||Although excavated rectangular piles, i.e., barrettes, have been extensively used recently, relatively few investigations have been carried out to study the fundamental behaviour of barrettes. The objectives of this research are to improve the understanding of: 1) the ground deformation and soil stress change during the installation of a barrette; 2) the mobilisation behaviour of the shaft resistance of a barrette under vertical loading. These objectives have been achieved by carrying out full-scale field tests, analytical study and three-dimensional numerical modelling. A new, simple and explicit plane stress analytical elastic solution was derived and calculation charts and tables were developed for calculating the horizontal stress changes and displacements caused by the installation of a barrette. It has been shown that for a long and thin barrette, stress relief and stress increase are generally induced in the soil around the faces and the ends of the barrette trench during the bentonite stage, respectively. The larger the aspect ratio (length to width) of a barrette, the greater is the magnitude and the more extensive is the influence zone of the horizontal stress changes and displacements caused by the installation. The maximum horizontal displacement is approximately proportional to the length of the barrette. The applicability of the proposed two-dimensional analytical solution to three-dimensional barrette installation problem was subsequently examined by using a three-dimensional numerical technique. In addition, the loading test on a 40 m deep barrette constructed at Kowloon Bay provided a comprehensive case history of the soil response to barrette installation. It has been observed that during installation of the barrette, the total lateral pressure on the sidewalls of the barrette trench was first reduced from the total earth pressure at rest to the hydrostatic bentonite pressure and then increased to the wet concrete pressure given approximately by a theoretical bilinear envelope. Also observed are small ground movements. During the vertical load testing of the barrette at Kowloon Bay, the installed total earth pressure cells and piezometers provided a valuable opportunity to measure the lateral earth pressure and pore water pressure changes at the barrette-soil interface. When loading caused significant slippage of the barrette, a substantial reduction in the lateral earth pressure was observed, together with a sudden drop of excess pore water pressure. Moreover, a detailed database has been developed from 15 comparative full-scale compression loading tests on barrettes founded in granitic saprolites in Hong Kong. Test results from the database barrettes were critically reviewed and analysed with the emphasis on the mobilisation behaviour of the shaft resistance of the barrettes and construction effects. According to the limited data available, a minimum local displacement (i.e., relative pile-soil movement) required for the full mobilisation of the shaft resistance capacity of barrettes in granitic saprolites was about 20 mm. It was found that the post-grouting technique improves the shaft resistance of barrettes; however, construction time, including the duration of excavation and trench standing time, has an adverse effect on the shaft resistance. According to three-dimensional numerical analyses, a uniformly distributed shaft resistance is not mobilised across the barrette until the failure of the soil nearest the barrette has been reached.