Journal of Biogeography (J. Biogeogr.) (2009) 36, 1384–1397

SPECIAL Phylogeographic patterns in Hypochaeris ISSUE section Hypochaeris (, Lactuceae) of the western Mediterranean Marı´aA´ ngeles Ortiz1*, Karin Tremetsberger1,2, Tod F. Stuessy2, Anass Terrab1,2, Juan L. Garcı´a-Castan˜o1 and Salvador Talavera1

1Departamento de Biologı´a Vegetal y Ecologı´a, ABSTRACT Universidad de Sevilla, Sevilla, and Aim To analyse phylogeographic patterns in the four species of Hypochaeris sect. 2Department of Systematic and Evolutionary Botany, Faculty Centre Biodiversity, University Hypochaeris, evaluating possible areas of origin and the microevolutionary of Vienna, Vienna, Austria processes that have shaped their morphology, genetics and distribution. Location Western Mediterranean area. Methods We applied amplified fragment length polymorphism (AFLP) markers to a total of 494 individuals belonging to 82 populations of Hypochaeris arachnoidea, H. glabra, H. radicata and H. salzmanniana to determine population structure. Results Populations with the largest proportion of private and rare AFLP fragments were found in . This region was consequently inferred to be the ancestral area for H. arachnoidea, H. glabra, H. radicata and H. salzmanniana. The Guadalquivir River (southern Spain) was inferred to be an effective dispersal barrier for H. glabra and H. radicata. The Strait of Gibraltar was inferred to be a somewhat weaker barrier than the Guadalquivir River for H. radicata and a much weaker barrier for H. glabra. The main barrier for H. salzmanniana coincides with the extension of the Rif Mountains to the Atlantic coast in Morocco, and the Strait of Gibraltar is a much weaker barrier for this species. Hypochaeris arachnoidea appears to have originated in the Atlas Mountains. Main conclusions The highest levels of genetic variation in La Mamora forest (H. glabra and H. salzmanniana) or the adjacent central Middle Atlas (H. arachnoidea and H. radicata) in Morocco suggest that these areas were a centre of origin of Hypochaeris sect. Hypochaeris. All three potential barriers – the Guadalquivir River, the Strait of Gibraltar, and the Rif Mountains – have been *Correspondence: Marı´aA´ ngeles Ortiz, important in shaping genetic diversity in species of section Hypochaeris. ´ ´ Departamento de Biologıa Vegetal y Ecologıa, Keywords Universidad de Sevilla, Apdo-1095, 41080 Sevilla, Spain. AFLP, Atlas Mountains, Guadalquivir River, , , E-mail: [email protected] population genetic variation, rare fragments, Rif Mountains, Strait of Gibraltar.

which now separates the old Palaeozoic lands of the Sierra INTRODUCTION Morena to the north from the Tertiary lands of the Betic The western Mediterranean region has undergone dramatic Sierras; the Riffian Corridor occupied the modern Loukos and geomorphological and environmental changes during the past Sebou river valleys in Morocco, which separate the Rif 8 Myr (Thompson, 2003). The principal events were the Mountains to the north from the Atlas ranges to the south. closure of the connection to the Atlantic (7–5.33 Ma) and the These sea channels constituted formidable barriers for the opening of the Strait of Gibraltar (5.33 Ma). migration of and animals between North and The pre- had two seaway connections Europe. This situation changed dramatically with the closure with the Atlantic Ocean: the Betic and the Riffian corridors. of the two Mediterranean–Atlantic channels (at 7–5.33 Ma), The Betic Corridor became the future Guadalquivir valley, creating the so-called Messinian Salinity Crisis, with the drying

1384 www.blackwellpublishing.com/jbi ª 2009 The Authors doi:10.1111/j.1365-2699.2008.02079.x Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris out of much of the Mediterranean Basin, followed by the 1995, 2008; Weiss et al., 2003; Weiss-Schneeweiss et al., 2003, opening of the Strait of Gibraltar (at around 5.33 Ma). Until 2007, 2008), DNA sequences (Cerbah et al., 1998; Samuel c. 3.2 Ma, the climate of the south-western Mediterranean area et al., 2003; Tremetsberger et al., 2005), AFLP population seems to have been subtropical, although cold and arid analyses (Stuessy et al., 2003; Tremetsberger et al., 2003a,b, conditions were apparently established in North Africa from 2004, 2006; Muellner et al., 2005; Mraz et al., 2007; Ortiz around 3.9 Ma (Estabrook, 2001). A mediterranean-type et al., 2007, 2008) and reproductive biology (Ortiz et al., climate, with warm, wet winters and hot, dry summers, seems 2006). To date, South American species have been the focus to have become established during the Pliocene (around of research because they represent the greatest concentration 3.5 Ma), before the onset of the Northern Hemisphere of species diversity. glaciations at around 2.4 Ma. The latter gave rise to oscillating We review available data for Hypochaeris sect. Hypochaeris changes in sea level in the Strait of Gibraltar area that exposed and focus on four biogeographical issues: (1) the possible area and submerged islands in this area during the glacial cycles. Up of origin of this section; (2) the impact of the Strait of to now, relatively few studies have been concerned with the Gibraltar on populations of H. glabra, H. radicata and evolutionary and biogeographical consequences of Pleistocene H. salzmanniana; (3) the impact of the Guadalquivir River glacial cycles on northern African and southern European taxa. in southern Spain on population divergence in H. glabra and There were three important biogeographical barriers for H. radicata; and (4) patterns of genetic divergence in terrestrial species in the western Mediterranean area after H. arachnoidea in Morocco. the Messinian age: the Guadalquivir valley, the Strait of Gibraltar, and the Loukos and Sebou valleys in Morocco. The MATERIALS AND METHODS most important bioclimatic changes in the Mediterranean Basin over the last 5 Myr were: (1) the establishment of the Hypochaeris sect. Hypochaeris mediterranean climate during the Pliocene (c. 3.5 Ma), and (2) the glacial periods of the Pliocene and Pleistocene. The onset of Section Hypochaeris is a monophyletic group (Tremetsberger the mediterranean climate caused very dramatic changes in the et al., 2005) composed of four species: H. arachnoidea Poir., woodlands, particularly in the understorey, with selective H. glabra L., H. radicata L. and H. salzmanniana DC. In pressure to shorten plant life cycles, or to modify underground addition to having morphological differences, these species rhizomes to form corms. As a consequence, more than 50% of differ in a range of other parameters: (1) life-form: H. radicata the plant species of the Mediterranean Basin are annuals is perennial whereas H. arachnoidea, H. glabra and H. salz- (Talavera, 1991), and rhizomatous plants are confined to the manniana are annuals; (2) somatic chromosome number: more humid environments (e.g. streams, lagoons and springs) H. glabra has 2n = 10 chromosomes, whereas H. arachnoidea, or to the mesic understorey. The Quaternary glacial periods H. radicata and H. salzmanniana have 2n = 8 chromosomes shaped the expansion and diversification of plant populations (Tremetsberger et al., 2005); and (3) distribution: H. glabra along the western Mediterranean. During glacial periods, the and H. radicata are widespread in the Mediterranean region sea level was lowered, for example by 130 m in the Last Glacial and also occur as weeds world-wide, whereas H. arachnoidea is Maximum, some 20,000 years ago, and the temporarily endemic to North West Africa (Morocco and Algeria) and emergent coastlines occupied by plant populations were used H. salzmanniana is restricted to the Atlantic coast of Morocco as land-bridges for colonization. During interglacials, the sea and south-western Spain (Ca´diz). The natural habitat of the transgressed and coastal populations suffered drastic reduc- former three species is the understorey of open Quercus tions and possibly extinctions. All of these influences have led woodland (although weedy invasive populations of H. glabra to complex evolutionary and biogeographical patterns in the and H. radicata occur in different habitats), but H. salzman- biota of the Mediterranean region, especially in the western niana occurs principally on coastal dunes. All four species can part (Jong, 1998; Ve´la & Benhouhou, 2007). hybridize in the greenhouse, but in nature hybrids have never Further assessment of the impact of these environmental been found in the natural range of the species. However, changes in the western Mediterranean requires the study of Parker (1975) found sterile hybrids between H. glabra and additional groups, such as the genus Hypochaeris, which can H. radicata in England, where both species are non-native. As serve as a model system. The genus consists of c. 58 species with many composites, Hypochaeris species have sporophytic world-wide, with only 15 confined to the Mediterranean self-incompatibility (SSI), requiring obligatory cross-pollina- region, three in Eurasia, and more than 40 in South America. tion (usually mediated by solitary bees) for fruit-set. Within Of those in the Mediterranean area, section Hypochaeris, with section Hypochaeris, SSI is found in H. arachnoidea and four species, is centred in the western Mediterranean. This H. radicata and in most populations of H. salzmanniana, section is monophyletic (Tremetsberger et al., 2005) and whereas H. glabra is self-compatible (Ortiz et al., 2006). diversified between c. 1.7 and 2.0 Ma (K. Tremetsberger et al., unpublished data). In general, Hypochaeris is a suitable Sampled populations genus with which to evaluate biogeographical patterns because of the many previous studies conducted on its We sampled a total of 494 individuals belonging to 82 cytology and cytogenetics (Cerbah et al., 1995; Ruas et al., populations of the four species of section Hypochaeris (see

