Stabilization of a Sanitary Landfill to Support a Highway

Transcription

1 46 Transportation Researh Reord 754 REGRESSION. The proposed model was found to be statistially valid and aurate. Equation 1 allows one to predit or ontrol the given response in terms of important ompation variables-namely, water ontent end ompation pressure. ACKNOWLEDGMENT We are grateful to the Joint Highway Researh Projet of Purdue University, the Indiana State Highway Commission, and the Federal Highway Administration for their finanial support of the researh reported here. We also appreiate the assistane of our Purdue olleague, A.G. Altshaeffl, REFERENCES 1. M.F. Essigmann. An Examination of the Variability Resulting from Soil Compation. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept , Aug J.C. Sott. Examination of the Variability of the Soaked Strength of a Laboratory-Compated Clay. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept. 77-8, May. 3. J.T. Prie. Soil Compation Speifiation Proedure for Desired Field Strength. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept. 78-7, May Y. Yoshimi. One-Dimensional Consolidation of Partially Saturated Soil. Northwestern Univ., Evanston, IL, Ph.D. thesis, Aug. 1958, 149 pp. 5. L. Barden. Consolidation of Clays Compated Dry and Wet of Optimum Water Content. Geotehnique, Vol. 24, No. 4, 1974, pp R.J. Hodek. Mehanism for the Compation end Response of Kaolinite. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept , De I. Garie-Bengohea. The Relation Between Permeability and Pore Sie Distribution or Compated Clayey Silts. Purdue Univ., West Lafayette, IN, Joint, Highway Researh Projet Rept. 78-4, May A. DiBernardo. The Effet of Laboratory Compation on the Compressibility of a Compated Highly Plasti Clay. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept. 79-3, May G. A. Leonards and P. Girault. A Study of the One-Dimensional Consolidation Test. Pro., 5th International Conferene on Soil Mehanis arid Foundation Engineering, Paris, Vol. 1, 1961, pp o. R.J. Hodek and C.W. Lovell. A New Look at Compation of Landfills. Presented at Assn. of Engineering Geologists Conferene, Hershey, PA, Ot. 18, M.A. Reed. Frost Heaving Rate of Silty Soils as a Funtion of Pore Sie Distribution. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept , Aug R.A. Abeyesekera. Stress Deformation and Strength Charateristis of a Compated Shale. Purdue Univ., West Lafayette, IN, Joint Highway Researh Projet Rept , May J.E. Jennings and J.B. Burland. Limitations of the Use of Effetive Stress in Pertly Saturated Soils. Geotehnique, Vol. 12, No. 2, 1962, pp L. Barden end G.R. Sides. Engineering Behavior and Struture of Compated Clay. Journal of the Soil Mehanis and Foundations Division, ASCE, Vol. 96, No. SM4, 1970, pp N.H. Nie and others. Statistial Pakage for the Soial Sienes, 2nd ed. MGraw-Hill, New York, 1975, 675 PP 16. R. J. Hodek end C. W. Lovell. Soil Aggregates and Their Influene on Soil Compation and Swelling. TRB, Transportation Researh Reord 733, 1979, pp Publiation of this paper sponsored by Committee on Earthwork Constrution. Stabiliation of a Sanitary Landfill to Support a Highway RONALD E. SHEURS AND RAJ P. KHERA The results of the stabiliation of a sanitary landfill by use of surharges are reported. The New Jersey Department of Transportation has undertaken the onstrution of two experimental roadways-1 85 eastbound and 1-85 westbound-diretly over landfills that ontain partially deomposed garbage. Field measurements of settlement and pore-water pressure are presented. The data indiate that the settlement response of a landfill area is similar to that of finegrained soils. Stress history plays an important role in this response. When the ratio between the stress inrease i:aused by surharge load and the existing stress was <:1, the measured strain.