This paper presents a systematic measurement of packet delivery performance in three different environments: an indoor office building, a habitat with moderate foliage, and an open parking lot, using up to 60 Mica motes. The study focuses on the packet delivery performance at the physical layer and the medium-access layer, and examines how environmental factors, transmit power, and physical layer coding schemes affect packet loss and reception rates. At the physical layer, packet delivery performance is influenced by environmental factors, physical layer coding schemes, and the spatial separation between sender and receiver. The study finds that in all environments, a significant portion of links experience high packet loss, with some links experiencing over 30% packet loss. The variability in packet loss is attributed to the existence of a "gray area" within the communication range of a node, where receivers are likely to experience choppy packet reception. This gray area is also characterized by significant variability in packet reception over time. While more sophisticated physical layer coding schemes can mask some of this variability, they may do so at the expense of bandwidth efficiency. At the medium-access layer, packet delivery performance is affected by interfering transmissions and the effectiveness of mechanisms such as carrier sense and link-layer retransmissions. The study finds that packet losses at the MAC layer also exhibit heavy tails, and that the efficiency of the MAC layer is low, with 50% to 80% of communication energy wasted in overcoming packet collisions and environmental effects. In harsher environments, nearly 10% of the links exhibit asymmetric packet loss. The study also examines the relationship between signal strength and packet delivery performance, finding that signal strength can provide some indication of link quality, although it is not a perfect predictor of packet loss. The study further explores the impact of different physical layer coding schemes on packet delivery performance, finding that SECDED coding performs better than 4b6b and Manchester coding in terms of packet loss reduction, although at the expense of bandwidth efficiency. The study concludes that packet delivery performance in wireless sensor networks is highly variable and influenced by environmental factors, transmit power, and physical layer coding schemes. The findings suggest that topology control mechanisms that carefully select neighbors based on measured packet delivery performance could significantly improve packet delivery in sensor networks. The study also highlights the importance of understanding the temporal and spatial characteristics of packet delivery performance for the design and evaluation of sensor network communication protocols.This paper presents a systematic measurement of packet delivery performance in three different environments: an indoor office building, a habitat with moderate foliage, and an open parking lot, using up to 60 Mica motes. The study focuses on the packet delivery performance at the physical layer and the medium-access layer, and examines how environmental factors, transmit power, and physical layer coding schemes affect packet loss and reception rates. At the physical layer, packet delivery performance is influenced by environmental factors, physical layer coding schemes, and the spatial separation between sender and receiver. The study finds that in all environments, a significant portion of links experience high packet loss, with some links experiencing over 30% packet loss. The variability in packet loss is attributed to the existence of a "gray area" within the communication range of a node, where receivers are likely to experience choppy packet reception. This gray area is also characterized by significant variability in packet reception over time. While more sophisticated physical layer coding schemes can mask some of this variability, they may do so at the expense of bandwidth efficiency. At the medium-access layer, packet delivery performance is affected by interfering transmissions and the effectiveness of mechanisms such as carrier sense and link-layer retransmissions. The study finds that packet losses at the MAC layer also exhibit heavy tails, and that the efficiency of the MAC layer is low, with 50% to 80% of communication energy wasted in overcoming packet collisions and environmental effects. In harsher environments, nearly 10% of the links exhibit asymmetric packet loss. The study also examines the relationship between signal strength and packet delivery performance, finding that signal strength can provide some indication of link quality, although it is not a perfect predictor of packet loss. The study further explores the impact of different physical layer coding schemes on packet delivery performance, finding that SECDED coding performs better than 4b6b and Manchester coding in terms of packet loss reduction, although at the expense of bandwidth efficiency. The study concludes that packet delivery performance in wireless sensor networks is highly variable and influenced by environmental factors, transmit power, and physical layer coding schemes. The findings suggest that topology control mechanisms that carefully select neighbors based on measured packet delivery performance could significantly improve packet delivery in sensor networks. The study also highlights the importance of understanding the temporal and spatial characteristics of packet delivery performance for the design and evaluation of sensor network communication protocols.