RELATION OF VARIOUS PARTICLE SEZE umrs . m m 3m 3125 RANGE TO SELECTED ansm— ’ PROPERTEES Xhesis for the Degree 6f M. S. MICE-£56m STATE UNIVERSITY GARY CARL STEENHARDT 1968 THESIS LIBRAR 3:" Michigan Sta“? University 1 I us.- ‘- -. *l.‘ ' '4‘ ABSTRACT RELATION OF VARIOUS PARTICLE SIZE LIMITS IN THE SILT SIZE RANGE TO SELECTED PHYSICAL PROPERTIES By Gary Carl Steinhardt Various size limits were correlated with selected physical properties. Little relation was found between silt fraction and the liquid limit, plasticity index, shrinkage limit. shrinkage ratio, optimum moisture, maximum density, shear stress, 1/2 compressive strength, natural density and dry density. A close relationship, however, was found between the (.OOme and (.005mm content and the liquid limit and plasticity index. The highest correlations between size fractions and various measures of available water were found with coarser silt fractions. For .06 to 6 atmospheres as a measure of available water. the .125-.020mm size fraction was best correlated. When .06 to 15 atmospheres was used as a measure of available water. the .lOO-.020mm fraction was best correlated. The .33 to 15 atmospheres measure of available water was best correlated with .lOO-.OlOmm size fraction. Both capillary rise and hydraulic conductivity were found to be more dependent on the selection of a lower size limit than of an upper size limit. It was concluded on the basis of results that separa- tions should be made at these particle size limits in the silt to fine sand range: .002mm .020mm .062mm or .074mm . 105mm or . 125mm RELATION OF VARIOUS PARTICLE SIZE LIMITS IN THE SILT SIZE RANGE TO SELECTED PHYSICAL PROPERTIES By Gary Carl Steinhardt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Soil Science 1968 ACKNOWLEDGMENTS The author wishes to express his sincere appreciation to Dr. E. P. Whiteside for his guidance during the course of this study and earlier for his initial introduction of soil science to the author. The author also expresses his appreciation for the advice of Professor 1. F. Schneider. Acknowledgment is due the Michigan State Highway Department for providing data which.were used in this study. Particular thanks go to Mr. G. O. Kerkoff of that department whose assistance was vital in obtaining these data. The author gratefully acknowledges the encouragement and help of his family during the course of this study. ii I. II. III. IV. TABLE OF CONTENTS INTRODUCTIONOOCOOOOO0000......IO..IOOOOOOOIOOOOOOOO LITERATURE REVIEWOOOOOOOOOOOOOOOCIOOOOOOO0.0.0.0000 A. B. C. D. PARTICLE SIZE CLASSIFICATION................... 1. Agricultural Particle Size Classification and Particle Size Analysis................. 2. Geological Particle Size Classification.... 3. Particle Size Classification in Engineering MOISTURE RELATIONSHIPS AND PARTICLE SIZE IN SOILSOOOOOOOOO0.0.0....OOOOOOOOOOOIIOOOOIOOI CAPILLARY RISE AND PARTICLE SIZE IN SOILS...... PERMEABILITY AND PARTICLE SIZE IN SOILS........ METHODSOOOOOOOOOO...0.0.0.000....OOOOOOOOOOIOIOI... A. B. C. ENGINEERING PROPERTIES OF SOILS AND PARTICLE SIZE LIMITSOOOOCOOOOCCOOOOOOCOOOOIOOOOOOOOOCOOO AVAILABLE WATER AND PARTICLE SIZE LIMITS....... PARTICLE SIZE OF SILT FRACTION AND EFFECT ON HYDRAULIC CONDUCTIVITY AND CAPILLARY RISE...... 1. HYdffiUliC condUCtiVitYoococo-00000000000000 2c capillary Rise........................o.... RESULTS AND DISCUSSIONOOOOOOOOOIIOOOIIOOOOOOI..0... A. ENGINEERING PROPERTIES OF SOIL MATERIALS AND SILT PARTICLE SIZE LIMITS...................... lo LiQUid Limit and PlaStiC1ty IndeXooooOooooo 2. Shrinkage Limit and Shrinkage Ratio-0000000 3. Natural Density and Dry Density............ 4. Optimum Moisture and Maximum.Density....... iii Page 35 49 57 63 66 75 75 78 79 80 81 82 82 83 86 89 90 V. VI. 5. Shear Stress and % Compressive Strength.... 6. Relation Between Particle Size Limits...... B. AVAILABLE WATER AND SILT PARTICLE SIZE LIMITS.. C. PARTICLE SIZE OF SILT FRACTION AND EFFECT ON HYDRAULIC CONDUCTIVITY AND CAPILLARY RISE...... lo HYdrRUIiC condUCtiVitYQOOOooooOooococo-0000 2. capillary Rise......................o...... CONCLUSIONSCC0.000.0.00....O.................0"... LITERATURE CITEDOOOOOOCO0.000......0.00.00.00.00... iv 90 95 98 117 117 120 124 125 1. 2. 3. 4. 6. 7. 8. LIST OF TABLES Correlation coefficients of the relation between various particle size limits and physical propertieSo00000000000.0000000000000000. Correlation coefficients of the relation between various particle size limits and phy81cal properties-00000000000000000000000.0000. Correlation coefficients and equations relating various size groupingS........................... Correlation of particle size ranges and water released between .06 and 6 atmospheres tension... Correlation of particle size ranges and water released between .06 and 15 atmospheres tension.. Correlation of particle size ranges and water released between .33 and 15 atmospheres tension.. Hydraulic conductivity of various size fractions Correlation of hydraulic conductivity with various fractionS..........................o..... Capillary rise data for various size fractions... Page 82 89 95 102 103 104 118 119 121 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. LIST OF FIGURES Liquid limit vs. 1 (.OOme...................... Plasticity index vs. 1 (.002mm.................. Shrinkage limit vs. 1 <.005mm................... Shrinkage ratio vs. 1 (.005mm................... Dry density vs. 1 (.005mm....................... Optimum moisture vs. 1 <.005mm.................. Optimum moisture vs. 1 .020-.002mm.............. Maximum density vs. 1 (.005mm................... Shear stress vs. 1 <.074mm...................... % Compressive strength vs. 1 (.020mm............ 1 .005mm vs. 1 <.002mm......................... 1 .062-.002mm vs. 1 .050-.002mm................. 1 .105-.002mm vs. 1 .050-.002mm,................ 1 Water released .06-6 atm. tension V30 % 0125-0020mm000000000000000.0000.o000000000 1 Water released .06-6 atm. tension V80 % .100-.020mm...............o...o........... 1 Water released .06-6 atm. tension V80 % 0125-0002mmoooo0.0000000000000000000.0000. 1 Water released .06-15 atm. tension V80 % 0125‘0020mmoso00000000000000.0000000000000 1 Water released .06-15 atm. tension V8. % .100-.020mm............................... 1 Water released .06-15 atm. tension V80 % olzs-ooozmmo00000000000000.000000000000000 1 Water released .33-15 atm. tension V80 % 0020-0002nmL00.000000000000000...000000000 vi Page 84 85 87 88 91 92 93 94 96 97 99 100 101 105 106 107 108 109 110 111 21. 22. 23. 1 Water released .33-15 atm. tension V8. % .050-.002mm...........o.................. 1 Water released .33-15 atm. tension V30 % 0125‘0002mmosoo00000000000000000000000000 1 Water released .33-15 atm. tension V80 % olOO'oOlommoooOoo0000000000000000000000so vii 112 113 114 I. INTRODUCTION This study is an attempt to evaluate various lower and upper size limits which have been suggested as silt separa- tions in particle size classifications of earthy materials. Particle size has long been recognized as a factor in determining physical properties of earthy materials. The exact nature of these relationships has not always been clear. This study will attempt to show quantitative rela- tionships between physical properties and various lower and upper particle size limits of fractions in the .002-.125mm particle size range. The properties studied include Atterberg limits. maximum density. optimum moisture. shear stress. 1/2 compressive strength. natural density. dry density. moisture retention, capillary rise and hydraulic conductivity. II. LITERATURE REVIEW Soil scientists. engineers and geologists are using particle size distribution information in their work. These professional groups. however. use different size limits in the classification of particles. particularly in the silt size class. By examining the development of the classifica- tion and the means of measuring particle sizes of each discipline. significant decisions and reasons for given size limits can be shown more clearly. Pr0perties related to the size distribution will be examined particularly. A. PARTICLE SIZE CLASSIFICATION 1. Agricultural Particle Size Classification and Particle Size Analysis Physical properties of soils which are a direct result of or closely correlated with particle size distribution in the soil materials were noted by early workers in agriculture. The first attempt at a quantitative approach to particle size and physical property relationships was by Houghton. His experiments including procedures and results were published in 1681 in ”A Collection of Letters for the Improvement of Husbandry and Trade.” The experiments were later cited in Mbrtimer's ”The Whole Art of Husbandry" published in 1706 from which the section quoted below was taken and cited by Keen (1931): An Experiment of Mr. Houghton. to know what quantity of Sand any Earth or Marle is mixed with. which may 2 be of use to try the nature of several sorts of Land by: He took a piece of Clay such as the Brewers stop their Cash with which is commonly a sort of yellow Tyle-clay. which weighed 4 Ounces % Averdupois, which he dissolved in Water. and poured off the thick into another Bason till all was gone but the Sand; which stuck together, but lay loose in the Water. and when it was dry would run like Hour-glass-sand. of which he had about the quantity of an Ounce. being of a yellow colour and something glistening, and some little Stones and other foul matter was with it; and when the Clay was settled he poured off the Water. and left the rest to dry in a Pewter-bason, which hung together. only 'twas full of Cracks. He tryed likewise Fullers-earth.which left a thick settlement. which when dry would easily break to powder. but he could find no Sand in it; neither by the Microscope. nor any Grittiness by rubing of it between his Fingers. This experiment was the first recorded particle size analysis and was the basis for many others that followed. More importantly, Houghton attempted to establish relations between his results from particle size analysis and physical properties which he could observe. He was particularly interested in the relation of particle size and moisture content. He attempted to explain soil moisture content not only on the basis of sand and earth in the soil. but also plant roots and the globular nature of water. His concepts are not in all cases correct by our explanations today. but they represent a crucial first step. Little use was made of this finding until Sir Humphrey Davy included an improved version of Houghton's system of particle size analysis in his Royal Institution Lectures to the Board of Agriculture. published as Elements of Agriculture in 1813. Davy, who was one of the first agricultural chemists, was quick to realize the importance of particle size to the chemical aspects of soils. A German named Wanschaffe prepared the following system of particle size limits in 1814 (Roderick 1963). SizeI mm fine gravel > 2 very coarse sand 2-1 coarse sand l-O.5 medium sand 0.5-0.2 fine sand 0.2-0.1 very fine sand 0.1-0.05 silt 0.05-0.01 fine clayey portion (0.01 It is significant to note that studies in soil physics were then being organized for the first time into a single body of information and pursued by another German named Schubler (Keen 1931). His studies in the physical properties of soils went far beyond the English work at the time which was being done by Davy. It was at this same time that Liebig in Germany. Boussingault in France. and Louis and Gilbert in England were making great strides in soil chemistry. As a result. much of the work on the physical properties of soils including particle size was neglected. Sieves are commonly used for particle size analysis of soil materials at the present time. but early work in agriculture did not make extensive use of them. Davy included them in his early procedure. but relied on sedimentation for separation of sand fractions (Keen 1931). Sieves. however. have been used for particle size separations far back into antiquity. The first written record of the use of sieves is found in Greek and Roman accounts of mining operations about 150 B.C. These writings describe sieves made out of planks or hides punched full of holes or woven from horse hair. reeds or even human hair. The first woven wire sieves were developed by Germans in the 15th century. Other improvements were made to use machines rather than hand shaking. The principal use for these sieves initially was for mining and other extractive industries (Tyler 1967). The idea of elutriation as a means of separating particle size fractions was originated by Schulze in 1839 according to Krumbein (1932). Schulze's apparatus consisted of a "champagne glass" with a long thistle tube extending from a water source to the bottom of the glass. The top was sealed with a tube leading out into a beaker. The procedure was to sieve out the pebbles and boil the remaining finer fraction for 10 minutes. The soil was then poured into the glass. and water run into the thistle tube until the water in the glass was clear. Schulze noted the effect of the current and the size of particles which remained in the glass. Prior to his initial work all particle size analysis had been based on the principle of sedimentation in still water. The first mention of the term elutriation came in 1843 in a paper by Fownes. As it was used by Fownes it referred to Davy's sedimentation procedure and not the technique with which the term is associated today. Ten years later, Schulze published a paper on particle size analysis which included refinements in his elutriation method as well as a grade size classification (Krumbein 1932). He was not satisfied with the sedimentation techniques that were being used for particle size analysis at the time. He was particularly disappointed with the reproducibility of some of the methods, particularly that of Schubler which was impossible to duplicate within 4 percent in the fine fraction. His grade size classification made these separations: Dimension flamg )1 2011 stone 1 - 1/2 Zoll gravel 1/8 2011 - 1/3 lime coarse sand 1/3 lime - 1/12 lime strewsand ‘ 5 coarse gravel 5-3 15 fine gravel 3-2 very coarse sand 2-1 coarse sand l-0.5 fine sand O.5-0.25 very fine sand (0.25 At this time in Russia. significant progress was being made on a grain size classification. ‘Fadejeff developed a system which was presented by Williams in 1895 to the Agricultural Academie Petroffskaja (Roderick 1963). The system suggested by Fadejeff made these separations: SizeI mm QEQEE stones & pebbles >10 