In this study, the effects of argumentation-eliciting interventions on conceptual understanding in evolution were investigated. Two experiments were conducted: In the 1st, 76 undergraduates were randomly assigned to dyads to collaboratively solve and answer items on evolution; half of them were instructed to conduct an argumentative discussion, whereas control dyads were only asked to collaborate. In the 2nd experiment, 42 singletons participated in 1 of 2 conditions: Experimental students engaged in monological argumentation on their own solution and a confederate's solution in response to prompts read by the confederate, whereas in the control condition they merely shared their solutions. Conceptual gains were assessed on immediate and delayed posttests. In both experiments, students in the argumentative conditions showed larger learning gains on the delayed posttest than control students. Students in argumentative conditions were able to preserve gains that were obtained immediately following the intervention, whereas control participants either lost immediate gains (dialogical condition) or did not improve their conceptual understanding at any time (monological condition).