The Mediterranean region – a hotspot for plant biogeographic research

Plant evolution in Mediterranean climate zones – IX^th Meeting of the International Organization of Plant Biosystematics (IOPB), Valencia, Spain, May 2004

The Mediterranean region has been recognized as one of the 18 world hotspots where exceptional levels of biodiversity occur (cf. Blondel & Aronson, 1999). Representing only 1.6% of Earth's dry land, this region harbours more than 25 000 known vascular plant species, or about 10% of the world total – and more species continue to be described (e.g. Lihováet al., 2004). In fact, more than half the plant species are endemic, and 80% of all European plant endemics are Mediterranean. Such high levels of plant diversity and endemism are generally thought to result from four major factors: biogeography, paleogeological/-climatic history, ecogeographical heterogeneity, and human influence (Blondel & Aronson, 1999). Accordingly, a comprehensive understanding of plant differentiation and speciation in the Mediterranean – though still elusive and lacking a book-length treatment – requires consideration of patterns and processes at a variety of spatio-temporal scales and taxonomic levels, thus fostering discussions among various subdisciplines within plant evolutionary research.

The IX^th IOPB Meeting (May 2004, http://www.jardibotanic.org/iopb.html) did just that, and provided broad-based and up-to-date views of the subject, including: discussions of phylogenetic origins, character evolution, and times of diversification of Mediterranean plant lineages; their response to the climatic vicissitudes of the Pleistocene as evidenced by molecular phylogeographic reconstructions; and detailed discussions about the roles of ecology (Debussche et al., 2004), hybridisation/reticulation (Gutiérrez Larena et al., 2002), and reproductive isolation (Pérez et al., 2003) in the origin of Mediterranean plant species. Genetic diversity, gene dispersal, and metapopulation dynamics of rare and endemic species were also covered (Hardy et al., 2004), with their direct relevance to conservation biology and resource management. While the focus of the symposium was largely on the Mediterranean region (including Macaronesia, and with many studies restricted to the western and central portions of the Basin), all of the topics addressed hold promising lessons for other parts of the world, especially where Mediterranean-type climates (with a predominantly winter rainfall regime) exist, namely parts of California, the Cape Province of South Africa, central Chile, and two disjunct regions in southern and south-western Australia. Here I will focus on some emerging biogeographic themes from this symposium with an emphasis on Macaronesia and the Mediterranean Basin.

Macaronesia

The Macaronesian Islands comprise the Azores, Madeira, Canary Islands, Selvagems, and the Cape Verdes. Inferring the origin of the flora of these Atlantic archipelagos has been the focus of many molecular phylogenetic studies during recent years (for review see Vargas, in press). From a modern biogeographic perspective, there is hardly any other subregion within the Mediterranean realm better explored. At least in part, this reflects the archipelagos’ long history of taxonomic research and the relative ease of access to comprehensive sources of plant material, both factors of which form the foundation of molecular phylogenetic studies. Based on those, there is compelling evidence now that: (i) the majority of species endemic to Macaronesia are sister to species/clades that have a predominantly western or, more rarely, eastern Mediterranean distribution (e.g. Convolvulus; Mark A. Carine, NHM, London, UK); (ii) many of the taxa studied occupy derived rather than basal positions in phylogenetic trees; and (iii) woodiness is often a novelty in these plants, derived from herbaceous ancestors (Javier Francisco-Ortega, Florida International University, Miami, USA).

Despite the large number of molecular phylogenies available that include Macaronesian taxa, participants in the symposium emphasized that further research is needed, especially with respect to a broader taxon (and gene) sampling on the continent, and the possibility of incorporating new and advancing approaches to molecular dating analysis (e.g. Sanderson, 2002). Still, the picture appears incomplete since I am not aware of studies that have included taxa most relevant to the long-standing ‘Tertiary relict hypothesis’ (Vargas, in press). Arriving at an answer will require the production of good species-level phylogenies (and calibrated molecular clocks) of those representatives of the Macaronesian laurel-forest vegetation whose closest living relatives today are restricted to (neo)tropical regions, but which are well-documented in the European Late-Tertiary fossil record, such as members of the laurel family (Laurus, Apollonias, Ocotea, Persea) and the olive family (Picconia).

The Mediterranean Basin

Compared with Macaronesia, far less attention has been devoted to the origin of the complex admixture of plant taxa found in the Mediterranean Basin itself, and whose biogeographic origins, respective age, and evolutionary histories are thought to vary enormously. Located at the crossroads between Europe, Asia, and Africa, this region has been considered as a huge ‘tension zone’ for plant lineages of various biogeographic origins (i.e. northern/central Eurasia, south-west/central Asia, Saharo-Arabia, and tropical Africa), plus an indigenous component, corresponding to species that supposedly differentiated in situ within the limits of the Basin (Blondel & Aronson, 1999). Traditionally, designation of these five main ‘elements’ dominating the Basin's extant flora followed from fossil data or the historical interpretation of extant distribution patterns of taxonomic and/or phytogeographical entities (e.g. Quézel, 1985).

