Résumé
Carob (Cetatonia siliqua L.) is a well-known Mediterranean tree whose domestication was coetaneous to that of the first trees (eg. olive, fig) probably in the Middle-East. Since then, carob has played a crucial role in Mediterranean societies due to its edible fruits used for cattle forage and subsistence agriculture. Carob plants are able to grow on stressful rocky soils increasing the socio-economic value of many Mediterranean dry lands. Despite its economic importance, the origins and history of carob populations are still under debate. Since De Candolle, the wildness of carob populations has been questioned on the basis of paleo-botanical, archeological and philological evidences and , unfortunately, the scarce studies on the ecology and genetic diversity of wild populations have not thrown much light on this topic. The extremely low cold-stress tolerance of carob plants constituted the main argument against a long-term persistence of natural populations throughout Pleistocene glaciations in the Mediterranean. Under this scenario, the current distribution would be explained by human dissemination. However, a global phylogeographic study covering the entire distribution of carob is still lacking. In this context, we aimed at exploring the two main hypotheses about the origin of carob populations: their possible persistence in unknown refugia during the Pleistocene or their putative naturalization after human dissemination throughout the Mediterranean from a single origin. We used Environment Niche Modeling (ENM) under present and past climatic conditions (MidHolocene – aprox. 6 kya; Last Glacial Maximum, LGM – aprox. 22 kya; and Last Interglacial Maximum, LIG – aprox. 120-140 kya) to investigate the potential range changes that carob could have undergone driven by climatic oscillations. Additionally, we sequenced three plastid regions from both natural and cultivated populations covering the whole current distribution of carob to explore its phylogeography based on coalescent methods. Our results point towards two distant and separated two distinct and separated phylogroups atat southern boundaries of carob range during LIG, which left a strong genetic footprint within carob natural populations. This was followed by a reduction of the potential distribution area during LGM, which subsequently expanded during MidHolocene up to the current Mediterranean known distribution of carob. The current potential distribution modeled for carob is extremely restrictive to the coastal areas of the Mediterranean, and it actual distribution is probably linked to strong selection pressures at the margins of its range. Climatic variability and long term persistence within the carob range suggest that evolutionary legacy could be a pivotal factor to take into account in both breeding and reforestation programs.