Journal of Biogeography 36, 1384–1397 1385 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

Appendix S1 in Supporting Information). Most populations replicated individuals of H. radicata (Bonin et al., 2004), were in the western Mediterranean region (the centre of amounted to 1.12% (Ortiz et al., 2008). diversity of the group – see below), but we also included other populations from throughout the distributional ranges: 15 Data analyses populations of H. glabra (five from Morocco, G1–G5; eight from the Iberian Peninsula, G6–G13; one from the Canary AFLP markers can be used to infer phylogenetic relationships Islands, G14; and one from Chile, G15); 44 populations of based on measures of genetic distance for closely related H. radicata [11 from Morocco, R1–R11; five from the Central species (Beismann et al., 1997; Mueller & Wolfenbarger, 1999; Mediterranean area (, and Tunisia), R12–R17; 17 Zhang et al., 2001; Despre´s et al., 2003). This method was from the Iberian Peninsula, R18–R34; three from West and successful in determining phylogenetic relationships among Central Europe (, the Netherlands and Austria), R35– South American species of Hypochaeris (Tremetsberger et al., R37; three from Asia (Taiwan and South Korea), R38–R40; 2006). For this purpose we scored one individual from five and four from South America (Argentina and Chile), R41– populations of each species, covering as much as possible of R44]; nine Moroccan populations of H. arachnoidea (A1–A9); the natural distributional range (H. glabra: G4, G5, G8, G10 and 14 populations of H. salzmanniana (six from Morocco, and G12; H. radicata: R7, R21, R26, R27 and R28; H. arach- S1–S6; and eight from Ca´diz, Spain, S7–S14). Of these, all noidea: A1, A4, A6, A7 and A9; H. salzmanniana: S3, S6, S9, populations of H. arachnoidea, and Moroccan and central and S11 and S13). From this presence/absence matrix, we con- south-eastern Spanish populations of H. glabra were newly structed a dendrogram using the neighbour-joining (NJ) analysed in this study. Populations of H. radicata are those method in conjunction with Nei & Li’s (1979) genetic of Tremetsberger et al. (2004) and Ortiz et al. (2008), and distances in paup* (ver. 4.0b10; Sinauer Associates, Sunder- populations of H. salzmanniana are those of Tremetsberger land, MA, USA). Support for each node was tested by 10,000 et al. (2004) and Ortiz et al. (2007). The number of individuals bootstrap replicates. sampled in each population is shown in Appendix S1. Fresh We used famd ver. 1.1 (Schlu¨ter & Harris, 2006) to leaves of the plants were collected at least 1 m apart and dried interchange between different file formats and calculate the in silica gel. Vouchers of all sampled populations were proportion of private AFLP fragments (i.e. those confined to deposited in the Herbarium of the University of Seville only one species; Fpriv). Pairwise exclusive shared fragments (SEV, Spain) and/or the University of Vienna (WU, Austria). (i.e. fragments exclusively shared by a pair of species that

are not present in any other species; Fsh) and pairwise fixation indices based on Euclidean distances [analysis of DNA isolation and AFLP analysis molecular variance (AMOVA)-derived FST values; arlequin Total genomic DNA was extracted from dry leaf material ver. 3.01 (Excoffier et al., 2005)] were assessed for each following the CTAB (cetyl trimethyl ammonium bromide) species. protocol (Doyle & Doyle, 1987) with modifications. The For each species independently, the AMOVA-derived pop- amplified fragment length polymorphism (AFLP) procedure ulation pairwise FST matrix, based on the squared Euclidean followed established protocols (Vos et al., 1995) with modi- distances, was calculated with arlequin ver. 3.01 (Excoffier fications (Tremetsberger et al., 2003a, 2004; Ortiz et al., 2007). et al., 2005) and imported into splitstree ver. 4.6 (Huson & The six primer combinations for the selective polymerase chain Bryant, 2006) so as to construct a NJ population dendrogram. reaction (PCR) selected by Tremetsberger et al. (2004) were As a measure of within-population diversity, we assessed the applied to all four species: MseI-CTCG/EcoRI-ATC (Fam), percentage of polymorphic fragments (Fpoly), as well as the

MseI-CAC/EcoRI-ACG (Hex), MseI-CTA/EcoRI-ACC (Ned), number of private fragments (Fpriv, confined to one population MseI-CTG/EcoRI-ACA (Fam), MseI-CTC/EcoRI-AGG (Hex) or group of populations) for all populations of the four species and MseI-CTGA/EcoRI-AAC (Ned). In addition, three more of Hypochaeris sect. Hypochaeris. We also assessed the number primers selected by Ortiz et al. (2007), namely MseI-CAC/ of fragments that were shared exclusively between pairs or

