This thesis investigated ways of teaching manual signs to institutionalized mentally retarded persons. A series of seven experiments are reported. In Experiment 1, the differential effects of positive practice overcorrection alone and combined with positive reinforcement were compared against a no-training control condition. While the overcorrection procedure was superior to the no-training control condition, the addition of positive reinforcement produced the best results. Thus, positive practice overcorrection plus positive reinforcement was used as the basic training procedure in the rest of the experiments, Experiments 2 to 7 investigated the effects of using flashcards, photographs of signs, photographs of objects and real objects as training stimuli on the learning and generalization of new sign words. While the addition of training cues did not produce detrimental learning effects for any of the subjects, some of the cues enhanced the learning of some subjects. Finally it was found that the use of real objects resulted in the learning and generalization of signs equal to but not better than flashcards or photographs of objects. These results were discussed in terms of teaching manual signs to mentally retarded persons.