This paper introduces a geographical adaptive fidelity (GAF) algorithm to reduce energy consumption in ad hoc wireless networks. GAF conserves energy by identifying nodes that are equivalent from a routing perspective and turning off unnecessary nodes, maintaining a constant level of routing fidelity. GAF uses application- and system-level information to moderate this policy, keeping nodes that source or sink data active while intermediate nodes monitor and balance energy use. GAF is independent of the underlying ad hoc routing protocol and has been simulated over unmodified AODV and DSR. Analysis and simulation studies show that GAF can consume 40% to 60% less energy than an unmodified ad hoc routing protocol. Simulations also show that network lifetime increases proportionally to node density, with a four-fold increase in node density leading to a 3 to 6 times increase in network lifetime. GAF is an example of adaptive fidelity, a technique that extends the lifetime of self-configuring systems by exploiting redundancy to conserve energy while maintaining application fidelity. The paper discusses the energy consumption of four ad hoc routing protocols (AODV, DSR, DSDV, and TORA) using a simple traffic model. It finds that on-demand protocols like AODV and DSR consume much less energy than a priori protocols like DSDV. Energy use is dominated by routing protocol overhead, with overhearing being the major source of extraneous energy consumption. The paper explores approaches to avoid overhearing, such as the PAMAS protocol and TDMA protocols. The paper presents the GAF algorithm, which uses location information to associate nodes with a "virtual grid" and determine which nodes to turn off. Nodes in the same grid coordinate to determine who will sleep and how long, using application and system information. Nodes periodically wake up to accomplish load balancing. GAF interacts with the underlying ad hoc routing protocol and uses information from above the MAC-layer to control radio power. The paper evaluates GAF's performance in terms of network lifetime, energy conservation, and data delivery quality under low mobility. It shows that GAF extends network lifetime considerably, with 30-40% of nodes still alive after 900 seconds. GAF also saves energy, with 40% lower energy consumption than AODV. GAF maintains a constant level of routing fidelity, with data delivery ratios of 99% and mean delays similar to AODV. The paper also evaluates GAF's performance under high mobility, showing that GAF-b (without mobility adaptation) has worse data delivery quality than GAF-ma (with mobility adaptation). GAF-ma maintains better data delivery quality and energy savings under high mobility. The paper also shows that GAF's network lifetime is proportional to node density, with a four-fold increase in node density leading to a 3 to 6 times increase in network lifetime. GAF is also robust to location error and shadowing propagation models.This paper introduces a geographical adaptive fidelity (GAF) algorithm to reduce energy consumption in ad hoc wireless networks. GAF conserves energy by identifying nodes that are equivalent from a routing perspective and turning off unnecessary nodes, maintaining a constant level of routing fidelity. GAF uses application- and system-level information to moderate this policy, keeping nodes that source or sink data active while intermediate nodes monitor and balance energy use. GAF is independent of the underlying ad hoc routing protocol and has been simulated over unmodified AODV and DSR. Analysis and simulation studies show that GAF can consume 40% to 60% less energy than an unmodified ad hoc routing protocol. Simulations also show that network lifetime increases proportionally to node density, with a four-fold increase in node density leading to a 3 to 6 times increase in network lifetime. GAF is an example of adaptive fidelity, a technique that extends the lifetime of self-configuring systems by exploiting redundancy to conserve energy while maintaining application fidelity. The paper discusses the energy consumption of four ad hoc routing protocols (AODV, DSR, DSDV, and TORA) using a simple traffic model. It finds that on-demand protocols like AODV and DSR consume much less energy than a priori protocols like DSDV. Energy use is dominated by routing protocol overhead, with overhearing being the major source of extraneous energy consumption. The paper explores approaches to avoid overhearing, such as the PAMAS protocol and TDMA protocols. The paper presents the GAF algorithm, which uses location information to associate nodes with a "virtual grid" and determine which nodes to turn off. Nodes in the same grid coordinate to determine who will sleep and how long, using application and system information. Nodes periodically wake up to accomplish load balancing. GAF interacts with the underlying ad hoc routing protocol and uses information from above the MAC-layer to control radio power. The paper evaluates GAF's performance in terms of network lifetime, energy conservation, and data delivery quality under low mobility. It shows that GAF extends network lifetime considerably, with 30-40% of nodes still alive after 900 seconds. GAF also saves energy, with 40% lower energy consumption than AODV. GAF maintains a constant level of routing fidelity, with data delivery ratios of 99% and mean delays similar to AODV. The paper also evaluates GAF's performance under high mobility, showing that GAF-b (without mobility adaptation) has worse data delivery quality than GAF-ma (with mobility adaptation). GAF-ma maintains better data delivery quality and energy savings under high mobility. The paper also shows that GAF's network lifetime is proportional to node density, with a four-fold increase in node density leading to a 3 to 6 times increase in network lifetime. GAF is also robust to location error and shadowing propagation models.