EcoRI-ACT (FAM), MseI-CTC/EcoRI-ATC (HEX) and MseI- groups of populations (Fsh), and we calculated another index CTG/EcoRI-AAC (NED), were used to obtain better resolution of diversity DW (‘frequency-down-weighted marker values’) in H. salzmanniana. The fluorescence-labelled selective ampli- with AFLPdat (Ehrich, 2006). Population values were esti- fication products were separated on a 5% polyacrylamide gel mated by making a table with the presence of markers by with an internal size standard (GeneScan-500 ROX; PE population, dividing each marker by the total number of Applied Biosystems, Foster City, CA, USA) on an automated occurrences of this marker in the dataset, and summing these sequencer (ABI 377; Perkin-Elmer, Waltham, MA, USA). relative values to give the rarity index for this particular Amplified fragments of 60–500 bp were scored and exported as population (rarity 2). Private and rare fragments accumulate a presence/absence matrix using ABI Prism genescan through time and are, therefore, a measure of population analysis Software 2.1 (PE Applied Biosystems) and genog- antiquity (Stehlik et al., 2002; Scho¨nswetter & Tribsch, 2005). rapher (ver. 1.6.0 Montana State University 2001; available As another measure of genetic variability, we also calculated at: http://hordeum.oscs.montana.edu/genographer/). The the average gene diversity (HD; arlequin ver. 3.01; Excoffier error rate, based on phenotypic comparisons among the 37 et al., 2005).

1386 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

In order to examine the population structure of the H. salzmanniana (95% BS) and H. radicata is sister to this four species, H. arachnoidea, H. glabra, H. radicata and latter group. H. salzmanniana, we conducted an approach based on statistical The species with the largest number of private fragments inference with Bayesian clustering methods using baps ver. 5.1 (76, of which 10 were fixed) was H. salzmanniana (see (Corander et al., 2003, 2004; Corander & Marttinen, 2006; Table 1). Hypochaeris glabra exhibited the fewest private available at http://www.abo.fi/fak/mnf/mate/jc/software/baps. fragments (52), but the highest number of fixed private html), which uses stochastic optimization instead of Markov fragments (20). Hypochaeris radicata shared most exclusive chain Monte Carlo (MCMC) to find the optimal partition. The fragments with the other three species (a total of 45), and simulation was run from K =2toK = N + 1, where N is the H. glabra had the lowest number of exclusive fragments shared number of populations analysed in each species, except for with other species (a total of 26). Most exclusive fragments H. radicata, for which the simulation was run to 20, with five were shared between H. salzmanniana and H. arachnoidea replicates for each K. We used the option ‘clustering of (20), indicating a close relationship. AMOVA-derived FST individuals’ to estimate the admixture coefficients for the values indicated a weaker relationship between H. glabra and reference individuals, and this was performed with the following H. salzmanniana (FST = 0.67). settings: minimal size of clusters at four individuals, 100 iterations to estimate the admixture coefficients for the Hypochaeris glabra individuals, 200 simulated reference individuals from each population, and 20 iterations. The six AFLP primer combinations applied to H. glabra A post hoc Tukey–Kramer honestly significant difference yielded a total of 242 fragments, of which 81.8% were (HSD) test was applied to detect differences in private polymorphic. The unrooted NJ dendrogram and Bayesian fragments and DW among populations. We considered analysis, conducted with baps, for the 15 populations of H. differences significant at a 5% confidence level (Bonferroni glabra (Fig. 2a), showed three main clusters: the first (70% BS) correction applied). includes the populations from Morocco (populations G1–G5) We used AMOVA (arlequin ver. 3.01; Excoffier et al., and the Spanish populations from areas south of the Guadal- 2005) to distribute genetic variation into portions assignable to quivir River (G6–G7); the second cluster includes populations differences between predefined hierarchical groups (FCT), from central and south-eastern Spain (G8–G10; 100% BS); and among populations within these groups (FSC), and among the remaining populations form a distinct Bayesian cluster populations across the entire study area (FST) (Turner et al., (but without NJ bootstrap support) that groups the popula- 2000). We tested with AMOVA for possible effects of tions from Don˜ana and south-western Sierra Morena (south- potentially major geographic barriers (the Strait of Gibraltar, western Spain, G11–G13) together with those from the Canary Guadalquivir River, and extension of the Rif Mountains) for Islands (G14) and Chile (G15; the latter two supported by the three species H. glabra, H. radicata and H. salzmanniana. 75% BS).

FST values were calculated with arlequin ver. 3.01 (based Thus, the main splits within this species were between the on squared Euclidean distances), and, in order to counter- populations from Morocco and from the south of balance unequal sample sizes, qST values [based on samples the Guadalquivir River in Spain, those from Don˜ana, from of two individuals, i.e. qST(2)] were calculated according to the south-western Sierra Morena, and from the extra- the rarefaction method (Hurlbert, 1971; Mousadik & Petit, European introduced accessions (Canary Islands and Chile), 1996). and those from central and south-eastern Spain. Genetic diversity measures for each population of H. glabra are shown in Table 2. The highest numbers of polymorphisms and RESULTS average gene diversity were in the Moroccan populations (%F = 28.61 and H = 0.091) and the central and Phylogenetic relationships among the four species poly D south-eastern Spanish populations (%F = 28.55 and of section Hypochaeris poly HD = 0.091; Fig. 2b). Moroccan populations had the highest The AFLP primer combinations applied to five populations number of rare fragments and private fragments (DW = 18.0

(one individual per population) of each of the four species of and Fpriv = 4.8), and the lowest numbers were in Spanish section Hypochaeris generated 32–118 fragments, of which a populations south of the Guadalquivir River (DW = 6.3 and high percentage (72.5–100%) were polymorphic. The total Fpriv = 0.0). number of fragments was 428, of which 406 (94.8%) were polymorphic. The NJ dendrogram with the four species of Hypochaeris radicata section Hypochaeris (Fig. 1) had the same topology as the rps16 intron tree (Tremetsberger et al., 2005): H. glabra is sister to The population structure of this species was analysed in detail H. arachnoidea, H. radicata and H. salzmanniana. Each of the by Ortiz et al. (2008). These data were used to construct a NJ four species is well supported, with bootstrap support (BS) unrooted tree and perform Bayesian clustering (Fig. 3a). The ranging from 89% (for H. radicata) to 100% (for H. glabra main clusters were: Morocco (R1–R11); south of the Guadal- and H. salzmanniana). Hypochaeris arachnoidea is sister to quivir (Spain; R18–R24); Don˜ana (Spain; R25–R26); the

Journal of Biogeography 36, 1384–1397 1387 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

Figure 1 Neighbour-joining (NJ) dendro- gram of 20 individuals of Hypochaeris arachnoidea, H. glabra, H. radicata and H. salzmanniana analysed for AFLP and using Nei & Li’s genetic distances. Bootstrap values based on 10,000 permutations are indicated at each node (if > 50%).

