We conducted a three-phase study to evaluate the use of mands and the occurrence of problem behavior for 2 children with developmental disabilities. During Phase 1, a functional analysis identified the variables maintaining problem behavior. During Phase 2, functional communication training was implemented within a concurrent schedules design. The children were reinforced for using either a novel mand (communication card) or other existing mands (vocal speech, manual signs) that were not specifically trained but were observed to be part of the children's existing repertoire. We then conducted an assessment of mands and problem behavior across different stimulus conditions (card absent, card present) within an ABAB design (Phase 3). Results showed that during Phase 2, problem behavior decreased and participants used the card more frequently than they used other existing mands. Phase 3 showed that problem behavior remained low across both stimulus conditions. When the card was absent, the children used other existing mands; when the card was present, they primarily used the card. These results suggested that the presence of a communication card may function as a discriminative stimulus for a specific topography of manding, but that training with the card did not inhibit the use of other mands when the card was absent. (Contains 2 figures.)