Management alternatives that could help mitigate the effects of water withdrawals on streamflow in the Ipswich River Basin were evaluated by simulation with a calibrated Hydrologic Simulation Program--Fortran (HSPF) model. The effects of management alternatives on streamflow were simulated for a 35-year period (1961-95). Most alternatives examined increased low flows compared to the base simulation of average 1989-93 withdrawals. Only the simulation of no septic-effluent inflow, and the simulation of a 20-percent increase in withdrawals, further lowered flows or caused the river to stop flowing for longer periods of time than the simulation of average 1989-93 withdrawals. Simulations of reduced seasonal withdrawals by 20 percent, and by 50 percent, resulted in a modest increase in low flow in a critical habitat reach (model reach 8 near the Reading town well field); log-Pearson Type III analysis of simulated daily-mean flow indicated that under these reduced withdrawals, model reach 8 would stop flowing for a period of seven consecutive days about every other year, whereas under average 1989-93 withdrawals this reach would stop flowing for a seven consecutive day period almost every year. Simulations of no seasonal withdrawals, and simulations that stopped streamflow depletion when flow in model reach 19 was below 22 cubic feet per second, indicated flow would be maintained in model reach 8 at all times. Simulations indicated wastewater-return flows would augment low flow in proportion to the rate of return flow. Simulations of a 1.5 million gallons per day return flow rate indicated model reach 8 would stop flowing for a period of seven consecutive days about once every 5 years; simulated return flow rates of 1.1 million gallons per day indicated that model reach 8 would stop flowing for a period of seven consecutive days about every other year. Simulation of reduced seasonal withdrawals, combined with no septic effluent return flow, indicated only a slight increase in low flow compared to low flows simulated under average 1989-93 withdrawals. Simulation of reduced seasonal withdrawal, combined with 2.6 million gallons per day wastewater-return flows, provided more flow in model reach 8 than that simulated under no withdrawals.