Table 1 Pairwise exclusive shared fragments (Fsh; values below Hypochaeris salzmanniana diagonal), pairwise fixation index (AMOVA-derived FST; above diagonal), and private fragments (Fpriv) applied to five populations The population structure of this coastal species was analysed in (one individual per population) of each species of Hypochaeris detail by Ortiz et al. (2007). The results of the unrooted NJ sect. Hypochaeris. The values are based on analysis of a total of 428 dendrogram (Fig. 4a) showed five main clusters: Algeciras Bay AFLP fragments. Numbers in parentheses refer to fixed fragments. (Spain; S13–S14), Sierra San Bartolome´ (Spain; S11–S12), south of the extension of the Rif Mountains to the Atlantic FST coast in Morocco coincident with south of the Loukos River

Fsh H.glabra H.radicata H.salzmanniana H.arachnoidea (S3–S6), north of the extension of the Rif Mountains (Morocco; S1–S2), and Barbate (Spain; S7–S10); the latter H. glabra – 0.56 0.67 0.60 two grouped in the same Bayesian cluster. The averaged H. radicata 13 (1) – 0.44 0.34 genetic diversity parameters are shown in Fig. 4b. The H. salzmanniana 5 18 – 0.45 H. arachnoidea 8 14 20 (2) – population from south of the Loukos River in Morocco had the highest diversity. Fpriv 52 (20) 64 (3) 76 (10) 62 (4)

Hypochaeris arachnoidea Central Mediterranean (R12–R17); and a cluster including south-western Sierra Morena, northern, central and eastern The six AFLP primer combinations applied to H. arachnoidea Spain, and all the introduced accessions of the species (R27– yielded a total of 499 fragments, of which 93.2% were R44). The averaged genetic diversity parameters are shown in polymorphic. This species presented four main clusters Fig. 3b. Morocco is seen to be the group with the highest (Fig. 5a), the first from Taza, in the Rif Mountains (A1; diversity. 100% BS), the second from Essaouira, on the southern

1388 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

(b)

(a)

Figure 2 Sampling localities and results of AFLP analysis of Hypochaeris glabra: (a) neighbour-joining (NJ) unrooted dendrograms, based on analysis of molecular variance (AMOVA)-derived FST values (scale bar indicates an FST value of 0.1). Bootstrap values based on 10,000 permutations are indicated at each node (if > 50%). (b) Sampling localities of H. glabra in the western Mediterranean region. The bars indicate the mean number ± SE of the private fragments (Fp; max–min: 4.8–0), the rare fragments index (DW; max–min: 18.0–6.3) and the genetic diversity (HD; max–min: 0.091–0.034). Colour-coding of populations indicates results of Bayesian clustering (baps).

Moroccan Atlantic coast (A2–A4; 60% BS), and the third Mountains; Table 3) show that all three potential barriers from the Anti Atlas Mountains (A5–A6; 70% BS). The last may have been important in shaping genetic diversity in Bayesian cluster (not a group in the NJ tree) comprises species of section Hypochaeris. For H. glabra and H. radicata, geographically dispersed populations from Nador, on the the main breaks coincide with the Guadalquivir River in Mediterranean coast of Morocco (A7), Tiznit, in the foothills Andalusia and the Strait of Gibraltar. The Guadalquivir River of the Anti Atlas Mountains, close to the Atlantic coast (A8), accounts for 39.1% of the variance in H. glabra (compared and Tizi-N-Test from the High Atlas Mountains (A9). with 28.2% across the Strait of Gibraltar). For H. radicata, the Genetic diversity measures for each population of Guadalquivir River and the Strait of Gibraltar account for H. arachnoidea are shown in Table 2. The highest number 18.4% and 16.0% of genetic variance, respectively. For of polymorphisms, the highest average gene diversity, and the H. salzmanniana, the Strait of Gibraltar accounts for very highest number of rare fragments were found in the Anti little variance (11.9%), whereas the extension of the Rif

Atlas populations (A5–A6; %Fpoly = 71.24; HD = 0.150; Mountains to the Atlantic coast in Morocco accounts for DW = 53.65; Fig. 5b). The highest number of private frag- 17.6% of the variance. ments was found in Taza (A1; Fpriv = 24). The qST pairwise values correspond with the AMOVA-

derived FST values (data not shown). A post hoc Tukey–Kramer HSD test found significant differences in the number of private Evaluation of biogeographical hypotheses fragments and DW between populations from Morocco and all The analyses of molecular variance between populations on other populations (in H. glabra and H. radicata) and between either side of the potential biogeographical barriers (Strait of populations from south of Loukos River and all other Gibraltar, Guadalquivir River and extension of the Rif populations (in H. salzmanniana).

Journal of Biogeography 36, 1384–1397 1389 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

Table 2 Total number of fragments (Ftot), percentage polymorphic fragments (Fpoly), private fragments (Fpriv), rare fragments (DW), average gene diversity (HD) and number of samples (NAFLP) in populations of Hypochaeris glabra (G1–G15) and H. arachnoidea (A1–A9) sampled for AFLP.

Species/Pop. Localities Ftot Fpoly (%) Fpriv DW HD NAFLP

H. glabra G1 Morocco, Tanger 141 22.69 2 13.7 0.0677 5 G2 Asilah 133 10.53 1 11.2 0.0306 4 G3 Larache 162 40.74 5 20.9 0.1339 5 G4 Kenitra, La Mamora 1 167 40.12 12 (1) 27.2 0.1298 5 G5 Kenitra, La Mamora 2 145 28.96 4 17.1 0.0939 5 G6 S Guadalquivir, Spain, Barbate 112 – 0 5.7 – 2 G7 Punta Paloma 119 10.08 0 6.8 0.0489 2 G8 C & SE Spain,Ca´ceres 145 12.41 0 14.0 0.0326 5 G9 Jaen, Santa Elena 142 47.18 5 18.1 0.1404 5 G10 Almerı´a, Tabernas 142 26.06 0 14.6 0.1007 3 G11 SW Sierra Morena, Spain, Valverde 126 3.17 0 10.4 0.0163 2 G12 Don˜ana, Spain, Hinojos 119 5.04 0 8.0 0.0245 2 G13 P.N. Don˜ana 129 15.50 2 11.1 0.0816 2 G14 Introduced, Spain. Canary Islands 125 6.40 1 (1) 10.2 0.0326 2 G15 Chile, N˜ uble 124 10.48 0 9.0 0.0354 3 H. arachnoidea A1 Taza 160 60.62 24 (2) 40.4 0.0836 9 A2 Essaouira, Moulay-Bouzerktour 198 67.17 8 34.4 0.1106 10 A3 Essaouira 197 64.97 4 31.9 0.1049 10 A4 J. Amsittene 217 66.36 12 45.6 0.1140 10 A5 Anti Atlas, Biougra 230 69.56 5 44.8 0.1549 7 A6 Tafraoute 240 72.92 23 62.5 0.1441 10 A7 Nador, High Atlas, Anti Atlas, Nador 190 60.00 9 34.2 0.0998 9 A8 Tifnit 224 70.09 17 50.0 0.1331 10 A9 Tizi-N-Test 235 73.62 20 54.1 0.1516 9

