water. This initial dip of the sediments is usually small, but in areas of variable currents and on deltas it may be appreciable. In such places the dip varies in direction and amount and gives rise to cross-bedding. Varia- tions in velocity of water ordinarily are greater in shallow water than in deep water; consequently cross-bedding is more likely to be indic- ative of a shallow-water origin of sediments than of a deep-water origin. Cross-bedding, however, can form in water of any depth, so long as variable currents exist at or near the bottom. When streams enter the sea or lakes, the velocity is checked and part of the transported*load is deposited. Unless longshore currents transport the debris away, deltas form. Deltas are flat and gently shelving on top near where the stream enters, but, at some point out- ward from shore, the flat top of the delta gives way to a relatively steep front, which is constantly built forward at essentially the same 12 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 angle of slope as particles are swept forward off the flat part of the delta. At the base of the slope the layers flatten out. The three types of beds formed in this process are called top-set, fore-set, and bottom- set beds. In lakes or the sea, where the bottom slopes rapidly away from shore, the inclination of the fore-set layers, along the forward part of the delta, may be great; but in large bodies of water, where the water deepens gradually, the inclination of the fore-set beds is likely to be small. In fact, difficulty may be encountered in distinguishing fore-set beds from ordinary cross-bedding caused by variable currents. Geologists, when trying to determine the extent to which the angle of inclination of ancient sediments is due to initial dip, should bear in mind that initial dip in lakes is likely to be materially greater than in the sea. Steep initial dips commonly reflect diastrophism prior to dep- osition, which has led to the presence of relatively deep water close to shore or to the formation of an uneven bottom on which sediments are deposited. WIND Wind-blown or eolian deposits are essentially a characteristic of arid or semi-arid regions rather than of humid regions, because in arid regions the soil is poorly protected by vegetation and can be blown about by the wind. In present arid areas most eolian deposits consist of sand. However, during the Ice age, great masses of rock flour com- posed of fine particles of silt size, formed by the grinding action of the elaciers, were washed out by streams flowing from the ice and were de- posited in great outwash plains (Thwaites, 1941). In the absence of vegetation these silt particles could be blown away by the wind. The extensive deposits of loess, which are found around the southern border of the glaciated area, in part are believed to have formed from such wind-blown silt. These loess deposits commonly weather in vertical cliffs, in contrast with wind-blown sand, which forms slopes of 30° to 35°, representing the normal angle of repose of incoherent material. Wind-blown sand is well sorted because of winnowing action of the air which blows the fine particles away and transports the larger par- ticles essentially by rolling or jumping along the ground. For a full discussion of eolian deposits see Bagnold (1941), Reiche (1945), and the papers in the Loess symposium (1945). Wind-blown sand is de- posited mainly in dunes or hummocks. In ‘this process the grains blow up the side of the dune on a comparatively gentle slope and then, as they pass over the crest where they are not supported by the slope, many particles come to rest on the lee side. Thus dunes are con- tinually progressing in the direction in which the wind blows. If the Cu. 1] ICE 13 wind changes, the inclination and direction of the layers of sand like- wise change. Consequently dune deposits are characterized by cross- bedding. IcE Extensive deposits of glacial sediments are found in northern lati- tudes, in both North America and Europe. These deposits consist of three principal types: (1) groundtill, (2) morainal material, and (3) outwash. As the glacier moves along, it plucks up rock fragments and boulders of all sizes. In addition it scours the rocks over which it passes, producing much fine debris of silt and clay size. This frag- mental material becomes embedded in the ice, and, when the ice melts, the debris settles on the surface of the land, forming deposits called groundtill. The thickness of this till ranges from a few feet to several hundred feet, depending to some extent upon the depth and past history of the ice mass. This till covers the ground like a blanket and is composed of particles of all conceivable sizes, from huge bould- ers many feet in length to finely comminuted clay. The rock fragments are angular in outline but commonly have rounded edges. The par- ticles of clay size in some areas contain relatively little mineral clay. Thus the stability of the till is likely to be materially different than that of normal fine-grained deposits of lakes, rivers, or the sea, a point foundation engineers should bear in mind. An essential char- acteristic of till is the presence of numerous impermeable layers or zones which prevent the free flow of water. A similar type of deposit called morainal material is laid down at’ the sides or in front of glaciers. The front of a glacier is maintained by a balance between the rate of melting of the ice and the rate of forward movement of the glacier. Commonly this balance is so even that the front of a glacier stays at one position for a long time. As the ice is constantly moving, the material carried with it is trans- ported to the front of the glacier and is dumped when the ice melts. Thus big ridges called moraines are built up. Similar ridges form at the sides of the mass of moving ice. The material in these moraines is similar to grounditill. Water-laid deposits also are associated with glaciers. The ice melts during summer, and the melted water finds its way along the surface of the ice, in places dropping down cracks to the interior or bottom. Eventually this water emerges from the side or front of the glacier and deposits its load in broad plains. As the water flows across these plains rapidly, the deposits are well sorted, and ee of the particles of gravel size are well rounded. 14 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 Glacial deposits thus are characterized by a heterogeneous mixture of impervious and poorly sorted deposits interbedded with well-sorted, highly porous, and permeable sands and gravels. The geologist is of material service to the engineer in locating deposits of sand and gravel, close to their place of use, thus saving transportation costs. Moreover, the buried lenses of sand and gravel are good sources of underground water. (See Chapter 6.) Lakes are common in glacial country. In the spring and summer, melting water transports a large amount of debris to the lakes. The coarse particles soon settle out and form sand or coarse silt deposits. In the winter, the lakes freeze so that no sediment can enter. Deposi- tion is therefore confined to material largely of clay size which has stayed in suspension in the water during the preceding months. Thus a series of alternating coarse and fine deposits are formed (Antevs, 1922). These annual pairs of layers are called varves. The sand lay- ers, being permeable, facilitate consolidation under load. GRAVITY Gravity is also an agent of deposition. In cold climates or in arid regions where differences in temperature between night and day may be extreme, fragments of rock are dislodged owing to differential thermal expansion and contraction of the rock constituents and of the water in cracks and pore spaces. If slopes are sufficiently steep or smooth, dislodged fragments roll down and come to rest at some lower altitude, forming piles of debris which generally have a slope of 30° ‘to 385°. Deposits of this type are called talus and may attain con- siderable thickness if the blocks are not removed by other processes of erosion. Talus deposits oddly are rare in many arid regions, perhaps because the rocks disintegrate and occasional large floods wash them away. In humid and temperate climates, alternate freezing and thawing of water in the pore spaces of soil or alternating periods of wetting and drying of the soil result in a slow creep of soil particles down slopes (Capps, 1941). In this process of creep, the constituents of the soil slowly change their position with respect to other constituents, thus changing the strength and stability of the soil. Landslides are a manifestation of the effect of gravity. For de- tails see Sharpe (1938) and Chapter 13 in this symposium. Soil and rock resting upon a slope start to slide when the weight of the over- burden becomes greater than the strength or cohesion of the earth materials. Landslide deposits commonly are heterogeneous in nature because of the jostling caused as the material slides. Thus the sedi- Cu. 1] CHEMICAL ACTION 15 ment or mass of earth material is likely to be weaker than it was be- fore the slide took place. Landslides commonly leave characteristic scars upon the hills, showing the places from which they slide, and the surface of the landslide deposits is hummocky. Factors that affect landslides are the slope of the land, the strength of the earth materials, and the load that exists or is imposed. Loads can be imposed by adding material to the upper parts or by under- cutting the lower parts. The type of clay minerals and the amount of water in the soil materially affect the strength of the soil. VOLCANISM Volcanic sediments consist mainly of layers of wind-blown ash. As eruptions of ash come at irregular intervals and as the distance to which the ash is transported varies with the wind, successive layers of ash vary in grain size. They also vary in chemical or mineral composition if the source of the volcanic material changes. Ash de- posits commonly are interspersed with lava flows. Most deposits of voleanic ash contain bombs or fragments of rock and lava blown from the voleano. When these bombs land on newly deposited ash they bury themself in the ground, in many places more than one or two feet, and thus deform the layers upon which they land. The presence of these buried bombs leads to irregular consolidation when the overlying soil is loaded. Furthermore, voleanic ash commonly gives rise to montmorillonite clay, which has a great affinity for water and pro- duces relatively weak deposits when wet. ORGANISMS Sediments of biologic origin are essentially a product of environ- ment, because living organisms are dependent upon environment for existence. Furthermore, since deposits consisting chiefly of organic material can form only when the rate of production of organic material is significantly greater than the influx of inorganic material, deposits of organic origin, such as coal, diatomite, oil shale, bog iron ore, or algal limestone, require special conditions of formation (Bradley, 1931; Bramlette, 1946; Cady, 1942; Harder, 1919). Deposits rich in or- ganic matter commonly are relatively rich in radioactive material (Beers and Goodman, 1944). CHEMICAL ACTION Chemical action is essentially a question of degree of saturation. If water is undersaturated it can dissolve material; if it is oversaturated 16 Trask. PRINCIPLES OF SEDIMENTATION [Cu. 1 it can precipitate material. The principal factors are ionic concentra- tion, temperature, oxidation-reduction potential, amount of water, mi- crobial action, types of earth materials in contact with the water, and rate at which conditions change. Pressure usually is not a critical factor. Climate is a dominant influence because of its effect upon temperature, rainfall, and microbial action. Variations in moisture and temperature throughout the year or from one year to another strongly influence chemical reactions because of their effects upon equilibrium conditions and the degree of saturation of dissolved sub- stances. Evaporation changes the concentration of dissolved salts and thus changes equilibrium relationships. The distribution of water in the ground affects the oxidation-reduction potential. Earth materials lying above the water table tend to be more oxidized than correspond- ing materials lying below the water table. Changes in water level thus can significantly affect chemical reactions in the earth. Rate of flow influences the degree of saturation. Slow movement of water through pore spaces favors the development of a relatively high degree of saturation of dissolved substances, but, if the rate is so slow that the water becomes saturated, no further material can be dissolved. This feature has been brought out by Davis (1930) and others in respect to the development of cavities in limestone. The concentration of the different ions, particularly the hydrogen- ion concentration, is an important factor. The presence of even a small amount of some minor element may significantly affect the solu- bility relations of other substances. Boron, for example, materially affects solubility of caletum carbonate in sea water. The solution of many substances is aided by the activity of microorganisms or soil- forming processes (Waksman, 1936; Jenny, 1941; Zobell, 1946). M1- crobial action is favored by warm and humid climate. Parent earth materials vary greatly in their susceptibility to solu- tion. Limestone is readily dissolved by acid solutions and voleanic ash by basic or alkaline solutions. Salt, gypsum, nitrate, and other products of chemical evaporation are dissolved by moving water of almost any character. Chemical deposition is essentially a question of environment. The chief requirement is the development of a condition of oversaturation. The temperature, pressure, or concentration of dissolved substances must change before material can be precipitated. If the solutions are supersaturated, foreign material sometimes needs to be added to start the precipitating action. Evaporation is the chief cause of changes Cu. 1] INTERRELATION OF FACTORS 17 in degree of saturation, though microbial activity or the addition of other substances, such as iron, manganese, or silica from submarine or surface springs, may also influence the solubility relationships. The oxidation-reduction potential, as mentioned above, is also a factor, especially with respect to the development of red color in rocks (Tom- linson, 1916). ENVIRONMENT INTERRELATION OF FACTORS Deposition of sediments, as has been pointed out above, is influenced very strongly by environment. An environment of deposition may be considered an area in which the combined effect of the fundamental factors of topography, temperature, water, organisms, and rates of change of conditions are similar. Depending on the degree of simi- larity of factors, subenvironments exist within general environments. The ocean has many characteristics in common throughout the world, but different parts vary. The Gulf of Mexico, for example, is quite different from the Bay of Fundy. The fundamental factors that affect environment are interrelated. Changes in one factor result in corresponding changes in other factors, sometimes to a marked degree. The shape of the land influences pre- cipitation and flow of water and its resulting effect upon erosion, trans- portation, and deposition of debris. Topography affects the climate and hence the activity of organisms. The distribution of water influences the shape of the ground by its action on erosion and deposition; it affects the climate by its influence upon precipitation, evaporation, and the transfer of heat from the tropics toward the poles through the medium of oceanic circulation. Moreover, the distribution of water influences the activity of organisms by its effect upon food supply and distribution of oxygen and carbon dioxide. Temperature profoundly influences all other factors, primarily be- cause it is a source of energy. Climate is a direct result of the distri- bution of temperature and precipitation. Wind and rain are closely dependent upon temperature, and these two agents in turn profoundly affect erosion, transportation, and deposition and thus the shape of the land. Even diastrophism may be regarded as resulting from changes in distribution of temperature within the earth. Living organisms and plants influence the configuration of land because of their effect on erosion and on chemical action, which results in changes of the shape of the land. Organisms affect the distribution of water by transpiration and by increasing the infiltration of water 18 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 into the soil. Their plentifulness or scarcity modifies the rates at which environmental conditions change. Changes in environmental conditions in turn influence the environ- mental factors themselves. Diastrophism, or deformation of the earth, exerts a dominant influence upon environment, either directly by alter- ing the shape of the land and the distribution of water, or indirectly by modifying the climate and thus in turn the rate of erosion, trans- portation, and deposition. The movement of water, particularly cur- rents in the sea, is greatly influenced by diastrophism (Bailey, 1936; Krynine, 1940). OcEANIC PROCESSES The chief locus of deposition of sediments is the ocean. Any person attempting to study deposition in the sea should be familiar with the general principles of oceanography. For a comprehensive description see The Oceans, by Sverdrup, Johnson, and Fleming (1942). The surface temperature of the ocean is high near the tropics and low near the poles. The temperature of the water decreases with depth in all latitudes, but below a depth of 5,000 feet it is remarkably con- stant with respect to latitude, being everywhere less than 40° F. at that depth. As the temperature of sea water decreases, its density in- creases. Heavy water cannot long overlie light water; it must sink to a level where it encounters water of its own density. Thus, once cold and heavy water has sunk to a lower level in the sea, special con- ditions must develop to cause it to rise to higher levels. Hence the © water in the ocean is essentially stratified with respect to density. Water within individual layers is free to move within that layer. In fact, in some places water at different depths is flowing in different directions. Water moving at a given level may be deflected upward as it passes over a ridge. The effect of submerged ridges on the turbu- lence and the vertical motion of water is similar to the effect of moun- tain gaps on wind. This disturbance of water over ridges results in winnowing of the finer particles and causes the sediments on ridges to be more coarse-grained than those upon adjoining slopes in deeper water (Trask, 1932). No matter what the depth of water, the sedi- ments on the Mid-Atlantic Ridge in more than 10,000 feet of water are materially coarser than those on either side of the ridge (Bradley et al., 1942). Evaporation of ocean water increases the density. If the density increases, a column of water of unit height and cross section weighs more than a column of lighter water of the same height and unit cross Cu. 1] OCEANIC PROCESSES 19 section, with the result that water tends to flow from the more dense area to the less dense area. This process of adjustment of mass to- gether with the wind causes the great ocean currents, such as the Gulf Stream and the Japanese current. Owing to the rotation of the earth, currents are deflected to the right in the northern hemisphere and to the left in the southern hemisphere. The wind generates currents and tends to pile water up ahead of it, thus overbalancing the water in front with respect to the water be- hind. As a consequence the surface water in front sinks and the water behind rises, bringing the deep water to the surface. This deep water is rich in mineral nutrients needed for the microscopic plants, the plankton, which are the basic source of food in the sea and of organic matter in the sediments beneath. Upwelling of this sort is common off the coast of California. The ocean currents convey warm water from the tropics toward the poles and thus affect the temperature of the air, which in turn influ- ‘ences the movement of the wind, which then affects the motion of the water. Thus the ocean, like a dog chasing its tail, is constantly trying to attain a condition of equilibrium but never does. However, in many parts of the ocean the rates of change on the average are fairly con- stant; so a condition of dynamic, in contrast with static, equilibrium is obtained. This general circulation of the water has profound effects upon climate, the growth of organisms, the concentration of dissolved substances in the sea, and thus in turn upon the deposition of sedi- ments. Currents, however, are produced in the sea in several other ways. Waves approaching the shore diagonally pile the water upon the beach at an angle with the shore, and the water descending from the beach reaches the sea at some point downwind from where it first struck the beach (Munk and Traylor, 1947). Thus a longshore current develops which may have a profound effect upon transporta- tion and deposition of debris, particularly in areas in which the wind blows mainly from the same general region. Tides produce currents in shallow water. The large tsunamis which are reported to be caused by earthquakes conceivably could also set up currents. In water less than 600 feet deep waves can stir up the bottom and winnow the sediments (Stetson, 1938; Shepard, 1932, 1948). Internal waves, developed in layers of denser water beneath the surface, conceivably could generate currents. Coarse sediments in deep water in some areas may have resulted from such currents (Revelle, 1944; Sverdrup Anniversary Volume, 1948, pp. 673, 683). 20 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 Derr SEA The ocean contains several distinct environments of deposition with respect to supply of mechanical and chemical constituents, distance from land, character of water, and configuration of the basin of dep- osition (Murray and Renard, 1891). In the deep sea far from land, the rate of supply of mechanical debris is low. The sediments consist principally of material of organic or wind-blown (eolian) and cosmic origin. Relatively little material is now being precipitated in the deep ocean. Water shallower than 15,000 feet, however, apparently is satu- rated with calcium carbonate in some areas. Thus calcium carbonate perhaps is now being precipitated chemically, though most of the car- bonate seems to be of organic origin. The ocean water below a depth of 20,000 feet at present appears to be undersaturated with calcium carbonate. As a result, particles of calcium carbonate seemingly dis- solve as they settle to the bottom. The sediments contain little cal- cium carbonate. The undersaturation of the deep water in the ocean, however, should not be considered a reliable index of past condi- tions, because the present state almost certainly is due to the addi- tion of cold and dense water to the bottom of the ocean during the Ice age. As this cold water sank during Pleistocene time it must have displaced upward, or cooled, the water that formerly lay at the bottom of the ocean. The deep water in other geologic periods may well have been materially warmer and more thoroughly saturated with calcium carbonate than is the present water. Consequently some ancient lime- stones may be of deep-water origin. Radiolarian chert is considered by some people to be of deep-sea origin, because at present radiolarian deposits are found principally in deep water. Siliceous deposits of this type represent a relatively rapid rate of deposition of silica with respect to terrigenous debris (Davis, 1918). Such conditions most certainly can prevail in deep water far from shore, but they also could exist in basins near shore if the adjoining land areas supplied little debris and if an excess of silica were supplied the sea from submarine sources. Some of the radiolarian deposits of the Franciscan formation of California may be of shallow-water origin. Coarse sediments can be found far from shore in deep water, as indi- cated by the poorly sorted glacial marine deposits reported by Bradley and Bramlette (Bradley et al., 1942) from the North Atlantic. These sediments contain terrigenous constituents of ice-raft origin. Cu. 1] BASINS 21 BASINS Deposits in basins, either in the ocean far from land or near shore, are of many types. The depth of the sill in these basins influences very materially the conditions in the basin and thus in turn the de- posits. The sill may be defined as the outlet of the basin if sea level were lowered to the extent that the water in the basin would be entirely enclosed by dry land. If the sill is deep, the water below sill depth is more likely to escape than if the sill is at shallow depth. The water in basins with deep sills, therefore, is less apt to become stagnant. The oxygen content of the water in any basin, however, ordinarily will be less than in the overlying water, and the sediments are likely to be in a higher state of reduction than similar deposits laid down in more open places on the sea floor. The organic content is apt to be relatively high, as the lowered oxygen content inhibits the decomposi- tion of organic matter. Basins favor chemical precipitation, because the lack of circulation or the poor circulation favors the increase in concentration of dissolved substances to such an extent that a state of saturation develops. The addition of chemical substances such as silica, iron, or manganese from submarine springs is more likely to result in the formation of chemical deposits in basins than in other places. If the sill is relatively shallow and the basin relatively deep, verti- eal circulation, or ventilation of the water as it is called, is much in- hibited, and the water becomes stagnant (Strgm, 1939). Hydrogen sulphide is formed, and highly reduced sediments rich in organic deposits are laid down. If the rate of influx of terrigenous debris is small, deposits largely of a chemical nature form in basins. Extensive deposits of salt, gypsum, potash, phosphate, and limestone may form in this way (Loétze, 1938). Alternations of the level of water above or below sill level or specially favored areas of evaporation, like the Gulf of Kara Bugaz in the Caspian Sea, favor the influx of chemical ingredients and ensuing concentration to the point of precipitation. Unfortunately examples of basins of many of the types of chemical deposits formed in the past do not exist today. Certainly no areas are known where dolomite is now forming (Sander, 1936). Geologists in endeavoring to determine the mode of origin of saline and other chemi- cal deposits would do well to consider carefully the fundamental prin- 22 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 GEOSYNCLINES Geosynclines have many characteristics of basins (Jones, 1938), though they are not necessarily enclosed in character. In fact they are simply low areas in which thick deposits of sediments accumulate. Because of the presence of cross-bedding, ripple marks, and alterna- tions of sand, silt, and clay, many people regard geosynclines as essen- tially shallow, but they may be deep in part. Recent work in the Gulf of Mexico (Sverdrup Anniversary Volume, 1948, p. 683) has shown that cross-bedded sand and silt are found in deep water in the middle of the Gulf of Mexico. More likely, the depth at which the deposits form in a geosycline represents a balance between rate of sinking of the basin and rate of influx of debris. When studying the origin of sediments in ancient geosynclines, ge- ologists should consider the fundamental factors of quantity and qual- ity of debris supplied the basin, particularly the size of the particles, and the probable currents that influenced the deposition of the sedi- ments, especially the possibility of longshore currents which could transport debris far from the mouths of rivers that supplied the detri- tus. An understanding of the shape of the geosynclinal basin should also help in interpreting the mode of origin of the sediments. CONTINENTAL SLOPE Deposits on the continental slope, that is, on the slope from the flat continental shelf down to the abyssal deeps, are conditioned by the distance from land, the existence of currents or turbulence along this slope, the angle of slope, and the type of debris supplied. The deposits in general are fine-grained, but they may consist of sand. The sedi- ments also may have unusual skewed size distributions in places where material swept from the adjoining flat continental shelf is added to the normal supply of debris that is deposited. The angle of slope ranges from less than 1° to more than 10°. If the deposits are relatively fine-grained, as they are likely to be, conditions of instability of slope may prevail, with the result that considerable slumping may take place (Fairbridge, 1946). Deposits in which the layers of sediments seemingly have been deformed contemporaneously with sedimentation, therefore, may be continental slope deposits. Some of the Franciscan sediments, which are so badly distorted, may have such an origin. CoNTINENTAL SHELF A continental shelf extends outward from shore, 10 to 500 miles in many parts of the world (Shepard, 1932). This shelf slopes only a Cu. 1] BEACHES 23 few feet per mile and is characterized by relatively coarse, well-sorted deposits, mostly of sand or coarse silt size. As the water at the outer edge of the shelf is 300 to 600 feet deep, the floor of the shelf is within the reach of wave action during times of heavy storms. Undulations in the continental shelf influence materially the texture of the deposits. The fine material tends to concentrate in depressions on the shelf. Ridges or masses of sediment that rise above the general level of the platform are found on the continental shelf in several places, especially the Gulf of Mexico, the Atlantic Coast, and the North Sea. Some of these ridges may represent former shore lines when the sea stood at a lower level than it now does, but they also may represent submerged masses of sand which migrate across the shelf just as dunes march over the land (Liiders, in Trask, 1939, p. 337). BEACHES Beach deposits are of many different types, depending primarily on the configuration of the land and the strength with which waves can strike the beach. Where the ground slopes very gently both seaward and landward from shore, as in the Gulf of Mexico along the Texas coast, the beach deposits are very fine-grained. Not only is the size of particle supplied the sea of relatively small size, but also the gentle slope of the sea bottom interferes with the progress of large waves toward the beach. However, whether the flat slope is the result or the cause of the distribution of sediments is a question. Geologic text- books speak of a profile of equilibrium, but they do not correlate the angle of slope of the profile with the distribution of energy in the waves, or with the size and amount of material available for transport by the waves. It seems as if the profile depends very much upon the supply of debris, which in turn is influenced by the distribution of currents and the overall configuration of the locus of deposition. Cer- tainly the angle of slope of beaches is related to the size of the con- stituent particles. Steep slopes tend to have coarse particles, and gentle slopes fine particles; so why does not the same relation apply to the shallow water off shore? Beaches are influenced greatly by longshore currents, which in turn are influenced by the direction of prevailing wind (Munk and Traylor, 1947). Particles of sand are transported along the shore and form spits and bars (Gilbert, 1890). The waves and currents tend to cause the edge of the beach to migrate shoreward, particularly during time of high water and strong waves. Whether or not a beach erodes or builds seaward depends on whether sand is supplied the beach at a 24 Trask. PRINCIPLES OF SEDIMENTATION i@zet slower or faster rate than it is removed from the beach. (See Chapter 15.) In general a sort of dynamic equilibrium prevails. Bays In areas where sea level is rising or has risen recently, the mouths of streams become drowned and form bays. The tides pass in and out of the mouths of these bays. The water in most bays is shallow, and the diurnal movement of the tide keeps the water muddy. (See Chapter 16.) The position of places of maximum velocity of tidal currents changes from time to time, thus causing continual erosion and deposition of sediment in the bay. Silt and clay are deposited in quiet water, and sand or coarse silt is laid down in or near the channels. The sediments are poorly sorted and vary greatly in tex- ture from place to place. The tidal flats are exposed during low water, and the sediments alternately take on and lose water as the tide comes and goes. Organisms ingest the sediment in search of food. The deposits, therefore, are irregular in character, bedding, and strength. AREAS OF CALCIUM CARBONATE DEPOSITION Limestone is formed in several types of environments. Saturation or relatively high concentration of calcium carbonate in the water, however, is common to all types. In addition the rate of influx of terrigenous debris must be comparatively slow with respect to rate of deposition of calcium carbonate. Concentration of calcium car- bonate in water is favored by increase in temperature and salinity and decrease in hydrogen-ion concentration (increase in pH) (Watten- berg, 1933; Trask, 1937). Calcium carbonate may be precipitated directly from supersaturated water; it may be formed by bacteria or algae as a part of the metabolic processes by which the organisms get food and energy from the water; or it may be formed from detrital shell particles. Precipitation of caletum carbonate from saturated or supersaturated solutions is favored by the presence of solid par- ticles of calcium carbonate or other material. In places this process produces small spherical pellets which are called oolites (Brown, 1914). The most favorable environment of deposition of calcium carbonate is a shallow flat-bottomed sea adjacent to a low-lying coast in a tropical or semi-tropical climate, as for example the southern coast of Florida (Vaughan, 1910). Other environments are lagoons, atolls, coral reefs (Emery, 1948), or even the deep sea where the water is less than 20,000 feet deep. (See Chapter 33 in this symposium.) Calcium carbonate may even be deposited from saturated water in lakes (Rus- Cu. 1] RIVERS AND DELTAS 25 sell, 1885). The activity of organisms is a potent influence in the deposition of calcium carbonate in lakes. RIVERS AND DELTAS Many different types of environments of deposition are found on land. For the world as a whole, terrestrial deposits are less plentiful than marine deposits; but in places, notably in Wyoming and Colo- rado, thick non-marine deposits have accumulated in the past. These sediments are mainly the result of stream action. The essential re- quirement is the formation of a place of deposition, particularly a flat surface upon which the rivers can deposit their load at times when they get out of their channels. Most deposits laid down by running water on land are poorly sorted and poorly stratified, except in channels where currents are relatively strong and uniform. Most river and flood-plain deposits vary in thickness and are cross-bedded. Alternations between coarse and fine deposits are common. River deposits are characterized by cut-and-fill action during flood. (See Chapter 18.) When the discharge is great and the velocity of the water is high, the bed of the stream is scoured to a greater depth than during normal periods of flow. As the flood abates and the stream returns to normal levels, the depth of scour diminishes and the bed fills up to its customary position. Large streams may be scoured more than 50 feet, as is indicated by boards found during constructions of piers, or by the undermining of deep piers during floods. In con- structing caissons in rivers, geologists and engineers should endeavor to determine the depth to which the river scours during floods and de- sign foundations accordingly. At time of flood, streams rise above their channel and spread across the adjoining flat land, leaving deposits of silt and sand (Fisk, 1944). Deposits of this type ordinarily are not well sorted and tend to vary in grain size and thickness from place to place both longitudinally and laterally across the flood plain. Deposits commonly are thickest at the boundary between the channel and the flood plain, because de- crease in velocity of the water at such places favors the deposition of sediment. ‘This process results in the formation of natural levees. The.deposits at the mouths of large streams likewise are irregular. Sand alternates with silt and clay almost indiscriminately both later- ally and vertically. The sediments are poorly sorted and commonly are mixed with plant fragments (R. J. Russell, 1936). Thick deposits build up off the mouth of the stream unless longshore currents carry the debris away. A factor in the development of longshore currents is 26 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 the rate at which the delta sinks. If the rate of sinking is rapid, the mouth of the delta becomes embayed and longshore currents cannot effectively remove the debris until the delta builds forward to a point where the longshore currents can sweep the shore. In the Mississippi delta in some places, masses of clay rise up through the sediments to form mud lumps or mud boils (Shaw, 1913). The mud is pushed up in much the same way as mud rises in front of an embankment which fails because of overloading. Perhaps a similar overloading due to natural causes takes place in deltas. At any rate, displacement of masses of mud in this manner, if found in ancient sediments, might well be interpreted as evidence of contem- poraneous diastrophism when it really was only a readjustment of unstable sediments due to imposed loads. Desert AREAS In desert or semi-arid regions, rainfall is characterized by infre- quent but intense storms of local extent. After such storms the water commonly flows down the slope of the land in a sort of sheet, re- sulting in the formation of gently sloping cones (Paige, 1912; Law- son, 1915; Woodford, 1925; Bryan, 1936; Gilluly, 1937). The shape of the cone seems to be a function of the average load and the average velocity of water during time of flood. If irregularities protrude above the level of the surface of equilibrium they are worn away; if irregu- larities extend down below the surface of equilibrium they are filled up. The upward ends of the fans abut against a steep mountain slope or merge into a stream coming down through a valley in the mountains. As the mountain surface is eroded backward or as the upward end of the fan moves mountainward, the relationships of load to average dis- charge and to profile of equilibrium change, with the result that de- posits tend to accumulate in the lower parts of the fan and to be thin on the upper parts. The thinly covered rock surfaces on the upper parts of fans are called pediments. The deposits generally consist of two sizes of particles. The larger eroup of particles consists of angular or subangular fragments of rock, commonly ranging from 1 to 10 inches in diameter; and the smaller group consists of particles of coarse sand or fine gravel size. The larger particles ordinarily form less than 25 percent of the weight of the rock, and the finer particles more than 75 percent. Apparently the size of the smaller particles is a function of the transportive abil- ity of the water. The size of the larger particles seemingly is gov- erned by the capacity of the water to push them along over the smaller particles, which act somewhat like roller bearings. Cu. 1] SWAMPS 27 At the lower end the fan merges into a more or less broad valley filled with alluvium, or the fan may pass gradually into a playa which, during times of flood, is converted into a lake. The playas are exceed- ingly level and are dry most of the time. The lakes that occupy them during and after a flood are broad and shallow (Russell, 1885). The relative frequency of floods, the supply of debris, and the size of the playa govern the types of deposits that are laid down. The sediments vary in texture from layer to layer, but in general they are fine- grained. A variety of organic deposits depending on food supply and length of time required for the lakes to dry up is laid down. Some of the playas, such as Searles Lake in California, give rise to economic deposits of potash, salt, and borates (Foshag, 1926). Many of the streams in the semi-arid parts of western United States occupy broad valleys filled with alluvium. This alluvium consists of several different kinds of layers, some of which are discolored brown or black on top, and they probably represent more or less long inter - vals during which soil developed. Some of the layers are remarkably uniform in thickness for a distance of a mile or more. The layers are lenticular and presumably represent deposits of single floods or of single periods of similar climate. The texture of the deposits ranges from clay to sand. For further information see Chapter 23. This alluvium has been subjected to alternate periods of fill and scour. During periods of scour, deep gullies form which in time swing laterally to erode some of the fill previously deposited in the valley. At the present time a period of scour characterizes most of the South- west, particularly New Mexico, Arizona, Wyoming, and Utah, but only to a relatively small extent central and eastern Nevada and cen- tral California. The cause of the present scour has been attributed to overgrazing, but it can also develop from natural causes, because as Peterson shows in Chapter 23 of this symposium there have been at least three periods of scour and subsequent fill since Late Pleistocene time, caused by geologic agencies alone. SwAMPS The stagnant water in swamps favors the accumulation of woody and other types of plant material that give rise to coal deposits (Cady, 1942). Such environments commonly are found near the outer edges of coastal plains, particularly in areas bordered by barrier beaches or bounded by deposits laid down in distributaries of deltas. The essen- tial feature is the empoundment of water in a broad, low area in which vegetation, particularly trees, can flourish. Depending upon the rela- tive distribution of wooded and open areas, different types of organic 28 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 material accumulate. In the wooded areas the material consists prin- cipally of the partially decomposed fragments of wood embedded in a matrix of more or less completely decomposed woody material. This matrix has some of the characteristics of a colloidal gel of organic material. With time, such deposits give rise to ordinary coal. In the open areas pollen and spore material may collect. Such deposits ulti- mately form cannel coal. In places, particularly in open water, algae may develop luxuriantly and give rise to beds of boghead coal or oil shale. The essential feature in the formation of any particular type of coaly material is the relative abundance of woody, spore, and algal material preserved in the sediments. Thus distribution of wooded and open areas, types of plants present, length of time stagnant con- ditions prevail, rate of influx of terrigenous material, and subsequent ceologic history are essential factors in the formation of coal. LAKES Lake sediments are influenced materially by the nature of the water in the lake. Geologists interested in lacustrine deposits should be familiar with the fundamental principles of limnology, as there are many different types of lakes (Welch, 1935; Bradley, 1948). Criti- cal factors are average temperature of water, extremes of tempera- ture, quantity of rainfall, seasonal and cyclic distribution of rainfall, activity of microorganisms, concentration of dissolved materials, size and shape of lake, depth of water, and supply of detritus. The sea- sonal variations in temperature of the water profoundly influence the type of deposits. Fresh water attains its maximum density at 39° F. As the surface water approaches 39° F. it becomes heavy and sinks to the bottom of the lake, forcing the bottom water upward. In the more northern latitudes this phenomenon happens twice a year; once in the spring as the surface water warms up to 39° F., and again in the fall when it falls to 39° F. As the deeper layers of water com- monly are relatively rich in organic matter and mineral nutrients, plant life tends to increase when this deep water rises to the surface in the spring. However, if during the summer months the activity of microorganisms has consumed all the dissolved oxygen in the lower water, hydrogen sulphide is formed, which, when it rises to the surface during the spring and fall turnovers, may kill a considerable part of the life in the lake. In more tropic areas, temperature of the surface water never falls below 39°, with the result that the deep water does not rise to the sur- face. Thus, a permanent state of stagnation may be produced in the deep water, in which hydrogen sulphide can be continually present and Cx. 1] DIAGENESIS 29 only a few types of organisms, particularly certain anaerobic bacteria, can live. A heavy storm or a rare cold spell may cause some of the deep water to rise to the surface, killing off much of the life. This influence of the temperature on the density of water influences materially the type of organisms that grow in the lake and hence the chemical equilibrium of the dissolved materials. Lakes that are periodically ventilated by overturning of the water are in a relatively high state of oxidation compared with those in which the water does not turn over. The depth of water, however, is a factor, because, the shallower the lake is, the more likely the bottom water is to be stirred up during a storm. In some lakes algae develop luxuriantly, giving rise to deposits of organic matter, which in the course of time may form deposits of oil shale (Bradley, 1931). The organic matter, when first deposited, has a very high water content, but oil shale generally is found in very thin layers, which indicates a high degree of consolidation or compaction during the course of geologic time. As Bradley has shown, layers of oil shale commonly are of a dual nature, one part representing summer conditions, and the other winter conditions. Thus, oil shale, at least the Green River oil shale, resembles varves formed in glacial lakes. The deposits in lakes are likely to be cross-bedded and highly vari- able near shore, particularly on the deltas where streams flow into lakes. The larger lakes have bars, spits, and beaches, which have essentially the same characteristics as corresponding features in the ocean. The sediments laid down in the central parts of the large lakes, like sediments in the open sea, are relatively fine in texture, and are comparatively well stratified; however, lacustrine deposits commonly change in texture and character from one layer to another. DIAGENESIS Sediments change after they have been deposited. The water is squeezed out of pore spaces owing to the effect of superimpesed loads, thus causing a reduction in thickness. This compaction of the de- posits presses the constituents more firmly together and gives the sediments greater strength. Materials are deposited in the pore spaces and cement the particles together. New minerals form and old min- erals grow by the addition of new materi#l. Ultimately the sediment is consolidated into rock. The sum total of these processes is called diagenesis. Diagenesis, however, is to be distinguished from meta- morphism, which results in the more or less extensive formation of new 30 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 minerals as a result of dynamic stress, comparatively high tempera- ture, and mineralizing solutions. At present the processes of diagenesis are not well understood. Com- paction of sediments, with resulting decrease in porosity, is clearly re- lated to the imposed load and starts as soon as the sediments have been deposited. Compaction is used by geologists in essentially the same sense as the word consolidation is used by soil mechanics engi- neers. The water content of the sediments, or, more appropriately, the porosity, is directly related to the texture of the sediments. Clays have a high initial water content compared with silts or sand and thus compact more than silt or sand. For sediments to compact, the water must be squeezed from the deposits. The permeability, therefore, is a big factor in the rate of consolidation. Sands, being permeable, have a high initial rate of compaction compared with clays, which are rela- tively impermeable. In fact, sands ordinarily compact very quickly to the poit where each grain is in contact with one or more other grains; after this they compact very slowly as grains are deformed or rotated so that adjoining grains can nestle more effectively into avail- able pore space. Clays likewise compact more rapidly at first, but, after the initial period of rapid forced removal of pore water, they compact very slowly for a long time, as the pseudo-anticlinal structure of sediments over buried hills indicates (Athy, 1930; Hedberg, 1936). Essential factors in this compaction are load, grain size, permeability, and time. During consolidation of the sediment into rock, that is, while water is being squeezed out, the strength or ability of the sediment to sup- port a load is affected to a considerable extent by deformation of min- eral constituents, which causes the individual constituents to be more tirmly locked together. While this process is going on, but particu- iarly after the initial stage of settlement, material is precipitated from the pore water and thus binds the constituents together. The concentration of dissolved substances, the rate of movement of the water, the solubility relationships of the dissolved materials, base ex- change, rock pressure, activity of microorganisms, size and shape of the pore spaces, and mineral composition of the rock particles all in- fluence the precipitation of cementing materials. Commonly these materials are laid down in pore spaces between grains, but in places zones of cement are deposited around some focus of precipitation un- til a nodule or concretion is formed. The causes of localization of pre- cipitated material in the pore spaces of sediments is at present poorly understood, but it is of great importance in interpreting the migration of oil and also the origin of ore deposits in permeable rocks. (See Cu. 1] CLASSIFICATION 31 Chapter 29.) The mineral constituents of sediments dissolve or grow to such an extent that as time goes on the grains are locked more firmly together, with resulting increase in strength of the rock. A great variety of minerals are known to be generated in sediments, but the controlling factors of formation of new minerals are not yet com- pletely known. (See Chapter 25 in this symposium, and Pettijohn, 1949, pp. 476, 500.) CLASSIFICATION Nomenclature of sediments is a complicated subject, principally because so many different attributes are used in classifying sediments. Many systems of terminology have been described among which may be mentioned those of Allen (1936), Casagrande (1947), Grabau (1913), Jenny (1941), Krumbein and Pettijohn (1938), Krynine (1940), Pettijohn (1949), Rutledge (1940), Twenhofel (1932, 1937), Udden (1914), Wentworth (1935), and Wentworth and Williams (1932). Most systems represent a combination of descriptive and genetic factors. However, in so far as possible, a classification should be descriptive, because a genetic classification is based on some sort of conception of the mode of origin of the deposit based on descriptive properties. Three main groups of sediments are generally recognized: (1) elastic deposits, composed principally of particles that have been transported mechanically and then deposited; (2) chemical deposits, composed mainly of materials precipitated from solution; and (3) organic deposits resulting from the activity of organisms. Some deposits such as limestones overlap all three groups. Lime- stones can be composed of particles transported by water, ice, or wind from some limestone area and then deposited; they may form directly as a result of chemical action; or they may result from the metabolic process of organisms. The end products in all three processes may be similar. The mode of origin, therefore, might be ascertained only after careful study, and perhaps not then. Siliceous deposits may also be difficult to classify, particularly after they have been subjected to percolating waters which have dissolved and reprecipitated various materials. However, most deposits of chemical origin, such as salt, potash, gypsum, chert, and some iron deposits, as well as organic de- posits of coal, oil, shale, and even petroleum, can be classified so clearly that no further discussion is needed here. The principal confusion ex- ists among the clastic or mechanical deposits. The clastic deposits are classified mainly upon the basis of the grain size of the major constituents. The names gravel, sand, silt, 32 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 clay, and colloid are universally adopted for classifying individual constituents by size. Fairly good agreement regarding boundary be- tween size groups has been achieved, particularly for sand particles, which almost everybody agrees range between 64 microns (0.064 milli- meter) and 2 millimeters in diameter. The lower limit for silt gen- erally is set at 4 or 5 microns. Workers interested in classifying sedi- ments into size groups whose upper and lower limits have geometric dimensions prefer 4 microns as the upper limit for clay because it makes the classification so much easier to handle statistically; but some workers interested in the physical properties of clays prefer to use 5 microns for the upper limit of clay because that size marks a division point between recognizable physical properties, based prin- cipally on the properties of clay minerals, which commonly do not reach much beyond 5 microns in size. Most people set the boundary between clay and colloid between 0.5 and 1 micron; but, as particles of this dimension are difficult to meas- ure accurately and as the actual content of colloids is not yet a matter of major importance for most practical applications of sediments, people have not argued much over the lower limit of clay size. Sediments are composed of particles of many different sizes. Geolo- gists are accustomed to classify the deposit upon the basis of the di- ameter of the average (median) particle, the Ds 9 of the engineer. The median is the diameter for which one half of the weight of the sedi- ment is composed of particles of larger size and one half of particles of smaller size. It represents the mid-point on the grading curve. If the sample is well sorted, that is, if most of the particles are close to the median, little difficulty in classification exists, but, if the sample is poorly sorted and contains a large amount of material larger or smaller than the median, the tendency is to use a modifying adjective. Thus a sample having a median diameter of 15 microns and containing 25 percent clay would be called a clayey silt. Difference of opinion ex- ists about how much of the lower or larger grade size need be included to warrant the use of a modifying adjective. In practice the amount ranges between 10 and 30 percent. As a rule workers use the name that most conveniently fits the purpose for which the work is done. Casagrande (1947) has adopted an arbitrary series of limits of par- ticle size and has given corresponding names to sediments of given size distributions. Soil mechanics engineers find this classification useful. The shape of the constituents is used in classifying sediments com- posed chiefly of particles larger than sand in size (Wadell, 1935). If rounded, the deposits are called gravel; if angular, they are called brec- Cu.1] IMPACT OF SEDIMENTS ON PRACTICAL ENDEAVOR 33 cia. Mineral or rock composition also is used for some of the coarser deposits. Coarse deposits consisting primarily of material of granitic origin, in which the rock and mineral fragments are distributed in essentially the same proportion as in the original rock, are called arkoses. The term graywacke is used by some for a sandstone com- posed mainly of feldspar and quartz embedded in a clay matrix, but others use the term to indicate material derived from basic rocks in contrast with arkose, which is derived from acid rocks. The proportion of caletum carbonate in a sediment causes con- fusion in terminology. A sediment should contain at least 30 per- cent calcium carbonate to be called limestone. If it contains between 10 and 30 percent the sediment is commonly called a calcareous sandstone or shale, depending upon the average grain size of the clastic constituents. Some workers object to the use of the term calcareous for this purpose, as they would like to distinguish a sandstone or shale cemented together by calcium carbonate from a sediment in which detrital particles of calcium carbonate are intimately mixed with par- ticles of terrigenous origin. Confusion also exists with respect to names for the sediments after they have been consolidated into rock. A consolidated sand is almost invariably called a sandstone, but a silt is classed as a siltstone or shale, depending on the degree of lamination of the rock. If massive, the rock is called a siltstone; if laminated, it is called a shale. De- posits of clay particles similarly are designated as mudstone and shale, respectively. If a worker is concerned about misinterpretation of the terms, he should describe the basis for the classification he uses. IMPACT OF SEDIMENTS ON PRACTICAL ENDEAVOR The practical applications of sediments depend on the properties of the sediments and the processes that affect them. The strength of sediments and the changes in strength under added or reduced stress are of fundamental importance to the construction engineer, whether he be interested in bearing strength of foundations, stability of slopes, or the construction of tunnels (Terzaghi and Peck, 1948; Taylor, 1948). The strength of loosely consolidated sediments and earth ma- terials depends on the water content, the grain size, the mineral com- position (particularly the type of clay minerals), the stresses to which the sediments have been subjected, and the length of time the stresses have been applied. The water content depends on, among other factors, the load, the grain size, and the clay mineralogy. The greater the load and the 34 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 longer the load has been applied, the lower is the water content and the less the response of the sediment to added load or added stress. The more permeable a bed is, the more rapidly the water is forced out of it and the more quickly the sediment comes into equilibrium with the stress applied. The finer the sediments and the more poorly they are sorted, the higher is their uniformity coefficient and the less permeable they become. When considering the release of pore-water pressure under added stress, the engineer and geologist would do well to look on this con- dition as a dynamic state and not as a static state. The water moves between the constituent grains, thus forming a greater proportion of the unit volume of a sediment in some places and a lesser proportion in others. If the stress continues long enough or is sufficiently strong, shear cracks develop through which water can move. Thus any measurement of shearing stress made in the laboratory must be inter- preted in light of conditions as they exist in the ground. A laboratory test may indicate a high angle of internal friction, but, when a load is placed upon the sediment in the field and the water begins to move through the constituent grains, the relations of water to individual grains change. The engineer should realize that, if the water forces its way between the grains so that the grains do not come to rest upon other grains in as many contacts per grain as they did previously, the shearing stress at such places becomes less and, if several such places should suddenly develop in a plane, a very different condition of shear- ing stress results than during a laboratory test in which the movement of the water within the sediment sample may be materially different from that in the field. It is perhaps for this reason that the apparent angle of internal friction is low along the planes on which sediments sometimes fail. Clay minerals differ in their affinity for water. Individual mineral grains or flakes of clay react differently to water under different con- centrations of cations in pore water (Grim, 1942; Kelley, 1939). This subject has not yet been thoroughly explored, but it is one that should be considered seriously by the engineer when working in areas where previous construction experience is scanty. (See Chapter 25.) Very likely the different relationships between liquid limit and plasticity index described by Casagrande (in Terzaghi and Peck, 1948, p. 36) depend upon the type of clay mineral. Otherwise why should all the sediments in a given area show such a constant relationship between liquid limit and plasticity index, whereas all the sediments in another area exhibit a different but equally consistent relationship? Alternating freezing and thawing of the ground in cold climates Cu.1] IMPACT OF SEDIMENTS ON PRACTICAL ENDEAVOR 35 materially affects the strength of the sediments. As Black shows in Chapter 14, this question is essentially one of dynamic relationship between water and ice within a sediment. The permeability and porosity of sediments are of great practical importance to petroleum and mining geologists, to ground-water hy- drologists, and also to construction engineers, because the movement of oil and water through ground depends so much upon the permeabil- ity (Meinzer, 1923; Muskat, 1937; Tolman, 1937; also Chapters 4, 6, and 32 in this symposium). Once again it should be mentioned that permeability is influenced by the size and degree of sorting of the con- stituent particles; porosity depends upon the size and shape of par- ticles and upon the load or degree of compaction (consolidation) of the sediments. For example, it is well known that the water content of a clay may be higher than that of a silt or sand but the permeability may be less. Clays normally do not have a low porosity unless they have been compacted for a long time. Permeability is of material significance to the mining geologist because the ease with which min- eralizing solutions and gases can penetrate rocks, particularly sedi- mentary rocks, in many places is a major factor in determining whether commercial ore deposits will be precipitated in the sediments. (See Chapter 29.) Both the physical and chemical properties of the constituent min- erals concern mining geologists in search of ore minerals. The manu- facturer also is interested in the nature of the sediments as sources of raw materials, as McKelvey points out in Chapter 27. The type of mineral, that is, the relationship of molecules within the unit cell, is of special concern to the manufacturer of clay products, as well as to the geologist and engineer working upon construction, because of the variation in properties with respect to type of clay mineral. At times the geologist uses mineral composition as a means of tracing sedimentary strata from one drill hole to another (correlation prob- lems), or as an index of the environmental conditions of deposition of the sediments. Physical properties of sediments are of special interest to manufacturers of abrasives, insulation materials, or aggre- gates for concrete. The mass properties of sediments, such as density, resistivity, radio- activity, or compressibility, concern the geophysicist and geologist in prospecting for oil, as indicated by Beers in Chapter 4. The processes that affect sediments concern all people who use sedi- ments. The character of sediments and their response to stress depend on the processes that lead to the formation of the sediments. In addi- tion, the processes themselves have certain very special applications, 36 Trask. PRINCIPLES OF SEDIMENTATION [Cx. 1 such as soil erosion, gullying of arable lands, degradation and aggrada- tion of stream valleys below dams, silting in reservoirs, beach erosion, development of harbors and breakwaters, stream channel control, and silting in irrigation canals (Brown, 1948). Lastly, sediments are of great use to the geologist interested in de- termining the mode of origin of sediments, or the environmental condi- tions under which they were deposited. For, if the geologist knows the mode of origin and if he understands the distribution of sediments and their attributes in the environment of deposition, he is in a position to predict what the nature of the sediments will be at some place be- neath the surface of the ground. Using information obtained from sediments exposed in some outcrop or drill hole, the geologist is con- stantly being called on to indicate the nature of the sediments at some distance from these exposed points. If the geologist cannot predict, he has to guess. Hence he should endeavor to learn all he can about sediments, so that he can foretell, with desired reliability, the nature of the deposits in the unexposed places. Geologists have a long way to go in this respect, but the goal is clear. It is hoped that this chapter will help some people along toward that goal. REFERENCES Allen, V. T. (1936). Terminology of medium-grained sediments: Report, Com- mittee on Sedimentation, 1935-1936 (National Research. Council), pp. 18-47. Alling, H. L. (1945). Use of microlithologies as illustrated by some New York sedimentary rocks: Bull. Geol. Soc. Amer., vol. 56, pp. 737-755. Andrée, K. (1916). Wesen, Ursachen und Art der Schichtung: Geol. Rundschau, Bd. 6, pp. 351-397. Antevs, E. (1922). The recession of the last ice sheet in New England: Amer. Geogr. Soc., Research Ser. 11, 120 pages. Athy, L. F. (1930). Density, porosity, and compaction of sedimentary rocks: Bull. Amer. Assoc. Petroleum Geol., vol. 14, pp. 1-24. Bagnold, R. A. (1941). The physics of blown sand and desert dunes: Methuen and Co., Ltd., London, 265 pages. Bailey, E. B. (1936). Sedimentation in relation to tectonics: Bull. Geol. Soc. Amer., vol. 47, pp. 1713-1726. Barrell, J. (1917). Rhythms and the measurements of geologic time: Bull. Geol. Soc. Amer., vol. 28, pp. 745-904. Baver, L. P. (1948). Soil physics: John Wiley & Sons, New York, 2nd ed., 398 pages. Beers, R. F., and Goodman, C. (1944). Distribution of radioactivity im ancient sediments: Bull. Geol. Soc. Amer., vol. 55, pp. 1229-1253. Blackwelder, W. (1940). The hardness of ice: Amer. Jour. Sci., vol. 238, pp. 61-62. Boswell, P. G. H. (1933). Mineralogy of sedimentary rocks: Thos. Murby, London, 393 pages. DH. 1] REFERENCES 37 Bradley, W. H. (1931). Origin and microfossils of the oil shale of the Green River formation of Colorado and Utah: U.S. Geol. Survey, Prof. Paper 168, 58 pages. (1948). Limnology and the Eocene lakes of the Rocky Mountain region: Bull. Geol. Soc. Amer., vol. 59, pp. 635-648. et al. (1942). Geology and biology of north Atlantic deep-sea cores be- tween Newfoundland and Ireland: U. 8S. Geol. Survey, Prof. Paper 196, 163 pages. Bramlette, M. N. (1946). The Monterey formation of California and the origin of its siliceous rocks: U. S. Geol. Survey, Prof. Paper 212, 57 pages. Brown, C. B. (editor) (1948). Proceedings of the Federal Inter-Agency Sedi- mentation Conference, Denver, Colo., May 6-8, 1947: U.S. Bureau of Reclama- tion, Washington, D. C., 314 pages. (A symposium on the effect of sedi- mentary processes on human endeavor.) Brown, T. C. (1914). Origin of oolites and oolitic texture in rocks: Bull. Geol. Soc. Amer., vol. 25, pp. 745-780. Bryan, K. (1936). The formation of pediments: Reports, Sixteenth Internationa: Geological Congress, vol. 2, pp. 765-775. Bucher, W. H. (1919). On ripples and related surface forms and their paleogeo- graphic interpretation: Amer. Jour. Sci., 4th ser., vol. 47, pp. 149-210, 241-269. Cady, G. H. (1942). Modern concepts of the physical constitution of coal: Jour. Geol., vol. 50, pp. 337-356. Capps, S. R. (1941). Observations of the rate of creep in Idaho: Amer. Jour. Sci., vol. 239, pp. 25-32. Casagrande, A. (1947). Classification and identification of soils: Proc. Amer. Soc. Civ. Engrs., June, 1947, pp. 783-810. Clarke, F. W. (1924). The data of geochemistry: U. S. Geol. Survey, Bull. 770, 841 pages. Davis, E. F. (1918). The radiolarian cherts of the Franciscan group: Bull. Univ. Calif. Publ. Dept. Geol. Sci., vol. 11, pp. 235-432. Davis, W. M. (1930). Origin of limestone caverns: Bull. Geol. Soc. Amer., vol. 41, pp. 475-628. Emery, K. O. (1948). Submarine geology of Bikini Atoll: Bull. Geol. Soc. Amer., vol. 59, pp. 855-860. Fairbridge, R. W. (1946). Submarine slumping and the location of oil bodies: Bull. Amer. Assoc. Petroleum Geol., vol. 30, pp. 84-92. Fisk, H. N. (1944). Geological investigation of the Alluvial Valley of the Lower Mississippi River: Mississippi River Commission, Vicksburg, Miss., 78 pages. Foshag, W. F. (1926). Saline lakes of the Mohave Desert region: Econ. Geol., vol. 21, pp. 56-64. Gilbert, G. K. (1890). Lake Bonneville: U. S. Geol. Survey, Mono. 1, 438 pages. Gilluly, J. (1937). Physiography of the Ajo region, Arizona: Bull. Geol. Soc. Amer., vol. 48, pp. 323-347. Goldich, S. 8. (1938). A study in rock weathering: Jour. Geol., vol. 46, pp. 17-58. Goldschmidt, V. M. (1937). The principles of distribution of chemical elements in minerals and rocks: Jour. Chem. Soc. London, 1937, pp. 655-673. Grabau, A. W. (1913). Principles of stratigraphy: A. G. Seiler and Co., New York, 1185 pages. Grim, R. E. (1942). Modern concepts of clay minerals: Jour. Geol., vol. 50, pp. 225-275. 38 TRASK. PRINCIPLES OF SEDIMENTATION [Cx. 1 Harder, E. C. (1919). Iron-depositing bacteria and their geologic relations: U.S. Geol. Survey, Prof. Paper 113, 89 pages. Hatch, F. H., Rastall, R. H., and Black, M. (1938). The petrology of the sedi- mentary rocks: George Allen and Unwin, London, 3rd ed., 383 pages. Hedberg, H. D. (1936). Gravitational compaction of clays and shales: Amer. Jour. Sct., vol. 31, pp. 241-287. Jenny, H. (1941). Factors of soil formation: McGraw-Hill Book Co., New York, 281 pages. Jones, O. T. (1938). On the evolution of a geosyncline: Proc. Geol. Soc. London, vol. 94, pp. ix—ex. Kelley, W. P. (1939). Base exchange in relation to sediments: Recent Marine Sediments, American Association of Petroleum Geologists, Tulsa, Okla., pp. 454465. Krumbein, W. C., and Pettijohn, F. J. (19388). Manual of sedimentary petrog- raphy: D. Appleton-Century Co., New York, 549 pages. Krynine, P. D. (1940). Petrology and genesis of the Third Bradford Sand: Penn. State Coll. Bull. 29, 134 pages. Lawson, A. C. (1915). The epigene profiles of the desert: Bull. Univ. Calif. Publ. Dept. Geol., vol. 9, pp. 23-48. Leith, C. K., and Mead, W. J. (1915). Metamorphic geology: Henry Holt and Co., New York, 337 pages. Loess (1945). Symposium on loess: Amer. Jour. Sci., vol. 243, pp. 225-303. Lotze, F. (1938). Steimsalz und Kalisalze: Die Wichtigsten Lagerstitten der “Nicht Hrze”: Geologie, Bd. 3, Tl. 1, Gebriider Borntraeger, Berlin, 936 pages. McKee, E. D. (1939). Some types of bedding in the Colorado River delta: Jour. Geol., vol. 47, pp. 64-81. Meinzer, O. E. (1923). The occurrence of ground water in the United States, with a discussion of principles: U. S. Geol. Survey, Water Supply Paper 489, 321 pages. Munk, W. H., and Traylor, M. A. (1947). Refraction of ocean waves—a process linking underwater topography to beach erosion: Jour. Geol., vol. 55, pp. 1-26. Murray, J., and Renard, A. F. (1891). Report on deep sea deposits based on the specimens collected during the voyage of H.M.S. Challenger in the years 1872 to 1876: Challenger reports, 525 pages. Muskat, M. (1937). The flow of homogeneous fluids through porous media: McGraw-Hill Book Co., New York, 763 pages. Paige, S. (1912). Rock-cut surfaces in the desert ranges: Jour. Geol., vol. 20, pp. 442-450. Payne, T. G. (1942). Stratigraphical analysis and environmental reconstruction: Bull. Amer. Assoc. Petrolewm Geol., vol. 26, pp. 1697-1770. Pettijohn, F. J. (1949). Sedimentary rocks: Harper and Brothers, New York, 526 pages. Polynov, B. B. (1937). The cycle of weathering (translated by Alex. Muir): Thos. Murby and Co., London, 220 pages. Reiche, Parry (1945). A survey of weathering processes and products: Univ. New Mexico Pub. Geol., No. 1, 87 pages. Revelle, R. R. D. (1944). Marine bottom samples collected in the Pacific Ocean by the Carnegie on its seventh cruise. Carnegie Inst. Washington, Publ. 556, Oceanography II, Pt. 1, pp. 1-180. Cu. 1] REFERENCES 39 Rittenhouse, G. (1944). Sources of modern sands in the middle Rio Grande Valley: Jour. Geol., vol. 52, pp. 145-183. Rubey, W. W. (1930). Lithologic studies of fine-grained Upper Cretaceous sedi- mentary rocks of the Black Hills region: U. S. Geol. Survey, Prof. Paper 165-A, pp. 1-54. (1938). Force required to move particles on a stream bed: U. S. Geol. Survey, Prof. Paper 189-E, pp. 121-141. Russell, I. C. (1885). Geological history of Lake Lahontan: U. S. Geol. Survey, Mono. 11, 288 pages. Russell, R. D. (1939). Effects of transportation upon sedimentary particles: Recent Marine Sediments, American Association of Petroleum Geologists, Tulsa, Okla., pp. 32-47. Russell, R. J. (1936). Lower Mississippi River delta: Louisiana Geol. Survey, Bull. 8, 454 pages. Rutledge, P. C. (1940). Description and identification of soil types: Proceedings Purdue Conference on Soil Mechanics, Lafayette, Ind. Sander, B. (1936). Beitrage zur Kenntniss der Anlagerungsgefiige: Tschermak. Mitt., Bd. 48, pp. 27-139. (Origin of dolomite and limestone.) Sharpe, C. F. S. (1938). Landshdes and related phenomena: Columbia Univer- sity Press, 138 pages. (Contains extensive bibliography.) Shaw, E. W. (1913). The mud lumps at the mouths of the Mississippi: U. S. Geol. Survey, Prof. Paper 85, pp. 11-27. Shepard, F. P. (1932). Sediments of the continental shelves: Bull. Geol. Soc. Amer., vol. 43, pp. 1017, 1039. (1948). Submarine geology: Harper and Brothers, New York, 348 pages. Shrock, R. R. (1948). Sequence in layered rocks: McGraw-Hill Book Co., New York, 507 pages. Sorby, H. C. (1908). On the application of quantitative methods to the study of the structure and history of rocks: Geol. Soc. London, Quart. Jour., vol. 64, pp. 171-233. Stetson, H. C. (1938). The sediments of the continental shelf off the eastern coast of the United States: Papers in Physical Oceanography and Meteorology, Mass. Inst. Techn. and Woods Hole Ocean. Inst., vol. 5, No. 4, 48 pages. Strgm, K. M. (1939). Land-locked waters and the deposition of black muds: Recent Marine Sediments, American Association of Petroleum Geologists, Tulsa, Okla., pp. 356-372. Sverdrup, H. U., Johnson, M. W., and Fleming, R. H. (1942). The oceans: their physics, chemistry and general biology: Prentice-Hall, Inc., New York, 1087 pages. Sverdrup Anniversary Volume (1948): A symposium on general oceanography, Jour. Marine Res., vol. VII, No. 3, pp. 127-686. Taylor, D. W. (1948). Fundamentals of soil mechanics: John Wiley & Sons, New York, 700 pages. Terzaghi, K., and Peck, R. B. (1948). Sozl mechanics in engineering practice: John Wiley & Sons, New York, 566 pages. Thwaites, F. T. (1941). Outline of glacial geology: Published privately, Mad- ison, Wis., 119 pages. Tolman, C. F. (1937). Ground water: McGraw-Hill Book Co., New York, 593 pages. Tomlinson, C. W. (1916). The origin of red beds: Jour. Geol., vol. 24, pp. 153- 179. 40 TRASK. PRINCIPLES OF SEDIMENTATION [Cu. 1 Trask, P. D. (1932). Origin and environment of source sediments of petroleum: Gulf Publishing Co., Houston, Texas, 323 pages. (1937). Relation of salinity to the calcium carbonate content of marine sediments: U. S. Geol. Survey, Prof. Paper 186-N, pp. 273-299. (editor) (1939). Recent marine sediments: American Association of Pe- troleum Geologists, Tulsa, Okla., 736 pages. Twenhofel, W. H. (editor) (1932). Treatise on sedimentation: Williams and Wilkins Co., Baltimore, Md., 2nd ed., 926 pages. Twenhofel, W. H. (1937). Terminology of the fine-grained mechanical sedi- ments: Report, Committee on Sedimentation, 1936-1937 (National Research Council), pp. 81-104. (1939). Principles of sedimentation: McGraw-Hill Book Co., New York, 610 pages. Udden, J. A. (1914). Mechanical composition of clastic sediments: Bull. Geol. Soc. Amer., vol. 25, pp. 655-744. Vaughan, T. W. (1910). A contribution to the geologic history of the Floridian plateau: Carnegie Inst. Washington, Pub. 133, pp. 99-185. Wadell, H. (1935). Volume, shape, and roundness of quartz pebbles: Jour. Geol., vol. 43, pp. 250-280. Waksman, 8S. A. (1936). Humus: Williams and Wilkins Co., Baltimore, Md., 494 pages. Walther, J. (1894). EHinleitung in die Geologie als historische Wissenschaft ; Lithogenesis der Gegenwart: G. Fischer, Jena, pp. 535-1055. Wattenberg, H. (1933). Kalziumkarbonat und Kohlensiuregehalt des Meeres- wassers: Wiss. Ergeb. Metecr Exped., Bd. 8, Tl. 2, pp. 1-338. Welch, P. S. (1935). Limnology: McGraw-Hill Book Co., New York, 535 pages. Weller, J. M. (1930). Cyclical sedimentation of the Pennsylvanian period and its significance: Jour. Geol., vol. 38, pp. 97-135. Wentworth, C. K. (1935). The terminology of coarse sediments: Natl. Research Council Buli. 98, pp. 225-246 (Report, Committee on Sedimentation 1932-1934). , and Williams, H. (1932). The classification and terminology of pyro- clastic rocks: Natl. Research Council, Bull. 89, pp. 19-53 (Report, Committee on Sedimentation 1930-1982). Woodford, A. O. (1925). San Onofre breccia, its nature and origin: Bull. Univ. Calif. Publ. Dept. Geol. Sci., vol. 15, pp. 159-280. Zobell, C. E. (1946). Marine microbiology: Chronica Botanica Co., Waltham, Mass., 240 pages. CHAPTER 2 ORIGIN OF SOILS Hans JENNY Division of Soils University of California Berkeley, Californa WHAT IS SOIL? To the geologist and the engineer Ramann’s definition of soil ap- pears familiar: “The soil is the upper weathering layer of the solid earth crust.”” However, Ramann’s definition fails to stress the con- cept of the soz profile. Contrary to common belief, soil is not a ran- domized aggregate of inorganic and organic particles. Soil is a body possessing definite organization. Soil has vectorial properties. Along an imaginary line extending from the surface of the soil toward the center of the earth (z-axis), the sequence of soil properties differs pro- foundly from that along lines parallel to the surface. Soils have a profile. Profile characteristics of soils may be conveniently displayed by sou property-depth functions as illustrated in Fig. 1. The zones of maxima and minima are designated as horizons. The surface soil, which has lost material to the lower strata, is labeled A horizon (Aj, Az), whereas the lower soil stratum, which has gained substances, is denoted B horizon (B,, Bz). The zone underlying the B horizons is called C horizon (Cy, Cz). It may correspond to the original rock or parent material. FACTORS IN SOIL FORMATION On the earth as a whole, there are millions of different soils or, more specifically, different soil profiles. Yet, in spite of an apparent hit- and-miss pattern of soil distribution, a certain regularity in the occur- rence of soils is discernible. It is of two kinds. First, similar types 41 42 JENNY. ORIGIN OF SOILS [Cx. 2 of soils are found in various regions. Thus the podsol profile (Fig. 1) occurs in Sweden, in Russia, in Germany, in the Alps, in New England, in Canada, in California, in equa- Horizons torial South America, and in many Al A} B; other parts of the world. Second, dis- similar soils may be grouped into (ae ee ee le sequences such that the properties of | | IAAL ano , 7 ou es = i the soils within a sequence vary in a 9 Sh bay systematic manner. se see tes The genetic and geographic rela- Spf tionships among soils may be con- a ientl d : Saya aialealc tea veniently expressed as 2 Hs Ll s = f(dl, 0, r, p, t) (1) Cogs 2ese The letter s denotes an — | y soil property sleet las EAM aia such as color, reaction (pH), clay con- a bel vesletayte alist tent, nitrogen content, or lime. The TES esreas symbol f designates “function of,” or 0 10 20 30 40 “dependent on.” The letters in paren- Depth of soil in inches : : va : theses represent the soil-forming fac- Fic. 1. Variation of soil proper- c fuse Saenclentmarl soil uiGeneel tors. The specific symbols have the profile). following significance: cl = air climate (environmental climate) 0 = species of organisms, that is, flora and fauna r = topography, including certain hydrologic features p = parent material, defined as the state of the soil at the soil formation time zero = time of soil formation (age of soil) Equation (1) is employed in two ways. First, in a qualitative sense, as a shorthand notation for stating that soils are affected by climate, organisms, topography, parent material, and time. The second mode of interpretation treats the soil-forming factors as independent vari- ables, which define the state of the soil system. This approach permits studying functions of individual soil-forming factors, as follows: Time functions or chronofunctions: = HG akan Soil properties are related to time (age of soil) under conditions of constancy of cl, 0, 7, p. Cu. 2] TIME 43 Parent material functions or lithofunctions: SiS Detar Soil properties are related to parent material under conditions of con- stancy of cl, o, 1, ¢. Topography functions or topofunctions: Sa ii Deno Soil properties are correlated with topographic and drainage features when cl, 0, p, t are constant. Climatic functions or climofunctions: o> Ti QD escanes Soil properties are related to climatic variables under conditions of constancy of 0, 7, p, ft. Organism functions or biofunctions: eS IO) ethers poet Soil properties are dependent on organic species. These functions deal with relationships between soil properties and organisms when cl, 7, p, t are held constant. The soil-forming factors cl, 0, 7, p are multiple factors and yield groups of functions. Soils have many properties: s1, Se, 83, s4, etc. All properties taken together constitute a collection, assemblage, or ensemble of properties which is the soil. If the ensemble of s values is designated by the symbol Es), Soll = es SS INGE OP 0, 1) (2) Just as each individual s property is a function of the soil-forming factors, so is the entire ensemble dependent on cl, 0, r, p, and t. In practice, the variations of the ensemble are recognized as profiles, soil types, soil series. In accordance with the five pedologie functions, the ensembles may be arranged in five sequences: chrono-, litho-, topo-, climo-, and bio-sequences. In contrast to equation (1), equation (2) is qualitative since “soil’’ cannot be assigned a single numerical value. Time (CHRONOFUNCTIONS AND CHRONOSEQUENCES) The rate of soil formation varies widely. It is often stated that it takes thousands of years to produce one inch of soil. As judged from weathering of dated buildings and tombstones, this estimate is 44 JENNY. ORIGIN OF SOILS [Cx. 2 probably correct provided that very resistant rocks such as granite, porphyry, quartzite are considered. Softer rocks, like certain sand- stones and shales, weather much more rapidly. In unconsolidated materials such as loess, sand dunes, moraines, alluvial deposits, and voleanic ash layers, visible profile development may take place in a few centuries or even decades. Examples of the chronofunction Ss = fume) ci o.7.0 are given in Fig. 2 for the s-property calcium carbonate (CaCOs). The curves refer to the leaching of CaCOs from the surface layer of English Percent CaCog ORO 00 Mm IS0m200 250 300 350 400 Age of soil in years Fic. 2. Time functions of leaching of calcium carbonate in English sand dunes (Salisbury) and Dutch clay polders (Hissink). sand dunes and Dutch clay polders (Jenny, 1946). In the cool, humid climates of England and Holland, about 300 years are required to free the surface soil of lime. An example of a chronosequence of depth functions Es) = time) cio,7,0 is presented in Fig. 3. The three soils Yolo, Zamora, and Hillgate were formed on alluvial material derived from sedimentary rocks, in a climate having mild, humid winters and hot and dry summers (Cali- fornia). The vegetation is grass. The relative age of these soils is in- ferred from the physiographic positions of the alluvial fans and ter- races. Yolo is the youngest soil; Hillgate the oldest. It may be no- ticed that the density of the subsoil (B horizon) increases as the soils become older. Cu. 2] TIME 45 Examination of numerous time functions indicates that the soil properties s have rates of change with time, ds/dt which become smaller as the age of the soil increases. Soils which have become rela- tively stabilized in relation to time are often designated as mature soils. Soil maturity does not imply complete arrest of soil develop- ment; it merely indicates relatively slow reaction rates. In climates which are not extremely dry (deserts) or cold (arctics), mature soils have well-developed profile features. The time necessary to develop 1g) = ee) Hillgate Zamora Mean apparent density i) ~s 1.6 0 10 20 30 40 50 60 Depth in inches Fig. 3. Chronosequences of soil profiles (Harradine). The age of the soil in- creases as follows: Yolo < Zamora < Hillgate. mature soils varies with the constellation of the soil-forming factors. In soft, porous, parent materials stabilized humus profiles may evolve within a few centuries. To produce claypan soils, which have B hori- zons rich in clay, probably tens of thousands of years are required. Soils are sometimes studied in accordance with the equation Evs) = f(time) (3) Here the subscripts are missing. The soil-forming factors are not kept constant. This formulation represents the historic approach. Soils are studied in relation to time, irrespective of time changes of climate, and the biotic factor. The historic approach is sometimes designated as the study of soil evolution, as it may occur during cycles involving long geological periods. In contrast, the chronofunctions deal only with those soil changes which take place under relatively constant cl, o, r, and p. 46 JENNY. ORIGIN OF SOILS [Cu. 2 PareENT MATERIAL (LITHOFUNCTIONS AND LITHOSEQUENCES) In the time functions in Fig. 2, the left end of each curve shows the value of the soil property CaCO; at the beginning of soil formation. The soil at zero time is designated as parent material. It may repre- sent consolidated or unconsolidated rock in the broadest sense of the word. Contrasting soil formation on granitic rocks with soil formation on basaltic rocks has little pedologic significance unless the magnitudes of the remaining soil-forming factors are indicated. Theoretically, the role of different parent materials in soil formation can be assessed only if cl, o, r, t are either constant or ineffective. An illustration of a parent material function or lithofunction is given in Table 1. It refers to soils derived from Winona glacial till in north- TABLE 1 LITHOFUNCTIONS OF GLACIAL SOILS IN ILLINOIS (Kellogg et al., 1949) Composition of Parent Propetties}orcel Material (Till) Type of Soil Depth of Clay CaCOs Coy gon ent leaching of of B horizon carbonates Op OF q inches Saybrook ey 27.8 <38 36.1 Elliot 25-35 28.2 38-46 32.0 Swygert 35-44 24.1 46-53 28.8 Clarence >44 20.6 >53 25.9 eastern Illinois. It is proper to assume that cl, 0, 7, t are the same for all profiles. The higher the clay content of the till, the higher is the amount of clay in the B horizon of the derived soil. The Winona glacial tills also vary in their content of calectum carbonate (CaCO3). Dur- ing soil formation, calcium carbonate is leached from the surface into lower strata. The depth of leaching is controlled by the clay content of the till. Its influence overshadows the inverse relationship of the carbonate content of the till. Cu. 2] TOPOGRAPHY 47 In humid, temperate, and cool climates, soils formed from granite are frequently coarse-textured and acid, and they often have a super- ficial humus layer of the mor type. Under similar climatic conditions, soils derived from diorite and the most basic gabbros are usually deeper and better supplied with calcium and phosphorus. They have mull types of humus layers (Kellogg et al., 1949). Owing to the interplay of soil-forming factors, the systems of soil classification based upon geologic features lack generality. They may 40 60 80 100 Feet Fic. 4. A slope sequence. Left: sketch showing deep road cuts through chaparral- covered hills of “soft” quartz diorite. Right: exposed soil mantle illustrating depth of soil in relation to slope. be valuable in regions of relatively uniform climate and constancy of other soil-forming factors. If comparisons of soils of entire conti- nents are attempted, the geologic systems do not portray the soil rela- tionships correctly. TopoGRAPHY (TOPOFUNCTIONS AND TOPOSEQUENCES) The topography or relief factor is complex, for it includes, in addi- tion to degree of slope, length of slope, shape of slope, and exposure, certain hydrologic features commonly referred to as drainage. A pure slope sequence, in absence of ground-water influences, is depicted schematically in Fig. 4. On soft quartz diorite, south of San Francisco, California, the dark soil mantle varies in thickness in relation to slope features. These variations in soil depth are the re- sult of erosion, soil creep, seepage, ete., all being functions of slope. Slope sequences become especially marked when capillary rise from ground-water tables influences the soil profile. In humid regions, tem- 48 JENNY. ORIGIN OF SOILS [Cu. 2 porary or permanent ground-water influences produce peat and bog conditions. In arid regions, alkali soils and saline soils may result. Figure 5 illustrates such a hydrosequence in the semi-arid Coalinga area of California. On an expansive, gently sloping, colluvial fan, consisting of outwash from softly consolidated calcareous sandstone and shales, all soil-forming factors—save one—are constant. This Fic. 5. Schematic illustration of a toposequence of soils influenced by ground water in arid California. variable factor is represented by the plumb distance from the surface of the soil to the temporary water table which forms during the rainy season. As a result of capillary rise of the salty ground water, the uniform fan material became differentiated into Panoche soil, Oxalis soil, and Levis soil. The Panoche soil, lying on the upper portions of the fan, is nearly free of alkali. The wet end of this hydrologic se- quence, represented by the Levis soil, is strongly impregnated with salt. The Oxalis soil, situated between the two extremes, possesses a nearly salt-free surface soil but has a slight to moderate salt content in the subsoil. Combinations of slope and ground-water sequences are often classi- fied as catenas (Bushnell, 1944). CLIMATE (CLIMOFUNCTIONS AND CLIMOSEQUENCES) Hilgard (1912) in this country and Dokuchaev in Russia showed that soils derived from the same parent material may have widely different properties, depending on the climate in which the soils are formed. Hilgard’s comparison of chemical analyses of soils of arid and humid regions (Table 2) has become a classic. In general, soils from arid regions contain more acid-soluble materials than soils from humid regions. The differences are especially pronounced for Ca, Mg, K, and Na. The fundamental difference in mode of formation of soils of arid and humid regions is conditioned by the moisture regime. In regions of low rainfall, water penetrates the soil to a limited depth only; weathering and soil formation do take place, but the products are not Cu. 2] CLIMATE 49 TABLE 2 ANALYSIS OF SOILS FROM ARID AND Humip REGIONS (Hilgard, 1912) Hydrochloric Acid Extracts No. of Total Region iss Soluble | SiO, | Al,O3 | FexO3} CaO | MgO | KeO | NaoO Analyses ; Material % % % % 0 % % % Arid 573 30.84 | 6.71 | 7.21 | 5.4 1.43 | 1.27 | 0.67 | 0.35 Humid 696 15.83 | 4.04 | 3.66 | 3.88 | 0.138 | 0.29 | 0.21 | 0.14 removed from the profile. The soils remain neutral or slightly alka- line. The bulk of the water returns to the atmosphere by evaporation and transpiration, the latter mode predominating. Contrary to many beliefs, this water regime will not produce alkali or saline soils. As is discussed above, special ground-water conditions are necessary for their development. Under high rainfall, water percolates through the soil profile and finally finds its way into rivers, lakes, and oceans. Soluble and dis- persible substances are continuously removed from the soil. Soils in humid regions tend to become leached and acid. With respect to the climate function, the complex factor cl may be conveniently split into a moisture variable m, and a temperature variable T, both being treated, mathematically, as varying independ- ently of each other. Accordingly, one may speak of soil property- moisture functions and soil property-temperature functions: Si GD epee SS ILO) asc teepof Good illustrations of s-m functions have been reported from the cen- tral portion of the Great Plains area, especially the states of Kansas and Nebraska. Here one may select localities having nearly identical mean annual temperatures but considerable variations in mean annual rainfall. Vegetation consists of grass communities; the parent mate- rial is loess and related wind-blown materials. By selecting samples from level ridges, topography also may be kept constant. Figure 6, based on analyses by Alway (Jenny, 1941), portrays the variation of the calcium content of virgin soil profiles. As one pro- 50 JENNY. ORIGIN OF SOILS [Cu. 2 ceeds from the semi-arid to the semi-humid zone, soil caletum declines in exponential fashion. Figure 7 shows the trend of total soil nitrogen along the 11° C. an- Percent CaO 0 18 20 22 24 26 28 30 32 Rainfall in inches Fic. 6. Variation of calcium content of soil with rainfall in Nebraska. The total CaO content is the sum of the acid-soluble and acid-insoluble CaO content. nual isotherm which lies slightly south of Alway’s regions. Each dot represents the total nitrogen content of a soil sample taken to a depth of 10 inches. As mean annual rainfall increases, soil nitrogen also in- creases. Since soil nitrogen and organic carbon are closely related, 0.25 0.20 2015 Cc g 2 0.10}-- 0.05 0 0 10 20 30 40 Rainfall in inches Peet el ont Ente eet Peele | 0 25 50 75 100 Rainfall in cm. Fic. 7. Increase in soil nitrogen (and organic matter) with increasing mean annual rainfall in the Great Plains area. the curve also indicates the trend of soil humus. Multiplying the ni: trogen percentage by 20 gives the approximate humus percentage o: the soils in this region. Cu. 2] CLIMATE 51 Whereas the declining content of calcium must be interpreted as a result of leaching, the increasing nitrogen content must be attributed to the greater production of vegetation organic matter, especially roots, as a consequence of increased precipitation. The Great Plains area also lends itself to the evaluation of s-T functions. Figure 8 shows the decrease of the total nitrogen content of the surface soils with increasing annual temperature. The samples e Bottom-land soils © Terrace soils Percent N Missouri Arkansas 40 50 60 70 Annual temperature in degrees F. Fic. 8. Soil nitrogen (and organic matter) temperature relationships in the central United States. were collected along a transect having fairly uniform annual moisture conditions and extending from Canada to the Gulf of Mexico. In the Appalachian mountain chain of the eastern United States, relationships between average clay content of soils (to a depth of 40 inches) and latitude (or annual temperature) have been reported (Jenny, 1941). Figure 9 illustrates the increase in soil clay with ris- ing annual temperatures for soils derived from basic igneous rocks, mainly diorite and gabbro. This transect extends from southern New Jersey to Georgia. The annual rainfall varies from about 40 to 50 inches. Each dot represents one soil profile. Not only does the amount of clay vary with temperature, but so does the chemical composition of the clay. In the northern portion of the above-mentioned belt, the silica-alumina ratio of the soil clay is 52 JENNY. ORIGIN OF SOILS [Cu. 2 greater than 2.0; in the southern portion it is considerably below 2.0. A few decades ago the reaction against geologic concepts in soil formation was extreme, and many attempts were made to classify soils solely according to climate. Although such systems have many at- tractive features, the oversimplification 60 leads to gross misrepresentations. The interplay of all soil-forming factors can- not be ignored. ($y) ie) Brotic Factor (BIOoFUNCTIONS AND Bio- SEQUENCES) a oO In soil investigations, the biotic factor is usually restricted to aspects of vegeta- tion. The vegetational factor refers to kinds of species of plants (flora) and not to the abundance and yields of plants. The latter aspect is a dependent variable, ie) (oe) Clay content of soil in percent (0%) (o) 10 being itself conditioned by soil and en- 0 vironment. 