At JET the error field correction coils can be used to generate an n = 1 or n = 2 magnetic perturbation field (Liang et al 2007 Plasma Phys. Control. Fusion49 B581). Various experiments at JET have already been carried out to investigate the mitigation of ELMs by resonant magnetic perturbations (RMPs) (Liang et al 2010 Nucl. Fusion50 025013, Liang et al 2011 Nucl. Fusion51 073001). However, the typical formation of a secondary strike point (strike point splitting) by RMPs observed in other machines (Jakubowski et al 2010 Contrib. Plasma Phys.50 701–7, Jakubowski et al 2004 Nucl. Fusion44 S1–11, Nardon et al 2011 J. Nucl. Mater.415 S914–7, Eich et al 2003 Phys. Rev. Lett.91 195003, Evans et al 2007 J. Nucl. Mater.363–365 570–4, Evans et al 2005 J. Phys.: Conf. Ser.7 174–90, Watkins et al 2009 J. Nucl. Mater.390–391 839–42) has never been observed at JET before. In this work we will present discharges where for the first time a strike point splitting by RMPs at JET has been observed. We will show that in these particular cases the strike point splitting matches the vacuum edge magnetic field topology. This is done by comparing heat and particle flux profiles on the outer divertor plate with the magnetic footprint pattern obtained by field line tracing. Further the evolution of the strike point splitting during the ramp up phase of the perturbation field and during a q95-scan is investigated, and it will be shown that the spontaneous appearance of the strike point splitting is only related to some geometrical effects of the toroidal asymmetric magnetic topology.