We tested the effects of an intensive tact instruction procedure on numbers of tacts emitted in non-instructional settings (NIS) using a multiple probe design across 3 participants (3- and 4-year old boys with autism). The dependent variable was tacts emitted in NIS before/after the mastery of sets of 5 different stimuli. The non-instructional settings included the toy area of the classroom, lunchtime, and the school hallway during transition. All probe sessions were conducted daily for a cumulative 15 minutes, 5 minutes in each NIS. Intensive instruction involved increasing the tact instructions to 100-tact learn units above the daily learn units students were receiving daily. The intervention increased vocal verbal operants (tacts and mands) emitted by the target students in NIS. (Contains 5 tables and 4 figures.)