stony coarse gravel & grits 10-7 7 medium gravel & grits 7-5 - gravelly fine gravel & grits 5-3 _ coarse sand 3-1 E medium sand 1-0.5 - sandy fine sand 0.5-0.25 J dust 0.25-0.01 ‘ coarse silt 0.01-0.005 medium silt 0.005-0.0015 r earthy fine silt (0.0015 Williams considered this a good system. but changes were needed based on the physical properties which he found characteristic of certain size particles. His suggested changes resulted in this substitution in the names of the earthy group: l6 Size. mm coarse silt 0.25-0.01 medium silt 0.01-0.005 fine silt 0.005-0.0015 clay '<0.0015 The change in the system was the removal of the dust category on the upper part of the silt range and the addition of clay as the smallest particle size. Williams was not arbitrary in his approach to the classification of particle size. His classification was based more than any other before him on physical properties which were observed related to a given grain size. He suggested the clay designation for particles less than .0015mm. His observations had shown that it was this fraction that was responsible for the cohesive properties of the soil. Particles .0015mm and smaller increased the cohesiveness of the soil. while larger particles tended to decrease it. This is more specific than any previous statement on particle size. Cohesive properties were by no means the only consideration that Williams made in setting his criteria for the particle size limits. He also noted a difference in specific gravity between the clay size and other sizes. Specific gravity of clay particles is less than the others. Total water content at saturation is also different for particles less than .0015mm and the silt size. just as the silt particles differ from the sand particles. The clay, of course. holds more total water than the silt and a great deal more than the sand. From a permeability stand- point, he noted that sand was very permeable, silt less so l7 and clay much less permeable. He also noted differences in the mineralogy. The sand and silt fractions were made up of primary minerals while that less than .0015mm which he called clay was made up of secondary minerals as a result of weathering. Despite the overwhelming acceptance of the Osborne system by government agencies in the United States. there were those who felt the limits were not selected with sufficient study of the implication of particle size limits to the usefulness of the classification. One of those who spoke against the arbitrary nature of the Osborne system was Hopkins of the U.S. Department of Agriculture, Bureau of Chemistry (1899). He was concerned that the comparison between particle size in the Osborne system would not be meaningful because the ratio between sizes was not constant. His approach to particle size classification was not based on a desire to match physical properties with the particle size. Rather his intent was to provide a suitable theoretical framework in which to fit the classification. He evaluated the Osborneclassifica- tion with an assumed soil of uniform gradation. The comparisons which he made are cited in the following tabulation: 18 Ratio of Ratio of Ratio of Theoretical Div. No. Size, mm diameters surfaces volumes 1 composition 1 >1. - - — 12.50 2 l -0.5 2 4 8 9.68 3 0.5 -0.25 2 4 8 9.68 4 0.25-0.10 2.5 6.25 15.6 12.10 5 0.10-0.05 2 4 8 9.68 6 0.05-0.01 5 25 125 24.20 7 0.01-0.005 2 4 8 9.68 8 (0.005 - - - 12.50 It is obvious from the study that consistent or significant relationships from the mathematical point of view are not readily available in this system. Osborne stated in corre- spondence with Hopkins the arbitrariness of his particle size limits and his surprise that they had been adopted by so many people (Truog et a1. 1936). In working out the beaker method of soil analysis I employed the limits of the various grades with reference simply to convenience in using my eyepiece micrometer. I have always thought that the limits of the various grades should be determined by a careful consideration of the various conditions involved in the problem of proper mechanical analysis of a soil, and have been surprised to see that the arbitrarily chosen limits of the various grades employed by me have been followed by others in applying the method in practice. Hopkins' solution to the difficulties was to devise a system which provided for a constant ratio of diameters, surfaces and volumes between particles and a constant theoretical percentage of composition from an assumed uniform grade. This in Hopkins' view would allow significant features of the particle size distribution to be more easily seen. For his classification, Hopkins chose the square root of 10 as the factor on which the increments were based. This value is approximately 3.2. His classification and the relationship between the groups are shown in the table below. Theoretical Div. Ratio of Ratio of Ratio of percentage of NoL Name SizeI mm diameters surfaces volumes composition 1 gravel Z>l - - - 12.5 2 coarse sand l-O.32 3.2 10 32 12.5 3 medium sand 0.32-0.10 3.2 10 32 12.5 4 fine sand 0.10-0.032 3.2 10 32 12.5 5 coarse 6 medium Silt 000100-000032 3.2 10 32 1205 7 fine silt 0.0032-0.001 3.2 10 32 12.5 8 clay ‘<0.001 - - - 12.5 In addition, he also suggested that further subdivisions could be made based on the fourth root of 10. This, in effect. divides each subdivision into two parts which maintain the relationship between classes that Hopkins felt was important. Although Hopkins had not designed his system as such, the separations which he made were related in many cases to capillarity and porosity of the soil. This classification. while including many of the elements which have been acknowl- edged as desirable, was never used extensively in soils studies. In 1904 a paper was written by Briggs, Martin and Pearce 20 which would have great impact on particle size studies for agricultural purposes. In this paper, they compared the grain size classification of Hilgard. Osborne and Hopkins. The conclusion that they reached was that the Osborne scale was best suited to their needs. In 1905 Atterberg. from Sweden, began publishing his system of particle size classification. Perhaps no other man or system of particle size classification has had such an effect on the particle size work of agricultural soil scientists. The studies of soils by engineers has also been affected by his work on the measurement of physical properties of soils. His system was well received in Europe when first proposed. It was adopted by the International Society of Soil Science in 1913 (Baver 1956) and has been used exten- sively for soil studies. It should be noted that while the United States and England report data in terms of the Inter- national System, they also work in grade scales which have been developed in their own countries. The Atterberg classification had two features that led to its wide acceptance. First. it was related to physical properties. There might be some question about some of the separations which Atterberg made, but certainly 2.0mm and .002mm are generally accepted today. The second feature is that the divisions are of logarithmic nature and can be better analyzed by statistical means (Krumbein & Pettijohn 1938). Atterberg's classification was set up in this manner: 21 Size 'Klipp block > 2m Boulders - Stenblock 20-6dm _ Blocksten 6-2dm Fcoarse rock 20-6cm Pebbles - _broken stone 6-2cm ' coarse gravel 20-6mm gravel - g fine gravel 6-2mm ' coarse sand 2-0.6mm sand - _ fine sand 0.6-0.2mm ' very fine sand 0.2-0.06mm very fine sand « __rock flour 0.06-0.02mm P Silt OsOZ‘OoOOém silt ~ 1 slime. silt, mud 0.006-0.002mm clay <0.002mm Atterberg made his main separation at 20, 2, .2. .02 and .002mm or two times powers of 10. It was at these that he saw changes in the physical conditions of soil. Further separa- tions were made at six times powers of 10. These are based on the fact that two times 110 = 6.32. This was rounded off to 6 and used as an even division in a logarithmic plot. These latter subdivisions were not significant physical sepa- rations as originally considered by Atterberg. Atterberg considered his groupings for these reasons (Krumbein & Pettijohn 1938): The largest category, boulders, included material larger than 200mm or thereabouts was not moved by wave activity on the shore under conditions he 22 observed. Material smaller than this was freely shifted by waves and current. Material between 20 and 200mm was called pebbles. They could be displaced by currents. but their packing properties were not desirable for roads. Material between 2 and 20mm was called gravel. This size was sepa- rated from larger material by packing characteristics and from finer material because of water relationships. Sand was defined as particles between 2 and .2mm. This was different from the very fine sand which.was between .2mm and .Ome in that the coarser sand was water permeable sand and the finer was a water retaining sand. Atterberg chose .02 as the separation between sand and silt. His observations noted a difference in the rate of capillary movement in materials finer than .02mm. It was not so much the height of capillary rise as it was the speed with which it took place. This size seemed to be an upper limit to aggregates which formed from finer material in the soil. It also appeared that this was the limiting size for penetration of roots into interparticle spaces. As a last point. this is also a limit to observation of the particle with the naked eye. The separation of silt from clay was at .002mm. The reason for this was the effect of Brownian movement on the particles. At this point. the effects can be barely noticed. but in larger particles it is nonexistent and in finer particles it is much more apparent. Other reasons for the significance of this particular separa- tion have been made and will be discussed later. Atterberg was not totally satisfied that his particle 23 size classification was as good as it could be. He suggested after his initial paper that .3. .03 and .003mm would be better points at which to make the separations (Roderick 1963). He cited the same reason for each division and its physical significance. He was faced with a real problem in his classification. He wished to maintain the even divisions in a logarithmic plot and still make separations which.were physically significant. The fact is that rarely do these coincide. Later, further observations had shown him that .2mm was a better measure than .3mm for separating the water retaining and permeable sand. No changes were made in the Atterberg system because of its wide use and the reluctance to change once the system was adopted. Even the argument that less time was required to measure a system of coarser particles failed to convince anyone. Widespread acceptance of this method of particle size classification in Europe did not stop the search for a better system. especially in the United States and Britian. Atterberg thought that the U.S. Department of Agriculture classification had too many separations in the sands and not enough in the silts and clays. He also felt that the .005mm clay limit was unrelated to physical properties. On the other hand. Hilgard was not in favor of Atterberg's large range for coarse sand. He thought more separations should be made in the sand fraction than Atterberg had allowed. In 1914. an International Commission from those interested in mechanical and physical investigations of soils 24 met to consider a uniform grade scale to be used by all scientists. They had several scales from which to choose. One scale which was considered was that used by Frosterus who had done work in the podzol zone. His system as reported by Schucht (Roderick 1963) made these separations: §izeI mm gravel 20-2 coarse sand 2-0.2 fine sand 0.2-0.l very fine sand 0.1-0.02 silt 0.02-0.002 clay (0.002 The classification was much like that of Atterberg's. except that there was an additional separation in the finer sand fraction. The American Society of Agronomy also suggested a system and was represented by its chairman. Coffey. Coffey's main interest was in soil classification. and he had worked on a soil map of the United States. The classification suggested set up these limits: M coarse sand 2-0.7 medium sand 0.7-0.2 fine sand 0.2-0.07 coarse silt 0.07-0.02 medium and fine silt 0.02-0.002 clay <0.002 25 The origin of this particular classification is not certain. It has a remarkable similarity to the Atterberg classification with its secondary subdivision inserted in the coarser fractions. Also available to the commission was the U.S. Department of Agriculture Bureau of Soils Classification which had been in use for several years at this point. In the end. Atterberg's classification was accepted by the commission and has since been referred to as the Inter- national System. The limits which are used in the classifi- cation are shown below: W gravel > 2 coarse sand 2-0.2 fine sand 0.2-0.02 silt 0.02-0.002 clay <0. 002 At the same time. developments in the techniques for particle size analysis were also being made by several workers. Decantation. one of the first methods. was shown by Hall (1904) to be effective in clay separations after a 24-hour settling period. Briggs. Martin and Pearce (1904) in the same year showed that a centrifuge could be used to make separations in the fine fraction after the sand had been separated. This resulted in a considerable timesaving step in this method. Elutriation was also improved and made more precise. 26 The work of Kopecky in 1901 (Krumbein 1932) stands as the model for further work in rising current elutriators. He arranged in series three conical elutriators. each with a precise diameter. The series was so arranged that water entered the bottom of one. exited at the top and then went to the bottom of another to repeat the process through the final elutriator. For this prototype. he chose diameters of 3.0cm, 5.6cm and 17.8cm. For his particle size analysis. Kopecky recommended an expanded particle size classification similar to Atterberg's. He used these separations: §i§24_mfl gravel > 2 coarse sand 2-O.6 medium sand 0.6-0.2 fine sand 0.2-0.06 coarse sand 0.06-0.02 medium silt 0.02-0.006 fine silt 0.006-0.002 clay < 0. 