ABSTRACT Laboratory model pullout tests were conducted in a "rod analogy" soil model, dense dry sand, dense submerged sand, and Gulf of Mexico marine sediments at pullout inclination angles from 900 (vertical) to 450 and at two burial depths. Equations are, developed from the experimental data that 0can used for design purposes. INTRODUCTION The expansion of the field of coastal and ocean engineering has resulted in a corresponding increase in applications for floating equipment, anchored in shallow or deep water, many of which impose a stringent station-keeping requirement. One of the solutions frequently used in these applications is a taut-line mooring system. However, the design of these systems demands consideration of the problem of loads on the anchor that are always more or less inclined from the vertical, due to the forces ocean currents and wind often exert against floating equipment. The effects of such loap' inclinations are not well known, since the phenomena of the resulting seafloor soil failure caused by excessive inclined loads have not been observed and understood. Lacking this information, the practicing ocean engineer must estimate the holding capacity in his anchor system design very conservatively whenever inclined loads can possibly occur. MODEL EXPERIMENTS Test Series The initial laboratory tests were designed to allow the observation in a two-dimensional array of particle motion, in a simulated, dense, cohesionless, granular material during pullout of a plate under the application of inclined loads. The failure phenomena that occurred during these modes of loading could be observed, which would aid in understanding them. The second series of tests was performed in a three-dimensional tank of dense, air-dry Ottawa sand. Values of maximum pullout loads for varying conditions of load inclination, and depth of burial, were measured. The third series of tests duplicated the second series except that the dense, Ottawa sand samples were submerged under fresh water. This series of tests was analogous to the second series, but showed the effects of submergence. The fourth series of tests was performed in the same manner and with the same equipment, except that the soil sample was composed of actual seafloor sediments obtained from the Gulf of Mexico. 'lhus, the differences in pullout force and failure meclianisms that occur between a fine-grained marine material and the granular materials examined in the second and third series, could be compared. Two-Dimensional Tests The two-dimensional tests utilized a plexiglas tank 3 ft high by 4 ft wide by 3/4 in. thick. The tank accommodated an array of approximately 150,000 steel rollers 1/8 in. in diameter and 3/4 in. in length. 'they were stacked with all of their cylindrical axes parallel and with their ends constrained by the plexiglas faces of the tank. 'the ends of the rollers were large enough so that photographs of the array taken through the plexiglas plainly show each roller and its displacements.