On April 11, 1986, at about 0600 UT a long Pc 1 wave event of the hydromagnetic chorus type started on the ground, as registered by the Finnish pulsation magnetometer network. The main pulsation band at about 0.3 Hz was observed for several hours. Soon after start, this band smoothly extended to higher frequencies, forming another separate wave band which finally reached up to 0.5 Hz. During the event the Viking satellite was on its southbound pass over Scandinavia, close to the MLT sector of the ground network. From 0650 until 0657 UT, Viking observed a chain of Pc 1 bursts with increasing frequency. The strongest bursts could be grouped into two separate wave regions whose properties differed slightly. The higher-latitude region had a frequency of 0.3 Hz, well in agreement with the main Pc 1 band on the ground. The lower-latitude region contained the highest frequencies observed on the ground at about 0.5 Hz. The latitudinal extent of both wave regions was about 0.5¿. They had slightly different normalized frequencies, Alfv¿n velocities, and repetition periods. Most interestingly, the repetition periods of both wave sources were too short for the bursts to be due to a wave packet bouncing between the two hemispheres. The results give new information about the high-latitude Pc 1 waves, showing that they can consist of separate repetitive but nonbouncing bursts. We suggest that the long-held bouncing wave packet hypothesis is generally incorrect and discuss two alternative models where the burst structure is formed at the equatorial source region of the waves.¿ 1997 American Geophysical Union