In this study, we conducted eight reciprocal peer tutoring classes where each student took either role of a tutor or a tutee to study covariational reasoning in ninth graders. Students were given the opportunity to teach their peers with their covariational reasoning as tutors, and at the same time to learn covariational reasoning as tutees. A heterogeneous group was formed so that scaffolding could be provided in the teaching and learning process. A total of eight reciprocal peer tutoring worksheets were collected: four quantitative graph type questions and four questions of the qualitative graph to the group. The results of the analysis are as follows. In reciprocal peer tutoring, students who experienced a higher level of covariational reasoning than their covariational reasoning level showed an improvement in covariational reasoning levels. In addition, students enhanced the completeness of reasoning by modifying or supplementing their own covariational reasoning. Minimal teacher intervention or high-level peer mediation seems to be needed for providing feedback on problem-solving results.