002 Elutriation using water as a medium had become a fairly standard technique. It is not. however. the only application of the rising current separation. Air can also be used as was shown in an apparatus described by Gary in 1906 (Krumbein 1932). This was a cylindrical arrangement with a space at the bottom for an air blast. The force of the blast was sufficient to blow away small sizes while leaving the coarser. This has never been reported in extensive use. 27 Another more usable air elutriator was developed by Cushman and Hubbard (1907). It was similar in design to the water elutriator developed at approximately the same time. There were five vessels connected together in a series. The first container had a capacity of three gallons: the second two gallons; and the remaining three. one gallon each. The path of the air current was much the same as that of the water in the elutriators that have previously been discussed. The air enters the bottom of the vessel and leaves through the top and on to the next. Care must be exercised just as with water elutriation to avoid loss of particles. To prevent loss of particles and to adjust the current. a suction valve was placed on the last container in the series. In 1912. a much debated development was made in sedimen- tation cylinders. Atterberg modified the sedimentation cylinder which had been used up to that point. He added a glass tube near the bottom controlled by a pinchcock that allowed him to remove a sample after a given period of settle- ment. This was somewhat like a method of analysis proposed by Schubler earlier and had certain features which related it to the pipette method of analysis. The procedure was to take a measured sample and find the amount of particles contained. and the size of the particles could be determined by calculation from Stokes' Law. The method was controversial. and several people particularly in Europe attempted to evaluate the validity of the method. but no consistent conclusion was reached by these workers (Krumbein 1932). 28 Up to this point. all attempts to evaluate the particle size distribution in soils had involved a physical separation of particles. A new technique involving a continuous sedi- mentation balance was introduced by Oden in 1915 (Krumbein 1932). The object was to measure the weight of particles. P. sedimented in a given time. t. By plotting P vs. t. a particle size distribution curve results by calculating the size on the curve from the t values. The apparatus for this continuous measurement was by weighing P on a platform balance in the sedimentation cylinder. The other arm of the balance held a weighing pan to which weight could be added to measure the amount of sediment collected on the other pan in a given time. Oden added a feature to this system by putting an electric switch a slight distance above the pan used to balance the sedimented material. This switch was turned on when too little weight was on the pan and it automatically added more weight to the pan. This method proved effective in giving accurate particle size distribution curves. It along with Oden's mathematical analysis of sedi- mentation did much to advance the study of particle size. Oden also suggested a classification of particle size which is discussed in a later section. Perhaps the next significant development in particle size measurement came with the development of the continuous sedimentary cylinder by Wiegner in 1918 (Krumbein 1932). In the method the difference between the hydrostatic pressure of the suspension and water was recorded through time. With 29 this a particle size distribution curve could be plotted. The equipment for this determination was quite simple. It consisted of a sedimentation cylinder with a parallel manometer tube attached to the cylinder near the bottom. A stopcock separated the tubes and was opened to record the progress of the sedimentation. Efforts were made in 1921 by Scales and Marsh (1922) to evaluate the amount of material in a soil suspension from the turbidity of the suspension. Later work by Stemm and Svedberg (1925) showed that particle size distribution could also be measured by this technique. This method was checked by micro- scopic techniques and similar results were obtained. The principle on which the method was based held that a soil suspension would prevent transmittance of light through it proportional to the content of the suspended material. The pipette analysis. which is commonly used today for particle size analysis. was introduced in three independent papers during the period 1922 to 1923. Robinson (1922). Krauss (Baver 1956). and Jennings. Thomas and Gardner (1922) all published accounts of a similar method of particle size analysis. The pipette method was an attempt to measure the variation in density of a suspension during the progress of sedimentation. It was dependent on Stokes' Law for the sizes of particles which were being measured. The method consisted of drawing out a known volume of the suspension and drying to find the weight of particles. Calculation of the depth at which the sample was taken and the elapsed time of sedimentation using Stokes' Law will show the size of the material being measured. The three papers which first developed the pipette method did not differ greatly in theory. There were. however. differences in the techniques employed to take the sample of the suspension. Robinson in his method used a pipette to draw out a sample. This could, of course. be done at various levels in the suspension. Jennings. et al.. devised a fixed pipette with several openings inside the sedimentation cylinder which drained out the bottom of the cylinder through a valve. Krauss used three pipettes in his analysis. They were suspended so as to be initially 10cm 'below the surface. The Agricultural Education Association of Great Britian (1926) compared several means of particle size analysis. The pipette analysis was selected as the best for routine soil analysis. Up to this point. Great Britian had been using a particle size classification that was developed by Hall and Russell (1911). It was a system based on arbitrary selection of particle size limits but was none the less widely used in Great Britian and was as follows: §i§24_flfl fine gravel :>l. coarse sand 1-0.2 fine sand 0.2-0.04 silt 0.04-0.01 fine silt 0.01-0.002 clay (0.002 31 This system was replaced in 1928 with the Atterberg system by the Agricultural Education Association of Great Britian (1928). There was no great attachment to the Hall- Russell classification. and since it was not related to physical properties or a mathematical relationship. it was dropped to foster unityamong international workers. In 1927, Bouyoucos presented the hydrometer method of analysis which measured the density of the soil suspension (1927a. 1927b). It has been criticized for its limitations but with modifications it remains today one of the most used procedures in particle size analysis. Initially. Bouyoucos used a hydrometer calibrated in g/l which had been standard- ized from previously analyzed soils. Initially. the hydrometer reading was made at the end of 15 minutes' sedimentation time. The value which was obtained at this time was correlated with the heat of wetting. It was the conclusion of Bouyoucos that this lS-minute reading represented the colloidal fraction of the soil, because of its relation to the heat of wetting. He later changed the approach to a measure of the sand silt and clay in the Bureau of Soils particle size classification. In the method that he used. readings were taken at 40 seconds. which.was the time for settling of sand grains with .05mm equivalent diameter. and at one hour which was the time for settling material coarser than .005mm. In addition. Bouyoucos also included a third reading at two hours which gave the <:.002mm material. These values were based on empirical evidence rather than Stokes' Law. although there was a 32 relationship. Later. Bouyoucos applied Stokes' Law to the hydrometer method. In this. he sought to justify his one- hour and two-hour readings by showing that there was a difference in the concentration between the top and bottom of the sedimentation cylinder. The hydrometer measured an average value which approximated the answer by canceling errors as was done with the pipette method. It was in 1928 when Kohn. as reported by Baver (1956). in studies of solutions of lead iodide found that a spherical sample was withdrawn by the pipette method. Prior to this study, it had been considered that a layer of the suspension was being taken. The spherical sample did not affect the accuracy of the determination because of the canceling effect of larger and smaller particles on each other. There was considerable criticism of the hydrometer method, particularly from the effect of experimental procedure on the theoretical aspects (Keen 1928, Olmstead, et a1. 1931). In answer to this. Casagrande (Baver 1956) developed a revised theory to explain the action of the hydrometer. Bouyoucos (1937) at this time developed a streamlined hydrometer which was more in harmony with the theoretical considerations. The significance of the particle size limit between silt and clay was still a problem in the U.S. bureau of Soils particle size classification. It was noted that discrepancies existed between the silt and clay texture classes determined by field parties and those determined by laboratory procedure. 33 Shaw and Alexander (1936) attempted to resolve this problem by examining material in the following size groups: .05- .005mm. .005-.002mm and.‘(.002mm. They asked field soil scientists to classify these materials. They found that several soil scientists classified the .005-.002mm particles as silt. This showed that these particles acted physically more like silt than clay. Chemical tests to determine the mineralogical content showed that the .005-.002mm fraction was more related to the .05-.005mm size group than the {.002mm fraction because of higher silica-sesquoxide ratio and percent silica. Shaw and Alexander concluded that the limit between silt and clay should be set at .002mm rather than .005mm because of a closer relationship with observed properties. In that same year. Truog. et al. (1936a, 1936b) also recommended a change in the U.S. Bureau of Soils particle size limits for the clay fraction. The arguments for the change from .005mm to .002mm were based on mineralogical evidence. It was their finding that .002mm represented a significant separation between primary and secondary minerals. Below .002mm. there are very few primary minerals and nearly the entire mass of material is made up of secondary minerals which are resistant to further decomposition and alteration. By setting the size limit of clay at .002mm. much can be inferred about the mineralogy and the state of weathering in the soil which cannot be made with the .005mm limit. The U.S. Bureau of Soils classification was changed in 34 1938 so that the clay limit was placed at .002mm instead of .005mm. However. the other agencies using this system in the U.S. Department of Agriculture continued to use the former particle size limits for clay without modification. It was not until 1947 that another change was made in this system of classification. The name of the fine gravel size class which included particles between 1.0mm and 2.0mm was changed to very coarse sand. This was in line with ideas of slightly cohesive properties when referring to sands and noncohesive properties referring to gravel (Soil Survey Staff 1951). This was the last change to date made in the agricultural classification of particle sizes in the United States. The two major systems in use today are the Atterberg. or Inter- national System. and the U.S. Department of Agriculture system. Recent improvements in particle size analysis have taken the following forms for the major methods of analysis: the pipette method has relied on improvements in dispersion procedures: the hydrometer method has been improved by better design. more detailed study of the theory behind the measure- ment (Day 1950) and improved dispersion procedures. It is still generally recognized that the hydrometer results are not as precise as the pipette method (Black et a1. 1965). While the foregoing is a survey of particle size classi- fication and particle size analyses from the agricultural viewpoint. it should be noted that developments in the techniques of particle size analyses have been made by individuals in many fields. It is not as may have been 35 .implied a result of the efforts of agricultural workers alone. 2. Geological Pargicle Size Classificatign Geologists. particularly those concerned with sedimenta- tion, have an interest in particle size classification. Their classifications have usually tended to emphasize ease in treatment of data rather than relating the classification to physical properties (Krumbein & Pettijohn 1938). A summation of the geologists' uses of grade scales is found in Krumbein and Pettijohn's Manual of Sedimentary Petro- graphy (1938). They cite two functions of grade scales on page 86: descriptive and analytic. The descriptive function is. in their words: Descriptive function. The first and perhaps the most important function of a grade scale is a descriptive function. which serves to place nomenclature and terminology on a uniform basis. If one reads the term coarse sand in a report. he would prefer to understand by the term exactly what the writer intended to convey. As long as there is no uniform terminology. each writer coins his own meanings. which may or may not be precisely defined. If. however. the reader knows that the writer is using the Atterberg classification. he may understand that material having a range of sizes from 2.0mm to 0.6mm diameter is meant. LikeWise. if the Wentworth scale is being used. the term refers to material from 1mm to %mm in diameter. Obviously. no "justification" whatever is required for the descriptive function of a grade scale. The particular choice of such terms as coarse sand, fine sand. cla , and the like need to be based on no other criterion than mutual agreement. If the limits chosen for each grade are also related to the physical properties of sediments. that fact may be taken as an added advantage. 