1. Nesting behaviour and interactions between the bee Chelostoma florisomne (L.) (Megachilidae) and its nest parasite Sapyga clavicornis (L.) (Sapygidae) were studied through continual observations of individuals and dissections of bee nests. Protection of bee offspring is based on (1) the bee’s discovery and removal of parasite eggs deposited prior to the construction of a cell closure, (2) minimising the time when fully provisioned cells might be parasitised successfully, and (3) the construction of empty cells in front of brood cells.

2. An empty cell was found in front of 64.4% of all brood cells and, if the outermost brood cell in a nest was excluded, in front of 74.3% of inner brood cells. A vestibule closure is most often constructed in front of the outermost brood cell.

3. Following oviposition, the bee made only five flights, which together lasted 6–13 min, to construct a cell closure. A cell closure does not prevent the nest parasite from oviposition inside the brood cell, however, and parasite eggs deposited through the cell closure are not detected and removed by the bee. Only an additional cell closure, i.e. the formation of an empty cell, may protect a brood cell when the bee is not in the nest. The nest parasite often oviposited through the additional cell closure but its offspring were then trapped in the empty cell and starved to death.

4. Only 5.4% of the inner brood cells that were protected by an empty cell were parasitised, compared with 28.9% of those without an anterior, empty cell; 27.4% of the empty cells contained dead parasite offspring (eggs and larvae). Thus, the empty cells provided significant protection and, combined with additional means of protection of brood cells, led to a low degree of parasitism. More than 77% of the wasp offspring died at an early stage due to intraspecific interference competition within brood cells and as result of the wasps’ oviposition into empty cells.