Limited available energy is one of the main challenges in Wireless Sensor Networks (W). Wireless nodes in W are typically powered by a limited battery which its energy needs to be consumed for sensing, processing, and communications of the sensor node. At the same time, a very long lifetime is usually expected which means the wireless nodes are expected to work for months or ever years without any intervention like battery recharge or replacement. However, such requirement cannot be fulfilled taking into account the current power consumption profile of state of the art low-power transceivers and embedded processors. One of the prominent techniques for enlarging the lifetime of wireless sensor nodes in W is duty cycling in which the nodes periodically fall into sleep mode and wake up for sensing, processing and communication in a short duration of time. The exact value of duty cycle directly influence the energy consumption of the wireless nodes, and picking the right value has a great impact on the quality-of-services provided by the WSN. If the wake-up duration is too short, the WSN performance such as end-to-end latency is degraded. If this duration is unnecessarily long, the energy consumption of the nodes increases which leads to shorter lifetime of the nodes and the WSN. In general, duty cycling can be static or adaptive. In static duty cycling, the sleep and wake-up durations are set and fixed at design time. In adaptive duty cycling, the duty cycle may change during network-operation based on the node conditions in the network. This thesis aims at designing an efficient adaptive duty cycling mechanism for W with very long sleep periods. Such W may, for instance, be exploited in environment monitoring applications such as water or air quality monitoring systems. In such application, the sampling rate of sensors is very low in the order of one sample per one or more hours. In this situation, the sleep periods between consecutive wake-up periods are very long. Thus, if the data generated by a sensor node is not delivered to the base station during a single wake-up period, its latency will be dramatically high, which cannot be tolerated by the application. We propose and develope a mechanism using which each sensor node gathers useful global information while it is contributing to the multi-hop data propagation in the network. Such global information about the network and the node’s role in data propagation is then used to set a proper value for the wake-up period of the node. The proposed technique is implemented and applied in various W with different topologies. Simulation results show that our adaptive duty cycling does not deteriorate the end-to-end data delivery latency, but saves up to 50% energy consumption is some networks.