36 The analytical function is. in their words: Analytical function. In addition to the use of grade scales to establish uniformity of terminology. the classes or grades are used as units in performing various kinds of analyses on the sediment. The classes are used. for example. in determining the mechanical composition of the sediment. and in addition they may be used as units during statistical analysis. It is in connection with these analytic functions of grade scales that most of the confusion arose regarding the merits of one or another of the proposed scales. The recognition of the fact that histograms vary in form according to the grade scale used has led various writers to the conclusion that some single scale should be used for all analyses. so that the unfortunate variation of the histogram could be avoided. Perhaps the first system devised specifically for geological work was by Orth in 1875. according to Wentworth (1922). His suggested particle size classification included these separations: M grave l ) 3 very coarse sand 3-1 coarse sand l-0.5 medium sand 0.5-0.25 fine sand 0.25-0.05 dust 0.05-0.01 finest dust *(0.01 It is certain that this classification would not be the most desirable for geologists given the criteria cited previously. This classification, like the early agricultural classifications. was not concerned with the finer fractions. A somewhat more complete particle size classification 37 was proposed by Diller (1898) in a U.S. Geological Survey Bulletin. §i£24_flfl gravel < 2 fine gravel 2-1 coarse sand 1-0.5 medium sand 0.5-0.25 fine sand 0.25-0.10 very fine sand 0.10-0.05 silt 0.05-0.01 finest silt 0.01-0.0005 clay < O. 005 This classification bore a close resemblance to the particle size classification suggested by Whitney in 1892 which became the basis of the U.S. Bureau of Soils system. In 1898 Udden (Roderick 1963) made perhaps the most important contribution to particle size studies in geology. His was the first classification based on a geometric progression. He had based his particle size classification on earlier work done by agricultural soil scientists. He disagreed, however. with the specific limits set up by the soil scientists. because they did not follow a geometric progression throughout the system. The advantage of this was in the analytical function of a grade scale and not necessarily in the descriptive phase. Udden's initial grade scale made these separations: 38 §i£24_flfl coarse gravel 8-4 gravel 4-2 fine gravel 2-1 coarse sand 1-1/2 medium sand 1/2-1/4 fine sand 1/4-1/8 very fine sand 1/8-1/16 coarse dust 1/16-1/32 medium dust 1/32-1/64 fine dust 1/64-1/128 very fine dust 1/128-1/256 In this geometric progression, each size from top to bottom had to be multiplied by 1/2 and the ratio between each separation and the next higher separation was equal to 2. It was also apparent in this particle size classification that Udden was interested in getting close to the descriptive classification of the soil scientist, while maintaining the geometric progression that he considered essential to analysis of the distribution curve. Udden (1914) added to his size limits at the top and the bottom of his classification system. At the top he included: Size. mm very coarse gravel 8-16 very small boulders 16-32 small boulders 32-64 medium boulders 64-128 large boulders 128-256 39 And in the lower segment of his classification. he added these classes: .§l§2i_flfl coarse clay 1/256-1/512 medium clay 1/512-1/1024 fine clay 1/1024-1/2048 These additional classes became necessary because of the nature of the material that Udden had begun to study. His initial classification was developed for a study of wind deposits. These later developments were a result of his studies of clastic sediments which afforded a larger range in particle sizes. It was at this time that the Atterberg classification was gaining popularity among agricultural soil scientists. It also appealed to geologists because: first. it was in the form of a geometric progression, and second, it was in some ways related to descriptive properties. Many geologists at this time used the Atterberg system. In 1908. Keilhack (Wentworth 1922) suggested a particle size classification for geologists. It was not as complete as the Udden or Atterberg systems and was not extensively used. 40 w gravel > 2 very coarse sand 2-1 coarse sand l-0.5 medium sand 0.5-0.2 fine sand 0.2-0.l superfine sand 0.1-0.05 dust 0.05-0.01 finest dust <0.0l Grabau (1913) reviewed the systems of particle size proposed by Diller. Keilhack and others. As a result of his study, he developed a classification of his own. He included these fractions in his grade scale: m boulders > 150 cobbles 150-50 very coarse gravel 50-25 coarse gravel 25-5 fine gravel 5-2.5 very coarse sand _2.5-1.0 coarse sand l.0-0.5 medium sand 0.5-0.25 fine sand 0.25-0.10 superfine sand 0.10-0.05 rock flour 0.05-0.01 superfine flour 0.01-0.005 clay size 0.005-0.00l 41 This classification is closely related in the finer fractions to the U.S. Bureau of Soils classification. In 1918. Boswell (Milner 1952) published a study of sands in various industries in Great Britian. For this study. he selected these particle size limits: SizeI mm gravel > 2 very coarse sand 2-1 coarse sand 1.0-0.5 medium sand 0.5-0.25 fine sand 0.25-0.10 superfine sand or coarse silt 0.10-0.05 silt .05-.01 clay or mud (0.01 Above .05mm this is the same as the U.S. Bureau of Soils classification with some modification in terminology. The publication of wentworth's particle size classifi- cation in 1922 is commonly hailed as the most important event in geological study of particle size. His classifi- cation of grain sizes was in effect a modification of the grade scale originally developed by Udden. Wentworth made these separations in his size classifi- cation: 42 SizeI mm boulder gravel > 256 cobble gravel 256-64 pebble gravel 64-4 granule gravel 4-2 very coarse sand 2-1 coarse sand 1-1/2 medium sand 1/2-1/4 fine sand 1/4-1/8 very fine sand 1/8-1/16 silt 1/16-1/256 clay <11/256 The system was like that of Udden's, based on a geometric progression with a factor of 2 and using 1mm as the base with some intervals left out. Wentworth felt this system was good for two reasons. First. as a geometric progression. it provided what he considered meaningful comparisons between the size groups. It also was a simple matter to graph and interpret these even units on a logarithmic scale. These were also consid- erations in the earlier Atterberg system. Secondly. he felt the separations which he made were as close as possible to those already being made by most workers, while also being. for the most part. a geometric progression. It is significant to note that Wentworth was also concerned to some degree about the relation of physical properties to the size classes which he had determined. In 43 a paper in 1933. he showed how he assumed the sizes were related to mode of transportation of the particles in geologic processes. His conclusion was that even though the separations may have been of an arbitrary nature initially. they were closely related to some critical properties as shown in the following chart. / Tract/ea /:r :':A fl); (a ’70! day] .5“): a." (on \ :‘K ,W/‘I’I. I‘CMH inpo 4'”- ‘0 4+4 L itex ”I carafe/- I! ”I; nefgr; 254 - i f It seems to assume that there are sizes which are not moved in the same continuum as others. This does not agree with observations of river deposits cited later in this work. A change was made when Krumbein (1934. 1936) suggested the modification of the Wentworth and Atterberg grade scales to the phi and zeta grade scales, respectively. The modifi- cation is nothing more than a logarithmic transformation of these scales using a transformation equation on each size limit diameter. For the transformation of the Wentworth grade scale. this equation is used: ¢ = -log 2 2; (when Z, 'is the diameter of the size limit in millimeters). For the transformation of the Atterberg grade scale, this equation is used: Z, = 0.301 - log 10 2, (when I, is the diameter of the size limit in millimeters on the Atterberg scale). The result of these transformations is shown as follows: 44 Phi and Zeta Grade Scales Wentworth ¢ Atterberg 2; Grades Grades 32mm -5 2000mm -3 16mm -4 8mm -3 200mm -2 4mm -2 2mm -1 20mm -1 1mm 0 l/2mm +1 2mm 0 l/4mm +2 l/8mm +3 0.2mm +1 l/l6mm +4 l/32mm +5 0.02mm +2 l/64mm +6 1/128mm +7 0.002mm +3 1/256mm +8 1/512mm +9 0.0002mm +4 1/1024mm +10 The theory behind the transformation to these scales is that any geometric scale can be changed to a scale with arithmetic intervals which is equivalent, provided that logarithms of diameter sizes are substituted for diameter sizes. The intent of these transformations to the 49 and Z, scales is to provide a better. more easily manipulated framework for the application of statistical methods of analysis. 45 Hilgard had earlier developed a classification of grain size for agricultural purposes based on the speed of elutri- ation. Along this same line. Rubey (1930) developed a classification for geologists based on settling times. This was developed in relation to his study of fine grained Upper Cretaceous sedimentary rocks in the Black Hills. His classi- fication made these separations: Rubey's Size Classification Based on Settling Velocities Settling Velocity Grade (in microns/sec. very fine sand >-3.840 coarse silt 960-3.84O medium silt 240-960 fine silt 60-240 very fine silt 15-60 coarse clay 3.75-15 medium clay 0.9375-3.75 fine clay (0.9375 Rubey arrived at these limits by plotting settling velocities vs. diameters of particles on double log paper. He then plotted the size limits of Atterberg, Wentworth and Udden on the graph. The settling velocities cited in the table were obtained from this graph. It is significant to note at this point that Rubey's classification is a geometric progression. Each settling velocity value is related to the next higher by a factor of 4. The particles themselves are related to the next higher 46 particle by 2 because of the nature of Stokes' Law. This classification has not been used to any great extent and no transformation such as the phi and zeta scale has been worked out. although an equation could be easily developed. Alling (1943) proposed another classification of particle sizes for geologists. His scale was developed from studies of thin sections and polished blocks. Alling was not considering particle size analysis in the usual sense of sedimentation or sieving. He was interested only in the two dimensional aspects of particle size. Alling in his article clearly stated four fundamental properties he considered essential to a good particle size grade scale. They are: l. The grain sizes should constitute a continuous series. 2.. Any division of the series will be arbitrary. 3. Convenience of use is a criterion. 4. Statistical analysis requires the use of a constant geometric ratio. In line with those principles, he proposed this grade scale: Size, mm _ Coarse 560-1000 medium 320-560 Boulder a fine 180-320 _ very fine 100-180 47 ' coarse 56-100 medium 32-56 Cobble - fine 18-32 - very fine 10-18 ' coarse 5.6-10 “ladium 302-506 Gravel q fine 108-302 L very fine l.0-l.8 P coarse 0.56-l.0 mdium 00 32-0. 56 Sand . fine 0.18-0.32 ..very fine 0.10-0.l8 ’ coarse 0.056-0.10 medium 0.032-0.056 Silt 7 fine 0.018-0.032 _ very fine 0.010-0.018 ’ coarse 0.0056-0.010 medium 0.0032-0.0056 Clay - fine 0.0018-0.0032 L very fine 0.0010-0.0018 ” coarse 0.00056-0.0010 medium 0.00032-0.00056 Colloid - fine 0.00018-0.00032 _ very fine 0.00010-0.00018 It was based on Hopkins' work which.was an alternative to the U.S. Bureau of Soils classification proposed in 1899. Hopkins had suggested using a geometric progression with a 48 factor of 10 from one size limit to the next higher size limit. He also made subdivisions based on {10 and suggested 4410 as a possibility for further subdivision. Alling used 4'-J']-.-0-to develop the system which he proposed. The American Geophysical Union has attempted to get uniform acceptance of a grain size terminology. In 1947. after thorough study of many systems. a subcommittee of the American Geophysical Union recommended this system: very large boulders 4096-2048mm or l60-80in large boulders 2048-1024 80-40 medium boulders 1024-512 40-20 small boulders 512-256 20-10 large cobbles 256-128 10-5 small cobbles 128-64 5-2.5 very coarse gravel 64-32 2.5-1.3 coarse gravel 32-16 1.3-0.6 medium gravel 16-8 0.6-0.3 fine gravel 8-4 0.3-0.l6 very fine gravel 4-2 0.16-0.08in very coarse sand 2-1 coarse sand 1-1/2 1-0.500mm medium sand 1/2-1/4 0.500-0.250 fine sand 1/4-1/8 0.250-0.125 very fine sand 1/8-1/16 0.125-0.062 coarse silt 1/16-1/32 0.062-0.03l medium silt 1/32-1/64 0.031-0.016 fine silt 1/64-1/128 0.016-0.008 very fine silt 1/28-1/256 0.008-0.004 coarse clay size 1/256-1/512 0.004-0.0020 medium clay size 1/512-1/1024 0.0020-0.0010 fine clay size 1/1024-1/2048 0.0010-0.0005 very fine clay size 1/2048-1/4096 0.0005-0.00024mm 49 It is basically the system of Udden with extensions at both ends of the scale. 3. Particle Size Classification in Engineering Engineers have been concerned with the particle size classification and distribution in earthy materials. Although perhaps beginning the study of particle size later than most groups. they have been keenly interested in the practical aspects of the particle size information. The first particle size classification specifically suggested for engineers was advanced by Terzaghi in 1925 (Glossop & Skempton 1945). It was a compilation of earlier work done by Romann and Atterberg. He used Atterberg's size limits for the fine fractions and Romann's size limits for the coarse fractions. His size classification made these separations: SizeI mm very coarse sand 2-1 coarse sand l-0.5 medium sand 0.5-0.2 fine sand 0.2-0.l coarse mo 0.1-0.05 fine mo 0.05-0.02 coarse silt 0.02-0.006 fine silt 0.006-0.002 coarse clay 0.002-0.0006 fine clay 0.0006-0.0002 ultra fine clay (0.0002 The particle size Mo comes from a Scandinavian word for very fine sand or silt. It not only is used to refer to the 50 particle size. but also the condition of land since it refers to sandy barren heath and a country drill field. This system became known as the Continental System and an addition was made to Terzaghi's particle size classifica- tion to include larger material. .s_i_z.e_._r_n_m stone > 30 coarse gravel 30-15 medium gravel 15-5 fine gravel 5-2 The U.S. Bureau of Public Roads developed guidelines for particle size classes for soils study early in the 19205. These were based on the work of Boyd (1922) and Goldbeck (1921). An arbitrary classification was evolved that was referred to as the Bureau of Public Roads system (Hogentogler 1937): SizeI mm Sigyg gravel > 2.0 (No. 10) coarse sand 2-0.25 (No. 10 - No. 60) fine sand 0.25-0.05 (No. 60 - No. 270) silt 0.05-0.005 clay < 0.005 colloids 2 coarse sand 2-0.6 medium sand 0.6-0.2 fine sand 0.2-0.06 coarse silt 0.06-0.02 medium silt 0.02-0.006 fine silt 0.006-0.002 clay < 0. 