ancestral variation of more recent climatic events (such as DISCUSSION glacial cycles). Hypochaeris arachnoidea is distributed throughout the Atlas Morocco: the ancestral area for section Hypochaeris Mountains in Morocco and Algeria in dry, open woodland Morocco has been inferred as the ancestral area of H. radicata (Oberprieler, 2002; Oberprieler & Vogt, 2002; Fo¨rther & and H. salzmanniana based on the presence of many private Podlech, 2003). Its probable origin is in the Atlas Mountains, and rare AFLP fragments in this region (Ortiz et al., 2007, although the exact location is difficult to determine, mainly 2008). In the case of H. radicata, the greatest numbers of because of limited sampling (populations from Algeria and private and rare fragments occur in the Rif, where it is found in parts of the Moroccan Middle Atlas have not been collected). wet pastures associated with Quercus suber forests, and in the Moreover, unlike the case for the other three species of section central Middle Atlas, where it grows in wet pastures associated Hypochaeris, there is no clear geographical genetic pattern. with forests of Cedrus atlantica. Ortiz et al. (2007) inferred Bayesian clustering groups several very distant populations, H. salzmanniana to have originated in the southern part of its and we interpret this as being the result of dispersal over long distributional range in Morocco, namely in the once more distances, possibly associated with human migrations. Moun- extensive Q. suber forests of the Sebou valley, which occur tain populations of H. arachnoidea (A1, A4–A6, A9) have today as remnant vegetation in La Mamora forest, close to higher mean levels of private and rare fragments (DW) than do Kenitra in the north-western foothills of the Middle Atlas. populations closer to the sea (A2–A3, A7–A8; 16.8 vs. 9.5 and Hypochaeris glabra also exhibits its highest genetic diversity in 49.5 vs. 37.6), and this pattern is also seen in populations the La Mamora area (G3–G5), where the species occurs in the around Essaouira (A4 vs. A2–A3). We hypothesize, therefore, understorey of the Q. suber forest, in mixed populations with that H. arachnoidea originated in a mountain habitat, and that H. salzmanniana. The Sebou valley seems to have played an the Atlas Mountains, at a lower elevation, served as a important role in diversification within the section. This Pleistocene refugium. The Atlas Mountains are known for diversification might be related to arid–wet cycles starting their high level of endemism (Que´zel, 1978; Fennane & Ibn 2.3 Ma (Suc, 1984). However, we do not know the effects on Tattou, 1998), also evident in other species of Hypochaeris, for

1390 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

(b)

(a)

Figure 3 Sampling localities and results of AFLP analysis of Hypochaeris radicata: (a) neighbour-joining (NJ) unrooted dendrograms, based on analysis of molecular variance (AMOVA)-derived FST values (scale bar indicates an FST value of 0.1). Bootstrap values based on 10,000 permutations are indicated at each node (if > 50%). (b) Sampling localities of H. radicata in the western Mediterranean region. The bars indicate the mean number ± SE of the private fragments (Fp; max–min: 4.1–0.5), the rare fragments index (DW; max–min: 9.4–4.8) and the genetic diversity (HD; max–min: 0.091–0.050). Colour-coding of populations indicates results of Bayesian clustering (baps). example H. angustifolia and H. leontodontoides (Gala´n de Mera to lagoons, and the two populations analysed (R25/El & Vicente Orellana, 1998a,b). Corchuelo and R26/El Acebro´n) are strongly genetically and morphologically divergent (Ortiz et al., 2006, 2008). This is not the case for H. glabra, which is widespread in the Don˜ana Phylogeographic patterns in H. glabra, H. radicata area. The similarity of populations of H. glabra and H. radicata and H. salzmanianna on both sides of the Strait of Gibraltar cannot be linked to the Hypochaeris glabra and H. radicata show very similar biogeo- closing of the Mediterranean Sea during the Messinian Salinity graphical patterns of genetic diversity. However, in H. glabra Crisis 6–7 Ma, because section Hypochaeris is estimated to be these results must be considered with caution because unequal of only Pliocene or Pleistocene age [1.7–2.0 Ma (95% Bayesian sample sizes were analysed. Based on high numbers of private highest posterior probability density interval = 0.6–3.5 Ma); and rare fragments, we infer a Moroccan origin for H. glabra Tremetsberger et al., 2005; K. Tremetsberger et al., unpub- and H. radicata (Ortiz et al., 2008). From here, H. glabra lished data]. appears to have first dispersed to the region north of the Hypochaeris salzmanniana is also hypothesized to have Guadalquivir River (south-western Sierra Morena, Don˜ana, originated in southern Moroccan Quercus woodlands and central and south-eastern Spain) and then world-wide. The migrated north (Ortiz et al., 2007). The main genetic division same basic pattern can be inferred for H. radicata (Ortiz et al., within the species is between populations on either side of the 2008). The second dispersal of both H. glabra and H. radicata extension of the Rif Mountains to the Atlantic coast in from Morocco across the Strait of Gibraltar was to the region northern Morocco. From northern Morocco, the taxon may south of the Guadalquivir River, as evidenced by strong genetic have crossed the Strait of Gibraltar during the Pleistocene, similarity between Moroccan populations and Spanish popu- when sea levels were lower (Ortiz et al., 2007). It developed lations south of the Guadalquivir River in both species (for differences in its mating compatibility system during north- more details, see Ortiz et al., 2008). The Don˜ana populations ward migration, from completely self-incompatible individuals of H. radicata are isolated and restricted to humid zones close in all Moroccan and Algeciras Bay populations, through mixed

Journal of Biogeography 36, 1384–1397 1391 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

(b)

(a)

Figure 4 Sampling localities and results of AFLP analysis of Hypochaeris salzmanniana: (a) neighbour-joining (NJ) unrooted dendro- grams, based on analysis of molecular variance (AMOVA)-derived FST values (scale bar indicates an FST value of 0.1). Bootstrap values based on 10,000 permutations are indicated at each node (if > 50%). (b) Sampling localities of H. salzmanniana in the western

Mediterranean region. The bars indicate the mean number ± SE of the private fragments (Fp; max–min: 18.3–3.0), the rare fragments index (DW; max–min: 50.7–15.8) and the genetic diversity (HD; max–min: 0.146–0.044). Colour-coding of populations indicates results of Bayesian clustering (baps).

