Abstract The Paraná Basin covers 1,600,000 km2 including parts of southern Brazil, Paraguay, Uruguay and Argentina. It contains rocks ranging from the Ordovician to Tertiary in age. The present study is focused on the Itacurubí Group, of Llandovery age, from Paraguay. The sedimentology and palynology have been studied in three boreholes from the central part of Paraguay. The Itacurubí Group consists of the Euzébio Ayala, Vargas Peña and Cariy Formations. It corresponds to a complete transgressive – regressive cycle with maximum flooding in the Vargas Peña Formation. The sediments of the group were deposited in a glaciomarine regime and are included in the second order sequence. The palynofacies are dominated by acritarchs and chitinozoans, with miospores rare. Among the latter, cryptospores are most abundant and trilete spores very rare. The miospore assemblages are typical of the Llandovery. The presence of Laevolancis divellomedia and the successive first occurrence of the trilete spore genus Ambitisporites, followed by Archaeozonotriletes, allow the recognition of three biozones: divellomedia I, divellomedia II and chulusnanus. Correlations between the three boreholes based on miospore biostratigraphy and sequence stratigraphy techniques are similar. The miospore assemblages have important palaeogeographic implications: Ordovician / Silurian miospore assemblages, identified in South America are similar to those described in palaeogeographically distinct regions such as China, UK, Belgium, USA etc. This suggests that the phytogeographical differentiation proposed by Gray et al. [1992], with a Malvinokaffric Realm characterised by smooth tetrads and an extra-Malvinokaffric Realm characterised by ornamented tetrads, has to be reconsidered. Clearly, the same miospores are recovered from both regions, no matter if they are close to the palaeo-equator or the palaeo-pole, as is the case in Paraguay. These findings suggest that the same vegetation could survive under various climates. In theory, large oceans could be impassable geographic barriers for land plant miospores larger than 25 μm in diameter. As similar miospore assemblages occur, around the Ordovician – Silurian boundary, on the Gondwana, Avalonia and Laurentia plates, it seems likely that these palaeo-continental plates were in close proximity permitting the expansion of the vegetation through these continents. Therefore, palaeogeographic reconstructions with narrow oceans between the continents better explain early miospore biogeographies in early Silurian times. Maps produced by Dalziel et al. [1994] better explain the diachronism of trilete spore first appearances: Hirnantian in Turkey, Rhuddanian in Saudi Arabia, early Aeronian in Libya, late Aeronian in Paraguay, latest Aeronian in UK, and possibly Telychian in USA. The Baltica plate could possibly have been isolated by a geographic barrier during the Ashgill and the Llandovery, as only simple naked tetrads are known from that plate. The first appearance of the trilete spores in Götland during the early Wenlock could correspond to the end of this geographic isolation. This apparent isolation of the Baltica plate could be due to a lack of data.