002 This classification also contained the same separations as the expanded Atterberg classification. The relationship between particle size and physical properties will hold for this calssification. Further work by Glossop and Skempton (1945) showed that engineering properties were related to those size separations as well as those advanced by Atterberg and others earlier. Specifically. Glossop and Skempton said 2.00mm was considered to be the upper size limit of capillary forces. 54 Below 2.00mm the particles cling together when moist. This is the sand fraction. Above 2.00mm the particles are too large to be held by these capillary forces. The separation between silt and clay was made at .002mm because of the colloidal and base exchange properties cited by Russell. These properties increase below .002mm and are practically nonexistent in coarser fractions. The separation between sand and silt is not nearly as clear as the other breaks but study in three areas shows that .06mm represents a good point to make the separation. From a textural standpoint there appears to be a break in the grittiness of the material at .06 or .07mm. There is also a break in the size of wind- blown sands at approximately .06-.08mm. Frost heaving is another of the properties considered. It was shown that silt was more likely to frost heave than was clay. It was also found that ground water lowering is easily accomplished in material above .06mm and very difficult below .06mm. They also mentioned the increased consolidation and settlement in silts over sands. Glossop and Skempton found on the basis of natural samples that .06mm was the value which was most closely related to these several properties separating silt and sand. In 1947, the American Society of Engineering Education through its civil engineering division advanced a size classification (ASCE 1947). They made these separations: 55 SizeI mm .gigyg coarse gravel 76.2-25.4 3"-1" medium gravel 25.4-9.52 l“-3/8" fine gravel 9.52-2.0 3/8"-No. 10 coarse sand 2.0-0.59 No. lO-No. 30 medium sand 0.59-0.25 No. 30-No. 60 fine sand 0.25-0.074 No. 60-No. 200 coarse silt 0.074-0.02 No. 200 fine silt (0.02 non-plastic No Lower Limit clay ‘<0.074 plastic No Lower Limit This group backed their classification with claims that it is for the most part based on properties of particle size groups which can be recognized either in the field or the laboratory. In 1947, Casagrande, who had been responsible for the development of the Unified Soil Classification System during World War 11. proposed a grain size classification. His grain size classification was based on sieve separations of coarse particles and the fine fraction was classified on the basis of Atterberg constants. He made these separations: SizeI mm Sieve cobbles > 76.2 3" coarse gravel 76.2-19.5 3"-3/4" fine gravel l9.5-4.76 3/4"-No. 4 coarse sand 4.76-2.00 No. 4-No. 10 medium sand 2.00-0.42 No. lO-No. 40 fine sand :<0.42-0.074 No. 40-No. 200 fines (silt and ‘<0.074 (classified ‘(No. 200 clay) as to plasticity and cohesion) 56 In 1950. American Association of State Highway officials revised their classification of particle size to adopt many of the Casagrande particle size limits (Roderick 1963). Their revised classification makes these separations: Size mm sears particles larger than coarse sand fine sand silt clay colloids 2-0042 (NO. 10 " NO. 40) 0.42-0.074 (No. 40 - No. 200) 0.074-0.005 (0.005 (0.001 In 1958, the American Society for Testing and Material also adopted the above classifications. In 1961. further revision of the particle size classification resulted in an expanded system. as follows (Roderick 1963): gravel coarse sand medium sand fine sand silt clay colloids SizeI mm 76.2-4.76 4.76-2.00 2.00-0.42 0.42-0.074 0.074-0.005 (0.005 (0.001 £19.12 3"-No. 4 Sieve No. 4 - No. 10 No. 10 - No. 40 No. 40 - No. 200 57 B. MOISTURE RELATIONSHIPS AND PARTICLE SIZE IN SOILS Moisture relationships have long been considered important considerations in particle size classification and particle size distributions. Atterberg was perhaps the first to set forth the relatioships of particle size and soil moisture as reviewed above. He discussed elements of fluvial hydraulics as well as water holding properties. It has been clearly shown that there is a relationship between particle size and a location on a traverse along a river (Leviavsky 1955). as shown in the data presented by Sternberg and the analysis by Schiklitsch. They showed a semilogarithmic relation between distance from source toward the mouth and particle size. Coarser particles are found at or nearer the source and finer at or nearer the mouth. There is still debate as to the cause for this relationship. Leopold. Wolman and Miller (1964) cite a near linear relation between particle size and distance traveled along the Mississippi and Missouri Rivers. Hack (1957) also worked on this subject and attempted to find a relationship to both particle size and lithology. Water holding capacity has been of real interest in agriculture. It has been particularly critical in relation to droughtiness of soils. irrigation and other water manage- ment practices. Several studies have been made of the moisture holding properties and various particle size group- ings in an attempt to establish relationships between them. At this point, some of the results appear contradictory but 58 it seems that a definite relationship has been found by several workers. Initially the work on soil moisture and soil texture stressed measurement of field capacity and moisture equiva- lent and the relation of each of these to texture and to each other. Perhaps the first study of this type was by Briggs and Shantz in 1912 (Baver 1956). They found a relationship between the moisture equivalent of soils and the percentage of the sand. silt and clay fraction. Problems with irrigation have created further need for study of available water capacity and readily available water capacity. Wilcox and Spilsbury (1940a, 1940b) observed that available water found by deducting wilting coefficient from field capacity 24 hours after irrigation was related to silt and clay contents. According to them, an increase in clay content up to 60 percent increased the available water in the soil. It appeared too that over 60 percent clay. the water is too firmly bound in surface films to be avail- able. Peele et a1. (1948) found in a study of the irrigation requirements of South Carolina soils that the available water determined as field capacity minus permanent wilt point was primarily dependent on the type of clay and the amount of organic matter content. Briefly stated, three relationships were found. Soils low in both clay and organic matter had lower moisture holding capacities. Soils high in organic matter were high in water holding capacity. Soils high in 59 clay had high and relatively low water holding capacities dependent on the type of clay mineral that was present in the profile. If the clay exhibits low swelling and shrinkage on wetting and drying or has low base exchange capacity, it will also have a low available water capacity. A study made by Lehane and Staple (1953) on Brown and Dark Brown prairie soils in Canada reports that clay is the dominant controlling factor in water availability. They found that more available water was held by clay textured soils than was stored in silt loams and silty clay loams. Available water in this study was considered as the field capacity 72 hours after saturation minus the wilting coefficient. In Holland, Ferrari et a1. (Hill 1959) showed clay to be the most significant measure of available water holding capacity. It should be noted that his clay determinations included material less than .016mm in diameter. This view that clay is the particle size in control of available water is no longer held by many workers in the field. In fact. much of the work done in recent years has tended to discount the role of clay in predicting available water capacity. Some results show that increase in clay content actually decreases the available water capacity (Jamison & Kroth 1958). Recent research shows the importance of the silt fraction to the available water capacity. Perhaps the first publication to cite silt as being more effective than clay in holding available water was by 60 Jamison (1956). Data in that article showed that silt loam held more water available to plants than did clay soils. Later. Jamison and Kroth (1958) studied soil profiles from several geographic locations in Missouri. They found that the available water storage capacity of a soil increased with increasing silt content. The coarse silt fraction (.05mm-.02mm) was more effective in increasing the available water capacity than the fine silt (.02mm-.002mm) fraction. In most cases they found higher clay contents actually lowered the available water capacity. In some soils with between 13 and 20 percent clay, however. the available water capacity increased with clay content. Jamison and Kroth suggested that silt size soil aggregates formed with organic matter may have been responsible. This clay relationship reminds one of Peele et a1. (1948) and the differences they found in available water capacity associated with clay mineralogy. Jamison and Kroth contend that available water is dependent on particle size and is not affected by organic matter. Others such as Peele et a1. (1948) and Jacks (1964) contend that organic matter increases available water holding capacity. In 1959, Lund (1959) in a study of Alluvial soils in Louisiana found results similar to Jamison and Kroth. He found a direct correlation between available water measured between 1/3 and 15 atmospheres tension and total silt. Lund also found a direct relation between the clay content and the total moisture content at 15 atmospheres tension which 61 was considered by him to be the wilting point of most crops. In addition, he found no correlation between clay and available water capacity and a negative relation between sand and available water. He also reported a positive correlation between pore size and permeability. His conclu- sion was that total water was related to the clay fraction but available water was related to the silt fraction. Bartelli and Peters (1959). describing moisture characteristics as unique and of great utilitarian value. sought to relate them to the classification of soils. Their study was made in Illinois on soil types representing Gray Brown Podzolic. Brunizem and Planosol Great Soil Groups. They found that the Gray Brown Podzolic soils averaged more available water regardless of texture than the other Great Soil Groups in the surface horizon. They found, as had Lund, that clay was related to the water content at 15 atmospheres tension but was not related to the available water capacity. They also found that the silt content was related to the available water capacity but not to the water content at 15 atmospheres. Hill (1959) in a bulletin on the storage of moisture in Connecticut soils used some different measures of soil water. Through these he showed that silt was the most important particle size class in determining the available moisture holding capacity. He also found that moisture release at high tensions was closely related to clay content. He also reported a relation between capillary porosity and the silt 62 content of soils. The correlation coefficient was .73. From his data. organic matter was not highly correlated with avail- able water. He found that cultivated soils had higher available moisture holding capacity than did forested soils. This seems to be in agreement with the finding of Jamison (1953) and with Feustal and Byers (Baver 1956). It is not however in agreement with the findings of all researchers. Millar. Turk and Foth (1965) stated clearly that organic matter increases the available moisture holding capacity. Earlier. Alway and Russell (1916) found organic matter added to moisture holding capacity. Later. Peele et al. found with soils. all of sandy loam texture, that organic matter content and available water was highly correlated. The question of organic matter effect on moisture holding capacity is still not completely clear. Franzmier et al. (1960) found that soils high in silt and very fine sand in the USDA particle size classification are also high in readily available water. Readily available water was defined in this case as water held between .06 and 6 atmospheres tension. He also found that the clay content was highly correlated with the amount of water present at 6 atmospheres tension. This would indicate two regions of available water: one below 6 atmospheres controlled by silt and very fine sand, and another above 6 atmospheres controlled by the clay content. Previous studies had centered on capacities at 15 atmospheres with no measurement between it and field capacity. 63 Salter and Williams (1965a, 1965b, 1966) in a study of the influence of soil texture on the moisture characteristics of soils showed clearly the effect of particle size on moisture holding properties. They found that medium textured soils held the most available water. They also found that silt size particles were most effective in holding available water. C. CAPILLARY RISE AND PARTICLE SIZE IN SOILS Capillary rise was a property which was recognized early in soil studies (Dougrameji 1965). This has an effect on plant growth (Luthin 1957, Hogen et a1. 