populations with self-compatible and self-incompatible indi- H. salzmanniana. However, at least two dispersals from viduals in Barbate, to self-compatible individuals in Sierra San Morocco to the Iberian Peninsula are inferred for H. glabra Bartolome´ and Zahara (Ortiz et al., 2006, 2007). and H. radicata (this paper and Ortiz et al., 2008). Hypochaeris salzmanniana is a coastal species growing on sand dunes along the beaches and might have been better adapted to conditions Pleistocene glacial impact on H. glabra, H. radicata presented by the exposed sea floor during glacial periods than and H. salzmanniana were H. glabra and H. radicata. The normal habitat of H. glabra is European and African landmasses were not directly connected on sandy soils in woodlands, rather than coastal dunes. by a continuous land bridge when sea levels were lower in the Hypochaeris radicata is confined to more humid habitats in the Pleistocene, but their coasts were considerably closer than they understorey of Mediterranean Quercus woodland. are today, especially on the Atlantic side of the Strait of Gibraltar, where the sea floor is not as deep as it is on the Impact of the Guadalquivir River on H. glabra and Mediterranean side (Pou, 1989; Yokoyama et al., 2000; Patar- H. radicata nello et al., 2007). Emergent islands present periodically during glacial periods in the Strait of Gibraltar area also Hypochaeris glabra and H. radicata share a principal genetic favoured contact between the two continents (Collina-Girard, split across the Guadalquivir River (Spain) in the western 2001). This periodical closeness of the African and European Mediterranean region. In H. radicata, the morphologically and landmasses probably facilitated the colonization of H. glabra, genetically divergent Don˜ana populations have an intermedi- H. radicata and H. salzmanniana from North Africa to the ate position between the two groups north and south of the south-west of the Iberian Peninsula (Ortiz et al., 2007, 2008). Guadalquivir River. We hypothesize, on the basis of our The Strait of Gibraltar appears to be a rather strong barrier to results, that ancestral populations of H. glabra and H. radicata, gene flow in H. radicata and a weak one in H. glabra and like those of H. salzmanniana, expanded out of northern

1392 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

(a) (b)

Figure 5 Sampling localities and results of AFLP analysis of Hypochaeris arachnoidea: (a) neighbour-joining (NJ) unrooted dendrograms, based on analysis of molecular variance (AMOVA)-derived FST values (scale bar indicates an FST value of 0.1). Bootstrap values based on 10,000 permutations are indicated at each node (if > 50%). (b) Sampling localities of H. arachnoidea in the western Mediterranean region.

The bars indicate the mean number ± SE of the private fragments (Fp; max–min: 24.0–8.0), the rare fragments index (DW; max–min: 46.2–37.3) and the genetic diversity (HD; max–min: 0.150–0.084). Colour-coding of populations indicates results of Bayesian clustering (baps).

Africa across the Strait of Gibraltar area into the southern of acidic Precambrian and Palaeozoic terrain, whereas the Betic Iberian Peninsula during the Quaternary (Ortiz et al., 2007). Cordillera is predominantly calcareous. In the Sierra Morena, First, both species arrived north-west of the Guadalquivir H. radicata grows frequently and abundantly in the understorey River, probably because the coastline was more extensive and of Quercus forests. In the rest of the Iberian Peninsula north of consequently closer to Morocco at that time. These popula- the Guadalquivir River, the species grows in anthropogenically tions expanded, corresponding to the modern populational modified sites on different substrates. systems in Don˜ana, Sierra Morena, northern, central and Differentiation across the Guadalquivir Basin has been eastern Spain, and possibly . In a second more recent documented in several animal genera (Busack, 1986; Garcı´a- dispersal event, H. glabra and H. radicata reached a more Parı´s et al., 1998, 2003; Garcı´a-Parı´s & Jockusch, 1999; San- easterly area of the Iberian Peninsula, giving rise to the martı´n, 2003) and in two plant genera apart from Hypochaeris: southern Guadalquivir populations. Anthoxanthum (Pimentel et al., 2007) and (Comes & Interestingly, the Guadalquivir River is still a modern barrier Abbott, 1998). In Discoglossus and Salamandra (Amphibia), it is preventing the admixture of populations, as it is an important thought that lineages were isolated by the opening of the Betic agricultural area in which H. glabra and especially H. radicata Strait or later as a result of the formation of the fluvial system are seldom found. The Sierra Morena to the north-west of the during the Pliocene, and that this isolation has been maintained Guadalquivir River and the Betic Cordillera to its south-east until recently by the Guadalquivir River Basin (Garcı´a-Parı´s offer rather different soil conditions. The Sierra Morena consists et al., 1998; Garcı´a-Parı´s & Jockusch, 1999).

Journal of Biogeography 36, 1384–1397 1393 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

Table 3 Comparison of analyses of molecular variance (AMOVA) across three major geographic barriers, the Strait of Gibraltar, Guadalquivir River and Loukos River, for the three species Hypochaeris glabra, H. radicata and H. salzmanniana. Groupings of populations are shown in brackets.

Source of variation d.f. Sum of squares Variance components Variance (%) F-values 95% confidence interval

H. glabra [G1–G13]

Populations 12 789.45 15.54 60.21 FST = 0.602 Individuals 34 349.32 10.27 39.79 Strait of Gibraltar [G1–G5], [G6–G12]

Groups 1 251.72 8.36 28.23 FCT = 0.282 0.22–0.32

Populations 11 537.73 10.96 37.05 FSC = 0.516

Individuals 34 349.32 10.27 34.71 FST = 0.653 Guadalquivir River [G1–G7], [G8–G12]

Groups 1 326.95 12.39 39.13 FCT = 0.391 0.32–0.44

Populations 11 462.50 9.00 28.42 FSC = 0.467

Individuals 34 349.32 10.27 32.45 FST = 0.675 H. radicata [R1–R34]

Populations 33 3830.74 19.85 48.67 FST = 0.487 Individuals 129 2701.48 20.94 51.33 Strait of Gibraltar [R1–R17] [R18–R34]

Groups 1 619.71 7.17 16.04 FCT = 0.160 0.11–0.20

Populations 32 3211.03 16.58 37.10 FSC = 0.442

Individuals 129 2701.48 20.94 46.86 FST = 0.531 Guadalquivir River [R1–R24] [R25–R34]

Groups 1 669.53 8.41 18.44 FCT = 0.184 0.13–0.22

Populations 32 3161.21 16.25 35.64 FSC = 0.437

Individuals 129 2701.48 20.94 45.92 FST = 0.541 H. salzmanniana [S1–S14]

Populations 13 2065.29 13.35 34.45 FST = 0.344 Individuals 126 3200.30 25.40 65.55 Strait of Gibraltar [S1–S8], [S9–S14]

Groups 1 468.80 4.90 11.93 FCT = 0.119 0.08–0.15

Populations 12 1507.88 10.76 26.22 FSC = 0.298

Individuals 126 3200.30 25.40 61.86 FST = 0.381 Loukos River [S1–S10], [S11–S14]

