Abstract of "Internal Models of Virtual Objects". Human subjects learn to control a computer-generated virtual object. A new experimental paradigm  is described to quantify the information that subjects acquire about the dynamics of the virtual object. This is the subjects' internal model of the object. An experiment is presented to identify the strategy that subjects use to interact with the object. An equation is proposed to describe the subjects’ asymptotic behavior. This equation is used to estimate the subjects’ internal model. Finally, the results of a transfer experiment are presented to demonstrate that the subjects’ internal model contains information about the dynamics of the virtual object.