For a mobile robot to be able to communicate usefully with others, the symbols it uses to communicate must be grounded to entities in the environment, and those groundings made consistent among agents. While it is common practice to hand-construct such groundings, this does not scale to large problems. In particular, when communicating about useful spatial references, there are a large number of potentially relevant groundings, even for a basic task such as navigation. This paper describes the development and evaluation of an approach that allows a group of robotic agents to develop consistent shared groundings for locations in an environment over time. This approach is based on local communication and interaction, and does not rely on the ability to broadcast references to all agents, and so is suitable for domains in which communication may be sporadic, such as robotic rescue. The evaluation of this approach, which compares several different grounding techniques, shows that shared groundings can be developed effectively over time, and that these improve the effectiveness of communication in a multi-robot setting.