Soil reinforcement by roots is studied by considering the contribution of the tensile force in a root segment that intersects a potential slip surface in a soil‐root system. To evaluate the tensile force when the system is subjected to a shear displacement, the root segment is analyzed as a beam on elastic‐plastic support and as a cable for small and large displacements, respectively. Equilibrium and displacement compatibility at a branch point are used to analyze the distribution of forces between two root branches of a root system. Laboratory model tests and in situ root tests were performed to verify the analytical models. The analytical models and in situ root tests are used as means of evaluating the shearing resistance of the reinforced soil.

In situ shear tests were performed on soils reinforced by roots. Analytical models of soil‐root interaction presented in a companion paper were used to compute the forces in the roots and the contribution of the roots to the shearing resistance of the soil‐root system. General agreement between measured and computed shearing resistances was taken as verification of the analytical models. Results of the computations also showed the influence of the dimensions of the shear box on the measured shear strength. The shear strengths measured by the in situ tests were found to be less than the shear strength along a potential slip surface in slope stability problems. Methods for estimating the shear strengths for stability analysis are presented.

This paper summarizes the experimental results of cyclic triaxial and resonant column tests with artificially cemented sand specimens. Since the samples for both types of tests were prepared with one‐to‐one correspondence of parameters such as cement content, curing period, relative density, and effective confining pressure, it was possible to directly correlate the results of both types of tests. The existence of good correlation between dynamic moduli and liquefaction resistance determined in the laboratory has been confirmed by this study. Though the correlations are dependent on type of sand, level of cementation and effective confining pressure, they provide laboratory estimates of cyclic strength of artificially cemented sands and possibly naturally cemented sands from measured moduli or wave velocities.

During the Central Chile earthquake of 1985, two small dams suffered substantial embankment deformations indicating liquefaction and major strength loss in the embankment or foundation soils. However, the configuration of the embankments after failure suggested that the soil had retained a small but significant residual strength after initial liquefaction $(r_u\approx 100\%)\text{.}$ Because of the scarcity of field values of residual strength after initial liquefaction, a study of these dams was undertaken. Standard penetration test (SPT) values were determined and normalized for effective confining stress and hammer energy. Silty sand layers of low density were identified near the base of both embankments. Subsequent analysis of the results of the field study gave upper and lower bounds of residual strength in these loose layers. Calculated values of residual strength were found to vary with the inferred pre‐earthquake densities of the materials. These values were compared with those determined from previous case studies and shown to be in reasonable agreement.

Closed-form expressions for the earthquake-induced shear stress magnitude in soils and its phase shift with respect to the ground acceleration are presented. They are expressed as functions of the depth, the damping ratio of the soil, and the thickness of the resonant layer. In the development of these expressions the soil was treated as a viscoelastic material with a horizontal surface and harmonic surface displacement. The natural frequency is determined by assuming that a known depth of soil is in resonance. The shear stress induced in the soil by the earthquake is then evaluated from the governing equation of motion and expressed in complex form. As in the simplified method of evaluating soil liquefaction potential, the magnitude of the shear stress is presented as the product of the rigid-body shear stress and a stress-reduction coefficient. By doing this an analytic expression for the stress-reduction coefficient is derived.

Some initial results of centrifuge model tests to examine the mechanism causing flow failure of a less permeable layer resting on a more permeable layer, subjected to seismic loading conditions, is presented. The centrifuge model that was subjected to a base motion using a servohydraulic shaker, consisted of a clay layer overlying a sand embankment submerged in water. During shaking the clay layer almost completely flowed off the sand and collected at the toes of the embankment. Settlement of the sand was also observed during shaking. As the soil overlying the sand is relatively impervious before flow failure, sand particles rearrange without any global change in volume. Locally, however, it is possible that loosening of sand close to the interface between sand and clay did occur due to the densification of the sand beneath it. This may leave a liquid film or a layer of very loose sand at the interface thus creating an unstable situation which will result in the creation of a flow slide of the impermeable material overlying the sand.