This investigation compares the results of three different methods of teaching tumbling, the Trampoline Method, Mental Practice, and the Traditional Method. The study also investigates whether sex and ability level significantly affect the results of the teaching methods. The subjects were the ninety members of the junior-level gymnastics classes for physical education majors at New Mexico State University during the fall and spring semesters of the 1972-1973 school year. There were forty-five female and fortyfive male subjects. A stratified random sample was constructed to insure equality of the teaching-method groups. The motor educability scores of the Johnson-Metheny Test were used to develop high-, medium-, and low-ability groups. The subjects in the ability groups were randomly selected and assigned to one of the teaching methods.