8 10 12 14 16 18 In nature it is difficult to evaluate the Annual temperature : , 4 in degrees C. role of vegetation on soil formation. It Fic. 9. The average clay is necessary to locate soils which carry content of soils derived from different kinds of plants, but which, at basi¢ igneous rock Increases the same time, have identical conditions from north to south (eastern ; i Waited States) of climate, parent material, topography, and time. An extensive region which sat- isfies these conditions is found in the Middle West, more specifically, the prairie-timber transition zone. Accurate comparison of prairie and forest soils shows that forests tend to accelerate soil formation. The soils are more acid, lime is leached to greater depth, and trans- location of clay is enhanced. PROCESSES OF SOIL FORMATION The functional relationships between soil properties and soil-form- ing factors hitherto discussed are formalistic in nature. They record observed dependencies among variables. They are not concerned with mechanisms of soil formation, and they are not based on physical, chemical, or biological theories. In contrast, the elucidation of processes of soil formation requires the application of knowledge and concepts developed by the basic sciences. Cu. 2] NITROGEN AND ORGANIC MATTER IN PROFILES 53 NITROGEN AND OrGANIC MATTER IN PROFILES Soluble nitrogen compounds in the rain water increase the nitrogen content of the soils to the extent of a few pounds of nitrogen per acre per year. More important is the fixation of atmospheric ni- trogen by soil bacteria, living either non-symbiotically (Azotobacter group and Clostridium group), or symbiotically in association with leg- umes (nodule bacteria). The contribution of biological fixation of atmospheric nitrogen may amount to a hundred pounds or more of ni- trogen per acre per year. The total amount of nitrogen and organic matter * in soils assumes substantial magnitudes (see Table 3). Its rate of accumulation is TABLE 3 Toran Amounts or NITROGEN AND OrGANIC MarTeR IN SELECTED PROFILES Depth of : : Organic Type of Soil Profiles Nitrogen Matter inches | lb. per acre | lb. per acre Grassland soil Cultivated (Yolo soil, California) 60 10, 700 180,400 Pastured (Cayucos soil, California) 36 8,880 147 ,000 Forest soil Under oak (Shaver Lake, California) 50 5,650 104,000 Under pine (Shaver Lake, California) 50 5,800 154,500 Tropical forest soil (Chinchiné, Colombia, S. A.) 50 31,400 404 ,000 Soil from tropical rain forest (Calima, Colombia, 8S. A.) 30 22,400 328 , 000 conditioned by the rate of addition of organic matter by vegetation and by the rate of decomposition by soil microorganisms. Under con- ditions of relatively constant vegetation, a quasi equilibrium of soil organic matter is reached in a few centuries. The annual rate of de- composition is then nearly equal to the annual addition of vegetative material. It is estimated that annual decomposition rates of soil or- * Humus represents the dark fraction of organic matter (Waksman, 1938). It consists of compounds synthesized by microorganisms from dead plant remains. As there is no standardized method of determining humus, soil scientists prefer to report total organic matter, obtained by multiplying organic carbon by the _ factor 1.742. 54 genNY. ORIGIN OF SOILS [Cu. 2 ganic matter amount to a few percent. In other words, in a stabilized organic matter profile annually 1 to 2 percent of the soil organic matter is lost and, eo zpso, replenished by decay of organisms. In grassland soils the bulk of soil organic matter is derived from the decomposition of the root system. In forest soils a considerable portion may be acquired by infiltration of humus from decomposing leaf layers lying on top of the mineral soil. PEDOCALS AND THE FoRMATION OF Lime Horizons Many soils in arid and semi-arid regions are characterized by lime horizons. Throughout the profile there are seams and nests of lime concretions, the individual concretions varying in size from pinheads to pea and nut size. The number of concretions per unit soil mass is greatest in the subsoil (lime horizon). Chemical analyses of such soil profiles reveal a high CaO content of the soil in the lime horizon and relatively low contents in the horizons above and below. Such soils - were designated by Marbut as pedocals. If a uniform parent material containing some calcium carbonate is assumed, the formation of the lime horizon may be visualized as the consequence of calcium carbonate-bicarbonate equilibria which are regulated by the carbon dioxide pressure of soil air. Root respiration and decay of vegetable matter which are very active in the surface soil produce large amounts of carbon dioxide. It con- verts the relatively insoluble calcium carbonate to the much more soluble calcium bicarbonate. Percolating rain water translocates the bicarbonates from the surface soil to the subsoil. There, owing to reduced COz pressure of the soil air, which is the result of a low biologi- cal activity, calcium carbonate is precipitated as lime concretions. In areas of low annual rainfall the carbonate horizon is close to the surface. As annual rainfall increases, the lime horizon moves to greater depth and, finally, above 40 inches of rainfall—in the tem- perate region—completely disappears from the soil profile. On uplands and high terraces, lime horizons will develop only if the parent material is high in bases. Thus, in semi-arid California, soils derived from basic igneous rocks frequently possess calcareous sub- soils; but soils derived from acid igneous rocks (for example, granite) rarely do. Likewise, non-caleareous sandstones do not produce cal- careous profiles. In these soils Ca exists as Ca-clay rather than CaCOs. On the other hand, soils of arid regions which are under the influ- Cu. 2] WEATHERING AND CLAY FORMATION 55 ence of ground water usually contain carbonates regardless of the nature of the parent material. WEATHERING AND CLAY ForRMATION Many of the common soil-forming minerals, such as feldspars, micas, pyroxenes, consist of chains and networks of tetrahedra and octahedra whose corners are occupied by O~ and OH” ions. The small inter- stices in the centers of the tetrahedra are occupied by Sitt** or Al+++ ions. Inside the octahedra are located Alt**, Mgtt, Fett+, and Fet* ions. These negatively charged oxygen and hydroxyl polyhedra share corners and edges, and they are balanced and held together by positive cations, especially KT, Nat, Catt, Mgt™. Whereas the interior of any crystal is in electrical equilibrium (Paul- ing’s rules), the surfaces of many crystals are composed of ions whose valences are not completely satisfied. For an orthoclase crystal, which consists of joined Si- and Al-tetrahedra and K ions in intertetrahedral cavities, the surface may be schematically depicted as in Fig. 10 (left side). H20 H20 HOH H H KOH y y Hf Onno” —0-Si-O-Al-O-K-O-— —> —0O-Si-O-Al-O-H-0O- OY ———————— Crystal surface Crystal surface Fic. 10. Schematic presentation of orthoclase surface reacting with water. Hydration of oxygen ions not shown. Upon the addition of water two reactions may occur. First, hydra- tion, whereby water molecules (dipoles) are attracted to the unsatis- fied valences of exposed Si and Al ions. The polarization of the at- tracted water molecules may become so strong that some of the H ions are expelled. They may become attached to exposed O ions which are thus converted to OH ions. The exposed polarizing Si and Al ions become surrounded by OH ions (water molecules minus H). The second reaction, proceeding simultaneously and independently of hydration, consists of an ion exchange (hydrolysis) between ex- posed K ions of the lattice and H ions of water, as follows: [Orthoclase] Kt + HtOH- = Ht + K+OH~ The liquid phase acquires alkalinity (pH 9-11), and the crystal sur- face gains H ions, which tend to combine with O= to form OH-. As a result of hydration and hydrolysis, the exposed oxygen tetra- hedra become partial hydroxyl tetrahedra. Aluminum tends to attract 56 JENNY. ORIGIN OF SOILS [Cu. 2 further OH ions to assume its preferred octahedral configuration of hydroxyl ions. Coupled with the absence of stabilizing intertetra- hedral K ions, the surface layer becomes unstable and polyhedra peel off. Tetra- and octahedra liberated from feldspars and other minerals aggregate among themselves to form clusters of colloidal size, namely, colloidal silica, colloidal aluminum hydroxide, and mixtures, the col- loidal alumino-silicates. Although young colloidal particles are prob- ably amorphous, upon aging the polyhedra orient themselves to defi- nite crystal lattices such as cristobalite, diaspore, bauxite, goethite, gibbsite (hydrargillite), clay minerals of the montmorillonite-beidel- lite-nontronite group, clay minerals of the hydrous mica-illite-vermicu- lite group, and clay minerals of the kaolinite-halloysite group. (Com- pare Chapter 25.) Fe ions of the original minerals tend to be excluded from incorporation into clay particles. They form the stable oxyhy- droxides, such as limonite, goethite, and hematite. ; It does not appear necessary that primary minerals undergo com- plete breakdown into individual polyhedra. Fragments of chains and sheets of tetra- and octahedra may recombine. Sometimes mere ionic substitutions bring about fundamental alterations, such as the conver- sion of biotite into vermiculite. The specific conditions which control the formation of the various clay minerals in soils are not well known. Long ago Mattson showed that the $10.:Al,03 ratio of colloidal alumino-silicates is influenced by the pH of the medium in which they are formed. As the reaction changes from alkaline to acid, the ratio increases. In general, as stated by Ross and Hendricks (1945), “Alkaline feldspars and the micas tend to alter to kaolin minerals, whereas ferromagnesian minerals, cal- cic feldspars, and volcanic glasses commonly alter to members of the montmorillonite group.” The roles played by climate, time of weath- ering, vegetation, and drainage conditions have not yet been eluci- dated. FORMATION OF CLAYPAN SOILS AND THE MIGRATION OF COLLOIDAL CLAY PARTICLES Numerous soils exhibit accumulation of colloidal clay particles in the B horizon (Fig. 1). Extreme cases of clay accumulation produce soils known as claypan soils. Their B horizons may be so tight and sticky that they are nearly impermeable to water and air. The source of clay in the B horizon is twofold. First, clay is formed Cu. 2] FORMATION OF CLAYPAN SOILS 57 in place as a result of weathering. It is possible that the subsoil region is especially conducive to clay formation, owing to favorable moisture conditions. Second, clay accumulates in the subsoil as a result of downward migration of colloidal clay particles from the surface soil. The downward migration of clay is governed by colloid chemical principles, especially dispersion and flocculation. Colloidal clays such as montmorillonite clays tend to form stable suspensions, provided that free electrolytes (salts, acids, bases) are absent. Such clays are said to be highly dispersed. Their particle sizes are very small—of the order of a few hundred Angstrom units (1 Angstrom unit = 10-8 centi- meters). Addition of suitable amounts of electrolytes to dispersed clay sys- tems will produce flocculation. When observed under the ultramicro- scope, it is seen that the tiny individual clay particles unite to form ageregates or flocks which may become so large that they settle read- ily under the influence of gravity. The flocculation of clays by mono- and divalent cations is usually reversible. Removal of excess electro- lyte will restore the system to its dispersed state. The phenomenon of protective action of humus also must be taken into consideration. Leaching a soil with dilute ammonium hydroxide yields a dark-brown extract which contains colloidal humus particles. This humus extract possesses the power of protective action. If a small amount of colloidal humus extract is added to a dispersed clay system, its flocculation value becomes higher. In other words, a higher amount of electrolyte must be added to produce clay floccula- tion. Conversely, the addition of colloidal humus to a flocculated clay often results in dispersion of clay. The clay aggregates separate into the ultimate clay particles. These aspects of colloidal chemistry aid in the understanding of clay migration. Let us postulate a uniform parent material of medium texture containing, say, 10 to 20 percent of clay, and 5 to 10 percent of calcium carbonate. Loess and many alluvial deposits closely corre- spond to such a hypothetical parent material. The climate is assumed to be humid. Owing to the presence of calcium carbonate and bicarbonate, the clay exists in the flocculated form. It is in a state of rest. As calcium bicarbonate is being leached downward, the surface soil’s electrolyte concentration is reduced below the flocculation value of the clay. Aided by the protective action of soil humus, the clay aggregates be- gin to disperse. The fine individual particles are carried by the perco- lating water to the subsoil. 58 JENNY. ORIGIN OF SOILS [Cu. 2 In the lower part of the subsoil, the high electrolyte concentration reflocculates the clay particles. As colloids settle out, the zone of illuviation becomes denser, the pores smaller, and additional migrating clay particles may be retained by mere sieve action, even in absence of excess electrolyte. The surface soil becomes depleted of clay, and the accumulation zone (B horizon) becomes thicker. The resulting extremely slow passage of water preserves the claypan over very long periods of time. LATERITIC Soits AND LATERIZATION More than 100 years ago, in 1830, Buchanan described the red soils of India which are used locally for making bricks (“later”). In ensuing years the word “laterite” was applied to all red soils occurring in tropical regions; in fact, sometimes to all red soils anywhere. In recent years, especially under the influence of Pendleton, the concept of laterite has tended to be restricted to specific soil strata rich in © sesquioxides, possibly formed under the influence of ground-water conditions. The literature on laterites and lateritic soils is very voluminous. Among pedologists the prevailing viewpoint stresses laterization as a widespread soil-forming process of humid regions in which silica and bases are lost from the soil profile. However, only in extreme cases would laterization produce an actual laterite as defined by Buchanan and Pendleton. According to Wiegner (1926), laterization is the direct result of normal weathering of rocks in absence of acid humus. In accordance with the ideas on rock decay presented in a previous section, one may write schematically colloidal silica /\\ i + colloidal aluminum hydroxide colloidal iron hydroxide . Feldspar | + HOH KOH, NaOH, Ca (OH)>- (high pH) Since colloidal silica is negatively charged, alkaline reaction produced by the free bases will disperse it, and silica will leave the soil profile. Often silica is found precipitated in lower soil strata as chalcedony. Iron and aluminum hydroxides form positively charged colloids, and they are flocculated at high pH. Accordingly, the sesquioxides remain in the soil. The preferential leaching of silica tends to limit the Cu. 2] PODSOLS AND PODSOLIZATION 59 formation of clays to those having narrow silica-alumina ratios. Ac- cordingly, laterization is characterized by preferential accumulation of sesquioxides and -hydroxides (gibbsite, limonite) and kaolinitic clays. Soil colloids extracted from lateritic soils have silica- alumina ratios of 2 and less, as illustrated in Fig. 11. According to Wiegner, lat- erization takes place, on a minute scale, in frigid zones, above the zone of vegetative growth. Its prevalence in trop- ical regions may be the result of a combination of warm hu- mid climate, which favors hy- drolysis and leaching, and length of weathering periods Silica-alumina ratio 0 Dep at soil in inches embracing the entire Pleisto- Fic. 11. Illustration of lateritic and podsolic weathering and soil formation. cene and a considerable part ‘i nae: P ae Silica-alumina ratios of clay colloids iso- of the Tertiary. Laterization lated from various soil horizons. should be especially preva- lent in deep horizons where infiltration of humic acids is neg- ligible. PoDSOLS AND PODSOLIZATION The podsol profile (Fig. 1) occurs under a great variety of environ- mental conditions in cold, temperate, and tropical regions. It consists of the following horizons: Ao: <1 to 5 inches. Partially decomposed forest litter and leaf mold. A,: 1 to 6 inches thick. Dark brown humus layer (raw humus) mixed with mineral soil; strongly acidic (pH 4). A»: 1 to 8 inches thick. Bleached horizon, grayish white; relatively rich in silica. By, By: 2 to 15 inches thick. Rust-brown horizon with accumulations of iron hydroxide and some humus; also known as ortstein. C: parent material, usually of sandy or loamy texture. The characteristic chemical feature of the-podsol profile is displayed by the trend of the silica-alumina ratio of the colloidal clay fraction isolated from various horizons (Fig. 11). The curve reaches a maxi- mum in the A», horizon and a minimum in the B horizon. 60 JENNY. ORIGIN OF SOILS [Cu. 2 To comprehend the formation of a podsol profile, or, in other words, the process of podsolization, we may resort to ideas expressed by Wiegner (1926). He stresses the role played by the acid humus col- loids in bringing about the reversal of the process of laterization. Wiegner reasons as follows: From the thick organic Ag and A, hori- zons, acid humus colloids enter the mineral soil and neutralize the bases as rapidly as they are being formed. This important step shifts the weathering process from lateritic to podsolic. The soil solution, being acid, flocculates negative colloidal silica, which thus remains in the surface soil. Acidity disperses positive colloidal aluminum and iron hydroxides, a process which is encouraged by the protective action of the acid humus colloids. Accordingly, colloidal iron and alumina are removed from the surface soil. They are reprecipitated in the sub- soil (B horizon) where electrolyte concentration and pH are relatively high. In contrast to mere clay migration, as it occurs in claypan soils, podsolization involves a differential behavior of colloidal silica and colloidal sesquioxides. Colloids extracted from A, horizons have high silica-alumina ratios. Those in the B horizon have low ones (Fig. 11). Podsol formation postulates the presence of acid colloidal humus. Accordingly, it can occur wherever large amounts of acid humus are formed, be it as a result of low temperature, as in northern regions, or as a result of high rainfall, as in certain parts of the tropics. REFERENCES Bushnell, T. M. (1944). The story of Indiana soils: Purdue Univ. Spec. Cire. 1. Harradine, F. F. (1949). The variability of soil properties in relation to stage of profile development: thesis, University of California. Haseman, J. F., and Marshall, C. E. (1945). The use of heavy minerals in studies of origin and development of soils: Missouri Agr. Exp. Sta. Res. Bull. 387. Hilgard, E. W. (1912). Soils: The Macmillan Co., New York, 593 pages. Jenny, H. (1941). Factors of soil formation: McGraw-Hill Book Co., New York, 281 pages. (1946). Arrangement of soil series according to functions of soil forming factors: Soil Sczence, vol. 61, pp. 375-391. Joffe, J. S. (1949). Pedology: Rutgers University Press, New Brunswick, N. J., 575 pages. Kellogg, C. E., et al. (1949). Soil classification: Sozl Science, vol. 67, pp. 77-191. Lutz, H. J., and Chandler, R. F. (1946). Forest soils: John Wiley & Sons, New York, 514 pages. Marbut, C. F. (1935). Soils of the United States: Atlas of American Agriculture, U.S. Department of Agriculture, Washington, D. C., vol. III. Mohr, E. C. J. (1944). The soils of equatorial regions: Edwards Bros., Ann Arbor, Mich., 766 pages. Cu. 2] REFERENCES 61 Ross, C. S., and Hendricks, S. B. (1945). Minerals of the montmorillonite group: U. S. Geol. Survey, Prof. Paper 205-B, 79 pages. Sigmund, A. A. J. (19388). The principles of soil science: Thos. Murby, London, 362 pages. Waksman, S. A. (1938). Humus: Williams and Wilkins Co., Baltimore, Md., 526 pages. Wiegner, S. (1926). Boden und Bodenbildung: Th. Steinkopff, Dresden. Wilde, S. A. (1946). Forest soils and forest growth: Chronica Botanica Co., Waltham, Mass., 241 pages. Yearbook of Agriculture (1938). Sozls and Men: U.S. Department of Agricul- ture, Washington, D. C., 1232 pages. CHAPTER 3 THE LAWS OF SEDIMENT TRANSPORTATION H. A. EINSTEIN AND J. W. JOHNSON Division of Mechanical Engineering Unwersity of California Berkeley, California It appears basically impossible to separate the description of sedi- ment transport from that of erosion and deposition of the same par- ticles, since erosion represents the initiation, and deposition the termi- nation, of any sediment motion. However, if the term sediment trans- portation laws is used in the specific sense, that is, restricting the de- scription of the relationships which link the rate of sediment motion at any flow section to the parameters of the flow and its boundaries, this separation is fully justified. It is in this sense that the transportation laws may be defined for the purpose of this chapter. It has been shown in Chapter 1 in this symposium that sediment particles may differ in many respects, such as size, shape, specific grav- ity, roundness, mineralogical composition, to mention only a few. For the purpose of transportation studies, a knowledge of the size, settling velocity, and specific gravity of a grain is sufficient. Even at that, this may represent grain types which behave in three different man- ners.* The realization that each sediment mixture may contain all these different grains at any possible frequency makes it clear that a sediment mixture may be a very complex unit. It is not surprising that under these conditions sediment mixtures cannot always be treated as a unit, but that some part of a mixture may be moved in the same flow according to entirely different laws than is some other part. * The terminology proposed by the American Geophysical Union (see Transac- tions, Vol. 28, Dec. 1947, p. 936) has been adopted in this chapter. 62 Cu. 3] WASH LOAD 63 TYPES OF LOAD TRANSPORTED BY FLOWING WATER In attempting to describe the various relationships between the sediment rate and the flow of the fluid which produces the transport of the material, it is always important to remember that every par- ticle moving through a reach of a stream must satisfy two conditions: (1) It must have been eroded or otherwise have been made available in the watershed upstream from the reach. (2) It must have been moved down to the reach and through it by the flow. The rate at which each kind and size of particle is moving, therefore, is limited to its actual rate by either the first or the second condition. Needless to say, the laws derived from the two conditions restricting the rate of flow are inherently different. It has also proved to be helpful to introduce different terms for the two parts of the sediment load of a stream; thus that part of the load the rate of which is governed by its availability in the watershed is termed “wash load,” and that part which is governed mostly by its ability to be moved in the stream channel is termed “bed-material load.” Wasu Loap The rate at which the various sizes of the wash load of a stream are moving through a reach depends, according to definition, only on the rate with which these particles become available in the watershed and not on the ability of the flow to transport them. This may be inter- preted to mean that, if the rate of flow is greater, more particles of given size can be transported if they are available. From this, again, one may conclude that wash-load particles are not deposited in the stream channels on their way from the place of erosion down to a point of measurement. They travel with the same velocity as the flow. The rate of transport of the wash load may be found to depend upon the different factors determining erosion, such as intensity of precipitation, rain-drop size, soil type, vegetal cover, and previous soil conditions, but not upon the discharge at the reference section, as has been shown by Vetter (1937). As these different parameters effecting erosion are very complex, and as the value of each parameter is not known for most of the larger watersheds, the rates of wash-load transport usually have not been analyzed in detail, but are given as an average annual load as determined from either lake surveys (Eakin, 1936) or sus- pended-load sampling programs. As the supply of the wash load never reaches the capacity of the channel to transport it, it is never deposited in the main flow channel, 64 EINSTEIN-JOHNSON. SHDIMENT TRANSPORT [Cu. 3 and its rate, therefore, has no significance in determining the stability of the stream channel itself. As it represents usually a large part of the total load (90 percent in the Rhine River above Lake Constance), and as it is usually deposited at a lower volume-weight than the bed- material load (Iowa University, 1943), it is the most important part of any lake or reservoir deposit. Several government agencies which are interested in the manage- ment of reservoirs, or in the rates of erosion, are today engaged in a very active program of sediment sampling. These agencies use suspended- load samplers for this purpose and thus sample the total load; thus some of the bed-material load may be included in estimates of the sediment load. All information so obtained is assembled by the Inter- Agency Committee on Sedimentation, which meets regularly in Wash- ington, D. C., and which is responsible for the publication of the data. With this committee serving as a clearing house, duplication of effort within the government service is eliminated. Some attempts have been made to map the sediment productivity of different areas in the country (Brown, 1945), but it is well known today that even the rate of wash load (measured in tons per year per square mile of drainage area) varies inversely with increasing size of the watershed area. The reason for this inverse relationship may be found in the deposition of wash-load material in overbank and dead- water areas. Brp-MatTertaL Loap Basically the movement of bed-material load behaves differently from that of the wash load of a stream. In the past it has been fur- nished, and usually it still is being furnished, by the watershed at a rate that is higher than the capacity of the channel to transport the material. A certain percentage of this load has been and is still being deposited along the channel, reducing the actual transport down to the capacity load. If for any reason the sediment inflow into a given reach of a channel is smaller than the capacity load, the flow imme- diately scours some of the formerly deposited sediment from the bed and keeps the load constantly at capacity level. Bed-material load is thus always transported to capacity. For a given channel reach, the flow conditions usually can be described in terms of the stage or of the discharge. It must be expected, therefore, that the capacity to transport bed material can be given also in terms of the discharge. In this respect the relationship which gives the capacity of the stream to transport the various sediment sizes of the bed material at different flows is termed the “sediment function” (Fig. 1). Cu. 3] BED-MATERIAL LOAD et 4 puodes Jad 399} D1qnd ul as1eYydsIG 10° 10° 10° 10* 10° 10° 10 Transport in tons per day Bed-load function, Rio Grande at Bernardo, New Mexico. 1. Fic. 65 66 EINSTEIN-JOHNSON. SHDIMENT TRANSPORT [Cu. 3 In an existing channel the sediment function may be determined ex- perimentally. Instruments have been devised to measure that part of the bed-material load which moves along the bed as bed load or surface creep (Iowa University, 1940b); normal suspended-load sam- pling instruments and procedures (Iowa University, 1940a, 1948) may be used to measure that part of the bed material which moves sus- pended in the upper layers of the stream cross section. The total rate of these relatively large particles has been found to be actually a well-defined function of the discharge (Einstein, 1947), and this rela- tionship may well be used to predict future or past bed-material motion in this same channel if the flow rates are known. Whenever a reach of channel which has a sediment function, and which is usually called an alluvial reach, shows a low or zero rate of deposition or scour, the sediment function may be interpreted over a given length of time to determine the total or average bed-sediment supply of the watershed above. Basically, the sediment supply is naturally the primary factor determining the behavior of the stream, and in the course of centuries the stream channel has been built up by sediment deposits until it has finally become able to move the bed sediment at the rate of supply. The rate of transportation of load, therefore, may be used as an indicator of the sediment supply. If the different types of bed sediment move in a given channel at rates which are a well-defined function of the discharge, it must be possible to determine this relationship analytically. Attempts in this direction date back to the last century, when the first bed-load equa- tions were developed by DuBoys (1879). A bed-load equation, in contrast to what is termed a bed-load function herein, is a local rela- tionship between the rate of sediment motion per unit of width and unit of time and the local bed and flow conditions. In such instances the bed usually is described by an average or representative grain size. The flow in any vertical section, both the average and the ve- locity distribution, is defined basically by the local shear stress and by the total water depth. In the usual cases of bed-material load, where most of the load moves near the bottom, the flow velocities may be derived from the local shear stress alone. All the early bed-load formulas, and most of those that are used today, describe the flow by its shear stress only. In some of these formulas one or two coefficients have been introduced which may change in value somewhat for different depths and sediment sizes (Vanoni, 1947). Recent attempts to eliminate the necessity of using these variable coefficients have led to an interpretation whereby the bed-load equation is used only to express the motion of bed material Cu. 3] BED-MATERIAL LOAD 67 as surface creep in a very narrow layer above the bed, whereas all the bed material moving in higher layers of flow is interpreted as suspended load. The existing suspended load theory (Vanoni, 1946) permits caleu- lation of the sediment concentration at any point throughout the depth of a given flow, provided that the concentration at only one point in the depth is known. It has been found that the surface creep in the narrow layer which it occupies near the bed defines such a reference concentration. Conditions are encountered, especially in large streams with comparatively fine bed sediment, where the sedi- ment goes into suspension easily; hence the rate of suspended-load transportation is many times larger than that of surface creep of the same bed particles. In these cases the material moved by surface creep may be negligible in amount in itself, but it still retains its large significance as a valve or trigger controlling all suspended-load concentrations. All bed-load equations that are in existence have been derived from flume measurements. Experimental flumes that have been used for this purpose have ranged from a few inches to 7 feet in width and depth and from 5 to 200 feet in length. Most equations were derived originally from experiments with uniform sediment and then checked for applicability to sediment mixtures. Most mixtures for which the range of particle sizes is not excessive and which do not contain a large percentage of particles finer than 0.2 millimeter have been found to move as a unit. This finding indicates that the mechanical com- position of the bed sediment is about the same as that of the material being transported. If very small rates of transport are included in the experiment, or if one of the two limitations given above is trans- gressed, a considerable segregation of the particles becomes apparent, and an overall appraisal of the total transport by means of a repre- sentative grain diameter is impossible. Thus it appears to be possible that the normal bed-load equations can be applied to the different components of a bed if the basic as- sumption is made that every particle moves according to a law that includes its own availability, its ability to move, and the capacity of the flow to move it. It appears that this law does not depend to a large degree on the presence of other sediment particles. This same assumption of non-interference between sediment particles in motion is also one of the basic assumptions of the existing suspended-load theory (Vanoni, 1946). The composition of the bed, however, deter- - mines the bed roughness and, indirectly, the flow pattern. Publica- 68 EINSTEIN-JOHNSON. SHDIMENT TRANSPORT [Cu. 3 tions on this type of bed-load equation may be expected to appear in the next few years. PRACTICAL APPLICATION OF BED-LOAD EQUATIONS After a bed-load equation has been developed from flume experi- ments, it is very important to determine its applicability to natural streams by making measurements in such streams. Such measure- ments of sediment load should combine bed-load measurements (Ein- stein, 1944) with a suspended-load-measuring program in order to determine the bed-load function of the stream. For a comparison it is necessary, therefore, to integrate the specific load as obtained from the bed-load equation over the entire cross section. The inte- eration of the transport over the cross section may be made on a basis - of either local or average specific load; that is, either the hydraulic conditions from which the load is calculated may be averaged and the load calculated from this average flow, or the load may be calcu- lated locally and then integrated over the cross section. Both methods lead to practical difficulties and call for various assumptions to be made. At the present stage of development in the field of sediment transportation, it is not clear in which cases one or the other method will give the best results. The predominant significance of wash load in the silting of most reservoirs has been mentioned. In very few cases would this problem justify the separate and detailed study of bed-material load. Its main significance appears in problems of stream-channel stability, for which it has been shown that the wash load has no influence. If a stream is depositing annually a certain amount of sediment in its channel bed, flood damage must be expected to develop. The usual question in such a problem is whether or not any countermeasures are economically feasible to prevent all or part of the flood damage. This problem may be approached by either of two methods or by a combi- nation of them. The rate of sediment supply may be reduced by re- taining the material at its point of origin or in sediment-retention basins; or the sediment capacity of the channel in question may be increased by channel rectification and elimination of unnecessary flow resistance (Einstein, 1944), or by a combination of the two methods. ~ ROUGHNESS A discussion of the laws of sediment transportation would not be complete without some remarks about the roughness conditions along Cre 3] TRACTIVE FORCE 69 a movable bed. This friction will determine the relationship between stage and discharge in the river, and it will determine the average velocity and the velocity distribution. It has been found that, basi- cally, this friction may be described by von Karman’s logarithmic equations, using the constants which Keulegan (1938) has derived from Nikuradse’s experiments for the flow along rough walls. The roughness diameter k in these relationships is represented by the erain diameter in the case of sediment of uniform size. For sediment mixtures the diameter used in computing the roughness factor k is the diameter known to engineers as De;5. That is, 65 percent of the weight of the sediment is composed of particles smaller than Deg; and 35 percent of particles larger than this diameter. Additional flow resistance is caused by the ripples and bars of the sediment bed and by the irregularities of the channel. Both may be expressed in terms of the specific sediment rate. From flume studies where channel irregularities are practically eliminated, one can learn that the additional friction due to ripples and bars is small for the very lowest and for extremely high rates of sediment transport, reach- ing a maximum value at an intermediate rate where sediment bars are most commonly developed. In natural rivers not constricted by artificial banks, the reduction of the additional friction at low sedi- ment rates is counteracted by channel irregularities such that the overall friction coefficient is highest at lowest flows. These relation- ships are not sufficiently understood yet but are just as important for the calculation of an alluvial river as is the bed-load equation itself. The above discussion pertains primarily to the rate of sediment transportation. Another phase of the sediment problem that has been investigated extensively in the past is the critical condition control- ling the commencement of movement of the bed material. TRACTIVE FORCE The most common conception of the mechanics of bed-load move- ment is that a dragging force is exerted on the bed of a stream by the flowing water. This force, termed the tractive force, is the dragging or entraining force exerted at the base of a prism of water of unit area of the bed and of height equal to the water depth sliding, under the influence of gravity, down an inclined plane having a given slope. Critical tractive force is that tractive force which creates “general movement” of the bed material. General movement, as commonly used in this sense, is that condition under which sand grains up to and including the largest size available are in motion. This critical con- 70 EINSTEIN-JOHNSON. SHDIMENT TRANSPORT [Cu. 3 dition for movement is related to the specific gravity of the bed ma- terial and to such mechanical properties as grain size, grain distribu- tion, and grain form. Numerous flume experiments have been conducted to obtain a rela- tionship between these various factors at the point of general move- ment. Kramer (1935), one of the first to attempt to formulate a criterion for defining critical conditions, used experimental data from all available sources and developed a formula in terms of sand size and distribution. In later investigations at the U. S. Waterways Experi- ment Station (1935), Kramer’s criteria and data were used and a modified formula was derived. Still later Chang (1939) analyzed all available data and also pre- sented a formula for the critical tractive force in terms of the mechani- cal properties of the material. Numerous experiments were conducted at the University of Iowa (Mavis, Ho, and Tu, 1935) to establish a relationship between a competent bottom velocity, corresponding to impending motion of the stream bed, and the size and specific gravity of the bed material. Somewhat later Mavis and Laushey (1949) re- analyzed the Iowa data and presented a new relationship between critical bottom velocity and size and specific weight of the material. The practical purpose of relationships between critical conditions for movement and the sediment characteristics is to establish permis- sible velocities in the design of earth canals. Also, in many bed-load formulas the rate of transportation is expressed as a function of the difference between the tractive force and the critical tractive force of the material in the bed. Although the theory underlying this rea- soning, namely, that the transportation is a function of the “residual tractive force,” is perhaps correct, in the practical case the critical tractive force is relatively small compared to the total tractive force, and its inclusion in a bed-load formula becomes primarily one of aca- demic interest. REFERENCES Brown, C. B. (1945). Rates of sediment production in southwestern United States: U. S. Dept. Agr. Soil Cons. Service, SCS-TP-58, Jan. 1945, 40 pages. Chang, Y. L. (1939). Laboratory investigation of flume traction and transpor- tation: Trans. Amer. Soc. Civ. Engrs., vol. 104, pp. 1246-1313. DuBoys, P. (1879). Le Rhéne et les riviéres a lit affouillable: Ann. ponts et chaussées, 5 ser., tome 18, pp. 141-195. Eakin, Henry M. (1936). Silting of reservoirs: U. S. Dept. Agr., Tech. Bull. 524, 168 pages (revised by C. B. Brown, 1939). Einstein, H. A. (1944). Bed-load transportation in Mountain Creek: U. S. Dept. Agr. Soil Cons. Service, SCS-TP-55, 54 pages. Cu. 3] REFERENCES 71 Einstein, H. A. (1947). Determination of rates of bed-load movement: Pro- ceedings of the Federal Inter-Agency Sedimentation Conference, Denver, Colo., May 6-8, 1947, pp. 75-90, U.S. Bureau of Reclamation, Washington, D. C. Iowa University (1940a). Feld practice and equipment used in sampling sus- pended sediment, Report No. 1: Published at St. Paul Engineer District Sub- office, Iowa City, Iowa, Aug. 1940, 175 pages. (1940b). Equipment used for sampling bed-load and bed material, Re- port No. 2: Published at St. Paul Engineer District Sub-office, Iowa City, Iowa, Sept. 1940, 57 pages. (1943). Density of sediments deposited in reservoirs, Report No. 9: Published at St. Paul Engineer District Sub-office, Iowa City, Iowa, Nov. 1943, 60 pages. (1948). Measurement of the sediment discharge of streams, Report No. &: Published at St. Paul Engineer District Sub-office, Iowa City, Iowa, March, 1948, 92 pages. Keulegan, G. H. (1938). Laws of turbulent flow in open channels: Jour. Res. Natl. Bur. Standards, vol. 21, pp. 707-741. Kramer, Hans (1935). Sand mixtures and sand movement in fluvial models: Trans. Amer. Soc. Civ. Engrs., vol. 100, pp. 798-838. Mavis, F. T., and Laushey, L. M. (1949). Formula for velocity at beginning of bed-load movement is reappraised: Civil Eng., vol. 19, No. 1, Jan. 1949, pp. 38-39. Mavis, F. T., Ho, Chitty, and Tu, Yun-Cheng (1935). The transportation of detritus by flowing water—I: Univ. Iowa Studies in Engin., Bull. 5 (new ser. 294), 53 pages. U. S. War Department, Corps of Engineers (1935). Studies of river bed ma- terials and their movement, with special reference to the lower Mississippi River: U. 8S. Waterways Exp. Sta., Vicksburg, Miss., paper 17, 161 pages. Vanoni, Vito A. (1946). Transportation of suspended sediment by water: Trans. Am. Soc. Civ. Engrs., vol. 111, pp. 67-102. (1947). Development of the mechanics of sediment transportation: Pro- ceedings of Federal Inter-Agency Sedimentation Conference, Denver, Colo., May 6-8, 1947, U. S. Bureau of Reclamation, Washington, D. C., pp. 209-222. Vetter, C. P. (1937). Why desilting works for the All-American canal? Eng. News-Record, vol. 118, No. 8, pp. 321-326. CHAPTER 4 GEOPHYSICAL PROBLEMS IN APPLIED SEDIMENTATION Rouanp F. BEERS President, The Geotechnical Corporation Dallas, Texas One approach to problems in applied sedimentation cuts across the lines of development of the geophysical sciences. In following this application it is important to avoid the many preconceptions and prejudices one encounters in approaching geophysics for the first time. Many problems in applied sedimentation will benefit by the use of geophysical methods, but there is great danger in assuming that geo- physics will provide a perfect solution. The history of many phases of geophysical science has been repeated too often: first, a brilliant idea showing exceptional promise of success, without practical limita- tions; second, an extravagant program of application without ade- quate buttressing of the contributory sciences; third, the inevitable disappointment following the misguided application; and, finally, a wholesale condemnation of the method employed, perhaps of geo- physics entirely. It is difficult to understand why some applications of geophysics have been developed in a scientific manner to produce valuable results, whereas other equally important applications have followed the un- happy course outlined above. Some people become emotional about geophysics. Either they ride the crest with unbounded enthusiasm, or they sink to the depths of despair when confronted with failures. This pattern has been retraced altogether too often during the past three decades for it to be passed by without comment. It is hoped that these remarks may spare the reader from similar unprofitable experi- ences. Geophysics is no panacea. It has purpose, scope, and limitations. It can be used to advantage for many problems, and the overall costs are usually favorable. Its successful use involves broad knowledge and experience in many sciences. Strictly speaking, there is no single department of geophysics. It transcends the boundaries of many 72 Cu. 4] GEOPHYSICAL PROBLEMS 73 sciences, such as physics, mathematics, engineering, economics, hy- drology, oceanography, seismology, meteorology, geodesy, terrestrial magnetism and electricity, and practically all phases of geology. If _ one approaches a problem in geophysics without recognizing the breadth of this segment of human knowledge, it is highly improbable that a satisfactory solution will be obtained. Restricting one’s point of view too severely has accounted for many of the failures of the past. It is only through the complete integration of all related fields that a completely satisfactory solution may be achieved. One may ask: “How can one person cover such a broad field of knowledge?” ‘The problem is not new. It is one which the geologists have cultivated, one which accounts for their success in many fields of endeavor. Their guiding principle was stated many years ago by T. C. Chamberlain * as the principle of working hypotheses. Under the philosophy of this idea the investigator pursues the problem with an open mind. He does not formulate a final solution until he has attacked the problem from many angles. In the design of a founda- tion, for example, he will first study the geologic setting, the structure and stratigraphy of bedrock. Then, from the viewpoints of sedimen- tation and soil mechanics, he will examine the soil and overburden, viewing them not only as structural materials to support the founda- tion. As a geologist he will recognize that the soil in his hand is a result of many active processes, resulting in the soil as it is today, and constantly changing its properties. At some future time it may not possess the same characteristics. The scientist will look for new facts and new processes which in the future may modify present conditions. These researches lead him into the fields of seismology and earth- quakes, oceanography, geodesy, hydrology and meteorology, geology and sedimentation. The illustration depicts the principle of multiple working hypotheses. From each independent point of view a preliminary estimate of perti- nent factors is achieved. When all possible points of view have been examined, they are integrated into an evaluation of the relative im- portance of each separate element. When this integration is made, a new factor will be noted, namely, that the whole is greater than the sum of the parts. Herein lies the great power of this working prin- ciple. When several independent factors are correlated, new factors appear which are the result of the interrelation of two or more com- ponents. In communication networks this element is known as “cou- pling.” It represents a measure of the mutual and reciprocal rela- *T. C. Chamberlain, The methods of the earth-sciences, Popular Science Monthly, Vol. 66, pp. 66-75, Nov. 1904. 74 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 tionships existing among the various parameters. It is important to estimate the probable significance of these coupling factors, first, as they are independently established and, then, as they are arranged to show their several relationships. The development of a final solution is a series of probability deter- minations. At each successive stage a closer approach to the end result is achieved. With the addition of each sample of new data and the interpretation and correlation thereof, relative priorities of each partial solution become established. The final solution expresses the order of probabilities which the respective solutions bear to one another. One solution may rank very high and be employed as the first working basis. This method of attack will stand in sharp contrast to that employed by some engineers and scientific workers. In many fields it is estab- lished practice to formulate a single, explicit, unique solution. The possibility of other solutions is not admitted. In geophysical sciences the number of variables is so large that a single solution cannot be depended on. It is necessary to use the multiple-hypothesis method to be certain that nothing has been overlooked. Sometimes not even this method will cover the field. OUTLINE OF GEOPHYSICAL METHODS The broad nature of geophysics, as indicated above, requires one to approach the subject slowly. A division can be made into: (1) “pure science” geophysics; (2) exploration geophysics for petroleum and minerals; (3) engineering geophysics (including soil mechanies) ; (4) geophysical aids to geologic mapping, surface and subsurface; (5) development of ore deposits; (6) petroleum production—primary and secondary recovery; (7) military and naval problems. The advantage of viewing all these areas of interest together is that principles, techniques, and instrumentation developed in one area may find useful applications in others. Borrowing of ideas is important because of the high cost of geophysical research and development. A reflection seismograph used for petroleum exploration may cost $100,000 in capital outlay and an operating charge of $600 per diem. If there is any operating principle, technique, or instrument which the petroleum industry has already developed it is expedient to adapt it to new problems. The cost of geophysical research and development for new problems of sedimentation may be prohibitive. The wisdom of employing expert consultants in study groups is indicated by these costs. The experts bring with them not only a working knowledge Cu. 4] GEOLOGY 75 of scientific principles, but also an intimate acquaintance with the implementation thereof. Geophysical methods require the collection of substantial amounts of field data. For this purpose specially designed equipment is used, operated by trained technicians and scientists. The task of gathering field data with this equipment is, in itself, a special skill. Not only is the quality of data better, but also the unit cost is lower when the data are gathered by competent field units. In many areas of applied geophysics competence is widespread, but, for the past ten years, there has not been enough personnel to meet the demands. If geophysical activities continue at the current rate, little improvement of this con- dition can be expected. If more geophysicists are required, it is neces- sary that more schools of geophysics be established. In addition to their formal education, geophysicists also require a comprehensive program of field training. Excepting the petroleum industry, no such program of training is in effect. For these reasons many problems in sedimentation which would benefit from the attention of experienced geophysicists are now being neglected. The following discussion contains the principal topics covered by the cited divisions of geophysics. These divisions exist mainly in the published literature. For convenience of reference they are retained herein. “PURE SCIENCE” GEOPHYSICS SEISMOLOGY The subject of seismology includes earthquakes, their cause and their effect; frequency and places of occurrence; the internal constitu- tion of the earth; long-distance propagation of seismic waves; micro- seisms, their origin and propagation, their correlation with meteor- ology, and their use in storm forecasting. Applications to sedimentation. Seismology has many practical ap- plications: the design of quakeproof structures; soil mechanics and foundation problems; selection of sites for installations, buildings and public works in consideration of the structural properties of the ground; amplitudes of vibrations from natural and man-made sources; soil tests and test equipment. GEOLOGY Sediments are one of the three fundamental rock types, and they cover three-fourths of the surface of the land areas of the world. Sedimentation comprises the following subjects: the fundamental laws of sedimentation; the work of water, wind, and ice in removing, trans- 76 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 porting, and depositing sediments; the natural history of sediments and sedimentary rocks; provenance studies, the source rocks of sedi- ments; alteration of sediments by burial, compaction, chemical, ther- mal, physical and igneous metamorphism; the effects of wind, frost, heat, water, ice and chemical agencies; the sedimentation cycle; the evolution of soils; the geologic processes at work on soils; structural geology, mountain and continent building as factors in the evolution of sediments and soils. Applications to sedimentation. Geophysical applications include problems of installations: dam site, highway and public works con- struction; water supply; the properties and usage of soils. HypDROLOGY Water and its effects constitute the subject of hydrology. Topics considered are: principles of origin of ground and surface water; dis- tribution and recovery of water; hydrometeorology; precipitation, distribution, and variations; surface evaporation and transpiration of water; the permeability of rocks and soils; snow surveys; transporta- tion of sediments by running water; sedimentation in reservoirs, lakes, and ponds; the water content of soils. Applications to sedimentation. Applications of hydrology include water-supply storage; ground water; hydroelectric projects; flood control; storm protection; land drainage; irrigation; runoff control; stream flow. OCEANOGRAPHY The ocean and its characteristics constitute the subject of oceanog- raphy. Topics included are: ocean waves, tides, currents, and swells; water levels: diurnal and secular variations; tsunamis and seismic waves; forecasting of breakers and surf; interactions by the ocean and the atmosphere; physical properties of sea water, sea-surface temperatures; beach composition and construction; ocean channels, their structure and alteration. METEOROLOGY Meteorology covers the subjects of weather forecasting; climatol- ogy; hydrometeorology; mean annual precipitation rates, runoff, storm and flood forecasting; water levels and water supply; mean annual temperature distribution. Applications to sedimentation. Meteorology is related to sedimen- tation in many ways: water project system operations; determination of maximum rainfall, flood crests, reservoir capacities, flood storage Cu. 4] EXPLORATION GEOPHYSICS 77 basins; hydroelectric developments; reservoir and dam construction projects, irrigation projects, land usage surveys. REFERENCES TO LITERATURE At the end of this chapter is a list of references which the reader may consult for further information on the foregoing subjects. There is a wealth of knowledge in the texts and periodical literature cited in these references, much of which is in suitable form for immediate utilization by engineers and scientists who are approaching problems in sedimentation without previous experience in the field. EXPLORATION GEOPHYSICS (For PrerroteEuM AND MINERALS) The United States petroleum industry spent approximately $500,- 000,000 during 1948 in search of new deposits of oil and gas. At least one quarter of this amount was spent in scientific methods of exploration employing geology and geophysics. Most of the latter sum was spent on exploration geophysics through the activities of approximately six hundred field parties in the United States alone. It may be inferred, correctly, that the expenditure of sums of this order implies the existence of a high degree of scientific and technical competence. It is true that the expenditures provide a return in new discoveries of oil and gas equal in value to many times the exploration costs. The success of the petroleum industry during the past twenty- five years is ample testimony to the philosophy of multiple working hypotheses. Although most of the money is spent on exploration geo- physics, the strategy of long-range campaigns embraces every imagi- nable point of view in petroleum geology. The consequence of this great effort has been the development of an extremely high degree of scientific and technical skill in the pur- suit of new discoveries. The success of an exploration program, when distributed over a sufficiently broad base, is now taken for granted. There is a direct relation between the amount of oil discovered and the funds applied to the task. It is true that the unit cost of discovery is constantly rising, but the law of diminishing returns is not yet too seriously felt. Exploration by scientific means is still profitable on a large scale. The technical basis of petroleum exploration geophysics can be expressed briefly. Through many years of trial and error the industry has found that many physical properties of rocks are systematically distributed with reference to oil and gas. Four or five of these prop- erties are significant enough to warrant serious prospecting techniques. 78 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 More than half the money spent for geophysics is devoted to the seis- mograph, an instrument which measures the travel time of sound waves into the earth and return. The principle is as simple as the echo returning from a sharp discontinuity: a brick wall, the side of a cliff, or the edge of a forest. The principle was first developed in echo depth sounding after the sinking of the steamship Titanic by an iceberg in 1913. Since then the echo principle has been developed in many fields, finally resulting in the evolution of radar. In its application to oil finding, the seismograph employs miniature earthquakes set off by small explosions of dynamite just below the surface of the ground. Sound waves traveling through the earth are in part returned to the surface by sudden changes in sound velocity. The sequence of layered rocks commonly found in oil provinces makes an ideal setting for the return of a series of echoes. Each reflection is recorded on a moving film which can later be analyzed to identify the origin of a long series of echoes. The reflection seismograph makes echo determinations over a network of points spaced from a few hundred feet to one mile apart on the surface of the ground. At each of these points there results a sub-sea-level datum value of each of the reflection horizons which appear on the prospect under survey. The numerical datum values, under the guidance of a competent ge- ologist, may be contoured to show the attitude at depth of one or more reflecting horizons. If a horizon is related to the occurrence of oil and gas in the form of a confining trap, this fact will be inferred by the geologist from the contours. He will then recommend the drilling of a well upon a favorable site to investigate the conditions of permeability and saturation in the objective horizon. By established refinements in the reflection technique it is possible to achieve a quantity of subsurface data which are equivalent to drilling a well at each reflection point. An accurate structural picture can be obtained as well as valuable stratigraphic data. If lateral changes in sedimentation occur, they may often be inferred from the character of reflection records received along the transition zone. Full use of the resolving power of the reflection seismograph in this application is rarely employed, but if fine detail is required it is available at additional cost. There are a number of controls over the resolving power which enable one to delineate fine structure within the grosser features if the expense is warranted. Among the refinements which the reflection seismograph may em- ploy with advantage is the accurate determination of the velocity of sound through the sequence of beds. For this purpose a special seismic detector is lowered into a well drilled through the rock formations. Cu. 4] EXPLORATION GEOPHYSICS 79 Small explosions at the surface are recorded by the well detector at various depths so that individual formation velocities can be deter- mined. It will be found that the formation velocities offer valuable guidance to the identification of the beds, becoming, thereby, a veloc- ity signature. This property is used in the refraction seismic method which maps only the major velocity horizons. The method is useful in large-scale reconnaissance surveys where fine detail is not required. It is also adaptable to small-scale problems such as the measurement of the thickness of soil and overburden, the depth to bedrock, and the dis- tribution of these quantities over sites for buildings and public works construction, highways, reservoirs and storage basins, and in the search for new supplies of ground water. In seismology for petroleum exploration, the general principles of earthquake seismology are employed in a restricted degree. It is sufficient to note that the velocity of sound in rocks is the variable parameter upon which the success of the seismic methods depends. Their use is restricted to the delineation of structure and stratigraphy in rocks. If these factors are systematically related to the occur- rence of oil and gas and of ore deposits, the method may achieve economically useful results. There is no implication that an accurate seismic survey directly indicates the occurrence of valuable treasure. This inference must be based upon other considerations, principally geological. The use of the seismograph is comparable with the use of the transit in surface surveying. Both collect numerical data which may be contoured to show relative relief over the area surveyed. Both methods are precision forms of engineering application, and as the indications of either are established by the presentation of numerical data, other realms of science become involved. It is outside the juris- diction of the map maker to interpret the significance of the data listed thereon. Other properties of rocks whose distribution may have bearing upon the occurrence of petroleum and minerals are specific gravity, magnetic susceptibility, electric conductivity, and radioactivity. By gathering data at the surface of the earth, or in bore holes drilled especially for the purpose, it is possible to infer something of the distribution of rocks beneath the surface from a map of the observed and reduced data. Variations in subsurface rock densities reveal themselves at the surface by the distribution of contours of measured values of gravity. By making very precise measurements of gravity at a series of closely spaced points it is possible to derive significant indications of the 80 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 distribution of rocks at depth. The method has a high resolving power when all possible corrections are made. These include corrections for sea-level elevation of each station, distribution of surface rock densi- ties and their elevation, topographic corrections, and instrumental drift corrections. When properly gathered and interpreted, gravity data are of great value in detecting the presence of buried structures and mineral deposits. Interpretation of the data requires an intimate knowledge of the rock characteristics in the layered sequence, as well as of the general geologic setting in which the gravity survey is con- ducted. The method is of value to the sedimentologist in locating placer deposits such as buried river channels, filled basins, faults and dikes, and many geologic anomalies based upon density contrasts. Variations in the magnetic susceptibility of rocks are almost en- tirely determined by the magnetite content. Although there are many magnetic minerals in sediments, none have such widespread occurrence as magnetite. If the geological and mineralogical associations in sedi- mentary rocks are well understood, it is possible to map the distribu- tion and attitude of such formations. For this purpose, measurements of the earth’s magnetic field in a bore hole, at the surface, or from aircraft may reveal valuable information. The magnetometer is the oldest prospecting device used by man which has continued to the present day. Like other geophysical instruments, it does not lead directly to the discovery of treasure, excepting for magnetic minerals; therefore, its usefulness is confined to the indications which it gives on structure, stratigraphy, and the distribution of magnetic materials. In applied sedimentation the magnetometer may be used for the broad scope of mapping the distribution of magnetite-bearing rocks or for local surveys for ore bodies. Its use has been most widespread in prospecting for ore, relatively less for petroleum surveys than other methods. Because of its low cost and speed of operations it is well suited for broad reconnaissance surveys immediately after geo- logic mapping. Favorable areas isolated by the magnetometer are usually explored in further detail by other methods. The electrical conductivity of rocks in a place is determined by two principal factors: (a) their water content, and (b) concentration of mobile ions. If the presence of conducting ions is systematically re- lated to the occurrence of oil, gas, or minerals, electric resistivity surveys will be of value. These have been successfully conducted from the surface of the earth by employing a profile method which traverses lines of exploration related to the area under investigation. The presence of a conducting layer, such as a sand formation saturated with conducting water at moderate depths (not over 5,000 feet), may Cu. 4] EXPLORATION GEOPHYSICS 81 afford a convenient datum to which the resistivity measurements may be referred. By using the conventional four-electrode method, values of formation resistivity may be plotted along parallel lines of explora- tion to show the subsurface attitude of the conducting horizon. Modi- fications of this simple geologic setting reveal complicated pictures of fine detail. The presence of a layer of high conductivity near the surface will effectively mask more deeply buried features. A modification of the electrical method is found in electric well surveys. Here direct measurements of formation resistivity and elec- tric self-potential are taken from electrodes lowered into a well. A log of the variations in these two quantities is of great value in the development of petroleum production. It shows the presence of oil or gas, fresh or salt water, and striking variations in the properties of different types of formations, sandstones, limestones, and shales. The electric logging method has great promise in near-surface appli- cations, as yet relatively undeveloped. The instrumentation is simple; its operation involves no great cost. Problems to which the method might be applied are the location of water supply, determination of bedrock beneath the surface and overburden, the location of faults, dikes, and other structural features related to construction engineer- ing problems. The measurement of the radioactivity of sedimentary beds in place offers striking correlation with electric measurements on the same for- mations. The reasons for these correlations reside in the particle size distribution of sediments and sedimentary rocks. In general, the fine-grained rocks, shales, and clays show the greatest radioactivity, whereas rocks comprised of coarser fragments, such as sandstones and conglomerates, show lower activity. Induced radioactivity em- ploying neutron radiation is also of value in identifying sedimentary sequences. Both methods have undeveloped potentialities in the same series of problems outlined for electric applications. The search for minerals and ore deposits by geophysical means in the United States has not greatly advanced the science in recent years. Within the mining industry there seems to exist an attitude which is difficult to explain. In the past the merits of any geophysical survey have been largely judged by the amount of ore the survey produced. Surveys failing to produce ore have been condemned without critical examination, and there has been no opportunity to profit from these failures. The mining industry has not operated under the philosophy which the petroleum industry has found so successful, that is, to de- velop geophysics on as broad a base as possible. The result is that full development of mining geophysics in the United States has not taken 82 BEERS. GEHOPHYSICAL PROBLEMS [Cu. 4 place, and we now find ourselves practically in the same position as in 1925. In Canada more success has attended the applications of geophysics. Magnetic and electric methods are in general use. Much new ore has been found from these applications. The airborne magnetometer of- fers considerable promise in the development of large virgin ter- ritories which are otherwise inaccessible. The advances to be ex- pected in this field in the immediate future are those in transpor- tation and operational facilities. Probably no revolutionary discov- erles In geophysical principles should be expected. ENGINEERING GEOPHYSICS (Inctupine Som Mercwuanics) In the preceding section it was intimated that some exploration techniques might be applied with advantage to problems in engineering. There are many problems, closely related to each other, which would benefit by a modest effort in research and development. These prob- lems are recognized by many, but there seems to be no widespread understanding that geophysics might be of aid in their solution. Un- like the petroleum industry, there has been no sponsoring benefactor to pay for development costs. It is possible that state and federal agencies or other institutions may attack these problems to advantage. The general principles outlined here are intended to guide those wish- ing to pursue future courses of action. These principles will be il- lustrated by a few examples. In the selection of a site for the foundation of a dam, bridge, aque- duct, viaduct, highway, or a large building construction, the con- struction engineer and the contractor seldom have adequate data on the subsurface underlying the site. Current practice is to subcon- tract for a series of boreholes to bedrock, or to such depths that piles can be driven, to support foundations. In cases where bedrock is found it may happen that the rock encountered is a first layer, under which other unconsolidated material may lie, hidden from sight. The location of the boreholes is usually such that the profile of bed- rock surveys is assumed to be smooth between boreholes. It is rare that enough boreholes are drilled because their cost is very high. Where piles are driven, the bearing capacity is calculated from em- pirical formulas based upon assumptions and experiences which may be quite removed from those of the site under consideration. Sample cuttings of boreholes in unconsolidated materials display properties quite unlike those encountered in situ. Foundation sites for many large structures have often been selected Cu. 4] GEOPHYSICAL AIDS TO GEOLOGIC MAPPING 83 without knowledge of subsurface geology, subsurface geological fea- tures such as faults, buried channels, and basins filled with uncon- solidated materials. Many famous structures are in danger because of these oversights. Some of the errors could have been prevented by the employment of competent geologists when the selection was made, but they might have been unable to detect hidden features which could be brought to light by geophysical methods. The properties of soils as components of structures and foundations require more convenient and precise methods of determination. Con- sultants in the field of soil mechanics are casting about for devices which they may use for testing soils and their properties. It is diffi- cult to prescribe one formula to meet all these needs. This is es- pecially true because prevailing concepts in soil mechanics generally omit the geological point of view, the concept of active processes which affect the soil constantly throughout its life. The foregoing examples illustrate the need for research and de- velopment in the field of applied sedimentation. In all these problems it is clear that an application of the existing techniques of geophysics will bring early assistance to bear. Already there are competent re- sources of personnel and facilities if the way is made clear for their utilization. The difficulty is to find sponsoring agencies who will co- ordinate these efforts. This is a new borderline science which it seems important to develop. It is hoped that these remarks may be brought to the attention of those in position to chart future courses of action. GEOPHYSICAL AIDS TO GEOLOGIC MAPPING One of the functions of the U. 8. Geological Survey and of state geological surveys is to map the geology and the mineral resources of our country. In this function the surveys have for many years carried out a comprehensive program of geologic and topographic map- ping. The locations where these surveys have been completed are those where the needs were greatest. At first, immediate develop- ments followed closely upon the completion of adequate maps. This function has been extended to subsurface mapping where underground workings have been opened by mining operations and by the drilling of wells for oil, gas, and water. We have come to a phase in the development of our country’s mineral resources where these methods of geologic mapping greatly restrict the nation’s development. It is now time to think of ways of large-scale mapping of subsurface geology, without competing with 84 BEERS. GEOPHYSICAL PROBLEMS [Cu. 4 the functions in this realm normally exerted by commercial operating companies. To expand the mapping function it is inevitable that resort be made to geophysical methods. The seismograph has dem- onstrated well its ability to portray the attitude and depth of key horizons in sedimentary provinces. If structure and stratigraphy are concealed beneath a cover of overburden, alluvium, or valley fill, there is no better method of unveiling subsurface conditions than by the use of geophysics. Its cost is insignificant compared with that of drilling an adequate number of holes in a virgin territory. The distribution of data yields a much better density than other methods permit. The prevailing lack of appreciation of geophysical methods may be at- tributed in part to lack of understanding of the method, but also to lack of facilities and trained personnel in the United States. Since it is the proper function of the geological surveys to develop new methods of exploration and survey, it is hoped that this area of investigation may be developed in the near future. Already we have valuable ex- amples of the contribution made to the nation’s resources by the air- borne-magnetometer maps which the U. 8. Geological Survey has pro- duced in the Lake Superior region, in the magnetite, lead, and zine deposits of New York, in the central portion of the state of Pennsyl- vania, and elsewhere. These reconnaissance maps will be of value for many years to those who follow with ground surveys of greater detail. DEVELOPMENT OF ORE DEPOSITS Since the mining industry has not developed geophysical methods for exploration to a high degree, it is conceivable that a beginning might be made in the development of existing ore deposits. There are many problems in the extension of working properties to which geophysics would make substantial contributions. These would employ the conventional instruments and techniques of the principal kinds discussed herein. Close correlation by the geophysicist and the min- ing geologist is likely to develop profitable areas of application through the extension of known ore deposits. None of the principal geophysical techniques has been exhausted in this connection. It seems likely that valuable advances could be made in this field at relatively small costs. DEVELOPMENT OF PETROLEUM PRODUCTION In sharp contrast to the mining industry, the petroleum industry has expanded large numbers of geophysical techniques to the benefit of producing properties everywhere. The industry is quick to take Cu. 4] SUGGESTED PROGRAM OF ATTACK 85 advantage of every scientific fact related to the occurrence and pro- duction of oil and gas, with the consequence that scientific aids to oil production have been developed to a high degree, a degree that is rare in all American industry. The wisdom of this policy is found in the excellent returns of producing divisions. Properties measured in the course of petroleum production follow: (a) Electrical conductivity of reservoir (7) Acoustic impedance. rocks and their contained fluids. (k) Density. (6) Electrical self-potential. (1) Magnetic susceptibility. (c) Spectrographic analysis of reservoir (m) Colorimetric determinations. rocks and contained fluids. (n) Porosity. (d) Radioactivity of rocks and fluids. (0) Permeability. (e) Induced radioactivity of rocks and (p) Fluid saturation. fluids. (q) Composition of connate fluids, in- (f) Fluorescence of reservoir fluids. cluding pH. (g) Mineralogical composition. (r) Temperature distribution. (h) Insoluble residues of sedimentary (s) Seismic wave velocity. rocks. (z) Clay mineral content and identifica- tion of species. These measurements of conditions of sedimentation suggest the large number of possibilities which have not been developed in geophysical applications to problems outside the petroleum industry. MILITARY AND NAVAL PROBLEMS The application of geophysics to these problems involves many sub- jects which are classified for military security. Some of the published articles on the subject show possibilities of use in other areas which ap- proach those of the petroleum applications in number. It is promis- ing to observe that under the stimulus of military necessity many established geophysical techniques have developed new values in a different setting. Some of the spectacular performances of World War II involved highly coordinated programs in the geophysical sciences. If it had not been for combat teams in oceanography, mete- orology, and military geology, many of our task forces would have suffered severely at the hands of the enemy. SUGGESTED PROGRAM OF ATTACK For the sedimentologist who wishes to employ geophysics as an aid _ to a solution of his problems, it is recommended that the principles 86 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 discussed above be applied through the study-group method. Unless one is contemplating a long period of geophysical activity it is in- advisable to recommend that full knowledge of the contributing fields be acquired. This would be a difficult task in itself, not to speak of the time involved. In many organizations where special points of view are indicated, good progress is made in relatively short time by or- ganizing a small group of specialists with a leader whose knowledge and experience are broad enough to give him a general understanding of the language of each member of the group. In this way no possible interrelations will be lost, and the productivity of such a group, under able leadership, will be surprising. It should be the policy of the study- group leader to encourage all worth-while suggestions during a stage in the program where criticism and evaluation are not present. Until the group is accustomed to working together, there may be some whose natural timidity will prevent them from delivering all their ideas for the benefit of the meeting. It has been found practicable to cover the entire field of research in a preliminary stage as suggested. After the material so presented is organized, it will be desirable to evaluate individual ideas. For this purpose it is helpful to have one member of the group who is outspoken and straightforward, and who commands the respect of others for his clear thinking and power of analysis. The function of the group at this stage is to screen out all valuable material and to discard that which is unsound or irrelevant. The third stage arrives when the sound material has been organized and resubmitted to the group. At this point important decisions will be made, strong convictions will be formed, and there will arise a surprising unanimity among the members of the group as to the logical procedures for the future. It may develop during the course of the meetings that new data or new interpretations are required. For these functions the chairman may appoint small working groups or individuals who will assemble and present the new materials in such manner that all members of the group may make use of them. The timing of the group meetings will depend upon the amount of this de- tailed work required. One brief caution may be noted. In some group meetings normal scientific procedure and good judgment have been overridden by an emotional factor. Some member of the group should be given the responsibility of watching for this element. In one case a group of eminent scientists was completely won over by a “‘spellbinder” presen- tation which lay undiscovered for many weeks. Only under long and careful examination did the weakness of the group’s conclusions be- Cu. 4] ENGINEERING GEOPHYSICS 87 come evident. It is unlikely that any member of that group will ever become a victim of similar circumstances again. PROBLEMS SUGGESTED FOR FUTURE RESEARCH AND DEVELOPMENT ENGINEERING GEOPHYSICS The examples outlined above in this field are largely confined to problems of the ground, ranging in depth from the surface of the earth to a few hundred feet. In terms of the accuracy and resolving power of geophysical methods applied to exploration, many of these near- surface problems offer great promise of valuable solutions. Sedimen- tologists confronted with these problems should be readily able to de- velop methods of attack when equipped with a good working knowl- edge of applied geophysics. The barrier to rapid progress in this field of development seems to lie in the transfer of knowledge and under- standing from existing areas of application to those requiring geo- physical aids for the first time. It is obviously impossible to im- part in a short time the full scope and limitations of geophysical methods to problems of applied sedimentation. The outlook is not hopeless, however, and it is believed that, equipped with a general out- line, the student of applied sedimentation may select from the litera- ture the educational material he needs for his tasks. From a brief reading of the material of interest in the list of references at the end of this chapter, the geologist interested in sedimentation should quickly place himself in position to discuss the problems in further detail with expert geophysicists. By these means it is believed that an adequate dissemination of geophysical knowledge may be accomplished. Seismograph. The principal function of this instrument is to meas- ure depths to key horizons of sediments and rocks. A derived func- tion is to identify such horizons, enabling one to establish correlations of stratigraphic or lithologic equivalents over their lateral extent. It may also show the presence of discontinuities and transitions in sediments and rocks. From seismic data it is possible to infer the dis- tribution of sedimentary and rock formations, their dimension, shape, attitude, and identity. Often it is possible to infer something of the detailed nature of a formation such as, for example, the distinction between rocks which are massive and those which are composed of a series of stratified layers. The gravitymeter. This instrument may be adapted to near-surface ~ problems in the detection of deposits of sediments, rocks, and minerals 88 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 which exhibit density contrasts. The resolving power of the method depends primarily upon the numerical magnitude of the gravity con- trasts and the size and distance of the anomalous formation from the surface. It gives its best results on flat terrains; uneven topography introduces errors which cannot always be accounted for. If the gravity contrasts are associated with structural features, these may often be correctly inferred by the intelligent interpretation of the data and their correlation with geologic data on the prospect. The successful application of the gravitymeter extends to density contrasts in excess Gi Oeil Ground magnetometer surveys. These surveys show a high degree of resolving power for buried deposits of magnetic rocks and min- erals, if the effective magnetite content of such minerals is at least 0.1 percent. Masses of such formations can be detected and outlined satisfactorily. Negative anomalies may also prove significant, be- cause they show the presence of formations of low susceptibility in a region of normally higher values. Estimates of depth based upon magnetic data are, like those of gravity data, not entirely satisfactory unless the limits of depth estimates are supported by good geologic data. Electrical and electromagnetic methods. Such methods are of im- mediate value in the determination of depth of overburden, thickness of soil, and depth to ground water, and in the differentiation of types of rock and sediments. The accuracy of depth and thickness measure- ments is good, depending upon the amount of detail applied in the field. In locating water supply, buried gravel deposits, placers, and geologic discontinuities involving changes in conductivity, these meth- ods are cheap, rapid, and effective. The tools are readily available, awaiting only their application to problems in applied sedimentation. REFERENCES Texts and Handbooks Berry, F. A., Bollay, E., and Beers, N. R. (editors) (1945). Handbook of meteor- ology: McGraw-Hill Book Co., New York, 1068 pages. A valuable reference text for data, fundamental theory, weather forecasting, and analysis. Empha- sizes scientific and engineering aspects of meteorology. Bullen, K. E. (1947). Introduction to the theory of seismology: Cambridge Uni- versity Press, 276 pages. A concise treatise of the theoretical phases of seismic-wave propagation over long distances of the earth’s crust and interior. Contains valuable list of published works on the subject. Byerly, Perry (1942). Seismology: Prentice-Hall, New York, 256 pages. An ad- Cu. 4] REFERENCES 89 vanced text on “pure” seismology, containing data and discussions on earth- quakes, their nature, cause, effects, distribution and frequency. Byers, H. R. (1944). General meteorology: McGraw-Hill Book Co., New York, 645 pages. Aims at a complete presentation of general meteorology. Intended for professional students, engineers, and physical scientists. DeGolyer, E. (editor) (1940). Hlements of the petroleum industry: American Institute of Mining and Metallurgical Engineering, New York, 519 pages. Written for the layman who is interested in the operation of the petroleum and gas industry. Non-technical, descriptive. Contains chapter on introduc- tion to the literature of oil and gas. Authentic and valuable for a general review. Dunstan, A. E., Nash, A. A., Brooks, B. T., and Tizard, Henry (editors) (1938). The science of petroleum: Oxford University Press, London, Toronto, New York, 4 vols., 3192 pages. A comprehensive treatise of the principles and practice of production, refining, transporting, and distribution. A monu- mental work, the chief treatise on the subject. Finch, V. C., Trewartha, G. T., Shearer, M. H., and Caudle, F. L. (1942). Ele- mentary meteorology: McGraw-Hill Book Co., New York. Emphasizes clima- tology, non-mathematical. Best for geologists and earth-science students. Fleming, J. A. (editor) (1939). Terrestrial magnetism and electricity: McGraw- Hill Book Co., New York, 778 pages. Gutenberg, Beno (editor) (1939). Internal constitution of the earth: McGraw- Hill Book Co., New York, 413 pages. Contains valuable articles on the origin, composition, and structure of the earth’s crust and discussions of the data on the interior with interpretations. Heiland, C. A. (1940). Geophysical exploration: Prentice-Hall, New York, 1013 pages. A complete treatment, primarily for the benefit of the practicing geo- physicist. An introductory section (Part I) surveys the field comprehensively and serves as an introduction to the lay reader, the geologist, the manager of exploration. Part II discusses the technical bases and data of all methods of exploration geophysics, with ample theoretical development to give under- standing of the subject. The text is full of details which will delight the men who operate in the field. Petroleum and mining fields are covered. Herold, S. C. (1928). Analytical principles of the production of oil, gas and water from wells: Stanford University Press, 659 pages. Jakosky, J. J. (1940). Hzploration geophysics: Times-Mirror Press, Los Angeles, 786 pages. Like Heiland’s book, this one is also written from the viewpoint of the practicing geophysicist, serving as both reference and text. Details of theory, laboratory testing and calibration, and operational features are well covered in an authoritative manner for petroleum and mining geophysics. Jeffreys, Harold (1929). The earth, its origin, history and physical constitution: Cambridge University Press, 2nd ed., 346 pages. A classical treatise by one of the world’s foremost geophysicists. Discusses the evidence and interpreta- tion for the theories of the earth’s origin, age, figure, composition, and struc- ture. A scholarly volume. Lahee, F. H. (1941). Field geology: McGraw-Hill Book Co., New York, 4th ed., 853 pages. A field manual for geologists and others who deal with the collec- tion of field data, its representation and interpretation. Contains a chapter on geophysical methods of exploration and an extensive bibliography on the subjects covered. 