Using pre- and post-intervention non-concurrent multiple probe designs across participants, I conducted 2 experiments that tested the effects of imitation instruction using a mirror on the emergence of both basic and advanced forms of generalized imitation (GI) involving physical actions with preschool students diagnosed with developmental delays. Experiment I was first conducted to test the effectiveness and efficiency of using a mirror to induce GI with 4 young developmentally delayed preschoolers. The dependent variable was the number of untaught, unconsequated basic physical imitative responses (e.g., nodding head) emitted by participants prior to and following intervention conditions while directly facing the experimenter. The independent variable involved teaching basic imitative actions (e.g., clapping hands) while looking at an adult perform the actions in a mirror until a mastery criterion was met. The results of Experiment I showed that a functional relationship was demonstrated between mastering the imitation of basic physical actions using a mirror and the emergence of GI for all 4 participants. Experiment II was then conducted to further examine the effectiveness and efficiency of using a mirror to induce more advanced forms of GI with older preschoolers diagnosed with developmental delays as well as examine the effects of the absence and presence of advanced GI on the rate of participant learning after an adult provided a how-to demonstration. The primary dependent variable was the number of instructional sessions required for each participant to meet a mastery criterion on a novel dressing skill before and after the participants acquired GI. The second dependent variable was participant responding during block structure duplication and transcription tasks before and after the participants acquired all advanced forms of GI at mastery level. The independent variable was the induction and/or presence of advanced, mastery level GI using mirror imitation instruction across bodily actions (e.g., gross motor, fine motor, oral motor, and multiple-step motor actions). The results of Experiment II showed that the presence of mastery level GI accelerated learning for all 4 participants, which may have facilitated the participants' capability of learning through demonstration.