1967) and on the engineering uses of soils (Terzaghi 1948, Hough 1957. Capper & Cassie 1953). The effect of capillary rise results in ques- tions which have practical as well as theoretical significance. The first correlation of particle size and capillary rise was found by Schnaskey in 1864 (Dougrameji 1965). It was based on the data collected earlier by Schubler in his studies of soil physics. He is also credited with developing the concept of capillary and noncapillary pores and their effects on moisture retention. He realized the practical aspects of this on plant growth and maintenance of good soil conditions. Wollney. as reported by Baver (1956). also found a relation between particle size and capillary rise. He found that capillary rise was affected by several factors including particle size. An increase in capillary rise was noted with 64 increases in temperature, looseness of packing and original moisture content of the soil. He found that mixtures of sand sizes were more effective in raising capillary water than were narrow ranges of sand sizes. He also found that the capillary rise was highest with a capillary diameter of .O5-.10mm with the lowest limit of capillary rise at 2.0mm. In addition to this. he also established the highest total rise with the finest particle size. The findings of Wollney were borne out later in experi- ments by Loughbridge (1894). He too found a relation between coarse silt size particles and higher capillary rise. He went further to say that in the sand fractions. both capillary and noncapillary water is a factor since both kinds of pores are available. Capillary rise is reduced in dry soil and increased in a moist soil. King (1907) thought that if a thin. dry layer of soil was maintained at the surface. evaporation would be reduced. It had been recognized earlier by Johnson (Baver 1956) that total capillary rise was not as significant as the rate of capillary rise. In this he of course had reference to plant utilization of the water. Studies of the rate of capillary rise in relation to particle size were made by Atterberg (1908, from a trans- lation by Dr. R. W. SimonsonL He studied the capillary rise of selected size fractions in 24 and 48 hours as well as the time required to reach the maximum capillary rise. He found 65 that the most rapid capillary rise was in the fraction between .02mm and .05mm in both 24 and 48-hour periods. He also found that the days required for maximum capillary rise were greatest in the .05mm to .lOmm fractions. In the study of capillary rise in soils, two explanations of capillary rise have been offered. First was the capillary size hypothesis of Briggs and later the capillary potential hypothesis of Buckingham. Much work has been done to explain these relationships. It seems that most modern work shows the capillary potential to be most useful (Baver 1956). Regardless of the theoretical approach, it can be seen that different conditions affect the result. Several studies have been conducted on the effects of stratification of soils on capillary rise (Dougrameji 1965). Disruption and variation of capillary rise have been noted in such materials but no specific size differences necessary have been cited. However. general size limits have been set in the sand fraction by Mamanian and later by Felitsiant (1959, 1961). Dougrameji found a 2cm layer of 2.5-lmm diameter particle will stop capillary rise. A similar layer of 1-.5mm particle will not stop capillary rise. relative to adjacent sizes. Later work on capillary rise in the sand and coarse silt fractions by Dougrameji and Erickson in 1962 showed a similar trend to the results obtained by Atterberg. They showed that the coarse silt fraction gave the highest capillary rise in 48 hours. Earlier work reported by Franzmeier et a1. (1960), while choosing different size limits of separations than 66 Atterberg, showed a similar relation between silt sizes and capillary rise in 48 hours. They were not able to obtain as high a rise in 48 hours as did Atterberg. This variation could be due to a difference in techniques or conditions of the experiments. D. PERMEABILITY AND PARTICLE SIZE IN SOILS Permeability measurements in soils with saturated and unsaturated flow of water have been used to help evaluate the physical condition of soils by many authors. Some consider permeability to be the best criterion for measuring soil structure (Fireman 1944). Others approach it from the point of view of measuring capillary and noncapillary pores (Baver 1938). Others propose that permeability in disturbed samples is supplemental to capillary rise measurement. the latter reflecting the amount of capillary pores and the former the noncapillary pores. Permeability and a closely related property percolation rate are of real interest to sanitarians (US Dept. HEW, PHS 1967, Bender 1961) and to engineers (MSHD 1960). This discussion of permeability will be limited to a study of saturated flow. Initially, studies of saturated flow were related to the work of Darcy (Baver 1956). The law which he developed, Darcy's Law, can be mathematically stated as follows: 67 Where v is the velocity in cubic centimeters per second, h is the difference in pressure head in centimeters, l is the length of the column in centimeters and k is the proportionality constant. It does not relate permeability to the particle size of the material concerned. Slichter thought that a more effective formula for saturated flow would include soil properties as well as the external factors: as a result, he arrived at this equation: q = 10.22 pdzs fAhk Where q is the quantity of water in cubic centimeters per second, p is the difference of the pressure head in centi- meters, d is the mean diameter of the soil grains in centi- meters, 3 is the cross sectional area in square centimeters, h is the height of the soil column in centimeters. V is the coefficient of viscosity in poises, and k is a constant. Later, Zunker (Baver 1956) also found that soil properties were important in calculating permeability. He modified Slichter's equation of expected permeability as follows: a Q—Lpfu (.29.) F When Q refers to amount of transmitted water, h is the pressure difference, q is the coefficient of viscosity. .1 the length of the column, U the effective specific surface. ft the type and arrangement of particles, p the total pore 68 space, po the tension free pore space, and F the cross sectional area. Zunker placed the values for p from 2.3 for round particles to .5 for disk-shaped particles. Since most soils are a mixture, 1.0 was the value placed on most soils. He also developed a formula for finding the tension free pore space, p0. P0 = p - (W/IOO) (l-p)s In the formula, W is twice the hygroscopicity. s is the specific gravity of soil, and p is total pore space. Because of its relevance for engineers, several studies have been made to show the relationship between particle size and permeability. As Ritter and Paquette (1960) stressed. permeability is a result of void ratio, grain size and structure. The void ratio is defined as the ratio of the volume of voids contained in a soil mass to the volume of solids. Some data specifically relating permeability and grain size of some materials are found in the following table from Terzaghi and Peck (1948). 69 Formation River Deposits Rhone at Genissiat Small streams, eastern Alps Missouri Mississippi Glacial Deposits Outwash plains Esker, Westfield, Mass. Delta, Chicopee, Mass. Till Wind Deposits Dune sand Loess Loess loam Lacustrine and Marine Offshore Deposits Very fine uniform sand, Bulls's liver, Sixth Ave.. N.Y. Bull's liver. Brooklyn Clay Value of k (cm/sec) Up to 0.40 0.02 to 0.16 0.02 to 0.20 0.02 to 0.12 0.05 to 2.00 0.01 to 0.13 0.0001 to 0.015 Less than 0.0001 0.1 to 0.3 0.001: 0.0001: 0.0001 to 0.0064 0.0000 to 0.0050 0.00001 to 0.0001 Less than 0.0000001 They have also included a classification of soil materials based on permeabilities. 70 Degree of Permeability High bbdium Low Very low Practically impermeable Value of k (cm/sec) Over 10"1 10-1 to 10 10‘3 10'5 Less than 10' Along with this Terzaghi and Peck included an inter- pretive table based on work by Casagrande and Fadum. 71 Venezuou saddened :xo manmuwpawsoo .mummu coqupHHOm :cou mo muasmou do women nonpouDQEoo mam>muw was modem mmoHsoawmnoo sumac ou keno manmoaaad< .sowusnauumap ouamuswmuw Scum compoundsoo x so none pmuasomu umaaeumump huwmmoums nmuasomu moamwuomxm uumuaosu muamaumaxm manmumpam mocwauoaxm £032 mapped .manmwa :sou .oaameaou magnum .manmaamus: .woumE :om .umuoEmQE .uouofimefiumd osonuwswaamm :mmsumm pmonuwcaaamm :uma psmnawsaaamm pouwadmu mucowwmaxm mauqu .uouoEmeuma pmoauucmumcoo pmuasvmu oucoauomxm :wcwwuwmwu manmuopamnou .pouodnaoo mammaoua . p ; uuowao we manmwamm .muwou wsHaEdau:soauamoa . _Hmcwwwuo mum cw HwOm mo wswumou uoouwa wcwumaummz was sownmumwm> mo mundane An nowmaoos mawom :maow>uanH: weaponuwoB .ouo .munmomop mega . woman no snow sonn F nwauwumnum Heap Hwaomaw o mango msoocmonon knee was pawn pawn mo woman wthfluxHE Ha m ..w.m .maaom :xas .muawm owsmwuosa was Ho>muw use case :msow>uomEH: oasmwuo .mUGMm scam huo> sumac .mp:Mm sumac Hm>mnw cameo mnow>ummEH haemoauomum uoom nooo owmsaoun once muoa muoa ouoH muoH euoa muoH Nuoa Huoa o H HOH Noe Aegean onV can use So cw x Auwaanmoeuma mo ucmwuwumooo maaom mo monumaumuumumno mwmcwmun was muwawnmmEumm 72 In England, Capper and Cassie (1953) also cited a relationship between particle size and permeability: Permeability Ranges value of k 102 1o 1 10'1 10'2 10'3 10‘4 10"5 10"6 10'7 cm/sec 105 104 103 102 10 1 10"1 10‘2 10'3 lO-4ft/day Gravels Sands Silts Homogeneous - Clays Fissured & Weathered Clays Good drainage Poor drainage Impervious Further, Hough (1957) also discussed permeability and related this to particle size: TURBULENT FLOW LAMINAR FLOW 73 Typical Values of Permeability Coefficients Particle Size Range Permeability Millimeters Coefficient-k Dmax Dmin ft/mo cm/sec Derrick STONE -- -- 100x105 100 One-man STONE -- -- 30x105 30 Clean, fine to 5 coarse GRAVEL 80 10 10x10 10 Fine, uniform 5 GRAVEL 8 1.5 5x10 5 Very coarse, clean, 5 uniform SAND 3 0.8 3x10 3 Uniform, coarse 5 SAND 2 0.5 0.4x10 0.4 Uniform, medium 5 SAND 0.5 0.25 0.1x10 0.1 Clean, well-graded 5 SAND & GRAVEL 10 0.05 0.01x10 0.01 Uniform, fine SAND 0.25 0.05 400 40x10"4 Well-graded, silty _4 SAND & GRAVEL 5 0.01 40 4x10 Silty SAND 2 0.005 10 10'4 Uniform 3111 0.05 0.005 5 0.5.110"4 Sandy CLAY 1.0 0.001 0.5 0.05x10'4 Silty CLAY 0.05 0.001 0.1 0.01x10-4 CLAY (30 to 50% -4 clay sizes) 0.05 0.0005 0.01 0.001x10 Colloidal CLAY _4 _ (-2/x;;50%) 0.01 108 10 10 ’ Hough also felt that at the 2.00mm limit there was a change in flow within the soil mass. flow between particles was turbulent and below 2.00mm, was laminar. Above 2.00mm size, the it 74 The Portland Cement Association Soil Primer (1962) also suggests this relation between particle size and permeability: Size class R value cm/sec sands 1.0 - 10'3 silty and chgaysand 10"3 - 10"7 cohesive clays Less than 10'8 The Asphalt Institute (1966) cites the following figures showing the relation between particle size and permeability: Approximate coefficient of bhterial permeability, feet per day 0.000001 Silty clay 0.00001 0.0001 0.001 Fine sands and silt 0.01 0.1 l Sands 10 100 1,000 Gravel 10,000 100,000 They also showed the relationship of grain size summation curves and permeabilities. III. METHODS A. ENGINEERING PROPERTIES OF SOILS AND PARTICLE SIZE LIMITS In the course of its operations, the Michigan State Highway Department tests for certain soil properties. In addition to materials below a given size from a grain size distribution curve, several other physical properties are evaluated. These tests are run in accordance with various current ASTM and AASHO standards (MSHD 1960). Included in this study are the following properties: liquid limit, plastic index, shrinkage limit, shrinkage ratio, maximum density and Optimum moisture. An attempt is made here to show the relation between these properties and certain portions of the particle size distributions. The techniques used are those of simple correlation and regression analyses. For purposes of these analyses, samples were selected which were accompanied by particle size distribution curves that were plotted during the particle size analyses. In most cases the <.002mm.value was not represented on the curve and it was necessary to extrapolate to get this value. MOre data on physical properties and particle size distributions were available in previous Reports on Results of Tests by the Michigan State Highway Department. However, these reports did not include the particle size distribution curves drawn during the particle size analyses in connection 75 76 with the other tests. Only the percentages of .005, .050, .074 and .105mm, read from the size distribution curves, are reported in the silt size range. To evaluate the accuracy of the values obtained when a particle size distribution curve was replotted from the above mentioned data, 10 samples that represented the approximate range of particle size distributions in the data were selected from among the samples for which the original size distribution curves were still available. In the table below are the results of particle size classes read from the curves redrawn from the data reported together with those read directly from the original graphs. 77 Comparison of Percentages of Particle Size Classes Read From Curves Constructed During Mechanical Analyses (MSHD), and Those Redrawn (RDC) From(.005, (.050, (.074, (.105 and Sand Sizes Reported: ‘72 M ax . ‘fé Max . Sample maximum Size Sample maximum Size No. Size, mm NSHD RDC No. Size, mm NSHD RDC 6481331 .002 1,2 6 668897 .002 7 4 .010 l6 16 .010 10 10 .020 20 19 .020 12 13 .062 25 25 .062 l6 16 6481 .002 66 74_ 6581130 .002 2 l .010 89 86. .010 4 3 .020 92 90 .020 5 4 .062 94 95 .062 9 9 648242 .002 67 52. 6783148 .002 26 30 .010 80 78 .010 42 40 .020 84 87_ .020 50 46 .062 98 97 .062 59 58 6582917 .002 23 23 6781773 .002 .28 35 .010 37 37 .010 56 55 .020 47 48 .020 76 72 .062 70 70 .062 97 97 6481334 .002 9 7 6682659 .002 43 44 .010 16 15 .010 50 51 .020 20 18 .020 53 53 .062 23 23 .062 58 58 Underlined values differ by more than 2%. It seems clear from this table that a particle size distribution curve drawn from the data usually reported in the silt and sand sizes is not reliable for extrapolating and interpolating all size fractions. Note for example the divergence in the‘<.002mm,values. However, it is good to note too that the .062mm size is easily and with good accuracy available from this data. The intermediate size percentages never differed by more than 4 percent. 78 As a result of this study, it was decided to avoid use of the data which did not include the original particle size distribution curves. This was done to avoid possible misleading results of comparisons of silt sizes with physical properties because of the inaccuracy in the lower limit, .002mm, of the silt sizes. In addition to the data from the Michigan State Highway Department. other data from the Federal Housing Agency (1961). with essentially the same properties given, were analyzed. The difference between the two sets of data was in mineralogy. Values reported from the Michigan State Highway Department were of mixed mineralogy. The 124 observations selected from the Federal Housing Agency data were high in kaolinite, a coarse platy clay mineral. For the Federal Housing Administration data, simple correlation and regression analyses were calculated with only: (.002mm, <1 % ($7014 0 <1 44:72:?“ 1 'H <1 23:10:1< seduces sun .m .wae \Y,‘ O. . flUt? LOG CC" Aatsuop £10 '7'?" (KY. -'.l C LL“? 92 1 “- 8 ‘ o 4 —< ; <<1 .4 <1 <1 C; 5 q ‘ .. 3,: Q. :::‘g 8 V 0‘} R --1 g“ ; In > o o ‘5! -w n' u -3 U .4 3 o. 'd U 8' 9. ' 8 6 <1 ‘ ‘ . a :9 ‘ .1 d h ”'1 <1 <1 as a J- o 1 1 L J J 9 1 1 C'CE' 3‘03 01:1 3 CT (701: 0313.10“ mido 93 . OOH .