Aided by molecular phylogenetic reconstructions and estimates of absolute divergence times, researchers have now the opportunity to confirm long-held hypotheses about single or multiple biogeographic origins as well as the possible paths and periods of migration of plant lineages invading the Mediterranean Basin. For example, Caujapé-Castells et al. (2001) employed cpDNA restriction-site analysis to demonstrate that Mediterranean species of Androcymbium (Colchicaceae) are of Afro-tropical origin, and that the northward migration routes followed by this genus were at times restricted to the mountain ranges of eastern Africa, which probably acted either as stepping stones or as refugia. Interestingly, the genus likely arrived and started to diversify in the area under review in the Middle Miocene (c. 12 million years ago, Mya), hence well before the establishment of the seasonal Mediterranean-type climate, about 3.2–2.8 Mya (Suc, 1984). This suggests that the geophytic life-form of Androcymbium species in the Mediterranean did not specifically evolve as adaptation to present-day Mediterranean conditions but rather represents an ancient, plesiomorphic trait supposedly maintained by natural selection. Overall, results obtained for Androcymbium are consistent with traditional biogeographic views about ancient contacts between southern/eastern Africa and the Mediterranean. This is further illustrated by recent phylogenetic work, e.g. in Senecio (Asteraceae; Coleman et al., 2003), as well as a symposium presentation dealing with cpDNA haplotype variation in the Olea europaea complex, supporting an ancestral origin of eastern Mediterranean variants in subtropical Africa (Guillaume Besnard, University of Lausanne, Switzerland).

As to the temperate Eurasian element, relevant molecular phylogenetic data from Mediterranean buttercups were presented (Elvira Hörandl, University of Vienna, Austria). Both nuclear (ITS) and chloroplast (matK) sequence data suggest that Mediterranean members of Ranunculus sect. Ranunculastrum derive from two distinct lineages in the western and eastern parts of the Basin, respectively, and therefore must have colonized the region independently. Such parallel evolution of independent lineages sets the stage for further inquiry into multiple-trait parallelisms promoted by natural selection (Levin, 2001). If, for example, such lineages evolved after the establishment of the Mediterranean climate, one may ask whether they share the same set of morphologies and life-history traits in response to the similar environmental conditions in which they both occur now. Clearly, if common, such parallel changes may frequently result in faulty estimates of morphology-based classifications but, from an evolutionary perspective, will contribute to an increasing support for determinism in evolution (Wood et al., in press).

However, it is the south-west Asian (‘Irano-Turanian’) stock that most likely donated the greatest number of derivatives to the Mediterranean Basin and even more remote regions. Here, the genus Hordeum (barley) is of particular interest. Based on dated phylogenies derived from three nuclear gene regions, the genus is thought to have originated about 12 Mya in the eastern Mediterranean and adjacent Mesopotamian regions (Frank Blattner, IPK Gatersleben, Germany). From there, it spread not only to Europe and Central Asia but also to the Mediterranean climate zones of South Africa and the Americas, often involving colonization of novel habitats at high elevations and/or latitudes. This startling range expansion now poses the challenge to identify the genetic, genomic and ecological changes associated with it, as well as the primary isolating barriers involved in the adaptive divergence of species in this system (see also Ramsey et al., 2003).

It is certainly premature to make general statements about the patterns, processes, and times of arrival and settlement of the present-day flora of the Mediterranean Basin. This will require coordinated efforts of generating dated phylogenies for a wide diversity of evolutionarily independent plant taxa, perhaps similar to recent molecular studies of the comparative biogeography of European high mountain plants (Comes & Kadereit, 2003).

Illuminating large-scale biogeographic trends is also leading to more detailed intraspecific studies at the regional level, for example, involving populations north and south of the Gibraltar Strait (separating the Iberian Peninsula from North Africa, 14 km) to test for genetic discontinuities as they may have been impacted since the Messinian salinity crisis (c. 6.0–5.3 Mya) when most of the present-day Mediterranean Sea dried-up. However, in contrast to what has been found in animal systems (Castella et al., 2000), the Strait is probably less likely to act as a severe barrier to historical or current gene flow for plants, as supported by a recent AFLP survey in Hypochaeris salzmanniana (Asteraceae; Tod F. Stuessy, University of Vienna, Austria).

Paleoendemism, radiations, and Quaternary climate change

One noteworthy feature of the Basin's extant flora concerns patterns of very limited or disjunct geographical distribution, and low species-genus ratios, with many primarily long-lived taxa restricted to island or mountain habitats (Blondel & Aronson, 1999). In numerous instances, this has fuelled speculations about their status as ‘paleoendemics’ of likely Tertiary origin. One rare endemic from the Maritime Alps, Saxifraga florulenta, does serve as a good example. Here, molecular evidence indicates that the species evolved in the Late Miocene, most likely between the Tortonian (c. 11.6–7.2 Mya) and the Messinian (Elena Conti & Frank Rutschmann, University of Zürich, Switzerland). Precariously, if current trends of global warming continue, the strict ecological adaptation of this saxifrage to siliceous substrates at the highest altitudes of the Maritime Alps might represent an extinction risk. In general, barring the presence of unrecognised sibling/descendant species, such Mediterranean paleoendemics should be of considerable interest to evolutionists as they imply long periods without speciation. In turn, identifying factors that prevent these plants from diversifying (e.g. low dispersal ability) may provide clues to what is required for speciation (Coyne & Orr, 2004).