-was only 5-7 perent. When this ratio was 1.4 or greater, the strain varied between 11 and 17 perent. The ompression ratio for the sanitary landfill was found to range between 0.16 and Pieometri heads were found to be errati and frequently muh higher than the projeted values. This was attributed to the expulsion of methane gas from the inner pieometri tube. Relatively little is known about the behavior of a sanitary landfill subjeted to loadings suh es buildings or embankments. Extensive foundation problems an result unless the landfill has been stabilied before strutures are plaed on it. Studies have been reported on a range of stabiliation tehniques, from hemial injetions of grout and fly ash to appliations of surharges (!). Most of the work done to date has been done only under laboratory onditions or small, ontrolled field onditions. The New Jersey Department of Transportation (DOT) has undertaken the onstrution of two roadways diretly over landfills that onsist of partially deomposed garbage. The projet is loated on the north side of an ative sanitary landfill. The experimental roadways ere I-85 eastbound and I-85 westbound. These roadways have a total length of 975 m (3200 ft) end are part of the I-280 onstrution projet in Kearny, New Jersey. Stabiliation is being attained by the use of 1.8-m (6-ft) surharges over a minimum period of 24 months. Settlement plates and pieometers are being used extensively to monitor the progress of the work. This is an experimental approah to highway onstrution that the New Jersey DOT has not previously attempted. In the past, it hes used overloads with and without the aid of vertial send drains to aelerate the stabiliation of natural deposits of lays and marshlands but not of landfill areas. If adequate stabiliation of the

2 Transportation Researh Reord Figure 1. Soil profile for 1-85 eastbound. ORIGINAL GROUND ELEVATION PROPOSED FINAL PAVEMENT PROFILE ~ s r ~;::;;:============"=f.--=;::=;::=;::===;:::: i LI.I LI.I ::IE -- - LIMIT OF EXCAVATION 3 t------;j~-:::::;;~::::::::r~~~:,,,,.;;;;;;:: :;;"""'~~~~~~,..:::;;:;;;;=::=-i 0 > LI.I -' LI.I f<'o _..... _ Note: 1 m = 3.3 ft STATION Figure 2. Soil profile for 1-85 I westbound... E ORIGINAL GROUND :I ELEVATION ~ 3 2 CXISTIN(J > -' LI.I ,, STATION Note: 1 m = 3.3 ft. sanitary landfill an be ahieved in the present projet, then a similar or a modified approah an be used on future highway projets. In April, atual onstrution and exavation work began at the easterly end of 1-85 eastbound, and a general leveling proess to a maximum ut of about 9 m (30 ft) began at its westerly end (see Figures l and 2). Beause of original ground onditions, 1-85 westbound required the addition of sanitary landfill. The material used for this fill was obtained from the exavation of 1-85 eastbound. All material plaed along the westbound roadway was ompated by using from three to five passes of a 312-kN (35-ton) sheepsfoot ompator. Work progressed at an average exavation of 1913 m '/day (2500 yd' /day). This rate dereased onsiderably following days of rain beause of slippery ground onditions and large differential settlements. A total of m' ( yd') of sanitary landfill was exavated along the experimental roadway. To add to the overall diffiulty in exavation operations, approximately m ( yd') of oil-saturated material was enountered and had to be hauled to speially prepared and lined disposal sites. SOIL CONDITIONS Soil borings and laboratory testing of soils were done between 1966 and The ground elevations and the elevations of the groundwater table reorded on the boring logs and those determined at the time of onstrution were not in agreement. Most of the landfill material enountered was between 5 and 15 years old. The age of the material was determined from onversations with loal authorities responsible for the landfill and the personnel working at the site. This was onfirmed from visual examination of the state of the landfill during the exavation of suh materials: Newspapers were still readable, ans were not ompletely rusted, and various types of household refuse were still distinguishable. The materials in the landfill inluded normal household refuse, truk bodies, paints, plastis, and hemials in various states of deomposition, with thin layers of over material. The underlying material onsisted of an average of 1.25 m (4 ft) of gray-blak layey organi silt. Below the organi silt was about 0.9 m (3 ft) of brown peat and fibrous vegetation, followed by m (20-25 ft) of dense, oarse to fine sand and gravel with traes of silt. The soil profile and the amount of garbage removed and/or replaed along

