This paper analyzes packet loss in a 38-node urban multi-hop 802.11b mesh network. The study identifies that packet loss rates are relatively uniform across the range of loss rates, with no clear threshold between "in range" and "out of range." Most links have stable loss rates, though some exhibit bursty behavior. Signal-to-noise ratio and distance have limited predictive value for loss rates, and intermediate loss rates are likely due to multi-path fading rather than attenuation or interference. The findings suggest that the neighbor abstraction is a poor approximation of reality, as most node pairs have intermediate loss rates. The paper explores the causes of packet loss, including multi-path fading, signal strength, and interference, and highlights the implications for MAC and routing protocol design. The study also shows that packet loss is influenced by time scale, with bursty behavior observed in some links. The results indicate that signal-to-noise ratio alone is not sufficient to predict packet loss, and that multi-path fading is a significant factor. The paper concludes that the neighbor abstraction is not well-suited for wireless networks, and that routing protocols must account for the prevalence of intermediate loss rates. The study provides insights into the behavior of wireless networks and the design of future MAC and routing protocols.This paper analyzes packet loss in a 38-node urban multi-hop 802.11b mesh network. The study identifies that packet loss rates are relatively uniform across the range of loss rates, with no clear threshold between "in range" and "out of range." Most links have stable loss rates, though some exhibit bursty behavior. Signal-to-noise ratio and distance have limited predictive value for loss rates, and intermediate loss rates are likely due to multi-path fading rather than attenuation or interference. The findings suggest that the neighbor abstraction is a poor approximation of reality, as most node pairs have intermediate loss rates. The paper explores the causes of packet loss, including multi-path fading, signal strength, and interference, and highlights the implications for MAC and routing protocol design. The study also shows that packet loss is influenced by time scale, with bursty behavior observed in some links. The results indicate that signal-to-noise ratio alone is not sufficient to predict packet loss, and that multi-path fading is a significant factor. The paper concludes that the neighbor abstraction is not well-suited for wireless networks, and that routing protocols must account for the prevalence of intermediate loss rates. The study provides insights into the behavior of wireless networks and the design of future MAC and routing protocols.