The modeling of human movement is vital for a complete understanding of complex human–computer interaction. As three-dimensional collaborative tangible user interfaces (TUIs) evolve, research is needed to understand how people physically interact with each other within a virtual environment. Previous study of physical collaboration in virtual environments has utilized Fitts' law to model gross upper-extremity movement in a passing task. However, no study has modeled passing tasks that require precision grasp with the human hand, an important feature of human–computer interaction in TUIs. The purpose of this study was to evaluate the validity of Fitts' law in modeling movement time for a precision passing task in a 3D TUI, and to assess the coordination between passer and receiver using kinematic parameters. In this experiment, 12 participants (six male, mean age 22.6 years) performed a prehensile passing task within a desktop virtual environment. Results detail the kinematic events required to achieve the necessary temporal and spatial coordination specific to the passing task. Further, results indicate that Fitts' model does not adequately explain movement time for this task (R2 = .51). This finding challenges the external validity of previous results. We argue that the task-specific complexity of human neuromotor control should be considered when using predictive models in 3D TUI design.