The Plymouth-Carver aquifer underlies an area of 140 square miles and is the second largest aquifer in areal extent in Massachusetts. It is composed primarily of saturated glacial sand and gravel. The water-table and bedrock surface were mapped and used to determine saturated thickness of the aquifer, which ranged from less than 20 feet to greater than 200 feet. Ground water is present mainly under unconfined conditions, except in a few local areas such as beneath Plymouth Harbor. Recharge to the aquifer is derived almost entirely from precipitation and averages about 1.15 million gallons per day per square mile. Water discharges from the aquifer by pumping, evapotranspiration, direct evaporation from the water table, and seepage to streams, ponds, wetlands, bogs, and the ocean. In 1985, water use was about 59.6 million gallons per day, of which 82 percent was used for cranberry production. The Plymouth-Carver aquifer was simulated by a three-dimensional, finite difference ground-water-flow model. Most model boundaries represent the natural hydrologic boundaries of the aquifer. The model simulates aquifer recharge, withdrawals by pumped wells, leakage through streambeds, and discharge to the ocean. The model was calibrated for steady-state and transient conditions. Model results were compared with measured values of hydraulic head and ground-water discharge. Results of simulations indicate that the modeled ground-water system closely simulates actual aquifer conditions. Four hypothetical ground-water development alternatives were simulated to demonstrate the use of the model and to examine the effects on the ground-water system. Simulation of a 2-year period of no recharge and average pumping rates that occurred from 1980-85 resulted in water-level declines exceeding 5 feet throughout most of the aquifer and a decrease of 54 percent in average ground-water discharge to streams. In a second simulation, four wells in the northern part of the area were pumped at 10.4 million gallons per day in excess of rates simulated in the steady-state model for the four wells. This resulted in water-level declines of 2 feet or more in an area of 25 square miles and a decline in average ground-water discharge to streams of 6 percent. When this pumpage was simulated as recharge to the aquifer, water levels beneath the recharge area rose more than 40 feet, and ground-water discharge remained equal to average discharge in the calibrated steady-state model. In a third simulation, all 21 existing production wells were pumped at nearly the design capacity of 17.8 million gallons per day; this pumping rate produced water-level declines of less than 2 feet throughout most of the aquifer. When simulated pumpage was increased to 32.8 million gallons per day from existing wells and from 15 additional wells, the area where water-level declines exceeded 2 feet significantly increased. In another set of simulations, a well field close to a stream was pumped at rates of 2, 4, and 6 million gallons per day. At a pumping rate of 6 million gallons per day, ground-water discharge to the stream decreased 34 percent during periods of normal precipitation and 56 percent during drought conditions.