Phylogeographical studies use molecular markers to trace the historical expansions of species that have led to their current distribution range. The vast majority of these surveys have focused on population dynamics that took place during the Quaternary (spanning the past two million years) and especially after the last glacial maximum (18 000 years bp). However, examples are now accumulating of species that have retained the genetic imprints of much more ancient dynamics (e.g. Dick et al. 2003; Petit et al. 2005; Grivet et al. 2006). A tree species studied by Magri and colleagues in this issue beats all records of stability and population genetic stasis.
The authors investigated the phylogeography of cork oak (Quercus suber L., Fig. 1), an emblematic Mediterranean tree exploited since the Antiquity for its bark. They found that its range-wide population structure is consistent with the break-up and separation of several microplates that took place in the Oligocene and Miocene, between 25 million years and 15 million years ago. The current distribution of cork oak spans both sides of the western Mediterranean Basin, including all the major islands. Magri and coworkers collected more than 100 populations throughout the species range to investigate geographical patterns of chloroplast DNA (cpDNA) polymorphism. Using an extensive set of 14 cpDNA microsatellites, they detected just five haplotypes, but with extremely clear-cut distributions. Two haplotypes characterize all stands from Morocco, the Iberian Peninsula, the Balearic Islands and southwestern France, while another two haplotypes occupy the Italian Peninsula. The remaining haplotype shows a striking distribution: it occurs in a narrow belt ranging from Tunisia and Algeria via Sardinia and Corsica to Southeast France. The geographical overlap between the three groups of haplotypes is minimal and population fixation is almost complete (GST = 0.97). Curiously, the authors found that the two Italian haplotypes and one of the two Iberian-Moroccan haplotypes originate from Quercus cerris L. and Quercus ilex L. through introgression.
These results by themselves were not completely novel. Lumaret et al. (2005) had already described very similar patterns of geographical variation, including traces of introgression, based on restriction fragment length polymorphism (RFLP) variation over the whole chloroplast DNA molecule. The innovation of Magri et al. (2007) is their new interpretation of the observed haplotype distributions based on the geological history of the region. Rejecting the hypothesis of repeated long-distance colonization events, the authors show that the genetic patterns reflect the geological history of the western Mediterranean Basin. A number of tectonic microplates connected in the early Tertiary broke up in the Oligocene and drifted apart, eventually integrating the Algerian, Spanish, French, and Italian coasts or forming the Balearic Islands, Sardinia and Corsica. Magri and coworkers show that the current geographical distribution of haplotypes can largely be explained by these early land bridges, along with subsequent range expansion across Iberia and Morocco and along the Italian Peninsula. In addition, cork oak populations have introgressed cpDNA haplotypes from two other oak species (Q. cerris in Italy and Q. ilex in parts of Iberia and Morocco), probably at an early stage of this expansion. Interestingly, cork oak is not the only western Mediterranean tree species that has preserved traces of these Tertiary events: a similar haplotype distribution — although with somewhat lower resolution — has been described in a study of mitochondrial DNA in the maritime pine (Pinus pinaster Ait.) (Burban & Petit 2003).
These studies and a previous one on European beech (Magri et al. 2006) highlight the amazing capacity of some tree species to conserve traces of ancient population dynamics. Several factors could explain the low rates of evolution of cork oak. First, it is a long-lived species. Generation time has been shown to reduce rates of evolution, in plants just as in animals (Kay et al. 2006; Petit & Hampe 2006). Second, previous studies based on allozyme markers indicate that the species is relatively uniform, with no indication for incipient speciation (Toumi & Lumaret 1998). The only oak species suggested to derive from cork oak (Quercus afares Pomel, in North Africa) appeared by hybridization with another oak species (Mir et al. 2006), not through allopatric divergence. Third, geographical persistence of tree populations could have resulted in molecular stasis. In the Mediterranean Basin, the rugged topography and the reduced impact of climatic oscillations, compared to higher latitudes, should have favoured long-term population persistence (Hampe & Petit 2005), thereby preserving ancestral molecular variants. In fact, examples of the great antiquity of lineages in this region are starting to accumulate (Petit et al. 2005). The same effect might be expected in other regions with similar climate and topography, such as California (Grivet et al. 2006) or in the tropics (Dick et al. 2003). Further analyses of the effects of demography on rates of evolution may help explain why some species persist for tens of millions of years while many others last much less than a million years (Levin & Wilson 1976).