Horizontal drains are commonly used to stabilize slopes or embankments in which ground‐water seepage is the main cause of instability. When using horizontal drains to remedy an unstable slope, the length, spacing, and position of the drains must be chosen. These drain‐placement parameters are usually chosen on the basis of engineering judgment; sometimes they are obtained from published design charts. This paper presents a methodology that can be used to develop or experimentally verify such charts. Model slopes are tested in a geotechnical centrifuge to evaluate the influence of horizontal drains on the distribution of piezometric levels within slopes. The observed results, from those experiments in which drains were installed in the model slopes, were compared to reference conditions obtained from experiments without drains. The limited capacity of the centrifuge used in this research precludes a complete verification of the published design charts, but the results demonstrate the usefulness of the centrifugal modeling technique in studying the process of slope stabilization.

The deformation behavior of sand under shear loading is studied, based on the concept of particulate mechanics with consideration of the slip and nonslip contacts between the sand grains on a macroscopic shear plane. The “fabric factor” of a sand, which depends on the distribution of the contact angles between sand grains, is introduced to correlate the state of stress and the fabric of sand under shear loading. The relationship between the shear‐stress ratio and the fabric factor, and the relationship between the fabric factor and the dilatancy rate of a sand are developed. A stress‐dilatancy relation for sand under shear loading is then derived. The proposed model can better describe the deformation behavior of sand under cyclic loading. The results of slow, cyclic, drained, direct shear tests and drained, cyclic, triaxial shear tests agree reasonably well with the proposed stress‐dilatancy relation. The experiments also confirm the concept of fabric factor and existence of nonslip contacts between sand grains.

A highly plastic clay liner was compacted in the field with full‐sized equipment to an average of more than 100% of standard Proctor maximum dry density. The average field permeabilities of two sections of the liner, as measured with four sealed double‐ring infiltrometers and a $16\times 16\mathrm{ft}(4.88\times 4.88m)$ pan lysimeter, were $3\times {10}^{-5}$ and $10\times {10}^{-5}\mathrm{cm}/s\text{.}$ Laboratory permeability tests were conducted on samples compacted in the laboratory by three procedures, only one of which gave a reasonable prediction of the permeability of the liner. The high permeability of the liner was caused by the failure to achieve the two “basic” compaction objectives that govern the permeability of compacted clays. The basic objectives (destruction of soil clods and bonding between lifts) and 17 other key factors that influence the permeability of compacted clays explain the performance of the liner and provide a basis for understanding how to construct liners to avoid such high permeabilities.

Laboratory triaxial compression tests are performed to determine the static stress‐strain response of sands reinforced with discrete, randomly distributed fibers, and to observe the influence of various fiber properties, soil properties, and test variables on soil behavior. In addition to the experimental program, a model is developed, based on a statistical theory of strength for composites, to predict the fiber contribution to strength under static loads. Randomly distributed fiber inclusions significantly increase the ultimate strength and stiffness of sands. The increase in strength and stiffness is a function of sand granulometry (i.e., gradation and particle size and shape) and fiber properties (i.e., weight fraction, aspect ratio, and modulus). The sand‐fiber composites have either a curved linear or a bilinear failure envelope, with the break occurring at a threshold confining stress called the “critical confining stress.” The magnitude of the critical confining stress decreases with an increase in sand gradation, particle angularity, and fiber aspect ratio, and increases with an increase in fiber modulus. The critical confining stress is insensitive to changes in sand particle size and fiber content. Predicted strength increases from fiber reinforcement using a theoretical model based on a statistical theory of strength for composites agree reasonably well with measured values.

