A single subject alternating treatment design was used to compare the effectiveness of embedded and small-group instruction to teach vocabulary word definitions to four middle school students with developmental disabilities. Embedded instruction was implemented in the students' general education classes. Students were taught to verbally define five vocabulary words drawn from the general curriculum. Instructional trials were distributed within and across the ongoing activities of the general education class. Instructional procedures included constant time-delay, differential reinforcement, and systematic error correction procedures. Small-group instruction was implemented in the students' self-contained special education class. Students were taught to verbally define five vocabulary words drawn from the curriculum in their general education classes. Instructional procedures for small-group instruction were identical to those used during embedded instruction. The small-groups included the target student and two peers who were randomly selected from his or her special education class. Small-group instruction employed an intrasequential format with spaced-trials (Collins et al., 1991; Reid & Favell, 1984). Results suggest that embedded and small-group instruction were equally effective in promoting the acquisition and generalization of the target skill. Results are discussed in terms of future research and implications for supporting the inclusion of students with developmental disabilities in general education classes. (Contains 1 table and 4 figures.)