We study an event of structured electromagnetic ion cyclotron (EMIC) waves observed by the Polar satellite in good conjunction with the Finnish ground stations on April 25, 1997. Polar observed two EMIC wave bands around the plasmapause. He+ band waves consisted of repetitive bursts which were observed on ground as a classical Pc1 pearl band. H+ band waves occurred over a large latitude range of more than 5¿ invariant latitude and were observed on ground as a broad, diffuse Pc1 pearl band with several subbands. We found the same repetition period of ~100 s for the Polar He+ band bursts and ground Pc1 pearls, in conflict with the bouncing wave packet (BWP) model. Comparing the burst structure of He+ band waves in Polar and on ground, we found a transit time of ~45 s and an average group velocity of ~500 km s-1. Within the BWP model this velocity would lead to a pearl repetition period of more than 250 s, in dramatic contradiction with the observed repetition period. Moreover, the bursts of the two Polar bands were roughly simultaneous with no significant dispersion, contrary to the expectation of the BWP model. These results clearly reject the classical BWP model, i.e., that Pc1 pearls are generated by one wave packet bouncing from one hemisphere to another. Instead, we find that EMIC waves were accompanied by long-period ULF waves which had a period very close to the repetition period of the simultaneous EMIC bursts. Interestingly, plasma density showed simultaneous fluctuations with roughly the same period. As an alternative to the BWP model, we discuss models where the EMIC wave packet structure and ULF waves are connected. We note that the suggested relation of EMIC wave packets and ULF waves offers a new explanation to the well-known preference of Pc1 pearls for the plasmapause. We have also estimated the full Poynting flux of EMIC waves for the first time, using the three components of electric and magnetic fields. The magnitude of the total Poynting flux of the He+ band waves was ~20--25 &mgr;W m-2 and strongly directed downward away from the equator. ¿ 2001 American Geophysical Union