90 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 Leet, L. Don (1938). Practical seismology and seismic prospecting: D. Appleton- Century Co., New York, 430 pages. Macelwane, James B., S.J. (1947). When the earth quakes: Bauer Publishing Co., Milwaukee, 288 pages. An authentic treatise for the general reader by an outstanding seismologist. , and Sohon, F. W., S.J. (1936). Introduction to theoretical seismology: John Wiley & Sons, New York, 2 vols., 366 and 149 pages. A comprehensive authentic treatment of theoretical, instrumental, operational and interpretive phases of seismology. Meinzer, O. E. (1923a). The occurrence of ground water in the United States, with a discussion of principles: U. S. Geol. Survey, Water Supply Paper 489, 321 pages. Dr. Meinzer’s famous statement of fundamental principles. Now out of print, but available at libraries. (1923b). Outline of ground-water hydrology with definitions: U. S. Geol. Survey, Water Supply Paper 494, 71 pages. (15¢ from Superintendent of Docu- ments, Government Printing Office, Washington 25, D. C.) A classic treatise, one of the few texts available in the field of hydrology. Muskat, M. (1937). The flow of homogeneous fluids through porous media: McGraw-Hill Book Co., New York, 763 pages. An exhaustive treatise, theo- retical and mathematical, of the technical problems encountered in oil-field production. National Research Council. Physics of the earth: a series of related monographs prepared under the direction of the council. The following were published as bulletins of the National Research Council and may be obtained from The Publications Office, National Research Council, Washington 25, D. C. No. 77 (1931). Volcanology. 77 pages. No. 78 (1931). The figure of the earth. 286 pages. No. 79 (1931). Meteorology. 289 pages. No. 80. (1931). The age of the earth. 487 pages. (Contains much informa- tion on sedimentation, stratigraphy, and a complete account of age-deter- mination methods, including radioactivity measurements.) No. 85 (1932). Oceanography. 581 pages. No. 90. (1933). Sezsmology. 223 pages. Nettleton, L. L. (1940). Geophysical prospecting for oil: McGraw-Hill Book Co., New York, 444 pages. Less specific in technical details than Heiland’s and Jakosky’s books, although these features are not neglected. This volume is the best introduction to petroleum geophysics yet published, and is widely used as a textbook in courses for geologists and others who have not yet arrived at the professional stage. The treatment of all phases of geophysical theory and practice is well balanced. It is the best first text for sedimentologists, yet by no means lacking in authenticity and completeness. Sverdrup, H. U. (1942). Oceanography for meteorologists: Prentice-Hall, New York, 250 pages. , Johnson, M. W., and Fleming, R. H. (1942). The oceans: their physics, chemistry and general biology: Prentice-Hall, New York, 1087 pages. Contains a wealth of data and their interpretation on the vast subject of the title. Uren, L. C. (1939). Petroleum production engineering, oil-field exploitation: McGraw-Hill Book Co., New York, 756 pages. U. 8. Department of Agriculture. (1941) Climate and man: 1941 Yearbook of Ag- Cu. 4] REFERENCES 91 riculture, Washington, D. C., 1248 pages. A comprehensive outline of climatol- ogy, its relation to man, with a valuable section on the scientific approach to weather and climate and a section on climatic and weather data for the United States and the world. Willet, H. C. (1944). Descriptive meteorology: Academic Press, New York, 310 pages. General treatment of atmospheric processes for engineers and physical scientists. Periodicals American Association of Petroleum Geologists, Bulletin, published monthly by the Association, 624 South Cheyenne Ave., P.O. Box 979, Tulsa 1, Oklahoma. Contains occasional papers on geophysical exploration for petroleum and gas. Index, 1917-1945, contains over 150 references published in the Bulletin in the period cited. Case histories, theoretical and applied techniques, problems of correlation with geologic data, interpretation and significance. American Geographical Society (1944). Geographical Review, July, 1944, New York. Contains a review of about 40 meteorological books by Chapman and Brooks. American Geophysical Union, Transactions, published by National Research Council, Washington, D. C. Six numbers per annual volume. Contains articles, news items, reports of Union meetings and papers, bibliographies, ab- stracts and reviews, and an annual index. The American Geophysical Union was established by the National Research Council in 1919 to “promote the study of problems concerned with the figure and physics of the Earth, to initiate and coordinate researches which depend upon international and na- tional cooperation, and to provide for their scientific discussion and publica- tion.” Sections of the Union are: meteorology, hydrology, oceanography, seismology, volcanology, terrestrial magnetism and electricity, geodesy, and tectonophysics. Indispensable to all geophysicists with scientific interest in the field. American Institute of Mining and Metallurgical Engineers, T'ransactions, New York, published irregularly as separate volumes: Geophysics and geophysical prospecting, vol. 81 (1929), 676 pages; vol. 97 (1932), 510 pages; vol. 110 (1934), 583 pages; vol. 1388 (1940), 489 pages; vol. 164 (1945), 426 pages; also vol. 74, pp. 3-28, contains one paper on electrical prospecting. Professional articles on all phases of geophysical exploration under a series of Technical Papers sponsored by the A.I.M.E. Committee on Geophysics. Covers case histories, theoretical, mining, laboratory and field tests, with a series of papers on geophysics education. Petroleum development and technology, published annually as Transactions. Contains professional articles on the theory, research, and engineering of petroleum production and reservoir studies. For late years see Transactions, vol. 160 (1945), 665 pages; vol. 165 (1946), 292 pages; vol. 170 (1947), 279 pages. Beginning in 1948 a new series called Statzstics of oil and gas devel- opment and production was started; 514 pages in the 1948 volume. American Meteorological Society, The Teaching of Meteorology in Colleges and Universities, Bulletin, vol. 27, No. 3, March 1946, pp. 95-98. An article on various approaches to teaching meteorology to professional meteorologists, to non-professional science students, and to laymen. American Petroleum Institute, Annual Meeting Proceedings, Mid-Year Meeting Proceedings, Drilling and Production Practice Yearbook, New York. List of 92 BEERS. GHOPHYSICAL PROBLEMS [Cu. 4 Publications available from American Petroleum Institute, 50 West 50th St., New York. Bibliography of Seismology, by E. A. Hodgson. Publications of the Dominion Observatory, Ottawa, Canada. Canadian Geophysical Bulletin, published by the National Research Council of Canada, Ottawa, Canada, quarterly. “Copy free upon request to anyone with a bona fide imterest.” Subjects reviewed and abstracted are: general geo- physics, meteorology and hydrology, oceanography, seismology, volcanology, terrestrial magnetism, atmospheric ionization, geodesy (including gravity), tec- tonics, geothermometry, and geochemistry. Carnegie Institution of Washington, Washington, D. C. List of publications available on request. A comprehensive subject list of scientific geophysics, containing reports by staff members and others sponsored by the Carnegie Institution under the direction of the Geophysical Laboratory and the De- partment of Terrestrial Magnetism. Invaluable for fundamental researches on the Earth. Journal of Geophysical Research (formerly Terrestrial Magnetism and Atmos- pheric Electricity), published by the William Byrd Press, Sherwood Avenue at Durham Street, Richmond 22, Virginia. Contains papers on all subjects in scientific geophysics (as distinguished from practical or industrial geophysics). Royal Astronomical Society, Monthly Notices, Geophysical Supplement, London, England. Contains brief articles and notes on scientific geophysics. Seismological Society of America, Bulletin, published by the University of Cali- fornia Press, Berkeley, California, quarterly. Among the objects of the Society are: “to promote research in seismology, the scientific investigation of earth- quakes and related phenomena.” Society of Exploration Geophysics, Geophysics, a journal of general and applied geophysics, Tulsa, Okla. Contains articles on theory, applications, mstru- mentation, techniques, methods, case histories, strategy and tactics of exploration by geophysical and geochemical methods. A gold mine of tech- nical detail for the practicing geophysicist. Some general articles, such as presi- dential addresses each year. U. S. Coast and Geodetic Survey, Sezsmological Publications, Department of Commerce, Washington, D. C. List of publications available on request. Covers many scientific phases of seismology and other geophysical subjects. Valuable detailed information on design, installation, and operation of ob- servatory equipment; the collection, reporting, reduction, and interpretation of data. U. S.. Geological Survey, Geophysical Abstracts, quarterly, Washington, D. C. Covers “world literature of geophysics contained in periodicals, books and patents. It deals with exploration by gravitational, magnetic, seismic, elec- trical, radioactive, geothermal, and geochemical methods and with underlying geophysical theory, research, and related subjects.” CHAPTER 5 PRINCIPLES OF SOIL MECHANICS AS VIEWED BY A GEOLOGIST * Cuirrorp A. KAyE Geologist, U. S. Geological Survey Washington, D.C. Soil mechanics, or the study of the mechanical properties of un- consolidated earth materials, is one of the youngest and most promis- ing of the border sciences that lie between the quantitative realm of engineering and the qualitative domain of geology. Soil, as used in soil mechanics, is not simply the organic-rich surface layer of the agriculturist and the geologist. The scope of the word has been broad- ened to include all granular earth materials which cannot be called hard rock. Unconsolidated sediment, regolith, and mantle rock, ir- respective of proximity to the surface, are classed as soil. Although a few short decades ago soil mechanics was only an academic specialization in the sprawling technology of civil engineer- ing, it has rapidly grown to become one of the indispensable tools of modern engineering. Within this time the more scientific methods of soil mechanics have almost completely replaced the empirical meth- ods that had traditionally been employed for the design of founda- tions, retaining walls, earth slopes, earth dams and dikes, and high- way subgrades. As part of this new engineering activity, site in- vestigations and soil-sampling techniques have been developed to a point of high refinement, and soil-testing laboratories have sprung into being in many parts of the world. Soil mechanics already has had some impact on geology. Most engineering geologists today are obliged to deal quantitatively with the mechanical properties of unconsolidated sediments, and, as a re- sult, geologists are developing an active interest in soil mechanics. This is particularly so because the close collaboration between geology and engineering, which characterizes much of today’s engineering plan- ning, demands that both groups speak the same language and recog- * Published by permission of the Director, U. S. Geological Survey. 93 94 KAYE. PRINCIPLES OF SOIL MECHANICS [Cy. 5 nize the same problems in the same terms. However, a fact less well recognized by geologists is that soil mechanics can be of great value to other fields of geological investigation. Tools have already been fashioned by engineers which geologists could profitably use in such problems as estimating the former depth of burial of clays, the critical conditions of instability in landslides, and the active pressures neces- sary to produce shearing and failure in soils. But perhaps the greatest value soil mechanics can have for geology is the introduction of mechanical concepts into geological thinking. The mechanical tech- niques are already well-advanced in soil mechanics. The geologist may be able to extend these methods to hard-rock problems and thereby in time derive a more refined understanding of the nature and intensities of earth stresses. On the other side of the ledger, there is no doubt that soil mechanics at this stage of its development needs the geologist’s sensitive under- standing of the variations of earth materials. Furthermore, there is a growing recognition that additional progress in soil mechanics awaits the solution of a number of problems concerning the funda- mental properties of soils which fall into a common ground between geology and soil science. In view of these common needs it does not seem unrealistic to expect a close harmony between geology and soil mechanics in the near future. It is therefore the purpose of this chapter to arouse the geologist’s curiosity in soil mechanics. For a _ formal introduction and ensuing acquaintanceship with the subject the student is referred to the list, at the end of this chapter, of several excellent textbooks on soil mechanics which have appeared in recent years. WHAT IS SOIL MECHANICS? The quantitative appraisal of the stress-strain relationships of soils in the engineering works of man is the subject of soil mechanics; or, to quote Professor Terzaghi (1943, p. 1), “Soil mechanics is the ap- plication of the laws of mechanics and hydraulics to engineering prob- lems dealing with sediments and other unconsolidated accumulations of solid particles produced by the mechanical and chemical disin- tegration of rocks, regardless of whether or not they contain an ad- mixture of organic constituents.” The investigation of the mechanics of soil behavior is the logical extension of the engineer’s interest in the strength of materials. The strength and elastic constants of steel and concrete, for example, have long since been determined with satisfactory precision. This knowl- edge has permitted engineers to design highly economical and safe Cu. 5] SCOPE OF SOIL MECHANICS 95 structures using these materials. Because soils form an integral part of most structures, whether as foundation or as construction material, it follows that modern structural design should include a rational analysis of soils. However, of all the impertant construction mate- rials used by man, soil is the only one which, because of its complexity and variability, cannot be reduced to relatively simple and universal numerical values. This is apparent when the complexity of factors affecting soil behavior is recognized. The mechanics of soil certainly must include such considerations as compressibility, rigidity, perme- ability, and the elastic and plastic properties. Furthermore, the aniso- tropic nature of most soils and the difficulty of deriving a statistical expression for the deviations from homogeneity complicate still further any attempt to reduce soil behavior to simple terms. It was, however, the initial problem of soil mechanics to decide, first, if any sort of reduction was possible and, then, how such a re- duction, simple or otherwise, could be made. The problem initially re- solved itself into the following questions: (1) What soil properties af- fect mechanical behavior? (2) How can these properties be measured? (3) What are the stress distributions within a soil mass? (4) How do soils act when subject to stress? What was needed, in short, was a knowledge of the stress-strain relationships in any soil mass. All soils consist of a three-phase system: solid, water, and gas. A thorough understanding of soils must therefore involve an under- standing of the interaction of all phases. Although there are notable exceptions, it can be said that soil mechanics is mainly concerned with the interaction features, or the aggregate properties, of soils. Re- search into the properties of the individual phases, and in particular into the properties of the solid or granular phase, is carried on mostly in mineralogy, sedimentation, soil physics, and soil chemistry. PRINCIPLES OF SOIL MECHANICS GENERAL The many soil conditions, with their attendant engineering prob- lems, that are analyzed by soil mechanics can be grouped into three categories: (1) those in which overstressing produces rupture or failure of the soil, (2) those in which the moderate stress conditions produce only deformation of the soil, and (3) those in which permeability of the soils is the important factor (Table 1). These problems can be further classified into those in which the hydraulic properties (the condition of the pore water) of the soil play the principal role, and _ those in which the strength of the entire soil mass is involved, that is, 96 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 TABLE 1 Tue PrincrpaL PRoBLEMs oF Som MEcHANICS Soil Properties Strength properties Hydraulic properties SLOPE STABILITY, evaluation of critical height of slopes in cuts, fills, ete. EARTH PRESSURES, prediction | SEEPAGE FORCES, effect of Stability of the magnitude and distri- pore-water pressures on sta- problems bution of soil pressures on bility of dams, dikes, slopes, walls, timbering, and other walls, ete. retaining structures BEARING CAPACITY, evaluation of safe bearing capacity of footings and piles SETTLEMENT, elastic and plas- | CONSOLIDATION, compression of Deformation tic deformation of soil under soil by external load or cap- problems footings, piles, and other illary pressure due to extru- concentrated loads sion of pore water PERMEABILITY, computation of rate of flow of water Permeability through dams, dikes, and problems natural embankments, or through subsoil of water- retaining structures. where both the solid and the liquid phases of the system are of decided importance. In soil mechanics theory all soils are considered to be granular ag- geregates consisting of discrete solid particles and interspaced voids. The solid particles are mutually supporting, each grain pressed against, and kept in place by, neighboring grains. The voids, or pore spaces, of soils are generally filled with water and air, and more rarely with minor amounts of other gases. Beneath the water table the voids are entirely filled with water except for dissolved and entrapped gases. The granular nature of soil structure is found in clays as well as in the more obviously granular silts, sands, and gravels. However, the Cu. 5] SCOPE OF SOIL MECHANICS 97 simple granular concept becomes somewhat complicated in clays, for it is known that in these colloidal soils the boundary between the solid grains and the fluid-filled voids is not always sharply defined. The presence of relatively thick layers of adsorbed water molecules on the surfaces of clay colloids introduces forces in the soil system which cannot be readily explained in terms of an ideal granular soil. Most soils possess appreciable shearing strength. In the more coarsely grained and non-cohesive soils, the source of this strength is the friction developed at the points of contact of the soil grains. The greater the pressure between soil grains, the greater is the force neces- sary to displace one grain in relation to another. The importance of intergranular friction on soil strength can be demonstrated by a simple laboratory experiment. A soft rubber blad- der is filled with loose, dry sand. As can be readily visualized, the sand inside the bladder is easily displaced by small stresses, such as the prodding of a finger. However, when a suction pump is applied to the bladder and the air from inside the bladder is excluded, the mass of sand is seen to develop a rather surprising rigidity. It will be found to resist not only the prodding of a finger but also the weight of a heavy book placed on top of it. This sudden development of strength is the result of increased intergranular pressure in the sand which is brought about by the sand grains being made to bear the weight of the atmosphere pressing against the outside of the bladder. The intergranular pressures developed in this little experiment are roughly of the same magnitude as those in dry sand buried at a depth of about 20 feet, for at that depth the weight of overburden is about equal to the atmospheric-pressure load in the experiment. This experiment demonstrates another important property of soil, namely, that the stresses in the pore fluids have a direct effect on the strength of the soil. Indeed, the strength of a soil is affected by all stresses within the soil system. This interdependence can be simply demonstrated in the laboratory in another way. The apparatus con- sists of a tank containing loose sand. Water is poured into the tank to a level above the top of the sand. When we attempt to dig a hole in the submerged sand, we discover that it is impossible to maintain a steep face in the excavation. Sand flows in from all sides, and the slopes immediately assume a low angle of repose. On the other hand, when water is allowed to flow down through the sand and out an open- ing in the bottom of the tank, we discover that we are able to dig a steep face in the submerged sand, and, in fact, with care we can make a vertical face. Finally, if the direction of flow in the experi- _ ment is reversed and water is forced up through the sand, allowing the 98 kaye. PRINCIPLES OF SOIL MECHANICS [Cu. 5 overflow to drain out the top of the tank, the sand suddenly becomes a “quicksand.” This is shown by the fact that a small weight, which had previously rested on the submerged surface of the sand, now sinks down into the sand. The explanation of these three related states of strength lies in the effect of the pore fluids on the intergranular pressure. The friction of downward-moving water on the sand grains acted as an additional downward stress and thus increased intergranular pressures. The re- sulting inerease in strength was expressed by the steep slopes that were maintained in the sand. On the other hand, the upward-moving pore water produced the same frictional effect, but in the opposite direction. The upward stresses it induced cancelled the effect of the gravitational stresses. If the water moved upward with a great enough velocity, the sand grains would go into suspension as sediment and be carried out of the top of the tank. In addition to intergranular friction, cohesion is another property which contributes directly to the shearing strength of soils. As is well known, it is futile to attempt to mold a sand castle in dry beach sand; whereas the same sand, when moistened, can be fashioned into near- vertical walls and turrets. Clay, on the other hand, can be molded into vertical walls, and even into intricate overhangs. Moreover, dry clay maintains its shape and, unlike sand, develops a high strength. This ability of soil granules to hold together when unsupported is called cohesion. Cohesion can result from a number of conditions in a soil. For ex- ample, the cohesion of the moist beach sand is the result of capillary tensions, whereas that of the clay is ascribed to the cohesive bonds between adsorbed layers of oriented water molecules on the surfaces of the clay particles. Cementation of different kinds may also be con- sidered a form of cohesion. It is apparent, therefore, that the shearing strength of soils is the sum of the two factors, cohesion and intergranular friction. This was expressed by Coulomb (1776) in an equation that has become classical in mechanical theory: s=c+ntan ¢ where s is shearing strength along any plane, c is cohesion, n is the pressure normal to the plane, and tan ¢, or the tangent of the angle of internal friction, is an expression of the frictional properties of the material. However, today it is known that the validity of this equa- tion is somewhat limited. The relationship between s and n for co- hesive soils is more complicated than the equation indicates. Cu. 5] SAMPLING AND TESTING a9 SAMPLING AND TESTING The several fundamental mechanical properties of soils that have just been described serve as a point of departure for much of soil mechanics theory. In practice there are many ways of applying soil mechanics to quantitative soil problems. Each set of geological con- ditions, as well as each type of construction, demands separate con- sideration and a different treatment. In general, hewever, soil in- vestigations follow a pattern of sampling > testing = analysis. The prediction of soil behavior that is the result of this procedure is there- fore dependent on (1) the representativeness of the samples, (2) the pertinence of the testing, and (3) the pertinence of the mechanical theory used in the analysis. Soil samples are of two types: (1) undisturbed samples, in which the soil is removed as an integral lump, thus preserving its intergranular relationships; and (2) disturbed samples, in which no attempt is made to preserve the structure of the soil. Much thought has gone into the design of undisturbed samplers for use in boreholes. A large variety of types have been built, most of them based on a removable sampling tube with a minimum of side friction and with some type of valve arrangement to prevent loss of the sample on withdrawal. Soil testing is of two basic types: field testing and sample testing. Field testing is the measuring of a soil property directly in the field without isolating a sample. The driving of penetration cones, test piles, and other resistance devices into the soil to determine strength or bearing capacity is an example of field tests. Results from such tests are generally empirically applied. Field testing is much more common in the countries of northern Europe, where widespread uniform soft Quaternary deposits prevail, than in the United States. Sample testing can be divided into three categories: (1) classifica- tion tests for index properties; (2) empirical properties tests—direct application; (3) basic properties tests—indirect application. The second and third types are simulative tests. The common classification tests employed in the United States include such procedures as mechanical analysis for grain-size dis- tribution and liquid and plastic limit, to mention only a few. The principal value of classification tests is for correlation and record. The accumulation of index data, such as plastic index or grain-size dis- tribution, when tied in with observations of soil performance, con- tribute to the building up of an empirical understanding of soil proper- ties. In fact, much soil mechanics specification for foundation design is done without the aid of the more complicated soil tests and is based 100 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 TABLE 2 SomME oF THE More Common Sort Mecuanics Tssts ! Simulative Tests Index Properties Tests 1 Empirical properties tests | Basic properties tests Mechanical analysis Standard ‘Proctor’ com- | Consolidation Sieve paction (moisture-density | Direct shear Hydrometer relations) Triaxial shear Elutriation California bearing ratio Permeability Specific gravity of solid par- | Various bearing tests ticles Natural water content Degree of saturation (cohe- sionless soils) Liquid limit Plastic limit Unconfined compressive strength Atterberg limits 1 See textbooks in bibliography for description of tests. primarily on the experience and the sensitivity of the soils engineer to differences in soil types. The recognition of similar soils as ex- pressed in classification tests is often sufficient to recommend similar soil treatment for different projects. More interesting from the standpoint of soil mechanies techniques are the properties tests. These are tests to determine isolated soil properties which have a direct bearing on soil behavior. It is noted that they are simulative, which means that they attempt to repro- duce, on a reduced and measurable laboratory scale, phenomena which occur or will occur in the prototype soil mass. The triaxial shear test is a good example. Here a cylindrical sample of soil is stressed axially under controlled lateral confinement. This is a close approximation of an unbalanced stress system operating on a buried cylindrical element of soil. This type of soil test yields strength moduli (cohesion and angle of internal friction) which are then applied to the analysis of the full-scale stability problem. The triaxial shear test is, therefore, a basic properties test. The empirical properties tests are also simulative. The usefulness of Cu. 5] ATTERBERG LIMITS 101 these tests lies in the empirical correlation between field performance and test data. Inquiry as to “why” does not form a necessary part of its application. As an example, the standard “Proctor” compaction test may be cited. This test is used to determine the compaction properties of soil when placed as fill. The basis for this test lies in the well-known fact that a given soil, depending on its moisture con- tent, will compact to different densities with the same amount of roll- ing. The Proctor test is, therefore, a standardized test which deter- mines the optimum moisture content for maximum density by pound- ing into a cylindrical mold with a standard weight, dropped from a standard height, representative soil samples with different moisture contents. This operation simulates the energy transmitted by the tamping of a sheep’s-foot roller on the soil in the field. The data from the Proctor compaction test are applied directly in construction, with- out further analysis, merely by specifying that all fill be compacted with the determined optimum moisture content. ATTERBERG LIMITS With increasing water content a clay changes consistency and passes from a solid state through a plastic state to, finally, a liquid state. Hach soil possesses a rather characteristic set of limits to these three states. These limits of consistency are arbitrarily fixed by a standard- ized testing procedure which was first proposed by Atterberg and which has come to be called the Atterberg limits (Terzaghi and Peck, 1948, pp. 32-36). The water content defining the upper limit of the plastic range is called the liquid limit, and that defining the lower limit is the plastic limit. The numerical difference between these two limits for any soil is the plasticity index of that soil. Statistical studies of the Atterberg limits of many clays (Casa- grande, 1947) have shown some interesting relationships among a num- ber of soil properties. It has been noted, for example, that the larger the plasticity index of a soil, the greater is its plasticity, its compressi- bility, and its dry strength. In addition, it has been found that, when the plasticity indices and liquid limits for a large number of clay sam- ples coming from the same bed or from geologically related deposits are plotted on a graph, the data define a straight line (Fig. 1). Fur- thermore, the linear plots of clays of different geologic origin occupy different areas on the graph. It is also noteworthy that all the lines in Fig. 1 are roughly parallel. 102 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 Gumbo clays (Miss., Ark., Ie all Wig gy vA Organic silt 200 and clay Z as ushing 4 nil 50 100 + W, ene earth 0 (Calif. and Mass.) a = > 0 100 200 300 400 500 SG Liquid limit = 20 Kaolin-type clays ms (Vera, Wash. and a South Carolina) 20 Glacial clays Prisiz GEHL Organic silt and clay {besten Detroit, ‘iy we ‘ ene Meadawe: EE 1 icago, Canada) rganic clay ew London, Conn.) 10 go, Canada). Lae 0 10) 10 20 30 40 50 60 70 80 90 100 110 Liquid limit Fic. 1. Relation of liquid limit to plasticity index. (After Casagrande, 1947, p. 803.) SLOPE-STABILITY ANALYSIS When a deep road cut is designed, how does the engineer know that the slope will be stable? Or, for that matter, how can the geologist determine the quantitative conditions for landslides and the formation of land forms developed from mass displacements of soil? To answer this type of question a number of methods for the analysis of slope stability have been devised in soil mechanics. The following exposi- tion is only a brief account of the basic principles; the standard textbooks should be referred to for more detailed explanation. To determine the stability of a soil slope against sliding, the dis- tribution of soil types behind the slope and in the toe of the slope must be known, as well as the position of the water table. Undisturbed samples of each soil type are obtained, and laboratory tests are made to determine their unit weights (weight per unit volume), cohesion, and angle of internal friction. In general, it can be said that a slope will slide if there exists some surface within the embankment along which the resultant of all shear- ing stresses exceeds the total shearing resistance along that surface. To determine whether this is a possible condition for a given slope, a graphical method is widely employed. On a geological cross section (Fig. 2) taken at right angles to the slope, a possible sliding surface in the form of a circular are is drawn more or less at random. In Fig. 2 this circle has its center at O. The circular-are sliding sur- Cu. 5] SLOPE-STABILITY ANALYSIS 103 face is considered valid because detailed landslide studies have con- firmed that most surfaces of sliding are strongly concave upward, al- though their exact form may be conditioned by such geological factors as bedding, fissures and fractures, the shape and distribution of dif- ferent types of rock and soil, and the local concentration of pore-water pressures. The force tending to produce sliding in Fig. 2 is the weight W of the W=Weight of soil Disturbing moment WD, S= Shearing strength Resisting moment=S)Li+S2 L2+S3L3+W2 D> resisting moment Factor of safety =———_—_—_ disturbing moment Fic. 2. Slope-stability analysis. See text for description; also Taylor (1948, pp. 406-479). wedge of soils lying above the arcuate sliding surface and to the right of the vertical through the center O. It is measured as a moment about the center O of the circle. The force tending to resist sliding, or the resisting moment, is the total shearing resistance which can be mobilized along the entire length of the sliding surface plus the weight W’ of the soils lying to the left of the vertical through the center O and above the sliding surface. It is also expressed as a moment about the center O. The lever arms for both moments are the horizontal distances between the centers of gravity of the two soil masses and the vertical through the center O. In this analysis the total shearing resistance is computed by using Coulomb’s equation, in which the values for cohesion and angle of in- ternal friction are known from the testing results, and the normal pressures on the sliding surface are scaled or computed from the draw- 104 KayE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 ing. The ratio of the resisting moment to the disturbing moment is the factor of safety of the circle. To complete the analysis, other circles are drawn and their factors of safety are computed. The circle having the smallest factor of safety is called the critical circle and is the surface most likely to fail. The factor of safety of the entire slope is that of the critical circle, and for a stable slope the factor of safety must exceed 1. A number of factors limit the accuracy of the stability analysis. In the first place, the cohesion and angle of internal friction of many clays are far from constant, and the determination of their values is subject to much uncertainty. In fact, studies of many slides in homo- geneous clays have indicated that the stability analysis can be ap- preciably simplified by assuming that the shearing resistance along the sliding surface is equal to half the unconfined compression strength of the clays. Secondly, in some circumstances the exact value of the stresses that are effective in any plane though the soil mass cannot be ~ readily appraised, particularly when the soil is subject to seepage forces. In fact, when any part of the sliding mass lies beneath the water table, seepage forces have to be taken into consideration in the stability analysis. This calls for the construction of a flow net, an operation often characterized by considerable uncertainty. Thirdly, it is difficult to evaluate the many geological factors, such as the structural heterogeneities (cracks, fractures, cemented layers, small pervious beds, etc.) present in many soils. However, despite these limitations, shearing strength analyses play an important role in soil mechanics and have been used with considerable effectiveness. CONSOLIDATION The slow compression of saturated clays under load is a deformation problem based on the hydraulic properties. rather than on the strength properties of the soil system. There has been deep interest in the solution of this type of problem in soil mechanics, and of direct interest to the geologist is that out of it has been developed a technique of value to historical geology. It will be recalled that soils are a skeleton of solid particles with fluids filling the interspersed voids. If a volume of soils with voids entirely filled with water is confined laterally and subjected to vertical pressure, any reduction of volume (strain) occurs only if either of the following conditions is fulfilled: (1) if there is a reduction in volume of the solids making up the soil skeleton, or (2) if there is a reduc- tion in volume of the voids. As it is assumed that the soil grains and the water within the voids are incompressible, a reduction in volume of Cu. 5] CONSOLIDATION 105 the soil mass can come about only by the extrusion of some of the pore water. This process of the reduction in the volume of soils due to the extrusion of pore water is called consolidation in soil mechanics, a usage not to be confused with that of geology. Because the rate at which water flows through clays is generally very low, there is a considerable time lag between the moment of application of a stress to the clay and the completion of the strain. This time lag is a function of the permeability of the soil. _In the laboratory, saturated sam- ples of clays are tested for consolida- tion in a device called the consol- idometer. In this device a small disk of soil, confined on the sides and bounded top and bottom by porous stones allowing free drainage of the sample, is subjected to load- ing. The consolidation, or change in volume, is measured by the displace- ment of the top surface of the sam- ple. The loading is done in incre- ments, and each succeeding load is appled only after consolidation has ceased for the previous load. The sleaen pieeacale) data from this test are generally peg Wesehdae AP dsm plotted on semi-logarithmic paper, preconsolidation pressure. (After with the total consolidation for each Casagrande, 1936.) load expressed by void ratio e (the ratio of total volume of voids to total volume of solids in the sample) as the ordinate and the load per unit of area or pressure p as abscissa. For the geologist, the interesting point is that it has been established that the shape of this e-log p curve (Fig. 3) reflects to a certain extent the history of loading to which a clay had been subjected during the geologic past. Clays, which at one time had been covered by the weight of a thick overburden or of an ice sheet, or in which there have been fluctuations in the water table, undoubtedly retain residual effects of this former early consolidation. Figure 3 represents a typical e—log p curve of a clay with a previous history of consolidation. The break in the slope of the e-log p curve is considered to be the expression of this earlier preconsolidation, and the pressure value corresponding to this break in the curve is roughly that of the former load. To establish the preconsolidation pressure more precisely, Casagrande (1936) has ' suggested a construction in which point b (Fig. 3) indicates the pre- Void ratio e O 106 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 consolidation pressure p. This point is found in the following manner. At point a, which is the point on the e-log p curve with the smallest radius of curvature, a tangent aB and a horizontal line aD are con- structed. Point 6 hes on the bisector aC of the angle between these two lines and is the intersection of the bisector with the upward ex- tension of the steeply dipping straight part of the e-log p curve. In an interesting paper on some postglacial clays, Skempton (1948) has recently demonstrated that this method for estimating precon- solidation loads is essentially correct. Other workers, however, have experienced difficulties in reconciling the computed preconsolidation loads with what was known of the histories of the clays. Greater interest of geologists in this work may reveal that present imperfec- tions are due to limitation in our knowledge of the properties of clay and in our knowledge of the complexities of the geologic past. RATIONAL BASIS OF SOIL MECHANICS The rational method of soil mechanics is a process that should be familiar to the geologist. The long road that starts with facts, leads through operations, and terminates with inductive inferences—the road that is typical of much of geological thought—also characterizes soil mechanics. At the beginning of every new job the engineer is confronted simply with a certain volume of soil, its properties unknown, its exact dimen- sions depending more upon the nature of the construction and to a lesser extent upon the nature of the soils; for instance, heavy structures stress soils to greater depths than light structures, ete. The engineer’s problem is, therefore, to determine the pertinent physical properties of the unknown, which is the soil mass. This is done, as we have seen, by (1) observation, (2) sampling, (3) testing, and (4) analyzing. As a result of the sampling, testing, and analytical procedure, the en- gineer hopes that he has found out all he has to know about the stress- strain relationships of the soil mass in order to insure the safety of his structure. However, he arrives at this position by virtue of a number of assumptions, which, it must be acknowledged, may affect the ac- curacy of his prediction. It is assumed, for example, that all sig- nificant soil types in the mass are known and that the samples tested are typical of each type and are in the same condition as the soil in place. Furthermore, it has to be assumed that the mechanical theories applied are pertinent to the problem and that the rigors of the mathe- matics are applicable to the statistical qualities of the soil mass. These qualifications do not, however, affect the validity of the analytical Cu. 5] THE ROLE OF GEOLOGY IN SOIL MECHANICS 107 process. All applied sciences involve a similar chain of facts and in- ferences. If soil mechanics not infrequently falls short of the ac- curacy that numerical answers imply, so does geology in its more quantitative aspects. Ore is not always found where predicted by the mining geologist on the basis of his geometrical projection of strike, dip, and fault displacement. Geologists recognize that in such cases the error does not lie with the geometry or with the process of analysis, but rather with the fact that, for lack of a thorough understanding of what happens to rock under all conditions, the geologist is forced to rely on simplified and idealized assumptions. THE ROLE OF GEOLOGY IN SOIL MECHANICS There is little need to emphasize the close relationship between geology and soil mechanics. Soils form the raw material of both sciences. Soil mechanics has, however, limited its interest principally to the behavior of soil, whereas geology has confined its interest to the origin and, to a lesser extent, to the substance of soil. The division between these interests is not always clear-cut; and the origin, sub- stance, and behavior of soils are probably as intimately interrelated as are the human body and the human mind, two man-devised sub- divisions which tradition has somewhat arbitrarily relegated to the physician and the psychiatrist. A study of the most recent soil mechanics literature reveals that soil mechanics is looking more and more to geology to explain cer- tain phases of soil behavior which heretofore had been taken some- what for granted. Unfortunately, the soils engineer is rarely a geol- ogist. Indeed, he generally has all he can do to pursue the com- plexities of his own specialty without embarking on those of geology. Geologists are, therefore, needed to cooperate with soils engineers if modern soil studies are to be carried out effectively. There are three ways in which the geologist can help the soils engineer to arrive at a better estimate of behavior: (1) The geologist can determine the type and degree of anisotropy (variability) in soils. This is especially important in stability and hydraulic studies. Minute details of fissuring, stratification, and changes in texture, which may not be noticed by an untrained observer, are commonly of great importance to the strength of a soil mass. Such details can be detected more readily by the geologist, partly because of his habitual concern for such detail and partly because of his ability to deduce such detail from considerations of the origin of the deposit. The slow deepening and widening of shrinkage fissures in some clays 108 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 has caused many serious slides in deep cuts that had been perfectly stable for years. Testing of samples of these clays for shearing strength gives absolutely no indication of the specific weakness from which the soil eventually failed. Many, if not most, landslides and soil failures are due to the de- velopment of high pore-water pressures. The evaluation of the hy- draulic properties of soil is based primarily on geological conditions which often are so insignificant as to escape detection by all but the trained geological observer. Small pervious or impervious layers may influence the path of pore-water movement and the localization of pore-water pressures. The eventual failure may occur in a way en- tirely neglected in the stability analysis. In such cases, the strength of the predominant soil has no bearing on the strength of the slope, and it is the geological interpretation of structural relations that gives the clue to specific weaknesses. (2) The geologist can reconstruct the history of a clay deposit and in this way roughly compute the type, duration, and amount of the preconsolidation load. Knowledge of this kind is of value in inter- preting consolidation, settlement, and shearing-strength data in many clays. Surficial clays may possess very much higher strength than deeper clays, owing to changes of the ground-water table. With each lower- ing of the ground-water table, surface clays experience a renewal of consolidation due in part to desiccation and in part to the increase in load resulting from the reduction of buoyancy. This process may not necessarily affect the deeper clays. In situations of this sort, the engineer, when adequately forewarned, can resort to deeper sampling, where he may find important differences in soil strength. The geologist is inclined to wonder whether some examples of recent earth movement are not perhaps due more to the consolidation of clays than to tectonic disturbances. The recently reported findings of a Roman city many fathoms beneath the surface of the Mediterranean, just off the mouth of the Rhone River in France, may be an example of the extreme consolidation of thick, soft deltaic deposits rather than a case of crustal subsidence of the area. (3) The geologist can study the granular characteristics of fine- erained soils, particularly clays, and correlate his observations with data on mechanical behavior. The mineralogy of clays, the inter- eranular structure of clays, and the physicochemical characteristics of clays are frontiers of research which will in time yield information of the greatest usefulness to soil mechanics. It is well known that the shape of sand and silt grains affects to Cu. 5] THE ROLE OF GEOLOGY IN SOIL MECHANICS 109 some extent permeability, internal friction, and even the elastic prop- erties of soils. The intergranular structural arrangements of clay crystals are only imperfectly understood, although it is known that they have an important effect on strength. What are the formational environments, and what clay minerals produce the different cellular and flocculated arrangements of clay crystals? The mortar structure of many clays, that is, a structure characterized by an open skeleton of large non-clay mineral grains embedded in a matrix of clay minerals, may produce hybrid mechanical properties. Frictional strength may be imparted to the soil by the non-clay framework. A shifting of the contact points of these grains may suddenly destroy the strength of the framework, and a weakening of the whole clay mass may result. The chemical activity of the clay minerals is known to affect perme- ability and plasticity. The permeability of montmorillonite clays has been reduced by introducing sodium ions into a calcium-rich clay. A famous example of this was the treatment of the clay lining of the lagoon on Treasure Island for the 1939 Golden Gate Fair in San Francisco (Lee, 1940). The surface activity of clay minerals affects strength and perme- ability. Plasticity is an index property of primary importance. Highly plastic clays almost always bring with them difficult engineer- ing problems. The nature of all the factors affecting the plasticity and the cohesion of clays is yet to be determined. Is the theory of the adsorbed water layer on the surface of the colloidal clay crystals the only and entire explanation of cohesion, or will further research suggest other forces? The correlation of plasticity data, as expressed by the Atterberg limits, with geological observations may be of far-reaching consequence in furthering knowledge of the mechanical properties of clays. Geolo- gists will do well to recognize that plasticity is a property as im- portant to study in cohesive soils as are grain size, rounding, and min- eral content in the study of sands. The loss of the cohesive bond of many clays on disturbance and manipulation and the slow redevelopment of cohesion on standing (thixotropy) is a soil property concerning which are many ques- tions. There are a number of examples of the practical importance of this property. For example, the movement of a large soil mass on the shores of Lake Gerzen, Switzerland, was brought about by the weak- ening of the cohesive bond of a lake marl due to the vibrations set up by the blasting of the stumps of trees (von Moos and Rutsch, 1945). The phenomenon of the movement of pore water in soil under elec- 110 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 tric and thermal impulses has led to a number of attempts in recent years to drain and stabilize soils by electrical methods. Although it is widely believed that this is essentially a phenomenon of the elec- tric double layer, or the charged layer of oriented water molecules on granular surfaces previously mentioned in connection with plasticity, there is still need for additional investigation into the cause of electro- and thermo-osmosis before large-scale application becomes practicable. The list of questions on fundamental soil properties is long, and the solution of any of them would undoubtedly be of value to both soil me- chanics and geology. In the field of applied soil mechanics there is no reason to advocate that geologists should, or even could, replace engineers in the purely engineering application of soil mechanics. The geologist’s role will probably continue to be that of a consultant on engineering matters, and the final decision regarding design will un- doubtedly always remain the engineer’s responsibility. Greater aware- ness, however, of geologists of the utility of soil mechanics would re- sult in the training of more geologists able to answer in a quantitative way not only the engineer’s quantitative questions on soils, but also some of geology’s quantitative questions on earth materials. REFERENCES American Society for Testing Materials (1939). Symposium on shear testing of souls. (1944). Procedures for testing soil. American Society of Civil Engineers (1940). Selected bibliography on soil me- chanics: Manual of engineering practise, No. 18. Boston Society of Civil Engineers (1940). Contributions to soil mechanics, 1925- 1940. Burmister, D. M. (1938). A study of the physical characteristics of soils, with special reference to earth structures: Columbia Univ. Dept. Civ. Eng., Bull. 6. Campbell, F. B. (1939). Graphical representation of the mechanical analyses of soils: Trans. Amer. Soc. Civ. Engrs., June 1939. Casagrande, A. (1936). The determination of the preconsolidation load and its practical significance: Proceedings First International Conference on Soil Me- chanics, vol. 3, p. 60. (1947). Classification and identification of soils: Proc. Amer. Soc. Civ. Engrs., pp. 783-810, June 1947. , and Carrillo, N. (1944). Shear failure of anisotropic materials: Jour. Boston Soc. Cw. Engrs., April 1944. , and Fadum, R. E. (1940). Notes on soil testing for engineering purposes: Harvard Univ. Grad. School of Eng., Soil Mechanics Series, No. 8. Coulomb, C. A. (1776). Essai sur une application des régles de maximis et minimis & quelque problémes de statique, rélatifs 4 l’architecture: Mémoires par divers savans étrangers, Académie des Sciences, Paris. Cu. 5] REFERENCES 111 Fahlquist, F. E. (1941). New methods and techniques in subsurface exploration: Jour. Boston Soc. Civ. Engrs., vol. 28, pp. 144-160. Fellenius, W. (1936). Calculations of the stability of earth dams: Tvransactions, Second Congress on Large Dams, Washington, D. C., vol. 4. Glossop, R., and Skempton, A. W. (1945). Particle-size in silts and sands: Jour. Inst. Civ. Eng. London, Paper 5492, pp. 81-105. Grim, R. E. (1940). The clay minerals in soil and their significance: Proceedings of Purdue Conference on Soil Mechanics, Lafayette, Ind. Harvard University Graduate School of Engineering. Soil mechanics series. Highway Research Board. Proceedings of annual meetings. Houser, E. A. (1945). Colloid chemistry of clays: Chem. Rev., vol. 37, pp. 287- 332. Houwink, R. (1937). Elasticity, plasticity and structure of matter: Cambridge University Press, Cambridge, England. . Hvorslev, M. J. (1940). The present status of the art of obtaining undisturbed samples of soil: Preliminary Report, Committee on Sampling and Testing, Soil Mechanics and Foundation Division, American Society of Civil Engineers. International Conference on Soil Mechanics, Harvard University (19386). Pro- ceedings: 3 vols. Johnson, H. L. (1940). Improved sampler and sampler technique for cohesion- less soils: Civ. Eng., June 1940. Krynine, D. P. (1947). Sow mechanics: McGraw-Hill Book Co., New York, 450 pages. Lee, Charles H. (1940). Sealing the lagoon lining at Treasure Island with salt, Proc. Amer. Soc. Civ. Engrs., vol. 66, No. 2, pp. 247-263. Mohr, H. A. (1943). Exploration of soil conditions and sampling operations: Harvard Univ. Grad. School of Eng., Soil Mechanics Series. No. 21, 3rd ed. Peck, R. B. (1942). Earth pressure and shearing resistance of plastic clay: Proc. Amer. Soc. Cw. Engrs., June 1942. Preece, E. F. (1947). Geotechnics and geotechnical research: Proc. Highway Re- search Board, pp. 384-416. Purdue Conference on Soil Mechanics, Lafayette, Ind. (1940). Proceedings. Ruedy, R. (1945). Bibliography on soil mechanics 1940-1944: National Research Council of Canada, Ottowa. Rutledge, P. C. (1940). Description and identification of soil types: Proceedings of Purdue Conference on Soil Mechanics, Lafayette, Ind. (1944). Relation of undisturbed sampling to laboratory testing: Trans. Amer. Soc. Civ. Engrs., vol. 109, p. 1155. Second International Conference on Soil Mechanics and Foundation Engineer- ing, Rotterdam (1948): Proceedings, 6 vols. Skempton, A. W. (1944). Notes on the compressibility of clays: Geol. Soc. Lon- don Quart. Jour., vol. C, pp. 119-135. (1948). A study of the geotechnical properties of some post-glacial clays: Géotechnique, vol. 1, No. 1, pp. 7-22. Taylor, D. W. (1937). Stability of earth slopes: Jour. Boston Soc. Civ. Engrs., July 1937. (1942). Research on consolidation of clays: Massachusetts Institute of Technology, Department of Civil and Sanitary Engineering. (1948). Fundamentals of soil mechanics: John Wiley & Sons, New York, 700 pages. 112 KAYE. PRINCIPLES OF SOIL MECHANICS [Cu. 5 Terzaghi, Karl (1936). Stability of slopes of natural-clay: Proceedings of Inter- national Conference on Soil Mechanics, Harvard University, vol. I, pp. 161-165. (1943). Theoretical soil mechanics: John Wiley & Sons, New York, 510 pages. (1947). Undisturbed clay samples and undisturbed clays: Jour. Boston Soc. Civ. Engrs., July 1947. and Peck, R. B. (1948). Soil mechanics in engineering practice: John Wiley & Sons, New York, 566 pages. University of Texas Bureau of Engineering Research. Proceedings of Texas Con- ferences on Soil Mechanics and Foundations Engineering. U.S. Bureau of Reclamation, Engineering and Geological Control and Research Division (1946). Laboratory procedures in testing earth materials for founda- tron and construction purposes. von Moos, Armin, and Rutsch, Rolf F. (1945). Ueber einen durch Ge- fiigesto6rung verursachten Seeufereinbruch, Eclogae Geol. Helvetiae, vol. 37, No. 2, pp. 385-400. Winterkorn, H. F. (1947). Fundamental similarities between electro-osmotic and thermo-osmotic phenomena: Proc. Highway Research Board, pp. 448-455. CHAPTER 6 SEDIMENTATION AND GROUND WATER * FRANK C. FoLey District Geologist, Ground Water Branch U. S. Geological Survey Madison, Wisconsin Ground water is the water that occurs below the surface of the earth in the zone of saturation. Few, if any, places on the earth are devoid of ground water, though the quantity of it may be small. In most places it is sufficiently plentiful to form all or an important part of the water used by man to sustain life, and in many areas there is enough to operate industry or irrigate crops. The location, develop- ment, and distribution of ground water is a major industry. In the United States the total amount of ground water used in 1945 was about 20 billion gallons a day, nearly double the amount used in 1935 (Sayre, 1948). The almost universal occurrence of ground water in the crust of the earth makes it an element, either beneficial or detrimental, to be con- sidered in almost all engineering construction projects, in mining, and in petroleum production. It may have to be drained, pumped, or walled out at considerable cost to permit construction or to allow op- eration of projects. Disposal of oil-field brines also is a serious prob- lem in some areas. Ground water has played an essential role in the deposition of many ore deposits. The occurrence and recovery of ground water for beneficial use will be discussed in this chapter, rather than its significance in other activities. OCCURRENCE OF GROUND WATER Though ground water in usable quantities may be found in any kind of rock in which openings occur of sufficient size and continuity to al- low its passage, by far the most prolific aquifers are eertain types of sediments, especially stream-laid sands and gravels and cavernous limestones. Volcanic rocks of the Columbia Plateau, the Hawaiian * Published by permission of the Director, U. S. Geological Survey. 113 114 FroLEy. SEDIMENTATION AND GROUND WATER [Cu. 6 Islands, and some other parts of the world are sources of large quan- tities of ground water. Weathered and fractured metamorphic rocks In some areas are the only source of ground water, as, for example, in parts of New England and the Piedmont Plateau of the southeastern United States, but the quantities of water obtained from them are small and sufficient usually for only small individual domestic sup- plies. Intrusive igneous rocks also generally yield only small supplies. As the sediments are the most important of all rock types as reser- voirs from which ground water may be extracted, knowledge of their classification, origin, and transportation is essential for the practical development of most ground-water supplies. WATER-YIELDING CAPACITY OF SEDIMENTS To be of practical importance as a water-bearing formation, or more properly a water-yielding formation, or aquifer, the material must be © of such a character that water can move through it rapidly enough to furnish water to a well or spring. The amount of water that a sediment contains when saturated is the same as its porosity, usually expressed as a percentage by volume of the total volume of the rock. The amount of water, also expressed as a percentage of the total vol- ume, that the sediment will yield by gravity drainage is the specific yield of the material; that retained by the material is its specific retention. Estimates of supplies available in any formation based only on porosity, without consideration of specific yield, are likely to be entirely wrong. A thorough discussion of principles of the water- yielding capacity of rock materials is given by Meinzer (1923). PERMEABILITY The permeability of a rock or soil with respect to water is its ability to transmit water under pressure. Darcey (1856) verified the earlier work of Hagen and Poiseuille, in which they demonstrated that the rate of flow of water through capillary tubes is proportional to the hydraulic gradient and showed its applicability to permeable filter sands. The principle is stated as Darcy’s law and is sometimes ex- pressed as Os 2104 in which Q is the quantity of water discharged in a unit of time, P is a constant which depends on the character of the material, J is the hy- draulic gradient, and A is the cross-sectional area through which the Cu. 6] PERMEABILITY OF SEDIMENTS 115 water passes. The constant P is usually called the coefficient of perme- ability. It has been expressed in various units by many investigators. The coefficient of permeability used by the U. 8. Geological Survey was defined by Meinzer, who selected gallon, day, and square foot as the units most applicable to ground-water work. Meinzer’s coefficient is the rate of flow of water, in gallons a day, through a cross-sectional area of 1 square foot under a hydraulic gradient of 100 percent at a temperature of 60° F. It is adapted for field use by correcting for the prevailing temperature of the ground water in the area under study. Wenzel (1942, p. 11) lists ten different permeability units in use in the United States and conversion factors for changing these units into Meinzer’s units, or Meinzer’s units into these units. The term “coefficient of transmissibility” was introduced by Theis (1935, pp. 519-524) and is now in common use in water-supply work. It is Meinzer’s field coefficient of permeability multiplied by the satu- rated thickness of the aquifer, and it is particularly useful because it describes the ability of an aquifer as a whole to transmit water, an essential factor in determining the actual amount of water that can be obtained from an aquifer. It is also readily determined from data collected during pumping tests. PERMEABILITY OF SEDIMENTS The permeability of the sediments ranges from extremely low for clay, the most nearly impermeable sediment as a source of water sup- ply, to very high for coarse, clean gravel and cavernous limestone, the most productive of all aquifers. Clay is useless as an aquifer not be- cause it contains no water, for many clays have a porosity of more than 50 percent and are saturated with water, but because the interstices between the grains are so small that essentially all the water is held tightly by molecular attraction. Permeability of a sediment depends not only on the absolute size of the constituent grains, but also on the sorting of the grains. A clean, fine-grained sand may have a higher permeability than a coarse sand or gravel in which the spaces between the larger grains are filled with finer material. Coefficients of permeability have been determined in the hydrologic laboratory of the U.S. Geological Survey on many hundreds of samples of material from many states. Ten samples are listed in Table 1 to illustrate the relation between grain size, sorting, and permeability. The table includes the highest and lowest coefficients determined so far in the hydrologic laboratory, but undoubtedly other materials exist ‘with higher and lower coefficients. 116 FoLEY. SEDIMENTATION AND GROUND WATER [Cu. 6 TABLE 1 RELATION OF GRAIN SIZE TO PERMEABILITY Size of Grain (Percent by Weight) ; Coefficient Lab- sui Poros- &: pee Larger Smaller ent ity ote tory ‘ean 2.0- 1.0- 0.5- | 0.25- 0.125- | 0.062- shan Specific %, ability No. 20 1.0 0.5 0.25 0.125 0.062 0.005 0.005 Gravity (gal. per day j mm. | mm. | mm. mm. mm. mm. i per sq. ft.) mm. mm. 1,001 Babe 0.6 1.0 3.0 5.7 24.7 44.0 21.0 1.62 58.2 0.0002 2,278 Sone Sine RET 2.5 0.9 1.0 45.3 49.3 1.20 55:5 0.2 2,286 19.6 24.2 17.4 25.9 8.8 1.3 1.51 tthe 1.54 37.0 30 1,382 15.4 15.2 20.2 19.5 16.4 7.0 4.5 1.5 1.92 26.3 150 1,374 29.7 16.9 18.9 17.1 15.4 1.3 0.4 0.2 1.90 27.1 480 1, 562 17.9 31.4 OZ em 14.0 2.0 0.8 0.41 ait 1.63 31.4 1, 200 1,393 40.4 15.5 22.8 16.5 3.9 0.4 0.3 0.1 1.87 Plo? 2,600 2,325 75.2 8.6 9.4 5,2 0.7 0.2 0.21 er 2.06 23.4 4,200 1,564 68.2 14.8 11.8 4.2 0.4 0.1 0.21 Bae 1.86 25.6 12,800 2,241 90.0 7.9 1.0 1.0 0.1 0.1 0.11 Bere Se 38.0 90,000 1 Includes clay. It is apparent from the table that, in general, the coarser grained materials are more permeable than those with predominantly finer grain. The table also shows that the proportion of large grains alone is not necessarily significant, but that samples having a higher pro- portion of small sizes generally have lower permeability. Wenzel (1942, pp. 21-50) lists a comprehensive bibliography on permeability and laminar flow prepared by V. C. Fishel. The water-bearing properties of various rock types are not described in detail in this chapter, and the reader is referred especially to Meinzer (1923, pp. 117-148) for detailed descriptions. Special mention should be made of the limestones and associated rocks as aquifers. Though normally they are compact fine-grained rocks, they are excellent aquifers in many places where extensive open- ings have been produced by fracturing and solution. The capacity of the limestones and associated rocks to produce water is likely to be erratic even within a small area, for if a well does not happen to en- counter many or large fractures or solution channels it will have a small capacity. Because of the continuous character of most of the openings, the water from limestone formations is much more subject to pollution than is the water that moves through the interstices be- tween grains, as in sand and gravel. Cu. 6] GROUND-WATER STUDIES 117 GROUND-WATER STUDIES The study of the ground-water resources of any area is fundamen- tally geologic. Extraction of ground water after it has been located is dependent upon the character and attitude of the rock formations that contain it, and upon the climate and opportunity for recharge. As the sediments are the most important of all rocks as aquifers, the sedimentary processes that produced the aquifers are of prime sig- nificance. The first step in examination of ground-water resources of an area is to map the geology and to become familiar with the kinds of materials to be found. The stratigraphy of the whole section must be studied, with special emphasis on the possible aquifers. Study of the outcrops of aquifers is especially important in relation to recharge, because outcrop areas are the places where the water enters to replenish that which is withdrawn. Possible impermeable cover that might seal the recharge areas is important. For example, relatively impermeable glacial till impedes recharge in parts of the Cambrian sandstone aqui- fers that supply the artesian water in eastern Wisconsin. Impermeable strata also act as barriers which confine water under artesian pressure, or they may cause the water to emerge as springs by preventing down- ward movement of ground water, thus forcing it to move laterally. Indispensable tools in water-supply studies are well logs. Accurate logs ean be obtained only by examination of drill cuttings, where iden- tification of sedimentary materials and their stratigraphic location for correlation must be made, but electrical logs are useful as an aid. The U.S. Geological Survey and most of the state geological surveys main- tain files of carefully identified samples of cuttings from water wells. They form an invaluable source of information for determining the character of the formations, both aquifers and non-aquifers; for choos- ing the best locations for new wells and for estimating probable yields of water from them; for forecasting the type of well construction that will be most satisfactory in the area; and for pointing out probable difficulties that may arise during well construction. Methods for the quantitative examination of water resources have been developed and improved greatly during the past fifteen years. Basic mathematical concepts of the nature of the movement of water through permeable materials are analogous to the transfer of heat through a homogeneous medium (Theis, 1935). No natural aquifer is isotropic; therefore corrections, determined from the examination and - evaluation of the sedimentary characteristics of the material, must be 118 FoLEY. SEDIMENTATION AND GROUND WATER [Cu. 6 made to compensate for departures of the actual aquifer from the ideal isotropic aquifer. Grain sizes and sorting vary, both parallel to and normal to bedding planes. No standard method of applying correc- tions can be set up because each area has its own characteristics, and at the present time the effect of sedimentary changes in aquifers on transmissibility is not easy to evaluate. Research needs to be con- tinued in many areas where detailed quantitative studies are in prog- ress to determine the effects on transmissibility of such changes in lithology. WELL DRILLING AND DEVELOPMENT The consulting engineer and the well driller are continually en- countering problems in sedimentation during well construction and de- velopment. There are few places where wells can be drilled where there is not at least one of the problems of caving walls, quicksand, or water-bearing material that must be screened out or retained with a gravel wall and then properly developed (Bennison, 1947). A thorough knowledge of the sediments and careful examination of them as drilling progresses can prevent much erief and, by indicating proper drilling and development methods, may make the difference between a poor well and a good one. The selection of the proper screen for a well in formations where influx of sand must be inhibited requires a careful sieve analysis of the material of the aquifer. The best screen size seems to be one that will allow the finer grains near the screen to pass through during development of the well, leaving the coarser grains outside. This process increases the permeability of the aquifer near the screen and increases the well capacity. If an artificially gravel-walled well is to be constructed, the same careful analysis is necessary to select the ideal gravel size or sizes that will retain the aquifer material and yet allow maximum flow of water to the well. SEDIMENTS OF VARIOUS ENVIRONMENTS AS SOURCES OF GROUND WATER In the following description of sediments of various environments and their importance as sources of ground water, Twenhofel’s classifi- cation (1932, pp. 785-871) has been followed in general. Sedimentary deposits of almost all environments are sources of ground water, some much more important than others. The younger sediments are usually more prolific in yield, because cementation has not progressed so far; thus they commonly have a relatively higher permeability. However, Cu. 6] THE GLACIAL ENVIRONMENT 119 some pre-Cambrian sandstones are usable aquifers; for example, late pre-Cambrian sandstones in the Lake Superior district of northern Wisconsin yield water to wells in several places, though some of the water is highly mineralized. The Paleozoic sediments of the Central States area are important aquifers, and in places some beds of sand- stone are almost completely uncemented. CONTINENTAL ENVIRON MENTS—TERRESTRIAL THE DrsertT ENVIRONMENT True desert deposits play a part in the water supply of desert areas. Wind-blown sand may be a source of usable ground water. Desert sediments in the United States are usually intermingled with piedmont deposits and are not readily distinguished from them. The saline deposits of the desert environment cause the ground water in many places to be highly mineralized and so are a detriment to development of ground-water supplies. THE GLACIAL ENVIRONMENT Deposits of glacial origin are very important sources of ground water, and in some areas the only source. Though the good aquifers are water-sorted, and so are fluvial, they are normally considered to be of glacial origin. In glaciated areas underlain by crystalline rocks in parts of New England and in most of the North American pre- Cambrian shield area, the glacial deposits are the only source of ground water, or the only source of a sufficient quantity to supply more than small domestic wells. In much of northeastern North Dakota and northeastern Montana, the glacial deposits are the only source of ground water for municipalities and most farms. Glacial deposits have a wide range in their capacity as aquifers. Till is so heterogeneous in its composition that normally it has a very low permeability. The gravel and sand deposits of water-sorted glacial material form some of the finest aquifers known. The large quantities of water escaping from melting ice sheets laid down extensive deposits of sand and gravel in front of advancing ice, at the edges of ice sheets, and from streams flowing on and in the ice itself. Many preglacial valleys have been filled with sand and gravel and now form excellent aquifers of great economic importance. The buried valley of the pre- glacial Rock River in southern Wisconsin and northern Illinois is an example. The water supply for Dayton, Ohio (Norris, 1948), is ob- - tained from a buried preglacial valley. 120 roLey. SEDIMENTATION AND GROUND WATER [Cu. 6 Outwash plains, eskers, kames, and kame terraces all form important aquifers. The extensive outwash plains off the front of the Wiscon- sin end moraines in Wisconsin are good examples. In the Antigo area in northern Wisconsin the outwash plains furnish much water for municipal, domestic, and irrigation use. The problem of locating sup- plies of water from glacial deposits is a special field which requires a thorough knowledge of glaciology. CONTINENTAL ENVIRONMENTS—FLUVIAL PIEDMONT ENVIRONMENT Some of the sediments deposited on piedmont slopes are important aquifers. In much of the basin-and-range country of the southwestern United States they, and sometimes associated desert deposits, are the only source of ground water, and frequently the only source of water of any kind. Most piedmont sediments, however, are not good aquifers, for they are formed principally by sudden floods, which deposit large quantities of relatively coarse, blocky material, and in turn are overlain by finer sediments that clog the openings, making them relatively impermeable. Most of the good aquifers seem to be stream deposits of sand and gravel formed during times of rather quiet flow in and near the stream channels themselves. Such deposits are very heterogeneous in dis- tribution and are difficult to correlate, even for short distances. Many of the aquifers crop out along the upper edges of the piedmont slopes and so receive annual recharge, whereas others receive recharge by slower percolation of the water from less permeable materials sur- rounding them. Coarse talus material near the upper edges of the piedmont slopes may allow much water to reach the sand or gravel aquifers that transmit it toward lower areas. In most piedmont areas, ground water obtained at lower elevations is under artesian pressure, and flowing or near-flowing wells are common. An example of ground water in piedmont areas is that in Tooele Valley, Utah, described by Thomas (1946). Many other areas in southwestern United States have been studied by the U. 8. Geological Survey and its cooperating agencies and the results have been pub- lished. THE VALLEY-FLAT ENVIRONMENT Included here are sediments deposited in stream channels as well as on the flood plains of streams. The character of the sediments Cu. 6] THE LITTORAL ENVIRONMENT 121 depends on the type of material supplied to the stream, the climatic and topographic conditions at the source, and the extent of transportation. Channel deposits usually have a higher proportion of coarse material and therefore constitute better aquifers. The deposits are usually thin, and always lenticular in cross section. If a stream is actively aggrad- ing, aquifers of considerable extent may be formed. There are ex- tensive sand beds, primarily of glacial-outwash origin, in the valley of the Ohio River in the vicinity of Louisville, Kentucky, where large supplies of water are pumped daily (Rorabaugh, 1948). Much water is pumped from valley-flat deposits of the Platte River in Nebraska, Wyoming, and Colorado. The very extensive Tertiary deposits of western Nebraska, North and South Dakota, eastern Montana and Wyoming, and western Kansas are good examples of valley-flat sediments. They contain much per- meable material, particularly that close to the mountain fronts, and they form good aquifers in many places. At some distance from the mountain sources of material, much clay is interbedded with sand and, though the sands yield some water, the permeability is usually rather low. CONTINENTAL ENVIRONMENTS—SWAMP Tue Swamp ENVIRONMENT Swamp sediments are usually poor aquifers, for the materials are generally fine-grained. Peat and coal beds may be permeable. Lig- nite beds in western North and South Dakota and eastern Montana provide water to many wells. The water is usually colored brown and in many places is rather highly mineralized. Coal beds provide water for many wells and springs in Pennsylvania. MIXED CONTINENTAL AND MARINE ENVIRONMENTS Tue Lirroran ENVIRONMENT Sediments of the littoral zone have a wide range in composition— from shale or limestone to boulder deposits. It is usually unnecessary to distinguish aquifers of littoral origin from marine sediments as sources of ground water, because they are of limited extent, particu- larly at right angles to the shore line. In some places, however, recog- nition of their origin may be significant in their evaluation as water- bearing beds. 122 FOLEY. SEDIMENTATION AND GROUND WATER [Cu. 6 Tue Detta ENVIRONMENT Delta deposits are extremely heterogeneous, with great ranges in composition within short distances, both vertically and horizontally. Sediments are generally rather fine-grained and consist of sand, silt, clay, and vegetable matter. Where sand has been deposited, deltas are important sources of ground water. Delta deposits of existing streams are likely to have good recharge from the stream itself. Delta plains constitute large areas of the earth’s surface at present, and many of them are densely populated. The deltas of the Ganges, Bramaputra, Nile, Hoang Ho, Rhine, and Mississippi rivers are ex- amples of large, heavily populated deltas where ground water is ob- tained in large quantities and has been one of the factors in the de- velopment of the areas. MARINE ENVIRONMENTS The widespread, relatively uniform marine sediments are the most extensive and continuous of all aquifers in areas where they have not been disturbed by faulting and folding. The great area of Paleozoic marine sediments of the Central States contains many very important aquifers. The Cambrian sandstones of eastern and southern Wiscon- sin, Illinois, and Iowa provide water to wells in large quantity. A well in Madison, Wisconsin, with 240 feet of screen in Cambrian sand- stone, produced more than 3,000 gallons a minute of water of excel- lent quality. Most of the municipalities in the area underlain by this water-bearing bed depend entirely on it for water supply. The St. Peter sandstone of Ordovician age has long been famous as an aquifer in the North Central States, but ground water attributed to it in some places probably actually comes from other formations. The St. Peter is a poorer aquifer in Illinois than in Wisconsin, ap- parently owing to an increase in fine material in the sand at greater distances from the Wisconsin arch, in the core of which pre-Cambrian rocks are now exposed. The permeability of the Cambrian sandstones in Wisconsin varies from area to area but has proved to be surprisingly uniform over dis- tances of 20 miles, as shown in the results of pumping tests conducted in the Milwaukee-Waukesha area, Wisconsin (Drescher, 1948). Marine limestones, where they have become cavernous as a result of solution, are among the world’s most productive aquifers. The Eocene Ocala limestone of Florida and Georgia yields many millions of gal- lons of water daily to wells and springs (Stringfield, 1936). Lime- Cu. 6] REFERENCES 123 stone of Pliocene age in southeastern Florida has an average trans- missibility in the Miami area exceeding 2 million gallons a day per square foot (Parker, Ferguson, and Love, 1944). The Tertiary lime- stones of Puerto Rico are capable in some places of yielding several thousand gallons a minute to individual wells (McGuinness, 1948). REFERENCES Bennison, E. W. (1947). Ground water, its development, uses and conservation: Edward E. Johnson, Inc., St. Paul, Minn., pp. 219-282. Darcy, Henri (1856). Les fontaines publiques de la Ville de Dijon: Victor Dal- mant, Paris. Drescher, W. J. (1948). Results of pumping tests on artesian wells in the Mil- waukee-Waukesha area: University of Wisconsin, unnumbered report, Feb. 1948. McGuinness, C. L. (1948). Ground-water resources of Puerto Rico: Puerto Rico Aqueduct and Sewer Service (mimeographed). Meinzer, O. E. (1923). The occurrence of ground water in the United States: U. S. Geol. Survey, Water Supply Paper 489, 321 pages. Norris, S. E. (1948). The water resources of Montgomery County, Ohio: Ohio Water Res. Board, Bull. 12. Parker, G. G., Ferguson, G. E., and Love, S. K. (1944). Interior report on the investigation of water resources in southeastern Florida, with special reference to the Miami area in Dade County: Florida Geol. Survey (mimeographed), June 1944. Rorabaugh, M. I. (1948). Ground-water resources of the northeastern part of the Lowisville area, Kentucky: Louisville Water Co., Louisville, Ky. Sayre, A. N. (1948). Ground-water investigations in the United States: Hcon. Geol., vol. 43, p. 552. Stringfield, V. T. (1936). Artesian water in the Florida peninsula: U. S. Geol. Survey, Water Supply Paper 773-C, pp. 115-195. Theis, C. V. (1935). The relation between the lowering of the piezometric sur- face and the rate and duration of discharge of a well using ground-water stor- age: Trans. Amer. Geophys. Union, Pt. 2, pp. 519-524. Thomas, H. EF. (1946). Ground water in Tooele Valley, Tooele County, Utah: Utah State Engr., Bienn. Rept., 1944-46, Tech. Pub. 4, pp. 97-238. Twenhofel, W. H. (1932). Treatise on sedimentation: Williams and Wilkins Co., Baltimore, Md., 2nd ed., pp. 785-871. Wenzel, L. K. (1942). Methods for determining permeability of water-bearing materials: U. S. Geol. Survey, Water Supply Paper 887, 192 pages. Part 2 ENGINEERING PROBLEMS INVOLVING STRENGTH OF SEDIMENTS CHAPTER 7 SEDIMENTATION AND HIGHWAY ENGINEERING ARTHUR B. CLEAVES Professor of Engineering Geology Washington University St. Louis, Missourt A knowledge of sediments, whether they be unconsolidated strata or bedrock, can be vital to successful highway construction. The science of soil mechanics in which the physical characteristics of soils are studied, measured, and evaluated has already achieved signal success in helping the engineer toward better construction. The geologists, as a whole, have not as yet made the fundamental contributions of which they are capable. Soil and sedimentary rock origin, age, mineralogic composition, de- gree of weathering, and attitude in the earth’s crust are all functions of the physical characteristics observed by the soils engineer. The sedi- mentary processes and natural or artificial agents which have brought the sediments into the position where the engineer finds them afford logical explanations for the non-isotropic characteristics of those soils and rock strata. The density, layering, intertonguing, porosity, per- meability, and similar features of soils and sedimentary rock are directly related to the processes responsible for their origin and place- ment. The physiographic history of an area is fundamental to an under- standing of the general physical characteristics of the recent overbur- den and often of the configuration of the bedrock profile. A knowledge of structural geology prepares the geologist for in- terpretations of the effects of joints, faults, inclined strata with their infinite variations, weathered zones, and similar features with respect to the structures placed above, or excavations made in, them. A petrographic study involving X-ray and differential thermal an- alysis will determine whether the clay in a cut is the dangerous mont- morillonite, or the petrographic study alone will determine whether the stream gravel or bedrock selected for aggregate has deleterious re- - action in the presence of high-alkali cement. 