The submeridional Dahar chain in southeastern Tunisia is over 200 km long. It is separated from the Mediterranean Sea by the Jeffara plain with some tens of kilometers in width. This landscape continues to the South into Libya, but to the North, the chain ends with the Matmata mountains which form a plateau slightly inclined to the west and some 10 km wide. The eastern scarp shows a mainly calcareous geological stratigraphy from Upper Permian to the Senonian. The Dahar-Matmata structure belongs to the Sahara platform and shows a hiatus during the whole Tertiary, since it was emerged since Upper Cretaceous. The Tunisian Atlas nearby shows a completely different paleogeographic evolution, with a complete Tertiary series and a later Plio-Quaternary structuration. These two paleogeographic domains of Southern Tunisia, the Sahara Atlas and the NE border of the Sahara platform, were influenced by the Messinian crisis (5.9 Ma to 5.3 Ma). This was expressed by the collapse of the Mediterranean Sea level, profoundly modifying the fluvial dynamics with an inversion of the erosional system, from normal erosion to regressive erosion. It results a deepening of canyons in the downstream part and a deepening of the watercourses in the upstream part. The geological structures in the Messinian have been deeply affected by these large eustatic changes, with an incision of cluses in the Atlas and the deposition of a thick clayeysandy series that we could recently link to deltaic systems and Gilbert deltas. The re-establishment of seaways between the Atlantic and the Mediterranean, and the subsequent infill in the Lower Pliocene (Zanclean transgression), with an important inpact in Southern Tunisia, had multiple consequences in that region. The newly adjusted sealevel, together with a more humid climate that was confirmed by faunal and floral extension oof tropical plants in Northern Africa, stimulated an important karstification of the limestone areas. In the Dahar chain, caves, dolines, karstic depressions or karstic dry valleys emerged, the most spectacular ones being found in the Matmata Mountains. The karstic depressions are the forms that represent best this Pliocene karstification that surely was interrupted in an early stage, because localized endokarstic forms had not enough time to develop. So the karstification seems to have been active in Matmata from 5.4 to 4.0 million years, i.e. two times as long than the duration of the Messinian crisis. The interruption of karstification is due to an increase in temperature and dryness, which even gets more intense during the Pliocene, pulverizing the soils. Already at the beginning of the desertification, a calcareous crust forms by rapid cristallization of dirt. It is immediately transported from the karstic zones to the Jeffara plain. This transfer fo dissolved calcite was the origin of the resistant calcitic crust well known in the Jeffara plain. We now identified the same crust in a karstic depression in the Matmata Mountains, opening the way to new geomorphologic and tectonic interpretations, and a review of the eolian silts formerly attributed to the Upper Pleistocene. Later, during Upper Pliocene-Gelasian, we observe a general tectonic uplift of the Dahar chain and the Matmata Mountains as well as the subsidence of the Jeffara plain at the Medenine fault (NW-SE), prolonging the large Gafsa fault towards the East. The karstic paleoforms were thus uplifted more than 500 m, but nevertheless remain open on the Jeffara plain, as seen by large depressions. As a consequence, the karstic depressions of Matmata played the role of traps for eolian silts blown from the Jeffara plain during the extreme desertification in the Upper Pliocene-Gelasian. The morphological reconstruction since the Messinian shows a succession of important events during the Pliocene that profoundly influenced the Quaternary. All indications permit to reject the hypothesis that the Matmata silts came from the West (Eastern Erg).