The complex history of the olive tree: from Late Quaternary diversification of Mediterranean lineages to primary domestication in the northern Levant

Abstract

The location and timing of domestication of the olive tree, a key crop in Early Mediterranean societies, remain hotly debated. Here, we unravel the history of wild olives (oleasters), and then infer the primary origins of the domesticated olive. Phylogeography and Bayesian molecular dating analyses based on plastid genome profiling of 1263 oleasters and 534 cultivated genotypes reveal three main lineages of pre-Quaternary origin. Regional hotspots of plastid diversity, species distribution modelling and macrofossils support the existence of three long-term refugia; namely the Near East (including Cyprus), the Aegean area and the Strait of Gibraltar. These ancestral wild gene pools have provided the essential foundations for cultivated olive breeding. Comparison of the geographical pattern of plastid diversity between wild and cultivated olives indicates the cradle of first domestication in the northern Levant followed by dispersals across the Mediterranean basin in parallel with the expansion of civilizations and human exchanges in this part of the world.

Figure 1.

Diversity of the three Mediterranean olive plastid lineages [10]. (a) The reduced median haplotype networks [17] for each lineage and for both wild and cultivated gene pools. Each haplotype is numbered and represented by a symbol with a definite colour and/or motif. Haplotype frequencies are proportional to symbol diameter. The missing, intermediate nodes are indicated by small black points. Small red squares on the networks indicate mutational steps at the homoplasious CAPS-XapI locus, which was excluded for the network analysis [16]. The frequency of each lineage in oleasters and cultivars is indicated in brackets. (b) The geographical distribution of haplotypes in oleaster populations. The size of pie charts is relative to the number of trees analysed per location. (Online version in colour.)

Figure 2.

Distribution modelling of suitable habitats for oleasters based on temperature variables. (a) Distribution model fitted to current climatic conditions. (b) Projection to a period of the LGM (ca 21 000 BP). Results under the Community Climate System Model (CCSM) [28] and Model for Interdisciplinary Research on Climate (MIROC) [29] general circulation model simulations are averaged (for separate CCSM and MIROC results, see electronic supplementary material, figure S6). The locations of available olive fossil/subfossil remains (macrofossils and pollen) during the LGM are indicated (see the electronic supplementary material, table S8). Note that palynological data show low congruence with our model predictions, but this is probably due to efficient pollen dispersal by natural agents over long distances [3]. (c) Projection to the LIG (ca 120 000–140 000 BP) [30]. (d) Areas inferred as continuously suitable under LIG, LGM and current climatic conditions (putative ‘long-term refugia’). In all cases, the average model of ten replicates is shown. (Online version in colour.)