127 128 cLteaves. HIGHWAY ENGINEERING PROBLEMS [CE 7 The applications of seismology and electrical resistivity methods for determining the depth to bedrock, and the position of the water table, when used in conjunction with check core borings, speed up preliminary subsurface surveys and reduce the guesswork in estimat- ing common and rock excavation. Through the utilization of these and other methods, and upon the basis of geological understanding, the geologist can determine the posi- tion of bedrock, the ground-water table, subsurface divides, possible perched ground water, and can fairly appraise the nature of the under- lying materials. Because the modern major highway is planned for the purpose of satisfying the principal traffic needs of an area, and because of the advent of modern excavation and earth-moving equipment, it 1s now mandatory to construct roads with gentler grades and straighter align- ment than ever before. In general, regional alignment is fixed for the modern high-speed highway, and only very minor local adjustments may be made. This means that, locally, sections must be built through terrain, often with undesirable physical characteristics such as swampy ground, mine subsidence areas, heavy cuts, and poor aggregates. Nev- ertheless, within the economic limits controlling the saan Dee the project must be built as laid out. The problems in the Appalachians are different from those on the West Coast, in the tropics or in the Arctic. It would be impossible to treat adequately the significant sedimentation characteristics of soils alone for such varied areas, yet an attempt will be made to outline some of the applications of sedimentation as they affect construction on and in the soil cover and sedimentary rock. Brief reference will also be made to slides and areas of subsidience in the hope that some of the lessons learned in recent years may be helpful to others. CONSTRUCTION ON SOILS In highway or airfield construction, soil is the overburden to the en- gineer, and he is less concerned with its origin and classification than he is with the equipment he can use to move it most efficiently and cheaply. The soils studies for airfields have, in general, been much more thorough than those for highways, but the time has come when equally careful studies of the soils must be made for roads. This ap- plies specifically to the subgrade and the placing of base courses be- neath the surfacing of modern superhighways. It follows naturally that the soils engineer will do a better job of interpreting his tests Cu. 7] RESIDUAL SOILS 129 if he knows whether he is dealing with residual, fluvial, lacustrine, glacial, or some other type of soil. There is no general agreement relative to the classification and identification of soils; hence no common usage of terms and their meanings exists. Soils may be residual or transported. If the former, they may be mature or immature; if the latter, they may be fluvial, lacustrine, eolian, voleanic, or glacial. In each of the various types there are infinite variations and rarely any approach to homogeneity. Jenny (1941) says that the soil is a physical system consisting of various properties that are functionally related. This relationship may be expressed as soil-function (climate, organisms, topography, parent material, time). Any one of the properties of a soil, porosity, density, etc., is determined by the independent variables listed in the parentheses. When one establishes and evaluates the independent variables, the soil type is fixed. However, by the very nature of their origin, there can be no single set of physical constants for any sedi- mentary deposit. A sample taken and tested from a layer located between bedding planes gives no hint of the weakness in a sample which includes the bedding plane. In an alluvial fan, glacial deposit, beach, delta, and many other types of deposits, the intertonguing, crossbedding, variable laminae, grain-size variability, mixing, and the infinite variations and successions of materials of different densities, porosities, and permeabilities make a typical sample a figment of the imagination. Continuous sampling and extremely close spacing of borings such as to show the true physical characteristics would not be an economic feasibility. Even then the weakest layer would be the determining factor governing the settlement of a fill or any super- posed structure. In spite of these apparent unsolvable obstacles, the soils-testing program is extremely important, and it certainly can and does indi- cate the direction that construction in and on soils must take. It must not be forgotten, however, that geologic factors associated with origi- nal deposition and, perchance, features located far beneath the tested zones may govern the permanent stability of a fill, foundation, or cut slope. RESIDUAL SOILS These vary from a few inches to many feet in thickness. They also vary in accordance with the nature of the parent rock, the slope, and the weathering processes acting on the bedrock. The types of proc- esses involved in the formation of residual soils do not evoke the con- cern of the engineer, but end products do. Irrespective of whether 130 cteaves. HIGHWAY ENGINEERING PROBLEMS [Cu. 7 the material acted upon is bedrock or a transported soil, the residual products that remain after weathering are very important. The prin- cipal end product is clay, and the clay minerals may be divided into three groups: montmorillonites, kaolinites, and illites (also called hydrous micas and bravisites) (Grim, 1942). In illites and montmoril- lonites a portion of the aluminum is replaced with iron, magnesium, or both. Potassium is an important element of the illites, which are thought to be the result of alterations occurring on the sea floor rather than after uplift above sea level. The montmorillonites and kaolinites are products of subaerial weathering but also may be produced by the action of hot solutions. The importance of determining which clay may be present is indicated in one respect by the characteristic tend- ency of montmorillonites to swell. Consequently the identity of a clay underlying a structure or a fill should be ascertained. If ordinary soil mechanics and petrographic methods do not suffice, the sample should be submitted to the geological clay specialist for identification. The depth to which weathering may penetrate in the formation of residual soils depends on many factors, but the principal one is time. In recent glacial deposits the effects of weathering may be negligible, but in the humid tropics it may extend to depths of 100 feet or more. It may be unusually deep along joints and faults in the bedrock. The structural attitude of the bedrock may profoundly affect, and cause extreme variations in, a residual soil cover. In flat-lying strata, if it is assumed that secondary structures and surface conditions exert no local control, the thickness of the overburden in a section may be more or less constant. A concentration of vertical joints, or even a single joint, may allow a “seamy” condition to develop. In a limestone, this can be serious inasmuch as these weathered joint areas, filled with residuum, may be from inches to many feet in width and can appreciably increase the costs of excavation. In steeply inclined strata, similar irregularities may exist. A series of limestone beds, steeply dipping, may have layers differ- ing greatly in solubility. Hence it is no uncommon condition to find limestone cropping out under one wing wall of a bridge, but at a depth of many feet under the adjacent wing wall. Recent electrical resistiv- ity exploration in Cumberland County, Pennsylvania, checked by bor- ings and later proved by excavation, showed beneath a single bridge abutment overburden thicknesses varying from 2 to 46 feet. Such irregularities proved to be the rule in this area and were attributable solely to the solubility of individual limestone beds often only a few feet thick. Sunilar conditions were found in road sections crossing steeply dip- Cu. 7] CONSTRUCTION ON SEDIMENTARY ROCKS 131 ping Triassic strata. The extremes, however, were not so great be- cause the series of strata consisted of clay shales, sandstones, and conglomerates. In this area some of the sandstones were most strongly influenced because of the solubility of the cementing ma- terial between the sand grains. In an area of sparse surface outcrops, a contractor must gamble in estimating the percentage of common and rock excavation, but a careful geological study of the area and intelligent interpretation of the subsurface findings can reduce the gamble to a negligible factor and, in consequence, the contract bid price. The variations in the physical characteristics of most transported soils may be extreme in both horizontal and vertical section. They are most significant in glacial deposits, alluvial fans, deltas, terraces, and beach deposits, but they may be less serious in eolian, flood-plain, lake-bed or deeper marine deposits. The rapid variations in the dep- ositional features and physical characteristics of these sediments in- troduce many complications involving moisture content, density, por- osity, permeability, and, ultimately, in bearing values. The many combinations embrace too many factors for detailing in this brief discussion. Nevertheless, an identification of the soil type and its origin in advance of construction eliminates many possibilities and permits intelligent analysis of those known to be associated with the particular soil types concerned. CONSTRUCTION ON SEDIMENTARY ROCKS The principal sedimentary rocks may be divided into those me- chanically deposited and those chemically or biochemically deposited (U. S. Bur. Reclamation, 1942). CHEMICALLY OR MECHANICALLY DEPOSITED BIOCHEMICALLY DEPOSITED Shale (consolidated clay) (A) Calcareous Siltstone (consolidated silt) Limestone (CaCOs3) Sandstone (consolidated sand) Dolomite (CaCO3-MgCOs) Conglomerate (consolidated gravel or Mar! (calcareous shale) cobbles, rounded) Caliche (calcareous soil) Breccia (angular fragments) Coquina (shell limestone) (B) Siliceous Chert Flint Agate Opal Chalcedony (C) Others Coal, phosphate, salines, ete. 132 ctEAves. HIGHWAY ENGINEERING PROBLEMS [Cu. 7 The ability to handle rock strata without the use of explosives de- pends on the physical characteristics and structural attitude of the layers. In the Appalachian region most sedimentary rock must be drilled and blasted. Exceptions are flat-lying coals and some shales and siltstones. These may often be loosened with a rooter and sub- sequently removed by carryall scrapers or elevating graders. In the shales and siltstones, the fissility of the rock and its horizontal posi- tion determine whether or not explosives are necessary. Many clay shales which must be blasted when initially encountered become de- hydrated so rapidly on exposure that they must be treated as soils shortly after uncovering or removal from their original sites. Many Cenozoic strata are so weakly cemented that they too may be exca- vated with conventional road-building equipment. Aside from drainage, slides, and subsidence problems, the chief ques- tions relative to sedimentary rocks involve the design of cut slopes. Until relatively recent years and the advent of modern excavation and earth-moving equipment, highways utilizing deep cuts in order to minimize grades were not practical. However, with the availability of such equipment and because of the tremendous demand made on roads by modern truck transportation, the excavation of deep cuts has become a necessity. The design of cuts depends mainly on the physical characteristics and structural attitude of the soil or rock to be excavated. Relative -to soils, Terzaghi (1929) states that . . . the stability of all our clay fills and clay cuts depends essentially on cohesion. Due to this fundamental fact, the factor of safety of slopes with a given inclination rapidly decreases beyond the critical height at which the soil can stand with a vertical face. Hence a stable fill of a certain height and con- sisting of a certain clay soil is no indication of stability in a fill of twice that height, with the same slope and consisting of the same material. In computing the factor of safety of a cut or fill, the curvature of the sliding surface must be taken into account, else the results of the computation may be very misleading. There does not appear to be, and may never be, sufficient empirical data to omit the advantages to be gained by soil mechanics testing of soils in proposed deep cuts. Fundamental data obtained relative to the stability of the slope in one soil cut may not apply to another adjacent to it because of the vagaries of normal sedimentation proc- esses. Such vagaries are more often the rule than the exception. In sedimentary rock cuts, even with horizontal strata, no general rule applies. Nevertheless, in massive-bedded, strongly cemented, horizontal sandstones, limestones, and dolomites, of generally homog- Cu. 7] CONSTRUCTION ON SEDIMENTARY ROCKS 133 enous materials and free from closely spaced, steeply inclined frac- tures and joints, slopes of 44 to 1 and even vertical are possible. In the Appalachians the strata are seldom devoid of joints; consequently in cuts 80 or 90 feet deep and deeper there is a growing tendency to excavate benches at varying heights above grade. The Clear Ridge Cut, near Everett on the Pennsylvania Turnpike, is one of the deepest highway cuts in North America. Here the bed- rocks are interbedded sandy shales and sandstones, striking normal to the direction of the road and dipping 53° from the horizontal. This cut is 153 feet deep, 2,600 feet long, 380 feet wide at the top, and 88 feet wide at grade. Two benches, the lower 23-feet wide and 30 feet above grade, and the upper 85 feet above grade, were provided for protective and drainage purposes. Access from the ends of the cut was provided to the benches for removal of accumulated debris. The slope from grade to the lower bench is % to 1, that between the benches °4 to 1, and above the upper bench *% to 1, flattening to 1 to 1 near the crest. Maintenance for falling rock fragments on the shoulders and benches has been negligible in nearly 9 years of operation. An unusual design involves a new cut on the Turnpike’s Philadel- phia Extension at the western approach to the Susquehanna Bridge. Here Triassic shales and sandstones dip toward the highway at angles varying from 34° to 45° and strike at an angle about 15° from the line of the road. A slope of 1 to 1 or steeper was originally planned, but it is apparent that such a slope would progressively cut off indi- vidual layers at grade. Because weak clay shales are interbedded throughout the sequence of strata, slides were invited by such a de- sign. Consequently the slopes are now designed to follow up the “ap- parent” dip on the “shingle-like” edges of the strata irrespective of whether the beds dip 34° or 45°. In this way every stratum on the slope is “toed in” below grade. A bench is planned at the top of rock about 45 feet above grade, and the 20 (plus or minus) feet of over- burden will be laid back on a 1% to 1 slope. In cuts excavated in more or less horizontal strata where massive- bedded hard rock is underlain by clay or soft shales, progressive de- terioration of the clays and shales permits undermining and collapse of the overlying strata. This condition is accelerated: when the over- lying hard rock is strongly jointed. A uniform slope design is obvi- ously an invitation to rock falls unless the slopes are extremely flat, which condition in cuts 60 feet and greater in depth may not be economically sound. Observations in western Pennsylvania indicate that a compound slope design is practicable. Here, by accident rather than by intent, one cut 90 feet high shows that relatively stable slopes 134 cLEAvVES. HIGHWAY ENGINEERING PROBLEMS [Cu 7 were developed in nodular clays and in weak clay shales on slopes between 2 to 1 and 3 to 1. In the overlying interbedded shales and sandstones, 14 to 1 and % to 1 slopes are stable. When sedimentary rock strata in a sidehill cut strike parallel to the cut but dip into the hillside away from the road, steep slopes are feasible. Vertical fractures and joints, on the other hand, may make it desirable to modify such slopes. Schultz, Cleaves, and Rutledge state (in press): . . . because experience indicates that, once started, there is no way to control deep deformational slides, the modern method of slope design is to determine the stable angle of slope in advance of excavation. It is simple to prevent the development of deformational slides by the expedient of adopting extremely flat slopes. Such practice could, however, involve the excavation of more material than if the slide were actually permitted to de- velop. Consequently it is essential to approach the problem from a quantita- tive point of view. Such a procedure involves close correlation of geology and soil mechanics. From the standpoint of soil mechanics the stability of slopes with respect to the possible occurrence of deep deformational slides depends chiefly on the following factors: shearing strength and density of the materials in question in relationship to height and slope of the banks. Shearing strengths are deter- mined in the laboratory, and the proper slope is found by correlating the shearing strengths of the various materials with the depth of the cut and their positions in the banks. It is obvious that weak rocks generally require flatter slopes than stronger materials. In regions of hard sedimentary strata such as the Appalachians the following statement of these three authors does not necessarily apply, but in other areas, especially those of less strongly compacted, consolidated, and cemented Mesozoic and Cenozoic strata it may be particularly apropos. If the rock is perfectly uniform with respect to shearing strength and density, the higher the banks the flatter the slopes required. If strong rocks are over- lain by weak materials, the stable slope for the former may be found by cor- relating their shearing strengths with the total height of the banks (weaker materials included). The slope for the overlying weak materials is determined in a similar manner, except that only this class of materials needs to be considered in determining the effective height of the bank. If weak rocks are overlain by stronger rocks, the slope must be the same as for a bank of equal height composed entirely of the weaker materials, but having the same densities as the materials in question. The above analysis takes no account of the influence of bedding planes, faults, fissures, and other geological factors. Bedding planes are zones of — Cu. 7] SLIDES 135 relative weakness and potential sliding, and it should be apparent that greater danger of sliding exists when they dip steeply toward the excavation than when the inclination is in the opposite direction. Close jointing may impair the shearing strength of an otherwise strong and competent rock. From these and numerous other examples which might be cited, it should be apparent that the influence of geology is very important. Hence the importance of corre- lating the results of soil mechanics studies with a detailed knowledge of geological conditions. SLIDES Landslides have always constituted a threat to engineering projects, and they cause great annual property damage every year. They as- sume greater importance in highway engineering than ever before be- cause of the deeper cuts and excavations essential in modern construc- tion practices. A large percentage of the literature is devoted to de- scriptions of slide occurrences, but relatively little embraces studies of the mechanics of slide movements and their classification. Ladd (1927, 1928) published analyses of slides and their relation- ship to highways, particularly types occurring in West Virginia, Ohio, and southern Pennsylvania. He classifies these as: flow-moment slides; slope-adjustment slides; small-scale adjustment slides; struc- tural slides; and slides involving artificial fills. In a later paper (1935) he discusses slides in relationship to railway construction. Here he sets up five major groups: (1) flows; (2) slope adjustments; (3) subsidence; (4) structural slides; (5) clay ejection from clay- filled caverns opened by cuts. Within these groups are 17 subdivi- sions based on material type and structural control. Ladd’s descrip- tions and discussions of control methods are more effective than his classification. The first distinctly scientific effort at a sound classification was made by Sharpe (1938). In this work Sharpe recognizes four principal groups, which are divided into subdivisions as follows: 1. Slow flowage 3. Sliding (a) Rock creep (a) Slump (6) Talus creep (6) Debris slide (c) Rock-glacier creep (c) Debris fall (d) Soil creep (d) Rock fall (e) Solifluction (e) Rock slide 2. Rapid flowage 4. Subsidence (a) Earth flow Sinking over mines, caves, etc. (b) Mud flow (c) Debris avalanche 136 cueaves. HIGHWAY ENGINEERING PROBLEMS [Cu. 7 The general treatment is excellent, and Sharpe shows how the classi- fication of mass movements of earth rest on variable factors such as (1) type, size, cause, and rate of movement; (2) water content; (3) type of material involved; (4) characteristics of internal friction and organization of material within the moving mass; and (5) relation- ship of the moving mass to surface material and substrata. The two principal types of mass movement, flows and slides, depend on the presence or absence of a slip plane. Within the limitations of a geolo- gist who makes no claim of also being an expert in soil mechanics techniques and the mechanics of soil movement, Sharpe’s work is excellent. Subsequent writers, including soil mechanics specialists, have not improved on his classification, although they have materially aided in presenting some idea of the mechanics of movement. Terzaghi and Peck devote considerable space not only to the theory relative to stability of slopes (1948, pp. 181-191) but also to stability of hillsides and slopes in open cuts (1948, pp. 354-371). They also suggest a classification for slides based on the types of soils in which they occur: (a) detritus, (b) sand, (c) loess, (d) fairly homogeneous soft clay, (e) clay flows, and (f) stiff clays. These are analyzed on the basis of the assumption of more or less homogeneity. In two ad- ditional groups more complex types are considered: (g) clay with layers or pockets of water-bearing sand; and (h) sudden spreading on clay slopes. They state that slides may occur slowly or suddenly, in any con- ceivable manner, but claim that they are usually due to excavation or to undercutting of the foot of an existing slope. Initiation of slid- ing may be caused by gradual disintegration of the soil, starting at hair cracks which subdivide the soil into angular fragments, or by increase of pore-water pressure in permeable layers, or by shock which liquefies the soil beneath the slope. They admit that, because of the variability of the factors and processes leading to slides, the condi- tions for slope stability generally defy theoretical analysis. In addi- tion they agree that secondary structures and conditions may invali- date the results of computations. For cohesionless dry sand, they express the slope factor of safety relative to sliding by the equation tan F tan 6 where the angle 8, made by the slope with the horizontal, is equal to or less than the angle of internal friction ¢ for loose sand. Cu. 7] SLIDES 137 They state that, in homogeneous, cohesive soils, material with . a shearing resistance, s=c+ptangd can stand with a vertical slope at least for a short time, provided the height of the slope is somewhat less than H, [the critical height]. If the height of the slope is greater than H,, the slope is not stable unless the slope angle £@ is less than 90°. The greater the height of the slope, the smaller must be the angle 8. If the height is very great compared to H,, the slope will fail unless the slope and @ is equal to or less than @. Terzaghi and Peck say that slope failures in cohesive material are preceded by tension cracks near the top of the slope and that at some subsequent time sliding along a concave surface occurs. When the failure on this surface intersects the toe of the slope or some point above the toe, the slide is known as a slope failure. If failure occurs on a surface some distance below the toe of the slope, it is: known as a base failure. In Sharpe’s classification both of these types would be termed slump. Terzaghi and Peck further point out that because failures of slopes are common during the construction period, such failures can be con- sidered as large-scale shear tests, thereby offering an opportunity for evaluating minimum shearing resistance and avoidance of similar failures by a design change in further slopes in that vicinity and ma- terial (see Fig. 1). Fic. 1. Slumping due to slope failure. (After Terzaghi and Peck, 1948, p. 182.) . . . The depth z, of the tension cracks and the shape of the surface of sliding are determined by field measurements. The line of sliding is then replaced by the arc of a circle having a radius r and its center at O. Equilibrium requires that lan Wil — Wale + sr dye from which — Wil — Wole aN a dye 138 ctEAvES. HIGHWAY ENGINEERING PROBLEMS [Cu.7 where W, is the weight of the slice akfe which tends to produce failure, and W2 is the weight of the slice kbdif which tends to resist it. For the determination of slopes with known shear characteristics the authors say: . it is necessary to determine the diameter and position of the circle that represents the surface along which sliding will occur. This circle, known as the critical circle, must satisfy the requirement that the ratio between the moment of the forces tending to resist the slide and the moment of the forces tending to produce it must be a minimum. Hence, the investigation belongs to the category of maximum and minimum problems exemplified by Coulomb’s theory . . . and the theory of passive earth pressure. .. . After the diameter and position of the critical circle have been determined, the factor of safety G; of the slope with respect to failure may be computed by means of the relation [see Fig. 1] (aN moment of resisting forces _ Wols + sr dies moment of driving forces Wilh 3s = wherein r represents the radius of the critical circle and dye, the length of the surface of sliding. Many variations are required in accordance with data such as soil differences and position relative to the water table. The computa- tions will serve as a guide for design in trying to ascertain probable failure. Because of the tremendous variations in soil, and soil and rock combinations, and even rock slides wherein the end results are similar to Sharpe’s “slump” for soils, or Terzaghi and Peck’s “slope failure,” the engineer must not place too much confidence in either soil mechanics or purely geological interpretations alone. In such problems these specialists form a natural team whose cooperation is essential to further slide studies. SUBSIDENCE As defined by Sharpe (1938), “Subsidence is movement in which there is no free side and surface material is displaced vertically down- ward with little or no horizontal component.” The causes of sub- sidence are varied, and their end result evokes serious concern in high- way construction. Chief among the types are settlement over coal mines, compaction of unconsolidated sediments through drainage of swamp lands, solution and removal by natural or artificial means of salt, gypsum, and sulphur, depletion of oil fields, lowering of the water table through excessive public and private usage, break-through over sinkholes in limestone areas, changes in the “active” layer in regions ee ee Cu. 7] SUBSIDENCE 139 of “permafrost,” and kettle-hole subsidence in recently glaciated areas. As problems relating to construction in “permafrost” areas are de- scribed in Chapter 14, they will not be considered here. It will suffice to say that no subsidence problems involving loss of bearing values in soils are more serious than those encountered in arctic “perma- frost” regions. See Muller (1947) and Hardy (1946). Subsidence in kettle-hole areas, because of their limited size and restricted regional distribution in glaciated areas, is of minor im- portance. Settlement is ascribed to melting of ice layers or lenses, insulated for long periods after general ice retreat by overlying ma- terials. Settlement over metal mines is not uncommon, but that over coal mine areas is of serious concern to railway and road con- struction, especially in the eastern part of the United States. In Illi- nois, where the coal layers are relatively flat-lying, subsidence has usually resulted in long, gentle sags. Break-throughs to the surface are rare, but ponding in the depressed areas is common. Here com- paction of the overlying sediments causes a squeezing of the water to the surface and annually results in considerable property loss, particu- larly in agricultural land. Structural damage to buildings occurs, and regrading of railways and highways is sometimes necessitated. Prob- ably nowhere in the United States has more study been given to this problem than in Illinois; see Rice (1940), Cady (1921), Herbert and Rutledge (1927), and Young and Stolk (1916). In Pennsylvania, cave-ins in the anthracite areas, where whole buildings and even sections of roads drop into the steeply dipping mine workings, are common. The bituminous coal fields in the same state have not caused such catastrophic conditions; nevertheless mine settlement areas and cave-ins have resulted in various construction safety measures. In the Pittsburgh region special grouting techniques have been developed to afford safety for major building construction. Not only are void areas present in the form of open rooms and gal- leries, but also in areas of total roof collapse and in “mine-gobbed”’ sections and in the cracked zones above settlement sections, different techniques, grout mixes, and variations in dry “slushing”’ are success- fully used. In the line of the original Pennsylvania Turnpike, where mined-out coals showed pitted cave-in areas, it was the practice to remove the coal, pillars, gobbed rooms, and collapsed roof debris to ' the full width of the highway and backfill with suitable material. The depth to which excavation extended rarely exceeded 40 feet. Gal- leries outside the line of the roadway were bratticed off, and, when necessary, transverse drains permitting the original mine drainage to function were installed. A number of bridges were also located in 140 cteaves. HIGHWAY ENGINEERING PROBLEMS [Cu. 7 these areas, and their foundation excavations were carried beneath the former mine workings and beneath the underclay when it was present. No subsidence has occurred under any of these structures where such work was carried out. Where the coal underlay the high- way at depths greater than 40 feet, the coal beneath the right-of-way was reserved, and the wisdom of such practice was observed when in one stretch coal at a depth of over 100 feet was permitted to be re- moved. As a result, serious damage to an underpass and substantial settlement of the road resulted within a few months. In one small mine beneath the highway, dry “slushing” was re- sorted to. Here the areas beyond the road berms were bratticed off and the confined area filled with agricultural slag which was blown into position by compressed air. Sand, when available, may be handled in the same manner. Anticipated subsidence beneath the piers of an important railway structure in the Middle West was taken care of in a similar manner. The mined-out sections 165 feet beneath the viaduct were bratticed off outside the calculated limits of the angle of draw, and an 8-inch well-drilled hole was driven from the surface into the critical area. Hoses were led down this cased hole, and lean cement grout was placed in the danger area, completely filling it. Sand or agricultural slag would have accomplished the filling as well and would have been cheaper. Sinkhole regions present a serious problem when highways must pass above them. In country where the limestones are relatively flat-lying or massively bedded, it may be necessary to skirt such features where major caverns occur. However, in many areas the sinks are of limited extent, sometimes having considerable lineal ex- tent in response to bedding or major joint direction, but limited lateral extent. In these, filling with field stone to preserve the natural sub- surface drainage and topping off with graduated stone and soils in the road’s base courses may be sufficient. In such a case a reinforced- concrete slab gives added security. Nevertheless, in some places ac- tual bridging of sinks may be required. The most serious conditions develop when the overburden of soils conceals the cavernous condi- tion and construction proceeds with no anticipation of the cave-ins likely to occur later. It is possible, when such caverns are relatively large, that an electrical resistivity survey would accurately determine the position of them. Compaction of sediments is due to an assortment of causes, but com- mon ones are drainage of swamp and marsh lands and the construc- tion of fills on such ground. The same results arise from lowering a Op Cu. 7] THE SUBGRADE AND BASE COURSES 141 of the ground-water table by natural causes or from excessive domes- tic and public usage in restricted areas. In regions of crystalline or otherwise hard rock, settlement may be negligible, but in soils and weakly cemented strata it can be a serious problem. In any area where liquids are withdrawn from the pore spaces in sediments, re- adjustments are inevitable and must be planned for in construction. When marsh or swamp lands are floored by impervious material which in turn is underlain by free-draining strata, drainage by means of vertical sand drains is feasible provided that the free-draining layer is above the water table. These sand drains may be in the order of 28 inches in diameter, spaced 10 to 20 feet apart, and placed to depths of 40 or 50 feet or more (Hewes, 1942, p. 181). Sand drains of smaller diameter may be used. This method may also be used in deep soil cuts for the prevention of slides. However, when the base of the cut is close to bedrock, or the water table is high, the vertical sand drains must feed into a longitudinal drainage tunnel or into drains leading from the base of the sand drain to the gutter or longitudinal sub- drains beneath the highway shoulder. Peat and organic silts, which are highly compressible, should be entirely removed when feasible, by the use of either conventional excavation equipment or explosives (du Pont de Nemours, 1939). Settlement otherwise may be antici- pated, even though a prism of suitable material is used to replace the upper few feet. Subsidence of the surface over shallow oil fields (Johnson and Pratt, 1926) or sulphur mines, where the oil or sulphur is brought to the sur- face in a liquid state, may cause subsidence in the form of long, gentle sags, but sometimes with offsets of a foot or two at the margins. Here, again, ponding is the chief danger, because the subsidence may be so gradual over a relatively long time that highway and railway services may not be interrupted. In extreme cases regrading may be necessary. THE SUBGRADE AND BASE COURSES The modern heavy-duty highway is subject to more frequent and heavier wheel loadings than ever before; hence it is essential that most critical attention be paid to the underlying soils and base courses. - The soil beneath the road surface is known as the subgrade. Base courses or foundation courses are often placed between the road sur- face and the subgrade. The load-carrying ability of the subgrade is of vital importance, and it is often necessary to prepare these soils for the types of loads anticipated. The thickness, gradation, and 142 cLeAves. HIGHWAY ENGINEERING PROBLEMS [Cx. 7 other physical characteristics of the base courses are governed by the expected loadings and topographic and climatic conditions. Empiri- cal rules accepted in the past do not necessarily apply to the modern superhighway when heavy traffic is focused on particular highways. For example: the Lincoln and William Penn highways in Pennsyl- vania used to share the heavy interstate trucking, but the Pennsylva- nia Turnpike has bled off most of this and in addition has drawn heavily on traffic from the National Highway (Route 40) and prob- ably also the Mohawk Trail in New York. This means not only that the heaviest truck loads permissible under Pennsylvania state law use the Turnpike, but also that the frequency of their passage over any point is greater than on ordinary major highways. The relaxation interval of the subgrade of the Pennsylvania Turn- pike between loadings is probably less than for any similar route in the country. Inasmuch as this road is a forerunner of others to come, soils testing for such arteries should receive greater attention than ~ ever before. No apt comparison is feasible between the Merritt Park- way in Connecticut, because that highway excludes truck traffic. Nor is a comparison with major airfield flight strips, because on them the traffic is far less although the wheel loadings are greater. It is ap- parent that the recovery time for the subgrade is longer for flight strips. Determination of the subgrade soil characteristics is essential from the point of view of drainage and stability. These features are gen- erally related. “Stability means, essentially, resistance to movement under conditions of moisture or load” (Hewes, 1942, p. 165). In ac- cordance with the soil type, topographic position, climatic environ- ment, and artificially or naturally changed conditions, the moisture content and load-bearing capacity may be extremely variable. Hittle and Goetz (1946), investigating the factors influencing the load-carrying capacity of base-subgrade combinations, considered such variables as (1) soil type, (2) type of granular-base material, (3) depth of base material, and (4) seasonal moisture. In this study a cyclic-loading technique was developed by which measurements of the base-subgrade combinations were made and the elastic and perma- nent deflection characteristics of both the base and the subgrade ascertained. Their study shows that base stability is determined principally by the grading characteristics and density of the base ma- terials. Whereas seasonal moisture changes are of vital importance, so also are long-term climatic changes. A road constructed during a stretch of dry years may suffer serious deterioration when the cycle moves on Cu. 7] THE SUBGRADE AND BASE COURSES 143 into a period of wet years and the position of the ground-water table rises. Obviously the moisture content of the subgrade and base courses may be drastically altered and, in consequence, the associated bearing values. Middlebrooks and Bertram (1942), in a study of soil tests rela- tive to design of runway pavements, arrive at the tentative conclu- sion that, for testing the subgrade, the time-tested California bearing- ratio test is most suitable for flexible pavements and that, for de- termining the modulus of soil reaction and the use of this value in Westergaard’s center-loading formula, the field-bearing-test method is the most satisfactory for rigid pavements. They also make a very pertinent statement to the effect that no accurate methods for evalu- ating the true bearing value in soils affected by frost action are known. They also state: “Highway experience indicates that, in areas - which are subject to frost action the base course of non-frost heaving material under a pavement used for highway loads, should extend to a depth of at least 50 per cent of the average frost penetration, in order to provide suitable subgrade reinforcement.” The most important effect of frost action is in the reduction of the bearing values of the soils. In a critique of the above study, Campen and Smith (1942) make a point of the fact that, in laboratory testing of saturated, compacted soil samples, it should be borne in mind that soils beneath pavements are chiefly affected by moisture that reaches them from below and not through the essentially waterproof road surfacing. The effect of steady traffic is essentially that of vibration and con- stant shocks which result in a densification of the underlying soils. This is less in clays, where intergranular slippage is less because of its cohesive bond, than in more granular materials. In quicksands, vibra- tion or shock destroys the natural sand structure or bond in the sur- face layers and causes an upward movement of the contained pore water until the surface grains may actually be suspended, with a consequent loss in the bearing value of the sand. It is believed that a similar condition is approximated under modern high-speed high- ways. If there is a time lag after the passage of a heavy vehicle dur- ing which the subgrade can relax and regain the moisture “squeezed” to the surface, no permanent damage results. However, if there is not - sufficient time for the soil to recover this moisture from the surficial layers, it may travel in the direction of traffic flow until it passes laterally from under the pavement or reaches a joint. If it is assumed that some of the finer particles in the base courses and subgrade move along with this moisture and are lost through “pumping” at the 144 cLEAVES. HIGHWAY ENGINEERING PROBLEMS (Cu. 7 joints, fracturing of the pavement is inevitable. This condition is heightened during excessively wet times and is seasonal as well as cycli- cal over longer periods. A free-draining surficial base course which is not confined laterally by shoulder material that is too dense may provide a solution. It appears obvious that, if the soil moisture brought to the surface by constant traffic vibration cannot be recov- ered by the subgrade, it should be permitted egress in such a manner that the pavement is protected from permanent damage. At the present time we do not know enough about what takes place below the road surfacing. Certainly samples taken from the shoulder at the pavement edge do not help in the understanding of changes beneath the insulated and protected part of the highway, away from the edges. A volume of case histories of surfacing failures in rela- tionship to the thickness, sequence, gradation, and particle size of the base courses might be very illuminating. One instance is known in which under a porous surface the base courses graded from coarse to fine downward to the subgrade. The deterioration of the road with respect to bearing capacity of the surficial layers was extreme. Dam- age to the lower layers, however, was negligible; hence resurfacing with essentially waterproof material eventually restored this road to service. Space does not permit expansion on the great variety of problems relative to the vast amount of qualitative data available on the sub- ject in the literature. However, under the leadership of such original thinkers in the field of soil mechanics as Terzaghi, Casagrande, Rut- ledge, Taylor, and others, and with the research and quantitative data coming from the Bureau of Public Roads and the highway test- ing laboratories in many states, remarkable progress has been made. Constant awareness of the importance of soils testing in highway construction is essential. FUTURE RESEARCH The possibilities and scope of research, like the topic of sedimenta- tion in highway construction, are of such magnitude that one can merely “brush” the surface. It goes without saying that continuing research involving soils in relationship to highway construction is necessary. Some basic work concerning slides in soils has been initi- ated, but the need for much more dealing with all types of dry or wet mass movements of soils, soil and rock, and rock alone in varied climatic and topographic environments is indicated. Basic research on the bearing values of soils affected by frost action es Cu. 7] REFERENCES 145 is essential if construction in the arctic and northern latitudes is to succeed. A beginning has been made, but “permafrost” studies in relationship to construction have in a sense only begun. Additional studies of precompressed soils, such as those unloaded by erosion, exposed at the base of deep cuts, found in the beds of drained lakes and swamps, and under long-departed ice loadings, would be helpful. Investigations of the subgrade beneath existing major highway and airstrip pavements in relationship to similar natural soils adja- cent to these features might be worth while. Refinements of geophysical exploration of soils and closer correla- tion between geophysical and soil mechanics investigations is neces- sary. Already seismic and electrical resistivity studies have proved themselves in highway construction and foundation work. In some quarters caution is indicated, because at present overly enthusiastic acceptance of electrical resistivity methods may obscure the fact that check borings are essential and it is not a single “tool” but must be used in conjunction with others. Because of the varied topographic and climatic environments as well as the countless variations in consolidated and unconsolidated sediments of diverse origins and histories, the research possibilities are infinite. So long as men travel on or take off and land aircraft on the surface of the earth, new and continuing research will be re- quired. REFERENCES Cady, G. H. (1921). Coal resources of District 11: Illinois Mining Investigations, Ill. Geol. Survey, Coop. Mining Ser., Bull. 26. Campen, W. H., and Smith, J. R. (1942). Discussion on soil test for the design of runway pavements: Proc. Highway Research Board, p. 174. EK. I. du Pont de Nemours & Co., Inc. (1939). Blaster’s Handbook: pp. 132-137. * Federal Works Agency, Public Roads Administration (1943). Principles of highway construction as applied to airports, flight strips, and other landing areas for aircraft. Grim, R. E. (1942). Modern concepts of the clay minerals: Jour. Geol., vol. 50, pp. 225-275. Hardy, R. M. (1946). Permanently frozen ground and foundation design: Eng. Inst. of Canada, vol. 29, pp. 4-10. Herbert, C. A., and Rutledge, J. J. (1927). Subsidence due to coal mining in - Illinois: U. S. Dept. of Commerce, Bur. Mines, Bull. 238. Hewes, Laurence I. (1942). American Highway Practice: John Wiley & Sons, New York, vol. I. Hittle, J. E., and Goetz, W. H. (1946). Factors influencing the load-carrying - * Articles of comprehensive scope or containing extensive bibliographies. 146 cLEAvVES. HIGHWAY ENGINEERING PROBLEMS Gist, 7 capacity of base-subgrade combinations: Proc. Highway Research Board, pp. 521-548. * Jenny, Hans (1941). Factors in soil formation, a system of quantitative pedol- ogy: McGraw-Hill Book Co., New York, 281 pages. Johnson, D. W., and Pratt, W. E. (1926). Local subsidence of the Goose Creek Oil Field: Jour. Geol., vol. 34, Pt. I, pp. 577-590. Ladd, G. E. (1927). Landslides and their relation to highways: Public Roads, U.S. Dept. Aer. Pie lvolyseNon2ppy clot (1928). Jbid., Pt. 2, vol. 9, No. 8, pp. 153-163. (1935). Landslides, subsidences and rock-falls as problems for the rail- road engineer: Proc. Amer. Ry. Eng. Assoc., vol. 36, pp. 1091-1162. Middlebrooks, T. A., and Bertram, G. E. (1942). Soil tests for design of run- way pavements: Proc. Highway Research Board, pp. 144-184. * Muller, Simon (1947). Permafrost or permanently frozen ground and related engineering problems: Edwards Bros., Ann Arbor, Mich., 231 pages. * Rice, G. S. (1940). Ground movement and subsidence studies in mining coal areas and non-metallic minerals (a review of the work of 15 years and sugges- tions for future studies): Trans. Amer. Inst. Min. Met. Engrs, Coal Div., vol. 139, pp. 140-154. Schultz, J. R., Cleaves, A. B., and Rutledge, P. C. (in press). Geology in engi- neering practice, Chap. X XI, John Wiley & Sons, New York. * Sharpe, C. F. S. (1938). Landslides and related phenomena: Columbia Univer- sity Press, New York, 186 pages. * Terzaghi, Karl (1929). The mechanics of shear failures on clay slopes and the creep of retaining walls: Public Roads, U. 8. Dept. Agr., vol. 10, No. 10, p. 192. , and Peck, R. B. (1948). Sozl mechanics in engineering practice: John Wiley & Sons, New York, 566 pages. * U.S. Bureau of Reclamation (1942). Concrete Manual. * U.S. War Department, Office of Chief of Engineers (1943). Engineering Man- ual. * Young, L. E., and Stolk, H. H. (1916). Subsidence resulting from mining: Univ. of Ill. Eng. Exp. Sta., Bull. 91 (exhaustive world bibliography on sub- ject). * CHAPTER 8 FOUNDATION PROBLEMS OF SEDIMENTARY ROCKS SHAILER S. PHILBRICK Division Geologist, Corps of Engineers Department of the Army Pittsburgh, Pennsylvania The geologic map of the United States shows the great expanse of sedimentary rocks which underlie the region from the Piedmont to the Rockies. Because igneous and metamorphic rocks appear through the sedimentary blanket only rarely, most of the major dams and many of the larger buildings in this area are founded upon sedimen- tary rocks. The design of these structures is influenced by the type, degree of weathering, and depth through the overburden to the sedi- mentary rocks. The present physical and structural properties of the rocks are influenced by their inherent geologic characteristics. Hence an understanding of the foundation problems associated with a given rock type at one locality will be of assistance in using that same rock type as a foundation elsewhere. This inferred oversimplification ap- plies to the general problem of each rock type on which is super- imposed the special problems coming from local conditions of weather- ing, geomorphology, geologic structure, the activities of man, and the purpose for which the rock is to be used. Although it may appear that heavy hydraulic structures impose severe requirements on founda- tions, it is not uncommon to find more intense local loadings developed by industrial structures and high buildings. Foundation problems may be created by such diverse causes as reversible loadings occa- sioned by filling and emptying of reservoirs, release of load caused by excavation, and reduction in support of the ground surface by mining. Regardless of the cause of the problem, its importance, frequency, _and magnitude will be related immediately to the particular type of engineering structure and directly to the type of rock upon which the structure will be founded. It is, therefore, difficult to dissociate the problem of foundations from the implications of sedimentation. In fact, by the application of sedimentation to foundation problems 147 148 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 of sedimentary rocks these problems are framed and reduced to more understandable terms. Foundation problems can be reduced from numerous individual problems to two general problems the correct solution of which is paramount to the successful completion of any undertaking involving foundations. These may be phrased in many ways, but they boil down to two questions of which the first is always present and the second may be present: (1) Will the foundation rock support the load? (2) Will the foundation rock leak? The answers to these questions may be automatically affirmative in rare cases, but in general the development of proved answers will re- quire the expenditure of considerable money, effort, and thought. It is toward the reduction of these expenditures that the following re- marks are directed, inasmuch as the writer believes that foundation — problems can be classified rationally in relation to the type and mag- nitude of the structure and its geologic environment. There is a pat- tern in the foundation problems which he has met in the cyelic sedi- ments of the Pennsylvanian of the Appalachian Basin, just as there is another pattern in the calcareous sediments of the Tennessee Val- ley, another pattern in the Permian sediments of the Mid-continent area, and another in the Cretaceous and Tertiary sediments of the upper Missouri Valley. These comments are restricted to the rocks as we find them, with- out regard to the processes which developed their present characteris- tics, unless such processes are continuing and will affect the rocks ap- preciably during the lifetime of the proposed structure. This chapter considers briefly the types of foundations, the problems of determin- ing the physical character of the foundation rock, and the problems arising from the geological character of the several types of sedi- mentary rocks. FOUNDATIONS A satisfactory foundation must support the load imposed upon it. The requirements of support may include resistance to compression, shear, sliding, and maintenance of cohesion despite repeated freezing and thawing and wetting and drying. Commonly adhesion to con- crete is required. In addition to supporting properties, the founda- tions for hydraulic structures, such as locks and dams, must be rela- tively impermeable or be such that they can be feasibly and econom- Cu. 8] APPRAISAL OF FOUNDATION 149 ically rendered relatively impermeable. All foundations must be in- soluble during the lifetime of the project. These requirements apply- ing to the rock mass upon which the structure is founded or abutted are stated in greater detail by Houk and Keener (1941, p. 1116). Consider a dam. The foundation rock is under compressive forces occasioned by the mass of the structure. Shearing stresses are im- posed upon the foundation by the hydrostatic head acting horizontally on the upstream side of the dam. This horizontal stress may be con- sidered to produce a tendency to slide either on the surface of the foundation rock or on a plane of weakness at some depth within the foundation rock. A combination of shearing and sliding is usually considered in the design of gravity dams. The impermeability of the foundation is a necessity if the dam is to hold water, and yet the en- tire foundation from the upstream side of the dam to the toe need not be impermeable. Ideally it would be desirable if the foundation at the heel were essentially impermeable and downstream therefrom were permeable. This relationship would reduce to a minimum the area of the base of the dam on which full reservoir head could be exerted and correspondingly increase the base area subjected only to tail- water uplift. This type of foundation would permit a theoretically thinner section with consequent reduction in concrete and cost. The conditions of loading of a dam vary with the variations in the eleva- tion of the reservoir, and the foundation rock must be sufficiently elastic to withstand these changes. Consider a bridge pier or column footing. The foundation rock is under compressive stress, but with the flow of traffic and wind load on the bridge shearing stresses are developed in the foundation. How- ever, it need not be impermeable, but it should be homogeneous and must be competent to support the load. Consider a retaining wall. The foundation is subjected to the same stresses as the foundation for a dam, although the variation in stresses is greatly reduced. As in the case of a bridge pier, it need not be impermeable but it must be competent to support the load, and it is better if homogeneous. APPRAISAL OF FOUNDATION There are three methods of approaching the problem of determin- - ing safe values in bearing, shearing, sliding friction, elasticity, and permeability of a foundation: (1) experience and general practice; (2) assumption of an idealized homogeneous mass, the properties of which can be determined from extrapolation of the properties of rep- resentative samples; (3) recognition of the heterogeneity of the mass. 150 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 the physical properties of which may be indicated only within ap- proximate limits by the physical properties of representative samples and which therefore must be considered as generally indeterminate because of the effect of geologic flaws. Regardless of the approach used, the investigation of the foundation should be contimued from the planning stage through the construction stage when the founda- tion rock is exposed and the design assumptions can be re-examined. Experience and general practice include all the past knowledge of foundations derived from the construction of similar structures on similar foundations. And who is to say that one foundation is the same as another? The only ones who can speak with authority are those who had intimate knowledge of the foundation upon which the past structure was built and who are as familiar with the proposed foundation. And therein lies the weakness of foundation design. One never knows as much about the foundation while it is being consid- ered during design as one does about the same foundation during © construction. However, within limits and depending upon the type and magnitude of the structure, past comparable experience and the generally considered “good practice” of the engineering profession are not to be discarded as an inapplicable approach, because they represent time-tested data. The reliability of this method is a func- tion of the similarity of the foundations and the structures built thereon. It can be used safely only when all the data obtainable on the proposed foundation indicate values equal to, or higher than, those of the compared foundation. This is the common method used in the design of low dams on insoluble rocks, and it seems to be the method employed in fixing allowable loads on certain rocks in city building codes. It usually results in setting ultraconservative values with apparent ultimate safety factors as high as 20. The second method is based on the assumption of an idealized homogeneous mass, the properties of which can be determined from extrapolation of the properties of representative samples. This is a typical mathematical approach to a problem, and it is beloved by the stress analyst for it yields a clean answer. But there isn’t a clean answer. In the first place, sedimentary rocks are not homogeneous and isotropic. In the second place, who is to say that the samples are representative of the mass? Certainly the geologist will not say so; for he has studied the surface geology, observed the action of the core drill, logged the cores, noted the behavior of the injected water and grout during the pressure testing and test grouting, and descended the exploration shafts and pits to vestigate the character and structure of the rock in place. He has seen the heterogeneity of the bedrock Cu. 8] SUBSURFACE INVESTIGATIONS 151 and the variations in rock types over the short distances between the borings. He is aware that there are variations in dips of the beds and possibly displacements along faults with resulting breccia and gouge. And what constitutes a representative sample of such a mass? Is it the almost unobtainable sample of the gouge or the easily re- coverable, smooth-sided, unbroken length of solid core from the massive rock? If the properties of the massive rock are assumed to be typical of the foundation, one would feel safe in recommending high unit loads. However, if one were to recover a sample of the gouge, it would have to be subjected to the laboratory techniques of soil mechanics to obtain an idea of the highest limit of allowable unit loads. With these unit loads the logical assumption would follow that the site should be abandoned for a normal design or that a radical departure from normal design would be required with consequent in- crease in cost. All that is desired here is to emphasize that it would be most unusual to find a foundation of sedimentary rocks so homogene- ous that the physical properties of representative samples thereof could be extrapolated to the mass foundation. The third method recognizes that, in sedimentary rocks, foundations are composed of heterogeneous materials, the physical properties of which may be determined within limits, although in the final analysis the physical properties of the rock mass are generally indeterminate. Indications may be developed, however, which will permit limiting, rational assumptions as to the safe, allowable loadings and the reason- ably anticipated rates of flow of water. Under this approach it is assumed that the presence and orientation of geologic flaws, such as planes of stratification, joints, faults, random fractures, and solution channels in distinction from inherently differing lithologic character- istics, will be compensated in the equations by safety factors. Or it may be assumed that the geologic flaws will be minimized to the ex- tent that they may be disregarded through inclination of base of the structure, localized design to offset their effect on the mass of the structure, or specialized treatment to raise their properties to accept- able minima. SUBSURFACE INVESTIGATIONS All three of these approaches require that the character of the _ foundation rocks be known, but in varying degrees. In the first case, in which the designer will rely on past experience and general practice to set the allowable loadings, there may be made little or no attempt to determine the details of geologic conditions other than to ascertain from the general geology of the area the probable rock types to be en- 152 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 countered in the foundations. However, it would be unusual now to attempt the design of anything other than a very minor structure without core borings. The cores would probably be no larger than a nominal 2 inches in diameter, and the logs would be based on the driller’s idea of what the rock is, supplemented by the engineer’s or architect’s interpretation of what that means with respect to the clas- sification in the building code or in some handbook. In general the results of this approach to loadings under small structures, and in some cases under much larger structures, has been entirely satisfactory be- cause of the conservatism inherent in building codes and handbooks. There have been sufficient examples of the unfortunate results of this approach in connection with structures of ordinary size to make it advisable and almost mandatory to consider with some care the foundation reactions under larger structures. As a result, emphasis is placed upon determining geologic conditions more clearly and exactly. Drilling tools are available with which almost 100 percent — recovery can be obtained in nearly all types of sedimentary rocks. This equipment ranges from the diamond-bitted, modified standard, double-tube core barrels operated by hydraulic-feed core drills for rock drilling to ingenious soil samplers used in the sampling of uncon- solidated soils and sediments. There are the large cores recovered from borings ranging from 6 inches to over 48 inches in diameter. These larger cores, formerly considered only awkward by-products of the drilling of exploration shafts, provide a source of samples for laboratory tests superior to those from smaller holes. Samples and check samples may be drilled from the large cores in the plane normal to the axis of the core, as well as parallel to the axis of the core. It would be possible to evaluate roughly the sliding friction along bedding planes with some degree of similarity to conditions underground with the use of contiguous samples of the larger cores. In other words, samples of almost any type of foundation rock can be recovered, even very soft ones. But, the samples having been recovered, there remains the evaluation of the representativeness of these samples. In a massive foundation the samples are probably representative. But most sediments are not massive. Or, if they are massive geologically, they may not be massive in the engineering sense that the entire column of foundation rock sub- ject to reaction to the imposed load will react as a homogeneous mass. Bedding planes in a horizontally bedded rock will affect the foundation reaction, and the closer the spacing of the bedding the more the re- action will be affected. The shear-friction factor can well be the con- trolling factor in the foundation design of a structure founded on hor- Cu. 8] MATERIALS TESTING 153 izontally bedded, average shale. In considering this factor it is neces- sary to obtain quantitative information on the shearing strength, and sliding friction along bedding planes. If the samples obtained from the borings are approximately the lengths of the intervals between the major bedding planes, the laboratory tests will not produce these data. The samples therefore would not be representative. The crite- rion, then, of whether a sample is representative of foundation condi- tions is not whether the sample came from a given elevation in a boring in the foundation but whether that sample contains the critical ele- ment—be it structural, stratigraphic, or lithologic—which is present at that elevation. MATERIALS TESTING It is common practice to investigate the physical properties of the foundation materials by laboratory tests in order to obtain data to be included in the computations relating to the stability of the struc- ture and to define the characteristics of the foundation rock. The stability calculations require the determination of the shearing and compressive strengths. It is desirable to include tests to indicate the coefficient of friction between concrete and the proposed bearing sur- face of the rock, as well as between rock surfaces on either side of the planes of weakness on which sliding friction may be a factor in the design. More recently there has been added to these three tests a test for the purpose of measuring the unit tensile strength of the rock. By means of this test, in conjunction with the unit compressive strength of the material, the angle of internal friction or cohesion of the rock may be calculated. Certain routine tests to determine the total ab- sorption, specific gravity, void ratio, and porosity are primarily de- signed to provide quantitative data on the general characteristics of the rock. These do not enter into the calculations involving stability, but they may be used in indicating consolidation under loading or ex- pansion with decrease in loading. A test to determine the soundness and durability of the rock, and its susceptibility to weathering when exposed, is called the weathering test and is of considerable use in in- dicating the relative cohesion and the necessity for protecting the - rock from deterioration during the period of construction. This test also provides an indication of the reliability of the bond that may be obtained between concrete and rock. Some compaction rocks weather rapidly and in a matter of a few hours may develop a thin coating of fine, discrete particles which would form a film between the rock 154 pHILBRICK. FOUNDATION PROBLEMS (Cu. 8 surface and the concrete, thus preventing the formation of a good bond between the concrete and the rock. The shearing test consists in mounting a test specimen in two cast- iron blocks in such a manner that a single shearing load may be freely applied to the required shearing plane of the test specimen. In a core of horizontally bedded rock drilled normal to the bedding plane, a load is applied parallel to the axis of the core by means of a ecal- ibrated spring, and the load parallel to the bedding is then applied to one of the shearing blocks through a compression machine at a rate of approximately 500 pounds per square inch per minute. The test is continued until the specimen fails, and the unit shearing strength is then computed by dividing the total load by the area of the plane of failure. The unconfined compression test is conducted similarly to the com- P=total load pression test of concrete cylinders, apart load on section A-A with the load applied to the speci- f,=unit tensile strength at r men at the rate of 1,000 pounds per Fic. 1. Diagrammatic sketch of ring SQuare inch per minute. Where test. (After Philippe, 1941.) possible, the angle «, which the sur- face of the compression break makes with the horizontal, is measured, and the angle of internal friction ¢ is computed by the formula ¢ = 2(a — 45). Mr. R. R. Philippe (1941, p. 3), Director, Ohio River Division Laboratories, Corps of Engineers, U. 8. Army, has developed a ring test for measuring the unit tensile strength of the rock. The test consists of applying a line load to a ring specimen of the rock. The specimen is prepared by cutting a narrow section of the rock core with the cutting machine, and then finishing the cut surfaces on a lapping stone with a carborundum grain to make them true, after which a hole is drilled through the center. The rings are so proportioned that the thickness of the ring and the diameter of the drilled hole are from 10 to 15 percent of the diameter of the specimen. [The method of applying the load and the re- quired dimensions to be measured are shown in the sketch of Fig. 1.] The dimensions of the ring and the total load at failure allow an evaluation of the unit tensile strength which, together with the unit compressive strength of a similar material, is used to calculate the angle of internal friction and cohesion of the rock. The equations used are as follows: Cu. 8] SAFE BEARING CAPACITY 155 By Mohr’s Theory of Rupture: emake o = are i 7 (1) Be Sy atid) (2) 2cos@ where f. = unit compressive strength, f; = unit tensile strength, @ = angle of internal friction, c = cohesion. Using the results from equation (1) and (2) the shearing strength s; for any given normal load P can be computed by Coulomb’s equation: ss =c+Ptand (3) The elastic properties of the rock are determined by the standard tests of modulus of elasticity and are pertinent to the determination of the question whether or not consolidation and settlement will occur. It is common to find in sedimentary rocks greatly varying elastic properties between different rock types. At the site where a structure is to be constructed upon rocks of different elastic properties, allowance for this condition may be required in the design. This would be perti- nent at a site where steeply dipping rocks strike diagonally across the site so that sections of the individual foundations would be located on different rock types. Elastic properties of the rock are also to be considered when repeated loading and unloading of the foundation oc- curs or when marked decrease in the loading may result from the execution of the proposed plan of construction, as, for instance, in the excavation of a spillway cut in which several hundred feet of rock may be removed and a relatively large area of concrete pavement may be constructed upon the rock thus exposed. SAFE BEARING CAPAciITy Philippe (1941, p. 7) has summarized the computation of bearing capacity as follows: In computing bearing capacities from the shearing strength of a foundation material, the following formulas based on the theories of elasticity and plasticity are recommended: (1) Theory of Elasticity: p= 35 (4) where p= the bearing capacity ss= shearing strength of the material The above formula is to be used with minimum or average shear strengths 156 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 obtained from the direct shear tests. When the formula is used with values of c and @, a factor of safety F of 4 is usually applied; then Ss =c+ ptangd From Equation 4 big! Ss == 1 Then bearing capacity p becomes ™C Bo eee eee eee 5 . F —wrtang (6) (2) Theory of Plasticity: Using a factor of safety, F = 4 Jae ee SCT (6) where sp = the yield strength of the rock in tension. In the above formula, the ultimate tensile strength of the rock is used for sp. This is an allowable approximation since for brittle materials the yield strength in tension is only slightly less than the ultimate tensile strength. The application of the preceding equations to test samples of shales indicated that allowable loads could exceed 40 tons per square foot. Existing building codes would probably rate similar shales in bearing capacities at 8 to 10 tons per square foot, indicating an apparent ultimate safety factor on the order of 16 to 20. In analyzing the foundation reaction under masonry dams, Creager, Justin, and Hinds (1945, pp. 295-304) and Houk and Keener (1941, p. 1126) have placed emphasis on the shear-friction factor of safety. In these calculations, ‘shear strength of the rock at the base of the dam, based on average shearing strength of that material, is included. The determination of the coefficient of friction is not usually subject to test but is assumed to be between 0.6 and 0.8. In order to obtain the desired critical data it may be necessary to take the test to the sample and conduct a field test on the rock in place. Should the field test indicate values in the same order as the laboratory tests, it may be assumed that the laboratory tests are yielding reliable data. In actual practice, before field tests are resorted to, one of the following conditions must arise: preliminary calculations of loadings must indicate that the estimated strength of the foundation rock will be approached closely by the loadings to be produced by the proposed structure; or the strength of the foundation rock must be unknown and questioned, as in the case of certain shales upon which no comparable Cu. 8] SAFE BEARING CAPACITY 157 structures have been built. Field tests have been conducted at Possum Kingdom Dam (Niederhoff, 1940) on the Brazos River in Texas, at the location of Watts Bar Dam (Rountree, 1940, pp. 1538-1543) on the Tennessee River, and at the Lake Lynn Dam on the Cheat River in West Virginia, to mention only a few. Tests of bearing capacity, shearing and sliding strengths, and reaction to repeated loading and unloading have been performed. Blocks of concrete are poured directly upon a surface of shale pre- pared as if it were to be the foundation for the project structure. Un- der a vertical load supplied as in a common load test on a bearing pile, the block is subjected to horizontal force by jacks. Both vertical and horizontal forces may be increased, or the horizontal force alone may be increased. Suitable instrumentation permits the recording of deformation and the applied forces. The test is continued usually until failure of the foundation is achieved either under continuously in- creasing load or repeated reloadings. After the failure, horizontal loading may be applied again and an indication obtained of the co- efficient of friction. If the failure occurred in the rock, the coefficient will refer to the friction of rock on rock and will be of value in analyz- ing the rock foundation; however, if the failure occurred at the con- tact of the concrete and the rock, the subsequent loading will indicate the frictional value of concrete on rock. The data derived from such field tests are indicative of the conditions prevailing at the location of the test block and may be extrapolated to the remainder of the foundation area in relation to the similarity and homogeneity of the rock. A characteristic of shales and other argillaceous sediments that is commonly overlooked in relation to foundations is the tendency to con- solidate under a load less than required to produce failure. In field and laboratory testing of the less elastic shales the rate and amount of consolidation are measured in relation to time and to the increased loading. Because many structures impose varying loads, the rebound or return toward original volume is also determined with the release of loading. Cyclic loading and unloading in the range below failure may be applied until the shale approaches a degree of consolidation where effects of successive cycles are essentially those of preceding cycles. In summation, it is a practical impossibility to define the ultimate mass strength of a foundation, although quantitative indications may be obtained of the strength of the rocks by means of field and labora- tory tests (Burwell and Moneymaker, in press). It is the writer’s im- pression that most foundations are considerably stronger en masse than the localized tests indicate. 158 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 PERMEABILITY Water may move through the intergranular pore spaces in sedi- mentary rocks as well as through channels along bedding planes, fractures, joints, faults, and solution channels. Two general founda- tion problems arise from permeability of the rock: uplift and leakage. Uplift is the buoyant effect of water on a submerged foundation. Leakage is the passage of water through a foundation and is of primary importance in the determination of the economic feasibility of a proj- ect. Uplift has been the subject of considerable thought and discussion recently, much of which has been summarized by Harza (1947) and others with the result that it is commonly assumed in the design of a structure with a submerged foundation that the entire base of the structure is subject to uplift pressure, regardless of the type of rock upon which it is founded. There remain then to be determined the in- tensity of the uplift at any point in the foundation and a means of re- duction of that pressure. The pressures may be closely determined by the construction of a flow net. Construction of a grout curtain will reduce the leakage but not the uplift pressure. The reduction of the uplift pressure can be effectuated only by provision of relief. This is usually in the form of nearly vertical drainage wells or drain holes drilled in the foundation rock downstream from the grout curtain, as close to the zone of maximum pressure as feasible and connected to a collector system in the basal portion of the dam. The direction of these holes, their size, spacing, and depth should be dependent upon the type and structure of the foundation rock and its post-construction characteristics, because the methods of construction and type and extent of foundation treatment may modify or obliterate the necessity for draining certain elements of the foundation rock. Leakage beneath structures has been encountered in varying degrees in projects located on all types of sedimentary rocks. Leakage can be reduced to a negligible minimum by finding and cutting off stratifica- tion, fracture, and solution channels that would allow water to by-pass the structure. The prime purpose of leakage reduction is to prevent the escape of water for whose storage the structure was built and upon which storage the economics of the project depend. The safety of the structure should not be endangered by the free passage of water through its foundation. Although limestone is only slowly soluble, it, like any other rock, is subject to mechanical erosion. Theoretically, then, it might be permissible to reduce leakage in a limestone foundation to a point where velocities of flow are mechanically non-erosive, if the Cu. 8] INSOLUBLE SEDIMENTARY ROCKS 159 economics of the project were not materially affected. Such could not be permitted in a foundation containing gypsum or salt. As a matter of good practice the intent of all grouting and waterproofing treatment of foundations is to reduce leakage to the absolute minimum feasible under the circumstances. The problems involved in control- ling leakage require the determination of the location, extent, and size and interconnection of actual and potential channels of leakage, and the means of cleaning and closing off or filling these channels. The solution of these problems requires the application of many geologic techniques including detailed investigation of the sedimentary features of the rock (Moneymaker, 1941). CLASSIFICATION OF SEDIMENTARY ROCKS The sedimentary rocks may be separated into innumerable types, depending upon their chemical composition, geologic history, mode of deposition, grain size, primary structure, texture, and mineralogical composition. However, in relationship to foundations they may be grouped into two broad classes: soluble rocks and insoluble rocks. The insoluble rocks include rocks composed of essentially insoluble minerals bonded with generally insoluble cement in which naturally cavernous conditions do not occur. Examples of this type of rock are sandstone, siltstones, shales, indurated clays, mudstones, claystones, coal and other carbonaceous rocks. Soluble rocks include limestones, dolomites, and evaporites in which cavernous conditions have occurred or may occur. Soluble and insoluble rocks may be closely associated as in an interbedded sequence. The determination of appropriate classifica- tion depends on the extent of past or potential solution and its in- fluence on the type of foundation problem. The typical association of insoluble and soluble rocks is the interbedding of shales and limestones. Another condition, less recognized from the standpoint of foundations, is the cyclic deposition found in the Carboniferous rocks of the Ap- palachian and Interior coal basins of the United States where non- marine and marine sediments contain limestones as well as insoluble members. INSOLUBLE RocKs Insoluble rocks can be separated into rock types based primarily on their grain size, mineralogical composition, cement, and frequency of bedding. The foundation problems of these rocks depend almost en- tirely on the physical properties of the rock which are closely related to the foregoing geologic characteristics. Except for the siltstones of the 160 PHiILBRICK. FOUNDATION PROBLEMS [Cu. 8 Appalachian Basin the physical strength and intergranular permeabil- ity decreases, in general, with decrease in grain size and increase in clay content. CoarsE-GRAINED, INSOLUBLE SEDIMENTARY ROCKS These rocks include the sandstones, conglomerates, and sedimentary breccias. Commonly these rocks have been well cemented. The bed- ding varies from an inch or two in thickness up to over 20 feet. Cross lamination is common. The rocks may be considered competent rocks which have reacted to deformation with the development of joint systems and fractures and, in the case of faulting, have produced brec- cia and gouge zones containing granular material. Structurally the coarse-grained insoluble sediments may be characterized by three sets of planes referable to bedding, cross lamination, and deformational jointing. On these may be superimposed faulting. From the stand- point of physical characteristics, these rocks are among the hardest and most durable materials upon which a structure may be founded. With the cementation of the rock by silica the sandstones assume the physical characteristics of a quartzite, and compressive strengths as high as 34,960 pounds per square inch have been recorded (Eckel et al., 1940). On the other hand, compressive strengths as low as 120 pounds per square inch have been found in the poorly cemented sandstones. In general, compressive strengths are of the order of 10,000 pounds per square inch, and shearing strengths, in pounds per square inch, of 770 + 2.60p, where p equals intensity of load on the plane of shear, are found. With these physical characteristics, the average sandstones, sedi- mentary breccias, and conglomerates are commonly acceptable founda- tion materials when considered within the bounding planes of the test specimen. The chief difficulty with the sandstones lies in their brittle character, which has caused them to fail under deformational forces and produce not infrequent cracks and occasional zones of pervious eranular material along faulting planes and occasionally along the major system of joints. Because of the high intergranular permeability of some sandstones, uplift pressures may approach the theoretical maximum. Mahoning Dam on Mahoning Creek, Armstrong County, Pennsyl- vania, is a gravity dam about 176 feet high above the stream bed founded on a sandstone which, in laboratory tests, indicated strengths many times those required to support the load of the dam. The usual small- and large-diameter core borings indicated no structural flaws in the foundation and showed a low anticlinal nose plunging down the dip of the regional structure. However, when the foundation was ex- Cu. 8] INSOLUBLE SEDIMENTARY ROCKS 161 posed, a branching thrust fault carrying about 18 inches of gouge was encountered just below the foundation level at one side of the spillway and some 20 feet deeper on the other side. The fault re- quired the excavation of the overlying material to a depth of about 20 feet, where the gouge thinned and became sufficiently granular to permit consolidation by shallow pattern grouting. The sandstone foundations are commonly acceptable, subject to geologic structural weaknesses not reflected by laboratory tests and related to the inherent characteristics of the rock en masse. MepiuM-GraINnED, BEpDED, INSOLUBLE SEDIMENTARY Rocks These rocks include siltstones and some of the coarser shales. The general characteristics of the siltstones are similar to those of the sand- stones with the exception of permeability, which is commonly lower in the siltstones. Physical tests of siltstones in the Appalachian Basin indicate compressive strengths of 8,800 pounds per square inch and shearing strengths, in pounds per square inch of 800 + 2.4p. The re- action of siltstones to deformation is similar to that of sandstones. In general, then, siltstones are similar to sandstones as foundation mate- rials, although somewhat weaker. FINE-GRAINED, LAMINATED, INSOLUBLE SEDIMENTARY ROCKS These rocks are the common shales, they vary greatly in character, and they constitute some of the most difficult foundation rocks. The shales are distinctively thin-bedded to laminar and composed of lam- inar minerals which lie parallel to the bedding. Shales are character- ized by an established plane of potential weakness on which may be superimposed regional and local systems of joints, fractures, folds, and faults. The bedding may be disturbed by prelithification dis- tortion, as in some of the near-shore deposits, or it may remain almost planar in its smoothness. The minerals are usually micas and clay minerals with varying amounts of silica. The intergranular cement may be siliceous, cal- careous, or, less frequently, ferruginous, or intergranular cement may be lacking entirely. The first type of shale has been termed by Mead (1938) a “cemented” shale, and the second type a “compaction” shale. The physical properties of these two pure types differ considerably. The difficulty of the foundation problems seems to increase with the de- crease in cement and resulting decrease in strength and durability. The cemented shales behave more like rocks and are variably elastic, whereas the compaction shales behave more like soils in relation to stress. There is no hard and fast line separating these two types of 162 pHitBRick. FOUNDATION PROBLEMS [Cu. 8 shale, nor is one possible inasmuch as most shales are partially cemented and partially compacted rocks. The simple weathering test, which is the behavior of the rock during five successive cycles of wet- ting and drying with water or 100 N ammonium oxalate, seems to be the most revealing of the coherent quality of the shale. Those shales which are reduced by this process to uncohering aggregates of ap- proximately grain-sized particles are compaction shales. Those which are entirely unaffected or reduced only to flakes are cemented shales. This test also indicates the behavior of the shale upon exposure to atmospheric conditions during construction. The chief problem with a shale is the determination of the physical characteristics of the rock from the standpoint of design criteria. Such a shale has been described by Rountree (1940, p. 1539) as follows: The shale underlying the easterly portion of the Watts Bar Dam is a part of the Rome formation and is classified as a cemented, clay and silty shale. It is fairly compact and, in general, unweathered. The shale is characterized as “fissile” and is comprised of very thin layers which are easily separated; the surfaces of these layers are smooth and glossy in appearance. The shale is interbedded with thin layers of hard sandstone which vary from % inch to 2 inches in thickness and which are normally from 1 inch to 24 inches apart. In general the shale has a dip angle of from 20° to 30°, but occasionally, in limited areas, the bedding planes are horizonal. The apparent weakness of the material against movement parallel to the bedding prompted these tests. The field load tests were conducted on six blocks of concrete poured on the shale which had been trenched to a depth of 2 inches around the base of the blocks. The shearing strength of the shale was appraised at a value in tons per square foot intermediate between 0.9 + 0.43q and 0.8 + 0.35q, where q is the intensity of load on the plane of shear. The value of the coefficient of friction of shale sliding on shale, parallel to the bedding planes, was determined as something less than 0.53 (Rountree, 1940, pp. 1542-1543). If it is assumed, then, that in a shale foundation the physical proper- ties of the shale are known, what effect do its geological properties have upon the design of the foundation? If the materials are homogeneous and infinite in extent, the geological properties are defined by the physical properties. But in practically all cases the shale is not homo- geneous and infinite in extent. ‘Therefore one is forced to consider the inherent geological properties. The most important are: the inter- bedding of softer, thin layers; dip; fracture systems; and durability upon exposure. Usually these can be defined only by visual investiga- tion of the rock in place through preconstruction shafts and tunnels, Cu. 8] INSOLUBLE SEDIMENTARY ROCKS 163 although their presence may be indicated by surface exposures. Not infrequently shales carry variably open bedding planes along which there may be very soft veneers of clay. With these conditions it is almost impossible to define physical properties exactly, and large safety factors are considered in the calculations of stability and built into the structure. As an example, the base of the foundation may be inclined so that the angle of the resultant of all forces tends to be more normal to the foundation surface, with the result that the founda- tion rock is engaged to a greater depth, thus reducing the influence of weaker zones immediately below a level, higher foundation surface. Similar results are obtained by the construction of shear walls suitably anchored to the structure and extending below the base of the re- mainder of the foundation of the structure. But in general the tend- ency in overcoming geological weaknesses in shale foundations is to re- duce the unit loading by broadening the base of the structure or re- ducing its weight and founding it well below the surface of sound bedrock. The concrete of the foundation is placed directly against the rock enclosing the excavation. Advantage is taken, thus, of the shearing resistance of the rock mass into which the foundation is keyed. | The foundation problems of the “immature” shales of the Fort Union formation represent the other end of the transition between the foundation problems which are controlled by the geological character- istics of shales and those which are controlled by the physical char- acteristics of shales. Golder, in a personal communication dated Feb- ruary 9, 1949, has described these materials at the site of Garrison Dam, North Dakota, as follows: The Fort Union is composed of nearly flat-lying fine sands, silts, and clays (with clayey phases predominant), limestone and lignitic horizons of Paleocene age, ranging from thin partings to beds over 15 feet thick. These sediments have been described both as clays and shales, but we feel that they are some- thing in between these two, and should be called immature shale. There is little cementing material in the Fort Union and most of the standard soils laboratory testing procedures are applicable to the sediments of this formation. Shear tests of undisturbed samples show that the cohesive strengths range from practically 0 to almost 2 tons per square foot with angles of internal friction ranging from 15° to 34°. The foundation problems arising from the necessity to construct massive structures on these sedi- ments include the determination of the safe bearing values by test pro- cedures and the preservation of the rock during construction and prior to the placement of concrete. 