Nevertheless, there are many primarily short-lived plant groups that have produced large numbers of ecogeographically differentiated species in the Basin and there is increasing evidence that adaptive radiation has taken place relatively recently. This is particular so for a Mediterranean clade of Senecio where a molecular phylogeny indicates a near simultaneous and relatively recent diversification of c. 10 species (of presently wide or restricted distribution) within the Late Quaternary (< 1.0 Mya; Coleman et al., 2003). Still, the number of such studies is not great, but is likely to grow in the near future as several basin-wide phylogenies presented during the conference are equally suggestive of relatively recent and ‘explosive’ speciation events, e.g. within Compositae-Anthemideae (Christoph Oberprieler, University of Regensburg, Germany) or Centaurium (Gentianaceae; Guilhelm Mansion, University of Neuchâtel, Switzerland).

Especially for short-lived taxa, perhaps this newly emerging picture of recent and rapid radiations is not surprising given the recent establishment of a Mediterranean-type climate at the Plio–Pleistocene boundary, which may have allowed the invasion of new and unoccupied habitats following the decimation of a Tertiary mesophytic flora (Coleman et al., 2003). Arguably, however, as more information from paleoenvironmental and genetic studies is accumulating (Willis et al., 2004), the question also arises of whether the Quaternary climatic oscillations (≤ 1.8 Mya) have exerted a major influence of patterns of geographic diversification and increased rates of speciation in the Basin, as suggested for Mediterranean warblers of the genus Sylvia (Blondel et al., 1996) or Anatolian mountain frogs (Veith et al., 2003).

Unfortunately, we have very few phylogenetic data of Mediterranean plant taxa that have been used to test hypotheses about differential rates of diversification within (or among) clades that may coincide with Quaternary climatic change or associated movements of sea level. This last factor, which in particular provides opportunities for allopatry, is likely responsible for accelerated speciation rates in Nigella (Ranunculaceae) from the Aegean archipelago (Christiane Bittkau, University of Mainz, Germany).

Finally, it is against this paleoclimatic background that the rapidly growing discipline of molecular phylogeography (Avise, 2000) comes into full swing, and several participants in the symposium drew attention to the importance of the Quaternary climatic changes in having shaped the present-day spatial genetic structure of a variety of plant species in the Mediterranean Basin, including ferns (Asplenium), conifers (Pinus), evergreen/deciduous oaks, and various coastal herbs (e.g. Cakile maritima). However, it is often difficult to draw firm inferences on the past history of populations entirely from genetic data when paleoecological data are non-informative or lacking. On the other hand, the power of the phylogeographic approach is illustrated by recent mtDNA studies in Pinus sylvestris, indicating the existence of refugia in the northern Iberian Peninsula that allowed the species to survive Pleistocene glaciations (Rachid Cheddadi, University of Montpellier, France). Significantly, this invalidates a common misapprehension that Iberian glacial refugia of woodland species were entirely confined to the southernmost parts of the peninsula.

Conclusions

In spite of having witnessed past periods of flourishing plant taxonomic and biosystematic research (e.g. Strid, 1970), it is only recently that the Mediterranean region, and the Basin in particular, have regained interest from evolutionists. As noted above, the area abounds in strikingly varying patterns of species distributions and species-genus ratios. This requires not only consideration of the temporal origins of a wide diversity of plant lineages but also their rates of species formation over time, as well as attempts to correlate these rates with various paleogeological, paleoclimatic or ecological factors, or inherent biological features of the organisms (key characters) that might promote speciation (Coyne & Orr, 2004).

Although these topics received relatively little attention during the symposium, theoretical and practical advances have been outlined for investigating the general causes and rates of speciation within (and among) clades based on phylogenies derived from molecular data (Barraclough & Nee, 2001; Kadereit et al., 2004). In addition, phylogeny-based methods now exist for inferring the geographic mode of speciation (allopatric, sympatric), e.g. by plotting the degree of range overlap of sister taxa against their age of divergence. This last approach, while criticised on grounds of past range shifts that may eradicate the signature of the biogeography of speciation, has been successfully applied to a diverse range of animals (see Coyne & Orr, 2004), suggesting that sympatric speciation is infrequent in nature.

Evidently, results obtained from all the above approaches must be qualified in view of their underlying assumptions, including the issue of (in)complete taxon sampling, the resolution and robustness of the underlying phylogenies (which too often are confounded by hybridisation, incomplete lineage sorting, or both), or the difficulties to calibrate branch lengths in real time. If reliable, these new analytical methods will provide an effective approach toward settling some long-standing questions about the origin and temporal course of plant diversification in the Mediterranean region and the forces involved.