S U M M A R YA collection of 306 oriented samples from 43 Permian and Triassic intrusions from the Ukrainian Shield, the southwest portion of the East European Craton, has yielded valuable palaeomagnetic directions with new ages obtained using the 40 Ar/ 39 Ar method. Andesitic intrusions have Late Triassic ages (six dating samples) of 204.2 ± 1.6 Ma to 215.7 ± 2.0 Ma and dual-polarity Dec/Inc = 60.1 • /+64.4 • , k = 96, α 95 = 4.5 • , derived from N = 12 sites. Trachyte dykes have an early Artinskian (mid-Early Permian) age (one dating sample) of 282.6 ± 2.6 Ma, and yielded Dec/IncBecause the Artinskian and younger Permian palaeopoles obtained from the Gondwana continents are subject to uncertainties, some studies have adopted a palaeopole (at 41 • S, 61 • E) largely based on results from ∼280 Ma Permian quartzporphyries in the Southern Alps of Italy as proxy to position Gondwana at that time. With this palaeoposition and our new ∼280 Ma palaeolatitude for Baltica, a Pangea A reconstruction cannot be supported. To avoid its inherent continental overlap of Africa and Eurasia, a Pangea B reconstruction has been favoured, wherein Eurasia's southern margin faces the northcoast of South America instead. However, the 280 ± 10 Ma palaeomagnetic data from the Gondwana continents themselves give a mean palaeopole at 30 • S, 59 • E, which results in a palaeogeographic position of Gondwana that allows a Pangea A type reconstruction. Thus, the choice between Pangea A versus B in the Artinskian hinges on data selection and reliability criteria and the assumptions about tectonic coherence of northern Adria (i.e. the Southern Alps) with Africa in Permian times. A more reliable mid-Early Permian palaeopole from cratonic Gondwana would provide a more definitive conclusion.