3 48 Transportation Researh Reord 754 the experimental roadway are shown in Figures 1 and 2. The soil profile within the experimental setions is basially onsistent throughout the projet areas, the major differene being that the landfill operations had extended only about one-third the length of the entire projet. Laboratory onsolidation tests were performed on undisturbed samples of various underlying natural deposits. Sine the number of onsolidation tests was limited, based on the nature of the materials present below eah of the settlement plates, appropriate test results were seleted to best represent the properties of the undisturbed underlying materials, suh as organi silt and peat strata. No laboratory onsolidation tests were done on landfill materials. PROPERTIES OF LANDFILL MATERIALS The density of a typial landfill an vary greatly. The density of the material, as delivered to a sanitary landfill, ranges from 120 to 419 kg/m' ( lb/ft'), and the water ontent ranges from 10 to 35 perent (~). After the garbage is deposited, it is spread by means of a bulldoer in layers as thik as 3 m (10 ft), whih are ompated to different degrees and overed with soil, as required by various loal authorities and ordinanes. In the past, virtually no ontrol has been exerised in landfill operations, but suh operations are being managed better on more reent and better-designed landfills. The total unit weight of the present landfill materials, whih have been in plae for several years, was taken as 1129 kg/m' (70 lb/ft'). Under buoyant onditions, the unit weight was assumed to be 484 kg/m' (30 lb/ft'). Where the existing landfill was exavated and then ompated (after a ertain amount of sorting, whih onsisted of removal of large artiles suh as refrigerators, truk bodies, and washing mahines), its unit weight was taken as 1450 kg/m' (90 lb/ft'). POTENTIAL PROBLEMS IN LANDFILL AREAS The major areas of onern from a geotehnial viewpoint were the large total and differential settlements, both short- and long-term. The short-term settlements our during onstrution as a result of the operation of the onstrution equipment. Suh settlements make the work with onstrution equipment very diffiult, time-onsuming, and expensive. The long-term settlements are the result of the weight of the material itself and of applied loads as well as the deomposition of the landfill materials (hemial and biologial). Other problems arise from the potential of spontaneous ombustion and possible ill effets on workers attributable to hemial ations and the generation of gases. SETTLEMENT PLATFORMS AND PIEZOMETERS All settlement platforms were plaed three in a group, and eah group was loated 61 m (200 ft) on enter. At eah loation, all platforms were plaed at the same elevation, one unit at eah shoulder and one along the enterline of the proposed finished pavement. The results reported were for those settlement platforms along the enterline of the roadways. Before the plaement of the settlement platforms along 1-85 eastbound, a 15- to 30-m (6- to 12-in) layer of ohesionless material was plaed to provide a suitable level base for the settlement platform. The settlement platforms onsisted of a 0.9-m by 0.9-m by 12.7-mm (3-ft by 3-ft by 0.5-in) steel plate attahed to a 1.22-m by 12.7-mm (4-ft by 0.5-in) standpipe. All plates were set level, and standpipes were set plumb. Base elevations of settlement plates are shown in Figures 3 and 4. After the settlement platforms had been plaed, heavy liquid pieometers were installed so that the pieometer tips were in the underlying natural soil deposits. The pieometers were intended to serve as a ontrol on the rate of plaement of surharge, whih was designed to be plaed at a weekly rate of 1.22 m (4 ft). The monitoring of the pieometers indiated that the pieometri heads did not onform to the expeted values but were highly errati, as Figure 5 shows. This response was attributed to the expulsion of methane gas from the inner pieometer tube. Thus, the effetiveness of the pieometers beame highly questionable and their readings unreliable. So far, however, Figure 3. Settlements for eastbound roadway...,,..,.. 0 a: - ~:II..J - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~--.., ORIGINAL GROUND ELEVATION \_SETTLEMENT PLATES,.. <n a: :11111,....J,.. :II, , COMPUTED SETTLEMENTS OF PEAT ll SILT 0.75'--~~~-'-~~~--L~~~~L-~~~.._~~~-'-~~~.&~~~~._~~~~ STATION Note: 1 m = 3.3 ft.

