Osteoclasts are commonly defined as terminally differentiated polykaryons, with a lifespan of days to weeks, during which time they resorb bone, before undergoing apoptosis. This linear model of osteoclast fate has been formulated primarily through dynamic examination of osteoclasts in vitro and static histological analyses in vivo. We have developed a novel intravital two-photon imaging methodology which allows us to directly image the intact endocortical surface of the tibia in live mice. This has allowed the visualisation of the dynamic behaviour of osteoclasts in vivo, revealing previously unappreciated cellular plasticity and an alternative cell fate. We showed that, in the steady-state, multi-nucleated LysM⁺Blimp-1⁺Osteosense⁺, Cathepsin K⁺, osteoclasts exhibit a stellate structure forming syncytial networks on the endosteal bone surface. Activation of bone resorption with exogenous soluble RANKL (sRANKL) induced rapid retraction of cellular processes, causing loss of cell-to-cell contact and syncytial networks. Subsequently, these activated osteoclasts were visualised undergoing cell fusion and, unexpectedly, cell fission. Daughter cells were observed to refuse with parent cells or other nearby osteoclasts, in a novel process we have termed osteoclast recycling. This was blocked by treatment with the RANKL inhibitor osteoprotegerin-Fc fusion protein (OPG-Fc). However, rather than undergo apoptosis, small round LysM⁺Blimp-1⁺ cells persisted after OPG-Fc treatment. Critically, 3-4 weeks following OPG-Fc treatment withdrawal recycling osteoclasts re-fused to form networks of active enlarged LysM⁺Blimp-1⁺Osteosense⁺ osteoclasts, similar to those seen following sRANKL treatment. These data demonstrate that intravital imaging of the intact endosteal bone surface in the tibia can be used to study osteoclast dynamics in vivo and that rather than a simple linear cell fate, osteoclasts can also recycle their cellular constituents. Intravital imaging provides an opportunity to examine bone cell dynamics in vivo, advancing our understanding of bone physiology under normal and diseased states, as well as determining the mechanisms underlying therapeutic responses.