Previous research in NVE mainly based their design on assumptions on how avatars behave, and use randomly generated movements to evaluate their design. We believe that there is a gap between these assumptions/randomly generated movements and how real avatar behave. To bridge the gap, we collected mobility traces of 84,208 avatars spanning 22 regions over two months in Second Life, a popular networked virtual environment. We analyzed the traces to characterize the dynamics of the avatars mobility and behavior, both temporally and spatially. We discuss the implications of the our findings to the design of peer-to-peer networked virtual environments, interest management, mobility modeling of avatars, server load balancing and zone partitioning, client-side caching, and prefetching.