—1r.— .0? aamoo.uo~o. R 0.93. a a J N a 4 Eamoo.uo~o. fl .a> eunuedoanasafiugo 0.0m... 0.01. 04H“. OAHW .n 0.0.... .wE 00.0 :J'UT U'L'Jr? canistomrmnmtado 0'92 0 94 <1 <1 <1 4 <3 41 Q .5. <1“ 8 .3? 9 <1 371174 («17 21 Q] It «1!:9 ‘54 . Sigiuq g :3<- 4373‘ ‘4 b .1 «3:1 4 *4 4 ‘..¢ Q g ‘23:: o a: = I <1 '0 4 ‘ 2 4 <1. 32 . (1515‘ Q «4 3? <1 <1 1 0 so 4 e 'l q 00 .4 4 [A L _l 1 *1 L138. 'CFS'I T‘U' fistsuap mm “10.0 "70 . .0 b0 ‘1- nuwnouuu cards-eunloo m .0." .wam 1 CU? a r}.- f th 3 {£12. TY}! :Iflu' QJSUOJJI 0A1 sseadmoo g 98 The results of three of these relationships are shown graphically in Figs. 11, 12 and 13. These would enable us to predict the amount of material expected in a given size if the amount of another size and the relationship between them is known. Good examples of this that may be useful. to soil scientists. geologists and engineers are the relationships cited in Table 3 based on data from the Michigan State Highway Department. The <.OOme.vs. (.005mm percentage relationships for the Kaolinitic materials cited by the Federal Housing Administration showed a correlation coefficient of .98 compared to .97 for the Michigan State Highway Department data in Table 3. This indicates a close relation which could be used to transfer data on one clay size into termm of the other clay size limit with the regression equation. B. AVAILABLE WATER AND SILT PARTICLE SIZE LIMITS The results of the statistical calculations presented in Tables 4. 5 and 6 and Figs. 14-23 show that a relation- ship does exist between silt content and available water content. This relation is shown in nearly all cases by the relatively large. >.70, correlation coefficients, and the validity of these relationships is borne out by graphs of available water vs. silt content for various definitions of available water and the various silt size fractions found in these undisturbed, naturally occurring soil materials. 99 5580 a o a? Emoo v.» 2: w; 'GJT n"'500') 1 100 E£N©o.10mo. R 0.23 0.0m 0.01 0.0m 0.0m aauoo.uomo. fl .a> ElNoo.|~oo. fl .3 .wE {LTD 3. I ' ” a on; ;v. 3'33 3‘0: 0'0". 5: "x: .3 umZOO'-Z90' % '0: 'CIJT o 1 .050- 0 002m 1 . 1051.002m vs F180 13. 101 J L 1 3'0“? C1115 0'0"». mmZOO"SOI° 1 “DID we. .0 50.53 70.0 53.0 . ‘L .050-.002m 5C] 0 102 Table 4 -- Correlation of particle size ranges and water released between .06 and 6 atmospheres tension. Correlation Size Range Coefficient Eguation* .125-.020mm .751 Y = 4.08 + .193X .125-.010mm .712 Y = 3.70 + .178X .125-.002mm .606 Y = 3.75 + .140X .100-.020mm .745 Y 8 4.84 + .204X .100-.010mm 0699 Y = 4053 + .180X 0074-0020m 0708 Y ' 5.38 + .216X 0074-0010111!“ .654 Y a 5.03 + 0187X 0074-0002m 051.7 Y 3 5025 + .130X .050-.020mm .537 Y = 6.14 + .228X .050-.010mm .421 Y = 6.40 + .180X .050-.002mm 0245 Y = 7.24 + .078X .020-.002mm -.123 Y = 9.72 - .064X * X = % of fraction in size range quoted and Y = % or readily available water holding capacity. available 103 Table 5 -- Correlation of particle size ranges and water releasedbetween .06 and 15 atmospheres tension. Size Range 0 125-0020"]!!! 0 125-001.0111!“ O 125-0002m e 100- a 020m 0 100‘- . 010m .100-.002mm e 0 74- a 020m 0 074- 0 010mm 0 074’" a 002m 0 050- a 020mm .050-.010mm a 050- a 002m .020-.002mm Correlation Coefficient .760 .755 .711 .777 .756 .696 .737 .714 .640 .587 .507 .398 .078 r< 0 nodes.» .Euo ouoo. voesodeu have: a. :3 .35 4 OUT 3'0?- uotsue: °ms 9-9o° pests-[ea assay; 1, 3'0”. 106 alouo.uooH. fl .e> Godunov .Euo ouoo. voesoaou noue3 a .ma .wem 1110 7101' 00"- uotsuea 9 'mas 9-9o° pesaetoa Jena“ 1 107 £83644 . a .004 0.03 0.00 0.00 0.03 0.00 0.03 0.00 0.00 0 04 00.0 11.71-- 14.--- - -4 1 «1|: —. 4 7| .1 - q, - -4 - _. 4 4 4 4 4 4 4 444 A4 4 4 «24 4 4 4 4 .4\\4\4. 4444 44444 4 1\1\\4 4 4 44 4 4 .444 ...-...\...-- 444 444 4.4 4 444 4 4%2fi544.444 444 4 4 444 .44 .4 444 4 44 4 4 4 4,344A4fl4 44 44 4 4.4 . .4 44 -44 44 4 44 4 mill... 4 44 4 4 .. 2.. . 4 4 .\s.\5)l\.\kku Q 5:15.111 AN 4 AV Q a--- an... 4 a 4 4 4 4 4 4 4 4 4 4 EENoo.nm~H. fl .e> soaesou .Euo ouoo. voeeoaou nous: fl .oa .wam U'UT 0°08 uotsuea °m3s 9-90' pestetea 10:35 1 TD". 108 ‘.III I I 1 553.441 x 0.0m DAMP OZOJ 0.0m aflomo.anuu. u .0) now-sou .auu mauoo. voodoaou hand: n .54 .n«~ O m o °uau g1-90° posvatoa Joann 1 notice: 109 .004 .00 0.00 .00 0.00 0.00 0 03 0 00 0.00 0 04 00 0 0 - . r -. r } . _ J. 44. .4. g . 4 .1 4 4 4 0 T .LII 4 4 4 44 4 4 44 4 44 4444 4 4 4 4 4 4 44 44.4 44 4 4444 . 4 44440.44 .4 44444 444 1 £444.24. 434444444 4. 444 44 4 4 .--4...\ 4 44 :44 4 4 4 4. 4 44 4 4 4 4 .\ 4 44 4 \\\\X 4 4 4 lomofooa. a .0» now-sou .luc nauoo. vo-ucaou noun: a. .3 .9: L 5101‘ CD 0 corn»: nun. g1-90° 900.010.! .103.“ 1, a a 3 110 £~oo.un~.~. fl 0.05 0.0.4 0.0m 0.03 0.0m 0.0m -1 4 H Q Q 4Q 4 E q 4431 4 4 111. f4 4 d. 4 4 4 luoofnua. a. .0) «04:3» .luu 3.60. vooaoaou noun: a. Ca” 3 h O‘OT :10}:an 'Iat gt-go' pentatoa Janna 1, J'O'I ‘— 1.11 llNoo.IONOw R 0 SJ 0 or ea~oo.-o~o. w ..> no4.aou .au. n4-mn. U . or K144 4.44 flamafl 4 444444444 4 44 4 4 4 4 4 4 44 4 4 Q 4 .ou .44m woodmaou noun: a Ill 3’ OT °m39 g1-gg° posaataa 1939M % "CF 0‘0” uotsuan 112 .004 C.D.U Oiflmm 0.0m- ri+|l.l!-J—?l. if- 4i ! .145} El . a! . Till- -1415 i 1---?! Efimwo.uomo. R 03 0 CT. 0.01 U41“, OZCH 0.04 4 4434 44444 4 g§444g4 m4 44. 4 4:} ’3 l6~oo.uomo. u .n> now-nou .luu manna. vomuoaou noun: R .HN .wam 3'01? 0 OT °m39 gI-gg' pasratea JSJPM {TUE notsuaa LY" % 113 .sa~oa.-n~4. a .00.". O.Q_U 0.0m 0.0? 0.03 0.0m 0403 CAM 0.8 EENoo.nm~A. a .n) Ionnlou .Eun manna. vouquou noun: fl .NN .uqm ll 0‘ OT 3'0‘. 0mg SI’EE. 9‘3'3193 3.3?“ 1 anyone: 114 4. g l ‘i 3 <3 4 Q C ' I o E . O '4 . 0 fl (5 <13 0 O > 8 O O «I . r\ I! 3 o! 0 C3" 5 3°. " 8' n 7 9" O O n n m" 0 '6 2 3 Q ‘ . 9 0 .4 o H O u . O 3 B ‘ 3 ,. Q . R M N , O. :9 9 (z. o 0. 1 1 i O can on: 0'02 0'0? 03° uoxsuoa 'Iau g1-cg° 900.010.: 10mm 1 115 While silt is an important factor in these relation- ships, it is not the only factor contributing to water retention. Other factors such as aggregation, compaction and distribution of particle sizes in the silt fraction as well as other factors. such as the kind of plant. should be considered to evaluate the total available water holding capacity. The correlation is greatly affected by the fraction of the particle size distribution which is used. For example. correlations are consistently lower. regardless of the tension values used to measure available water, when .OSOmm is used as the upper limit of silt as compared to limits up to .lOOmm or .lZSmm. The correlation improves as the lower silt size limit used becomes coarser up to .020mm. except when l/3 to 15 atmospheres is used as the measure of available water. In that case the .010-.100mm fraction gives the best correlation. Correspondingly. the coarser silt fractions are better correlated with available water, particularly when the lower tension, .06 atmospheres, is used as a measure of field capacity. It should be noted that as a group, the correlations with l/3 to 15 atmospheres tension as the measure of available water are least sensitive to differences in silt size ranges and give better correla- tions than the other tension limits with the less than .050mm silt fractions. These are not. however. the highest correlations found, nor are they most meaningful expressions of available water. 116 From these results it seems clear that .OSOmm does not represent the most significant upper size limit of silt in interpretations involving available water storage capacity. Regardless of the tensions used in evaluating available water. in all cases with .002mm as a lower silt size limit, higher correlations were found with size limits coarser than .050mm. Each definition of available water has a maximum correlation with a different silt size based on these data. For .06 to 6 atmospheres it is with .125-.020mm silt. For .06-to 15 atmospheres tension it is with the .lOO-.020mm silt fraction, and for .33 to 15 atmospheres it is with the .lOO-.OlOmm silt fraction. For the lower size limit of the silt fraction. correla- tion with measures of available water show the .020mm limit has the highest correlation coefficients with the .06 to 6 atmospheres and .06 to 15 atmospheres available moisture evaluations. These values seem to show a clearer pattern of relationship to particle size than does .33 to 15 atmospheres available water. In every case, for the former available water values, the highest correlation was obtained using .020mm as the lower limit of the silt size range. This is consistent with the idea that these coarse silt sizes are more significant in determining water availability at lower tensions. This specific result shows the significance of the .020mm size limit used by the International Society of Soil Science. It is thus significant for more than just completing a logarithmic ’ 117 scale based on 2, or as a limit of vision with the naked eye. Thus it is clear that an upper size limit greater than .050mm up to .125 or .lOOmm and a lower size limit of silt up to .010 or .020mm will give better correlations with available water holding capacities. If .002mm is accepted as the lower silt size limit. the best correlation between available water and upper silt size is at .125mm when .06 atmospheres is used as a measure of field capacity, and at .lOOmm when 1/3 atmospheres is used as a measure of field capacity. An upper silt size limit of .lOOmm will be near the optimum for any of the lower silt size limits or tensions tested. As suggested by Franzmeier et al. (1960) this seems to be a considerable improvement over the present upper silt size limit for soil scientists. C. PARTICLE SIZE OF SILT FRACTION AND EFFECT ON HYDRAULIC CONDUCTIVITY AND CAPILLARY RISE l. Hydraulic Conductivity Using the .002mm as the lower limit of silt. it appears that in terms of hydraulic conductivity. there is little difference between .050mm and .lOSmm as an upper size limit. as shown in Table 7. 118 Table 7 -- Hydraulic conductivity of various size fractions. lst 2nd 3rd Reading Reading Reading Average Size Limits cmlhr cmlhr cmlh; cmZhr inlhr .0023.050mm .23 .16 .16 .18 .07 .002-.050mm .16 .12 .12 .13 .05 .002-.105mm .16 .14 .09 .13 .05 .002-.105mm .16 .16 .15 .16 .06 .010-.050mm 1.16 1.14 1.16 1.15 .45 .010-.050mm 1.18 1.18 1.18 1.18 .46 .010-.105mm 1.57 1.59 1.61 1.59 .62 .010-.105mm 1.61 1.59 1.63 1.61 .63 .020-.062mm 2.35 2.29 2.21 2.28 .90 .020-.062mm 2.58 2.39 2.35 2.44 .96 .020-.105mm 4.38 4.38 4.37 4.38 1.76 .020-.105mm 4.64 4.64 4.66 4.65 l.83 The results were obtained on structureless prepared samples of the individual size fractions. There may be a difference between these two limits when considered in the total size distribution with sand and clay of less silty materials or in undisturbed samples where structural development may be present. Certainly there is a marked difference between the fractiomwhen coarser lower limits were used. There was a marked difference in the turbidity in the water that percolated through the samples with .002mm as the lower limit compared to samples with the coarser lower 119 size limits. The water from the .002mm samples was con- sistently muddy, while the others were clean. Also. when removing the material with .002mm as the lower size limit from the metal cylinders, the samples flowed in the saturated condition. The samples with coarser lower size limits had to be dug out. These may be significant in terms of sedimentation and the bearing or plastic properties of similar materials. On the soil samples studied by Erickson et al. earlier. the hydraulic conductivity in terms of in/hr was also measured on undisturbed and unfractionated samples. Simple correlations with material less than a given size showed these results: Table 8 -- Correlation of hydraulic conductivity with various fractions. Correlation Ftaction ngftitiegt (.002mm -.346 (.OlOmm -.380 (.020mm -.406 <.050mm -.486 <.074mm -.514 (.lOOmm -.522 (.125mm -.521 The correlations are not high but a definite trend is established in these values. Certainly particle size is 120 not the only factor that determines hydraulic conductivity but these correlations show that some values are more significant than others. 2. Capillary Rise Results of capillary rise over a 48-hour period are shown in the following table. 121 .Flluulflii. A 1 .1 1...”!!1nurhsurfiin- Aconcausoov 0.05 0.00 5.00 4.00 0.00 0.00 0.04 0.00 0.00 0.40 0.04 0* EE00.....000. 0.00 4.40 5.00 0.00 0.40 0.00 0.44 4.00 0.00 0.40 0.50 H4 EE00H.u000. 4.05 0.40 0.00 0.50 0.00 4.04 0.04 0.00 0.00 0.00 H.0H 04 EE000.1000. 0.00 5.00 0.50 0.00 4.00 0.44 0.00 0.00 0.00 0.00 0.00 0* 85000.u000. 0.00 0.H0 0.04 0.04 0.04 0.00 0.00 0.00 4.40 0.0a 0.0g 0* EE0OH.I000. 0.50 0.00 0.00 0.04 5.04 0.00 0.00 0.50 0.40 0.00 5.00 H§.EE000.0000. 4.00 0.00 5.04 0.04 0.04 0.00 0.00 0.50 4.00 0.0a 0.00 0* EEO00.t040. 0.50 0.00 0.04 0.54 0.04 0.04 5.00 4.00 0.40 0.04 4.00 0% 85000.3000. 0.00 0.00 0.00 0.04 0.50 0.04 0.00 0.00 0.0 0.5 0.0 04 EE00~.-000. 4.00 0.00 0.00 0.50 0.00 0.00 0.00 0.0 0.5 0.0 5.0 0* 86004.1000. 0.00 0.00 0.00 0.0a 0.00 0.40 0.00 0.0 0.5 0.0 4.0 0* EE000.:000. 0.00 0.40 0.00 4.00 0.00 0.00 0.00 0.00 0.0 0.0 0.0 H4 EE000.:000. PE PE PE mus PE man an; us 35 4.44... a g m 4 m.m m m.~ N m.4 4 me on ma .EU :4 mcoauomum onwm macaum> you came omen kumaaammo nu 0 manna 122 0.000 0.000 0.000 4.00 0.500 4.004 0.000 0.004 4.400 0.004 0.000 0.000 0.500 4.000 oun00mm 0.05 0.00 mludwlum 04 00 h ul.‘l a . 0.400 0.00 0.000 0.00 0.000 4.000 0.400 0.00 0.00 3.. 4. 11!: it. m.wo o.mo m.~o «.4o 4.0m ~.ww o.noa ~.m00 «.mo mu500oe «.mo o.wm o.~m 4.000 0.40 m.hw m.oo m.oo 4.4m ~.om m.mo 0.44 m.em 4.w4 w.m4 4.4m 4.04 ~.me mus00um m.m4 mus00mm o.m4 ggg 4N ON 40 0.00 4.00 0.00 0.00 0.05 0.45 4.05 0.05 0.50 0.50 0.00 0.44 was 00 0.40 0.00 0.00 0.50 5.45 0.00 0.05 0.45 0.40 0.00 0.00 4.00 was 00 0.05 0.05 4.05 0.00 0.00 4.00 0.00 0.00 0.00 0.00 0.00 0.40 an: 0 0.05 4.00 0.05 0.00 0.00 0.00 0.00 0.00 0.00 0.00 4.00 4.00 444 o 04 04 04 04 04 04 04 04 04 04 04 04 EE004.:000. EE004.:000. EE000.-000. EE000.-000. EE000.:000. EE000.-000. EE000....000. EE000.u000. EE004.:000. EE000.:000. EEO00.-000. 