Discrete trial training was delivered using English and Spanish languages to a student with autism from a Spanish-speaking family. An alternating treatments design was used to examine the effects of language of instruction on the child's response accuracy and challenging behavior. More correct responses and fewer challenging behaviors occurred when instruction was delivered in Spanish compared to English. Results suggest that the language of instruction may be an important variable even when a student initially presents with very little spoken language and comparable scores on English and Spanish standardized language assessments.