Following a review of known past experiments with pile groups in sand, this paper presents the results of a research project devoted to comprehensive testing of large-scale models of pile groups in sand. The testing program included testing of four- and nine-pile groups of instrumented piles, with and without caps, in sand. The 4-in.-diameter, 60-in.-long piles were placed by jacking at spacings ranging from two to six pile diameters (center-to-center) into artificial deposits of dry and submerged sand in two soil situations: (a) homogeneous, medium dense sand (relative density about 65 percent) and (b) two-layer mass, consisting of an upper stratum of very loose sand (Dr ≈ 20 percent), underlain by a stratum of dense sand (Dr ≈ 80 percent). Parallel tests with single piles were made to allow comparison with behavior of single piles under otherwise identical conditions. The results indicate practically no group effects on ultimate point loads. However, a significant increase of ultimate skin loads is detected when the piles are placed in a group. The pile caps contribute to the overall bearing capacity of the groups inasmuch as they are supported by sand along their outer rim. The overall group efficiency of the entire group (with cap) in medium dense sand increases to a maximum of about 1.7 at spacings of three to four pile diameters, becoming somewhat lower with further increase of pile spacing. A sizeable part of the increase in bearing capacity comes from pile caps. If the loads transmitted by the caps are deduced, the group efficiency drops to a maximum of about 1.3. Measurements of axial loads in individual piles indicate little variation of load distribution in different loading stages. In the case of nine-pile group, the center pile carried from 20 to 50 percent more load than the average. The settlements of the groups are larger than those of individual piles carrying the same load. It appears that the group settlement ratio increases roughly as the square root of relative width of the pile group.