Two experiments were conducted to characterize the system of beliefs that make up the naive impetus theory of motion and to determine what effects physics instruction has on students' conceptions of motion. Thirteen college students were asked to solve several quantitative problems and were interviewed about their answers in the first experiment. Results indicate that these subjects held an impetus conception of motion. Three of these impetus-related misconceptions are discussed (curvilinear impetus belief, resistance to gravity, and straight-down belief). In the second experiment, high school students' (N=60) concepts of motion were examined before and after they took a physics course. Results suggest that instruction often corrects some of the gross misconceptions about how moving objects behave, but may leave others unaltered. In addition, instruction frequently fails to alter students' impetus beliefs about why moving objects behave as they do. These findings suggest the need to discuss the impetus view of motion, explaining precisely how this view differs from the Newtonian view. In addition, resistance to gravity and straight-down beliefs should be taught. Physics courses should also emphasize student understanding of the fundamentals of qualitative principles of mechanics, even if this means teaching less about the quantitative formalizations of these principles. (Author/JN)