4 Transportation Researh Reord _._ Figure 4. Settlements for westbound roadway. --~-..,, OF OVERLOAD _ J TOP SETTLEMENT PLATES 1-., a 1- I-:: Note: 1 m = 3.3 ft _... _ STATION Figure 5. Pieometri heads for both roadways. 2!1 8 ~ I- 20 P-\ 6 > f \ I 1 \ 0 I- 0:: l!i I- I :I: 0:: 4 :I I I- I LEO END :I 0 10 PIEZOMETER STATION MAX. EXPECTED N NUMBER HEAD ELEV. ii:!i '" p I 'J EB FT ' I S EB 2o+oe 9.1 FT. 2 rf we HT. WB S.2 FT. o.....,.,. ~ APRIL JUNE AUG OCT DEC FEB 1978 APRIL 1178 JUNE 1171 failure has not ourred at this rate of appliation of surharge. SETTLEMENT DATA AND ANALYSIS Settlement-plate readings were taken for all settlement platforms from time to time. Figure 6 shows plots of settlement values for two of the settlement platforms for the eastbound setion. Plots for the westbound setion are shown in Figure 7. Settlement plots for stations under whih only landfill existed but no peat or organi material was enountered are shown in Figure 8. In general, all of these settlement-time urves are alike and not muh different in shape from those observed for fine-grained soils. In the early periods of time, the settlement inreases rapidly and then ontinues to inrease

5 50 Transportation Researh Reord 754 Figure 6. Settlement time for eastbound roadway. :II I.,, 0, SUltCHAltG[.,, ADDED D: :a I I :a TIME IN MONTHS Figure 7. Settlement time for westbound roadway. O -r------:= SETTLEMENT PLATE N0.39 STA I 85 WESTBOUND I- ::E I- I I- I-.,, ,, D: I 400 ::E I I ::E SETTLEMENT PLATE NO. 40 STA I 85 WESTBOUND SURCHARGE ADDEO TIME IN MONTHS at a dereasing rate. At some stations where additional surharge was applied at a later date-e.g., SP-26, SP-47, SP-8, and SP-39-as one would expet, the rate of settlement inreased. The thikness of the garbage fill along the westbound roadway varies from 5 to 9.5 m ( ft), and the thikness of the underlying peat and organi silt ranges from only 0.75 to 1.7 m ( ft). Therefore, a large proportion of the total settlement is ontributed by the garbage fill. In general, the total strain varies between 11 and 14 perent, with the exeption of three stations. The strain values are lower at the two end stations ( and ) beause the overload extended only about 15 m (50 ft) beyond the loation of the settlement platforms. The low strain value at station annot be explained. Below the settlement plates where there was only landfill and no peat or organi silt was found, as was the ase for station eastbound and station westbound, the measured strain ranged between 11.4 and 11.8 perent. This is in lose agreement with the values reorded for the westbound roadway. For both the roadways, the total settlement attributable to surharge varied, depending on the nature of the underlying materials. At the time surharging was ompleted, an average of about 66 perent of the total settlement had ourred, whereas the settlements one month and three months after the ompletion of surharging were, respetively, 77 and 83 perent of the total

6 Transportation Researh Reord Figure 8. Settlement time for landfill area only. o.e ::1 :...I I&. :. II) 1.8..I..I 800 : SETTLEMENT PLATE NO. 47 STA WESTBOUND SETTLEMENT PLATE N0.21 STA EASTBOUND SUltCHAIHE -ADDED SURCHARGE ADDED TIME IN MONTHS settlement. Chang and Hannon (!) reported that 80 perent of the settlement was attained within 30 days after the ompletion of overload on CA-52. In order to interpret the settlement response of the eastbound roadway, an analysis was arried out on the basis of the amount of landfill material removed or exavated and the amount of surharge added. In other words, the stress-history effets were onsidered: 1. Station has the lowest values of settlement beause there is neither landfill material nor organi silt and peat below it. The measured defletion of 60 mm (0.2 ft) is attributable to the settlement of newly plaed sandy bakfill material that the ontrator had end-dumped into water after having exavated all organi materials down to the sand and gravel stratum. 2. Four of the stations (18+00, 20+00, 24+00, and 26+00) show strains between 5 and 7 perent. For eah of these stations, the ratio between the inrease in stress attributable to the surharge load and the stress from the exavated landfill materials was!ii The remaining three stations for whih the strains vary between ll.4 and 16.8 perent had orresponding stress ratios of 1.4 and greater. SETTLEMENT OF NATURAL ORGANIC MATERIALS An attempt was made to separate the settlement ontribution of the landfill materials from those of the underlying natural deposits of peat and organi silts. Laboratory test results were used to ompute the settlement in the natural deposits. The differene between. the settlement-platform readings and those omputed for peat and organi silt was onsidered to reflet the settlement of the sanitary landfill. Sine onsolidation tests were available for a few typial samples, based on the lassifiation of underlying soils, the most appropriate soil parameters were seleted for omputation of settlement in the underlying natural deposits. One must keep this fat in mind in viewing these results. Initial alulations were made by using ground elevations and water levels as indiated by the boring logs. The results of this analysis yielded omputed settlements somewhat larger than the measured settlements. Upon reexamination, it was determined that groundwater and ground elevation, as shown on the boring logs, were different from those obs_erved at the time of onstrution. Comparative values. are given below, measured in feet in relation to New Jersey data at mean sea level. Water-table values indiate atual groundwater observed in the field at elevation plus 4.0 (1 ft = 0.3 m): Original Borings (%) Water Ground Atual Ground Station Table Elevation Elevation (ft) Eastbound o+OO Westbound o The error in the boring logs was believed to be aused by inorret inferenes by the boring rew, either beause of perhed water onditions or the water used during boring operations. After adjustments were made to ground and groundwater elevations, settlement omputations were revised. Both omputed and measured settlement values are given in Table 1. For the westbound roadway, the strain in the garbage fill and the underlying peat and organi silt is about the same. For the eastbound roadway, similar agreement is found for all but two stations. In the ase of station 26+00, the omputed settlements for the underlying natural deposits were equal to the total measured settlements, whih indiates that the landfill stratum that was 1.8 m (6 ft) thik had no settlement. Before the installation of the settlement platform, this area was used as a haul road. This onstrution ativity resulted in a high degree of ompation of the underlying garbage fill prior to the appliation of the surharge load.