164 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 Fine-GRrAINED, INSOLUBLE MASSIVE SEDIMENTS These rocks are fine-grained argillaceous sediments comparable in almost all characteristics to the shales with the exception that the bedding is infrequent. They seem to occur sporadically through the geologic column from the Paleozoic into the Tertiary. They are vari- ably cemented and, like the true shales, exhibit decrease in strength generally with a decrease in intergranular cement. Field tests of such materials, made in connection with Possum Kingdom Dam on the Brazos River, Texas, to obtain the limiting data upon which to base the criteria of design of a safe structure, are summarized by Niederhoff (1940). The Bear Paw shale at the site of Fort Peck Dam, Montana, is another of the massive argillaceous sediments which has been ex- tensively investigated. However, at Fort Peck the foundation problem was complicated by the presence of bentonitic beds, variable weather- ing, and considerable faulting. The problems there involved the test- ing of foundation rock in the laboratory by means of undisturbed cores, the preservation of the rapidly disintegrating rock during con- struction, and determination of permeability of bentonitic seams. As most of these massive argillaceous rocks are poorly cemented and owe their strength to compaction produced primarily by the weight of the overlying sediments, it is not unusual to find that such rocks expand with reduction in load resulting from removal of the overlying rocks. At Fort Peck the Bear Paw shale has expanded in the spillway area and heaved the spillway pavement. Foundation problems in these rocks are subject to closer determina- tion by methods of direct testing than problems in the other argil- laceous rocks as these rocks tend more toward homogeneity. Cyctic SEDIMENTS The concept of cyclic sediments has been applied only recently to foundation problems (Philbrick, 1947), although structures have been built on cyclic sediments for many years in the Carboniferous of the Appalachian and Mid-continental coal basins of the United States. Practically all the problems of both shale and sandstone foundations, together with those of coal and limestone foundations, occur in a single sequence in a cycle, and several cycles may be present at a site. For a low structure the problems are usually those of the rock type which happens to be the bearing bed. In a structure higher than the thick- ness of the strata composing the sedimentary cycle, the problems of the several rock types present in the cycle will be repeated. Difficulties encountered in the design and construction of Tygart Dam and Lake Cu. 8] CYCLIC SEDIMENTS 165 Lynn Dam in West Virginia (Crosby, 1941) were primarily the result of cyclic sedimentation. A cycle in-the Coal Measures usually, but not always, is composed of the following rocks, from the base upward: sandstone, shale, fresh- water limestone, under clay, coal, shale, marine limestone, and shale. Dissimilarity of physical properties of contiguous rocks is the dominant characteristic of the cycle from the standpoint of foundations. The sandstones may be hard massive rock. The shales and under clay may be soft to medium hard and fissile to massive. The limestones range from pure limestones to nodular limy shales or clays. The coal is brittle. To even minor geologic deformation the several rock types react differently. The sandstones, sandy shales, and carbonaceous shales behave as competent beds and fracture. The weaker members fail by plastic flow with the development of numerous, random, slicken- sided surfaces. The coal develops a blocky structure, and the lime- stones usually develop a fracture system. None of these systems neces- sarily parallels the system of a contiguous member. The rate and degree of weathering vary with each rock type. Localized zones of weathering are found at the base of the pervious members. Inherently soft materials are found at the top of the impervious members under- lying pervious members. The result is that, from the standpoint of design, the foundation is heterogeneous. Sampling of each bed or type of bed is required to produce data upon which to base the design of that portion of the structure controlled by the characteristics of that bed. The individual portions of the structure may require individual design. During construction, certain rocks must be protected from weathering and disintegration, whereas other members can be left com- pletely unprotected. The plan of excavation is dependent upon the extent of weathering in the several rock types and the quantity of material that may be removed without permitting deleterious expan- sion of the underlying compaction shale members. Bond between con- crete and foundation rock may be virtually unobtainable in some of the compaction shales and the under clay without very careful prepara- tion and cleaning. It is usually safer to place foundations only on the cemented shales and sandstone. This causes the foundation to be cut on nearly vertical slopes at monolith joints and requires the ex- ercise of great care in excavation and protection of these riser faces. The cyclic sediments require as much care in foundation investiga- tion, sampling, testing, design, treatment, and construction as any of the sediments, because almost all the problems of the other sediments are found in the cyclic sediments. Although the problem of extensive natural leakage is not commonly met, artificial leakage channels may 166 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 be present in the form of coal, clay, or limestone mines. In the case of structures requiring support only, it is not uncommon now to find the site of the structure underlain by a coal mine which may have been abandoned for many years or on which few or no data are available. Such a mine may be accessible only under dangerous and adverse con- ditions and not infrequently only through new shafts or entries. Such mines may be grouted (Philbrick, 1948) or backfilled with stable, free- draining materials as is common at present in the anthracite fields of Pennsylvania. If sufficient data are available on the extent of mining, the distribution of pillars, and the character and thickness of over- lying materials, it may be feasible to disregard the mine entirely. If the mine is on fire, it is wise to find another site rather than attempt to put out the fire and utilize that site. SOLUBLE SEDIMENTARY Rocks Much has been written on the foundation problems of soluble sedi- mentary rocks, and feasible solutions to the problems of leakage control have been developed (Lewis et al., 1941). Although tests of uncon- fined 6-inch cores of the Ocala limestone have indicated compressive strengths of as low as 64 pounds per square inch, in general the com- mon limestones and dolomites are ample in supporting capacities both in shear and in compression for a dam less than 300 feet in height. The nature of the rock and the usual rough surface developed during construction reduce to an academic consideration the problems of failure by sliding. However, limestones may carry, close beneath the base of the proposed structure, thin partings of shale which require consideration of their properties in the design of the structure or ex- cavation to below the shale. Certain limestones carry bentonite beds which require similar investigation and treatment. It is generally as- sumed that the rate of solution of limestone and dolomite is sufficiently slow to have no bearing upon the safety of the structure. Chalk, such as the Niobrara in Nebraska (Happ, 1948) poses a dif- ferent problem. Happ has described the fresh Niobrara as a fairly dark gray, fine-grained, compact, brittle, soft chalky limestone cut by thin horizontal layers of bentonitic clay ranging from a small frac- tion of an inch to 4 inches in thickness, the majority being very thin. It has an average dry weight of about 100 pounds per cubic foot and a compressive strength ranging from about 250 to 2,400 pounds per square inch and averaging about 1,000 pounds per square inch. The bentonitic clay layers in laboratory shear tests showed values in cohesion ranging from 0.1 to 0.4 ton per square foot with tan ¢@ ranging from 0.06 to 0.23. In this case the problem is not that Cu. 8] CONCLUSION 167 of providing effective cutoff of pre-existing water channels as in lime- stone terranes, but rather of designing a structure on soft rock cut by seams of softer rock, the latter being subject to disintegration upon exposure and to expansion upon release of load with the removal of overlying materials. Some structures have been built on gypsum and on salt, and others are being considered for sites containing these rapidly soluble sedi- ments. The problems concern the geologic dating of collapse of caverns leached in the gypsum, the determination of the present water table, and the effect of the increased height of the water table on the rate of solution of the soluble sediments. Where it has been proved that the gypsum is effectively cut off by impermeable beds from the additional hydrostatic head, reservoirs have been successfully constructed and operated over gypsum. For salt the situation is similar. The general rule in considering foundations containing rapidly soluble sediments is that another site is probably better. However, if no other site is available, the site should be considered as one on which a flexible structure capable of withstanding some settlement should be designed because there are as yet available insufficient data to warrant an as- sumption of the rate of solution of these sediments. The laboratory rate of solution of gypsum in a bath may be highly misleading, or it may be approximately the rate of solution in situ. CONCLUSION In many cases it may be more economical and equally desirable to design for the worst rocks in the foundation, even though this may mean an apparently more expensive structure, than to attempt to raise the values of these rocks or to remove them and found on stronger materials or choose another site where better rock may be present closer to the ground surface. It should be recognized that foundation treatment (the all-inclusive term applied to “monkeying” with a foun- dation) can be most astonishingly expensive even when compared with the cost of a major project. On the other hand, it may be considerably cheaper to invest a large sum in improving foundation conditions at a site than to reduce the magnitude of the structure and build other smaller supplementary structures, particularly in the case of a coor- dinated river development scheme or in the case of a high bridge. The economic treatment of foundation problems is intimately related to the general engineering and economic features of the entire project, and they may dictate the general permissible scope of the foundation treatment. 168 PHILBRICK. FOUNDATION PROBLEMS [Cu. 8 REFERENCES * Burwell, E. B., Jr., and Moneymaker, B. C. (in press). Geology in dam con- struction: Geol. Soc. Amer., Berkey Volume. * Creager, W. P., Justin, J. D., and Hinds, J. (1945). Engineering for Dams: John Wiley & Sons, New York, pp. 295-304. * Crosby, I. B. (1941). Geological problems of dams: Trans. Amer. Soc. Cw. Engrs., vol. 106, pp. 1181-1182. * Hekel, E. C., et al. (1940). Engineering geology of the Tennessee River System: Tennessee Valley Authority, Tech. Mono. 47, p. 259. Happ, 8. C. (1948). Geology of the Harlan County Dam, Republican River, Nebraska: Abstract, Bull. Geol. Soc. Amer., vol. 59, p. 1328. * Harza, L. F. (1947). The significance of pore pressure in hydraulic structures: Proc. Amer. Soc. Civ. Engrs., vol. 73, pp. 1507-1528. (This paper is discussed in succeeding volumes of the Proceedings.) Houk, Ivan E., and Keener, K. B. (1941). Basic design assumptions. Trans. Amer. Soc. Civ. Engrs., vol. 106, pp. 1116-1126. Lewis, J. S., Jr., Ross, Robert M., Gongwer, Verne, Fox, Portland P., and Hays, James B. (1941). “Foundation experiences, Tennessee Valley Authority, a sym- posium” and discussion: Trans. Amer. Soc. Civ. Engrs., vol. 106, pp. 685-848. Mead, Warren J. (1938). Engineering geology of dam sites: Transactions, Second Congress on Large Dams, vol. 4, p. 183. Moneymaker, B. C. (1941). Discussion of foundation experiences, Tennessee Valley Authority, a symposium: Trans. Amer. Soc. Civ. Engrs., vol. 106, p. 805. Niederhoff, August E. (1940). Field tests of a shale foundation: Trans. Amer. Soc. Civ. Engrs., vol. 105, pp. 1519-1534. Philbrick, S. S. (1947). Relationship of cyclothems to dam design: Abstract, Bull. Geol. Soc. Amer., vol. 58, pp. 1217-1218. (1948). Investigation and proposed treatment of caved mine beneath 20-story V.A. hospital: Abstract, Bull. Geol. Soc. Amer., vol. 59, p. 1344. Philippe, R. R. (1941). Physical properties of rock, Berlin Dam site: Unpub- lished report, Corps of Engineers, Pittsburgh, Pa., July 25, 1941. Rountree, Jack R. (1940). Discussion of field tests of a shale foundation: Trans. Amer. Soc. Civ. Engrs., vol. 105, pp. 1538-1543. * Articles of comprehensive scope or containing extensive bibliographies. CHAPTER 9 FOUNDATIONS FOR HIGHWAY BRIDGES AND SEPARA- TION STRUCTURES ON UNCONSOLIDATED SEDIMENT C. H. Harnep Geologist, Bridge Department California State Division of Highways Sacramento, California Bridges, in common with all other civil engineering works, must be dependent for support upon earth materials. Since bedrock support is the exception rather than the rule, it is appropriate that space be al- located in this volume to a brief discussion of foundation study tools and techniques applicable to the solution of problems involved in the frequent practice of transmitting structure loads to unconsolidated sedi- ment. The engineer is often confronted with serious problems posed by the design and construction requirements of safe economical support for bridges. This is particularly true when a dependence for support must be placed upon recently deposited sediment which has not been sub- jected to loads greater than those which exist at present. The purpose of this chapter is to present a description of some of the tools, methods, and personnel requirements of a modern founda- tion investigation. It is hoped that by so doing additional interest may be encouraged among workers engaged in this and related fields of ap- plied sedimentation, to the end that reflections may be observed in both the curricula of personnel training and ultimate practical achieve- ment in this important field of application. Normally the bridge engineer is allowed little choice regarding the location of his structure because such determinations are usually de- pendent upon the requirements imposed by highway grades and align- ments. This situation presents the problem of economic conformity to any type or variety of material types that occur at a specific site regardless of either the physical characteristics of the material or the magnitude of the problems involved. It follows that, if economically sound, practical substructure designs are to be consistently realized, they must be based upon accurate observations and interpretations of 169 170 HARNED. HIGHWAY BRIDGE FOUNDATIONS i@r.)9 the natural foundation conditions at each structure site. Obviously this goal can be accomplished only through the development and use of trained personnel and special-purpose tools. The list of references at the end of the chapter has been carefully selected for the benefit of those who may desire reference to technical discussions of the various aspects of foundation investigation and in- terpretation practices. PERSONNEL REQUIREMENTS Men charged with the conduct and interpretation of foundation in- vestigations for highway structures hold responsible positions. It is extremely important that an accurate picture of existing conditions be obtained for design purposes, and it is equally important that con- struction operations actually encounter predicted conditions in order that costly delays in construction, redesign requirements, contract change orders, legal suits, and, in extreme cases, structural failures may be avoided. The choice of the foundation type alone may seriously influence the economics of a project. For example, a careful estimate revealed that, for some seventeen of the many separation structures to be constructed on one of the freeways through the City of Los Angeles, the State of California would realize a saving in excess of one million dollars in the event that the sediment was capable of supporting the loads on footing foundations and pile supports could be eliminated. Determinations of this sort require precise measurements and careful consideration by well-trained and experienced personnel if the bridge engineer is to have sufficient confidence in the recommendations he receives to design and construct accordingly. It is anticipated that many universities and colleges will follow the recent example of the few that have inaugurated curricula in engineer- ing geology for the purpose of training interested students for this im- portant field of geologic application. It is hoped that the term en- gineering geology will be interpreted in its proper light and that steps will be taken to insure that students who enter this field through in- terest and choice will be encouraged to obtain an educational back- eround in civil engineering as well as geology. The “make it stout” complex so prevalent among substructure de- signers today is the natural outgrowth of works without faith. We can have little faith without understanding and little understanding without knowledge born of interest. Consequently foundation studies should be conducted by men whose prime professional purpose in life Cu. 9] FOUNDATION STUDY TOOLS 171 lies in the scientific study and interpretation of nature as applied to civil engineering. It is not important whether these men call them- selves engineers or geologists. The real requirement is that they be both. FOUNDATION STUDY TOOLS The foundation study for any highway structure should consist fundamentally in determining the proper methods of applying loads to earth materials in amounts that will not result in excessive settle- ment or failure. If the structure is to span a stream, additional con- sideration must be given to insure against structural failure due to scour in the channel or erosion of the banks. It is of prime importance that field parties charged with the deter- mination of foundation requirements be equipped with the modern tools of their trade. Hole-boring tools capable of penetrating and sampling all types of earth material to any required depth are an essential part of foundation study equipment. In most cases it is desirable that boring tools be power-operated and at the same time sufficiently light and portable as to insure access to sites located in areas of troublesome relief. The boring equipment should be capable of drilling rotary wash boreholes up to 3 inches in diameter and to depths of about 200 feet. Contrary to the general opinion among engineers, wash boring is by no means a worthless technique in so far as its applications to foundation studies are concerned. It is the writer’s experienced opinion that borings of this type furnish valuable information at the lowest possible cost per lineal foot when properly conducted under the close observation of adequately trained personnel. This technique, when used in conjunc- tion with spot sampling of the significant horizons, may in some cases fulfill all the requirements of an adequate study. This is particularly true in areas where the geology is well understood or where an ex- perience has been gained by observing the performance of existing structures within a geologic provenance. The drilling machine should in addition be equipped with bits, core barrels, and samplers capable of the efficient penetration and sampling of earth materials regardless of type or depth. A small drop hammer or other suitable device is required for the pur- pose of driving casing, penetrometers, and sediment samplers. Al- though light drop hammers are in general use for this purpose, it has been demonstrated that small-diameter undisturbed samples show less disturbance when samplers are forced into the sediment by means of a steadily applied hydraulic or screw-activated push or by a rapid, ex- 172 HARNED. HIGHWAY BRIDGE FOUNDATIONS Ore ea) plosive percussion blow (Creager, Justin, and Hinds, 1945, p.19). The resistance the sediment offers to tool penetration and extraction is com- monly an important measurement of sediment character. This is particularly true when procedures are standardized and measurements are recorded by properly calibrated gages as attested to by results of the standard penetration test (Terzaghi and Peck, 1948, p. 265). Undisturbed sediment or soil samplers capable of delivering samples in a reasonably undisturbed state are essential if some of the more significant soil mechanics tests are to be conducted. Of the many types of samplers on the market, the thin-walled types, in general, meet most of the sampling requirements. Certain situations, however, such as the procurement of samples of very soft mud, clean uniform sat- urated sand, or the sampling of sediment at the proposed tip elevation for bearing piles, may require the use of special-purpose sampling spoons. It should be kept in mind at all times that no amount of un- disturbed sampling or testing is of any great significance if the sampling procedure does not, or cannot on an economic basis, result in the pro- curement of representative samples of the sediment in a sufficiently un- disturbed condition to permit accurate evaluation of the control criteria by means of standard testing procedures. It is the engineering geolo- gist’s Job to determine whether or not the samples are representative of the true conditions. If the site is one of such complexity that samples are not diagnostic, the recognition of this fact is extremely important, and the problem may then be handled by means of an ad- justment in the safety factor or by incorporating extra flexibility in the contract specifications. The large array of boring-tool types, samplers, penetrometers, and other special tools designed for the foundation study purpose furnishes ample evidence of the fact that there is no universal tool panacea for work of this nature. Among the control factors regarding the choice and extent of tools are the amount of foundation investigation work _ to be done by any specific agency, the degree of geologic complexity or variability of the general area and specific site of operation, and the size of the project. Many thousands of dollars have been wasted for the want of proper equipment to conduct efficient foundation studies. Numerous attempts have been made to solve the problems presented by complex or erratic conditions of sedimentation by liter- ally drilling the site full of holes, taking continuous undisturbed samples, and conducting innumerable shear, unconfined compression, consistency limits, moisture, density, permeability, consolidation, ete., tests, when possibly all that was required to be determined was how Cu. 9] FOUNDATION STUDY TOOLS 173 far steel piles would penetrate, or perhaps the feasibility of supporting the load by means of friction piles. These answers might well have been obtained by driving penetrometers at a few selected locations throughout the structure site; the resistance offered to penetration could have been observed, and, after appropriately spaced time inter- vals, periodic pull tests designed to measure skin friction values could have been conducted. Test pits in large numbers and to great depths have been dug after the inability to penetrate a deposit of loose boulders and gravel to bedrock depth by means of available drilling equipment has been established. This discovery is usually made after at least one of every type of drill bit available has been destroyed in the attempt. Refraction seismic equipment or certain electrical geophysical apparatus could have revealed the answer for a mere fraction of the expended time, energy, and money. Although current researches have not resulted in the ultimate with regard to applying geophysical techniques to mantle exploration, much progress has been made in recent years, and the future looks very hopeful (conference with Dr. Thomas Poulter, January 1949). In addition to equipment capable of solving the drilling, penetrating, and sampling requirements of a foundation study program, it is equally important that the field party conducting such studies have available sufficient testing apparatus to permit on-the-job measurements of im- portant physical characteristics of the sediment. Pumping tests should be conducted in drill holes or wells in cases where footing foundations are to be situated below the water table or where high permeability is likely to exert a major influence on the builders’ operation. Quick shear or unconfined compression tests or both, when conducted in the field, often give results that may influence or control the conduct of the entire study program. Percent voids, unit weight, moisture content, the consistency limits, and grain-size classification can, and should, be determined in the field. This pro- cedure not only develops better-trained personnel but also places sampling and testing directly on a job-requirement basis. It is not suggested that all laboratory-type testing be done in the field. Tri- axial shear, time consolidation, petrographic analysis, and most chemi- cal tests require permanent laboratory facilities and can be properly conducted in no other place. Since the advantages offered in the form of checks on both dynamic measurements and the judgment of the engineering geologist far out- weigh those of permanent laboratory facilities for certain measure- ments, field testing is strongly recommended whenever feasible. 174 HaRNED. HIGHWAY BRIDGE FOUNDATIONS LC 9 EXPLORATION PROCEDURE The first step in the conduct of a foundation study for a highway structure should consist of a complete determination and understand- ing of the general geology of the area. The foundation condition at each structure site, whether complex or simple, is the direct result of the geologic processes, environment, and history of the area. If an understanding of these factors is not acquired first, the study becomes a routine matter of boring holes in the ground and conducting monoto- nous tests rather than an interesting interpretation of the complexities of nature. The second step consists in ascertaining the physical char- acter of the sediment for the purpose of determining specific founda- tion requirements at the sites of structure support. The depth to which borings or soundings should extend, and the spacing between borings, should be dependent entirely upon the com- plexity of the site, the size of the project, and the depth to compres- sible sediment occurring within the limit of surface-load influence. Any attempt to standardize boring depths or distances between borings is obviously dangerous and to be avoided. Each structure site re- quires its own individual explorational treatment, and any attempt to predetermine the study requirements or procedures usually leads to in- efficient operations and faulty results. There is no substitute for ade- quate training and experience when the proper study procedure for a foundation problem, or the accurate determination of the time when further study will cost more than it is worth, is being selected. FOOTING FOUNDATIONS In those cases in which scour or lateral erosion is not indicated and in which the boring logs and sounding patterns or both do not reveal the existence of soft compressible sediment within the significant depth of surface-load influence, the structure may be supported by footing foundations resting upon near-surface sediment (Taylor, 1948, p. 560; Terzaghi and Peck, 1948, p. 56). All footings must be located below the influence of frost and seasonal moisture. No precise statement can be made with regard to the depth to which footing foundations may extend and render structural support in economic preference to pile foundations, because this determination must be made for each substructure in the light of allowable bearing value of the sediment, depth of seasonal volume change due to moisture, frost, plants, ete., depth of scour, span lengths; live, wind, earthquake, Cu. 9] FOOTING FOUNDATIONS 175 and dead loads; and allowable settlement. This is, however, an im- portant phase of the economic study of the substructure and one that requires careful consideration. Every foundation should be designed for a factor of safety com- parable to that used in the design of superstructures. This factor should be applied against the possibility of the foundation failing as a result of applied loads exceeding the shear strength of the supporting sediment, sliding upon the contact between the footing block and the supporting earth material, sliding upon bedding planes of the sediment, or excessive settlement due to compressive materials. This factor of safety should not be less than 2%, and it rarely needs to exceed 3. It is not uncommon in present practice for factors of substructure safety to be as high as 20. Such a design policy rarely results in a founda- tion failure in the normal sense, but it does cost untold thousands and thereby represents economic waste. The total settlement of the structure should not exceed that which may result in differential settlements in amounts capable of either damaging the structure or developing a poor riding deck. Differential settlement between adjacent supports will usually not exceed 50 per- cent of the total settlement unless the loads are vastly different or the foundations are of different types; for instance, adjacent bridge piers supported by pile and footing foundations respectively. In such cases the differential settlement may approach 100 percent of the total settle- ment. Precise settlement computations are rarely justifiable for highway structures except when firm support cannot be reached by piles, col- umns, or caissons because of great depth, or when soft compressible sediment occurs beneath the footing at a depth sufficiently shallow to permit consolidation under the applied load (Terzaghi and Peck, 1948, pp. 413-456; Plummer and Dore, 1940, p. 192). Considerable money could be saved by a more extensive use of the flexible-hinge design for structures where serious differential settle- ments are anticipated. Hinge design, together with facilities for jacking, would, for example, make it possible on many occasions to use floating abutments on fills through which piles are commonly driven. Load tests on bearing plates, when accompanied by data that iden- tify the character of the sediment within the significant depth, fur- nish valuable information regarding the load-carrying ability of the supporting material. Such tests, although generally considered ex- pensive, can, with properly designed equipment, be conducted on a routine basis at low cost. Load tests at the sites of pier locations for 176 HARNED. HIGHWAY BRIDGE FOUNDATIONS [Cu. 9 the Plaza Garage Overhead structure on U. S. 99 north of Visalia, California, proved conclusively that footing foundations will safely support loads of 3 tons per square foot at a depth of 7 feet. The cost of the tests was only a small fraction of the saving realized as a re- sult of pile elimination. The load tests in this case were conducted in large-diameter holes drilled with a power-driven bucket auger on plates having a surface area of 2 square feet. The load was applied by means of a power-operated hydraulic pressure cell. Constant pressure was maintained by use of a nitrogen-loaded gas accumulator and a mer- cury pressure switch. Twelve-inch expanding anchors fixed in drill holes furnished adequate load reaction. Load tests on bearing plates must be used with a full appreciation of the limitations of the method. It should be kept in mind that the test reflects only the character of the sediment to a depth of about twice the diameter of the bearing plate, whereas the footing foundation will exert an influence on the underlying sediment to a depth at least as great as the footing width for ideal conditions, and to a depth several times the footing width if the sediment becomes softer and weaker at depth. The depth below the footing elevation to the water table is of con- siderable importance, particularly in the case of footing foundations supported upon cohesionless sediment. If the water table rises within the significant depth, the factor of safety should be increased from 2.5 to 4. Large footings settle more than small ones if unit loads are the same. For example, allowable loads should be reduced from 6 tons to approximately 5 tons per square foot on dense sand where the footing width is increased from 5 to 10 feet. It is not advisable to attempt to support footing foundations for highway structures on soft clay or plastic silt except as a last resort. Differential settlements up to 6 inches are commonly unavoidable in such cases, and total settle- ments amounting to nearly a foot have been measured. It is not advisable to apply loads to footing foundations supported by plastic sediment in amounts that exceed the unconfined compressive strength of the weakest material encountered. If weak plastic sedi- ment occurs at some depth below a firm stratum upon which footings are to rest, a calculation is required for the purpose of determining the load distribution at depth and, thereby, the amount of load that will be applied to the weak sediment. A safety factor of 3 should then be used against overstressing the weak stratum. Cu. 9] PILE FOUNDATIONS 177 PILE FOUNDATIONS Pile support for highway structures should be used only in those cases in which footing foundations cost more or are unsafe because of the possibility of erosion, or where settlements are apt to be ex- cessive. The engineer has, through experience, acquired a deep respect for the complexity and variable nature of foundation conditions. As a result, pile foundations occupy a hallowed niche in the engineering profession and are apt to be used, not necessarily when needed, but whenever possible. The limits of possibility, when confronted with engineering ingenuity, have, needless to say, been pushed back quite a way. Piles may be classified on the basis of load-transfer method into two general types: (1) Friction piles. Those which transfer the imposed structure load to the surrounding sediment largely through skin friction. (2) Bearing piles. Those which transfer loads to a firm or hard material at depth largely through the pile point contact. It should be noted that friction piles support a portion of the load in the manner of bearing piles and that bearing piles transfer a por- tion of their load to the sediment through skin friction. Therefore, in order to estimate the load-carrying capacity of a single pile, both fric- tion and point-bearing data must be obtained in most cases. Carefully conducted wash borings, together with a few spot samples, usually suffice for sediment classification purposes. Drive or push soundings when accompanied by penetration tests will yield valuable data regarding pile length requirements and point bearing values. Pull tests spaced at proper intervals of time and conducted at significant depths, and locations throughout the structure site as previously deter- mined by borings, will permit reasonably accurate estimates of skin- friction values. Friction values determined by the pull-test method in plastic sediment may be checked by means of direct shear or un- confined compression tests, and those in granular sediment by the standard penetration or triaxial shear tests or both. A few test piles and full-scale load tests should be insisted upon in all cases in which structural support is to be derived from friction piles. The use of pile-driving formulas for measuring the ultimate bearing capacity for individual piles should be avoided in all cases in which resistance to pile penetration is due to friction between the pile and 178 HARNED. HIGHWAY BRIDGE FOUNDATIONS [Cu. 9 cohesive silts or argillaceous sediment. Allowable loads in such cases should be determined by use of a safety factor of not less than 2 when used in conjunction with load tests approved by the American As- sociation of State Highway Officials, Standard Specifications for Bridges (1949). Settlement determinations must be made in all cases in which soft plastic sediment occurs within the significant depth below the pile tips, or in which the load-carrying capacity of the pile group or cluster is less than the bearing value of the cumulative capacities of the piles within a cluster (Seiler and Keeney, 1944; Moore, 1947, pp. 1341- 1358). CONCLUSIONS There is a growing insistence among bridge engineers that precise foundation data are a prerequisite to efficient substructure design. This insistence is commendable and completely justifiable within the limits of our present understanding of nature, and it is plainly up to us as engineers and earth scientists to accept the responsibility for our present state of knowledge, foster the extension of our knowledge and understanding at every opportunity, and encourage the prompt ap- plication of practical truths to the design and construction of current projects. The foundation specialist must have a keen sense of the economics involved in his work and must not become so carried away with his study program that he loses sight of his justification in a conscientious attempt to know all the answers prior to construction. A foundation study is not over when the borings are completed and the testing done; it should be continued through observation and study during the period of construction and ultimate use. It is only by so doing that we learn of our mistakes and are able to accumulate a stockpile of corrected data for our future use or the use of those who follow. REFERENCES Allin, R. V. (1935). The resistance of piles to penetration: Spon, London, 130 pages. (1941). Pile driving formulas: Proc. Amer. Soc. Civ. Engrs., pp. 853-866, May 1941. (1946). Pile foundations and pile structures: Amer. Soc. Civ. Engrs. Manual of Engineering Practice, No. 27. American Association of State Highway Officials (1949). Standard specifications for bridges. Anderson, Paul (1948). Substructure analysis and design: Irwin-Farnham Puh. Co., Chicago, 305 pages. Cu. 9] REFERENCES 179 Baver, L. D. (1948). Soil physics: John Wiley & Sons, New York, 2nd ed., 398 pages. Bibliography North American Geology (1785-1945). U. 8. Geol. Survey, Bulls. 746, 747, 823, 937, 938, 949, 952. Boston Society of Civil Engineers (1940). Contributions to Soil Mechanics, 1925- 1940. Casagrande, A. (1932). The structure of clay and its importance in foundation engineering: Jour. Boston Soc. Civ. Engrs., pp. 168-221, April 1932. (1947). Classification and identification of soils: Proc. Amer. Soc. Civ. Engrs., pp. 783-810, June 1947. Chellis, R. D. (1944). Pile-driving handbook: Pitman Publishing Corp., New York, 276 pages. Converse, F. J. (1935). Settlement of footings in alluvial soils: Eng. News- Record, Nov. 28, 1935. Creager, W. P., Justin, J. D., and Hinds, J. (1945). Engineering for dams: John Wiley & Sons, New York, vol. 1, 245 pages. Cummings, A. E. (1935). Distribution of stresses under a foundation: Proc. Amer. Soc. Civ. Engrs., pp. 823-834, Aug. 1935. (1940). Dynamic pile driving formulas: Jour. Boston Soc. Civ. Engrs., pp. 6-27, Jan. 1940. (1940). Pile foundations, Proceedings of Purdue Conference on Soil Mechanics, Lafayette, Ind., pp. 320-3838, Sept. 1940. Forrester, J. D. (1947). Principles of field and mining geology: John Wiley & Sons, 647 pages. Gilboy, G. (1936). Improved soil testing methods: Eng. News-Record, May 21, 1936. Hansen, V., and Kneas, F. N. (1942). Statice load tests for bearing piles: Civ. Eng., vol. 12, pp. 545-547. Housel, W. S. (1929). A practical method for selection of foundations based on fundamental research in soil mechanics: Univ. Mich. Eng. Res. Bull., 13, Oct. 1929. (1933). Bearing power of clay is determinable: Eng. News-Record, Feb. 23, 1933. Jacoby, H. S., and Davis, R. P. (1941). Foundations of bridges and buildings: McGraw-Hill Book Co., New York, 8rd ed., 535 pages. Johnson, H. L. (1940). Improved sampler and sampling technique for cohesion- less materials: Cov. Eng., vol. 10, pp. 346-3848. Krynine, D. P. (1937). Pressures beneath a spread foundation: Proc. Amer. Soc. Civ. Engrs., pp. 669-691, April 1937, Yale Univ. Bull., 19. Lahee, F. H. (1941). Freld geology: McGraw-Hill Book Co., 4th ed., 853 pages. Legget, R. F. (1939). Geology and engineering: McGraw-Hill Book Co., New York, 650 pages. Lutz, H. J., and Chandler, R. F., Jr. (1946). Forest soils: John Wiley & Sons, New York, 525 pages. Meinzer, O. E. (editor) (1942). Physics of the Earth. Part IX, Hydrology: Mc- egraw-Hill Book Co., New York, 711 pages. Miller, R. M. (1938). Soil reactions in relation to foundations on piles: Trans. Amer. Soc. Civ. Engrs., vol. 108, pp. 1193-1236. Mohr, H. A. (1943). Exploration of soil conditions and sampling operations: Harvard Univ. Grad. School of Eng., Soil Mechanics Series, No. 21, 3rd rev. ed., Nov. 1943. 180 HARNED. HIGHWAY BRIDGE FOUNDATIONS [Cu. 9 Moore, Wm. W. (1947). Experiences with predetermining pile lengths: Proc. Amer. Soc. Civ. Engrs., p. 1859, Nov. 1947. Newmark, N. M. (1935). Simplified computation of vertical pressures in elastic foundations: Univ. Ill. Circ. 24. Plummer, F. L., and Dore, S. M. (1940). Soil mechanics and foundations: Pit- man Publishing Corp., New York, 473 pages. Ries, H., and Watson, T. L. (1936). Engineering geology: John Wiley & Sons, New York, 5th ed., 750 pages. Rutledge, P. C. (1944). Relation of undisturbed sampling to laboratory testing: Trans. Amer. Soc. Civ. Engrs., vol. 109, p. 1155. Seiler, J. F., and Keeney, W. D. (1944). The efficiency of piles in groups: Wood Preserving News, Nov. 1944, AASHO Specifications. Shrock, R. R. (1948). Sequence in layered rocks: McGraw-Hill Book Co., New York, 507 pages. Simpson, W. E. (1934). Foundation experiences with clay in Texas: Cw. Eng., vol. 4, pp. 581-584. Straub, Lorenz G. (1942). Mechanics of rivers, in Physics of Earth, Part 9, Hydrology: O. E. Meinzer (editor), McGraw-Hill Book Co., New York, pp. 614-636. Swiger, W. F. (1941). Foundation tests for Los Angeles Steam Plant: Civ. Eng., vol. 11, pp. 711-714. Taylor, D. W. (1948). Fundamentals of sow mechanics: John Wiley & Sons, New York, 700 pages. Taylor, K. V., et al. (1935). The predetermination of piling requirements for bridge foundations: Ohio State Eng. Exp. Sta., Bull. 90, July 1935. Terzaghi, K. (1929). The science of foundations: Trans. Amer. Soc. Civ. Engrs., pp. 270-405. (1935). The actual factor of safety of foundations: Structural Eng., vol. 13, pp. 126-160. (1936). Failure of bridge piers due to scour: Proceedings of International Conference on Soil Mechanics, Cambridge, Mass., vol. II, p. 264. (Discussion by Irving B. Crosby, vol. III, p. 238.) (1943). Theoretical soil mechanics: John Wiley & Sons, New York, 510 pages. (1947). Shear characteristics of quicksand and soft clay: Proceedings of Seventh Texas Conference on Soil Mechanics and Foundation Engineering, Jan. 1947. , and Peck, R. B. (1948). Soil mechanics in engineering practice: John Wiley & Sons, New York, 566 pages. Wenzel, L. K. (1942). Methods for determining permeability of water-bearing materials: U. S. Geol. Survey, Water Supply Paper 887, pp. 74-191. Wilson, G. (1941). The calculation of the bearing capacity of footings on clay: Jour. Inst. Civ. Eng., pp. 87-96, Nov. 1941. CHAPTER 10 EARTH DAMS Tuomas A. MIDDLEBROOKS Chief, Soil Mechanics Branch Office of the Chief of Engineers Department of the Army Washington, D. C. The major civil works activity of the Corps of Engineers is the con- struction and maintenance of dams and levees for flood control and navigation purposes. More than one half of the dams and all the levees are earth structures founded on a variety of geologic formations varying from sound rock to overburden soils. An application of soil mechanics principles to the design of these earth structures has re- sulted in safe and economical construction. However, soil mechanics is by no means an exact science and is dependent in a large measure on a proper understanding of geologic formations to interpret the results of foundation explorations. In the writer’s opinion, close co- operation between soils engineers and engineering geologists is essen- tial in the design and construction of earth dams and levees. Investi- gational, design, and construction features of earth dams are covered in this chapter with specific reference to soil mechanics and geological aspects of all phases of the work. Details of exploration and design have been purposely omitted because they are adequately covered in other publications. GENERAL FEATURES The selection of a dam site involves many factors, among which the geologic and soil mechanics features are of prime importance. Usually a site study must be confined to a limited reach of river channel in order that the dam may perform its required function. Unless conditions are obviously favorable at a particular location, regional geology should be studied. During the preliminary studies the site having the most favorable topographic features for the lo- cation of the dam and the spillway is selected for more detailed 181 182 MippLEBROOKS. HARTH DAMS [Cx. 10 study. The detailed site study should develop the nature of the overburden and rock foundation in the river valley and abutments. Special attention should be given to foundation conditions at all pos- sible spillway locations in order that the most economical spillway that will perform the specified functions may be designed. Founda- tions for outlet works should be thoroughly explored to determine whether a “cut and cover” type of conduit or a tunnel is preferable. Selection of the type of dam, whether concrete, rock-fill, or earth, has been adequately covered in other publications (Creager, Justin, and Hinds, 1945; Middlebrooks and Bertram, 1948). FOUNDATION STUDIES A study of foundation features of an earth-dam site may be divided into three general classifications: (1) investigation of foundations for the spillway structure; (2) investigations of foundation for the earth dam; and (3) investigations of foundations for the outlet structure. After the general geological and soil study of the site, a detailed study should be made for each of the three structures enumerated. SPILLWAY STRUCTURE In order that sufficient geological information may be furnished the designers, the geologist and the soils engineer should have a gen- eral knowledge of different types of spillways and the foundation re- quirements for each. In Figs. 1 through 6 are shown a few examples of various types of spillways and typical foundations upon which they are used. The spillway type should be varied to fit not only the foundation but also the frequency of use of the spillway. At most earth-dam sites the spillway design presents a major problem and represents a high percentage of the total cost of the project. Therefore a careful foundation study of all possible spillway loca- tions should be made. A gravity overflow spillway as shown by Figs. 1 and 2 presents fewer operational problems because it discharges into the river chan- nel. However, it is generally more expensive than any of the other types, except possibly some chute spillways on earth. The flip-up bucket is used in this country principally on massive hard-rock foundations. In Europe, this type of energy dissipater is extensively used on practically all rock foundations and on some low-head dams on soil foundations. The practice in this country is more conserva- tive than the European practice relative to the design of energy dissi- Cu. 10] SPILLWAY STRUCTURE 183 paters; therefore stilling basins are used extensively. Foundation re- quirements are similar to those for any gravity dam. Side-channel spillways similar to Fig. 3 require extensive founda- ee pool a Max.F.C.pool 335 — — Max. power pool EI.3307=== pi necwet pool ELSES, A D 1 | | | Grout Foundation 1 | curtain drain Do | ence drains Fia. 1. Gravity spillway, Clark Hill Dam. Foundation is massive granite gneiss. Frequent operation. ne oe ication drain Slab anchored to curtain rock by tie rods Fic. 2. Gravity spillway, Whitney Dam. Foundation is argillaceous limestone. Frequent operation. tion investigations because their satisfactory performance depends to such a great extent on the stability of the rock. Deep cuts into the abutment are conducive to minor slides, and the steep side walls are subject to creep and spalling during construction. These minor geo- logic factors are difficult to evaluate from explorations, and as a re- 184 MIDDLEBROOKS. HARTH DAMS [Cu. 10 sult costly changes during construction are not unusual for this type of structure. Chute spillways on earth as illustrated by Fig. 4 have been em- ployed in a number of cases where rock was inaccessible at reasonable EI.550 Ss K > US 4" black pipe drain 15'c.c. zZ KS v 1 4) & 1"¢ anchor bars ” e 9'c.c. Coe so\\ E1525 ES =n SS al IS Ay 2 SS ES 30° EN NUNU SUSU NUSUSUNYNUSUS Fic. 3. Side-channel spillway, Surry Mt. Dam. Foundation is massive granite eneiss. Infrequent operation. Steel sheet pile cutoff EI.217 ye Longitudinal drain Transverse drain Low water pile cutoff E1.150 Fic. 4. Chute spillway, Sardis Dam. Foundation is soil and soft rock. Frequent operation. depths. Foundation investigations for this type of spillway should be more extensive than for any other type. A cutoff along the center line of the weir section and an elaborate drainage system under the chute and stilling-basin portions are required. The shearing strength, compressibility, and expansion characteristics of the soil are im- portant factors in design. Upheaval of concrete spillway slabs lo- Cu. 10] SPILLWAY STRUCTURE 185 cated on expansive formations such as the Bear Paw shale encountered at Fort Peck Dam presents a problem for further soil mechanics re- search. At Fort Peck the expansion of the shale foundation under the spillway chute and approach slabs due to release of load in the spillway cut has been extensive. Maximum upheaval in the center of the chute has been over 12 inches. Upheaval of the gate structure, which was constructed on caissons to a depth of 40 feet, has been nil. |}«——— 2,100’ to river at El,315 + —————> 1 350 eeoeaeemenme |smEmaEnENIRI OC Os | Backfill Fic. 5. Saddle spillway, Wappapello Dam. Foundation is hard, blocky limestone. Infrequent operation. To river 16,000'+ at El.390£————>- 200'+ 1,000'+ | 0.5% 50> Ob} " jal TIN y-2'_| IYYyyyy1y rm = 4 Gi Bin 5 ONG USSU) 7 [A GRORAWAL 7 2' 7 10' Fic. 6. Saddle spillway, Blakely Mt. Dam. Foundation is firm shale, highly folded. Infrequent operation. Chute spillways are often founded on weak rock. Drainage under the chute section and stilling basin similar to that supplied for soil is generally required. Usually it is possible to anchor the slabs into the rock, thereby decreasing the chute and stilling-basin slab thick- nesses over those required for soil. The upstream cutoff is usually in the form of a concrete key. Saddle spillways without concrete pavement in the chute section are employed extensively where flow occurs infrequently. This type of spillway varies in design with foundation conditions and elevation of foundation rock in the saddle. At the Wappapello Dam (Fig. 5) a small ogee concrete section with a flip-up bucket was employed. At Blakely Mountain Dam (Fig. 6) a 50-foot paved section with a cutoff upstream and downstream forms the control. There are numerous cases where the rock itself is adequate for the control structure and no concrete weir or paving is required. There are also a few cases 186 MIDDLEBROOKS. HARTH DAMS [Cu. 10 where a flat natural divide consisting of soil is used as a spillway. The Mosquito Creek Dam spillway is an example of this type. In this case no formal spillway structure was built. HartH Dams The two major foundation problems in earth dams are stability and permeability. Settlement of the dam proper is of concern only to the extent that adequate freeboard is maintained. A few consolidation tests, properly correlated with the observed settlement during the early stages of construction, will usually allow the establishment of a satis- factory gross grade. General geology of the site will prove most help- ful in determining the most critical seepage and stability areas. Stability of foundation soils is, with few exceptions, the determin- ing factor in the design of the outer slopes of an earth dam. Explora- tions should be planned first to locate and outline the critical regions in the foundations. This can be accomplished adequately by disturbed- drive-sample methods by which satisfactory samples for classifica- tion and moisture determination can be obtained. After the critical areas have been outlined, large borings, 5 inches or more in diameter, should be made to obtain undisturbed samples of the silts and clays for shear and consolidation tests. Consolidated, drained, direct-shear tests are recommended for general use in determining the shearing strength of the materials under the induced loads. Triaxial compres- sion tests are recommended for well-graded coarse-grained soils and for special studies. Consolidation tests are used to estimate the per- centage of consolidation expected during construction. The shearing strength available at the critical period during construction is then determined on the basis of the percentage of the total consolidation that will have occurred. In places where only a small percentage of the total consolidation is expected to occur during construction, un- confined compression tests or unconsolidated undrained (UU) triaxial tests can be employed to determine the shearing strength. In all cases where consolidation is a factor in determining the shearing strength, piezometers should be installed to check the variations in pore pres- sures during construction. Open-ended pipes with a well point or sand pocket at the bottom are satisfactory for most foundations. Where more accurate measurements are required the Bureau of Recla- mation type of hydrostatie pressure gage is reeommended. Measures for control of underseepage are seldom designed on the basis of adequate information. Natural soils vary widely in character in both vertical and horizontal extent and are often stratified. Exten- Cu. 10] CONSTRUCTION MATERIALS 187 sive explorations are usually necessary to arrive at a reasonable esti- mate of permeability values. Positive cutoffs should be employed wherever feasible; therefore the exploration should first determine whether a cutoff is feasible along the center line or upstream of the cen- ter line. Borings made for this purpose should extend a sufficient depth into rock and be tested for water loss to determine if grouting of the rock will be required. If a cutoff is not practical, extensive explorations must be made upstream of the dam site to determine the extent and thickness of natural blanketing material and downstream to obtain foundation data for design of the required seepage-control measures. A row of deep borings at the downstream toe is required from which the effective size or relative permeability of each strata should be determined as a basis for relief-well design. In addition, borings should be made on several sections in order that continuity and extent of the pervious strata shall be outlined in the direction of flow. CONSTRUCTION MATERIALS The soils and geological survey of the site and surrounding area should develop the availability of (1) impervious and pervious ma- terials for the embankment, (2) sand and gravel for drains and filter blankets, and (3) stone for riprap. In order to appraise properly the suitability of the local material, the soils engineer and the geologist should know how these materials will be used in the embankment. In the design of dams the terms impervious and pervious are employed in a rather broad sense to designate materials which are relatively imper- vious or pervious when compared with other materials in the dam and foundation. In the classification of different soils with regard to permeability, the following table may be used: Impervious K less than 0.01 X 10~* em. per sec. Semi-impervious K from 0.01-1.0 X 107+ em. per see. Semi-pervious K from 1.0-50 X 10~* em. per sec. Pervious K from 50-500 < 10~* em. per see. Very pervious K greater than 500 X 10~* em. per see. A brief discussion of embankment sections and the use of these ma- terials therein are given in subsequent paragraphs. Depending on the availability of the relatively impervious material, the embankment may have a narrow central core such as employed on Sardis (Fig. 7), Mariposa (Fig. 8), and Franklin Falls (Fig. 12) dams, or a full impervious section as on Hulah Dam (Fig. 11). If there is 188 MiIppLEBROOKS. HARTH DAMS [Cu. 10 aX a reasonably equal distribution of im- | pervious and pervious material, the im- pervious material might form the upstream SY. half of the dam as in Cottage Grove (Fig. IK 9) and Great Salt Plains (Fig. 10) dams. S These dams represent the more general cases in which the “impervious” zone con- sists of impervious or semi-impervious ma- terial as classified above. However, there are flood-control dams in which the shell materials are so coarse that semi-pervious material is used in the core. If loss of water is not a problem for consideration, the main criterion to be satisfied is that the core be tight enough to insure a low saturation line in the downstream portion of the dam. This is usually accomplished in a satisfac- tory manner if the downstream shell is 100 times more pervious than the core. In cases where the entire embankment is composed of impervious material, or where a satis- factory ratio of permeability cannot be ob- tained, drainage is used to lower the satura- tion line. Pervious material is a most valuable as- set at any earth-dam site. Extensive ex- plorations are justified in locating possible sources for borrow areas. All types of pervious material from fine sand to gravel can be employed advantageously in the embankment section. Even sands generally classified as semi-pervious (silty or clayey sand) are relatively free-draining when compared with very impervious core ma- BS terials and will, when properly placed in aS the dam, assure a low saturation line. fe Locally available coarse sands and gravels are usually economical for use as filter BS blankets and drains even when screening is necessary to meet a required gradation. Where foundation stability is not a prob- lem, the most economical embankment sec- Drain Relief wells and seeded 12" topsoil, sodded Natural impervious blanket 10' to 12’ thick Min. impervious core Rolled fill Deep sand foundation Fie. 7. Hydraulic fill, Sardis Dam, Little Tallahatchie River, Mississippi. 3:0"dumped rock on 12" of gravel * Pervious shell Max. impervious core Spillway crest El.282 onservation pool n v Cu. 10] CONSTRUCTION MATERIALS 189 20° 10° El. 455.5 Spillway crest El.439.5 " Cobbl 3 ar es shell shell Cobbles Transition zone Transition zone iti iti Ske INU NY V7 SUING V7 X / V7 X ae Ze eau o > = Impervious core ‘ Stream-bed soils Quartz-mica Schist Fic. 8. Rolled fill, Mariposa Dam, Mariposa Creek, California. Pervious river gravels Fic. 9. Rolled fill, Cottage Grove Dam, Coast Fork, Willamette River, Oregon. 16"riprap, 25 [= IEl1, 1 f Mian pool EL tee. ee eee 10’ Gonservation == =~, BE 1,140. SZ 4 1,139.0 Collector pool EI.1,125.0 gE 1 — ff ditch SS 1,121. Ep pee Ae es SUNY SY NY SY SUSY SUSUNYU SUSY SYUNU SY SUSY RY “renee strata sand, and clay shale, and sand stone 7 Fig. 10. Rolled fill, Great Salt Plains Dam, Salt Fork of Arkansas River, Oklahoma. 1 | / 31>] E1.779.5 1-6" Dumped rock 3 Dumped rock on 6" spalls 2'-6'k 2'-6" drains at 15'c.c. Flood-plain alluvium of sands, silts, and clay Fig. 11. Rolled fill, Hulah Dam, Caney River, Oklahoma. 190 MIDDLEBROOKS. HARTH DAMS [Cu. 10 fe . tion 1s one with a narrow central core 5 with large sand and gravel shells. Re- cent large-scale triaxial tests have 5 shown that the angle of internal fric- tion for well-graded sand and gravel is about 45°. This high value of shear- ing resistance cannot be fully utilized, as it is not practical to construct on slopes steeper than 1 on 1% for angular gravels and about 1 on 2 for smooth, rounded gravels. The Mariposa Dam (Fig. 8) furnishes a good example of the economical use of such materials. The slopes are protected both upstream and downstream with large stones screened out of the main embankment fill. Soils from required excavation can in most cases be employed in the embank- ment to economic advantage even though some processing is necessary, such as the crushing and mixing of the spillway excavation at the Youghio- gheny Dam (Philippe, 1948). At the Fort Randall Dam the excess chalk rock excavating from the spillway and portal cuts was used in berms to reduce the main compacted dam section, and a heavy blanket of chalk was used on the upstream slope in place of costly riprap. A thick blanket of weathered eranite, stripped from the aggregate quarry, was also used to replace rip- rap on the Clark Hill Dam. Riprap for upstream slope protection S is one of the most expensive features S of earth-dam construction. Extensive 5 field investigations are justified to lo- cate the nearest source of satisfactory material. In a recent report, the Corps of Engineers (1948) gives results of a survey of riprap on more than thirty UW Dumped rock W TC ras Re Areata by 0) EX 3’ gravel 4' sand and gravel Selected pervious RR Pervious zone Gravel ‘ Gneiss and Schist rock Rolled fill, Franklin Falls Dam, Pemigewasset River, New Hampshire. Pervious zone Sand foundation TRIN W, Dumped rock Selected pervious U7, Fig. 12. 3' rock on 8"gravel U7, Cu. 10] EMBANKMENT DESIGN AND CONSTRUCTION 191 dams. The major findings of the survey are: a wide variety of rock, including some very soft formation, have proved successful; thinner riprap layers and correspondingly small stone can be used than have generally been employed for fetches not to exceed 10 miles; well-graded filter blankets are essential but may be relatively thin (6 to 12 inches). Required rock excavation, even though composed of relatively weak rock, can in most cases be used effectively in the upstream portion of the dam to reduce the quantity of required riprap. EMBANKMENT DESIGN AND CONSTRUCTION Stability of the embankment and foundation and seepage through the foundation are the principal design and construction problems that must be carefully considered by the soil mechanics engineer and geologist. Only a general summary of these problems will be given in this chapter because they have been adequately covered in other publications (Creager, Justin, and Hinds, 1945; Middlebrooks and Bertram, 1948; Middlebrooks, 1948). Factors affecting embankment stability are directly or indirectly related to the shearing strength of the foundation and embankment. The most critical period in the construction of the dam is just before it is brought to grade or shortly thereafter. At this time, pore pres- sures, due to consolidation in the embankment and foundation, are at a maximum. If pore pressures do not develop, there is usually no question concerning the stability of the embankment, since most soils have adequate strength when fully consolidated. The effect of rapid drawdown on the upstream slope deserves care- ful investigation. However, owing to the difficulty of accurately determining the degree of saturation that occurs during the filling period and the amount of drainage during the drawdown period, a detailed analysis is not usually justified. The method of analysis most generally employed on flood control dams is the instantaneous- drawdown method. A minimum safety factor of 1.0 is considered satisfactory in checking the slope stability for drawdown from maxi- mum to minimum pool level. Where there is any doubt concerning the possibility of developing pore pressure during construction, pie- zometers should be installed in critical areas. In the analysis of both upstream and downstream embankment slopes for stability during construction, a minimum safety factor of 1.50 is specified. For the stability of the upstream slope with reser- voir empty and of the downstream slope with steady seepage from full reservoir head, a minimum factor of safety varying from 1.5 192 MIDDLEBROOKS. HARTH DAMS [Cu. 10 for clean granular materials to about 2.0 for highly cohesive clays is needed. Most geologists may not be interested in the soil mechanics features that have been included in this discussion; however, they have been included to encourage greater interest in this field among geologists. It is considered most important in the design and construction of earth dams that the geologist at least understand the general principles of soil mechanics and that the soils engineer have a broader understand- ing of the effect of geological features on the application of the soil mechanics principles. REFERENCES Corps of Engineers (1949). Slope protection of earth dams: U. S. Waterways Exp. Sta., Vicksburg, Miss. Creager, W. P., Justin, J. D., and Hinds, J. (1945). Engineering for dams: John Wiley & Sons, New York, vol. 1, 245 pages. Middlebrooks, T. A. (1948). Seepage control for large earth dams: Proceedings of Third International Congress of Large Dams, Stockholm. and Bertram, G. E. (1948). Corps of Engineers practice on earth dams design and construction: Proceedings of Second International Conference on Soil Mechanics, Rotterdam, vol. 7 Philippe, R. R. (1948). Adoption of locally available materials for use in con- struction of earth dams: Proceedings of Second International Conference on Soil Mechanics, Rotterdam, vol. 4. CHAPTER 11 GEOLOGIC ASPECTS OF SOFT-GROUND TUNNELING Karu TERZAGHI Professor of the Practice of Civil Engineering Harvard University Cambridge, Massachusetts A geological report on the site of a proposed earth tunnel should contain full information on all those known geological details and con- ditions which may conceivably influence the behavior of the ground in the tunnel during construction. From a scientific point of view, most of these details are likely to be irrelevant. Therefore a geologist will hardly be able to prepare an adequate report on the site of a pro- posed tunnel unless he is familiar with the factors that determine the tunneling conditions. This chapter contains general information con- cerning these factors. In the discussion of tunneling conditions it is necessary to make use of various elementary concepts which are asso- ciated with soil mechanics. (See Terzaghi and Peck, 1948; and Chap- ter 5 in this symposium.) EFFECT OF STAND-UP TIME ON TUNNELING CONDITIONS In connection with soft-ground tunneling, the difficulties and costs of construction of a tunnel with given dimensions depend almost ex- clusively on the stand-up time of the ground. The term stand-up time indicates the time that elapses between the exposure of an area at the roof of the tunnel and the beginning of noticeable movements of the ground above this area. The factors that determine the stand-up time of an unsupported roof area with given dimensions depend to a large extent on the position of the water table. Above the water table, the stand-up time depends essentially on the tensile and shearing strength of the ground. Below the water table it depends not only on the strength of the soil but also on its average permeability and, to a large extent, on the degree of continuity of the most permeable members of the formation. A lens of water-bearing sand in a silt stratum may be perfectly harmless; a continuous layer 193 194 TerzaAcHI. SOFT-GROUND TUNNELING [Cu. 11 of the same sand may give rise to a “blow,” and, if this layer communi- cates with a large body of water-bearing sand within a short distance from the tunnel, a catastrophe may result. TUNNEL MAN’S GROUND CLASSIFICATION Unconsciously using the concept of the stand-up time as a criterion, the tunnel man distinguishes between six principal categories of “oround,” namely, firm, raveling, running, flowing, squeezing, and swelling ground. In firm ground the tunnel heading can be advanced without any roof support, and the permanent lining can be constructed before the eround will start to move. Typical representatives of firm ground are loess above the water table and various calcareous clays with low plasticity such as the marls of South Carolina. In raveling ground chunks or flakes of material begin to drop out of the roof or the sides some time after the ground has been exposed. If the process of raveling starts within a few minutes, the ground is fast-raveling. Otherwise it is referred to as slowly raveling. Fast- raveling conditions are likely to be encountered in residual soil or in sand with a clay binder located below the water table. Above the water table the same soils may be slowly raveling or even firm. In running ground the removal of the lateral support on any surface rising at an angle of more than about 34° to the horizontal is followed by a “run,” whereby the material flows like granulated sugar until the slope angle becomes equal to about 34°. Running conditions pre- vail in clean, loose gravel and in clean, coarse or medium sand above the water table. In clean, fine, moist sand the run is preceded by a brief period of raveling. A ground with such behavior is cohesive- running. Flowing ground moves like a viscous liquid. In contrast to running ground, it can invade the tunnel not only through the roof and the sides but also through the bottom. If the flow is not stopped, it con- tinues until the tunnel is completely filled. The rush of flowing ground into a tunnel is sometimes referred to as a blow. Flowing conditions prevail in any ground with an effective grain size in excess of about 0.005 millimeter, provided that the tunnel is located below the water table. Above the water table the same ground has the character of a raveling or running ground. Squeezing ground slowly advances into the tunnel without any signs of fracturing and without perceptible increase of the water content of Cu. 11] CHANGING GROUND CONDITIONS IN TUNNELS 195 the ground surrounding the tunnel. Although the squeeze may not even be noticed, the loss of ground due to squeeze may be important enough to produce a conspicuous trough-like subsidence of the ground surface above the tunnel. Squeezing conditions are met with in every tunnel through soft or medium clay. Swelling ground, like squeezing ground, moves slowly into the tun- nel, but the movement is associated with a very considerable volume increase ‘of the ground surrounding the tunnel. Swelling conditions are likely to develop in tunnels through heavily precompressed clays with a plasticity index in excess of about 30, and in sedimentary formations containing layers of anhydrite. PROCEDURES FOR CHANGING GROUND CONDITIONS IN TUNNELS All the serious difficulties that may be encountered during the con- struction of an earth tunnel are directly or indirectly due to the perco- lation of water toward the tunnel. Therefore most of the techniques for improving the ground conditions are directed toward stopping the seepage. In cohesionless or slightly cohesive ground, such as clean or silty sand, a mixture of soil and water enters the tunnel through every gap in the lining. Such “flows” or “blows” can be avoided by three dif- ferent means. The water table can be lowered to a level located be- low the bottom of the tunnel by drainage; the tunnel can be filled with compressed air under a pressure equal to the water pressure at the bottom of the tunnel, or the voids of the ground surrounding the tun- nel can be clogged by the injection of suitable materials such as ce- ment or mixtures of chemicals. The water that occupies the voids of a porous material above the water table is retained in the voids by capillary forces. (See, for in- stance, Terzaghi and Peck, 1948, pp. 115-129.) Hence, if the water table is lowered by drainage to a level below that of the bottom of the tunnel, the water ceases to percolate toward the tunnel, though the voids of the ground may remain almost or entirely filled with water. Drainage changes flowing ground into raveling or running ground which can easily be handled. Another method of improving the properties of soft ground involves the use of compressed air. The principle of this method is illustrated by Fig. 1. Figure la is a vertical section through the heading of a tunnel in flowing ground, The tunnel is filled with compressed air 196 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 under a pressure equal to the water pressure at the bottom of the tunnel. In Fig. 1b the horizontal distances between ac and ab repre- sent the pressure in the water which occupies the voids of the ground. Since the air pressure on the heading is greater than the water pres- sure, the compressed air not only stops the flow of water into the tun- nel, but it also tends to drive the water out of the voids away from the tunnel. The rate at which the water is displaced depends on the air pressure and on the effective size Di of the soil particles. The value of Dy indicates that 10 percent of the total dry weight of the soil Ground surface UY SUSY SUSY SC Water table =< ———— ORG eee ee iP ane eee Water pressure ‘alae KK Air Excess of air pressure Roof lining leakage — over water pressure c We) | __ jir pressure——>] (a) (0) Fic. 1. Diagram illustrating the principle of tunneling with compressed air. consists of particles with a grain size equal to or smaller than Dyjpo. If Dyo of the soil particles is smaller than about 0.01 millimeter, the rate of displacement is likely to be zero, because the surface ten- sion of the water may prevent the air from entering the voids. If Dio is greater than about 0.01 millimeter, the rate of displacement rapidly increases with increasing Dy; if Do is greater than about 0.2 milli- meter, it becomes so high that a considerable quantity of air leaks out of the tunnel. Under such conditions the miners are compelled to plaster the joints in the temporary tunnel support and part of the ex- posed surfaces of the ground with mud to reduce the loss of air. If the heading strikes an exceptionally permeable layer, the air pressure in the tunnel may suddenly drop to zero, whereupon the tunnel is flooded. Such an accident is known as a blowout. In unconsolidated deposits with an erratic structure, the danger of a blowout is consider- ably greater than in regularly stratified ones, because they do not contain continuous layers of relatively impermeable material which interfere with the formation of air channels leading from the heading toward the water table. Cu. 11] TREATMENT OF SWELLING GROUND 197 In fine-grained soils such as silt or soft clay or in mixed-grained soils with a silt or clay matrix, the air does not enter the voids at all. These soils are changed by compressed air from flowing or rapidly squeezing into raveling or slowly squeezing ground. The reason is illustrated by Fig. 2, representing a vertical section through the center line of a tunnel in soft clay. The clay tends to flow along a curved sur- face of sliding ab into the tunnel. This tendency is due partly to the weight of the clay located above ab and partly to the seepage pressure exerted by the water which per- colates toward the heading. Bute Compressed air under a pressure NR equal to the water pressure at the level of point b not only stops the seepage as shown in Settlement Fig. 1, but it also resists the tendency of the solid clay par- Zone of maximum My ticles to descend toward the shear % x heading. "by Ne Squeeze A third possibility for improv- ing the ground conditions in 6 earth tunnels consists in clogging Fic. 2. Diagram illustrating the cause of the “squeeze” in tunnels through soft clay. The clay moves slowly along ab toward the heading. the voids of the ground sur- rounding the tunnel by grouting. So far this method has been used only for reducing air leakage through gravel strata and for reinforcing the ground located between the roof of shallow tunnels and the base of foundations located above the roof (Harding and Glossop, 1940). The limits which nature has imposed on the successful application of the different methods for improving the ground conditions in earth tunnels are discussed below in the section on “Methods of Soft-Ground Tunneling.” TREATMENT OF SWELLING GROUND Even a very dangerous swelling ground may behave like a firm ground at the heading. However, a few hours or days after exposure the ground begins to advance toward the tunnel, and the ground move- ment continues until the tunnel is closed up, unless the movement is stopped by the construction of a tunnel support. The movement is associated with a considerable increase of the volume of the material surrounding the tunnel. After the tunnel support is erected and back- 198 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 filled, the swelling ceases; but the pressure on the support increases first at an increasing, later at a decreasing, rate, and it approaches an ultimate value which may be considerably greater than the over- burden pressure. In tunnels through clay this is not possible un- less the initial horizontal pressure in the clay is considerably greater than the vertical pressure. The most common cause of such stress conditions is intense compaction under the weight of superimposed strata which are subsequently removed by erosion. Settlement due to loss of ground US YUSYUNUNG | A few weeks after mining Swelling clay 7 Zone of swelling / ZL Before mining 1 __->< tone of compression Fic. 3. Diagram illustrating the mechanics of the swelling of stiff clay in a tunnel. The change in the distribution of the water content caused by the mining opera- tions is shown on the right-hand side. According to a widespread belief the water that causes the swelling comes out of the air contained in the tunnel. Therefore many at- tempts have been made to prevent the swelling by covering the ex- posed surfaces in tunnels with a bituminous coating. Most of these attempts have failed because, as a rule, the water that causes the swelling comes out of the ground and not out of the air. The mechanics of this process is illustrated by Fig. 3. The left-hand side of the dia- gram shows the distribution of the water content of the clay over a horizontal section through the center line of a proposed cylindrical tunnel in a heavily precompressed clay prior to excavation. Excava- tion relieves the confining pressure on the clay adjoining the walls of the tunnel. According to the theories of soil mechanics, local re- moval of the load on a saturated porous material draws water out of the loaded portions of the material toward the zone of pressure re- laxation. In the tunnel the zone of pressure relaxation has the shape of a cylindrical mantle surrounding the tunnel bore. On account of Cu. 11] TREATMENT OF SWELLING GROUND 199 the pressure relaxation, water is drawn toward this zone out of the clay located at a greater distance from the tunnel. Therefore the water content of the clay adjoining the walls of the tunnel increases and the clay swells, whereas beyond this zone the water content decreases. On the right-hand side of Fig. 3 this change of the water content is in- dicated by shaded areas. The theory illustrated by Fig. 3 has been verified repeatedly by the results of water-content determinations on samples from the proximity of the walls of tunnels in the “Argile plastique” of the region of Paris, France, a stiff Eocene clay with a plasticity index of 60 to 80 percent. The clay is similar in many respects to the London clay. It rests on the Cretaceous shales of Meudon. It is overlaid by Eocene limestones, and the entire formation is gently folded. The thickness of the clay stratum varies between about 100 and 200 feet. The state of precom- pression was produced by the weight of superimposed Tertiary and younger strata which were later removed by erosion. The temporary load due to the weight of these strata was of the order of magnitude of 25 tons per square foot. A general description of the physical properties of the clay was published by Langer (1936). One of the subway tunnels of Paris was excavated in this clay. The average initial water content of the clay was 56 percent. Within a few weeks after exposure the water content of the clay at the walls of the tunnel had increased to values between 90 and 130 percent. With increasing distance from the walls the water content decreased, and it had a minimum of 46 percent at a distance of about 13 feet from the walls (observations by Langer, reported by Terzaghi, 1936). Because the water that causes the swelling of stiff clay comes out of the clay, the swelling of the clay can be prevented only by a tube- shaped tunnel support which is strong enough to sustain the swelling pressure. However, the ultimate pressure on the tube can be reduced considerably by providing a clearance between the walls of the tunnel bore and the extrados of the tunnel support, or else by using tunnel supports which can yield considerably under pressure without being crushed. Experience indicates that a clearance of about 6 inches commonly serves the purpose. A description of the different systems of yielding steel supports used in Belgian and German coal mines for re- ducing the ultimate load on the supports was given by Ernould (1934). Application of a watertight coating can do no more than prevent relatively harmless surface processes such as the raveling of jointed clay or shale due to desiccation. 200 TERZAGHI. SOFT'-GROUND TUNNELING [Cu. 11 METHODS OF SOFT-GROUND TUNNELING The methods used in constructing an earth tunnel must be adapted to the stand-up time ¢, of the ground and to the performance of the ground during the process of mining. The following paragraphs con- tain a brief review of the methods, illustrated by Fig. 4. In firm ground, with a stand-up time ¢t, of more than about one day, no temporary support is required (Fig. 4a). In raveling ground with a t, value between 1 day and 5 minutes and in squeezing ground, steel liner plates are mined in, one by one, and assembled into rings or “courses” as shown in Fig. 4b. This can be done without giving the ground an opportunity to move. The face does not require any lateral support. In cohesive running or running ground, the leading edge of the roof and side support must be kept ahead of the excavation. This is accomplished by means of “poling boards” or “poling plates” (Fig. 4c). The face must be supported by means of breast boards. As ex- cavation proceeds, the breast boards are carried forward, one by one, over a distance equal to the width of one course of liner plates, start- ing at the crown. Really serious difficulties are encountered only in flowing ground. It appears that the problem of mining through flowing ground was not adequately solved until the beginning of the nineteenth century, when coal-mining operations were started on a large scale. The technique that was developed under the pressure of necessity is illustrated by Fig. 4d. The roof and the sides are supported by “poling boards” which are driven at a slight angle to the center line of the tunnel. The face is breasted, and the tunnel bottom is covered with a timber floor strong enough to prevent a heave of the ground located beneath the tunnel. The procedure calls for expert carpenters, because no open joints between boards can be tolerated. “Where the material is very bad only small openings should be made and while one miner is making it, another must stand ready with hay and filling material to stop the hole as soon as enough ground has, for the time, been admitted; this stopping must often be done while the man is splashed over with the pouring mass and standing knee-deep in it, groping with his eyes shut to perform his task. It is hardly necessary to say that none but the best and most reliable workmen can be depended on to show the requi- site coolness and dexterity at the right moment” (Drinker, 1878). In spite of skill and utmost precaution, catastrophes can not always be avoided, and under difficult conditions the progress slows down to 1 foot per week, two twelve-hour shifts per day. Cu. 11] METHODS OF SOFT-GROUND TUNNELING 201 On account of the costs and hazards involved in mining through flow- ing ground, the attempts to simplify the procedure started very early. They led to various refinements of the methods of drainage and, as early as the middle of the nineteenth century, to the invention of the compressed-air and the shield methods of tunneling. The original method of draining the ground surrounding a large : (a) Firm Ground (d) Flowing Ground (original method) ¢,=0 Liner plates ; Poling plates Be Shield — SS S WIS Breasting ie >5 min. (6) Raveling or Squeezing Ground (e) Coarse-Grained Flowing Ground (shield method) min. (c) Cohesive Running Ground (f) Soft Flowing Ground (shield method) Fic. 4. Principal types of temporary tunnel supports. tunnel by means of drainage galleries located on both sides of the tunnel was superseded by pumping from filter wells or wellpoints. The wells are drilled on both sides of the tunnel from the surface of the ground to a depth of 10 or 15 feet below the bottom of the tunnel, or else they are installed on the bottom of the tunnel. New wells are drilled as the heading advances. Both procedures have been suc- cessfully used on tunnel jobs. However, drainage by pumping from wells is impracticable unless the effective grain size Dy of the most permeable layers in the ground surrounding the tunnel is greater than about 0.05 millimeter. Under exceptional conditions, finer soils can be drained by means of the vacuum or electro-osmotic method (Ter- zaghi and Peck, 1948, pp. 337-340). The function of the compressed air in tunnels is illustrated by Fig. 1. 202 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 Compressed air under adequate pressure has the same beneficial effects as successful drainage. It transforms flowing into running, raveling, or slowly squeezing ground. If the air pressure is lower than about 18 pounds per square inch, corresponding to a head of water of 36 feet, the compressed air has no detrimental physiological effects. If the air pressure is increased beyond 18 pounds per square inch, its effects on the human organism become more and more harmful; the number of working hours per shift must be reduced; at the same time the wages per shift go up, and an air pressure of about 50 pounds per square inch, corresponding to a head of water of 100 feet, is the highest the organism of a normal person can stand. Hence the use of compressed air is limited to tunnels located at a depth of less than 100 feet below the water table. Furthermore, the use of compressed air requires the installation of a compressor plant and of airlocks, which are expensive. For this reason, in short tunnels or when a tunnel is driven over a short distance through flowing ground, it may be more economical to drain the ground, to consolidate the ground ahead of the working face by means of injections, or, as a last resort, to try the original procedure for mining through flowing ground, shown in Fig. 4d. The shield method of tunneling is illustrated by Figs. 4e and 4f. A shield is merely a ring-shaped biscuit cutter which is shoved ahead by means of hydraulic jacks. If the shield travels through coarse-grained flowing ground, it is necessary to transform the flowing into running or raveling ground by means of compressed air. The ground must be ex- cavated ahead of the shield, the roof located beyond the cutting edge must be supported by poling boards or other convenient means, and the face must be breasted as shown in Fig. 4e. In a shield tunnel the pressure on the breast boards can easily be transferred onto the heavy steel beams that subdivide the central opening into smaller panels. Furthermore if a run or a blow occurs, the miner has at his disposal ready-made supports against which he can brace his emergency bulkheads. However, a shield is very expensive, and the shield method of tunneling requires a very much heavier tem- porary lining than does the method of hand mining, because the lining must be strong enough to sustain the heavy longitudinal pressure ex- erted by the jacks. Hence, if the ground conditions call for excavation ahead of the shield, it is commonly more economical to construct the tunnel without the assistance of a shield. In soft, flowing ground such as river silt, no excavation ahead of the shield is required. It suffices to push the shield ahead by means of hydraulic jacks. The ground ahead of the shield is displaced and Cu. 11] SUBSOIL EXPLORATION 203 enters the tunnel through the portholes shown in Fig. 4f._ This pro- cedure is very economical because it permits rapid progress and it practically eliminates tunneling hazards. Therefore, nowadays, long tunnels through soft, flowing ground are seldom constructed without the use of a shield. SUBSOIL EXPLORATION Reliable information on the subsoil conditions can be obtained only by means of test borings. The methods for making the borings were developed in connection with the exploration of coal measures, and they are more than a hundred years old. However, up to the beginning of the twentieth century, soil samples were secured by means of tools, such as augers or mudpumps, which destroy the structure of the soil and change its consistency. In tunnels the behavior of sediments with similar grain-size characteristics and mineralogical composition varies greatly with their porosity and water content and with the details of stratification. A varved clay, for instance, is likely to soften much more rapidly than a homogeneous clay with equal average water content. Hence, in connection with the subsoil exploration for earth tunnels, relatively undisturbed samples should be secured by means of one of the many sampling devices developed during the last decades. A complete report on the present status of the technique of sampling is being prepared by the Soil Mechanics Division of the American Society of Civil Engineers. The amount of information that can be derived from the results of the test borings is limited by the fact that it is seldom practicable to make more than one boring for every hundred feet of the length of the tunnel. Furthermore, if compressed air is to be used, borings should only be made well beyond the horizontal boundaries of the space to be occupied by the tunnel. Thus the test borings furnish informa- tion only on the sequence of the strata along a few vertical lines, widely separated and located beyond the boundaries of the tunnel. On the basis of this fragmentary information, the investigator is compelled to construct a continuous geologic profile through the center line of the tunnel. The results of this operation can be very misleading, unless the geologic structure is very simple or the profile is prepared by an experienced geologist and is accompanied by a report dealing with the possible differences between profile and reality. A thorough investigation of the unpredictable variations of the soil properties be- tween drill holes was made during the construction of the subway of Chicago (Terzaghi, 1943). 