A principle of similarity applicable to saturated sands under the so called one gravity (1-g) model testing, as opposed to centrifugal testing say, is developed. The study is a fairly rigorous treatment of some work reported earlier by the Author and his then supervisor, the late Professor K. H. Roscoe, more than three decades ago, Roscoe and Poorooshasb (1963) and a further expansion of a study presented recently, Poorooshasb and Ishihara (1993). The development of the principle is carried out using the mixture theory and two special constitutive models. The first approach uses the constitutive model CANAsand to express the mechanical behavior of the solid phase. The second constitutive model treats the solid phase of the soil-water system as a poro-elastic material. It is shown that with a proper choice of state parameters it is possible to achieve similarity between two systems, the model and the prototype say, under quasi static loading conditions in both cases (poro-elastic and CANAsand model). When dealing with dynamic loading, however, similarity can be achieved only if the soil skeleton is assumed to be poro-elastic. With the CANAsand model the behavior of the prototype can not be simulated by a model. This finding is of paramount significance : it casts a doubt over the validity of all shaking table tests carried out thus far.
