A series of recent studies are reviewed which demonstrate that behavior can be controlled by nicotine injections in different ways depending on the behavioral history of the subject and the schedule of reinforcement under which nicotine is administered. Lever-pressing responses by squirrel monkeys and beagle dogs were maintained well above saline-substitution levels by injections of 10 to 30 micrograms/kg of nicotine under fixed-ratio schedules of nicotine injection. Lever-pressing responses by squirrel monkeys also were well maintained by injections of 30 to 300 micrograms/kg of nicotine under a fixed-interval schedule of nicotine injection. The highest rates of responding were maintained by injections of 10 to 30 micrograms/kg of nicotine under second-order schedules in which responding by squirrel monkeys produced brief-light presentations which were only occasionally paired with nicotine injection. Under other conditions, however, response-produced injections of these same injection doses of nicotine (10 to 30 micrograms/kg) suppressed food-maintained fixed-ratio responding by squirrel monkeys during the punishment component of a multiple schedule. Finally, under a schedule of nicotine postponement, injection doses of 30 to 56 micrograms/kg of nicotine maintained responding that prevented, rather than produced, nicotine injections. These findings indicate that nicotine may control smoking behavior of humans in very complex and divergent ways depending on prevailing environmental conditions.