Groups 1 557.41 7.56 17.58 FCT = 0.176 0.13–0.21

Populations 12 1507.88 10.03 23.33 FSC = 0.283

Individuals 126 3200.30 25.40 59.09 FST = 0.409

ACKNOWLEDGEMENTS REFERENCES

This work was supported by a pre-doctoral grant to M.A´ .O. Beismann, H., Barker, J.H.A., Karp, A. & Speck, T. (1997) (BES-2003–1506), a Juan de la Cierva grant to K.T., a post- AFLP analysis sheds light on distribution of two Salix spe- doctoral grant to A.T. (Proyectos de Investigacio´n de Excelencia, cies and their hybrid along a natural gradient. Molecular Junta de Andalucı´a), and grants to S.T. (REN2002–04634–C05– Ecology, 6, 989–993. 03 and CGL 2006-00817) and M.A. (REN2002–04354–C02–02 Bonin, A., Bellemain, E., Bronken Eidesen, P., Pompanon, F., and CGL 2005-01951) from the Ministerio de Educacio´ny Brochmann, C. & Taberlet, P. (2004) How to track and Ciencia (Spain), Junta de Andalucı´a (group RNM-204) and assess genotyping errors in population genetics studies. Fundacio´n BBVA (BIOCON 04), and to T. F.S. (P15225) from Molecular Ecology, 13, 3261–3273. the Austrian Science Foundation (FWF). The authors are Busack, S.D. (1986) Biochemical and morphological differen- indebted to F. Ehrendorfer, S. Castroviejo, S. Ortiz, L. Villar tiation in Spanish and Moroccan populations of Discoglossus and C. Mix for the sample collections. They also thank and the description of a new species from southern Spain G. Kadlec for precious help in the laboratory and the Bioinfor- (Amphibia, Anura, Discoglossidae). Annals of Carnegie matics Laboratory of the University of Sevilla General Services Museum, 55, 41–61. (CITIUS).

1394 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

Cerbah, M., Coulaud, J., Godelle, B. & Siljak-Yakovlev, S. (1995) Garcı´a-Parı´s, M., Alcobendas, M. & Alberch, P. (1998) Influ- Genome size, fluorochrome banding, and karyotype evolu- ence of the Guadalquivir River Basin on mitochondrial DNA tion in some Hypochoeris species. Genome, 38, 689–695. evolution of Salamandra salamandra (Caudata: Salaman- Cerbah, M., Coulaud, J. & Siljak-Yakovlev, S. (1998) rDNA dridae) from southern Spain. Copeia, 1, 173–176. organization and evolutionary relationships in the genus Garcı´a-Parı´s, M., Alcobendas, M., Buckley, D. & Wake, D.B. Hypochaeris (Asteraceae). Journal of Heredity, 89, 312–318. (2003) Dispersal of viviparity across contact zones in Iberian Collina-Girard, J. (2001) L’Atlantide devant le de´troit de populations of fire salamanders (Salamandra) inferred from Gibraltar? Mythe et ge´ologie. Earth and Planetary Sciences, discordance of genetic and morphological traits. Evolution, 333, 233–240. 57, 129–143. Comes, H.P. & Abbott, R.J. (1998) The relative importance of Hurlbert, S.H. (1971) The nonconcept of species diversity: a historical events and gene flow on the population structure critique and alternative parameters. Ecology, 52, 577–586. of a Mediterranean ragwort, Senecio gallicus (Asteraceae). Huson, D.H. & Bryant, D. (2006) Application of phylogenetic Evolution, 52, 355–367. networks in evolutionary studies. Molecular Biology and Corander, J. & Marttinen, P. (2006) Bayesian identification of Evolution, 23, 254–267. admixture events using multilocus molecular markers. Jong, H. (1998) In search of historical biogeographic patterns Molecular Ecology, 15, 2833–2843. in the western Mediterranean terrestrial fauna. Biological Corander, J., Waldmann, P. & Sillanpaa, M.J. (2003) Bayesian Journal of the Linnean Society, 65, 99–164. analysis of genetic differentiation between populations. Mousadik, A. & Petit, R.J. (1996) High level of genetic dif- Genetics, 163, 367–374. ferentiation for allelic richness among populations of the Corander, J., Waldmann, P., Marttinen, P. & Sillanpaa, M.J. argan tree [Argania spinosa (L.) Skeels] endemic to Mor- (2004) BAPS 2: enhanced possibilities for the analysis of occo. TAG. Theoretical and Applied Genetics, 92, 832–839. genetic population structure. Bioinformatics, 20, 2363–2369. Mraz, P., Gaudeul, M., Rioux, D., Gielly, L., Choler, P. & Despre´s, L., Gielly, L., Redoutet, B. & Taberlet, P. (2003) Using Taberlet, P. (2007) Genetic structure of Hypochaeris uniflora AFLP to resolve phylogenetic relationships in a morpho- (Asteraceae) suggests vicariance in the Carpathians and logically diversified plant species complex when nuclear and rapid post-glacial colonization of the Alps from an eastern chloroplast sequences fail to reveal variability. Molecular Alpine refugium. Journal of Biogeography, 34, 2100–2114. Phylogenetics and Evolution, 27, 185–196. Mueller, U.G. & Wolfenbarger, L.L. (1999) AFLP genotyping Doyle, J.J. & Doyle, J.L. (1987) A rapid DNA isolation pro- and fingerprinting. Trends in Ecology and Evolution, 14, 389– cedure for small quantities of fresh leaf tissue. Phytochemical 394. Bulletin, 19, 11–15. Muellner, A.N., Tremetsberger, K., Stuessy, T. & Baeza, C.M. Ehrich, D. (2006) AFLPDAT: a collection of R functions for (2005) Pleistocene refugia and recolonization routes in the convenient handling of AFLP data. Molecular Ecology Notes, southern Andes: insights from Hypochaeris palustris (Aster- 6, 603–604. aceae, Lactuceae). Molecular Ecology, 14, 203–212. Estabrook, G.F. (2001) Vicariance or dispersal: the use of Nei, M. & Li, W.H. (1979) Mathematical model for studying natural historical data to test competing hypotheses of dis- genetic variation in terms of restriction endonucleases. junction on the Tyrrhenian coast. Journal of Biogeography, Proceedings of the National Academy of Sciences USA, 76, 28, 95–103. 5269–5273. Excoffier, L., Laval, G. & Schneider, S. (2005) Arlequin ver. 3.0: Oberprieler, C. (2002) Hypochaeris. Catalogue des plantes vas- an integrated software package for population genetics data culaires du Nord du Maroc, incluant des cle´sd¢identification analysis. Evolutionary Bioinformatics Online, 1, 47–50. (ed. by B. Valde´s, M. Rejdali, A. Achhal El Kadmiri, S.L. Jury Fennane, M. & Ibn Tattou, M. (1998) Catalogue des plantes and J.M. Montserrat), pp. 686–689. Consejo Superior de vasculaires rares, menace´es ou ende´miques du Maroc. Investigaciones Cientı´ficas, Madrid. Bocconea, 8, 5–243. Oberprieler, C. & Vogt, R. (2002) Hypochaeris arachnoidea Fo¨rther, H. & Podlech, D. (2003) Die Gattung Hypochaeris L. Poir., a hitherto neglected species in NW Africa. Willden- sect. Hypochaeris (Compositae) im westlichen Afrika. owia, 32, 231–236. Sendtnera, 8, 35–43. Ortiz, M.A´ ., Talavera, S., Garcı´a-Castan˜o, J.L., Tremetsberger, Gala´n de Mera, A. & Vicente Orellana, J.A. (1998a) Hypo- K., Stuessy, T.F., Balao, F. & Casimiro-Soriguer, R. (2006) chaeris angustifolia (Asteraceae): distribucio´n, ecologı´ay Self-incompatibility and floral parameters in Hypochaeris fitosociologı´a. Acta Bota´nica Malacitana, 23, 247–252. sect. Hypochaeris (Asteraceae). American Journal of Botany, Gala´n de Mera, A. & Vicente Orellana, J.A. (1998b) Lecto- 93, 234–244. typification of Hypochaeris leontodontoides and other notes Ortiz, M.A´ ., Tremetsberger, K., Talavera, S., Stuessy, T.F. & on the Hypochaeris laevigata group (Aseraceae). Taxon, 47, Garcı´a-Castan˜o, J.L. (2007) Population structure of Hypo- 115–116. chaeris salzmanniana DC. (Asteraceae), an endemic species Garcı´a-Parı´s, M. & Jockusch, E.L. (1999) A mitochondrial to the Atlantic coast on both sides of the Strait of Gibraltar, DNA perspective on the evolution of Iberian Discoglossus in relation to Quaternary sea level changes. Molecular Ecol- (Amphibia: Anura). Journal of Zoology, 248, 209–218. ogy, 16, 541–552.

