A functional analysis for a boy with Down syndrome and autism suggested that vocal stereotypy was maintained by automatic reinforcement. The analysis also showed that instructions and noncontingent attention suppressed vocal stereotypy. A treatment package consisting of noncontingent attention, contingent demands, and response cost effectively reduced vocal stereotypy. The treatment package remained effective even when noncontingent attention was removed, making the procedure easier to implement. Also, the presence of the therapist in the room with the participant was faded systematically. After completion of fading, vocal stereotypy remained low during conditions similar to the no-consequence phase of the functional analysis. (Contains 1 figure.)