The lateral load‐deflection response of bored piles in cemented sands was examined by field tests on single piles, and on small groups consisting of two piles. All 12 piles were 0.3 m in diameter and had a length of 3 m or 5 m. Two piles were instrumented with electrical resistance strain gages. The surface soils to a depth of 3.5 m were characterized as having both components of shear strength, c and ϕ. The measured lateral load deflection behavior was nonlinear, with a stiffer response for the restrained piles in the pile groups. Theoretical parabolic p‐y curves were employed in the cemented sand with conventional lateral earth pressure theories. These theoretical curves were in agreement with the p‐y curves determined from the strain gage measurements. Ignoring the cohesion intercept of cemented sand resulted in an incorrect and too flexible response compared with the measured behavior. Retesting the piles under lateral loads in a perpendicular direction indicated little difference in the lateral capacity and deflection.

A procedure is presented for estimating the small strain shear modulus $G_{\max }$ of soils based on flat dilatometer (DMT) test results. Previously published data from nine test sites at which both DMT and seismic tests were performed are used to develop a model. The nine sites represent various soil conditions ranging from soft clays to dense, overconsolidated sands. Earlier efforts had sought a relationship between Gmax and either the DMT modulus or the DMT horizontal stress index. The present study shows that a semiempirical expression for Gmax similar to Hardin's equation produces better results for a wider range of soil types. The expression includes terms for the effective confining stresses in the directions of wave propagation and particle motion and the DMT‐based total unit weight as a surrogate for void ratio. The average error between the observed Gm« and that determined by the model was 23%.

The stochastic finite element method is applied to study the effect of a spatially varying soil medium on the settlement of single‐pile foundations. In this study, the soil medium is assumed to be linear‐elastic, and it is modeled as a homogeneous random field. The random field captures the uncertainty in the soil modulus by means of the coefficient of variation and the scales of fluctuation. Using a first‐order second‐moment technique, the mean and coefficient of variation of the pile head settlement are determined. It is found that the uncertainty of the pile head settlement is dependent on both deterministic and stochastic parameters. Reliability analysis is then performed to compute the reliability index and the corresponding probability of unserviceable behavior. In order to facilitate the probabilistic design of pile foundations, the results of the analysis are presented in the form of concise charts suitable for use by practicing engineers.

The behavior of the pavement-subsoil system is analyzed from compaction due to traffic. A method of analysis is proposed based on the simple compaction theory and numerical algorithm that makes use of a standard finite element code for an elastic medium in plane strain conditions. Numerical results, computed for typical data, illustrate the method. Some conclusions of practical importance are drawn.

An analysis of the apparent overconsolidation properties of deep-sea abyssal plain sediments taken from holes DSDP 576A and MPG-1 is presented. Apparent overconsolidation is believed to be caused by interparticle bonding, or cementation. The cementation can be caused by any one or a combination of the following: silica, potassium, calcium carbonate, ferric oxide, manganese, or biological excretions. Plotting OCR versus depth is shown to result in a ``spurious correlation'' because both depth and OCR have the same independent variable in common, namely depth (z). An alternative method for presenting consolidation information is to plot preconsolidation stress versus depth, particularly when the depth is small. Using this procedure plots of overburden and preconsolidation stresses versus depth indicates that for hole 576A an overconsolidated condition exists down to a depth of 35 m. Below this depth an apparent condition of underconsolidation exists. Hole MPG-1 shows an over consolidated condition exists to the maximum depth investigated of 3.5 m.

The effect of stiffness of a shallow foundation on its bearing capacity has hardly been studied, although it is technically possible to construct a foundation with durable flexible interface. Reported here are the results of a preliminary experimental study aimed at observing the behavior of small-scale flexible and rigid foundation models, resting on dense sand, at their ultimate load-bearing capacity. The results indicate that reducing the foundation's stiffness at its soil interface may significantly increase its load-bearing capacity. This increase, however, is associated with increased settlement. No consistent differences in the slip surface geometries were observed. It is suggested that the apparent increase in bearing capacity is due to differences in the contact pressure distributions combined with the phenomenon of progressive failure. Upon further verification of this work's observations, its results are potentially applicable to foundations of structures such as mechanically stabilized earth or retaining walls founded on strip footing, where design is controlled by a factorized load-bearing capacity value.