204 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 POSITION OF THE WATER TABLE At the beginning of this chapter it was shown that the working con- ditions in tunnels through sediments of any kind other than silt and clay depend primarily on the position of the water table with reference to the bottom of the tunnel. Therefore the second prerequisite for a reliable forecast of tunneling conditions consists in securing conclusive information regarding the position of the water table. Even today it is by no means uncommon that the position of the water table is inferred from the height to which the water rises in the drill holes between shifts. A conclusion arrived at on such a basis can Moraine Caisson section (J Moist,fine sand [600° C) 20' Tunnel 2,600' Fic. 5. Geological section through a terminal moraine in northern Switzerland, along the center line of the Emmersberg Tunnel. (After Schweizer Bauzeitung, 1895, p. 135.) be utterly misleading, because in sand it may take only a few hours whereas in silt or clay formations it may take several weeks or years for the water level in the drill hole to arrive within a few inches of the position of the water table. The practical consequences of erroneous conceptions regarding the position of the water table are illustrated in Figs. 5 and 6. Figure 5 is a section across a terminal moraine along the center line of the Emmersberg Tunnel in northern Switzerland. The tunnel has a length of about 2,600 feet, and the crest of the moraine is located at an elevation of about 70 feet above the roof of the tunnel. The moraine consists of a dense mixture of rock fragments, grit, and rock flour. It contains large, irregular lenses of very fine sand which do not communicate with each other. Since the permeability of the sur- rounding moraine material is very low, the water table in the sand lenses is located at very different elevations. Some of the lenses do not contain any free water at all. In the first lenses that were en- countered, the sand was only moist. As a consequence it was so stable that no breasting was required. However, as soon as the heading ar- Cu. 11] POSITION OF THE WATER TABLE 205 rived at the first wet pocket, the sand exhibited the character of flowing ground. Large quantities of sand mixed with water invaded the tunnel. Sinkholes appeared on the ground surface above the tunnel. When attempts were made to continue the mining operations in the tunnel under compressed air, the disturbance of the ground surround- ing the tunnel was already so far advanced that the attempts failed, Elevation above sea level in feet 0 500 1,000 1,500 2,000 2,500 ft LZ Stiff clay Fine, silty sand 32] Coarse, clean, silty, or argillaceous sand Fic. 6. Geological section through a valley in the city of SAo Paulo, Brazil. The strata are part of a Tertiary delta formation and the relief is believed to be post-Plhiocene. (After E. Pichler.) and it was decided to finish the job by means of a series of compressed- air caissons which were carried down from the ground surface to the level of the bottom of the tunnel (Schweizer Bauzeitung, 1894-1895). Figure 6 is a simplified geological section through one of the dis- tricts of the city of Sao Paulo in Brazil. The city is located on Tertiary deposits which appear to have been formed by an aggrading, meandering river. As a result of a lowering of base level, the river deposit was dissected to a depth of at least 100 feet below the original elevation of the valley floor. The sediments consist of stratified layers of clay, silt, and silty sand and gravel. In horizontal directions there are imperceptible transitions from one stratum into another. There- fore the position of the water table is not well defined. When a hole 206 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 is drilled, the water is likely to rise from different strata to different elevations. The water table marked “principal water table” is the ap- proximate locus of the points to which the water rises in drill holes from most of the strata. A few years ago a municipal waterworks tunnel was built within the city limits. The tunnel is 9 feet wide and high and about 1,500 feet long. The bottom of the tunnel is located about 100 feet below the street surface and 50 feet below the principal water table. The outer parts of the tunnel, between the portals and the two ends of the middle portion, with a length of about 500 feet, are located in water-bearing, fine to medium sand with some clay. The tunnel was started as a free-air, hand-mined tunnel with timber supports. However, the ground moved so rapidly into the tunnel and the street surface settled so badly that air locks were installed. As the ground was already badly disturbed, the loss of air was excessive. Therefore the contractor did not succeed in drying up the bottom, and it was necessary to use forepoling and to breast the working face. The loss of ground was about 50 percent, and the buildings located above the tunnel settled by amounts up to 8 inches. The middle part of the tunnel, with a length of about 500 feet, is located in stiff, yellow clay. No difficulties were encountered in this section. In view of the decisive influence of the position of the water table on tunneling conditions, no guesswork regarding this position should be tolerated. If the subsoil contains continuous layers of fairly clean. sand or gravel, sufficiently reliable information concerning the ground- water conditions can be obtained by transforming the drill holes into observation wells and by keeping a record of the variations of the water level in these holes. If the subsoil consists of silty sand or silt, it is necessary to install in the drill holes piezometric tubes with a diameter of less than half an inch. Such tubes have been successfully used for observing the rise of the piezometric surface for a silty stratum due to the weight of a superimposed fill (Casagrande, 1949). However, if the ground surrounding the site of a proposed tunnel consists only of fine silt or clay, no accurate information regarding the water table is required, because the performance of these sediments in the tunnel depends only on their water content and consistency. FUNCTIONS OF THE GEOLOGIST IN EARTH-TUNNEL JOBS The foremost duties of the geologist on an earth-tunnel job are the construction of a geological profile along the center line of the proposed Cu. 11] FUNCTIONS OF THE GEOLOGIST 207 tunnel, based on the results of the test borings; selection of suitable sites for supplementary borings which are needed for eliminating the most objectionable uncertainties involved in the construction of the profile; and correlation of the results of the ground-water observations with the geologic structure of the site. On the basis of the results of a general geologic survey of the region surrounding the tunnel site, of a visual inspection of undisturbed samples, and of the soil tests performed by the engineer, the geologist ascertains the origin of the strata which were encountered in the test borings. In accordance with his findings he designates the strata as river, flood-plain, shore, lake, marine, glacial, fluvioglacial, wind-laid, or composite deposits. The next step is to construct a geological pro- file which is compatible with the geological origin of the strata. On account of his knowledge of the structure of sedimentary deposits, the geologist will be in a position to pass competent judgment on the degree of accuracy of the profile. In the report attached to the profile he will also inform the engineer about the possibility or probability of significant deviations from the profile such as the existence of pockets of water-bearing sand or of buried channels which were not encountered during the boring operations. Finally, the geologist should carefully examine and comment on the data that have been secured regarding the ground-water conditions and correlate them with the geological structure of the site. If there are indications of the existence of one or more perched water tables or of artesian conditions, not clearly revealed by the records, he should insist on the installation of all those supplementary observation wells or piezometric tubes which are necessary to clarify controversial issues. The effects of geological details on tunneling conditions depend en- tirely on the physical properties of the sediments and on the interaction between solid and water. Therefore the geologist will hardly be able to render satisfactory services on a tunnel job unless he knows the fundamental principles of soil mechanics. The geologist should also be acquainted with the techniques of boring and sampling and with the sources of the manifold errors that may enter into test boring records. On the other hand the engineer will hardly realize the uncertainties in- volved in the construction of geological profiles and the benefits he can derive from competent geological advice unless he is familiar with the elements of physical geology. Hence an active interchange of informa- tion between the two professions is urgently needed. 208 TERZAGHI. SOFT-GROUND TUNNELING [Cu. 11 TOPICS FOR FUTURE RESEARCH The most important subjects for future research activities in the field of soft-ground tunneling are the details of the structure of sedi- mentary deposits, the behavior of different sediments at tunnel head- ings, and the loads on temporary and permanent tunnel supports. A great deal of time and labor has been devoted by geologists to an investigation of the different processes of sedimentation, but the products of sedimentation have received relatively little attention. We need systematic collections of geological profiles showing the structural patterns which are produced by different processes of sedimentation. These patterns can be seen only on the slopes of open cuts, immedi- ately after excavation, on the slopes of sand, gravel, and clay pits, and at the headings of tunnels during construction. Familiarity with such patterns would discourage the engineer from constructing geological profiles without the assistance of a geologist, and it would help the geologist in the interpretation of test boring records. The behavior of different sediments at tunnel headings can be learned only from personal experience, which is inevitably limited, and from case records accompanied by an adequate description of the geology of the tunnel site. Yet, up to this time, very few records of earth tun- nels have been published, and in most of the published ones adequate geological information is conspicuous by its absence. The measurement of the load on tunnel supports is needed as a check on our methods of estimating the loads in advance of construction. Extensive observations of this kind were made during the construction of the subway in Chicago (Terzaghi, 1942, 1943). Equally important is quantitative information about the pressure exerted by swelling ground on rigid and yielding tunnel supports, because this pressure can be estimated only on the basis of precedents. Measurements of swelling pressures were carried out in various tunnels in the “Argile plastique” of the region of Paris (Langer, quoted by Terzaghi, 1936). At present, similar measurements are being made in coal mines in southern Holland (according to a personal communication received by the author from Mr. J. M. Hermes, Staatsmijn Emma, Treebeek, Hol- land), and, by the U. S. Army Engineers, in an experimental tunnel at the site of the Garrison Dam in North Dakota. REFERENCES Casagrande, A. (1949). Soil mechanics in the design and construction of the Logan Airport: Jour. Boston Soc. Civ. Engrs., vol. 36, pp. 192-221. Cu. 11] REFERENCES 209 Drinker, H. S. (1878). Tunneling: John Wiley & Sons, New York (out of print), 1031 pages. Ernould, V. (1934). Le souténement métallique dans les mines: L’ossature métallique, vol. 3, pp. 478-490. Harding, H. J. B., and Glossop, R. (1940). Chemical consolidation in railway work: Railway Gazette (London). Langer, K. (1936). The influence of the speed of loading increment on the pressure void ratio diagram of undisturbed soil samples: Proceedings of Inter- national Conference on Soil Mechanics and Foundation Engineering, Cam- bridge, Mass., vol. II, pp. 116-120. Schweizer Bauzeitung (1894 and 1895). Der Emmersberg Tunnel bei Schaff- hausen, vol. 24, pp. 67-69, 75-77; vol. 25, pp. 135-1387. Terzaghi, K. (1936). Discussion: Proceedings of First International Conference on Soil Mechanics and Foundation Engineermng, Cambridge, Mass., vol. III, pp. 152-155. (1942). Shield tunnels of the Chicago subway: Jour. Boston Soc. Civ. Engrs., vol. 29, pp. 163-210. (1943). Liner plate tunnels on the Chicago (IIll.) subway: Trans. Am. Soc. Civ. Engrs., Paper 2200, vol. 69, pp. 970-1007. and Peck, R. B. (1948). Sol mechanics in engineering practice: John Wiley & Sons, New York, 566 pages. CHAPTER 12 SEDIMENTARY GEOLOGY OF THE ALLUVIAL VALLEY OF THE LOWER MISSISSIPPI RIVER AND ITS IN- FLUENCE ON FOUNDATION PROBLEMS W. J. Turnpuit, E. L. Krinitzsxy, anp 8. J. JOHNSON Respectively Engineer, Geologist, and Engineer Soils Division, Waterways Experiment Station Corps of Engineers, Department of the Army Vicksburg, Mississippr ALLUVIAL DEPOSITS ALLUVIATION IN THE Lower MIssIssipPI VALLEY The most important recent geological influence on Mississippi River behavior has been the advance and retreat of continental glaciers. It has been estimated, and relatively accurately because the terminal moraine outlines are fairly well known, that 20 million square miles of land surface were covered by ice at its maximum extent. The estimates of thickness of this great ice cap are not in such common agreement, but a figure of 1 mile for average maximum thickness is often given. This would represent a volume of 20 million cubic miles of landlocked water. Theoretically, such a vast volume when re- moved from the ocean basins would be sufficient to lower sea level somewhere in the neighborhood of 450 feet, even if allowance is made for the present-day ice masses which have not yet been melted. With sea level at such a considerably lower stand during times of glacia- tion (Daly, 1929), the Mississippi River was flowing with an overly steep gradient and thus eroded deeply into its valley. But, with the rise of sea level during waning of glaciation and the consequent loss of gradient along the Mississippi River’s course, alluviation of its former incised valley took place. Since continental glaciation has occurred during five distinct times, there are an equal number of such alluvial deposits in the Lower Mississippi Valley. These are in the form of the modern Alluvial Val- ley fill plus four distinct terraces, which are recognized according to 210 Cu. 12] ALLUVIATION 211 their individual elevations (Fisk, 1944). The oldest of these terraces has the highest topographic expression, whereas the remainder have progressively lower elevations in order of decreasing age. This con- dition has resulted from a progressive continental uplift which has raised the deposits vertically, thus protecting portions of them from removal by subsequent erosive action of the Mississippi River. Their present areal expression is that of strips bordering the modern Al- luvial Valley and elongated “islands,” such as Crowley’s Ridge, which divide its northern portion. Southward, in southern Louisiana, these terrace formations have been downwarped as a result of sinking in the Gulf Coast geosyncline and plunge under Recent deposits of the Mississippi deltaic mass. The areal distribution of these sedimentary units in relation to the modern Mississippi River and to physiographic basins and ridges within its valley is illustrated in Fig. 1. Each of the terrace deposits reflects a sequence of deposition which, as a result of lesser modification by erosion, is ideally shown in the Recent alluvial fill. During its last period of overly steepened gradi- ent, the Mississippi River transported very coarse loads containing much gravel, derived chiefly from adjacent Paleozoic upland areas. However, with the slow decrease in gradient resulting from gradually rising seas, much of this coarse load was deposited, thus forming the present valley fill. Eventually, the Mississippi River’s gradient slack- ened to such a point that gravels were no longer being transported, although alluviation was continuing with further advances in sea level. Under these circumstances, the alluvial deposits were formed of finer materials, mostly sands. Thus it can be seen that, on the whole, the Recent alluvium can be divided sharply into two types: (1) the lower graveliferous section and (2) an upper, essentially non-graveliferous section. In the graveliferous section, about 95 percent of all samples con- tain some gravel; in the non-graveliferous section, less than 1 percent of the samples show any gravel whatsoever. In general, the gravel- iferous section contains gravels somewhat in excess of 25 percent by weight, the remainder being coarse sand. On the other hand, some parts of the graveliferous section may contain only fine- or medium- grained sands with little or no gravel. In general, the cobbles, where present, are larger at depth and smaller in the shallower layers. It is common, north of the Louisiana-Arkansas boundary, to find occasional cobbles 1 foot in diameter, but southward there is a rapid decrease in size, so that in Louisiana it is uncommon to find specimens over 3 inches in diameter, and toward the coast gravel over 1 inch in diameter 212 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 8) ay PE (~.~.) Le LEGEND FLOOD PLAIN FLOOD PLAIN AREA OF DELTAIC DEPOSITS RECENT ALLUVIUM LOCALLY ABOVE FLOOD LEVEL PLEISTOCENE ALLUVIAL 7 "AT CHAFAL Ara B8AY Fic. 1. Extent of flood plain and of adjacent deposits in Lower Mississippi Valley. Cu. 12] ALLUVIATION 213 is rarely found. Also, going north anywhere beyond the Louisiana- Arkansas line, the gravels lie close to the surface, often within 15 feet; however, coming south into Louisiana and southern Mississippi, the cover thickness increases so that the gravels are 100 feet below the surface in the vicinity of Houma, Louisiana. Thicknesses within the graveliferous sections as a result of slope controls during deposi- tion show somewhat less of a distinct north-south variation than is found in the overlying finer deposits. The gravel section near Sikeston, Missouri, is 180 feet in thickness. Near Memphis, Tennessee, the sec- tion is on the order of 100 feet but increases again to approximately 150 feet in the latitude of Yazoo City, Mississippi. Near Natchez, Mississippi, the value is about the same, but near New Orleans a thickening occurs in the deep-lying gravels so that 200-foot units can be measured. Also, there are important changes in the gravel sections within local portions of the valley, the gravels being thicker toward each tributary stream, where they assume a fan-shaped form. The non-graveliferous deposits are themselves divisible into two phases: (1) A pervious phase composed mostly of clean, well-washed sands exists in the lower part of the section. These thin out toward the north, where the whole section diminishes, and also thin southward because after alluviation became advanced sand was no longer carried in large quantities by the river below the central part of the valley. (2) The upper, less pervious phase consists of a layer of silts and clays which thin to the north but thicken southward and become especially thick in southern Louisiana, where they are part of the deltaic mass. These are the most typical materials of the modern Mississippi River flood plain, and they are the most important in their effects on modern river behavior. A special sort of sediment, the origin of which has been disputed in the Lower Mississippi Valley, is the loess that caps almost all terrace deposits along Crowley’s Ridge and the eastern Alluvial Valley walls. It extends from north of Cairo, Illinois, all the way southward to Bayou Sara in Louisiana. The loess is a homogeneous, unstratified, calcareous silt with some plasticity and has a tendency to vertical cleavage. Its thickness on hill crests varies from as much as 65 feet near the Alluvial Valley bluffs to about 10 feet or less on terrace margins. There it merges into a similar soil material known as brown loam. In Fig. 2 there is depicted a geological section through Recent and Pleistocene deposits of the Mississippi Valley in the latitude of 214 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 Natchez, Mississippi. The sedimentary sequence in both the valley fill and terrace deposits is indicated in the coarse to fine gradation of the materials in these sections. Loess is represented as restricted chiefly to crests of terrace hills of the eastern valley wall. MEANDER-BELT SEDIMENTS When sea level reached what is essentially its present stand, the Mississippi River ceased carrying coarse sediments into its valley and achieved what has been called a “poised” state. This is a condition in which the river has no pronounced tendency to do much valley deepening or filling. However, the behavior of the river is such that earlier sediments through which it flows become reworked to produce meander-belt deposits having distinctive properties. In a poised river, meandering or the development of “bends” is a trademark. And, as these loops in the river’s course are continuously enlarging and being abandoned while newer belts form, they result in reworked sediments that are left as superficial scars covering much of the flood plain. Resulting sediments of the following sorts are typical in the Mississippi River Valley (Fisk, 1947). Figure 3 depicts the various types of deposits. Point-bar deposits. As meander loops increase in size, the core of land in their interior expands through the addition of a series of arcuate ridges and intervening swales, which result from river deposition in a strip-like manner on growing bars. Most of the sediment contributed consists of sand and coarse silt. Topographically, these ridges may average 6 to 10 feet in height and usually conform to the channel curvature within which they were formed. As a result of frequent downstream migrations of channels which accompany bar development, these strips often truncate each other and otherwise complicate their patterns. Such truncation of bar ridges and the relation of these ridges to each other as individual accretions are shown in the point- bar deposits pictured in Fig. 3. Swale fillings. The swales, or topographic lows between accretion bars, often develop dense willow growths and trap fine sediments during high water long after their bordering ridges have developed. Thus these ridges become separated by intervening shallow strips of silts and silty clays which contain much organic matter. Within the point bar of Fig. 3 are swale strips situated between the bar ridges. The swales can be recognized by their heavy vegetation. Natural levees. Natural levees are the higher parts of a river’s flood plain, and they flank the immediate sides of its channel. These deposits are laid down in the form of thin layers resulting from deposi- 215 JSW 1334 NI NOILVA373 MEANDER-BELT SEDIMENTS Cu. 12] ‘Tddississt(l ‘ZoyoyeNT 1wou syisodep ouss04slo[q pUB JUSs0EY YSno1y} UOTYdes [woIBoToay °zZ ‘DI 'w-V NOILOSS S2711N Ni 3DNVLSIO Sb Or Se Oe sz 02 7 o1 S ° ose- wooua3e Auvinuai feASae4 002- sonvs snowsaranvus [Ee23] SNOILYWWHOS AYVWLNSWIGaS we (AN3DO0IW) AYVILYSL Wnivy1sdor snoiuzew) ESS} I ote $5307 Il EEE Tee ON3937 , at os- ° m ba m < os 3 fe) SYSNIL = 001 z a AVIAnITIY wey fe z uo oo2 © os2 “SSIW ‘ZFHILYN oo sLisogga SLISOd3a ee SDV yy aL ——S a, SBovyysaL SNSS0LsI3a7a eeainiien ves = SNSSO0LSIS Nd Jo, dVW NOILYS071 ©inquos\|sioy % ym La da 216 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 tion when the river overtops its banks during flood. The sudden loss of velocity, and consequent carrying power, of river water upon leaving its channel during flood results in deposition of thick, coarse silty materials nearest the river and successively finer grained and thinner deposits at a greater distance. The natural levees of the Mississippi River are approximately 10 feet in height, but near New Orleans and out onto the delta they are lower and extend to varying depths below the flood-plain surface, often more so below than above, as the load of these natural levee deposits has caused depression of soft underlying materials. Levee widths are extremely variable, from 14 mile or so to as much as 4 miles. Normally they are widest behind concave banks and are also likely to widen downstream from a bend. Surrounding False River (Fig. 3), distinct natural levees can be distinguished by their utilization for agricultural purposes. These form the firm land in this area and have the added advantage of being slightly above normal flood level. Crevasses. In areas directly behind natural levees there are fre- quent topographic irregularities caused by crevasses developed during floods. These features are almost like river distributaries, but they are smaller and much more temporary, serving only during very high water. Deposits laid down in this manner are generally thin but have many irregularities resulting from all sorts of braided-channel pat- terns. Silts of crevasse origin may form thin veneers over troublesome plastic clays and prove misleading in engineering work unless very careful borings are made. Scouring action by sudden crevassing some- times produces limited basins known as “blue holes,” which are filled with clays and fine silts after their channel has been abandoned. Crevasses are recognizable in the natural levee south of the point bar of Fig. 3. Here the channels can be seen to have developed small scour holes and to terminate with distributary courses. The crevasse channels in this area have become wooded, and their adjacent silty deposits have been utilized for farming. Channel-fill deposits. With abandonment of meander loops by de- velopment of natural cutoffs, the familiar oxbow lakes are formed. These, however, are only temporary features, as tributary creeks and river floods tend to fill them slowly. In time these crescent-shaped scars become filled chiefly with fine clayey sediments, so that they have earned the name “clay plugs.” Generally, in the basal part of a section, the clay is heavy and plastic. Above this, the material is more silty and less plastic, and, near the ends of the crescent, the mate- rial may be sandy or composed of coarse silts. The upper arm of a crescent generally accumulates the coarsest sediment, whereas the lower Cu. 12] MEANDER-BELT SEDIMENTS 217 arm fills more slowly and with finer material. Maximum thicknesses of these plugs generally occur in the bendways where 75 feet or so of sediment is an average accumulation in old Mississippi River scars. False River of Fig. 3 is such an oxbow lake of an abandoned meander FALSE RIVER CUT-OFF 1722. INT, BIAg E PIONS! 4D SWALES) EIN D Low t i > { < “) x U < 0 SCALE S000. 10000 ISHo0 FT : | REAR Fic. 3. Aerial view of typical meander-belt deposits and adjacent backswamp, False River, Louisiana. loop and is at present slowly filling with fine-grained deposits. From the illustration it can be seen that the crescent points have already become filled and that sediment continues to be brought in by small tributary creeks. Backswamps. Adjacent to the meander belts in the Lower Missis- sippi Valley are restricted low-lying areas known as backswamps, which receive considerable fine sediment during flood inundations. 218 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 The deposits are mainly very thinly laminated clays and silts which may contain organic matter, especially in southern portions of the valley. In addition, deltaic backswamp deposits contain much fresh organic material such as wood and roots of trees (Russell, 1933). The sediments are nearly always calcareous, as soluble materials tend to be concentrated in these basins. They vary in thickness from less than 20 feet in the latitude of Memphis to over 70 feet southward in Louisiana, where they merge into the deltaic deposits along an in- distinct line. Adjacent to the natural levees of Fig. 3, backswamps can be recognized from their low-lying, wooded appearance and their lack of directional drainage. | INFLUENCES OF FLOoD-PLAIN DEPosITs The flood plain in the Lower Mississippi Valley covers approximately 35,000 square miles, a value slightly greater than the area of Louisi- ana. In addition there is a non-flood-plain portion of the Recent al- luvium which consists of levels between the lowest terrace and the present flood plain. These surfaces are low fans which extend from valley entrances of tributary streams and are being dissected under present conditions. Areal extents of these fans are pictured in Fig. 1. Within its flood plain, the Mississippi River has distinct regional variations of behavior which are related to the sediments through which it flows. Southward from Cairo, Illinois, the river flows in a region of coarse deposits where its channel is chiefly in sands and silts. Un- der these circumstances the river tends to develop wide, shallow chan- nels and meanders extensively. Below Helena, Arkansas, sands and silts continue to predominate but with the influence of varied meander- belt deposits. Farther southward in the area of deep, fine-grained deltaic sediments, which are resistant to scouring action, channels tend to be narrower, deeper, and more stable. In general, as a result of regional sediment changes, the Mississippi River may be designated according to the following behavior units: Commerce, Missouri, to Helena, Arkansas. The river is broad and shallow and has many towheads and a high percentage of reaches. Channels shift rapidly. Helena, Arkansas, to Angola, Louisiana. Here there are deeper, narrower channels with fewer reaches and fewer towheads, although meandering still takes place rapidly. In this region the channels are complicated through influences of excessively variegated meander-belt sediments. Angola, Lowisiana, to the gulf. The channel is deep and narrow but with much less meandering. False River, Louisiana, is the southern- Cu. 12] DESCRIPTION OF ENGINEERING STRUCTURES 219 most cutoff, and Thompson Creek, Louisiana, is the southernmost point at which gravels are dumped into the Mississippi River. Below Baton Rouge, Louisiana, meanders and point bars form very slowly, whereas south of Donaldsonville, Louisiana, the channel is almost sta- tionary. Here the sediments are composed almost entirely of clay and fine silt which, owing to their cohesiveness, are extremely diffi- cult for the river to erode. Meanders in the modern delta are almost non-existent, with the few bends present showing very little change from the earliest surveys to the most recent. In meander-belt areas, the variety of sediments already discussed exert many material influences on river behavior where they are cut into by encroaching channels. The effects are generally directly related to grain size and plasticity of sediments and extent of in- dividual deposits. The following basic subdivisions of the alluvial fill materials may be made. Coarse-grained soils. SANDS AND GRAVELS. These materials wear away easily through the process of surface sloughing. Little resistance is offered to river migration, and channels are wider and shallower than those in finer grained soils. Fine-grained soils. Stuts (semi-plastic to non-plastic). These soils are more cohesive than sand and offer more resistance to meandering. The river is restrained to some extent by finer, more cohesive silts, so that some banks become comparatively stable. Cuays. Of all the flood-plain sediments, clays offer the maximum resistance to river erosion. Banks tend to be quite stable, and channels cutting into such deposits tend to be narrow, as evidenced by the Mississippi River in the southern portion of the valley. Since clay plugs exist in considerable numbers in the upper reaches of the valley and are of considerable individual size, they are often of great im- portance in halting river migration at various points. Integration of levee systems with existing clay plugs may serve to increase the use- fulness of the former. The knowledge of relative stability of river channels as determined from adjacent sediments is of valuable assistance in engineering plan- ning. ENGINEERING PROBLEMS DESCRIPTION OF ENGINEERING STRUCTURES Vast areas of land in the Alluvial Valley of the Mississippi River would be inundated were it not for the levees along the river and along some of its tributaries. When it is realized that the Mississippi River 220 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 is approximately 1,000 miles in length from Cairo to the gulf, and that levees lie on one or both sides of the river for the greater por- tion of this distance, it is apparent that the construction of levees is not only a major factor with respect to the economics and life of the large area within the Alluvial Valley but that also such a levee sys- tem presents some major engineering problems. Levees vary in height from relatively low structures up to heights rivaling many small earth dams, and levee heights in excess of 40 feet are not unknown. In addition to levees along the main river, flood-control structures on the tributaries are also required. In order to prevent flooding of the Alluvial Valley by tributaries of the river joining it below Cairo, dams have been and are being built to provide storage areas and thereby control flood stages. These dams include Sardis, Arkabutla, Enid, and Grenada in Mississippi; Blakely Mountain and Narrows in Arkansas; Ferrells Bridge and Texarkana in Texas; and other dams in Arkansas, Louisiana, and Texas. In order to provide for the requirements of navigation and local drainage, a large flood-control system must be supplemented by locks, floodgates, and miscellaneous drainage structures. All these structures are important to the proper integration of the overall system, and some present major design and construction problems, which will be discussed below. DESCRIPTION OF SoIL CONDITIONS The discussion of the geolcgical history of the Alluvial Valley has demonstrated that wide variations in soil types and states are to be expected as a result of sedimentation and fluvial action. These in turn give rise to radically different problems from the engineering view- point. The preponderance of sand in the northern portion of the Alluvial Valley indicates that, in general, the problems to be encoun- tered will involve underseepage and through seepage in levees. How- ever, even in this area, channel fillings, natural levee deposits, and other characteristic fine sediments deposited by the river give rise to the question of stability of the foundation and resistance to settlement. In the central portion of the Alluvial Valley the deep beds of sand and gravel are overlain by clayey and silty soils. In addition, the Mississippi River in this portion of the valley has meandered ex- tensively. Consequently, soil types and soil states are extremely vari- able and require an engineering and geological study to answer satis- factorily questions involved in the design of levees and structures, such as bearing capacity, settlement, seepage, etc. In the southern portion of the Alluvial Valley the soils, from the Cx. 12] LEVEE DESIGN AND CONSTRUCTION 221 surface down to a considerable depth, consist largely of clays and silty soils which in turn overlie deep beds of sand. The clay is highly plastic in some places. The Atterberg liquid limit generally is in ex- cess of 50, and values greater than 100 are not uncommon. The nat- ural water content of some of these fat clays ranges between 50 and 100 percent. It is evident therefore that primary engineering con- siderations in this region are stability and settlement. The loess deposits found along the east wall of the Lower Missis- sippi Alluvial Valley are unique in their physical properties and present many unusual problems when utilized as a foundation or construction material. Their marked tendency to lose shear strength when sub- ject to saturation or remolding and to erode easily when in an undis- turbed state introduces many problems. However, when properly compacted, they possess good shear strength and are quite resistant to erosion. The great difference in strength and erosion resistance of loess which has been properly compacted and loess which has received inadequate compaction at water contents substantially above or be- low optimum is particularly striking. Loess was used in the embank- ment of Arkabutla Dam with entirely satisfactory results. LivEE DESIGN AND CONSTRUCTION The wide variation in foundation soils and in borrow materials available gives rise to a multitude of problems. In the northern por- tion of the valley the sandy foundations are generally stable but are extremely pervious, and considerable underseepage that must be satis- factorily controlled may be present. In this region it is often difficult to find sufficient impervious material to cut off satisfactorily seepage through the levees, and to provide an impervious blanket on the river- side face. In addition, levees must often cross abandoned and filled stream channels, which filling may consist largely of fat clays and silts deposited in water and never exposed to any appreciable amount of drying or to any overburden load. In some cases these materials are not even consolidated under their own weight. Foundation stability becomes a major feature in these locations, requiring the taking of undisturbed samples that can be tested in the soils laboratory. Problems involved in levee design and construction in the central portion of the valley are somewhat similar to those in the northern portion except that, because the top stratum in many places is thicker and the river has meandered more in the past, difficult foundation problems involving stability are of more frequent occurrence. It is of importance not only to determine the locations of abandoned stream channels by geological investigation but also to determine the physical 222 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 properties of the materials, because the soils may be in a very weak state. For example, the shearing strength of clays as determined from unconfined compression tests may be as low as 0.2 ton per square foot or even less. Seepage beneath the levees is also of importance at many locations in the central portion of the valley, as in the northern portion, and a highly specialized technique has been developed for controlling sand boils which develop during flood stages of the river. Much engineering talent has been devoted to these techniques, which involve the proper method of bagging sand boils and the construction of sublevees, protection of caving banks, and other related problems. Fighting a flood brings as many and as difficult engineering problems as does the problem of design and construction of a levee system. Since the soils in the southern portion of the valley are generally clays to considerable depths, levee design and construction problems in this region are more concerned with foundation stability than with underseepage, although there are some notable exceptions. In this region the required heights of levees are less than in the central and northern portions; this is fortunate because the strength of the founda- tion soil, in general, also is less. Shear strengths below 0.2 ton per square foot are by no means uncommon, and it is evident that founda- tion stability becomes almost an insuperable problem in some loca- tions. These problems have been solved successfully in the past through determination of the properties of the soils by laboratory tests and the design of berms on both landside and riverside of the levee to give greater stability. In some cases construction has been prosecuted in stages which continue over intervals of years so as to bring the levee to grade gradually, thereby allowing the soils to con- solidate and gain strength. Occasionally, levees have been built by the displacement method; that is, construction is continued in a normal manner and the weight of the embankment squeezes out the unstable material in the foundation. As much as 200 percent of the material composing the normal levee section has been required in some cases before stability was reached and the levee finished to grade. REVETMENTS The meandering of the river frequently endangers the existence of the levees. In some instances this necessitates construction of a set- back levee, but in places this cannot be accomplished, owing to the location of a town or to the relative economy of bank protection in place of levee construction. The design and construction of revetment involves construction, soils, and hydraulic problems. Before a revet- ment is constructed, the engineering problems involved concern the Cu. 12] REVETMENTS 223 alignment of the river and related considerations. The revetment con- stitutes a protective coating laid on the bank to eliminate or reduce loss of material due to the river’s attacking the bank. Revetments have been quite successful in this respect, but it is apparent that a revetment in itself has no structural strength and will be carried away if the bank on which it is laid is unstable. Thus, important engineering problems are involved in the determination of the stability of a bank when subject to scour and seepage forces. The determination of a safe slope to which the bank can be graded is important, for the cost of the work involved is greatly dependent on this slope. Construction prob- lems increase greatly if the bank becomes flat, and slopes flatter than 1 on 3 or 1 on 3.5 are difficult to revet, from the construction view- point, and very expensive, from the economic viewpoint. Conse- quently, low factors of safety must often be used in determining a sat- isfactory slope to which the bank may be graded for placement of revetment. As a matter of interest, the Mississippi River Commission has under investigation at the Waterways Experiment Station a com- bined hydraulics and soils investigation of factors affecting the loca- tion, design, and construction of revetments. This includes an an- alysis of reasons for failures which may be due both to the large hydraulic forces involved and to the physical properties of the mate- rials, such as shear strength, resistance to seepage forces, structure, density of sand deposits, and susceptibility to flow slides. In the northern and central portions of the valley, revetments are important in protecting sandy river banks against scour. Seepage forces due to the natural draining of water out of the banks or to rapid fall in river stages become important, especially if the soils are not free-draining. Many drainage methods have been investigated by the Mississippi River Commission, and horizontal drain wells are now being considered. In the central portion of the valley, banks often consist of a top stratum of silts and clays 20 to 40 feet thick overlying deep beds of highly pervious sands and gravels which are readily attacked by action of the river, whereas the top stratum is more resistant. However, the top stratum may have a low shearing strength; then the problem be- comes complex, because the revetment must be laid on a bank, which must be graded to a slope that is stable, and must protect under- lying sandy sediments which are easily scoured out. In the southern portion of the Alluvial Valley the surface clays ex- tend to such depth that protection of silty sediments against attack by the river becomes of less importance. As a result of the resistance of these clays to erosion, the tendency of the river to meander is much 224 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cx. 12 less than in the central and northern portions of the valley. However, a much greater concentration of industrial facilities along its banks has taken place; this in turn causes minor tendencies to meander and local stretches of bank instability to become major problems. Wave wash does become of considerable importance in revetments in this area, as is readily apparent, because ocean-going vessels regularly travel as far upstream as Baton Rouge. EXCAVATIONS The making of excavations of any considerable depth in the Al- luvial Valley presents varied engineering problems, as these excava- tions may be made in very permeable sands or in weak, fat clays of low shear strength. Furthermore the bottom of an excavation may be located in impervious clays that overlie highly pervious sands of depths that may exceed 100 feet and that connect directly with the Mississippi River. Thus wellpoint systems are often necessary in the northern and central portions of the valley to lower the ground-water level in the permeable sands and to prevent loss of stability of banks and foundations due to seepage forces. The high ground-water level usu- ally found throughout the Alluvial Valley results in high heads which must be provided for. Excavation in clays where shear strengths may be less than 0.2 ton per square foot makes necessary an engineering study to determine the allowable slope to which the bank may be cut. Undisturbed samples of soil are regularly obtained and tested in the laboratory for large excavations. Slope analyses are made by the circular-are method or the method of sliding wedges, depending on which method may be most applicable. The presence of a pervious formation beneath the clay often makes it necessary to install wellpoints or deep wells to relieve uplift pres- sure against the bottom of the excavation, even though the excavation is in clay out of which a negligible amount of water flows into the ex- cavated area. Since the sand and gravel formation beneath the surface in the Alluvial Valley is so highly pervious, even in the southern por- tion, and connects with the Mississippi River, the possibility of exces- sive uplift must be taken into account, even though the river may be many miles away. An example of such variation in head in the per- vious sand and gravel formation in the southern portion of the valley near Baton Rouge is shown on Fig. 4, which shows the level in piezo- meters corresponding to variations in stage in the Mississippi River, which at this point is 10 miles away. The effect of these uplift pres- sures must be considered in planning the excavation, and methods must often be provided for their control and observation during con- EXCAVATIONS 225 ‘Oreiy 1A ‘ssuIpvel Jajawuozatd deep pue Jeary iddississtpy jo yduisoipAy “p ‘DIT Jaquiadeq JaqWiaAON 49q0}9Q iaquwajydes ysn3ny Aine aunr Kein judy yosew Asensiqay Alenuer GZOZSTOLS SZ0ZSTOTS ScOZSTOIS Sz0zSTOIS Sz0zSTOTS Sz0zSTOIS SzOZSTOIS Gz0ZSTOIS GzOZzSTOIS SzOzSTOIS Sz0zsTOIS Sz0zstor s Ls _, PENLEL. aieeen eS aS eN Jany Iddississi BOCs Jett oaned auiuanbej LEE rap. mc Besessse 4a ~) ~ JAA Iddtssissi|\y asnes aulwanbel4 Secseaeanenm “pee secs Beaas JSR Sa 2 Se es TOW 3224 Ul UOeNaly 226 TURNBULL ET AL. RIVER ALLUVIUM PROBLEMS [Cu. 12 struction. A particularly interesting case which involved the com- bination of excavation in very weak clays underlain by sand stratum is afforded by the construction of the Algiers Lock across the river from New Orleans, which is now under construction by the New Orleans District, CE. This situation was further complicated by the fact that gas was present in the sand stratum, and it was necessary that the gas pressure be relieved. The New Orleans District accomplished this by installation of pressure-relief wells which permitted the gas and excess hydrostatic pressures to be dissipated. STRUCTURES The many floodgates, locks, and other drainage structures which are required give rise to many foundation problems which require geo- logical and engineering studies for the economical and _ successful solution of design features. The geological study often results in re- location of the structure to avoid recent channel fillings and other undesirable foundation conditions. The engineering study is concerned with obtaining undisturbed samples of soil where required, testing samples in the laboratory, and evaluating such questions as allowable bearing capacity if spread footing is used, or the length of piles neces- sary if a pile foundation is required, estimate of settlement of the structure, and provisions for handling seepage beneath the structure, which often acquires considerable importance. Problems involved in making excavations for structures have been described in the pre- ceding section. REFERENCES Daly, R. (1929). Swinging sea level of the ice age: Bull. Geol. Soc. Amer., vol. 40, No. 4, pp. 721-734. (A pioneer project.) Fisk, H. N. (1944). Geological investigation of the Alluvial Valley of the lower Mississippi River: Mississippi River Commission, Vicksburg, Miss., 78 pages. (1947). Fine-grained alluvial deposits and their effects on Mussissippr River activity: Waterways Experiment Station, Vicksburg, Miss., 2 vols., 82 pages. Russell, R. J. (1933). Mississippi River delta: Louisiana Geol. Survey, Bull. 8, 454 pages. Part 3 APPLICATIONS OF PROCESSES OF SEDIMENTATION a) CHAPTER 13 RELATION OF LANDSLIDES TO SEDIMENTARY FEATURES * D. J. VARNES Geologist, U. S. Geological Survey Denver Federal Center Denver, Colorado Landslides have been the subject of considerable study for many years, and in many different countries, not only because they are of practical economic importance, but also because they are wide- spread, effective, and interesting agents in the shaping of land forms. There is now available a vast volume of observations by construc- tion engineers, geologists, and interested laymen covering the many diverse phases of landslides and other kinds of earth movement. Most striking is the growing literature on the application of soil mechanics to the stability analysis of certain types of landslides. The subject of landslides now has many aspects that cannot be covered in a short chapter. The discussion to follow, therefore, will be principally a description of those features of sedimentary deposits which are recog- nized as contributing to true landslides, or which must be considered in their prevention, analysis, or cure. Some of the features described are not limited to sedimentary deposits; likewise many of the dynamic processes operate as well in metamorphic or igneous terrains. To acquaint the reader with the terminology to be used, and with the typical form of the several kinds of landslides, C. F. S. Sharpe’s classification and one of his illustrations, taken from Landslides and Related Phenomena (1938), are reproduced in Figs. 1 and 2. Readers familiar with this excellent work will recognize that much of what follows is an enlargement on his “active” and “passive” causes as applied to sedimentary deposits. For the present purpose, the role of the major factors of (1) physical and chemical composition, (2) structure of the deposit, and (3) state of stress in the material in the production of some of the types of slides illustrated in Fig. 2 will be shown. * Published by permission of the Director, U. S. Geological Survey. 229 a Seat eT 4 CHIEFLY EARTH OR ROCK SE Oe Ory EARTH OR ROCK ICE PLUS ICE OF ICE OR WATER PLUS WATER Fig. 1. DEBRIS AVALANCHE DEBRIS AVALANCHE za 2 ke <= = oO (eo) oO. 22) Zz <= oO e — < > = — Le GLACIAL TRANSPORTATION Classification of landslides and related phenomena. (After Sharpe, 1938, by permission of the author and Columbia University Press.) NAS IS SS (2) Ber al. The five classes of landslides. author and Columbia University Press.) 230 Cu. 13] STRUCTURE 231 MAJOR FACTORS The major factors of composition, structure, and state of stress de- termine the equilibrium of an earth mass; the first two factors operate indirectly, and the last directly. These factors set the stage for slid- ing, and any change in one may itself disturb equilibrium or produce changes in other factors to lead to instability. For instance, a change in composition, such as increased water content, may change both the internal structure and the state of stress and lead to sliding. The variables comprising or influencing the major factors are commonly so interrelated that a discussion of their separate effects is impractical; therefore, they will be only briefly listed separately and their joint effects considered in more detail, with a number of examples. CoMPOSITION Any sedimentary deposit consists generally of three phases, solid, liquid, and gas. Important factors in the physical composition of the solid phase are the size, size distribution, shape, area, and nature of surface of the individual rock or mineral particles; the amount and kind of cement; and the mechanical strength of the particles and the cement. Some of these characteristics may be permanent, some are variable. To be considered in the chemical character of the solid phase are mineralogy, chemical stability, and surface effects such as hydration and base exchange. The relative abundance of the liquid phase is critical in determining the mechanical stability of fine-grained sedimentary deposits. The liquid, of course, is water. Many of the physical properties of sedi- ments are determined by the peculiar properties of water and the state in which it is held; that is, whether or not it has a free surface, or is held in capillaries, or bound into a semi-solid state around the mineral grains, or actually enters into the crystal lattice of certain clay minerals. Air or other gases within an unconsolidated deposit may affect its physical and chemical properties. STRUCTURE For convenience, structure may be classed as either gross or fine. Gross structure includes stratigraphic sequence, attitude, homogeneity, and discontinuities such as bedding planes, joints, and faults. Fine structure includes the arrangement of individual particles into loose or compact structures, resulting from modes of disposition, or the interaction of surface forces of the solid and liquid phases. 232 vARNES. LANDSLIDES [Cu. 13 STATE OF STRESS The state of stress determines directly whether or not an earth mass is stable. Gravity is, of course, the prime mover and is occasionally aided by other forces arising within the material. The state of stress is determined largely by topographic form, but is also influenced by changes in composition and structure as they affect the distribution of stress among the solid particles, the interstitial water, and en- trapped air. OPERATION OF MAJOR FACTORS TEXTURE The physical and physicochemical characteristics of sedimentary material and the relative abundance of water are of prime importance in the determination of stability. The size of individual particles in a sedimentary deposit in itself does not greatly affect its stability, unless the size is of fine sand or smaller, where capillary forces, surface effects, or the peculiar hydra- tion properties of clay minerals come into play. The effect of size in uniform conglomerate, gravel, coarse sandstone, and sand is primarily to produce high transmissibility, allowing water to migrate readily— possibly to adjacent shales where it may do real damage. The Tel- luride (San Miguel) conglomerate is believed by Cross (1899) to have so contributed to large landslides in the San Juan Mountains of Colo- rado. Coarse gravels capping shale mesas have been instrumental in guiding irrigation waters to slides along Cedar Creek, near Montrose, Colorado (Varnes, 1949). In even-grained deposits the permeability increases as the size increases (Fraser, 1935). In graded deposits, the filling of voids with fine material affects the permeability and may affect other physical properties of the sedi- ment. In dry cohesionless sands, size distribution apparently. has little effect on compressibility (Chen, 1948), but in a cohesive clay soil, addition of gravel has been found to decrease compressibility (Murdock, 1948). The shape of the constituent particles also influences the physical properties of soils and sedimentary deposits. In dry material, both the compressibility and the angle of internal friction (and hence shearing resistance) increase with increasing angularity of the con- stituent grains (Chen, 1948). The shape of the particles also influ- ences the fine structure developed during the process of sedimenta- Cu. 13] MINERAL CONTENT 233 tion and consolidation, and thus affects the density and shearing strength of the material. Laboratory tests show that, for sands of high roundness and sphericity, the density and porosity are quite sensitive to the rate of deposition (Kolbuszewski, 1948), slow deposi- tion producing the denser deposits. The shape of mineral particles may also affect stability through preferred orientation. The writer has seen numerous rockslides along bedding planes in dipping sand- stone near Glenwood Springs, Colorado. The bedding planes along which slippage occurred were covered with mica flakes lying in the plane of the bedding. MINERAL CONTENT Glauconite, a fairly common mineral in sedimentary rocks, may contribute to landsliding. An interesting example of a slide due to glauconite in the cliffs around Algiers Bay was described by Proix- Noé (1946) and by Drouhin, Gautier, and Dervieux (1948). The cliffs around the bay are composed of glauconitic marl underlying sandy limestone. Water percolating down through the limestone became charged with calcium carbonate. As the water entered the marl and contacted the glauconite, the calctum ions were fixed and potassium ions liberated The water became markedly more alkaline (pH 9), resulting in the deflocculation of the marl and the hydrolyza- tion of alumino-silicates. The marl was made highly fluid and could no longer support the overlying beds, so the cliffs collapsed. Glau- conitic mudstones and greensands have been involved in landslides in New Zealand (Benson, 1946) and along the seacoast of England near Folkestone (Toms, 1948) although no reference was made in either description to possible physicochemical action of the glauconite itself. Gypsum also may contribute to landsliding. It is moderately soluble, and its removal by circulating waters within a series of sedi- mentary strata may cause subsidence, or sliding if a steep face is ex- posed. It also has two other effects of quite different natures. The dissolving of gypsum disseminated through a shale, and its subse- quent recrystallization along minute fractures close to the ground surface, are potent factors in breaking up and rendering permeable an otherwise rather massive and impermeable clay or shale. This process has destroyed effectively a grout surfacing placed on Mancos shale slopes at Mesa Verde National Park to prevent the shales from being wetted by rain. At that place, water could get at the shale behind the roadcut from an overlying permeable and fractured sandstone (Varnes, 1949). Even in flat ground, the crystallization of gypsum 234 vaRNES. LANDSLIDES [Cx. 18 may cause sufficient swelling to damage structures (Joly and Ninck, 1935). Gypsum or, more properly, the calcium ion also contributed to landsliding in clays along the banks of the Volga River (Tchouri- nov, 1945). This reference to base exchange in clay leads to the subject of the clay mineral group, which as a factor in composition ranks in im- portance with water in the determination of stability or instability of slopes. Clayey soils and sedimentary deposits are widely used as foundation and construction materials; they are involved in land- slides throughout the world, and they pose problems of maintenance in countless railroad and highway cuts. The needs of engineering, economic geology, ceramic technology, and many other fields have resulted in intensive research on the physics and chemistry of clays, but only the briefest mention of a few of the results of this research can be made here. For further information on clays, the reader’s attention is directed to the many papers and their bibliographies in the First and Second International Conferences on Soil Mechanics and Foundation Engi- neering (at Harvard in 1936, and at Rotterdam in 1948); to papers in Soil Science of America; to texts on soil mechanics such as those by Terzaghi (19438) and Terzaghi and Peck (1948); and to articles by Grim (1942), Casagrande (1932), Terzaghi (1941), Hendricks (1942), Rutledge (1944), and Winterkorn (1942). PHYSICAL PROPERTIES The physical properties of earth materials that contain fine-size fractions depend very largely on the character of the smallest-size particles, especially those particles of less than 2 microns. The prop- erties of these small particles are influenced by the intensity of the negative electric charge on their surface. Because water molecules are polar, the positive (hydrogen) ends of the water molecules are attracted to the solid particles, where they orient into an adsorbed layer. Near the mineral particle, the film is practically solid, farther away the film is viscous, and near the outer surface the water is liquid. If little water is present, the rigid, adsorbed layer on one particle meets that on another, and the whole mass is rigid or highly cohesive. With more water, the films thicken and lubricate the par- ticles, making the mass plastic. With still more water the mass be- comes fluid. The thickness of the water films, their degree of orienta- tion, and their efficiency as binding or lubricating agents depend on the amount of water available, the charge on the mineral surface, and the surface area available among the mineral fragments. The clay Cu. 13] WATER CONTENT 235 minerals proper, through their habit of cleaving into minute flakes, present tremendous surface area. Moreover the arrangement of mole- cules on the crystal surfaces of certain clay minerals is favorable to the adsorption of water or to the adsorption of certain ions which themselves may attract thick water films. Montmorillonitic clays in which the exchangeable ion is Na* adsorb thick oriented films; they require somewhat more water to become plastic, and 5 or 6 times as much water to become liquid as do clays carrying exchangeable Cat + or H+. Thus they remain plastic over a considerable range of water content. Some clays also adsorb organic molecules and form gels in suitable organic liquids. Electron-microscope photographs show that clay particles tend to cluster around bacteria and bits of organic mat- ter (see Jackson, Mackie, and Pennington, 1946, and its bibliography). WATER CONTENT In fine-grain sedimentary deposits, the water films around clay par- ticles profoundly influence the forces holding the particles together, and also influence the arrangement of both the clay size and larger particles into structures that affect the porosity, shear strength, con- solidation properties, and permeability of the deposit. All these fac- tors in turn depend ultimately on the amount, kind, and size distribu- tion of the clay and non-clay size fractions, the kind and concentra- tion of exchangeable ions and soluble.salts, the type and amount of organic material present, the mechanics of deposition, the past loading history of the deposit, and the past history of wetting and drying cycles. Tchourinov (1945) has explained the slides in Cretaceous clays along the banks of the Volga River somewhat as follows. During the process of weathering, pyrite in the clay decomposed, calcium carbonate in the clay was attacked, magnesium ions increased through the decomposition of glauconite, and the ratio of sodium to calcium and magnesium in the clay decreased markedly. Presumably through base exchange involving the calcium or magnesium ions or both, the thickness of bound-water films decreased, and the clay assumed a granular structure leading to increased porosity, permeability, and free water content. Cohesion and shearing strength decreased and sliding occurred. Only the weathered clay was involved in movement. When unfissured natural earth materials are mechanically mixed and remolded at constant water content or even disturbed by vibra- tion, they commonly lose much of their shear strength. In clayey materials the mechanics of the change is imperfectly understood, but it is thought that loss of strength is due to a breaking down of more 236 varngES. LANDSLIDES [Cu. 13 or less loose structures in which the larger particles are cemented to- gether or supported by the fine clay-water particles, or by the shearing off of the water films which bind the clay particles together. Upon resting, even at the previous water content, disturbed clays may re- gain a considerable portion of their strength, probably through the reconstruction of the water films (Moretto, 1948). This sensitivity of fine-grained deposits to disturbance makes dif- ficult the control of those slides in which slump blocks turn rapidly into semi-liquid earth flows as the material becomes disturbed (see Hig. 2a)? VIBRATIONS Earth flows may be initiated by vibration. In describing earth flows that accompanied the San Francisco earthquake, Anderson (1908) says, “In certain cases the water seems to have risen with a gush, as if actually squeezed from the hills.” The author suspects that the vibrations caused the natural structure of the soil to break down and the soil to consolidate. Some of the water held tightly in small voids and bound around the mineral particles was freed, and it locally oversaturated the soil to cause flowage. A landslide along the banks of the Gerzenzee in Switzerland (von Moos and Rutsch, 1944) was due to the breaking up of the structure of a marl bed by the felling of trees and blasting of stumps. Here the marl bed beneath a cover of peat had remained stable, even though it was highly saturated, because of a gel structure involving organic material. When disturbed, the mass liquefied and flowed out into the lake, and the shore subsided. Fine-grained deposits of clay or silt size, but consisting of rela- tively fresh rock flour rather than clay minerals, are particularly susceptible to disturbance. Such deposits are common in lakes that have received glacial detritus and include varved “clays.” The water content of this material in its natural state commonly equals or ex- ceeds its liquid limit, and although it may be trimmed back to stable slopes, it has little cohesion and cannot be handled without loss of strength (Legget and Peckover, 1948; Tschebotarioff and Bayliss, 1948). Fine sand, if it is saturated and below its maximum density, may flow readily if support is removed. Peck and Kaun (1948) describe the liquefaction of sand deposited by a dredge behind sheet piling when two of the sheet piles were removed. Progressive flow slides in the Dutch province of Zeeland in sands below the critical density Cu. 13] PERMEABILITY 237 exposed to tidal scour have been described (Koppejan, Van Wamelen, and Weinberg, 1948). Gross STRUCTURE Clay minerology and the structure of fine-grained materials play a prominent role in the stability of poorly consolidated materials, but the previous discussion may have overemphasized these particular points because of the interesting advances in current research. Homogeneous deposits of clay or silt are uncommon, and it is gen- erally necessary to take into account many other factors that bear on the arrangement and grosser structures of the deposit before a rational approach to analysis and cure of a slide can be made. In particular, the stratigraphic sequence, the attitude of the strata, and the presence of discontinuities such as joints, faults, and bedding planes exercise their own kind of control. The influence of these large- scale structures is generally apparent after short study of any par- ticular slide. Their action is well-known and has been often described; hence it will not be greatly detailed below. Flat-lying stratigraphic series in which massive beds are interstrati- fied with or underlain by shales or clays and exposed by erosion into steep slopes are commonly subject to landslides if the shales are wet- ted. Climate and the stratigraphic sequence of alternating massive sandstones and limestones with shales have produced erosion forms particularly susceptible to landslides in the Colorado Plateau (Hinds, 1938; Reiche, 1937; Strahler, 1940; Varnes, 1949). Massive voleanics interbedded with or overlying shales have been involved in many slides in the canyons of Idaho (Russell, 1901) and in the San Juan Moun- tains of Colorado (Cross, 1899; Cross and Spencer, 1900; Howe, 1909). PERMEABILITY Permeable beds or lenses within fine-grained deposits are especially dangerous, for they allow water to reach and lubricate large surfaces. A slide in interbedded shale and sandstone at Soldier Summit, Utah, in a cut of the Denver and Rio Grande Western Railroad was prob- ably aided by water from an underlying porous sandstone. Owing to dip and impermeable cover, the water in the sandstone bed was under considerable pressure at the site of the slide. If the permeable beds are also loose, the danger of internal erosion and loss of support is added to the possibility of failure by rotational shear slip. An inter- esting example of the influence of a permeable sand bed on a landslip on the south coast of England is described by Ward (1945, 1948). 238 vaRNES. LANDSLIDES [Cu. 13 INCLINED STRATA If the strata have a component of dip toward the free face of a valley wall, the tendency for slippage along incompetent strata is, of course, greatly increased. A large rockslide of this type occurred in the Gros Ventre River valley in Wyoming in 1925 (Alden, 1928). Slippage took place obliquely down dip along clay beds in a series of sandstones and limestones. The slide mass dammed the river and formed a large lake. Two years later the dam washed out, causing a disastrous flood downstream. FRACTURES Joints and fissures of various kinds not only weaken any mass of earth material, be it solid rock or clay, but also allow water to enter. The properties of stiff, fissured clays are quite different from homo- eeneous intact clays. Stiff, fissured clays are usually very compact, but they contain innumerable fractures that run through the mass. The common characteristic of stiff, fissured clays is that if a lump of such clay, with a moisture content below the plastic limit, is dropped, it breaks into polyhedric fragments with dull or shiny surfaces that may differ in color from the mass (Cassel, 1948; Terzaghi, 1936). Clays of this type have been consolidated under considerably higher pressure than their present surcharge and have been lifted and exposed by erosion. Whereas the vertical pressure has been reduced, the hori- zontal pressure, corresponding to the previous high vertical stress, re- mains in considerable part. The fissuring is believed to result from the differential variation of the vertical and horizontal pressures dur- ing the geologic history of the clay and from the ensuing shear stresses. As soon as the horizontal resistance is decreased through excavation, the clay expands and the fissures open. If water enters the fissures, a progressive softening occurs, gradually reducing the strength of the clay mass. The rate of softening depends on many factors. Skemp- ton (1948) describes slips in London clay that occurred from 5 to 40 years after the cuts were made. Landslides in hard rock commonly occur along, -or are at least in part controlled by, joint surfaces. Many such slides, generally of the rockslide or rockfall type, have been recorded.: A notable example is the Turtle Mountain slide at Frank, Alberta, in which approximately 41,000,000 cubic yards of limestone, weakened by joints across the beds, descended into the valley, just touched the outskirts of the town of Frank, and killed 70 people (Daly, Miller, and Rice, 1912). Faults may exercise control similar to joints, with the added factor Cu. 13] STRESS DISTRIBUTION 239 of possible clayey gouge or loose breccia along the fault surface to aid movement. Wide fault zones, in which the material is highly brecci- ated, are always potential sources of trouble in deep excavations of any kind. Even geologic contacts between different types of rocks may act as sliding surfaces. Sharp (1931) describes a large slide in Wyo- ming in which Tertiary sand and gravel apparently slid along the sloping contact with underlying crystalline pre-Cambrian rocks. In short, the effect of discontinuities of any kind must be evaluated in the analysis of the stability of slopes. This is especially true if the fractures are so disposed as to create an unfavorable distribution of stress within the deposit. Srress DisTRIBUTION Some general remarks on the distribution of stress within earth ma- terials are perhaps in order. In determining the factor of safety of a slope, one must compare two sets of forces: those that tend to produce failure and those that tend to prevent it. The force of gravity is cer- _ tainly the prime source of stress tending to produce failure, though it may be aided by forces arising from frost action or from the hydration of such minerals as gypsum or the clays. Gravity is a force that affects failure because internal friction on any surface within the mass depends on the effective normal stress on that surface. The resistance to sliding on any arbitrary surface in general consists of two parts, cohesion and friction, and is given by the well-known Coulomb formula: s=c+ptand¢ The meanings of c, p, and ¢ vary somewhat with the material in- volved and the loading conditions. In clay and uncemented sands, the cohesion c is zero; p is the normal pressure due to the weight of overlying material; and ¢, the angle of internal friction, is about the same as the angle of repose of the sand. If the sand is cemented, it is regarded as having a kind of cohesion. If it is moist, the sand has an apparent cohesion due to eapillary tension of the moisture films between the grains. This vanishes if the sand is immersed. In addi- tion, if the sand is immersed, the effective normal stress is equal to the total normal stress lessened by the buoyant effect of the water on the overlying material and by the neutral stress or pore-water pres- sure. If the pore water is in motion, seepage pressures must also be considered. The shear strength of saturated sands depends on many factors; among the most important are the original state of aggregation or 240 vARNES. LANDSLIDES [Cx. 13 denseness of the deposit and the ability of water to migrate as stress is applied. If the density of the material is so low that shear stresses from additional loading tend to decrease the volume of the material, and if the load is applied so fast that the water content cannot change, part or most of the vertical load is transferred to the water, the ef- fective intergrain pressure and friction are lowered, and failure may result. If the sand is dense, the volume will increase on shearing and the effective intergranular pressure will remain. Recent work by Geuze (1948) and others indicates that, if the original void ratio of a sand is above a critical value characteristic of that deposit, the sand will compact on shearing and, if drainage is not immediate, as pore- water pressures can develop. Clays consolidate slowly, and the difference between shear strengths of saturated clays determined experimentally by quick or slow tests is frequently quite large. Some of the difference is due to the building up of pore pressures in the quick test, but the problem is still complex and subject to much current research. It has been found that, for soft clays, a satisfactory approximation to the shear strength of the ma- terial in place is given by: s = 4 unconfined compression strength The presence of gas bubbles within a fine-grained deposit decreases its permeability somewhat and, hence, the rate of drainage, but it accelerates the rate of consolidation. Also, if clay contains air, the shearing strength derived from a quick test increases with increasing normal stress (Terzaghi and Peck, 1948, p. 89). It is not the intention of the author to go further into the soil me- chanics of slopes than is necessary to indicate that the distribution of stress among the three phases of a natural deposit is critical to the stability of a slope. From the above discussion it 1s apparent that shear strength of earth materials is a complex function of the relative density of the deposit, its water content, its permeability, and the strength of the individual particles. For natural sands, the relative density depends largely on the conditions and rate at which it was deposited; the relative density of clays depends chiefly on the loads that the soils have carried and, in some instances, the rate at which the loads were applied. The permeability of a deposit is, of course, a function of many vari- ables of the material, including particle size, size distribution, miner- alogy, and structure. The development of pore pressures and the rate of consolidation depend in large degree on the permeability. Cu. 13] METHODS OF COMBATING SLIDES 241 Pore-water pressure has been studied principally in regard to the stability of earth dams. Several failures of earth dams, notably those of the Alexander Dam in Hawaii (Anonymous, 1930) and the Belle Fourche Dam (Anonymous, 1933) have been ascribed to the develop- ment of excess pore-water pressures through lack of drainage. Pore pressures developed in the walls of cuts or natural slopes sometimes produce very rapid and disastrous slides. Stratified deposits consisting of layers of sand and clay and masses of cohesive soil that contain irregular lenses or pockets of sand and silt are particularly susceptible (Terzaghi and Peck, 1948, p. 365). Glacial and glacial-outwash de- posits are commonly of this type. Beds or pockets of sand or silt be- come saturated during wet weather and may develop high pore-water pressure. These beds have little or no cohesion, and their shear strength is determined by (Terzaghi and Peck, 1948, p. 367) s = (p — U,,) tan As the pore-water pressure U,, increases, the shear strength de- creases until the porous bed can no longer support the overlying mass. Failure takes place by outward spreading and breaking up of the over- lying material on the underlying viscous bed. Such a failure by spreading may occur within a few minutes, in contrast to the slower action of slump in homogeneous clay. It is hardly necessary to comment on the natural agencies of erosion and the work of man in the production of slopes that may become unstable. It should perhaps be emphasized that the stability of a given slope is not a constant for all time. As erosion proceeds, as the ground-water conditions fluctuate, and as changes in the state of hydration and structure of clays occur with the passage of time, the state of stress throughout the slope may change greatly. The original slope design must take these possibilities into consideration. One must also consider the possibility of transitory stresses that arise from vibrations, blasting, earthquakes, or earth tides, triggering the failure of an otherwise stable slope. METHODS OF COMBATING SLIDES The methods employed in combating slides are generally curative rather than preventive. They depend on the kind of material, the area involved, and the type and rate of movement. These factors vary so greatly from slide to slide that the details of remedial measures are tailored to fit each slide. The design of high embankments and cuts, or the combating of 242 vARNES. LANDSLIDES [Cu. 13 large and serious slides, generally requires geologic mapping of both the surface and underground, the studies of ground-water levels and pressures, taking and testing of enough samples to be representative, and the analysis of the stability of the deposit. The purpose of a stability analysis is to aid in the design of safe slopes, to aid in the redesign of slopes that have failed, or to approxi- mate the original physical properties of material that has failed along a known surface. Assumptions are necessarily made regarding the gross and fine structure of the deposit and the shape of the sliding surfaces. The dangers of these assumptions are well recognized, but, in many instances, at least an approximate idea of the safety factor is obtained. It is generally assumed that sliding occurs along the are of a cylinder whose axis is parallel to the slope surface, and that the sliding block rotates about the center of the arc. The factor of safety is then defined as the ratio of the moments about the center of the are tending to produce movement and tending to resist movement (Terzaghi and Peck, 1948, p. 184): moment of resisting forces Safety factor = moment of driving forces Various refinements may be made, such as the assumption that the surface of sliding is composed of a series of ares of different radii. Many variables enter into the computation of the moments of the resisting and driving forces, and into the selection of centers and radii of the arcs. The factor of safety is computed for several sur- faces, and that having the lowest factor is the one along which sliding may most probably occur. Such analysis is most successful if the material is homogeneous and if data on the strength of the material and subsurface stresses are ample. Remedial measures for the cure of slides are generally aimed at changing one or more of the factors of composition, structure, or state of stress to restore equilibrium. Where slump or earth flow is involved, remedial measures include, if possible, changing the composition of the material by removal or diversion of water.. Methods of removal include open or rubble-filled trenches, drainage tunnels, wells, reverse filters, drying by heated air, and electro-osmosis. Methods used to prevent access include surface diversions, grouting openings, sodding, and oiling. Methods involving change of structure as well as composition, such as the introduction of gelling substances and stabilization by electric current, have been used in the stabilization of foundations, but the author is not familiar with their use in stabilizing slopes. Cu. 13] REFERENCES 243 Methods of changing stress distribution include drainage to relieve pore-water pressure, cutting back slopes and loading the toe of the slide to decrease the moment of forces tending to produce sliding, and the use of various types of barriers such as retaining walls, cribbing, and piles. Small rockslides in dipping strata have been successfully stopped by blocking or pinning loose slabs to the underlying firm rock (Laurence, 1948). Much has been written about the remedial measures applicable to landslides. The reader is referred especially to the works of Ladd (1935), Hennes (1936), and Forbes (1946); to the Proceedings of the First and Second International Conferences on Soil Mechanics and Foundation Engineering; to papers that appear occasionally in the Proceedings and Transactions of the American Society of Civil Engi- neers; and to many short articles in Engineering News-Record. FUTURE RESEARCH The cure for landslides is frequently difficult and costly, and all too often not completely satisfactory. Research in the problem of land- sliding should be directed not only at devising new and better methods of alleviating slides, but also increasingly toward developing criteria for the recognition of potentially unstable areas. Much has been done along this line in soil mechanics, especially in the criteria of stability for more or less homogeneous clay. Much remains yet to be done in the study of the natural fine structures of clays and silts, how they develop, how and why they change with time, loading, and water content, and the role of mineralogy and soluble salts on changes in structure. It has become increasingly clear that the stability of earth materials is a function of many factors, some of which are not easily recognized or evaluated. If the rank and file of engineers and geologists can be- come aware of the necessity for closer study of the slides they en- counter, and if soil engineers continue their researches at the present pace, perhaps sufficient data will become available to permit formula- tion of general principles of slope stability, even for non-uniform materials. REFERENCES Alden, W. C. (1928). Landslide and flood at Gros Ventre, Wyoming: Amer. Inst. Min. Met. Engrs., Tech. Pub. 140; Trans., vol. 76, pp. 347-361; abstract, Mining and Metallurgy, vol. 9, No. 262, p. 465, Oct. 1928. Anderson, Robert (1908). Earth flows at the time of the San Francisco earth- quake: Bull. Geol. Soc. Amer., vol. 18, p. 643. 244 vaRNES. LANDSLIDES [Cu. 13 Anonymous (1929, 1930). Alexander Dam: Eng. News-Record, vol. 103, No. 14, p. 517, 1929; vol. 104, No. 17, p. 703; vol. 104, No. 21, pp. 869-871, 1930. (1933). Slide in Belle Fourche dam follows reservoir drawdown: Eng. News-Record, vol. 111, p. 371, Sept. 28, 1933. Benson, W. N. (1946). Landslides and their relation to engineering in the Dunedin district, New Zealand: Econ. Geol., vol. 41, No. 4, Pt. 1, pp. 328-347. Casagrande, A. (1932). The structure of clay in its importance in foundation engineering: Jour. Boston Soc. Civ. Engrs., vol. 19, No. 4, pp. 168-209. Cassel, F. L. (1948). Slips in fissured clay: Proceedings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 2, pp. 46-50. Chen, Liang-Sheng (1948). An investigation of stress-strain and strength char- acteristics of cohesionless soils by triaxial compression tests: Proceedings of Second International Conference on Soil Mechanics and Foundation Engineer- ing, Rotterdam, vol. 5, p. 43. Cross, Whitman (1899). U.S. Geol. Survey Geol. Atlas, Telluride folio (No. 57), p. 10. , and Spencer, A. C. (1900). Geology of the Rico Mountains, Colorado, Chap. 5: U. 8. Geol. Survey, 21st Ann. Rept., Pt. 2, pp. 129-151. Daly, R. A., Miller, W. G., and Rice, G. S. (1912). Report of the commission appointed to investigate Turtle Mountain, Frank, Alberta: Canada Geol. Sur- vey, Mem. 27, 34 pages. Drouhin, G., Gautier, M., and Dervieux, F. (1948). Slide and subsidence of the hills of St. Raphael-Telemly: Proceedings of Second International Conference on Soil Mechanics and. Foundation Engineering, Rotterdam, vol. 5, pp. 104— 106. * Forbes, Hyde (1946). Landslide investigation and correction: Proc. Am. Soc. Civ. Engrs., Feb. 1946. Fraser, H. J. (1935). Experimental study of the porosity and permeability of clastic sediments: Jour. Geol., vol. 48, No. 8, pp. 910-1010. Geuze, E. C. W. A. (1948). Critical density of some Dutch sands: Proceedings of Second International Conference on Soil Mechanics and Foundation Engi- neering, Rotterdam, vol. 3, pp. 125-130. Grim, Ralph E. (1942). Modern concepts of clay minerals: Jour. Geol., vol. 50, No. 3, pp. 225-275. Hendricks, 8. B. (1942). Lattice structure of clay minerals and some properties of clays: Jour. Geol., vol. 50, No. 3, pp. 276-290. * Hennes, R. G. (1936). Analysis and control of landslides: Univ. Washington Eng. Exp. Sta. Series, Bull. 91, 57 pages. Hinds, N. E. A. (1938). Large landslides in the Colorado Plateau: Abstract, Proc. Geol. Soc. Amer. 1937, pp. 241-242. * Howe, Ernest (1909). Landslides in the San Juan Mountains, Colorado: U. S. Geol. Survey, Prof. Paper 67, 58 pages. Jackson, M. L., Mackie, W. Z., and Pennington, R. P. (1946). Electron micro- scope applications in soils research: Proc. Soil Science Soc. Amer., vol. 11, pp. 57-63. * Titles with asterisk deal with the general problem or contain good bibli- ographies. Cu. 13] REFERENCES 245 Joly, Henry, and Ninck, H. H. (1985). Un cas trés particulier de désordres causés par les modifications épigénétiques de roches liasiques (Schistes 4 posidonies du Taorcien de Lorraine): Congrés international des mines, de la metallurgie et de la geologie appliquée, vol. 2, pp. 529-532. (Translation by Severine Britt in open files of the U. S. Geological Survey.) Kolbuszewski, J. J. (1948). An experimental study of the maximum and mini- mum porosities of sands: Proceedings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 1, pp. 158-165. Koppejan, A. W., Van Wamelen, B. M., and Weinberg, L. J. H. (1948). Coastal flow slides in the Dutch province of Zeeland: Proceedings of Second Inter- national Conference on Sowl Mechanics and Foundation Engineering, Rotter- dam, vol. 5, pp. 89-96. * Ladd, George C. (1935). Landslides, subsidences and rock-falls: Amer. Ry. Eng. Assoc. Bull., vol. 37, No. 377, 72 pages. Laurence, R. A. (1948). Slides on Indian Creek road, Grainger County, Ten- nessee: In files of the U. S. Geological Survey, Sept. 29, 1948. Legget, R. F., and Peckover, F. L. (1948). Notes on some Canadian “silts”: Proceedings of Second International Conference on Soil Mechanics and Founda- tion Engineering, Rotterdam, vol. 3, pp. 96-100. Moretto, Oreste (1948). Effect of natural hardening on the unconfined compres- sion strength of remolded clays: Proceedings of Second International Confer- ence on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 1, pp. 137-144. f Murdock, L. J. (1948). Consolidation tests on soils containing stones: Proceed- ings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 1, pp. 169-173. Peck, R. B., and Kaun, W. V. (1948). Description of a flow slide in loose sand: Proceedings of Second International Conference on Soil Mechanics and Founda- tion Engineering, Rotterdam, vol. 2, pp. 31-33. Proix-Noé, Marthe (1946). Etude d’un glissenent de terrain di 4 la présence de glauconie: Comptes rendus hebdomadatres des séances de lacademie des sciences, vol. 222, No. 7, pp. 403-405. (Translation by Severine Britt in open files of the U. S. Geological Survey.) Reiche, Parry (1937). The Toreva-block—a distinctive landslide type: Jour. Geol., vol. 45, No. 5, pp. 538-548. Russell, I. C. (1901). Geology and water resources of Nez Perce County, Idaho; Part I: U. S. Geol. Survey, Water Supply Paper 53, pp. 75-79. Rutledge, P. C. (1944). Relation of undisturbed sampling to laboratory testing: Trans. Am. Soc. Civ. Engrs., vol. 109, pp. 1155-1216. Sharp, Henry 8S. (1931). A landslide scar in the Centennial Valley, Wyoming: Amer. Jour. Sci., 5th ser., vol. 21, pp. 453-457. * Sharpe, C. F. S. (1938). Landslides and related phenomena: Columbia Uni- versity Press, New York. Skempton, A. W. (1948). The rate of softening in stiff fissured clays, with special reference to London clay: Proceedings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 2, pp. 50-53. Strahler, Arthur N. (1940). Landslides of the Vermilion and Echo Cliffs, north- ern Arizona: Jour. Geomorphology, vol. 3, No. 4, pp. 285-301. Tchourinov, M. P. (1945). Modifications par altération dans la composition, la structure et les propriétés des argiles du Crétacé Inférieur: Comptes rendus (Doklady) de Vacadémie des sciences de VURSS, vol. 49, No. 5, pp. 364-368. 