7 52 Transportation Researh Reord 754 Table 1. Computed and measured settlement values. station Thikne ss (ft) Settlement Peat and Platform Landrlll Silt Settlement (ft) Strain(:') Total Peat and Silt (measured) (omputed) Landfill Total Landr!ll Eastbound ls+oo S+OO Westbound S Note: 1 ft 0.3 m. Figure 9. Strain versus stress for both roadways e 18l5 EASTBOUND G: UI O.ll O.Ol o. l5 0.6 loa (p 0 +t.p)/p 0, At station 22+00, after the adjustment was made to the total settlement, a muh greater strain was indiated to have ourred in the landfill. Sine most of the rest of the data indiate that the sanitary landfill had a strain of about 11 perent, inorret hoie of the soil parameters for the peat and organi silt may have been one of the reasons for this disrepany. SETTLEMENT OF LANDFILL MATERIALS Along the eastbound roadway, before the plaement of settlement platforms, it was neessary to remove about m (5-20 ft) of the landfill material while leaving several feet of the underlying landfill in plae. In the removal proess, suh artiles as the bodies of truks and washers and dryers were unovered. Along the westbound roadway, the top of the existing garbage fill was below the points where settlement plates were to be installed. The landfill material that was exavated from the eastbound setion was leared of larger artiles and reompated along the westbound setion. The ompated landfill material was as deep as 5.5 m (18 ft) in some setions. Beause of the sorting and ontrolled ompation onditions, the added garbage fill was onsidered to have a higher density [1450 kg/m 1 (90 lb/ft')] than the existing fill [1129 kg/m 1 (70 lb/ft')], whih was neither sorted nor ompated in any ontrolled manner. The relation between strain and stress for both roadways is shown in Figure 9. All but two points lie between two lines that have a slope-ompression ratio of 0.16 and As disussed before, the omputed strain value for the point that lies above the upper line that represents SP-11 is most likely inorret. Another point previously disussed is the fat that the low strain value for SP-39, whih lies below the lower line, ould not be explained. Thus, disregarding these two points, the average ompression ratio of the landfill is CONCLUSIONS Based on readings taken from several settlement platforms and pieometers along 1-85, whih is to be onstruted on a stabilied sanitary landfill material, the following onlusions an be made: 1. Pieometer readings were marred by the methane

8 Transportation Researh Reord gas that was still being generated in the landfill. 2. Settlement-time urves were found to have shapes similar to those of fine-grained soils. 3. Strain in the sanitary landfill was between 11 and 14 perent for the applied stress range. 4. The underlying deposits of peat and organi silt exhibited strains similar to those of the overlying landfill materials. 5. The ompression ratio for the sanitary and fill material, some of whih was undisturbed and some of whih was exavated and reompated, was between 0.16 and 0.20 and averaged The observed rates of settlement orrelate well with those reported for a highway onstruted on sanitary landfill in California (1) Refuse Landfills. In Geotehnial Pratie for Disposal of Solid Waste Materials, ASCE Speialty Conferene, Ann Arbor, MI,, pp G.F. Sowers. Foundation Problems in Sanitary Landfills. Journal of the Sanitary Engineering Division, ASCE, Vol. 94, No. SAl, 1968, pp R.E. Sheurs. Highway Embankment Constrution over a Sanitary Landfill. Department of Civil and Environmental Engineering, New Jersey Institute of Tehnology, Newark, J.C. Chang and J.B. Hannon. Settlement Performane of Two Test Highway Enbankments on Sanitary Landfill. In New Horions in Constrution Materials, Envo Publishing Co., In., Lehigh Valley, PA, Vol. 1, 1976, pp REFERENCES 1. S.K. Rao, L.K. Moulton, and R.K. Seals. Settlement of Publiation of this paper sponsored by Committee on Earthwork Constrution. Predition of Density and Strength for a Laboratory-Compated Clay D. W. WEITZEL AND C. W. LOVELL Preditions of ompated soil strength for a highly plasti lay are developed by statistially orrelating the resulu of unonsolidated undrained