The exploration of sand-filled arenas by workers of an entire colony of the Australian, subterranean foraging, tree-nesting termite, Coptotermes frenchi Hill was investigated under laboratory conditions. The first experiment tested whether termite exploration of sand was influenced by the presence of gaps or objects in the sand. Gaps and objects were chosen to represent soil heterogeneity in the urban environment: gaps to represent tunnels dug by other animals, perspex strips to represent cables and pipes, and wood strips to represent roots. Termites always chose to explore gaps thoroughly before they began tunnelling in the sand. Significantly more and longer tunnels were excavated from the end of gaps at the far end of the arenas, and relatively little tunnelling occurred around and along objects. Termite density was significantly greater around and along wood compared with perspex blocks. The second experiment tested whether termite exploratory tunnelling was influenced by soil moisture. The termites tunnelled slowly in dry sand, but after discovering a patch of wet sand, increased tunnelling five-fold until it was completely explored, after which activity declined. Energy and water conservation may be behind these patterns of exploratory tunnelling as well as those seen in large field studies, but caution is urged when interpreting small scale laboratory experiments to explain large scale field data.