Adolescents with disabilities have great difficulty with academic content in middle school, and their teachers have difficulty teaching them to understand and use academic language. We taught teachers of 6th grade students with Learning Disabilities, over half of whom were also English Language Learners (ELL), to implement about 15 minutes of daily interactive vocabulary instruction in their intact special education English/Language Arts classes. Three schools were assigned randomly to treatment (2 schools) or control conditions (1 school; 52 students total). We developed instructional routines to introduce four new words per week in three 4-week units to test for replicability. ANCOVAs (with each cycle's pretest and intelligence quotient (IQ) as covariates) were conducted on taught vocabulary, all of which favored the treatment condition with effect sizes ranging 0.6 to 0.7 per cycle. Near-transfer effects to vocabulary usage were weaker, with significant effects in the last two cycles. Effects were similar for students with disabilities who were ELL and native English speakers. Treated students maintained their knowledge of words 4 to 24 weeks following the close of treatment. [This article will be published in "Learning Disability Quarterly."]