Vision naturally occurs in the context of voluntary information gathering movements involving the eyes, head, and hand. However, much work in vision is dominated by trying to understand the events occurring within a single view of a scene, and we have only limited understanding of the consequences of eye and head movements for vision and visuo-motor coordination. The technology to look at performance in more natural circumstances now exists, and I am currently developing a human sensory-motor lab, in collaboration with Dana Ballard in Computer Science, for measuring unconstrained eye, head, and hand movements in the performance of natural tasks, and for developing a virtual reality display to allow controlled but visually complex stimulation. We also have the capability of providing force feedback for two finger grasping. The new instrumentation allows a large range of experiments not previously possible. My objective is to understand the demands placed on vision and motor systems by natural behavior and the nature of the representations that are required for visually guided tasks.