There are limitations in experimental studies on sand-gravel mixtures due to the small size of testing

specimens. Due to this problem, many researchers have worked on prediction of the shear strength of mixture by testing

the sandy fraction of soil alone and developed empirical relationships. Most of the previous relationships have been

determined for low surcharge pressures in which particle breakage does not affect the shear strength parameters.

However, the particle breakage affects the relationships in higher confinements. At the present study, the results of

large scale direct shear tests on sand and sand-gravel mixtures was used to investigate the shear behavior and

dilatancy characteristics in a wider range of surcharge pressures. The gravel content, relative density, surcharge

pressure and gravel grain size were considered as variables in testing program. The relationships between shear

strength characteristics of sand and sand-gravel mixtures were determined considering dilation characteristics of the

soil. In this regard, the minimum void ratio was found as a useful indirect index that relates uniquely to the critical

state friction angle independent of soil gradation. The relations between critical state or peak friction angles of the

mixture with minimum void ratio were determined as a function of surcharge pressure. The correlations could be useful

for determination of the strength parameters of sand-gravel composites by testing sandy fraction of mixture.

The work presented in this paper investigates the causes of size effects in structural-concrete members. It is

based on the use of a finite-element model found to yield realistic predictions of structural-concrete behavior in all

cases investigated to date. In fact, the previous use of this model in investigations of size effects in reinforced-concrete

beams indicated that such effects reflect the dependence of load-carrying capacity on small unintended eccentricities

of the applied load and/or load-induced anisotropy, rather than, as widely considered, on fracture-mechanics

characteristics. The present work extends the scope of the above investigation so as to include the case of reinforced

concrete flanged shear walls, the behavior of which is already established experimentally. It is found that, unlike the

flanged shear walls with a height-to-length ratio larger than 2, the shear walls investigated in the present work, in

contrast with the interpretation given to recently published experimental findings, are size-effect independent.