246 vaRNES. LANDSLIDES [Cu. 13 * Terzaghi, Karl (1936). Stability of slopes of natural clay: Proceedings of First International Conference on Soil Mechanics and Foundation Engineering, Har- vard University, vol. 1, pp. 161-165. (1941). Undisturbed clay samples and undisturbed clays: Jour. Boston Soc. Civ. Engrs., vol. 28, No. 3, pp. 211-281. (1943). Theoretical soil mechanics: John Wiley & Sons, New York, especially pp. 144-181. ES , and Peck, R. B. (1948). Soil mechanics in engineering practice, John Wiley & Sons, New York, especially pp. 9-17, 100-105, 181-191, 354-372. Toms, A. H. (1948). The present scope and possible future development of soil mechanics in British Railway civil engineering construction and maintenance: Proceedings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 4, p. 228. Tschebotarioff, Gregory, and Bayliss, John R. (1948). The determination of the shearing strength of varved clays and their sensitivity to remolding: Proceed- ings of Second International Conference on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 1, pp. 203-207. Varnes, Helen D. (1949). Landslide problems of southwestern Colorado: U. S. Geol. Survey, Cire. 31. von Moos, Armin, and Rutsch, Rolf F. (1944). Uber einen durch Gefiigest6rung verursachten Seeufereinbruch (Gerzensee, Kt. Berne): Eclogae Geologicae Helvetia, Zurich, vol. 37, No. 2, pp. 385-400. (Translation by Severine Britt in files of U. 8S. Geological Survey.) * Ward, W. H. (1945). The stability of natural slopes: Geographical Jour. Lon- don, vol. 105, No. 5-6, pp. 170-197. (1948). A coastal landslip: Proceedings of Second International Confer- ence on Soil Mechanics and Foundation Engineering, Rotterdam, vol. 2, pp. 33-38. Winterkorn, H. F. (1942). Applications of modern clay researches in construc- tion engineering: Jour. Geol., vol. 50, No. 3, pp. 291-306. CHAPTER 14 PERMAFROST * Rosert F. Buack Geologist, U. 8. Geological Survey Washington, D. C. Permafrost (perennially frozen ground) is a widespread geologic phenomenon whose importance and ramifications are rapidly becoming better known and more clearly understood. The problem is to under- stand permafrost in order to evaluate it in the light of any particular endeavor, whether practical or academic. To understand permafrost we need a precise, standardized terminology, a comprehensive classi- fication of forms, a systemization of available data, and coordina- tion of effort by geologists, engineers, physicists, botanists, climatolo- gists, and other scientists in broad research programs. These ob- jectives are only gradually being realized. This chapter is largely a compilation of or reference to available lit- erature. Its purpose is to acquaint geologists, engineers, and other sci- entists with some of the many ramifications and practical applications of permafrost. New data from unpublished manuscripts in the files of the U. S. Geological Survey also are included where appropriate for clarity or completeness. Inna V. Poiré of the U. 8S. Geological Survey has prepared numerous condensations of Russian papers on permafrost and made them available to the author. Others were made available through the National Military Establishment. The library of the Engineers School, the Engineer Center, Fort Belvoir, Virginia, has many abstracts, condensations, and translations of Russian works that are available to civilian readers. References in this chapter generally are only to later works, as most contain accounts of the earlier literature. The bulk of the literature, unfortunately, is in Russian and unavailable to the average reader, but some of it has been summarized by Muller (1945). The Arctic Institute of North Amer- * Published by permission of the Director, U. S. Geological Survey. 247 248 BLACK. PERMAFROST [Cu. 14 ica (Tremayne, 1948) is currently making an annotated bibliography of all arctic literature, including permafrost. A list of 190 titles of Russian articles dealing with permafrost is given by Weinberg (1940). The multitude of problems associated with frost action, as we refer to it in the United States, appropriately should accompany any dis- cussion of permafrost. However, limitation on space permits only a passing reference.to the relationship of permafrost to frost action. An annotated bibliography on frost action has been prepared by the Highway Research Board (1948). Thanks are due Lewis L. Ray, P. 8. Smith, Inna V. Poiré, Troy L. Péwé, David M. Hopkins, William 8. Benninghoff, Joel H. Swartz, and D. J. Cederstrom, of the U.S. Geological Survey, and to Stephen Taber and Kirk Bryan for critical reading of this manuscript. These and others in the Geological Survey have provided many valuable sug- gestions for which individual acknowledgment is difficult. The use of unpublished manuscripts and notes of P. 8. Smith and C. V. Theis is greatly appreciated. PERMAFROST The term permafrost was proposed and defined by Muller (1945). A longer but more correct English phrase is “perennially frozen ground” (Taber, 1943a). The difficulties of the present terminology are discussed by Bryan (1946a, b), who proposed a new set of terms. These are discussed by representative geologists and engineers (Bryan, 1948). Such terms as cryopedology, congeliturbation, congelifraction, and cryoplanation are being accepted by some geologists (Judson, 1949; Denny and Sticht, unpublished manuscript; Cailleux, 1948; Troll, 1948) in order to attempt standardization of the terms regard- ing perennially frozen ground and frost action. The term perma- frost has been widely adopted by agencies of the United States Gov- ernment, by private organizations, and by scientists and laymen alike. Its use is continued in this chapter because it is simple, euphonious, and easily understood by all. EXTENT Much of northern Asia and northern North America contains peren- nially frozen ground (Fig. 1) (Sumgin, 1947; Muller, 1945; Obruchev, 1945; Troll, 1944; Taber, 1943a; Cressey, 1939). The areal subdivision of permafrost into continuous, discontinuous, and sporadic bodies is already possible on a small scale for much of Cu. 14] EXTENT 249 Double hatching: Approximate extent of continuous permafrost. Ground temperature at a depth of 30 to 50 feet generally below —5° C. Diagonal hatching: Approximate extent of discontinuous permafrost. Ground tem- perature in permafrost at a depth of 30 to 50 feet generally between —5° and —1° C. Dotted diagonal hatching: Approximate extent of sporadic permafrost. Ground tem- perature in permafrost at a depth of 30 to 50 feet generally above —1°C. Reliability : Eurasia, good; Alaska, fair; all other, poor. (Eurasia after Sumgin and Petrovsky, 1940, courtesy of I. V. Poiré.) Fic. 1. Areal distribution of permafrost in the Northern Hemisphere. 250 BLACK. PERMAFROST [Cu. 14 Asia, but as yet for only part of North America. Refinements in delineations of these boundaries are being made each year. The south- ern margin of permafrost is known only approximately, and additional isolated bodies are being discovered as more detailed work is under- taken. The southern margin of permafrost has receded northward within the last century (Obruchev, 1946). Permafrost is absent or thin under some of the existing glaciers, and it may be absent in areas recently exhumed from ice cover. A greater extent of permafrost in the recent geologic past is known by inference from phenomena now found to be associated with permafrost (H. T. U. Smith, 1949b; Horberg, 1949; Richmond, 1949; Schafer, 1949; Cailleux, 1948; Poser, 1947a, b; Troll, 1947, 1944; Zeuner, 1945; Weinberger, 1944). Some of the more important phe- nomena are fossil ground-ice wedges, solifluction deposits, block fields and related features, involutions in the unconsolidated sediments, stone rings, stone stripes and related features, and asymmetric valleys (H. T. U. Smith, 1949b). The presence of permafrost in earlier geologic periods can be inferred from the cold climates accompanying many periods of glaciation and from fossil periglacial forms. In the Southern Hemisphere permafrost is extensive in Antarctica and probably occurs locally in some of the higher mountains else- where, but its actual extent is unknown. THICKNESS Permafrost attains its greatest known thickness of about 2,000 feet (620 meters) at Nordvik in northern Siberia (I. V. Poiré, oral com- munication). Werenskiold (1923) reports a thickness of 320 meters (1,050 feet) in the Sveagruvan coal mine in Lowe Sound, Spitzbergen. In Alaska its greatest known thickness is about 1,000 feet, south of Barrow. Generally it can be said that the frozen zone thins abruptly to the north under the Arctic Ocean. It breaks into discontinuous and sporadic bodies as it gradually thins to the south (Fig. 2) (Muller, 1945; Taber, 1943a; Cressey, 1939). In areas of comparable climatic conditions today, permafrost is much thinner in glaciated areas than in non-glaciated areas (Taber, 1943a). Unfrozen zones within perennially frozen ground are common near the surface (Muller, 1945) and are reported to occur at depth (Taber, 1943a; Cressey, 1939). They have been interpreted as indicators of climatic fluctuations (Muller, 1945; Cressey, 1939), or as permeable water-bearing horizons (Taber, 1948a). Cu. 14] CHARACTER 251 TEMPERATURE The temperature of perennially frozen ground below the depth of seasonal change (level of zero annual amplitude) (Muller, 1945) ranges from slightly less than 0° C. to about —12° C. (I. V. Poiré, oral communication). In Alaska the minimum temperature found to date is —9.6° C. at a depth of 100 to 200 feet in a well about 40 miles south- west of Barrow (J. H. Swartz, 1948, written communication). Repre- sentative temperature profiles in areas of (1) continuous permafrost are shown in Fig. 3a; of (2) discontinuous permafrost in Fig. 3b; and of (3) sporadic bodies of permafrost in Fig. 3c. Temperature gradients from the base of permafrost up to the depth of minimum temperature vary from place to place and from time to time. Measurement of four wells in northern Alaska resulted in gradi- ents between 120 and 215 feet per centigrade degree (data of J. H. Swartz, G. R. MacCarthy, and R. F. Black). The shape of a temperature curve indicates pergelation or depergela- tion (aggradation or degradation of permafrost) (Muller, 1945; Taber, 1943a). Some deep temperature profiles have been considered by Rus- sian workers to reflect climatic fluctuations in the recent geologic past. No known comprehensive mathematical approach has been attempted to interpret past climates from these profiles, although it seems feasible. Some of the effects of Pleistocene climatic variations on geo- thermal gradients are discussed by Birch (1948). CHARACTER As permafrost is defined as a temperature phenomenon, it may en- compass any type of natural or artificial material, whether organic or inorganic. Generally, permafrost consists of variable thicknesses of perennially frozen surficial unconsolidated materials, bedrock, and ice. Physical, chemical, or organic composition, degree of induration, texture, structure, water content, and the like, range widely and are limited only by the extremes of nature or the caprice of mankind. For example, perennially frozen mammals, rodents, bacteria, artifacts, beds of sand and silt, lenses of ice, beds of peat, and varied junk piles, such as kitchen-middens, mine dumps, and ships’ refuse heaps, are in- dividual items that collectively can be lumped under the term perma- frost. Ground perennially below freezing but containing no ice has been called “dry permafrost” (Muller, 1945). Permafrost containing much ice is abundant, particularly in poorly drained fine-grained materials. The ice forms thin films, grains, fill- 252 BLACK. PERMAFROST [Cu. 14 oe o o Ls o = Oe) 0 wm 92 =< oO = = c SS res Bie See eae x o (3) s x a= mt ©) So ee go ae = 2 S Syn = oy Ra) os 80! 90100 Relative erosion rates Fic. 10. Relative erosion rates under different vegetal covers in Pacific Northwest. Forest Ufiiees=a =e Permanent pasture _--________ Range or seeded pasture (good) Legumes-grass hayland_______ Range or seeded pasture (poor) NGG) = ee =e Small grain (standing or stubble) Wheat-peas (stubble not burned) Small grain (poor)--_________ Wheat fallow(stubble not burned) Wheat fallow (stubble burned)___ Orchards-vineyards (clean tilled) and structural differences of soils, an extreme variation of 33-fold in their erosion potential may be expected. By far the greatest relative difference in rate of erosion was found to result, however, from the character and density of vegetal cover. If the same soil, slope, and rainfall are assumed, the rate of erosion on land used for continuous row crops (such as uncontoured cotton, corn, or tobacco) was found to be more than 100 times as great as on land used for hay, pasture, woodland, and forest. Figure 10 shows relative erosion values for different vegetal covers in the Pacific Northwest. 400 BRowN. EFFECTS OF SOIL CONSERVATION [Cu. 22 These research data give a basis for estimating the potential re- duction in sheet erosion that might be achieved by modification of the causal factors. Although rainfall intensity and frequency cannot in themselves be changed, analysis of their monthly expectancy in any area gives a basis for planning maximum protection of the soil sur- face by cover crops, mulches, etc., in those months when the erosion potential from rainfall is greatest. The effects of length and degree of slope indicate the relative need for measures such as terraces, diver- sion ditches, strip crops, and benching to reduce the effective slope and break it into shorter lengths. Inherent differences in soil erodibility suggest adjustment of land use to afford greater vegetal protection of more erodible soils and more use of less erodible soils for row crops. It indicates also possibilities for reducing the inherent erodibility by changing the soil structure through tillage practices and by plowing under cover crops, mulches, and manure. Lastly, differences in effects of vegetal cover emphasize the need for rotation systems of farming, use of cover crops, and conversion of some row-crop land to permanent cover. Experiment station data have shown (Brown, 1948) that a good 3- or 4-year rotation will reduce sheet-erosion loss to 14 to 45 percent of that occurring under one-crop system of cultivation. Contour farm- ing, strip cropping, and terracing, under various soils, slopes, and rain- fall, effect reductions ranging from 10 to more than 90 percent, averag- ing generally about 50 to 75 percent (Stallings, 1945a, b, c). These experimental results have led conservationists to the conclu- sion that sheet erosion on agricultural land can be reduced 50 to 75 percent or more while agricultural production is sustained and im- proved by using every acre in accordance with its capabilities and treating every acre in accordance with its needs. Other types of erosion may require special forms of treatment. Stabilization of major gullies, valley trenches, or stream banks, for example, may be accomplished in some places by planting of trees and vines; or, if particularly unstable, they may require small dams, revet- ments, jetties, or other control structures. It is physically possible in areas of more than 20 inches of rainfall to stabilize completely most gullies and virtually eliminate them as a source of sediment. The less erosion can be controlled by vegetation, the more costly the control becomes, and the more likely that 1t must be accomplished through some federal program such as that contemplated in the Flood Control Act of 1936 (U.S. Congress, 1936). In areas of less than 20 inches of rainfall, more reliance must be placed on structural measures such as revetments, check dams and debris basins. Economic rather than phy- Cx. 22] POTENTIALITIES FOR SEDIMENT CONTROL 401 sical considerations, therefore, primarily govern the amount of re- duction in sediment production that can be achieved in many water- sheds, particularly those in the more arid sections of the country. The effects of changes in land use and treatment on sediment pro- duction from watersheds have been determined in a few representative cases by repeated reservoir-sedimentation surveys. In the Sangamon River watershed above Decatur, Illinois, more than 90 percent of the sediment is coming from sheet erosion. The rate of sediment production averaged 20 percent higher during the 10 years from 1936 to 1946 than during the preceding 14.2 years as a result of the greater use of land for intertilled row crops from 41 to 60 percent of the total watershed area. The writer and collaborators (1947) have estimated that, if farmers would adopt the recommended rota- tions, use contour farming, strip cropping, terracing, and other practices where needed in accordance with the capabilities of the land, the rate of sediment production from the area would be reduced by 62 percent from its rate in 1946 without reducing the level of net farm income. By reforestation and gully control on 83 percent of an 890-acre area of rolling and badly eroded land above the municipal reservoir at Newnan, Georgia, the rate of sedimentation was reduced 62 percent during the period 1937-1945, as compared with the earlier period 1925— 1937 (Brune, 1947). The reduction is progressively increasing as the watershed-treatment work becomes more fully effective. On the 62-square-mile watershed of the municipal reservoir at High Point, North Carolina, soil-conservation measures applied on about 35 percent of the total acreage caused a reduction of 24 percent of the rate of sedimentation for the period 1934-1938, as compared with the earlier period 1928-1934 (Brune, 1947). Still other records have been cited by the writer (1944), and more data are now being obtained. From the evidence now available, it is concluded that in the prin- cipal agricultural areas of the United States present rates of soil erosion could be reduced 50 to 75 percent or more by proper land use and treat- ment without decreasing the net agricultural income from the land (Brown, 1948). Rates of sediment production, as measured at a given point in a stream system, could be reduced even more by watershed- treatment measures that provide for control of channel erosion and selective deposition of erosional debris at many locations within the watershed. The limitation in controlling sediment production in the semi-arid to arid, non-agricultural regions of the United States is more economic than physical in that it depends primarily on structural con- trol works, accompanied by rigid regulation of land for grazing. These controls can be obtained, in general, only at public expense. The high 402 BROWN. EFFECTS OF SOIL CONSERVATION [Cu. 22 value of maintaining water-storage capacity in many western water- sheds indicates, however, that the cost of reducing sediment production by some 50 percent, on an average, through watershed-treatment measures, may be justified. RESEARCH AND PLANNING NEEDS In few fields is such a wide variety of research and planning needed as in the conservation of the soil and water resources of the nation. Needed studies fall within the scope of soil science (particularly soil physics and soil chemistry), agronomy, ecology, biology, forestry, hy- drology, geology (particularly geomorphology, sedimentation, and ground water), engineering (particularly hydraulic, structural, and ag- ricultural), economies, and law. Soil- and water-conservation research can be grouped functionally under the following headings: Studies of basic forces and resistances. These include, for example, studies of the forces involved in raindrop impact; the fluid dynamics of thin overland flows of water; the mechanics of sediment entrainment, transportation, and deposition in open-channel flow; aerodynamic forces causing soil movement; and the resistances afforded by the character of the soil, degree, length and aspect of slope, and various types of vegetal cover. Studies of the areal distribution and frequency of occurrence of forces and resistances. This research involves analysis of the geographic distribution in terms of amounts, intensities, and duration of precipita- tion, wind action, runoff, sediment transportation, soil erodibility, veg- etal cover, and watershed characteristics such as drainage density and relief. Development, testing, and improvement of measures and practices for the control of water erosion and motsture conservation on the land surface. Development of methods for controlling water and sediment move- ment and conveyance in channels, conduits, etc. Development of methods for water conservation and utilization in the soul and underground storage. Development of measures and practices for wind-erosion control. Studies of conservation economics affecting the operation and in- come of farms and other private land holdings, of drainage enterprises, of wrigation enterprises, and of public interests in flood and sediment control. Cu. 22] REFERENCES 403 Studies of the legal requirements and methods of organization for carrying out programs of soil and water conservation. Most of the physical requirements for controlling soil erosion on in- dividual farms and ranches are being developed by several thousand soil conservationists who have brought into being a new field of re- search, planning, and action founded on several of the older sciences. Farm- and ranch-conservation planning and technical assistance to landowners and operators in the application of sound land-use read- justments and conservation measures, such as terracing, strip crop- ping, contour farming, rotations, farm waterways, and farm ponds, are within the general field of practicing soil conservationists. In the more complex problems of planning adequate water disposal, flood and sediment control, and related farm drainage and irrigation on watersheds, the specialized experience of hydrologists, hydraulic en- gineers, soil scientists, and geologists is required. Sediment control, perhaps the most difficult of the conservation problems, requires the services of: (1) hydraulic engineers who are familiar with modern de- velopments in the fields of hydrodynamics and fluid mechanics, with particular reference to the effect of entrainment, transportation, and deposition of sediment on open-channel behavior; (2) geologists who understand: (a) the concepts of geomorphic development of land- scapes and the interpretation of land forms and drainage patterns in terms of their causal factors; (b) the geological aspects of sedimenta- tion, particularly those phases involving the statistical analysis of sediment characteristics; (3) soil scientists who can aid in relating the characteristics of soil texture, structure, and depth that affect erosion to rates of sediment production on the watershed; and (4) hydrologists who can aid in the correlation of hydrologic characteristics of water- sheds with rates of sediment production. The planning and design of control measures on the watershed also calls for the services of vegeta- tive specialists, on the one hand, and the structural engineer, on the other. Experience to date has shown that most sediment problems can be solved most effectively and economically by the collective ef- forts of a group of specialists who have an opportunity to study and plan jointly corrective measures to meet the needs of each individual watershed. REFERENCES Adams, C. (1944). Mine waste as a source of Galena River bed sediment: Jour. Geol., vol. 52, No. 4, pp. 275-282, July 1944. Bennett, H. H. (1939). Soil conservation: McGraw-Hill Book Co., New York and London, 993 pages. 404 BRowN. EFFECTS OF SOIL CONSERVATION [Cxu. 22 Brown, C. B. (1944). The control of reservoir silting: U. S. Dept. Agr., Misc. Pub. 521, 166 pages, U. S. Govt. Print. Off., 1943; slightly revised, Aug. 1944. (19452). Floods and fishing: The Land, vol. 4, No. 1, pp. 78-79, Winter 1945. (1945b). Rates of sediment production in southwestern United States: U.S. Soil Conservation Service, SCS-TP-58, 40 pages, Jan. 1945. (1945c). Sediment complicates flood control: Civ. Eng., vol. 15, No. 2, pp. 83-86, Feb. 1945. (1946). Aspects of protecting storage reservoirs by soil conservation: Jour. Soil and Water Cons., vol. 1, No. 1, pp. 15-20, 48-45, July 1946. (1948). How effective are soil conservation measures in sedimentation control? Proceedings of Federal Inter-Agency Sedimentation Conference, Den- ver, Colo., May 6-8, 1947, U.S. Bureau of Reclamation, Washington, D. C., pp. 259-266, Jan. 1948. , Stall, J. B., and De Turk, E. E. (1947). The causes and effects of sedi- mentation in Lake Decatur: Jil. State Water Survey Dviv., Bull. 37, 62 pages. , and Thorp, E. M. (1947). Reservoir sedimentation in the Sacra- mento-San Joaquin drainage basins, California: U.S. Soil Conservation Service, Spec. Rept. 10, 69 pages, July 1947. Brune, G. M. (1942). Island-formation and channel-filling on the upper Wabash River: Trans. Amer. Geophys. Union, vol. 23, Pt. 2, pp. 657-663. (1947). Effects of soil conservation on reservoir sedimentation in the Southeast: Jour. Soil and Water Cons., vol. 2, No. 2, pp. 103-105, 108, April 1947. (1948). Rates of sediment production in midwestern United States: U. S. Soil Conservation Service, SCS-TP-65, 40 pages, Aug. 1948. Eakin, H. M. (1939). Silting of reservoirs: U. S. Dept. Agr., Tech. Bull. 524, revised by C. B. Brown, 168 pages, U. S. Govt. Print. Off., Washington, D. C. Eaton, E. C. (1936). Flood and erosion control problems and their solution: Trans. Amer. Soc. Civ. Engrs., vol. 101, pp. 1302-1330. Edgecombe, A. R. B. (1934). Report on the silting of reservoirs and measures taken or projected for its prevention or control: Allahabad, United Provinces, India, 36 pages. Faris, O. A. (1933). The silt load of Texas streams: U.S. Dept. Agr., Tech. Bull. 382, 71 pages, Sept. 1933. Fippin, Elmer O. (1945). Plant nutrient losses in silt and water in the Ten- nessee River system: Sozl Science, vol. 60, No. 3, pp. 223-239, Sept. 1945. Fisk, H. N. (1947). Fine-grained alluvial deposits and their effects on Mississippr River activity: U.S. War Department, Corps of Engineers, Mississippi River Commission, Waterways Exp. Sta., Vicksburg, Miss., vol. 1, 82 pages, July 1947. Fortier, S., and Blaney, H. F. (1928). Silt in the Colorado River and its relation to irrigation: U. S. Dept. Agr., Tech. Bull. 67, 94 pages, Feb. 1928. Friedkin, J. F. (1945). A laboratory study of the meandering of alluvial rivers: U.S. War Department, Corps of Engineers, U.S. Army, Mississippi River Com- mission, Waterways Exp. Sta., Vicksburg, Miss., 40 pages, May 1, 1945. Garin, A. N., and Forster, G. W. (1940). Effect of soil erosion on the costs of public water supply in the North Carolina piedmont: U.S. Soil Conservation Service, SCS-EC-1, 106 pages, July 1940. Cu. 22] REFERENCES 405 Garin, A. N., and Gabbard, L. P. (1941). Land use in relation to sedimentation in reservoirs, Trinity River Basin, Texas: Tex. Agr. Exp. Sta., Bull. 597, 65 pages. Gilbert, G. K. (1917). Hydraulic-mining debris in the Sierra Nevada: U.S. Geol. Survey, Prof. Paper 105, 154 pages. Glenn, L. C. (1911). Denudation and erosion in the Southern Appalachian Region and the Monongahela Basin: U. S. Geol. Survey, Prof. Paper 72, 137 pages. Gottschalk, L. C. (1945). Effects of soil erosion on navigation in upper Chesa- peake Bay: Geog. Rev., vol. 35, No. 2, pp. 219-238, April 1945. Happ, S. C. (1944). Effect of sedimentation on floods in the Kickapoo Valley, Wisconsin: Jour. Geol., vol. 52, No. 1, pp. 53-68, Jan. 1944. (1945). Sedimentation in South Carolina piedmont valleys: Amer. Jour. Sci., vol. 248, No. 3, pp. 113-126, March 1945. (1948). Sedimentation in the middle Rio Grande Valley, New Mexico: Bull. Geol. Soc. Amer., vol. 59, pp. 1191-1216, Dec. 1948. , Rittenhouse, G., and Dobson, G. C. (1940). Some principles of accel- erated stream and valley sedimentation: U. S. Dept. Agr., Tech. Bull. 695, 133 pages. Hathaway, G. A. (1948). Observations on channel changes, degradation, and scour below dams: International Association for Hydraulic Structures Research, Stockholm, Sweden, Report on the Second Meeting, Appendix 16, pp. 287-807. Howard, C. 8. (1947). Suspended sediment in the Colorado River, 1925-41: U. S. Geol. Survey, Water-Supply Paper 998, 165 pages. Lowdermilk, W. C. (1934). Acceleration of erosion above geologic norms: Trans. Amer. Geophys. Union, Ann. Meeting 15, pp. 505-509. McLaughlin, T. G. (1947). Accelerated channel erosion in the Cimarron Valley in southwestern Kansas: Jour. Geol., vol. 55, No. 2, pp. 76-93, March 1947. Marshall, R. M., and Brown, C. B. (1939). Erosion and related land use condi- tions on the watershed of White Rock Reservoir, near Dallas, Texas: U.S. Soil Conservation Service, Watershed and Conservation Surveys, 29 pages. Musgrave, G. W. (1947). The quantitative evaluation of factors in water erosion —a first approximation: Jour. Soil and Water Conserv., vol. 2, No. 3, pp. 133- 138, July 1947. Osborn, F. (1948). Our plundered planet: Little, Brown and Co., Boston, 217 pages. Parshall, R. L. (1947). Riffle deflectors and vortex tubes remove suspended silt and sand: Civ. Eng., vol. 17, No. 12, pp. 34-35, 94, Dec. 1947. Rittenhouse, G. (1944). Sources of modern sands in the middle Rio Grande Valley: Jour. Geol., vol. 52, No. 5, pp. 145-183, May 1944. Seavy, L. M. (1948a). Sedimentation survey of Arrowrock Reservoir, Boise Project, Idaho: U.S. Dept. of Interior, Bureau of Reclamation, Denver, Colo., 43 pages, March 1948. (1948b). Sedimentation survey of Guernsey Reservoir, North Platte Project, Wyoming and Nebraska: U.S. Dept. of Interior, Bureau of Reclama- tion, Denver, Colo., 41 pages, Nov. 1948. (1949). Sedimentation surveys of Elephant Butte Reservoir, Rio Grande Project, New Mezxico-Texas: U. 8. Dept. of Interior, Bureau of Reclamation, Denver, Colo., Feb. 1949, 36 pages. Stallings, J. H. (1945a). Review of data on contour furrowing, pasture and range land: U.S. Soil Conservation Service, 10 pages, Dec. 1945. 406 BRowN. EFFECTS OF SOIL CONSERVATION [Cu. 22 Stallings, J. H. (1945b). Review of terracing data on crop yield, runoff and soil loss: U. S. Soil Conservation Service, 8 pages, Dec. 1945. (1945c). Summarization of strip cropping data on crop yield, runoff and soil loss: U. 8. Soil Conservation Service, 9 pages, Dec. 1945. Stevens, J. C. (1936). The silt problem: Trans. Amer. Soc. Civ. Engrs., vol. 101, pp. 207-288. Thornthwaite, C. W., Sharpe, C. F. S., and Dosch, E. F. (1942). Climate and accelerated erosion in the arid and semi-arid Southwest, with special reference to the Polacca Wash Drainage Basin, Arizona: U. S. Dept. Agr., Tech. Bull. 808, May 1942. U.S. Bureau of Reclamation (1948a). Report of river control work and investi- gations, Lower Colorado River Basin, Calendar Years 1946 and 1947: Region 3, Office of River Control, Boulder City, Nevada. U.S. Bureau of Reclamation (1948b). Proceedings of the Federal Inter-Agency Sedimentation Conference, Denver, Colo., May 6-8, 1947: Washington, D. C., 314 pages, Jan. 1948. U. S. Congress (1936). An act authorizing the construction of certain public works on rivers and harbors for flood control and other purposes: Public Law 738, 74th Cong., 2nd Sess. (1942). Little Tallahatchie Watershed in Mississippi: Letter from the Acting Secretary of Agriculture transmitting a report of a survey of the Little Tallahatchie Watershed in Mississippi, based on an investigation authorized by the Flood Control Act of June 22, 1936, 77th Cong., 2nd Sess., House Docu- ment 892, 74 pages. (1946). Pollution of navigable waters: Hearings before the Committee on Rivers and Harbors, House of Representatives, 79th Cong., 1st Sess., on HR 519, HR 587, and HR 4070, 347 pages. (1948). Long-range agricultural policy—a study of selected trends and factors relating to the long-range prospect for American agriculture: 80th Cong., 2nd Sess., House of Representatives Committee on Agriculture, 72 pages. U. S. Forest Service (1936). The western range: Letter from the Secretary of Agriculture ..., 74th Cong., 2nd Sess., Sen. Doc. 199, 620 pages. U.S. Forest Service and Soil Conservation Service (1940). Influences of vegeta- tion and watershed treatments of runoff, silting, and stream flow, a progress report of research: U. S. Dept. Agr., Misc. Pub. 397, 80 pages. U.S. War Department, Corps of Engineers, U. S. Army (1943). Interim report on sedimentation in Conchas Reservoir, South Canadian River, New Mexico: U.S. Engineer Office, Albuquerque, New Mexico, Sept. 1, 1943, 38 pages. Vogt, W. (1948). Road to survival: William Sloane Associates, New York, 335 pages. CHAPTER 23 THE PROBLEM OF GULLYING IN WESTERN VALLEYS * H. V. PETERSON Geologist, U. S. Geological Survey Salt Lake City, Utah To distinguish between gullying and other types of erosion, the term cully is generally applied to any erosion channel so deep that it cannot be crossed by a wheeled vehicle or eliminated by plowing. When ap- plied to the arid and semi-arid valleys of the West, this limitation re- garding minimum size seems somewhat inconsistent, for with many gullies having dimensions measured in tens of feet in depth, hundreds of feet in width, and scores of miles in length, a definition aimed at distinguishing them from canyons or small valleys might have been more appropriate. Generally the term arroyo has been used synonymously with gully, but by some (Thornthwaite et al., 1942, p. 72) arroyos are considered the result of natural conditions of erosion in contrast to gullies, which have developed under “culturally accelerated erosion” or, in other words, land misuse. Because it is difficult or impossible to learn from the appearance of the channel the conditions under which it developed, the distinction has little significance, and in this discussion the less euphonius but more descriptive term “gully” will be used. Gullies appear to have been a common feature of stream erosion in the recent geologic history of western valleys and possibly have always formed a part of the landscape in localities favorable to their develop- ment. Evidence of former gullies, similar in cross section and dimen- sions to existing ones, can usually be found wherever any -extensive section of fine-textured alluvial fill is exposed. Just as at present, gullies have developed in localities underlain by soft, easily eroded materials, which offered minimum resistance to downcutting. It is dif- ficult to determine the extent of this earlier gullying or to ascertain whether the channels were integrated into a continuous system wherein the gradients of all branches were accordant with the parent stream, * Published by permission of the Director, U. S. Geological Survey. 407 408 PETERSON. PROBLEM OF GULLYING [CH. 23 or whether they occurred as discontinuous segments, each developed individually with no relation with its neighbor, except for being aligned along a common drainage course. Both types exist today, so it may logically be presumed that they occurred previously. Although possibly the most destructive and certainly the most spec- tacular in appearance, gullies represent only one phase of water erosion. They may even be subordinate so far as silt removal from a given locality is concerned, being surpassed in this respect by rilling or by the much less prominent sheet erosion occurring on interfluviatile areas. What constitutes their critical and insidious threat is the lowering of the base level within the drainage system, which exposes to potential removal all material within the basin above this level. A long gully of narrow dimensions may represent the removal of only a few thou- sand acre-feet of silt, but potentially it has set the stage for removal of hundreds of thousands. That gullying is not confined to the arid and semi-arid West is shown by its widespread evidence in other localities, but when con- sideration is given to the problems involved in attempting to control or ameliorate western gullies, complications of unique nature im- mediately become apparent. As will be shown, the area affected by gullying is of tremendous extent, but the value of the land on a per-acre basis is extremely low, so low that treatment of only a minor nature can be justified on the basis of land improvement alone. Despite the low value, however, the lands are the major support of the western livestock industry, and practically every acre is utilized for grazing to the full limit of its forage production. Authority for such use is established on the basis of either direct ownership or legally recognized leases or permits. The vegetative cover that supports the livestock also forms the major, if not the only, effective protection against ero- sion. Sparse and erratically distributed precipitation in which droughts of several years’ duration alternate with flash floods of tremendous eroding power is characteristic of almost the entire area. The problem of maintaining a vegetative cover under these conditions is readily envisioned. Further to complicate the problem, many of the remedial measures applied thus far to a limited extent in an attempt to correct erosion have been of very doubtful success, thus raising the pertinent question whether our knowledge of the meteorologic, ecologic, hydrologic, and geologic aspects of the problem is extensive enough to permit designing a practical treatment program that will have a reasonable chance of accomplishing the desired results. As an example of western gullies, one embodying the common history Cu. 23] SAN SIMON VALLEY 409 and features associated with channels of this type, the San Simon gully of southwestern Arizona, might be considered typical. From its con- fluence with the Gila River near the town of Solomonsville, Arizona, this gully cuts southeasterly for nearly 70 miles through the heart of the lower San Simon Valley. The gully is not continuous, as one short reach of the valley floor, approximately 2 miles in length situated some 40 miles above the mouth, is uncut. It is not known if this reach has previously been dissected, but it is evident that unless remedial meas- ures are taken the two discontinuous segments will shortly be inte- erated into one continuous channel. In depth, the gully, in both branches, varies from 6 to 60 feet; in width from 75 to more than 1,000 feet. Surveys made by the U. 8. Soil Conservation Service * in the middle 1930’s show that silt removed from the main trench is of the order of 20,000 acre-feet. Additional excavations from gullied tributaries and from sheet and badland erosion on interfluviatile belts adjacent to the bank, both traceable in the main to rejuvenation re- sulting from development of the master gully, probably equal or ex- ceed this amount. Prior to the 1929 completion of Coolidge Dam, located on the Gila River some 70 miles downstream, silt disposal from the eroding area offered no particularly critical problem. Since that date it has become a very real threat to the life of the San Carlos Reservoir. The San Simon Valley, like many others of its kind located through- out the West, contains a deep alluvial fill which originated as outwash from the surrounding mountains. Older portions of the fill, repre- sented by both lacustrine and continental deposits, form a part of the Gila conglomerate of late Pliocene and Pleistocene age, described in the literature by Knechtel (1937). These beds, which are usually indurated to some extent, are exposed around the margins of the valley, and although extensively dissected they generally exhibit no serious recent erosion. Distributed along the central axis of the valley and extending moun- tainward along the tributaries, are deposits consisting of clay, silt, and fine sand intermixed with occasional lenses and stringers of coarse sand and gravel. These are typical flood-plain deposits obviously laid down by gently flowing streams capable of carrying fine-textured sediments only, except during periodic floods, when the coarse sand and gravel were brought in. Archaeologic evidence ~ indicates that * Unpublished data obtained through personal communication with earlier employees of U.S. Soil Conservation Service. + Personal communications from Dr. Emil Haury, Department of Anthropology, University of Arizona, furnished probable dating of artifacts found in the valley. 410 PETERSON. PROBLEM OF GULLYING [Cu. 23 these deposits were laid down within the past few thousand years, but that it was not a continuous process of aggradation is shown by the presence of numerous filled channels and well-defined erosional sur- faces within the fill. It is evident that, at various times in the past, erosion similar in character to, and possibly as extensive as, that taking place today has occurred. At the time of white settlement of the adjacent Safford Valley in the late 1870’s, the San Simon Valley is said to have presented a picture of pristine beauty.* Its floor was flat and unbroken. Reportedly large areas in the central portion were covered with grass thick enough and tall enough to be harvested for hay. San Simon Creek, dignified by some of the earlier explorers with the appellation “river,” was perennial throughout most of its length and meandered across the valley floor in a shallow channel, lined for most of its length with trees and willows. Stockmen naturally recognized this as an ideal grazing setup, and dur- ing the 1880’s, 50,000 head of cattle are said to have populated the valley. The present gully is reported to have started in the early 1880’s when farmers constructed a small drainage ditch to carry flood waters of San Simon Creek across farm lands adjacent to the Gila River. It reportedly did not reach serious proportions, however, until 1905. By then, grazing combined with the critical ten-year drought, extending from 1895 through 1904, had eliminated the protective grass cover- ing on the valley floor, leaving it “ripe” for gullying. The record wet winter of 1904-1905 with a winter seasonal index of wetness of 246 (the highest on record) provided the runoff necessary to accomplish this. Just how far the cutting advanced in this one season is not known, but, to judge by the meager progress measured in the past few years, probably the major portion of the present channel was excavated during this short period. Today’s picture of the valley, from both the conservation and the range-use viewpoint, is one of devastation. Except in the short uncut reach, the former grassy tracts now appear as barren flats, some com- pletely devoid of vegetation, others supporting only an occasional stunted bush or clump of grass. In places along each side of the chan- nel, belts up to several hundred feet in width and several miles in length have been stripped of topsoil to depths of 3 feet or more. Some of these remain essentially flat, and others have deteriorated into miniature badlands with a relief of 2 to 6 feet. The stream has long since lost any semblance of permanency, and the ephemeral flows are * Information obtained from interviews with several early settlers. Cu. 23] DISTRIBUTION OF GULLIES 411 marked by sudden peaks and rapid recession, the magnitude of flow depending on the intensity and duration of the accompanying rain since there is little vegetation to either impede or decrease the runoff by in- ducing percolation. The headcuts and gully banks remain vertical as a result of sapping and undercutting by these periodic flows. Al- though most of the major tributaries of the San Simon are still uncut, side drainage pouring over the vertical banks, particularly opposite the unprotected barren tracts, has incised literally hundreds of short deep channels and subterranean passages, leaving the terrain next to the bank cut into narrow fingers and isolated blocks of soil. Under these conditions the potential silt contribution to the Gila River from the San Simon Valley is enormous, being limited only by the amount of water available for transportation, since the sediments themselves offer little if any resistance to removal. Typical views of the present condition of the San Simon gully are shown in Fig. 1. With suitable modification for differences in length and dimensions of the gully, character of the valley fill, history of gully development, and other details, the picture of the San Simon could be transposed to a myriad of other valleys located throughout the West. The con- trol of these gullies is essentially the major erosion problem of the West. DISTRIBUTION OF GULLIES Gullies similar to the San Simon, some greatly exceeding it in size and destructiveness, are common throughout the Southwest. In New Mexico the gullied valleys of the Rio Grande, including the Rio Puerco, the Jemez, and the Salado, have been described in such detail and are so well known as to be familiar to those having only a casual acquaint- ance with the erosion problems of the area, but scores of others of al- most equally impressive dimensions also occur in the basin (Bryan and Post, 1937; National Resources Committee, 1938). In the Gila River drainage, the valleys of the Santa Cruz, the San Pedro, the Mangus, and the San Simon represent only the larger of numerous gullied tribu- tary valleys distributed from near the headwaters of the river almost to the mouth. Although possibly not so prevalent as in southern latitudes, gullies nonetheless occur in large numbers in the headwater areas of the Mis- sourl and its tributaries. Particularly is this true of the Bighorn and Powder rivers in Wyoming. One familiar with these streams will im- mediately call to mind such striking examples as Five-Mile, Muddy, E-K, and Badwater creeks in the Wind River Basin; Fifteen-Mile, 412 PETERSON. PROBLEM OF GULLYING [CH. 23 Cottonwood, and Dry creeks in the Bighorn Basin; and Mispah and Pumpkin creeks and Little Powder River in the Powder River Basin. These are but a few among many. Milk River in the north has a similar, though perhaps smaller, quota, the Willow Creek gully (some 25 miles in length and averaging approximately 20 feet in depth and 100 feet in width) being a prime example. Tributaries of the North Platte River, draining from Casper Mountain in central Wyoming, exhibit a network of gullies seldom duplicated in other localities. Parts of the Columbia River Basin are likewise checked with the tell- tale gully scars. The greatest development has already occurred in the semi-arid wastes of eastern Oregon and Washington, but many valleys in Idaho and northern Nevada show evidence of the same action. Probably the most advanced and critical development of gullying is found in the Colorado River Basin. Examination by the Inter-Moun- tain Forest and Range Experiment Station (Bailey, 1937) ,xeveals that, of the 115 major tributaries of the Colorado and Green rivers above Lees Ferry, 111 have been trenched by gullies. The combined total length of these channels and their associated tributaries is thousands of miles, and the material removed in these excavations probably ag- gregates hundreds of thousands of acre-feet. Stevens (1936, p. 1254) and Stabler (1936, p. 281) have focused attention on the paradoxical position of the Colorado Plateau in regard to silt and water contribu- tion to the Colorado River. The Plateau area, comprising about 65,000 square miles, or 45 percent of the drainage basin of the Colorado above the Grand Canyon, contributes less than 10 percent of the water but more than 75 percent of the silt entering the stream. Practically every valley within the area is gullied. Tributaries of the Colorado entering below the Grand Canyon, in- cluding the Bill Williams River, Meadow Valley Wash, Virgin River, and Kanab Creek together with most of their tributaries, present the same picture. Even valleys which at present make no pretense of pos- sessing creeks or rivers (these terms are used in the optimistic sense peculiar to the West, and one must be on the spot at the proper instant even to glimpse the flow of water) have not escaped being gullied, as evidenced by the spectacular trenches found in the lower reaches of the Sacramento and Bouse valleys located in some of the driest parts of western Arizona. An examination of numerous tributaries of some of the large Great Basin streams, such as Bear River in Wyoming and Idaho, Sevier River in Utah, and the Humboldt, Truckee, and Walker rivers in Nevada, reveal the same pattern of gullying. Even smaller streams that drain some of the driest parts of the Basin, as, for example, Cu. 23] PRACTICAL ASPECTS 413 Thousand Springs Creek in northeastern Nevada and Snake Creek in western Utah, have, in parts of their drainage area, developed gully systems that rival those existing in more humid areas. The fact that both cutting and the concurrent deposition in these remote areas affect only low-value lands accounts for the lack of attention thus far ac- corded the areas. PRACTICAL ASPECTS OF THE GULLY PROBLEM The above review of the widespread prevalence of gullying is pre- sented to emphasize that the phenomenon is not limited to any one locality or governed by any set of conditions relative to topography, geology, soils, or climate; it is obvious that, throughout this vast area, wide and contrasting variations in these features must occur. Moun- tain valleys, with slopes ranging up to one or two hundred feet per mile, exhibit the same gully characteristics as do others with slopes of only 10 to 15 feet per mile. And, although as noted, gullies are some- what more prevalent in areas such as the Colorado Plateau and por- tions of the High Plains and Rocky Mountain provinces, where the valley fills are derived from surrounding friable sandstones and shales, they also occur in almost equal numbers in southern Arizona and other localities where the alluvium originates from mountains composed of igneous, voleanic, or highly indurated sedimentary and metamorphic rocks. Dimensions and shapes of the valleys likewise appear to have little influence, for the small fan-shaped basin with limited drainage area commonly may be as deeply and extensively incised as its larger elongated counterpart. Although trenching and the destruction of land within the affected valleys constitute a serious phase of the problem, because the attack is aimed directly at the more productive portions of the range, this as- pect is possibly of less importance than the disposition of the removed silt. The few figures, such as the previously mentioned survey of the San Simon and a similar estimate by Bryan and Post (1937, p. 86), which shows that nearly 400,000 acre-feet have been removed from the main and tributary channels of the Rio Puerco, are significant only as indicators. Were the surveys extended to cover all the hundreds of other major gullies, aggregating thousands of miles in length, an even more disturbing picture would be revealed. Except where silt from these valleys is deposited without damage on low-value lands within interior basins, it enters and becomes the serious sedimentation prob- lem of the major streams. Thus the problem has a dual aspect: de- 414 PETERSON. PROBLEM OF GULLYING [Cu. 23 Fic. 1(a). View upstream near head of the San Simon gully, Arizona. Fic. 1(b). Badlands developing adjacent to San Simon gully, Arizona. Main channel not shown. Note brush and rock spreader in left foreground. Fic. 1(c). Erosional unconformity ex- posed in walls of San Simon gully, Ari- zona. Top layer is brown silt, bottom layer is pinkish gray clay. Cu. 23] PRACTICAL ASPECTS 415 Fic. 2(a). Headcut on a tributary of Rio Puerco, near Cabezon, New Mexico. Height 40 feet. Fic. 2(6). Headeut of Willow Creek gully, tributary of Milk River near Glasgow, Montana. Height 25 feet. Hay meadow directly above. Fic. 2(c). Recently gullied channel in Centennial Wash near Salome, Arizona. Height about 12 feet. 416 PETERSON. PROBLEM OF GULLYING [Cu. 23 struction of land, on the one hand, and constantly mounting diffi- culties relating to silt disposal, on the other. MECHANICS OF GULLY CUTTING Gullies follow a simple pattern of development which is character- istic in all localities. Almost invariably they are marked by vertical or nearly vertical headeuts and banks which give them a typical rectangular cross section. Once established, the headcuts advance, and the channels widen by sapping and undercutting the banks. Abra- sion by flowing water, either at the falls or above, is a relatively minor cause of enlargement. The depth of cutting and the gradient of the downstream channel vary widely, apparently dependent on a function of the flow, the character of the eroding sediments, and the slope of the valley floor. Depths ranging from a few inches to 60 feet have been observed. Advancing in this manner, gullies may cut for an indefinite distance without any disturbance of the surface, either laterally or upstream, be- yond the confines of the channel (Fig. 1a). Also, because cutting gen- erally takes place at depths below the reach of plant roots, the presence of even dense vegetation above headcuts has little if any influence on the rate of advancement (Fig. 2b). The upstream progress of headcuts occurs at highly variable rates which depend on the volume and duration of flow and on the character of the eroding sediment. A headcut can remain inactive for a period of years, then under proper conditions of flow it may progress hundreds or even thousands of feet within a few hours or days. Table 1 shows the measured progress of a number of headcuts during the past few years. The relatively insignificant progress made by the gullies during the past few years (which in most of the localities shown have been abnormally dry) strongly suggests that most of our long gully systems have been developed mainly during the few years that produced ex- traordinary floods. Unfortunately no records on growth are available for these years, but how else can one account for the San Simon mov- ing only a few hundred feet in 6 years when it has cut nearly 70 miles in the 44 years since its beginning in 1905? The condition is one that should be considered in making an estimate, on the basis of known records, of long-term silt carried by any particular stream, for, unless the period of record contains its quota of these wet seasons, the esti- mate can be highly erroneous. us 23 CAUSES OF GULLYING 417 TABLE 1 MraAsurED ProGress oF GuitLy ADVANCEMENT Date of Date of peat pie j ee Gully Designation River Basin Initial Check = —— Sivey ress Lees Gully of Gully | Started feet miles San Simon Creek, south- | Gila River Feb. 1944 | Feb. 1946 10 65 1905 western Arizona Feb. 1946 | Oct. 1948 50 Centennial Wash near | Gila River Feb. 1945 | Mar. 1947 75) 10 1920? Salome, Arizona Mar. 1947 | Dec. 1948 600 Deadmans Wash near | San Juan River Jan. 1944 | Oct. 1948 160 2 1910 Shiprock, New Mexico Hogback Wash near | Chaco River 1936 | Feb. 1946 200 1 1920 Shiprock, New Mexico Unnamed wash near | Mancos River 1936 | Feb. 1946 150 10 Unknown Shiprock, New Mexico Feb. 1946 | Oct. 1948 0 CAUSES OF GULLYING As gully cutting represents but one aspect of erosion, consideration of its cause involves an inquiry into the broad phases of the erosion problem. Specifically, erosion may be assumed to occur in a given locality when the resistance of the surface to erosion is less than the erosive power of the eroding agent. Thus, in gully formation, which is clearly the result of water action, the erosion may logically be at- tributed to a condition that either lowered the resistance of the sur- face in a given locality or increased the size or the velocity of the | stream. A combination of the two would make the action doubly ef- fective. As the fine-textured alluvium that floors practically all the western gullied valleys has little if any inherent cohesion, the greatest re- sistance to erosion is afforded by the protective vegetative cover. De- pletion or destruction of this cover in any manner sets the stage for incipient erosion under normal conditions of flow, and, likewise, flows of extraordinary magnitude might also be expected to accomplish the same action under normal condition of cover. Recognition of the critical nature of the erosion problem, particularly since the early 1930’s, has stimulated much study of both its cause and methods for curing or ameliorating it. As could be expected, the studies have evoked controversies on both points which, unfortunately, are no nearer settlement today than when first advanced. Doubtless 418 PETERSON. PROBLEM OF GULLYING [Cu. 23 these differences reflect the extreme complexity of the problem and the tenuous and uncertain nature of the evidence available for its interpretation. In general two theories have been advanced as the cause of the recent excessive erosion of the West. Advocates of the first ascribe gullying exclusively to land use, or misuse, the chief form of which is overgrazing, but establishment of roads and trails and other activities which locally destroy the protective vegetative cover are also included. From this conception has developed the term “accelerated erosion,” now so widely used in the literature on conservation. Advocates of the second theory consider the present cycle merely another in the sequence of similar events that happened previously, each ascribable to slow changes in climate wherein aggradation occurred in wet and degradation in dry periods. Adherents of this theory hold generally that overgrazing merely acted as a trip to set off in advance events which were ultimately bound to happen. It is to be noted that there is no disagreement among advocates of either school regarding the importance of vegetation in controlling erosion, nor does either side deny that vegetation has deteriorated. Opinions are split only on the cause of the deterioration. A third suggestion, that increase in stream gradient resulting from regional or differential uplift has been the cause of recent increased erosion, has been considered by some as completely untenable (Bryan and Post, 1937, p. 78) and, by others, has been accorded only minor consideration chiefly because of the difficulty of obtaining substantiat- ing evidence. Another possibility, that irrigation diversion from the many western streams has had an influence on initiating the gully cycle, has not, in the writer’s opinion, been given the study it merits. It appears entirely logical to conclude that interference with the regimen of a stream, particularly where diversions have affected the natural protective bank vegetation, might well lead to changes in the parent channel that would be reflected in the tributaries. ; The arguments for and against the two principal theories are vol- uminous, and there is no need for their review except to point out a few of the salient features of each. Lanp Misuse It would appear that the occurrence of widespread gullying shortly after the white man deployed his herds across the West was no coin- cidence. According to Bryan (1925), Gregory (1917, p. 130), Thorn- thwaite, et al. (1942, pp. 102-104), and the testimony of many living Cu. 23] LAND MISUSE 419 witnesses, most of the important gullies of the Southwest began in the 1880’s. Generally those in northern latitudes were cut at a somewhat later date, but most were cut before 1900. By this time, the livestock population equaled or was approaching that of the present, as indicated by Fig. 3, and advocates of the overgrazing theory conclude that range use had already been sufficient to disrupt the delicate balance between ageradation and degradation previously established in these valleys. Number of livestock in hundred thousand units 1860 1870 1880 1890 1900 1910 1920 1930 Fic. 3. Livestock population in western states, 1860-1930. Data from U. 8. Census reports. Livestock numbers have been reduced to common basis by con- sidering 5 sheep equivalent to 1 cow or horse. Prior to 1890, animals on the farm only were reported; beginning with 1890 the figures include animals both on the range and on the farm. It is natural to assume that the first herds concentrated on the valley floors, where the best feed existed and where water was available. This subjected these vulnerable areas to intense and perhaps destruc- tive use before more remote portions of the range were seriously af- fected. This suggests, therefore, that the initial cutting was due to lowering the resistance to erosion on the valley floors rather than to changes affecting the runoff. It is also known, however, that the Southwest at least was subject to large, if not unprecedented, floods during this same period, and prac- tically every account of gully cutting mentions a rain or a flood of extraordinary magnitude (see particularly Thornthwaite et al., 1942, pp. 102-104; Woolley, 1946, pp. 87-90). As no stream-flow records are available for the period, comparison with floods experienced since is precluded, but precipitation data at a few stations in the Southwest 420 PETERSON. PROBLEM OF GULLYING [Cu. 23 ANNUAL PRECIPITATION AT SELECTED STATIONS FOR THE YEARS 1881-1884, Station Benson, Arizona Bowie, Arizona Ft. Apache, Arizona Ft. Grant, Arizona Granite Reef Dam, Arizona Phoenix, Arizona Univ. of Arizona, Tucson Prescott, Arizona Silver City, New Mexico Deming, New Mexico E] Paso, Texas Ft. Bayard, New Mexico Lordsburg, New Mexico Roswell, New Mexico Santa Fe, New Mexico San Diego, California Los Angeles, California Riverside, California San Bernardino, California 1 Maximum of record. 2 Second maximum of record. TABLE 2 INCLUSIVE 1881 1882 1883 8.60 9.64 | 10.57 15.92 17.90 | 18.86 Sl WY | 7 GA |) Bil ss 18.96 14.82 | 15.48 7.24 9.10 | 10.61 8.91 6.94 7.40 14.92 15.59 | 17.53 15.45 14.26 | 16.13 30.822 | 19.27 | 20.28 20.80 8.71 9.36 18.17 8.27 | 12.92 30.821 | 19.27 | 20.36 17.46 8.74 6.42 19.90 9.91 | 17.04 22.25% | 11.37 | 14.76 5.00 9.74 8.01 5.53 10.74 | 14.14 3.95 5.78 5.54 5.46 10.67 | 12.76 1884 3 Fifth highest of record. Excelled in 1854, 1855, 1856, and 1865. Data from Climatic Summary and other records, U. 8. Weather Bureau. furnish indications of extraordinary rainfall and consequent runoff. Table 2 shows the precipitation at 19 stations in the Southwest for the period 1881 to 1884, inclusive, compared with the long-term mean. Eleven of the 19 stations had the maximum rainfall of record (the records at most stations being continuous to the present time) during one of the four years. Speculation about whether the valleys would have been cut by floods generated during this record precipitation had there been no previous grazing by imported livestock is of little value, because proof is im- possible, but that serious floods and erosion did occur in certain locali- Cu. 23] LAND MISUSE 421 ties prior to range depletion is shown by Woolley (1946, p. 87), who lists the occurrence of four floods in Utah by 1854, only 7 years after the arrival of the first pioneers. Two of the floods carried immense quantities of debris and mud, indicating advanced erosion even under practically virgin range conditions. The most serious criticism of attributing recent cutting exclusively to overgrazing is that it fails completely as an explanation for earlier periods of erosion which occurred long before the area was disturbed by the white man’s herds. Hack (1942), Sayles and Antevs (1941), Albritton and Bryan (1939), and Bryan (1926) have described the evidence of such previous erosion at various localities in the South- west. The writer has found similar evidence in the San Simon Wash of Arizona previously mentioned. The numerous filled channels that can be found outlined in the banks of most existing gullies are also con- sidered positive indications of former gully systems at least approach- ing the present ones in depth and extent. Overgrazing, as a prerequisite to erosion, likewise fails to explain the occurrence of erosion in areas that have never been used. Gregory calls attention to such occurrences in the Navajo Reservation (Greg- ory, 1917, p. 132). The writer has observed a similar condition in the Fort Bayard Military Reservation, New Mexico, where reportedly erazing has been excluded or rigidly controlled during the last several decades (Fig. 4c). Similarly an inconsistency is apparent in the con- dition found in many valleys, where certain portions have cut while other parts have not, although all have been subjected to the same grazing use. In the San Simon Valley, for instance, only a few of the main tributaries are gullied, although the others have obviously been subjected to equal, if not more intense, grazing. The literature cn conservation is replete with descriptions of the deterioration of the western ranges that has occurred since livestock arrival, and most of it is based on the premise that stock alone has been responsible for the depletion (see especially Bailey, 1935; Cooper- ider and Hendricks, 1937; Forsling, 1931; Bailey et al., 1934; Cottam, 1947). Much stress is placed on the descriptions of early explorers, who picture lush grass in contrast to the present barren conditions. These, however, can be considered fair comparisons only when they are made with due regard to precipitation experienced in the contrast- ing years. Occasional seasons of favorable rainfall still produce a cover approaching that described in the earlier accounts. That there has been a general deterioration is conceded, but the proof is not yet positive that it can be attributed to the effects of overgrazing alone. 422 PETERSON. PROBLEM OF GULLYING [@xan23 Fic. 4(a). Showing lack of vegetative recovery in fenced plot located in Free- man Flat near Safford, Arizona. Grazing has been excluded for 14 years. Altitude 3,000 feet. Average annual precipitation about 10 inches. Photo taken October 1948. Fic. 4(b). Shows excellent grass recoy- ery in Steamboat demonstration area, Navajo Indian Reservation, west of Ganado, Arizona. Area reportedly bar- ren in 1934. Grazing has been regulated but not excluded. Altitude 6,500 feet. Average annual precipitation about 16 inches. Photo taken September 1948. Fie. 4(c). Sparse vegetation on Ft. Bayard Military Reservation near Silver City, New Mexico. Grazing has been strictly regulated for past several dec- ades. Altitude 6,200 feet. Average an- nual precipitation 17 inches. Photo taken September 1948. Cu. 23] LAND MISUSE 423 Fic. 5(a). One edge of a filled channel exposed in the bank of Chinle Wash gully near Chinle, Arizona. Fie. 5(b). Outline of filled channel ex- posed in walls of San Simon gully, Ari- zona. Fic. 5(c). Gradual filling in unnamed eully north of Shiprock, New Mexico. Looking downstream. Depth of fill at lower end more than 6 feet. 424 PETERSON. PROBLEM OF GULLYING [Cu. 23 CLIMATIC CHANGES Inquiry into changing climatic conditions, as a prerequisite to ero- sion, is beset by many uncertainties. To begin with, there is no precise standard for identification of such a change. It could be thought of as a reduction in rainfall, an increase in temperature, a change in rainfall distribution, or merely the occurrence of unusually severe droughts or exceptional floods. Recognition and evaluation of changes affecting any item or combination of items is difficult or impossible with the meager data available. Weather observations in the West, extending back to the 1880’s or earlier, have been made at only a few stations. In general these show the expected yearly fluctuations in precipitation with no defined trend toward wet or dry cycles. Droughts and periods of excess precipitation, occurring during the past several centuries, have been identified and dated through tree-ring chronology (Schulman, 1945, pp. 42-47), but the indices used fail again to show definite cyclic trends. Despite the lack of positive evidence supplied by precipitation data, investigations at various localities in the Southwest indicate erosion periods which can be attributed only to a climatic change toward drier conditions. In the gullied Jeddito Wash, located in the Hopi Indian Reservation, Arizona, Hack (1942) has identified three periods of deposition and erosion. Remains of extensive sand dunes, and other features closely associated with the erosion periods, lead to the con- clusion that erosion occurred during a dry cycle. Albritton and Bryan (1939) describe a similar recurrence of deposition and erosion in west- ern Texas, the dates of which can be rather closely correlated with those of Hack (1942, p. 68). Sayles and Antevs (1941, p. 39) recognize former periods of erosion in the Whitewater Creek in southern Arizona, and, in summarizing the evidence pointing to climatic change, Antevs (1948) states, “Thus the same climatic evolution from a moist-Pluvial which has grown slightly drier, is recorded by various conditions in several regions distributed from Trans-Pecos Texas northwestward to the Sierra Nevada and Oregon.” Illustrations of previously filled gullies are shown in Fig. 5. Thornthwaite et al. (1942, pp. 88-89), in discussing the evidence used by these investigators in identifying former erosion periods, argue that climatic change is unnecessary to explain them. Their inter- pretation is that upstream migration of discontinuous gullies within a valley would result in the same features as widespread erosion, includ- ing both terraces and buried channels. It is assumed that development Cu. 23] CORRECTIVE TREATMENT FOR GULLIES 425 of such gullies would cause waves of deposition to migrate slowly up- valley, thus leaving what appeared to be a continuous fill of the same age but which in reality would vary greatly in age in different reaches of the valley. Development of terraces would naturally be associated with this process. Likewise a subsequent gully, cut along the line of the buried discontinuous channel, would expose filled sections duplicat- ing those found at present. Although this interpretation is logical in some respects, to accept it would simply mean admitting that erosion in past periods differed from that of the present only in extent. Under these conditions it hardly seems consistent to attribute one to natural causes, the other to overgrazing. Antevs has also pointed out this inconsistency (Antevs, 1948, p. 12). From this brief mention of the investigations directed toward ap- praisal of the effect of climatic fluctuations on erosion, it should be apparent that much research is still needed before its full importance, as applicable to the erosion problem of the West, can be determined. CORRECTIVE TREATMENT FOR GULLIES No generally successful method of gully treatments which has proved capable of preventing both lateral and headward cutting has yet been devised. Even the task of arresting headcuts, where the treatment can be concentrated on a small area, has proved both difficult and ex- pensive, and many efforts have ended in failure. The greater and more complex problem of stopping bank cutting where miles of raw, vertical walls composed of highly erodible alluvium are exposed is still more difficult of solution, and the final task of reversing the present phenomena to the extent of substituting aggradation for degradation in these narrow channels will doubtless prove most difficult of all. To date, the program of treatment has followed two general lines: (1) reduction of grazing use, and (2) installation of control structures of various types. Unfortunately, in most instances, these were not combined, so the full effect of the two acting in unison cannot be evaluated. Naturally, if it is assumed that overgrazing is responsible for erosion, the obvious treatment is reduction or complete exclusion of livestock from eroding areas. As such action directly affects the livestock in- dustry, it raises questions of a social and political nature which have not yet been resolved. Livestock growers, although fully conscious of the erosion menace, are generally not convinced that their herds are completely responsible for it or that removal of them will effect a cure, and such reduction as has been accomplished has generally been 426 PETERSON. PROBLEM OF GULLYING [Cu. 23 with the idea of increasing forage rather than as a treatment for erosion. Locally tracts have been fenced or otherwise protected to demon- strate the advantages of reduced grazing on both increased forage pro- duction and decreased erosion. The results vary widely, as indicated by Fig. 4, which shows typical demonstration areas. In some localities the recovery has been excellent; in others, insignificant. It is apparent from this that availability of moisture, condition of the soils, and other ecological factors may have as strong an influence on vegetative recovery as utilization. Structural treatment was extensively used during the C.C.C. pro- gram, which lasted from the early 1930’s to 1942. These structures were generally of simple design and, since the program was aimed at work relief, a large part represented hand labor. Most of the early treatment areas were confined to tributaries of the major channels or to areas in which smaller gullies had developed, and only occasion- ally was an effort made to treat or control the larger features. The structures utilized were of wide variety and included water spreaders of many different types, check dams, contour furrows and terraces, diversion and training dikes, and small to moderately large storage and silt-detention reservoirs (Fig. 6). Essentially the basic aim of the program was to induce vegetative recovery which would in turn furnish ereater opportunity for percolation and thus reduce floods to safe and non-eroding rates. As there was no precedent for this type of treat- ment, much of it can be considered strictly experimental in nature, and failure in many instances to achieve the hoped-for results is not to be considered a reflection either on the designers or on the idea of land treatment. Future practices should benefit from these mistakes. Appraisal of the results of the treatment programs on both erosion and revegetation presents, in general, a discouraging outlook. Detailed examinations by personnel of the U. 8. Geological Survey of seven treated areas comprising a total of 5,600 acres located in the Upper Gila River in Arizona and New Mexico show that, of the 1,094 in- dividual structures, 375 or 30 percent have failed for various reasons, the most prevalent being undercutting and piping in the foundations and lack of maintenance. More significant than the failures, however, which doubtless could be eliminated by a higher standard of construc- tion and careful maintenance, is the lack of any discernible evidence of soil stabilization or vegetative recovery in areas controlled by struc- tures that have not failed. In all cases vegetative recovery has been classed as unnoticeable or no different from adjacent untreated areas. Moreover, the treatment has had little effect in preventing soil move- Cu. 23] CORRECTIVE TREATMENT FOR GULLIES 427 ment or in reducing runoff crests to any appreciable extent. Old gullies and erosion scars have not healed, and in many instances new ones have developed. It is obvious from these results that in the more arid areas, like that represented by the Upper Gila Basin, this type of treatment is of little value. The reasons for this are not so obvious, but apparently the flashy type of runoff produced by the higher intensity summer storms characteristic of this locality is greater than these small-dimensioned structures can cope with. Lack of vegetative recovery may be at- tributed to drought and erratic distribution of rainfall together with continuous grazing. Effect of the latter factor is hard to determine, but the fact that herds were not reduced to any appreciable extent re- flects the lack of confidence of the stock grower in the success of this type of program. A more successful demonstration is needed before this confidence can be restored. Recently the practice of water spreading has been greatly expanded, and in some instances it has been used on major gullies. Success of this treatment depends on complete diversion of flood flows onto a spreading area sufficiently large to absorb all the water, because, if any is allowed to return to the channel, new cutting immediately starts at the bank. One of the most successful installations of this type has been on the deep Polacca gully in the Hopi Reservation, Arizona, where water has been spread a distance of 15 miles below the point of diver- sion. Sand dunes, distributed along the bank, act as a natural barrier to prevent water from returning to the channel for part of the distance; training dikes parallel with and close to the channel have been con- structed for the same purpose along the remaining part. Similarly treated areas located upstream on the Polaceca Wash and in the tribu- tary Wepo Wash utilize a system of dikes strategically placed so as to direct the water away from the channel at needed intervals. Other water-spreading areas using the same system are located in parts of the Tularosa Valley in New Mexico and in the Alzada district in the Little Missouri River basin in Montana. Although the projects have not yet been in operation long enough to permit decisive evaluation, results to date appear to be promising particularly from the standpoint of range rehabilitation. Increase in forage as a result of water spreading has been of the order of several hundred percent in some reported instances—enough to justify con- siderable expenditures for such treatment. The final effect this treatment will have on the gully has yet to be demonstrated. Naturally, erosion in the channel downstream from the point of diversion is stopped, but, since the supply of sediment is 428 peTeRSON. PROBLEM OF GULLYING [Cu. 23 Fic. 6(a). One type of erosion structure, utilized in C.C.C. program. Net wire spreader has caught silt but has failed to induce vegetative recovery. Freeman Flat area near Safford, Arizona. Struc- ture built about 1938. Photo taken October 1948. Fic. 6(6). Partially breached rock spreader built by C.C.C. about 1938. San Simon Valley near Rodeo, New Mexico. Photo shows slight increase in vegetation on alluvial fill behind spreader. Photo taken October 1948. Fic. 6(c). Spillway and drop structure leading from spreader dike constructed by C.C.C. Steamboat demonstration area, Navajo Indian Reservation near Ganado, Arizona. Structures have pre- vented deepening of the channel but will soon fail unless repaired. Cu. 23] CORRECTIVE TREATMENT FOR GULLIES 429 Fic. 7(a). Bank protection along Chinle Wash. Many Farms area near Chinle, Arizona. Ground water is within a few feet of the surface. Fig. 7(6). Vigorous growth of cotton- wood trees and willows form bank pro- tection in Keams Canyon gully, Navajo Indian Reservation, near Keams Can- yon, Arizona. Gully is 50 feet deep. Stream is perennial in this reach. Fig. 7(c). Deposition induced by tetra- hedrons with willows and other types of shrubs. Chinle Wash near Chinle, Ari- zona. 430 PETERSON. PROBLEM OF GULLYING [CuH. 23 also cut off, no opportunity is afforded for the channel to fill except from bank caving or from wind deposition. The reservoir behind the diversion dam provides limited storage for sediment carried by the stream, but once this is filled the problem of sediment disposal again is met. As the very act of water spreading reduces the carrying capacity of the stream, an alluvial fan immediately begins to form at or near the diversion point. The tendency is for this fan to increase in height until a new gradient on which the stream can carry its sediment load is established. Unless a protective vegetative cover becomes established, gullying is likely to develop on this higher gradient. Evi- dence of this trend is already becoming apparent in some of the older installations. One disadvantage in water spreading is the loss of water occasioned by the spreading operation. Where supplies are ample for all needs, the point has no significance, and the right to spread will probably never be contested. On the other hand, however, should the practice of spreading become widespread in the Colorado River Basin with its limited supply, compared with existing and planned demands, it is quite easy to visualize strong objections being raised from downstream users. This precise situation developed on the Gila River in 1940, where irrigators in the San Carlos District supplied from the San Carlos Reservoir took the stand that the C.C.C. structures in the Safford and Duncan valleys were interfering with and reducing the normal runoff on the river. A special investigation authorized by the National Resources Planning Board * showed that the supposed re- ductions were more fancied than real, and thus not significant enough to warrant action, but the attitude of irrigators, in this instance, is in- dicative of the developing conflict for the use of water in western streams, no matter for what purpose, and any gully treatment or other type of conservation program must eventually, if not at present, give it consideration. Efforts to prevent bank cutting and meandering in gullies has been tried on a limited scale and with varying success in a number of localities. Possibly greatest progress has been achieved in the Navajo Indian Reservation, where several miles of the Chinle Wash and Keams Canyon Wash gullies have been stabilized against widening under the conditions of flow thus far experienced. The treatment consists es- sentially of tree and willow plantings, some started without pro- * Upper Gila River Report by the Technical Committee, National Resources Planning Board, Oct. 21, 1940, unpublished. Cu. 23] CORRECTIVE TREATMENT FOR GULLIES 431 tection, others in association with tetrahedrons or other types of revet- ments. Figure 7 shows views of this treatment. Although plantings offer certainly the cheapest and most promising field for this type of treatment, their use is limited to reaches where water is available for plant growth. This condition automatically ex- cludes the vast majority of gullies from treatment of this sort, because most are cut in dry valley floors where the surface flow is too infre- quent or ground water is too deep to support vegetation. Protection of banks in this type of gully remains for future solution. A permanent solution of the overall problem presented by gullying in the western valleys can be achieved only where the gullies have been refilled and the valley floors restored to a condition approaching that existing before cutting occurred. This involves substituting aggrada- tion for degradation within the channels themselves. In theory, as shown by Lobeck (1939, p. 168), the construction of a barrier to the height of the gully wall or slightly above should accomplish this, and, given time enough, perhaps it will. However, the results observed at numerous barriers, some of which have been installed for more than 20 years, furnish little promise for results within the foreseeable future. Surveys conducted by the Soil Conservation Service * and the Geo- logical Survey + to determine the gradient assumed by the fill above these barriers show a minimum of 0.07 percent or 3.7 feet per mile and a maximum of 0.76 percent or 40 feet per mile. These results apply to localities where the fill was composed of sand and silt only. In every case the gradient was less than 50 percent of the slope of the valley floor. Textural analyses of the sediments constituting the deposits failed to furnish any information about the disparity in the fill gradients, although, as might be expected, coarser sediments were gen- erally associated with the steeper slopes. Numerous observations indicate that vegetation forms one of the most effective traps for sediment, and, where it is dense enough, even fine-textured sediments will deposit on the floors of gullies. This action is demonstrated in the Chinle and Keams Canyon washes, previously mentioned, where as much as several feet of sediment has been de- posited within the fringe of trees and willows planted primarily for bank protection. Gradual filling of numerous gullies tributary to the Powder River near Broadus, Montana, has been noted during the past few years of favorable precipitation, during which dense growths of *G. A. Kaetz and L. R. Rich. Report of surveys to determine grades of depo- sition above silt and gravel barriers, U. 8. Soil Conservation Service manuscript, 1939. ¥ Information in U. 8S. Geological Survey file in Salt Lake City, Utah. 432 PETERSON. PROBLEM OF GULLYING (Cher, OF rose briers have become established on the gully floors. Figure 5c shows filling on the floor of a gully in northern New Mexico. Here a moderately dense growth of grass, weeds, and shrubs has been suffi- cient to trap silt and clay originating from an active headcut located a few hundred feet upstream. Deposition in this locality is of unusual significance in that it has occurred during a drought period when other gullies in the vicinity were actively degrading. Studies of this and other aggrading reaches will perhaps furnish information regarding methods that might be employed to increase the gradient of deposition behind barriers. NEEDED RESEARCH ON GULLIES Irrigation and stock raising constitute the backbone of industry in the western states. With both being jeopardized at a constantly in- creasing rate by gullying and other forms of erosion, research aimed at developing feasible methods of control is patently a necessity. To be successful this research will need to be directed into all elements of the problem, from geology and ecology to the practical phases of engineer- ing and range management. The writer makes no pretense of being in a position to outline such a research completely, but on the basis of considerable experience within the area, the following generalized subjects are suggested as fundamental to the problem. Because re-establishment of aggradation within the gullies and on the valley floors offers the only permanent cure for erosion, geologic studies should be undertaken to determine the conditions existing dur- ing previous periods of valley filling. This will involve all phases of the sediment history of the valley, from the initial weathering of the source rock through its transportation to final deposition and possible alteration since. Pertinent questions which need to be answered in- volve the factors that were most important in establishing former pro- files of stream equilibrium: grade; magnitude of flow; sediment load; character of the sediments; vegetative cover on the watershed as a whole or more particularly on areas where aggradation took place. As reasonably accurate answers to these questions are found, com- parisons will be possible with existing conditions under which streams are actively degrading. The same questions relative to factors in- fluencing such action will again need to be answered, but with addi- tional ones aimed at determining the possibilities of favorably altering these factors by engineering structures or land-management practices. Information on the history of the eroding valleys, particularly with regard to the recurrence of degradational and depositional cycles, needs Cu. 23] REFERENCES 433 to be enlarged. If it can be shown that these periods were not local in extent but affected widespread areas simultaneously, and that they were related to some common cause such as increasing aridity of climate, of which the present might be an example, a long step will have been taken in formulating a plan of erosion treatment. As the record of the changes must be read from the sediments themselves, new tech- niques and standards for recognizing and evaluating the evidence must be developed. One urgent need at present is a method for determining the relationship between climate and the character of sediments de- posited during a given period. Another is for criteria that can be used in interpreting climate by the evidence of alteration in the sedi- ments since deposition. Each is fundamental in finding the causes of past changes in valley history. Because vegetation is so closely allied to the erosion problem, the need for research, not only on its relation to sediment transportation but also on means of propagating vegetation, are evident. Among the pertinent questions are: What is the minimum density of various types of cover required to prevent erosion on the many different soils and slopes commonly found in western desert valleys? What minimum density is required to induce deposition under the same conditions? Is the normal precipitation in the area sufficient to maintain this den- sity with or without grazing use? If not, what feasible measures, if any, including both engineering structures and changes in land-manage- ment practices, can be taken to insure protection of the most vul- nerable localities? These suggestions touch only broad phases of the research needed be- fore the critical erosion problem of the West, of which gullying is the most prominent feature, can be approached with any hope of success- ful solution. The prosecution of this research will necessitate detailed inquiry into the many interrelated aspects of erosion, each of which will involve long, arduous, and expensive effort before the final answer is obtained. REFERENCES Albritton, C. C., Jr., and Bryan, K. (1939). Quaternary stratigraphy in the Davis Mountains, Trans-Pecos, Texas: Bull. Geol. Soc. Amer., vol. 50, pp. 1423-1474. Antevs, E. (1948). The Great Basin with emphasis on glacial and post-glacial times: Bull. Univ. Utah, vol. 38, No. 20, June 1948, p. 12. Bailey, R. W. (1935). Epicycles of erosion in the valleys of the Colorado Plateau province: Jour. Geol., vol. 43, pp. 337-355. (1937). Watershed symposium: Jour. Forestry, vol. 35, No. 11, Nov. 1937. 