Journal of Biogeography 36, 1384–1397 1395 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd M. A´ . Ortiz et al.

Ortiz, M.A´ ., Tremetsberger, K., Terrab, A., Stuessy, T.F., Suc, J.-P. (1984) Origin and evolution of the Mediterranean Garcı´a-Castan˜o, J.L., Urtubey, E., Baeza, C.M., Ruas, C.F., vegetation and climate in Europe. Nature, 307, 429–432. Gibbs, P.E. & Talavera, S. (2008) Phylogeography of the Talavera, S. (1991) Flora y polinizacio´n. IV Congreso Nacional invasive weed Hypochaeris radicata (Asteraceae): from de Apicultura. Diputacio´n de Arago´n, Zaragoza. Moroccan origin to world-wide introduced populations. Thompson, J.D. (2003) Plant evolution in the Mediterranean. Molecular Ecology, 17, 3654–3667. Oxford University Press, Oxford. Parker, J. (1975) Aneuploidy and isolation in two Hypochoeris Tremetsberger, K., Stuessy, T.F., Guo, Y.-P., Baeza, C.M., species. Chromosoma, 52, 89–101. Weiss, H. & Samuel, R.M. (2003a) Amplified fragment Patarnello, T., Volckaert, F.A.M.J. & Castilho, R. (2007) Pillars length polymorphism (AFLP) variation within and among of Hercules: is the Atlantic–Mediterranean transition a populations of Hypochaeris acaulis (Asteraceae) of Andean phylogeographical break? Molecular Ecology, 16, 4426–4444. southern South America. Taxon, 52, 237–245. Pimentel, M., Sahuquillo, E. & Catalan, P. (2007) Genetic Tremetsberger, K., Stuessy, T.F., Samuel, R.M., Baeza, C.M. & diversity and spatial correlation patterns unravel the bio- Fay, F. (2003b) Genetics of colonization in Hypochaeris geographical history of the European sweet vernal grasses tenuifolia (Asteraceae, Lactuceae) on Volca´n Lonquimay, (Anthoxanthum L., Poaceae). Molecular Phylogenetics and Chile. Molecular Ecology, 12, 2649–2659. Evolution, 44, 667–684. Tremetsberger, K., Talavera, S., Stuessy, T.F., Ortiz, M.A´ ., Pou, A. (1989) La Erosio´n. Ministerio de Obras Pu´ blicas y Weiss-Schneeweiss, H. & Kadlec, G. (2004) Relationship of Urbanismo, Madrid. Hypochaeris salzmanniana (Asteraceae, Lactuceae), an Que´zel, P. (1978) Analysis of the flora of Mediterranean and endangered species of the Iberian Peninsula, to H. radicata Saharan Africa. Annals of the Missouri Botanical Garden, 65, and H. glabra and biogeographical implications. Botanical 479–534. Journal of the Linnean Society, 146, 79–95. Ruas, C.F., Ruas, P.M., Matzenbacher, N.I., Ross, G., Bernini, Tremetsberger, K., Weiss-Schneeweiss, H., Stuessy, T., Samuel, C. & Vanzela, A.L.L. (1995) Cytogenetic studies of some R., Kadlec, G., Ortiz, M.A. & Talavera, S. (2005) Nuclear Hypochoeris species (Compositae) from Brazil. American ribosomal DNA and karyotypes indicate a NW African origin Journal of Botany, 82, 369–375. of South American Hypochaeris (Asteraceae, Cichorieae). Ruas, C.F., Weiss-Schneeweiss, H., Stuessy, T.F., Samuel, Molecular Phylogenetics and Evolution, 35, 102–116. M.R., Pedrosa-Harand, A., Tremetsberger, K., Ruas, P.M., Tremetsberger, K., Stuessy, T.F., Kadlec, G., Urtubey, E., Baeza, Schlu¨ter, P.M., Ortiz Herrera, M.A., Ko¨nig, C. & C.M., Beck, S.G., Valdebenito, H.A., Ruas, C.F. & Mat- Matzenbacher, N.I. (2008) Characterization, genomic zenbacher, N.I. (2006) AFLP phylogeny of South American organization and chromosomal distribution of Ty1-copia species of Hypochaeris (Asteraceae, Lactuceae). Systematic retrotransposons in species of Hypochaeris (Asteraceae). Botany, 31, 610–626. Gene, 412, 39–49. Turner, T.F., Trexler, J.C., Harris, J.L. & Haynes, J.L. (2000) Samuel, R., Stuessy, T.F., Tremetsberger, K., Baeza, C.M. & Nested cladistic analysis indicates population fragmentation Siljak-Yakovlev, S. (2003) Phylogenetic relationships among shapes genetic diversity in a freshwater mussel. Genetics, 154, species of Hypochaeris (Asteraceae, Cichorieae) based on 777–785. ITS, plastid trnL intron, trnL-F spacer, and matK sequences. Ve´la, E. & Benhouhou, S. (2007) E´ valuation d’un nouveau American Journal of Botany, 90, 496–507. point chaud de biodiversite´ ve´ge´tale dans le Bassin me´di- