The phylogeography of five flowering plant species (Cakile maritima, Eryngium maritimum, Salsola kali, Halimione portulacoides, Crithmum maritimum) widespread along the European coasts was investigated across their entire European range using AFLP evidence. Both similarities and dissimilarities were found. All species contain a distinct Black Sea/Aegean Sea cluster, and all except E. maritimum contain a distinct Adriatic Sea cluster or group of genetically very similar clusters. All species except Cr. maritimum contain a distinct Atlantic Ocean/North Sea/Baltic Sea cluster clearly separate from the Mediterranean material. In Ca. maritima a distinct Baltic Sea subcluster was found. In the western Mediterranean, two species groups can be recognized. Whereas in S. kali and E. maritimum material from this area falls into only one cluster or a group of genetically very similar clusters, it falls into two clusters or groups of clusters of either more Atlantic or more central Mediterranean similarity in Ca. maritima, H. portulacoides and Cr. maritimum. Similarities and dissimilarities in patterns found are discussed in terms of a combination of historical and extant abiotic and biotic factors. Thus, the distribution range of all species in the eastern Mediterranean area was not affected by Quaternary temperature changes, resulting in phylogeographic congruency here. The existence of distinct Black Sea/Aegean Sea and Adriatic Sea clusters or groups of clusters is the result of sea currents isolating these regions from each other. In the western Mediterranean basin the more cold‐sensitive species (H. portulacoides, Cr. maritimum) but not the less cold‐sensitive species (S. kali, E. maritimum) had to retreat from northern coasts. Re‐colonization of these areas from two different directions is implied by their phylogeographic pattern. The existence of a distinct Gibraltar gap is explained in terms of extant sea currents. Comparison of phylogeographic patterns found with those observed in either marine or terrestrial organisms leads to the conclusion that marine dispersal is of overriding importance in these coastal plants.

No abstract

BackgroundPast clonal propagation of olive trees is intimately linked to grafting. However, evidence on grafting in ancient trees is scarce, and not much is known about the source of plant material used for rootstocks. Here, the Simple Sequence Repeat (SSR) marker technique was used to study genetic diversity of rootstocks and scions in ancient olive trees from the Levant and its implications for past cultivation of olives. Leaf samples were collected from tree canopies (scions) and shoots growing from the trunk base (suckers). A total of 310 trees were sampled in 32 groves and analyzed with 14 SSR markers.ResultsIn 82.7% of the trees in which both scion and suckers could be genotyped, these were genetically different, and thus suckers were interpreted to represent the rootstock of grafted trees. Genetic diversity values were much higher among suckers than among scions, and 194 and 87 multi-locus genotypes (MLGs) were found in the two sample groups, respectively. Only five private alleles were found among scions, but 125 among suckers. A frequency analysis revealed a bimodal distribution of genetic distance among MLGs, indicating the presence of somatic mutations within clones. When assuming that MLGs differing by one mutation are identical, scion and sucker MLGs were grouped in 20 and 147 multi-locus lineages (MLLs). The majority of scions (90.0%) belonged to a single common MLL, whereas 50.5% of the suckers were single-sample MLLs. However, one MLL was specific to suckers and found in 63 (22.6%) of the samples.ConclusionsOur results provide strong evidence that the majority of olive trees in the study are grafted, that the large majority of scions belong to a single ancient cultivar containing somatic mutations, and that the widespread occurrence of one sucker genotype may imply rootstock selection. For the majority of grafted trees it seems likely that saplings were used as rootstocks; their genetic diversity probably is best explained as the result of a long history of sexual reproduction involving cultivated, feral and wild genotypes.

In Israel Eruca sativa has a geographically narrow distribution across a steep climatic gradient that ranges from mesic Mediterranean to hot desert environments. These conditions offer an opportunity to study the influence of the environment on intraspecific genetic variation. For this, we combined an analysis of neutral genetic markers with a phenotypic evaluation in common-garden experiments, and environmental characterization of populations that included climatic and edaphic parameters, as well as geographic distribution. A Bayesian clustering of individuals from nine representative populations based on amplified fragment length polymorphism (AFLP) divided the populations into a southern and a northern geographic cluster, with one admixed population at the geographic border between them. Linear mixed models, with cluster added as a grouping factor, revealed no clear effects of environment or geography on genetic distances, but this may be due to a strong association of geography and environment with genetic clusters. However, environmental factors accounted for part of the phenotypic variation observed in the common-garden experiments. In addition, candidate loci for selection were identified by association with environmental parameters and by two outlier methods. One locus, identified by all three methods, also showed an association with trichome density and herbivore damage, in net-house and field experiments, respectively. Accordingly, we propose that because trichomes are directly linked to defense against both herbivores and excess radiation, they could potentially be related to adaptive variation in these populations. These results demonstrate the value of combining environmental and phenotypic data with a detailed genetic survey when studying adaptation in plant populations.