The motor cortex has long been regarded as a key node in the generation of motor output, based on observations of patients with motor seizures (Jackson 1889) as well as results from various kinds of studies employing lesions and electrical stimulation (Porter and Lemon, 1993). Evarts (1966, 1968, 1969) recorded the impulse activity of single cells in the motor cortex of behaving monkeys. He showed that cells typically changed activity before the first changes in muscle activity and that the frequency of discharge varied during movement or production of isometric force. These studies initiated research into the role of the motor cortex in the specification and control of motor output, a subject that has been hotly debated and widely speculated upon and that remains under active investigation. Simplistic notions of motor cortical function have been exemplified by the view that motor cortical cells are simply "upper motorneurons." This view is clearly untenable. Early observations from patients with motor seizures had already pointed to the complexity of motor cortical function (Jackson, 1889), as had the variable effects of electrical stimulation (Leyton and Sherrington, 1917). More recently, the technique of recording the activity of single cells during behavior (Lemon, 1984) provided a powerful tool by which to study the activity of motor cortical cells within the full-blown behavioral context of voluntary movement. This was an important advance and the only appropriate way to study the neural mechanisms of the initiation of voluntary movement. Indeed, a wealth of information as accumulated during the past 30 years concerning the involvement of the motor cortex in simple motor functions as well as in more complex aspects of motor behavior (see Georgopoulos, 1991, for a review).