35000.u000. g Avosc4ucoov 0 00049 123 Perhaps the most significant relationship found from this data is that the lower size limit of these particular fractions is more important in predicting capillary rise rate than is the upper limit. In the initial 36 hours, the .002mm fractiors had the least rise with the .010mm fractiors higher and the .OZOnun fractiors the highest. For the 36 hour reading the .010mm fractions equaled the .020mm and in the 48 hour WT! éflwm reading the .010mm fractions had the highest rise. For the .002mm fractions an upper limit of .050mm had consistently higher capillary rise than did those fractions . ‘-ml "‘1 7" ;~_ 1"- ' with .lOSmm as the upper limit. The other fractions showed mixed values for capillary rise between the upper limits. The differences between results with the same fraction may be due in part to insufficient packing in the tubes. Certainly the packing procedure needs to be more effective. Perhaps complete saturation of the soil in the tubes and drying before testing capillary rise would solve this problem. V. CONCLUSIONS The results point to the need to determine certain size limits in particle size analysis because of their relation to physical properties. ‘ Separations at 2.0mm and .002mm are generally regarded F as significant as cited previously. Results in this work note a need for a separation at “1):! ‘K '_ I g'ei _’ '. .020mm to provide a better measure of readily available water. And the amount of this material may affect capillary rise and hydraulic conductivity as well. A separation at .105mm or .125mm was shown to be important for evaluating readily available water. It appears, though more work is needed, that a separa- tion is necessary between .020mm and .105mm or .125mm because of shear stress and 8 compressive strength. Some value, either .062mm or .074mm. may be significant for these properties. 124 LITERATURE CITED Agricultural Education Association. 1926. The mechanical analysis of soils: a report on the present position and recommendations for a new method. Journal of Agticultural Scietce, 16:123-144. Agricultural Education Association. 1928. The revised official method for mechanical analysis. Journal of Agricultural Science. 18:734-737. Alling. H.L. 1943. A metric grade scale for sedimentary rocks. Journal of Geology, 51:259-269. American Geophysical Union. 1947. Report of the subcommittee on sediment terminology. Tra section ican Ge h sica Ufllgn. 283936-938. American Society of Civil Engineers. 1947. Report of Committee VII on foundation and soil mechanics. ASCE, Civil Engineering Division, Civil Engineering Bulletin 12. Asphalt Institute. The. 1963. §gils Man 1. College Park, Maryland: The Asphalt Institute. Asphalt Institute, The. 1966. Qtainage of Agphalt Pavement fittucturtg. College Park, Maryland: The Asphalt Institute. Atterberg, Albert. 1908. Studien auf dem Gebiete der Bodenkunde. Translated by R. W. Simeon. Landswitt- tthaftliche versuch-Stgtignen, 69:93-143. Bartelli, L.J. and Peters, D.B. 1959. Integrating soil moisture characteristics with classification units of some Illinois soils. ProceedingsI Soil Science Society 2f America, 23: 145-151. Baver. L.D. 1938. Soil permeability in relation to non- capillary porosity. Proceedings. §gil §cience Soctety f Americ , 3:52-56. Baver, L.D. 1956. Soil Physics. New York, N.Y.: J.W. Wiley and Sons Inc. Bender, William H. 1961. Soils suitable for septic-tank filter fields. U.S. Department of Agriculture. Soil Conservation Service, Agricultural Information Bulletin No. 243. Washington, D.C.: Government Printing Office. 125 .1? I'llflflllI-III 126 Black, C.A.: Evans, D.D.: White, J.L.: Ensminger, L.E.: and Clark, F.E. 1965. Methods 9f Soil Anatysis, Monograph No, 9. Madison, Wisconsin: American Society of Agronomy. Bouyoucos, G.J. 1927. The hydrometer as a new method for the mechanical analysis of soils. Soil Science, 23:343- 353. Bouyoucos, G.J. 1927. The hydrometer method as a new and rapid method for determining the colloidal content of soils. Soil Science, 32:319-330. Bouyoucos, G.J. 1937. A sensitive hydrometer for deter- mining small amounts of clay or colloids in soils. S '1 Science, 44:245-246. Boyd, J.R. 1922. Physical properties of subgrade materials. Canadian Engineer, 43:362-364. Briggs, L.J.: Nbrtin, F.O.: and Pearce, J.R. 1904. The centrifugal method of mechanical soil analysis. U.S.D.A. Bureau of Soils Bulletin 24. '15‘“ - Capper, P. Leonard: and Cassie, W. Fisher. 1953. The Mechanics of Engineering Soils. London, England: MCGraw-Hill Book Company, Inc. Casagrande, A. 1942. Classification and identification of soils. Ptgceedtggs, ASCE, 73:783-810. Cushman, A.S. and Hubbard, P. 1907. Air elutriation of fine powders. Jgurnal of the American Chemical Society, 29:589-596. Day, Paul R. 1950. Physical basis of particle size analysis by the hydrometer method. Sgit Sctence, 70:363-374. Diller. Joseph Silas. 1898. Educational series of rock specimens. U.S. Geological Survey Bulletin No. 150. Washington, D.C.: Government Printing Office. Dougrameji, Jamal Sharif. 1965. Soil water relationships in stratified sands. Unpublished Ph.D. thesis, Michigan State University. Felitsiant, I.N. 1959. Capillary movement of moisture in stratified soils. So iet 8 11 Science, 3:282-292. Felitsiant, I.N. 1961. Capillary movement and accumulation of moisture in stratified soils. Soviet Soil Science, 10:1099-1107. 127 Fireman, M. 1944. Permeability measurements on disturbed soil samples. Soii Science, 58:337-353. Franzmeier, D.P.: Whiteside, E.P.: and Erickson, A.E. 1960. Relationship of texture classes of fine earth to readily available water. h n e na 1 nal C n ress Soil §giente (Madison, Wisconsins, 1:354-363. Gill, William R. and VandenBerg, Glen E. 1967. Soil dynamics in tillage and fraction. Agricultural Handbook No. 316. Agricultural Research Service. U.S. Department of Agriculture, Washington, D.C.: Government Printing Office. Glossop, Rudolph: and Skempton, Alex Westley. 1945. Particle- size in silts and sands. Journal of the Institution of Civii Engineers, 2:81-105. Goldbeck, A.T. 1921. Tests for subgrade soils. Public Roads , 43 15-20. L Grabau, A.W. 1913. Principles of Stratigraphy. New York, N.Y.: A.G. Seiler and Company. Hack, J.T. 1957. Studies of longitudinal stream profiles in Virginia and Maryland. U.S. Geological Survey Profes- sional Paper 294B. Washington, D.C.: Government Printing Office. Hall, A.D. 1904. The mechanical analysis of soils and the composition of the fractions resulting therefrom. qurpal pf ghemicai Spteity Irapsactipnt, 85:950-963. Hall, A.D. and Russell, E.J. 1911-1912. Soil surveys and soil analysis. Jpptnai pf Agricuitural Science, 4:182-223. Hammitt, G.M. 11. 1966. Statistical analysis of data from a comparative laboratory test program sponsored by A.C.1.L. Miscellaneous Paper No. 4-785. U.S. Army Engineer Water- ways Experiment Station. Hilgard, E.W. 1873. On the silt analysis of soils and clays. The Apprican Jpptnal pf Science and Atts, 6:288-296. Hilgard, E.W. 1890. Soil investigation: its methods and results. University of California College of Agriculture, Agricultural Experiment Station Annual Report, 158-159. Hilgard, E.W. 1903. Methods of physical and chemical soil analysis. University of California Agricultural Experi- ment Station Circular No. 6. 128 Hill, David E. 1959. The storage of moisture in Connecti- cut soils. Connecticut Agricultural Experiment Station Bulletin 627. Hogan, Robert M.: Haise, Howard R.: and Edminster, Talcott W. ed. 1967. irrigatipn of Agricuiturai Lands. Madison. Wisconsin: American Soceity of Agronomy. Hogentogler, C.W. 1937. Engineering Propprties of Spi . New York, N.Y.: MCGraw-Hill Book Company Inc. Hopkins, C.G. 1899. A plea for a scientific basis for the division of soil particles in mechanical analysis. U.S.D.A. Division of Chemistry Bulletin 56:64-66. Hough, B.K. 1957. Basis Soil Epgipeering. New York, N.Y. Ronald Press Company. Jacks, G.V. 1965. The role of organisms in the early stages of soil formation. Experimental Pedology. London, England: Butterworths. Jamison, V.C. 1953. Changes in air-water relationships due to structural improvements of soils. Spii Sciencp, 763143-151. Jamison, V.C. 1956. Pertinent factors governing the avail- ability of soil moisture to plants. Soil Science, 81:459-471. Jamison, V.C. and Kroth. E.M. 1958. Available moisture storage capacity in relation to textural composition and organic matter content of several Missouri soils. Prpteedings, Spil Scienct Spciety pf Ametica, 22:189-192. Jennings, D.S.: Thomas, M.D.: and Gardner, W. 1922. A new method of mechanical analysis of soils. Spii §cipnce, 143 485-499. Keen, B.A. 1928. Some comments on the hydrometer method for studying soil. §pii Stience, 26:261-263. Keen, Bernard A. 1931. The Physicai Properties of thp Spii. London, England: Longmens, Green and Co. King, F.H. 1894. Physitt pt Agticplture. Madison, Wisconsin: Democratic Printing Company. Krumbein, W.C. 1934. Size frequency distribution of sedi- ments. Jpptna]. pf §edi£ntpty Ppttplpgy, 4:65-77. Krumbein, W.C. 1936. The application of logarithmic moments to size frequency distribution of sediments. Journal pf W. 5.35-47. 129 Krumbein, W.C. and Pettijohn, F.J. 1938. Manuai of Sedi- ‘gppt§£y_§gttpgt§ppy. New York, N.Y.: Appleton- Century-Crofts, Inc. Lehane, J.J. and Staple, W.J. 1953. Water retention and availability in soils related to drought resistance. Canadian Jpptnal pf Agricultural Science, 33:265-273. Leliavsky, Serge. 1966. An Introduction to Fluvial Hygtaulits. New York, N.Y.: Dover Publications, Inc. Leopold, Luna B.: Wolman, M. Gordon: and Miller, John P. 1964. Fipyiai Ptocespes in Gepmptphplpgy. San Francisco, California. W.H. Freeman and Company. Loughbridge, R.H. 1894. Investigations in soil physics: the capillary rise of water in soils. California Agri- cultural Experiment Station Annual Report, 1892-1894, 91-1000 Lund, Z.F. 1959. Available water-holding capacity of alluvial soils in Louisiana. PtpceedingsI Spii §cignte fipciety pt Ameticp, 23:1-3. Luthin, James N. ed. 1957. Drainage of Agricultural Lapds. Madison, Wisconsin: American Society of Agronomy. Michigan State Highway Department. 1960. Field Manuai pf So i e r . Lansing, Michigan: State of Michigan. Miller, C.E.: Turk, L.M.: and Path, H.D. 1965. Fundamentais p: Spii Sciepte. New York, N.Y.: John Wiley and Sons Inc. Milner, Henry B. 1952. Sedippntaty Pettography. London, England: Thomas Murby and Company. Olmstead, L.B.: Alexander, L.T.: and Lakin, H.W. 1931. The determination of clay and colloid in soils by means of a specific gravity balance. American Soil Survey Association Bulletin 12: 161-166. Osborne, T.B. 1887. The methods of mechanical soil analysis. Connecticut Agricultural Esperiment Station Annual Report: 141-158 a Peele, T.C.: Beale, O.W.: and Lesene, F.F. 1948. Irrigation requirements of South Carolina soils. Agriculturai Egg - ‘ppgtipg, 29:157-159. Portland Cement Association. 1962. P.C.A. Soil Primer. Chicago, Illinois: Portland Cement Association. 130 Ritter, Leo J. and Paquette, Radnor J. 1960. Highway Ehgihggtihg. New York, N.Y.: Ronald Press Company. Robinson, C.W. 1922. A new method for the mechanical analysis of soils and other dispersions. qurnal of hgricultptal Science, 12:306-321. Roderick, Gilbert L. 1963. A history of particle size limits. Review submitted to Soil Science Society of America Committee on Particle Size Distribution in Soils. Rubey, W.W. 1930. Lithologic studies of fine-grained upper cretaceous sedimentary rocks of the Black Hills region. U.S. Geological Survey Professional Paper 165A. Salter, P.J. and Willaism, J.B. 1965. The influence of texture on the moisture characteristics of soils. 1. A critical comparison of techniques for determining the available-water capacity and moisture characteristic curve of a soil. Ihp Jphthai of Soil Sciehce, 16:1-15. Salter, P.J. and Williams, J.B. 1965. The influence of texture on the moisture characteristics of soils. I1. Available-water capacity and moisture release character- istics. The Journal of Spii Science, 16:310-317. Salter, P.J.! Berry. Ga, and Williams. J.B. 1966s The influence of texture on the moisture characteristics of soils. III. Quantitative relationships between particle size, composition, and available-water capacity. The Joptnal pf Soil Sciencg, 17:9398. Scales, F.M.: and Marsh, E.W. 1922. The tyndellometer reading of soil dispersoids. Journal pf Industriai Engingtring and Chemistry, 14: 2- 4. Shaw, T.M. and Alexander, L.T. 1936. A note on mechanical analysis and soils texture. gtocpedingsI Soii Science W. 14303-304. Soil Science Society of America. 1967. Considerations relative to a common particle size scale of earthy materials. Committee Report. PtoceedingsI Soil Sgience §pcipty of Apgticp, 31:579-584. Stemm, A.J. and Svedberg, T. 1925. The use of scattered light in the determination of the distribution of size of particles in emulsions. Journal of the American Chemical Society, 47:1582-1596. Terzaghi, Karl and Peck, Ralph B. 1948. Soil Mechanics in Engihggting Practice. New York, N.Y.: John Wiley and Sons Inc. 131 Troug, E.: Taylor, J.R.: Simonson, R.W.: and weeks, M.E. 1936. Mechanical and mineralogical subdivisions of the clay separate of soils. P ceedin 3 Sci Sc ence Spciety of hpprica, 1:175-1 9. Troug, E.: Taylor, J.R.: Simonson, R.W.: and Weeks, M.E. 1936. Procedure for special type of mechanical and mineralogical soil analysis. Prpceedings, Soii Science Spciety of hmetica, 1:101-112. Tyler Sieve Company. 1967. Testing Sieves and Their Uses Hahdhpok 53. Cleveland, Ohio: W.S. Tyler Company. Udden, J.A. 1914. Mechanical composition of clastic sedi- ments. Bulletin of the Geological Society of America, 253 655-744. U.S. Federal Housing Administration. 1961. Engineeting Soii Cipssifitation f9; Rgsidential ngelopments. Washington, D.C.: Government Printing Office. U.Ss Public Health Service, Department of Health, Education and Welfare. 1967. Manupl pf Septic Tank Ptactice. Washington, D.C.: Government Printing Office. U.S. Soil Survey Staff. 1951. Soii Survgy Manuai. Agri- cultural Research.Administration, U.S.D.A., Washington, D.C.: Government Printing Office. Wang, Chwan-Chan. 1967. Comparisons of texture classifi- cation of earthy materials used by engineers and soil scientists in the United States. Unpublished M58. thesis, Michigan State University. Wentworth, C.K. 1922. A scale of grade and class terms for clastic sediments. Jpptnal pf Ggology, 30:377-392. Wentworth, C.K. 1933. Fundamental limits to the size of clastic grains. Sgignce, 77:633-634. Whitney, M. 1892. Some physical properties of soils in their relation to moisture and crop distribution. U.S.D.A. Weather Bureau Bulletin 4. Wilcox, J.C. and Spilsbury, R.H. 1940. Soil moisture studies. 11. Some relationships between moisture measurements and mechanical analysis. Stientific A riculture, 21:459-472. Wilcox, J.C. and Spilsbury, R.H. 1940. Soil moisture studies. 111. An application of soil moisture measure- ment to soil classification. Scientific Agricultpte, 21:473-478.