434 PETERSON. PROBLEM OF GULLYING Giza 22 Bailey, R. W., Forsling, C. L., and Bucraft, R. J. (1934). Floods and accelerated erosion in Northern Utah: U. S. Dept. Agr., Misc. Pub. 196, 21 pages. Bryan, K. (1925). Date of channel trenching (arroyo cutting) in the arid South- west: Science, n.s., vol. 62, pp. 3838-344. (1926). Recent deposits of Chaco Canyon, New Mexico, in relation to the life of pre-historic peoples of Pueblo Bonito: Jour. Washington Acad. Sciences, vol. 16, pp. 75-76. , and Post, G. M. (1937). Erosion and control of silt on the Rio Puerco, New Mexico: Unpublished manuscript. Cooperider, C. K., and Hendricks, B. A. (1937). Soil erosion and stream flow on range and forest lands of the upper Rio Grande watershed in relation to land resources and human welfare: U. S. Dept. Agr., Tech. Bull. 567. Cottam, W. P. (1947). Is Utah Sahara bound? Bull. Univ. Utah, vol. 37, No. 11, Feb. 1947. Forsling, C. L. (1931). A study of the influence of the herbaceous plant cover on surface run-off and soil erosion in relation to grazing on the Wasatch Plateau of Utah: U. S. Dept. Agr., Tech. Bull. 220. Gregory, H. E. (1917). Geology of the Navajo country: U. S. Geol. Survey, Prof. Paper 93, 161 pages. Hack, J. T. (1942). The changing environment of the Hopi Indians of Arizona: Reports of the Awatovi Expedition, Peabody Museum of American Archaeology and Ethnology, Harvard University, vol. 35, No. 1, 85 pages. Knechtel, M. M. (1937). Geology and ground water resources of the valley of Gila River and San Simon Creek, Graham County, Arizona: U.S. Geol. Survey, Water Supply Paper 796f, pp. 181-222. Lobeck, A. K. (1939). Geomorphology: McGraw-Hill Book Co., New York, 731 pages. National Resources Committee (1938). Regional Planning, Part VI. The Rio Grande joint investigation in the upper Rio Grande Basin in Colorado, New Mexico and Texas, 1936-37: National Resources Committee, Feb. 1938. Sayles, E. B., and Antevs, E. (1941). The Cochise Culture: Medallion Papers, No. 29, privately printed for Gila Pueblo, Globe, Arizona, June 1941. Schulman, E. (1945). Tree-ring hydrology of the Colorado River Basin: Univ. Ariz. Bull., vol. 16, No. 4; Laboratory of Tree-Ring Research, Bull. 2. Stabler, H. (1936). Discussion of a paper by J. C. Stevens, titled “The silt prob- lem”: Trans. Amer. Soc. Civ. Engrs., vol. 101, pp. 277-284. Stevens, J. C. (1936). Future of Lake Mead and Elephant Butte Reservoirs: Trans. Amer. Soc. Civ. Engrs., vol. 101, pp. 207-250. Thornthwaite, C., Sharpe, C. F. S., and Dosch, E. F. (1942). Climate and ac- celerated erosion in the arid and semi-arid Southwest with special reference to the Polacca Wash drainage basin, Arizona: U. 8. Dept. Agr., Tech. Bull. 808, May 1942. Woolley, R. R. (1946). Cloudburst floods in Utah 1850-1938: U. S. Geol. Survey, Water Supply Paper 994, 121 pages. Part 4 APPLICATIONS INVOLVING NATURE OF CONSTITUENTS i! vi CHAPTER 24 INFLUENCE OF SEDIMENTATION ON CONCRETE AGGREGATE RocEerR RHOADES Chief Geologist, U. S. Bureau of Reclamation Denver, Colorado Concrete aggregate, that is, the rock and sand that are mixed with cement to make concrete, may be natural sand and gravel, or quarried rock crushed to sand and gravel sizes, or any combination of crushed and natural material. Sand and gravel comprise slightly less than two thirds of the concrete aggregate produced annually in the United States. Quarried, sedimentary rock, predominantly limestone, will constitute about one fifth of the total production. The processes of sedimentation always influence and frequently control the develop- ment of the various properties that combine to determine the use- fulness of either sand and gravel or sedimentary ledge rock as aggregate for concrete. The remaining annual production is represented by non-sedimentary types, including crushed igneous and metamorphic rock. The latter category of aggregate is in no way influenced by processes of sedi- mentation and is therefore ignored in this presentation. Concrete in service is subjected to various external conditions which may impose stress, strain, saturation, desiccation, chemical attack by aggressive solutions, abrasion, impact, and temperature changes con- ducive to expansion and contraction. Internal conditions inherent in the chemistry of the cement itself subject the constituents of concrete to additional rigors. In the first place, cement is an aggressively al- kaline material capable of greater or lesser chemical interaction with any aggregate material. In the second place, the hydration involved in the setting of cement is accompanied by heat generation which, in large masses of concrete from which heat is dissipated very slowly, will cause significant expansion with attendant stresses and strains, followed ultimately by a contraction, with its own complement of stresses and strains, when cooling finally occurs. 437 438 RHOADES. CONCRETE AGGREGATE [CuH. 24 The different constituents of concrete are chemically and physically heterogeneous. Cements differ widely in composition and in physical character and behavior, and the rock and sand comprising the ag- eregate—excluding certain monomineralic types, such as limestone or quartz—will generally contain a variety of minerals, each with its own physical and chemical characteristics and with its own unique re- sponse to environmental conditions. These separate and diverse constituents of concrete react individu- ally to the internal and external influences to which the concrete is subjected, and they interact with each other and thus determine its durability and serviceability and its appropriateness and effectiveness for its intended use. Inasmuch as aggregates normally comprise about 75 percent of concrete, their physical and chemical properties pro- foundly influence the overall character and behavior of any concrete mass. Ideally, the selection of concrete aggregates should be based on a detailed appraisal of their chemical and physical properties, so that materials may be chosen that will react and interact harmoniously and compatibly and impart to the concrete properties consistent with its purpose. Deleterious reactions or incompatible interactions result in- evitably in premature deterioration or ineffectual service. Practically, it is not now possible in the manufacture of concrete to pay full deference to these considerations. In the first place, the variables involved are numerous and intricately interrelated, and the isolation of any separate variable for individual study, although prac- tical on a research basis, is impractical as a routine, and suitable laboratory tests have not been standardized for general application. In the second place, in the present state of concrete technology there are many factors of uncertain and obscure significance, and various current hypotheses lack certain confirmation. Furthermore, the use- fulness of concrete as a construction material depends in part on its relatively low cost. Economic considerations, therefore, prohibit costly beneficiation or the arbitrary rejection of a cheap local material and the costly importation of aggregate from a more distant source unless the advantages to be gained are assured and commensurate with the added expense. Current research is rapidly resolving the remaining technical un- certainties In aggregate selection; and the most enlightened current practice involves the formulation and progressive revision of specifica- tions and the development of acceptance tests that will exclude in- appropriate materials unequivocally, no matter how cheaply they may Cu. 24] PROPERTIES OF CONCRETE AGGREGATE 439 be obtained. But additional research will be required to define entirely the properties that are admissible and inadmissible, and to develop simple tests which will discriminate conclusively between acceptable and unacceptable aggregate materials. PROPERTIES OF CONCRETE AGGREGATE One important set of properties that influence the quality or suit- ability of concrete aggregate includes: strength, hardness, compres- sibility, durability, elasticity, particle shape, surface texture, specific gravity, porosity, volume change with varying thermal or moisture conditions, presence or absence of deleterious impurities or coatings, mineralogic composition, and a variety of chemical properties related to the stability of the aggregates in the aggressively alkaline environ- ment of concrete or their resistance to the agencies of natural weather- ing. These various properties apply to individual particles of aggre- gate. They may not always be equally important, their importance depending on the use, purpose, and environment for which the concrete is designed, but each influences the serviceability, durability, or cost of concrete for general or special purposes; and together they con- tribute fundamentally to the behavior of a concrete mass under dif- ferent service conditions. For the benefit of readers unfamiliar with concrete technology these properties are summarized briefly in the ap- pendix of this chapter from the standpoint of their significance to concrete. More complete discussion and extensive bibliographies may be found in American Society for Testing Materials (1948a, b, ec), U.S. Bureau of Reclamation (1949), Rhoades and Mielenz (1946), Blanks (1949). Another set of properties, related not to individual particles but to the overall assemblage of particles, includes minimum size, maximum size, size gradation, and degree of mineralogic or lithologic diversity. (See appendix of this chapter and ASTM, 1948b; U. S. Bureau of Reclamation, 1949; Twenhofel, 1932.) A third set of properties, applying to whole deposits of aggregate, that influence the quantity of material available, the optimum method of excavation, the kind of beneficiation required, and like considera- tions, defines the adequacy of any aggregate source and the cost of ag- gregate production. This category of factors includes the area, depth, variability or stratification, overburden, topography, and depth to ground water (ASTM, 1948c; U. S. Bureau of Reclamation, 1949). 440 rHoADES. CONCRETE AGGREGATE [Cu. 24 NATURAL SAND AND GRAVEL AGGREGATES The properties of individual particles of sand and gravel (for ex- ample, surface texture) or of assemblages of particles (for example, grading), which determine suitability as aggregate, as well as the characteristics of deposits as a whole which control feasibility and economy of production, are all influenced by sedimentation and related processes; indeed, the character of a natural sand and gravel usually results almost wholly from (a) the kind of original or source material from which it was derived and (b) the modifying influence of its sub- sequent experiences as it was transported and deposited (Twenhofel, 1932). PROCESSES OF SEDIMENTATION INFLUENCING SAND AND GRAVEL AGGRE- GATES Sand and gravel, potentially useful as concrete aggregate, may be transported and deposited by water, wind, ice, or gravity, or by any combination of these agencies. Inasmuch as different mechanisms are involved in transportation and deposition by these various agencies, each will impart to the material involved certain characteristic and recognizable properties. However, although the transportation and deposition may differ both in mechanism and rigor, all these agencies influence the material transported and deposited through some com- bination of the following processes: (1) impact, abrasion, or crushing; (2) sorting; (3) weathering, leaching, and chemical reaction. Weathering, leaching, and chemical reaction are not necessary ac- companiments to sedimentation, but they frequently are inextricably associated or occur concurrently with it, and in such cases their con- sideration is essential to any complete understanding of the rock prop- erties resulting as end products of the sedimentation process. These actions occur conjointly, and the final properties imparted to a sand and gravel will be the summation of their combined effect; but certain of the resulting properties are most influenced by one or another of these actions, as indicated in the following examples. Impact, abrasion, or crushing. This action, for example, is primarily responsible for modifying the initial shape of the particles in the direc- tion of increased roundness; for reduction in the size of all particles, and sometimes in the elimination of all particles above a certain size; for the elimination of soft constituents that may have been present in the initial material; or for the roughening of surface texture. Cu. 24] SIGNIFICANCE OF “SOURCE MATERIAL” 441 Sorting incident to transportation and deposition. This action con- trols the grading of the sand and gravel, subject, of course, to the gradation of sizes furnished at the source, and it usually governs the maximum and minimum sizes of particles occurring in the final deposit. This action also influences the variability of the deposit, its stratifica- tion, and the thickness and kind of overburden. Weathering, leaching, and chemical reaction. Through this action the strength or durability of the particles may be decreased, or the absorption increased through the leaching of soluble constituents or the chemical decomposition of vulnerable ingredients. Undesirable particles, if soluble or reactive, may be eliminated through solution, leaching, or chemical reaction; or the same processes may cause the formation of coatings about aggregate particles which may be in- nocuous if they are hard, chemically stable, and firmly adherent, but which may be deleterious if soft, chemically reactive, or loosely bonded. By a further extension of the same process particles may be cemented in a manner to hinder the economical exploitation of a deposit. Clay or other impurities may be formed as a consequence of the chemical de- composition of certain minerals (for example, feldspar); such clay may remain within the individual particles (contributing to volume change through wetting and drying) or accumulate as a separate com- ponent of the deposit and contribute to an excess in the very fine size orades. SIGNIFICANCE OF “Source MATERIAL” The final properties of a sediment, including those properties im- portant in concrete aggregate, are strongly influenced by the kind and condition of the source material. Different source materials will be affected differently by the various sedimentation processes, and a sedi- ment usually exhibits vestiges of initial properties—inherited char- acteristics, variously preserved and variously modified—as well as those superimposed during sedimentation. The properties contributed by the parent rock are those related to the mineralogic and petrographic composition of the rock or its initial physical and chemical condition (ASTM, 1948a; Bureau of Reclama- tion, 1949; Rhoades and Mielenz, 1946). For instance, the strength or the elasticity of an individual particle of sand or gravel is primarily determined by its mineralogy and the textural relationships of the component minerals. The initial mineralogy and texture may be altered slightly or profoundly during transportation and deposition, but they will be preserved in some degree unless the particle is wholly decomposed or disintegrated; moreover, the initial mineralogy and 442 RHOADES. CONCRETE AGGREGATE [Cu. 24 texture largely define the reaction and resistance to the various modi- fying processes and thus influence the kind and extent of the modifica- tion that will occur. For example, the shape of a sand or gravel particle is in part the result of attrition during transportation, but it is controlled also by the initial shape which in turn results from the spacing and patterns of joints and fractures in the parent rock. Similarly, the maximum size of particle in a gravel deposit will depend in part on attrition during transportation but can never exceed the dimensions of the particle originally supplied at the source, again a function of the spacing of joints or fractures. The mineralogy and petrography of the parent rock will also determine the initial clay content, which may be related to absorption and volume change under conditions of wetting and dry- ing. The initial porosity of the rock is related to its internal texture. Surface texture may be related directly to the mineralogy and texture of the parent rock, if governed by cleavage planes, as in feldspar, or by the fracture form characteristic of certain minerals as, for instance, the conchoidal fracture of quartz, both of these surface textures, unless subsequently rendered rugose by leaching or attrition, being too smooth for development of optimum bond with cement. The chemical characteristics of aggregate, although in part related to subsequent leaching, solution, weathering, and the like, which may attend the processes of sedimentation, are also in part controlled by the initial composition of the source material. Thus the mineralogy of the parent rock will dictate whether or not weakening, softening, in- creased porosity or absorption, or the development of secondary clay will be caused by these processes, and to what degree. The most important chemical reaction which occurs between agegre- gate and cement in concrete, the so-called alkali-aggregate reaction, is usually the direct result of the mineralogy or petrography of the parent rock (ASTM, 1948a; Rhoades and Mielenz, 1946; McConnell et al., 1948). The rocks and minerals susceptible to this type of reaction are the glassy volcanic rocks of acid to intermediate composition (rhyolites through andesites), such silica minerals as tridymite, opal, and chal- cedony (and cherty rocks of which they are constituents), and prob- ably a hydromica which occurs in some phyllitic rocks. These rock and mineral types, when exposed to the attack of the excess alkalies contained in a high-alkali cement, yield a silica gel that will imbibe water by osmosis and swell with the development of large, expansive pressures. Concrete so affected expands and cracks in an unsightly manner and not infrequently becomes unsafe or unserviceable. How- ever, the initial reactive potentialities of an aggregate may be lessened Cu. 24] SIGNIFICANCE OF “SOURCE MATERIAL” 443 by the processes of transportation and deposition, as, for example, through mixing and dilution with other rock types that are non-re- active; or a rock originally innocuous may become reactive by the ad- dition of secondary opal or chalcedony as coatings or intergranular cement during sedimentation. Many different geological formations may contribute to a sediment- ary deposit and supply a large variety of rock types representing many different physical and chemical “source” conditions. Table 1 indicates the complex composition of a gravel deposit on the upper Missouri River. Such mixing of materials is most pronounced and complicated in the sediments of rivers draining large basins. Although a sand and eravel deposit occurring near the head of a river may contain a limited number of rock types, the complexity of the mineral and petrographic assemblage will increase downstream as each successive tributary con- tributes the rock types available within its own watershed. Con- tributions of sound rock by a tributary to a stream whose sand and gravel load is mainly composed of inferior material will naturally be beneficial; but one major tributary contributing a large amount of inferior or deleterious material may render deposits farther down- stream less suitable as concrete aggregate (Spain and Rose, 1937). This relationship is well illustrated by the Colorado River between Hoover Dam and Parker Dam. Sand and gravel obtained upstream in the vicinity of Hoover Dam contain a complex assemblage of rock types, some of which are susceptible to alkali-aggregate reaction. However, the quantity of reactive types at this point is small, and Hoover Dam concrete containing this aggregate exhibits no distress from alkali-aggregate reaction after 20 years of service. The Colorado River below Hoover Dam traverses terrain that is predominantly vol- canic, and its sand and gravel become progressively enriched in deleteri- ous volcanic rocks. At Davis Dam, 67 miles downstream, the sand and gravel are sufficiently reactive that special measures were re- quired (use of low-alkali cement and pozzolanic admixtures) to fore- stall alkali reaction in the concrete. Farther downstream (156 miles below Hoover Dam) Bill Williams River contributes copious amounts of highly reactive aggregates—mainly andesites and rhyolites. Parker, Gene Wash, and Copper Basin dams built in this vicinity, and using local aggregate and high-alkali cement, exhibited extreme evidences of alkali reaction within 2 years after their completion. Any initial characteristic of a source material may be either im- proved or impaired by the rigors of sedimentary transportation and deposition. For example, as previously noted, improvement can result through the elimination of soft particles or the rounding of particles [Cne24 “queUIed TTR: FTE-YHTY YAM snowsyepaq "BOlB OWUBS BY} UT SefoY [Mp 10y}0 wou sojdures Joawis ul quoserd aire sad yoor osoyy, » puoq 100d ‘snoiod ‘pa1eA00-yx]eyO CONCRETE AGGREGATE RHOADES. 444 Tenotioee aC 100d * * poreyyeem Aydoeq qT ShOLezo[9q IIB i] olUOpss[Beyo ‘Usyo1q ‘petoyyvem Ajo,e1opour “WITT O'T sage persyyee j SnOle,a;9q A1I0{OBISIVBS dUOpeo[eyo ‘poureis-ouy ‘prez 9°¢ B'S S109 j Shor1ez9;eq, 100g peinzovsy ‘snosod ‘a[qeitqT sag an pereyyzem ATdeaq zy SNOI9z9T9q meq Assejs ‘poreyyveM Ajoyeilopoul ‘WITT Z9 peroyyeo Ay zSsnoweyajeq | Ar1ojoejstyEg Assejs ‘prey ‘poiey}vem Ajo}elepoul 04 AYYSITG ov soy10Bqy T SNOI19}9;9q 100g snoiod A1sA ‘e[qeliq 8°0 poreyyvem A[dooq, { SNOI1e}919q Ie snoiod ApJysI[s ‘pe1eyyveM Ajo} e1epoul ‘WITT 10 ZT per9y7B9 AA z sSnowezejaq | <--~---~-- Gravel---—-----+> Fic. 1. Comparison of gradings of natural and processed aggregates with ASTM specification limits. (Aggregate from North Platte River investigated for Kortes Dam, Wyoming.) quired must be processed and discarded in order to produce the requi- site quantity of coarser material. Attrition has had the effect in the Missouri Basin states, Kansas, Nebraska, and South Dakota, of eliminating entirely the coarse-gravel fraction of the river gravels (Scholer and Gibson). These gravels, derived initially from the Rocky Mountain area to the west, contain few, if any, particles over 1% inches in size at any location east of the Colorado state line, although to the west coarser materials become in- creasingly abundant, and near the front of the mountains the gravels are fully graded. River transportation normally causes a concentration of hard and firm particles through the selective elimination of softer materials by attrition. However, the strength and hardness of the particles are related closely to the source material from which the sediment was Cu. 24] RIVER AND STREAM DEPOSITS 447 derived and to the distance the material may have been transported. For example, the comparatively hard and firm gravels of the Colorado River above the mouth of the tributary Dolores River, in the vicinity of Dewey Dam site (Utah) contrast sharply with the gravels down- stream from the mouth of the Dolores, because of the large quantity of soft sandstone contributed to the main stream by that tributary. Farther downstream the soft Dolores sandstone is progressively re- duced in quantity. Improvement of surface texture also results from the attrition, im- pact, and abrasion incident to river transportation, so that river gravels will normally possess rugose surface textures conducive to good bond with cement. However, if large quantities of the pebbles that are furnished to the stream possess potential planes of breakage, as along cleavages, the impact and jostling during transportation may promote the division of larger pieces into smaller ones with comparatively smooth surfaces. It must not be overlooked that the environment of river transporta- tion and deposition is conducive to leaching of soft or soluble materials or to chemical reaction with materials susceptible to decomposition. These processes can aid the processes of mechanical attrition in promot- ing the selective concentration of hard and firm particles, but in other cases they can cause softening or weakening of particles that would be relatively invulnerable to mechanical attack alone. The field occurrence of river-channel deposits on the upper Missouri River is illustrated in Fig. 2 (see also the petrographic analysis, Table ae River terraces. River-terrace deposits possess the general character- istics of river-channel deposits and are widely used for concrete aggre- gate. Terrace deposits, in many places, are exploited more economi- cally than channel deposits, since, because of their higher topographic positions, they are normally not so subject to the production diffi- culties attending shallow ground water. In some areas, as along cer- tain reaches of the Colorado River in Texas, the rivers themselves are not now transporting gravelly material, but gravels may, nonetheless, be found in terraces mantling the highlands where they were deposited by gravel-bearing streams of earlier times. Terrace deposits commonly, although not universally, differ from channel deposits in exhibiting a greater prevalence of secondary coat- ing or cementation of particles. Coatings resulting from the action of ground water, or waters infiltrating from the surface, may permeate the entire thickness of a deposit, but commonly (particularly in arid regions) they are associated with transpiration of moisture toward the 448 RHOADES. CONCRETE AGGREGATE [Cu. 24 surface and are most prevalent in the upper few feet. Concentrated deposition in the upper layers of a deposit may cause the formation of cemented layers that hinder excavation. The most common coatings are calcium carbonate and, inasmuch as this material is chemically innocuous, they may in no wise impair the quality of aggregate if they are firm and tightly adherent to the Mixed terrace, alluvial fan nd Hood pigin deposits. Fic. 2. Deposits of natural sand and gravel on the Missouri River near Canyon Ferry Dam site, Montana. particles. Siliceous coatings—opaline or chaleedonic—are chemically reactive with alkalies in cement, and, no matter how firm and adherent, they will be detrimental to concrete unless used with cements low in alkalies (ASTM, 1948a; McConnell et al., 1948). Coatings composed of iron oxide, manganese oxide, or gypsum are not uncommon. Gyp- sum coatings, being soluble and pulverulent, are detrimental, but they are often removed by normal washing and screening. In many places, as at Davis Dam (Colorado River, Arizona-Nevada) the upper several feet of a terrace deposit must be removed and discarded because of del- eterious coatings or cemented layers. Terraces frequently exhibit postdepositional weathering. Extreme weathering may render a terrace deposit entirely unsuitable for use Cu. 24] RIVER AND STREAM DEPOSITS 449 as aggregate through decomposition of the individual particles. Ter- races in some parts of the Columbia River Basin, for example, were formed at two different periods; in some localities the older and younger terrace materials appear to occur in continuous sequence and are not distinguishable through superficial observation, nor is the con- tact between them clearly discernible. In the investigation of these materials to determine their suitability for concrete, anomalous and conflicting results arose from a failure to distinguish between the older and younger materials. Subsequent studies disclosed that durable concrete may be made of the firm and hard basaltic particles com- prising the younger terrace deposit, whereas the older, underlying material, which is similar in lithology and appearance, is composed of incipiently decomposed particles unsuitable for use in concrete. Typi- cal gravel terraces are shown in Fig. 2. Flood-plain deposits. Flood-plain deposits are characterized by variable and heterogeneous gradations. True flood-plain deposits caused by the periodic overflow of a stream from its normal channel are usually deficient in material larger than sand and hence have only a limited usefulness as concrete aggregate. Frequently such material mantles the upper surfaces of flood plains but is underlain by hetero- geneous assortments of sand and gravel deposited in former times as the river meandered randomwise from side to side. When favorable size gradations are found in flood-plain deposits, they are entirely suitable for aggregate and possess the same general character as chan- nel or terrace deposits. The flood-plain deposits shown in Fig. 2 are mixed with terrace and alluvial-fan deposits. They occur more typi- cally in the broad valleys of more mature rivers. Alluvial fans and cones. The principal characteristics of alluvial fan or cone deposits, particularly in arid regions, are heterogeneity and angularity. Since they were formed by successive torrential down- pours with brief but violent runoff, and since the volume and velocity of runoff differed from storm to storm, the deposits are usually rudely layered into zones ranging from very fine material to very coarse; and the transportation, although frequently violent, is usually too quick and too short for the development of rounded particles. Alluvial fans and cones deposited by intermittent streams may contain material suitable for the production of concrete but usually require elaborate beneficiation, including screening, and possibly other measures such as blending with imported materials to fortify certain deficient size erades. The alluvial fans shown in Fig. 2 were deposited under semi-humid 450 RHOADES. CONCRETE AGGREGATE [Cu. 24 conditions. A more striking series of alluvial cones and fans, typical of arid regions, is illustrated in Fig. 3. Alluvial fans similar to those shown in Fig. 3 occur along the Boulder-Kingman (Arizona) highway, and they were investigated in connection with postconstruction activities at Hoover Dam. These fans disclosed the following characteristics: (a) Individual particles Fic 3. Alluvial cones and fans formed under arid conditions near Muddy River, Nevada. are angular, having been transported quickly and only a short distance from the source material. (b) The deposit is composed of a limited number of rock types—those rock types that occur on the ridge at the head of the dry washes. (c) All sizes, including the sands, are es- sentially of the same composition, there having been no selective con- centration of the more durable types. (The softer materials occur in proportion approaching their original proportion in the parent rock.) (d) The deposit is heterogeneous as to grading within layers and from layer to layer. It was determined that this material would be suitable as concrete aggregate if locations and levels were chosen to avoid extremely coarse Cu. 24] WIND-BLOWN SAND 451 or extremely fine gradations, and if screening and washing were done thoroughly and with care. But the resulting concrete would be harsh and difficultly workable because of the angular nature of the individual particles. Tautus Deposits Talus deposits formed by gravity at the bottoms of steep slopes are usually composed of a single rock type or a limited number of rock types—the types that occur on the upper part of the slopes—which have accumulated as particles ranging from very fine to extremely large, of angular and irregular shape, and with a minimum of orderly distribution of the various sizes. Such deposits represent, in effect, simply the original parent rock broken into fragments but otherwise deposited with little of the rounding, sorting, or segregating actions that characterize to a greater or lesser extent most of the other agencies of sedimentary transportation. If the parent rock possesses characteristics suitable as concrete ag- sregate, the talus deposits will usually be suitable but will require the crushing and other beneficiation normally applied to quarried rock. Inasmuch as the rock has been fragmented by natural processes, it may on occasion be more economical to produce aggregate from a talus accumulation than from the parent-rock outcrop, from which the frag- ments would have to be produced artificially by blasting. In some cases the spacing and pattern of joints in the parent rock have con- trolled the maximum size of fragment that will form in a talus deposit, and in this way economical production may be facilitated through simplification of the plant installation required for crushing and re- lated processing. Four talus deposits have been recently investigated as possible sources of concrete aggregate: a granodiorite accumulation on the slopes of Ragged Mountain north of Spring Creek Dam site (Colo- rado); andesite porphyry on the slopes at Platoro Dam site (Colo- rado); and rhyolite from Sundance Mountain and phonolite porphyry- from Missouri Butte (Fig. 4), both in the vicinity of Keyhole Dam site (Wyoming). It was found that these materials would make satis- factory concrete but that, because of the expensive processing re- quired, the importation of river gravel and sand from other areas would be more economical. WinpD-BLown SAND The extreme rigor of transportation by wind frequently results in the preservation of only the hardest and most durable rock and mineral 452 RHOADES. CONCRETE AGGREGATE [Cu. 24 types. Hence wind-blown sands, by and large, are composed pre- dominantly of quartz. The grains are well-rounded and concentrated in a few size grades which reflect the average transporting capacity of the prevailing winds of the area. Such sands are useful in con- crete aggregate only for “blending” to augment other aggregates which are deficient in the finer size grades. A typical wind-blown sand, oc- curring near Pot Holes Dam site, Columbia Basin (Washington) is shown in Fig. 5. The gradation and composition of a typical wind-blown sand from near Kortes Dam (Wyoming) is shown in Table 2, sample A. A de- TABLE 2 GRAIN SIZE OF WIND-BLOWN SANDS A B C SETE Diameter pize Cumu- Cumu- Cumu- Retained ; Retained : Retained if lative lative lative microns % % q % q q 16 1,190 0 0 0 0 (0) 0 30 590 0 0 0 0 5 5 50 300 4 4. 9 9 6 11 100 150 50 54 76 85 9 20 Pan tan 46 100 15 100 80 100 A. Kortes Dam area, Wyoming. Mainly quartz and feldspar; some amphibole, mica, chlorite, chert, limestone, and siltstone. B. Columbia Basin, Washington. Derived from lavas. Composition is plagio- clase feldspar, with smaller amounts of basaltic glass, basalt, quartz, olivine, acidic glass, rhyolite, and opal. C. Pasco Pumping Plant area, Washington, near sample B. Derived from marine and glacial sediments that mantle adjacent Pasco slope. Composition is mainly quartz, feldspar, and muscovite. posit of wind-blown sand occurring in the Columbia Basin (Washing- ton) is noteworthy by reason of its mineralogic and petrographic com- position (sample B). Its heterogeneous ccmposition results from a predominantly local derivation from the Columbia Basin plateau lavas which comprise the prevailing bedrock. However, sample C, another blowsand from the same general area (near the Pasco Pumping Plant, Columbia Basin, Washington) exhibits quite a different composition, Cu. 24] WIND-BLOWN SAND 453 Fic. 4. Talus deposit at Missouri Butte, Wyoming. Fic. 5. Wind-blown sand near Pot Holes Dam, Columbia Basin, Washington. 454 RHOADES. CONCRETE AGGREGATE [Cu. 24 being derived from the marine and glacial sediments that mantle the adjacent Pasco slope. Deposits of voleanic ash and pumicite may be considered a special category of wind-blown sand. Glassy, volcanic tuff is useful as poz- zolanic admixture with cement to impart certain desirable character- istics to concrete (Lea and Desch, 1935; Lea, 1938). For this purpose the materials should possess a fineness comparable to that of cement (as a rough index, at least 90 percent of the material should pass a 325-mesh sieve). Some pumicites are prevailingly of this fine size as, for instance, the extensive accumulations of pumicite in the vicinity of Friant, California, from which material was obtained for use in Friant Dam. If a deposit contains material prevailingly coarser than this size or if it is not readily pulverulent, it must be pulverized arti- ficially; the additional cost of such processing is a frequent deterrent to the use of any given deposit of pumiceous material. Pumice and scoria in coarser sizes are coming into increasing use as ageregate for the construction of special-purpose concrete, as in “cin- der-block” construction or in the fabrication of acoustical or insulating concrete (Price and Cordon, 1949). GLACIAL DEPosITs Glacial deposits in the form of moraines, or tills, are typically char- acterized by extremely heterogeneous grading, frequently with large boulders and fine rock flour intimately admixed. Individual particles may be smooth or even somewhat rounded, but angular or subangular particle shapes are common. Glacial deposits normally exhibit some selective concentration of the harder rock types, but softer materials frequently are preserved through incorporation within the ice mass it- self, where they are shielded from the intense abrasion and crushing that generally characterize glacial transportation. The glacial envi- ronment discourages extreme chemical alteration and thus inhibits the selective removal of materials which are soluble or susceptible to chemical decomposition. Thus, whereas firm particles may be pre- served from the weathering and decomposition to which they might be susceptible in other sedimentary environments, soluble or chemically unstable materials may be preserved to the detriment of the deposit for use as concrete aggregate. Certain areas of the United States are extensively mantled with glacial deposits and must rely on this source for concrete aggregate (ASTM, 1948d). However, such deposits usually require expensive beneficiation. Elaborate screening may be required to correct the grading, and the necessity of wasting large amounts of oversized and Cu. 24] MARINE SANDS AND GRAVELS 455 undersized material frequently involves the handling of much excess material for the production of a given quantity of aggregate. In deposits that contain large boulders the expense of processing is fur- ther increased. Where crushing of the oversized material is necessary to augment finer size grades, the expense is again increased. Large accumulations of glacial erratic boulders may sometimes be used for concrete aggregate. In such cases the boulders are collected, crushed and screened in the same manner as though they were quar- ried rock. This procedure is economically feasible only in exceptional cases. In contrast to true glacial deposits, glaciofluvial deposits are widely used as concrete aggregate. Only moderate reworking of glacial mate- rials by fluvial agencies may suffice to form deposits comparable to true river sands and gravels. MarINE SANDS AND GRAVELS Marine sands and gravels are typically characterized by hard, firm particles as a consequence of the extreme natural selection of the durable and elimination of the non-durable materials consequent to the winnowing action by waves and currents on a beach. The particles usually are very well-rounded and restricted to a narrow range of size grades. Marine sands and gravels are not widely used as concrete aggregate, but where available in sufficient thickness and extent for economical production and where proper gradations may be obtained by no more elaborate processing than is economically feasible, ag- gregate of excellent quality can be produced. These conditions may be found in thick terraces lying above the level of the present shore. Such deposits not uncommonly exhibit vertical variations of grain size from layer to layer, but selective excavation may be feasible by which is removed material which possesses an approximation of the grada- tion required, thus minimizing the amount of artificial processing that must be done. Beach sands frequently exhibit characteristics similar to wind-blown sand, or indeed may owe their origin as much or more to the action of the wind as to the action of ocean waves and currents. These mate- rials, like wind-blown sand, discussed previously, are suitable for use as blending sands in combination with other aggregates. Sands and gravels occurring on beaches may be impregnated with deleterious salts and thus require rigorous washing. Such salts may have been entirely removed from older terrace deposits through leach- ing by surface and subsurface waters. 456 rHoADES. CONCRETE AGGREGATE [CxH. 24 Coral and algal reef materials, such as occur profusely in the South Pacific, have had extensive local use in the production of concrete, but the special considerations and techniques involved in this class of con- crete construction are beyond the scope of this discussion (Rasmusson, 1946). RESIDUAL DEPositTs Residual materials remaining in an area as the end products of weathering and erosion are rarely usable as concrete aggregate. Al- though they owe their existence to their durability and therefore are frequently hard and firm, they are normally ungraded in size and usu- ally limited in quantity. Characteristically, they are admixed with substances unsuitable for the production of concrete aggregate. Boulders of chalcedonic siltstone occurring over a considerable area in southwestern North Dakota and northwestern South Dakota furnish an example of the occasional usefulness of residual material. It was proposed that concrete aggregate for construction in that area be pro- duced through the collection, crushing, and screening of these materials. Although laboratory tests indicated that satisfactory and durable con- crete could be made with aggregates thus produced, excessive costs would be involved. The use of such material would be considered only in areas that lack material more suitable or more easily processed. QUARRIED SEDIMENTARY ROCK The suitability of a rock ledge for the production of concrete ag- geregate through quarrying and crushing depends on three general factors: (a) the intrinsic quality, physical and chemical, of the rock; (6) the uniformity of the available working face; and (c) textural and structural characteristics that influence the “crushing characteristics” of the rock (ideally, a rock should crush to firm particles, roughly equidimensional in shape, with a minimum of “powdering” or frag- mentation into extremely fine sizes) (ASTM, 1948a). The physical and chemical characteristics of quarried aggregate are those of the “source material” before the modifications—beneficial or harmful—which the processes of transportation and deposition would impose if the same material were to be transformed naturally into sand and gravel. Any beneficiation which a natural rock exposure may require to render it suitable as concrete aggregate must be applied artificially through crushing, screening, and washing. Different kinds of rocks will respond differently to such processing. Ledge rocks are commonly variable in hardness because of localized Cu. 24] CONCRETE AGGREGATE RESEARCH 457 leaching or cementation. Dolomites frequently are fractured, lime- stones frequently are leached and rendered porous along channelways penetrated by ground water. Concretionary structures representing local concentrations of mineral material deposited interstitially in sandstones, siltstones, or shales, or by replacement in limestone, occur sporadically in sedimentary formations. Limestones may contain in- terstitial clay, which renders the material susceptible to volume changes, or may contain chert lenses that are physically unsound or chemically reactive with the alkalies in cement, as in many localities in the southeastern United States. The rock at a given quarry site may be massive and uniform in com- position, or it may be stratified and composed of more than one rock type. Individual strata in a quarry may be hard or soft, porous or dense, jointed or unfractured. Limestones may be interbedded with shales, sandstones with siltstones, and quartzites with dolomites. Strat- ification may be vague or pronounced, the layers thick or thin. The thickness of layers may control the size and shape of the aggregate pro- duced: rocks with thin stratification or lamellar structure will crush into planar or elongated pieces. Limestone may possess good crushing characteristics if dense and massive, but may produce particles of poor shape or an excess of fines if it has a lamellar “grain” or is closely fractured or highly clayey. Similarly sandstone may be hard or soft and thus crush well or poorly depending on the kind and amount of cement that binds the grains together. These factors of petrography, texture, and structure are inherent in the sedimentary origins of rocks; the sedimentary processes by which a rock has formed will determine its intrinsic physical or chemical quality, the uniformity of the product which may be produced, and the practicability and ease with which it may be crushed and other- wise processed into aggregate suitable for concrete. CONCRETE AGGREGATE RESEARCH Past and current research on concrete aggregate has been a Joint venture by engineers, chemists, physicists, and, recently, geologists. In the future, even more than in the past, research in concrete aggregate will call for the joint preoccupation of scientists and engineers of various specializations, with geologists playing an increasingly im- portant role through analysis and interpretation of rock and mineral properties. Concrete aggregates have been appraised and selected in the past primarily through the application of a series of empirical tests which 458 RHOADES. CONCRETE AGGREGATE [Cu. 24 measure, for instance, their specific gravity, absorption, abrasive re- sistance, soundness under freezing and thawing (or under the more rigorous and accelerated test with sodium or magnesium sulphate). Such tests have been correlated with field service over many years, and on the basis of this long experience they serve as guide lines for estab- lishing specification limits and for the selection and rejection of pro- posed aggregate materials. Today there is an increasing recognition by concrete technologists that these empirical tests, however useful they may have been for the practical purposes of the past, do not measure specifically the funda- mental properties that define the “concrete-making” potentialities of rock materials (Blanks, 1949). Surface texture, for example, is generally admitted to be highly sig- nificant in the fabrication of concrete because of its profound effect on the bond that will develop between an aggregate and the enclosing cement. But that property is never measured except in research in- vestigations; there exists no test amenable to routine application by which roughness or smoothness of aggregate particles may be defined; there is no specification in which surface texture is used as a criterion for acceptance or rejection; and, indeed, there exists only a general and qualitative understanding of the effects of different kinds and degrees of surface texture. Many other properties of aggregates are understood in the same imperfect way: they are recognized as im- portant; their effect on concrete can be gaged qualitatively; but their significance lacks, and badly needs, quantitative definition. Concrete technology requires that research of the future be directed toward the determination and elucidation of the properties of rocks and minerals, in terms of their quantitative significance to concrete. An intimate knowledge ef the inner character of rocks and minerals will be a first requisite to this research. This knowledge will be con- tributed chiefly by geologists and petrographers, with their special knowledge of the origins, histories, compositions, and textures on which the properties of rocks and minerals depend. But, if these contribu- tions are to be directly useful, geologists and petrographers must also have insight into the concepts and technology of concrete. Research will progress only haltingly in this direction until the resources of the geological sciences are better mobilized for its prosecution—until more geologists acquire some competence in concrete technology, through academic training or practical experience, and become preoc- cupied with its problems. Cu. 24] APPENDIX 459 APPENDIX THE INFLUENCE ON CONCRETE OF VARIOUS PROPERTIES OF AGGREGATE STRENGTH The average crushing strength of rock types which would commonly be used as aggregate ranges from 10,000 to 30,000 pounds per square inch. Thus al- most any rock possessing the other characteristics necessary for an aggregate will be far stronger than concrete we expect to make. High strength in rock can be beneficial to concrete subjected to stress, but the degree to which the strength and elasticity of the aggregate contribute those qualities to concrete is controlled largely by the integrity of the bond between cement and aggregate. HARDNESS In the mineralogic sense, hardness of particles depends wholly on the hard- ness of the constituents, not on the firmness with which the constituents are knit together. When used this way, a convenient distinction can be made between “weak” particles, which may be composed of either hard or soft grains that are weakly joined, and “soft” particles, which abrade easily how- ever strongly the component grains may be joined. “Weakness” is never tolerable in concrete aggregate, but sometimes “softness,” 1f not extreme, may be admissible. Weakness and softness are frequently associated, and a safe rule is to keep both weak and soft particles at a minimum; in special-purpose concrete (for example, abrasion-resistant concrete) aggregate must be both hard and strong. COMPRESSIBILITY The compressibility of aggregate particles strongly influences the drying shrinkage of concrete. A compressible particle is one that is reduced in volume by contraction of the surrounding cement paste as the concrete dries. As the cement paste shrinks, aggregate particles are thrown into compression; if the particle is strong and the cement-to-aggregate bond is good, shrinkage of the cement paste is restrained, and, consequently, the volume change of the concrete is inhibited. On the other hand, if the particle can be com- pressed, the drying shrinkage of the concrete will be high. Drying shrinkage of concrete will be increased greatly by particles that expand and contract with wetting and drying. DURABILITY Durability of concrete is its resistance to disintegration, volume change, or loss of strength and elasticity under conditions to which it is subjected. 460 RHOADES. CONCRETE AGGREGATE [Cu. 24 Durability of concrete is controlled by the quality and gradation of the aggregate; the quality, composition, and fineness of the cement; the mix proportions; the methods of mixing, handling, placing, and curing; and the physical and chemical conditions that affect the concrete throughout its life. Aggregates which are physically unsound or chemically unstable, so that the particles disintegrate or change volume inordinately, can cause failure of concrete in service. Unsoundness or chemical instability of aggregate particles may be inherent to the constituent minerals or rock substances, or they may have been engendered by secondary processes of alteration. ELASTICITY Young’s modulus of elasticity of rock commonly used for aggregate ranges from 2 108 to 15 x 108. Young’s modulus of elasticity of hydrated neat cement paste ranges from about 1 x 10® to 4 X 108 as the load is sustained or rapidly applied and relieved. Although high elasticity is beneficial in particles used for ballast or road metal, aggregates of lower elasticity tend to reduce stress in concrete resulting from volume change or strain. PARTICLE SHAPE Angular particles make “harsh” concrete that is difficultly workable; rounded particles contribute to smooth, workable mixes. High proportions of flat or elongated particles also decrease workability and hence necessitate the use of more sand, cement, and water to produce satisfactory concrete. Also, they pack poorly, thus reducing bulk weight and decreasing compressive strength. Moreover, flat particles tend to orient themselves horizontally im concrete, permitting the accumulation of water beneath them, a condition that prevents development of good bond on their lower surfaces. Maximum AND MINIMUM GRADATION The gradation of an aggregate is its particle-size distribution and is con- veniently expressed as the percentage of material retained on standard screens. The grading of concrete aggregate has very pronounced influence on the workability of concrete and on the proportion of cement and water needed to produce concrete of desired quality. Because of economic considerations, variation in particle shape and texture, variation in different brands of cement, and variation in size and design of structures, no one gradation is satisfactory for all purposes. However, to be suitable for concrete manu- facture, aggregate gradation must fall within certain limits in order to achieve satisfactory packing of the particles. In practice, the gradation of aggregate is controlled by specifying the maximum and minimum quantities that can be retained on screens of various sizes of mesh. The grading of aggregate is so important that, when natural grading of an aggregate does not fall within specification limits, screening into size fractions and recombining to produce a satisfactory grading are justified. Cu. 24] APPENDIX 461 SuRFACE TEXTURE The roughness and pore characteristics of the surface and near-surface por- tions of aggregate particles create a surface texture which in large part deter- mines the integrity of the bond established between the particle and the cement paste. The surface texture reflects the original internal structure and com- position of the particles as well as the natural and artificial processes of im- pact and abrasion to which they have been subjected. SpeciFic GRAVITY Specific gravity of aggregate influences the unit weight of concrete but is of direct importance only where design or structural considerations require that concrete have an unusually high or low unit weight. Although low specific gravity of an aggregate is frequently considered an indication of unsoundness, numerous exceptions to this criterion prevent its application without confirma- tion by other means. Porosity A rock or mineral particle can be penetrated by water to the extent that interconnected voids or fractures are present. The amount of water absorbed depends on the abundance and size of the internal voids and fractures, and the rate of penetration depends on their size and continuity. Porosity, per- meability, and absorption strongly influence chemical stability, abrasion re- sistance, elasticity, and apparent specific gravity of particles, as well as the degree of bond between particle and cement. VoLUME CHANGE Volume change in aggregate resulting from changes in moisture content and temperature is a common source of injury to concrete. Shales, clays, and some rock materials expand by absorption of water and shrink on dehydration. In designing structural elements, allowance must be made for thermal volume changes of concrete, which are influenced by the average thermal expansivity of the aggregate. Damaging internal stresses may develop when the change in volume of an aggregate due to temperature is different from the change pro- duced in the cement paste. CoATINGS Coatings form on the surfaces of natural aggregates through deposition of mineral substances from ground water. Coatings usually are composed of clay, silt, or calcium carbonate; but opal, iron oxides, gypsum, manganiferous substances, and soluble phosphates can occur as coatings. Particles with sur- face coatings are generally undesirable for use as concrete aggregates. The bond between particle and coating may be weak and thus will decrease strength of aggregate-cement bond. Many natural coatings contain substances (such as opal) susceptible to reaction with alkalies in cement. Soluble alkali salts 462 RHOADES. CONCRETE AGGREGATE [Cu. 24 may augment alkalies in cement; other salts may alter the course of cement hydration. In some cases, coatings may be essentially inert and strongly bonded and may actually increase strength of aggregate-cement bond. DELETERIOUS IMPURITIES Natural aggregate is commonly contaminated by dirt, silt, clay, mica, coal, humus and other organic matter, and various salts. These substances act in a variety of ways to decrease strength and durability, cause unsightly appear- ance, or complicate processing and mixing operations. They may increase water requirements of the concrete; they may be physically weak or susceptible to breakdown by weathering; they may be so active chemically as to inhibit normal hydration of cement or react with cement constituents. Fortunately, excesses of deleterious impurities may frequently be removed by simple treat- ment. Silt, clay, soluble salts, and some lightweight substances are usually removable by washing. Special processes may be necessary for less amenable substances, or their removal may not be practicable. MINERAL AND CHEMICAL PROPERTIES The mineralogic and chemical composition and the internal texture and structure of particles control the physical and chemical properties of aggre- gates. Few, if any, rocks or minerals are inert when enclosed in concrete. Some may contain water-soluble constituents which can be leached, with attendant loss of strength and increase in porosity. Leached material may form unsightly efflorescence or modify normal hydration of cement. Un- stable minerals are susceptible to oxidation, hydration, or carbonation and may cause concrete to become distressed. Certain minerals are capable of base- exchange action in which alkalies in the minerals are exchanged for calcium in the cement solutions; released alkalies may then attack susceptible aggregate particles. Some minerals—opal, chalcedony, tridymite, cristobalite, and heu- landite, and rocks such as glassy or eryptocrystalline rhyolites, dacites, and andesites—and opaline and chalcedonic cherts and phyllites are reactive with high-alkali cements (cement containing more than 0.6 percent NazO + K20O expressed in equivalents of Na,O) and can cause deterioration of concrete through production of alkalic silica gels which subsequently absorb water osmotically from the cement paste, developing hydrostatic pressures In excess of the tensile strength of concrete. Thermal compatibility between mineral particles and cement paste signifi- cantly influences quality of cement when the linear thermal coefficient of ex- pansion of the mineral is decidedly lower than that of the cement paste. Some difficulty is experienced with minerals that have different thermal coefficients of expansions in different crystal directions. For example, with increasing temperature calcite expands greatly in one crystallographic direction and contracts in another direction. Cu. 24] REFERENCES 463 REFERENCES References marked by a single asterisk contain important general discussions; references marked by a double asterisk also contain comprehensive bibliographies. ** American Society for Testing Materials (1948a). Symposium on mineral ag- gregates: Special Tech. Pub. 83, 233 pages, Philadelphia, Pa. (1948b). Petrographic and mineralogic characteristics of aggregates: Roger Rhoades and Richard C. Mielenz, pp. 20-48. (1948c). Grading of mineral aggregates for portland cement concrete and mortars: Walter H. Price, pp. 184-151. (1948d). Production and manufacture of fine and coarse aggregates: Nathan C. Rockwood, pp. 88-116. (1948e). Distribution of mineral aggregates: K. B. Woods, pp. 4-19. ** Blanks, R. F. (1949). Modern concepts applied to concrete aggregate: Proc. Amer. Soc. Civ. Engrs., vol. 75, No. 4, pp. 441-468, April 1949. Lea, F. M. (1938). The chemistry of pozzolans: Proc. of Symposium on the Chemistry of Cements, Stockholm, pp. 460-490. , and Desch, C. H. (1935). The chemistry of cement and concrete: Long- mans, Green and Co., New York, 429 pages. McConnell, Duncan, Mielenz, R. C., Holland, W. Y., and Greene, K. T. (1948). Cement-aggregate reaction in concrete: Proc. Amer. Concrete Inst., vol. 44, pp. 93-128. ‘ Price, Walter H., and Cordon, William A. (1949). Tests of lightweight aggregate concrete designed for monolithic construction: Proc. Amer. Concrete Inst., vol. 45, pp. 581-600. Rasmusson, I. S. (1946). Concrete at advance bases: Proc. Amer. Concrete Inst., vol. 42, pp. 541-551. ** Rhoades, R., and Mielenz, R. C. (1946). Petrography of concrete aggregate: Proc. Amer. Concrete Inst., vol. 42, pp. 581-600. Scholer, C. H., and Gibson, W. E. (1948). Effect of various coarse aggregates upon the cement-aggregate reaction: Proc. Amer. Concrete Inst., vol. 44, pp. 1009-1032. Spain, E. L., Jr., and Rose, N. A. (1937). Geological study of gravel concrete aggregate of the Tennessee River: Amer. Inst. Mining Metallurgical Engrs., Tech. Pub. 840. Thoenen, J. R. (1932). Prospecting and exploration for sand and gravel: U. S. Bur. Mines, Information Circe. 6668, 53 pages. ** Twenhofel, William H. (1932). Treatise on sedimentation: Williams and Wilkins Co., Baltimore, Md., 2nd ed., 926 pages. *U.S. Bureau of Reclamation (1949). Concrete Manual: Denver, Colorado, 5th ed., 489 pages. CHAPTER 25 APPLICATION OF STUDIES OF THE COMPOSITION OF CLAYS IN THE FIELD OF CERAMICS * Rautey EF. Grim Geologist, Illinois State Geological Survey Urbana, Illinois This chapter states briefly the factors of composition which deter- mine the properties of clays and discusses the relation of the factors of composition to the particular properties of clays which determine their use for ceramic purposes. The practical application of studies of the composition of clays, particularly in the field of ceramics, is also discussed. THE COMPOSITION OF CLAY MATERIALS It is generally agreed by present-day students of clays that clay materials are composed essentially of crystalline particles of members of any one or more of a few groups of minerals known as the clay minerals (Grim, 1942). The clay minerals are hydrous aluminum silicates, frequently with some replacement of the aluminum by iron and magnesium and with small amounts of alkalies and alkali-earths. In some clay minerals, magnesium and iron completely replace alumi- num. In addition to the clay minerals, variable amounts of quartz, limonitic material, feldspar, pyrite, organic material, and a host of other minerals may be present as extremely minor constituents, or as prominent constituents in some clays. Most of the clay minerals occur in flat, flake-shaped particles. The unit shape of some clay minerals is lath- or fiber-shaped. The clay minerals occur in most clay mate- rials in particles less than about 0.005 millimeter in diameter. Table 1 lists the common clay mineral groups together with their composi- tion. * Published with the permission of the Chief, Illinois State Geological Survey. 464 Cu. 25] CONTROLLING FACTORS 465 TABLE 1 Common Cray MINERALS 1. Kaolinite group (a) Equidimensional flake-shaped units Kaolinite: (OH) gALSi4010 (6) Lath-shaped units Halloysite minerals: (OH)gAlSis4040 ; (OH)sALSi4019-4H20 2. Montmorillonite group (a) Equidimensional flake-shaped units Montmorillonite: (OH)4(Al,- Fea: Mg4)SigO20-2H2O (6) Lath- or needle-shaped units Nontronite: (OH)s(Fes- Mga4- Al4)Sig020-7 H2O Hectorite: (OH)4(Mg- Li)¢Sig020-n H2O 3. Lllite group Insufficient data to subdivide: (OH)4K,(Ala- Mga- Fea) (Sig_,- Alg)O20 4. Miscellaneous fiber-shaped units Attapulgite: (OH»2)4(OH)2Mg;SigO20 -4H20 Sepiolite-like: (OH)4MgeSigO20- ?7H2O 5. Clay minerals resembling chlorite and vermiculite are known to occur, but in- sufficient data are available for their classification. FACTORS THAT CONTROL THE PROPERTIES OF CLAYS The factors that control the properties of clays may be listed as follows: (1) Clay mineral composition: the kind of clay minerals, their relative abundance, and their particle-size distribution. (2) Non-clay mineral composition: the kind of non-clay minerals, their relative abundance, and the size distribution of the particles of each mineral. (3) Electrolyte content: the amount and kind of exchangeable bases and water-soluble salts. (4) Organic content: the amount and kind. (5) Miscellaneous textural characteristics such as shape of quartz grains, degree of parallel orientation of the clay mineral particles, silicification. The clay mineral composition is usually the dominant factor con- trolling the properties of clays, but in some clays the other factors play a dominant role. it is particularly significant that relatively small amounts of certain components may exert a very great influence on some of the properties of a clay (Grim, 1948). As will be shown, this 466 crim. CHRAMICS AND CLAY COMPOSITION [Cx. 25 applies to the clay minerals, electrolyte content, organic content, and non-clay mineral content. RELATION OF THE CERAMIC PROPERTIES OF CLAYS TO THEIR COMPOSITION The object of the following discussion is to provide some under- standing of the wide variation in the properties of clays and the causes of such variations. An understanding of the properties of clays is es- sential to any satisfactory evaluation of economic potentialities of a clay deposit. No attempt is made to consider the theory of such properties or the methods for their specific evaluation, since there is an abundance of information in the ceramic literature concerning these matters. PLASTICITY When mixed with certain moderate amounts of water, clays generally can be deformed under pressure without rupturing, and the deformed shape is retained after the pressure is removed. A clay that becomes plastic readily with a small amount of water and that deforms easily over a wide range of pressures is in general most usable. Plasticity values for various common clay minerals are given in Tables 2 and 3. The range of plasticity values is a result of variations due to particle size, character of exchangeable base, and the like. In the case of halloysite, the high and low hydration forms have sub- stantially no plasticity, whereas the intermediate stage has some plasticity. Clays composed of illite and kaolinite usually have good working properties, and the plasticity tends to increase as the particle size decreases. Montmorillonite clays are frequently exceedingly plastic and require large amounts of water to develop the plastic state. Small amounts (+5 percent) of montmorillonite in a clay material will gen- erally provide a material with considerably higher plasticity than it would otherwise have. Similarly, small additions of montmorillonite clay greatly enhance the plasticity of a relatively non-plastic material. A clay material that contains a considerable amount of non-clay material frequently has better working properties than a clay material composed solely of clay minerals. The application of this generality depends somewhat on the kind of clay, the type of ceramic ware to be provided, and the method of manufacture. It is most applicable to montmorillonite clays, and particularly to structural clay products. Clays containing as much as 50 percent non-clay minerals often have Cu. 25] PLASTICITY 467 TABLE 2 Tur Puastic AND Lieurp Limirs AND Puastic INDEX oF Some Cray MINERALS [After White (1947)] Clay Particle Size | Plastic | Liquid | Plastic Limit! | Limit? Index Illite Grundy Co., Ill. <1.0 micron 39.59 83.00 43.40 <0.5 micron p2.2¢ | 103.165 51.38 LaSalle Co., Ill. <1.0 micron 46.21 85.55 39.34 <0.5 micron 52.98 | 111.25 58.27 Vermilion Co., Il. 1.0 micron 44,44 95.05 50.61 Kaolinite Union Co., IIl. Whole 36.29 58.35 22.06 <1.0 micron 37.14 64.20 27.06 <0.5 micron 39.29 71.60 SP ail Dry Branch, Ga. Whole 30.0 35.0 5.0 Montmorillonite Belle Fourche, Wyo. Whole 97.04 | 625-700 | 528-603 Pontotoc, Miss. <1.0 micron | 109.48 175.55 66.07 Whole 81.41 117.48 36.07 Attapulgite Attapulgus, Ga. Whole 116.64 | 177.80 61.16 Halloysite Eureka, Utah Whole 51.81 62.30 10.49 1 The plastic limit is the water content at which the clay can just be rolled into a thin thread. The liquid limit is the water content at which the mass Just begins to flow. The water content is based on water loss at 110°C. For details of the test see Casagrande (1932). TABLE 3 WATER OF Puasticiry! ror VARIOUS CLay MINERALS [After White (1947)] Clay Minerals Water of Plasticity Kaolinite 8.89-— 56.25 Tllite 16.95- 38.5 Montmorillonite 82.9 —250.0 Halloysite 33° ~— 50 Attapulgite 92.6+ 1 The water of plasticity is the percent of water based on water loss at 110° C. required to develop a workable plastic state. 468 - GRIM. CHRAMICS AND CLAY COMPOSITION [Cu. 25 excellent working properties, and considerably less clay mineral may be present if the product is to be formed by dry pressing. Unsorted non-clay mineral particles provide greater density in the formed ware than particles of uniform size and are, therefore, most desirable. Small amounts of organic material enhance the plastic properties of some clays. However, not all types of organic material increase the plasticity, and it is not possible to predict generally the influence of organic material in a particular clay. The plastic properties of a clay vary somewhat with the content and character of exchangeable bases and soluble salts. Thus, the addition of soda ash (NasCO3) to a clay carrying exchangeable calcium tends to reduce the amount of water necessary to obtain a given degree of plasticity. SUSPENSION CHARACTERISTICS All the clay minerals form suspensions in water if their particle size is sufficiently small and if there is the proper electrolyte con- tent. The montmorillonite minerals, attapulgite, and perhaps the sepiolite-like minerals are more easily placed in suspension than the other clay minerals, because they break down very easily in water to exceedingly small particles, because of their adsorptive power for elec- trolytes, and because of their influence on the state of the water im- mediately surrounding them (Grim, 1942). When certain montmorillonite clays carry sodium as the exchange- able ion, their suspensions in water have a high degree of thixotropy. The thixotropic property would be detrimental in some ceramic ap- plications, as in slip casting, and advantageous in others, as in steel enameling. Small amounts of montmorillonite in a clay or added to a clay greatly enhance suspension-forming characteristics. BonpDING STRENGTH As shown in Table 4, the montmorillonite clay minerals have greater bonding strength than any of the other clay minerals studied so far. Montmorillonite clay carrying calcium (Mississippi) has higher green strength than a similar clay carrying sodium (Wyoming). Sodium montmorillonite clays have the highest dry strength. Green strength refers to the strength of a compacted test piece containing tempering water. Dry strength is the strength of the compacted test piece after drying at 110°C. for 2 hours. The bonding strength of illite and kaolinite clays increases as the particle size decreases, and halloysite clays in an intermediate state of hydration have higher strength than either the high- or low-hydration form. Cu. 25] DRYING SHRINKAGE 469 TABLE 4 BonpDING STRENGTHS FOR CLAy MINERALS [After Grim and Cuthbert (1945)] Opt. Max. Opt. Max. Clay | w0: | acs? | HO: | DCS? % % %o Kaolinite Grundy Co., Il. 8 ihe 7 14.6 4.5 77 Tllite Grundy Co., Il. 8 1.9 WE 7 4.5 90 Montmorillonite Belle Fourche, Wyo. 8 2.07 24.1 3+ 100+ Pontotoc, Miss. 8 2.65 30.5 5 90 Halloysite Eureka, Utah 8 2.95 21.3 5 30 1 Optimum H.O content is the amount of tempering water necessary for maximum compression strength in mixtures of 8 percent clay and 92 percent standard testing sand. 2 Green compression strength. 3 Dry compression strength. In ceramic processes, it is necessary to have only sufficient dry and green strength so that the ware can be handled without deformation and stacked in a kiln before firing. In general any clay material with a fair amount of clay mineral (+50 percent) and non-clay material that is not sorted or coarser than sand will have adequate strength. Strength can be given to a clay deficient in this property by the ad- dition of some organic binder or a strong natural clay like a montmoril- lonite clay (bentonite) or a ball clay. Ball clays are very fine-grained, light-burning clays that are composed largely of kaolinite but usually contain some organic material which is partly responsible for their strength. As noted for montmorillonite, the strength of a clay may be related to its exchangeable-base and soluble-salt composition. The addition of soda ash to some clays increases their strength, particularly the dry strength. DryInGc SHRINKAGE Economic use of a clay requires that the drying shrinkage be uni- form, small in amount, and develop without cracking the formed piece. 470 ckiM. CERAMICS AND CLAY COMPOSITION [Cu. 25 Clays which require large amounts of water to develop the plastic state will show high drying shrinkage. In general montmorillonite, when present in more than small amounts, increases shrinkage greatly. Clays composed of kaolinite and illite tend to have moderate drying shrinkage unless the com- ponent clay minerals are exceedingly fine-grained (minus 1 micron). Non-clay mineral components such as quartz and feldspar reduce the drying shrinkage when they are present in a clay. Shrinkage characteristics are in general improved by the presence of moderate amounts (25 to 40 percent) of such components. It is common practice to add non-clay material to reduce the high shrinkage of some clays. To be of commercial value a clay must form ware that will not be sensitive to considerable variation in drying conditions. The ware must not check or crack during drying, even though there is a large variation in time, temperature, or relative humidity during the drying operation. Actual tests are necessary to determine this point, but, in general, clays with very high clay mineral content (particularly if the clay mineral is very fine-grained) and montmorillonite are difficult to dry satisfactorily. FIRING CHARACTERISTICS Firing characteristics of a clay are: its shrinkage during firing; its color after burning; the temperature range during which it vitrifies; its resistance to heat or refractoriness; and the strength, texture, and other properties of the fired ware. Small quantities (a few percent) of certain components of clays, notably alkalies, alkali-earths and iron, may exert a controlling influence on the firing characteristics of a clay. The influence of certain of these minor components may depend on whether the component is present alone or mixed with some other con- stituent. Thus the effect of iron on color varies somewhat with the presence of lime and alkalies in a clay. Furthermore, the conditions under which a clay is fired, the rate of firing, and oxidizing or reducing conditions also influence the firing properties. It follows from the foregoing that variations in firing characteristics are exceedingly difficult to predict, and that actual firing tests are necessary for the evaluation of the burning properties of a clay. How- ever, a few general relationships seem to be established. Kaolinite clays tend to be light-burning (white if pure) and re- fractory. The color darkens, usually to a shade of red, and the re- sistance to heat decreases as the kaolinite is mixed with increasing amounts of illite and montmorillonite. Obviously iron in the non-clay minerals (limonite, pyrite, etc.) will also cause a change in color, and Cu. 25] AREAL EVALUATION OF CLAYS 471 alkalies and alkaline earths in feldspar, etc., will reduce the refractori- ness. Clays composed of a single clay mineral usually have a shorter temperature interval between the beginning of vitrification and com- plete fusion than clays composed of a mixture of minerals. This seems to be more applicable to illite and montmorillonite clays than to kaolinites. Factors other than mineral composition influence the vitrification range. Thus the vitrification range tends to increase as the range of particle size increases. The presence of alkali-earths fre- quently tends to shorten the vitrification range, and this is one reason why many calcareous clays are difficult to burn. Similarly, clays composed of a mixture of clay minerals are apt to have lower shrinkage during burning than substantially monomineral clays. This again is more applicable to illite and montmorillonite clays than to kaolinite clays. Certain clays tend to bloat or swell on firing when they are heated to a temperature at which considerable vitrification takes place. Il- lite and montmorillonite clays, particularly if they are relatively pure, have a tendency to bloat. Calcareous illite and montmorillonite clays seem to bloat very easily. A low-fusion temperature and short vitri- fication range are among the factors that favor bloating. INDUSTRIAL APPLICATION OF INVESTIGATIONS OF THE COMPOSITIONS OF CLAYS Studies of the composition of clays are of practical importance in several ways in the solution of problems involving the use of clays for ceramic and other purposes. EVALUATION OF CLAYS FOR CERAMIC USE IN AN AREAL STUDY Frequently there is the problem of determining if the clays in a given area are particularly suitable for any ceramic use. Obviously such a problem can be solved by making complete ceramic tests of samples of all varieties of clays. This procedure is expensive and time-con- suming if there are many varieties of clays to be tested, and if the area is large. A simpler and more rapid procedure is to determine the mineralogical composition of the clays and then select (on the basis of the analytical data and the general relation between com- position and properties) those clays that appear to be particularly suited for certain ceramic uses. The ceramic properties of those clays can then be determined in detail. The determination of the composi- tion of the clays does not take the place of determinations of ceramic 472 crim. CERAMICS AND CLAY COMPOSITION [Cu. 25 properties, but it presents a rapid method of eliminating the worthless materials. Although it is beyond the scope of this paper, it should be noted that there is a correlation between the composition of clay materials and uses of clay outside the field of ceramics, for example, in drilling muds, decolorizing oils, fillers, ete. Determinations of composition, therefore, also permit the selection of promising samples for com- mercial use in fields other than ceramics. EVALUATION OF A PARTICULAR CLAY DEPOSIT A geologic study of a clay deposit involves two distinct problems: first, the determination of the geologic setting; and, second, the uni- formity of the properties of the clay in the deposit. The solution of the first problem is straightforward geology; it provides data on the size of the deposit, overburden, ground-water conditions, large varia- tions in character, etc. The determination of uniformity of material in a clay body is apt to be difficult, but it is of great importance. The commercial utiliza- tion of a clay cannot be successful unless the deposit contains a large amount of clay of highly uniform properties or capable of simple beneficiation so that it can be made uniform. Any utilization of clay is inherently based on certain properties that the clay possesses, so that any change in the properties of the clay in a given deposit is likely to cause great difficulty in its processing or use. American producers of clay have not always realized the necessity of producing a uniform product, and only after this was accomplished were they successful in competing with clays that were imported into this country. Clay deposits seem to be inherently variable. A casual examination may suggest that a clay body is uniform, whereas detailed tests will reveal it to be quite variable. In other words, the outward appearance of uniformity does not necessarily indicate uniformity in properties. The fact that very small variations in composition may cause large variations in properties must mean that important changes in physical properties are often not accompanied by any changes in gross char- acteristics. This is an exceedingly important point, and geologists and others have been led into costly blunders because they assumed that clay was inherently a relatively inert, uniform material that could be evaluated by superficial study. Nothing could be further from the truth. The detection of variations in properties involves the study of the properties of many samples. Usually it is not enough to detect varia- tions in properties; the cause of the variations, their relation to com- Cu. 25] CERAMIC PROCESSING PROBLEMS 473 position, and the occurrence of the clay in the field are also essential. It is frequently necessary to predict the occurrence of clays with particular properties in a given deposit before it can be evaluated and mined. In addition, the variable properties of a clay cannot be con- trolled by beneficiation, and the variable properties cannot be com- pensated for in processing unless the cause of the variation is known. Evaluation of a clay deposit therefore should include a thorough study of all aspects of the occurrence, origin, and composition to re- veal the causes of any variations in properties of the clay. Many ex- amples could be given to illustrate the importance of this, but one will suffice. Certain beds of kaolin in the Georgia area contain small amounts of montmorillonite in addition to the kaolinite. The presence of montmorillonite does not change the appearance of the clay, and in fact is not revealed unless a careful analytical study is made, yet the presence of the montmorillonite changes the properties of the kaolin so that it is not usable for certain purposes. Obviously any informa- tion about the factors controlling the distribution of the montmoril- lonite in the kaolin is of great importance to the kaolin producers. SEARCH FOR CLAY OF A PARTICULAR TYPE The application of studies of the composition of clays to the search for particular types of clay can best be shown by an illustration out- side the field of ceramics. Some catalysts used in the making of gaso- line are prepared from clays. Only a very few deposits of clay suitable for this use have been found, and each clay has a particular composi- tion and an origin which requires a certain geologic setting. A knowl- edge of these factors obviously permits one to spot areas throughout the world in which such clays are most likely to occur. SOLUTION OF CERAMIC PROCESSING PROBLEMS A plant ceramic engineer spends considerable time studying process- ing problems which arise either because of variations in the processing technique or because of variations in the raw materials being used. The solution of problems of the latter type is expedited by a knowledge of the composition of the raw material being used and its relation to properties. Once the fundamental cause of the properties is known, the solution to the problem is often obvious. For example, a clay plant in a midwestern state operating on a series of glacial clays suddenly found its percentage of rejected brick to increase greatly because of lack of strength in the dried and fired brick. An examination showed the presence of a lens of silt which had gone undetected. More 474 crim. CHRAMICS AND CLAY COMPOSITION [CxH. 25 thorough mixing of the clay before processing to distribute the silt uniformly solved the problem. Frequently such processing problems are far more complicated, and the cause begins to appear only after very thorough analysis of all factors of the composition of the clay. PLANT CoNnTROL FOR UNIFORMITY OF MATERIAL It is obvious from the foregoing discussion that both producers and users of clay are interested in clay with uniform properties. They are, therefore, interested in tests, which usually must be rapid, to detect variations in properties and, if possible, the cause of the variation in properties. Simple determinations of properties are, of course, used in plant-control work, but frequently a test to show variation in com- position is more satisfactory because it shows something of the cause of the variations in properties 7n addition to the simple facts of property variation itself. Again an example will suffice to illustrate the point. Differential thermal analysis (Grim, 1942) has been found to be an excellent instrument for plant control because it provides information about several factors of the composition of a clay, so that, in addition to showing that a clay sample has varied, it reveals the nature of the variation. REFERENCES Casagrande, A. (1932). Research on the Atterberg limits of soils: Public Roads, vol. 13, pp. 121-130. Grim, R. E. (1942). Modern concepts of clay materials: Jour. Geol., vol. 50, pp. 225-275; Jil. Geol. Survey, Rept. Inv. 80. (1944). Differential thermal analysis of clays and shales; a control and prospecting method: Jour. Amer. Cer. Soc., vol. 27, pp. 65-76; Ill. Geol. Survey, Rept. Inv. 96. (1948). Some fundamental factors influencing the properties of soil ma- terials: Proceedings of Second International Conference on Soil Mechanics, Rotterdam, vol. III. , and Cuthbert, F. L. (1945, 1946). The bonding action of clays, Part I, ’ Clays in green molding sands: J/l. Geol. Survey, Rept. Inv. 102; Part II, Clays in dry molding sands: J1l. Geol. Survey, Rept. Inv. 110. White, W. A. (1947). The properties of clays: Master’s thesis, University of Il- linois. CHAPTER 26 FOUNDRY SANDS H. Riss Emeritus Professor of Geology Cornell University Ithaca, New York The term foundry sand is applied to those sands which are used for making the molds and cores in which metals are cast. The molds form the outside of the casting, and the cores make the hollows in it. These materials are of considerable economic importance. In 1947 the tonnage produced in the United States amounted to 8,308,434 short tons valued at $11,944,228. In addition to this, some sands were imported. Michigan, New Jersey, Illinois, Ohio, and New York were the leading producers. Foundry sands are widely distributed in the United States, but only special types of sands are suitable for this purpose, because the sands have to meet rigid requirements of permeability, strength, and volume change with temperature. For a more detailed discussion of foundry sands than given in this chapter, see Ries (1948). The predominating mineral in most foundry sands is quartz, and next to it in abundance is feldspar. In most foundry sands the feldspar content is relatively low, particularly in sands used for casting steel. Some of the glacial outwash sands of central New York contain ap- preciable amounts of feldspar. The sands of the Bridgeton and Pen- sauken formations of New Jersey contain as much as 15 percent. In California most of the soft Eocene and Miocene sands are strongly feldspathic, as are also the dune sands on Monterey Bay, California. Sands containing as much as 30 percent feldspar have been used in Illinois. Mica is present in small amounts in some foundry sands, but highly micaceous sands are not used. Many sands contain small amounts of dark minerals such as zircon, ilmenite, rutile, and mag- netite. Some sands contain minor amounts of leucoxene, staurolite, kyanite, tourmaline, and garnet. 475 476 riES. FOUNDRY SANDS [Cu. 26 The most common clay mineral in foundry sand is illite, but kaolin- ite and montmorillonite have also been reported. Some sands are practically free from clay, and the foundryman often refers to these as sharp sands. Those sands which contain clay mixed with the sand grains are referred to as naturally bonded sands, and those free from clay, to which the latter is added before use, are called synthetic sands. Bentonite and fire clay are the materials most commonly added to form the bond. Practically all steel foundries now use synthetic sands, and their use is extending to the casting of iron and other metals. PROPERTIES OF SAND GRAINS Shape of sand grains. Foundry sand grains vary in shape, size, and assemblage. In shape they may be angular, subangular, or round. They may also consist of aggregates of smaller particles cemented together by silica, iron, or calcium carbonate. For illustration of the separate sizes of a number of different sands see Ries and Conant (1931, p. 353). The degree of angularity varies in different sands, but most grains are either angular or subangular. Round grains are rare and occur usually only in sizes larger than 40 mesh (0.4 milli- meter). Most sand grains contain minute fractures, which may cause the grains to crack on heating, thus developing fines when the sand is ex- posed in the mold to the heat of the molten metal. The foundryman often refers to these cracks as cleavage. The surface of sand grains may be either smooth or rough, and the smooth surfaces are sometimes frosted. In addition, the surface may be clean, or it may be coated with a film of foreign matter such as clay or iron oxide. It is probable that rough or coated surfaces offer a better attachment for the bond. Fineness of sands. The size distribution of the grains affects the type of casting for which the sands are used. In making a fineness test to determine the percentage of different grain sizes, a. sample of 50 grams of dried sand is usually taken. Clay if present is first sepa- rated by stirring the sand in a special jar in water containing sodium hydroxide. The suspension is then allowed to stand, the grains larger than 20 microns settling in 5 minutes. The material still in suspension is re- moved by decantation, the settled material stirred again, and the process repeated until the supernatant liquid is clear. The grains up to 20 microns are known as AFS clay, which includes both true clay and fine silt. Cu. 26] PROPERTIES OF FOUNDRY SANDS 477 The settled material in the jar is then removed, dried, and screened, usually with sieves of the U. 8. Bureau of Standard Series, but some- times with the W. S. Tyler sieves. Grain-fineness number. After a sieve test is made, it is customary to calculate the average fineness, which represents the size of grain that would be formed if all the material were formed of grains of uni- form size. The method consists in multiplying the sand percent re- tained on each sieve by a certain factor. The sum of these products is divided by the sum of the grains, and the quotient represents the average fineness. The chief disadvantage of the grain-fineness number is that two sands may differ in grain-size distribution and thus differ in physical properties for foundry purposes, yet have the same fineness number. A better idea of the distribution may be gained by expressing the fineness graphically by means of the grading-size curve (weight- accumulation curve) or the size-frequency curve (histogram). The grading-size or cumulative curve is an S-shaped curve which presents percentage of particles passing given sieve sizes. The size-frequency curve is a peak-shaped curve similar to the probability curve. The advantages of the cumulative curve over the size-frequency curve are: (1) it gives a smooth curve; (2) sieves that retain little material can be eliminated or other sizes added without distorting the curve; (3) faulty sieves are indicated by a break in the curve when different samples are sieved; (4) it is more practicable for specifying types of sands required because it permits the specification limits to be plotted as two curves rather than as points falling between two limiting curves already plotted; and (5) the data for silt- and clay- size particles obtained with the hydrometer can be plotted on the same graph, as a continuation of the line presenting the sieve data. PROPERTIES OF FOUNDRY SANDS Sands used for molds and cores must possess certain properties, and unless these are developed to the right degree the sand is useless for foundry purposes. This fact is not always realized by those not familiar with the subject. The properties which the sands must possess are: (1) Sufficient cohesiveness to hold together when moist, and this necessitates the presence of some bonding material. (Clay is the com- mon natural bond and is present in the naturally bonded sands. If the sand has no clay bond, it must be mixed with clay or some artificial binder in the proper amounts. The types of clay commonly used are 478 ries. FOUNDRY SANDS [Cu. 26 fire clay, bentonite, and illite. Various oils and organic substances are also used. The sand mixture must have sufficient moisture added to it to cause the particles to cohere.) (2) Sufficient refractoriness to resist the heat of the molten metal. (3) Sufficient strength to resist the pressure of the metal. (4) Sufficient permeability to permit water vapor and gases to es- cape outward from the mold instead of being forced into the molten metal. (5) Proper texture so that the mold surface will be sufficiently smooth to produce a smooth surface on the casting and not develop surface defects. These are the main requisites, and the proper mixture is obtained by using sand of appropriate grain size, proper amount of bond, and correct amount of moisture. The various properties which control the conditions above mentioned can be determined by special tests, most of which have been standardized by the American Foundrymen’s Society (formerly the American Foundrymen’s Association). < © Zo cw OF S& O2z 20 wg w& fe) Sis} zacet ae = OnaAGa ! 1 a | | ! (a) (