During the last three decades, the robustness of experimental designs has been investigated under several criteria. The present paper surveys various results on robustness of block designs in three different directions, when some observations are missing. The criteria on robustness discussed here are taken up from a combinatorial point of view.