We present an analysis of high signal-to-noise ratio, 430-MHz total intensity single pulse observations from PSR 2303+30 recorded at the Arecibo Observatory in August 1986. We believe that the regular subpulse modulation observed in this pulsar can be explained only in terras of subpulse-associated emission beams rotating around the axis of the pulsar beam. This idea was first introduced by Ruderman and Sutherland (1975) in their spark model. We use the quantitative prediction of this model to calculate the pulsar magnetic field. For this reason we need to estimate the rotation period of a spark around the magnetic pole. We estimate this period by reproducing the observed pulse sequences within an angular beaming model of subpulse emission and fitting this model to the data.