1. Monkeys were trained to trace sinusoids with their index fingers on a planar surface. During this task, both the direction and speed of movement varied continuously. Activity of individual units in the precentral gyrus contralateral to the moving arm was recorded as the task was performed. These cells responded to passive movement of the shoulder and/or elbow. The relation between discharge rate and movement direction for these individual cells could be described with a cosine tuning function. 2. Data recorded as the sinusoid was traced were divided into 100 bins as each cell was studied during the experiment. In each bin, the activity of a particular cell was represented by a vector. The vector ("cell vector") pointed in the direction of finger movement that corresponded to the highest rate of neuronal discharge. This direction, referred to as the preferred direction, corresponded to the peak of the cosine tuning function. The direction of the vector was constant between bins, but the magnitude of this cell's vector was a function of the instantaneous discharge rate. 3. This cell vector is a hypothetical contribution of a single cell to the population response comprised of 554 similarly derived vectors from different cells. The population response was represented as the vector that resulted from forming the sum of the vector contributions from the individual cells. A separate calculation was made for each bin, resulting in 100 population vectors for each sinusoid. 4. Within a given time series of population vectors, their lengths and directions varied in a consistent relation to the tangential velocity of the drawing movement.(ABSTRACT TRUNCATED AT 250 WORDS)