triaxial tests with the ompation variables of water ontent, dry density, and om pative work. Work was alulated from measurements of the fore and displaement of a kneading-ompator foot as it loaded the soil. The statistial models for predition of density and strength gave results onsistent with well-doumented experimental evidene. The model for density predition dry of optimum inluded variables of water ontent and a ompative work ratio; for ompation wet of optimum, dry density was a funtion of water ontent only. The model for strength predition dry of optimum inluded variables of water ontent, dry density, degree of saturation, and onfining pressure. Wet of optimum, the logarithm of strength dereased linearly with initial void ratio. All of these models had good statistial validity. By using suh models, the de1igner an predit soil strength at partiular ompation levels or, if a minimum strength value Is required, the neessary levels of ompation values an be estimated. These relations were developed from laboratory ompation. Similar relations for field ompation are being developed with the intention of orrelating the two. Compation is ommonly used to improve the strength of embankments, although the plaement speifiation usually ontrols only ertain ompation variables. This study developed preditions of strength from ompation variables that are more ommonly and simply measured than strength. The as-ompated strengths of a laboratory-ompated, highly plasti lay were measured in unonsolidated-undrained triaxial tests. Samples were prepared by kneading ompation to densities that fit on three impat-energy urves, on eah of whih were four water ontents. Samples were then sheared at four levels of onfining pressure to simulate a variety of ompation onditions and embankment depths. In addition, the work expended to ompat the soil was estimated from measurements of fore and displaement for the ompator foot. The results of these tests were used in statistial regression analysis to develop predition equations for density and as-ompated strength in terms of the ompation variables. Correlation of these results with similar ones urrently being developed for field ompation will allow better ontrol of short-term shear behavior in embankments. Suh preditions are of partiular interest as (a) strength beomes more extensively used as a ompation speifiation element and (b) greater attention is paid to the potential overloading of the ompated soil by onstrution equipment. LITERATURE REVIEW The lay fabri established by ompation has an important effet on soil behavior. Modern fabri explanations were first introdued by Barden and Sides (1) and then by Hodek (2) and his "deformable aggregate" - theory (deformable aggregate is an agglomeration of lay partiles, or maropeds). The sie and distribution of the pore spae of ompated-lay fabri have also been studied onsiderably in reent times. Bhasin (3) defined pore-5ie distributions for several different lays at various energy levels. The distributions of the pore sies for soil at equal porosities wet and dry of optimum were very different: The dry sample had larger pores than the wet one. In addition, as ompative effort inreased at a onstant water ontent dry of optimum, the quantity of larger pores was vastly redued until a point was reahed at whih further hanges in the pore-sie distribution would not our. These results agreed with those of Sridharan and others (4) and Ahmed and others (5). Garia-Bengohea (6) also reahed similar onlusions for a silt-kaolin mixture. Suh studies deemphasie the role of individual lay partiles in the ompation proess and fous on the nature and ation of olletions of partiles into groups alled domains, pakets, maropeds, or aggregates. The arrangements of these aggregates vary signifiantly wet and dry of optimum. Dry of optimum, the aggregates are distint. The void spae is prinipally between aggregates, and a onsiderable quantity of it is in larger pores. As optimum water ontent is approahed, the soil gets loser to its plasti limit and the aggregates beome more deformable. Hene, in ompation the aggregates distort p.nd squeee together, and this redues the number of large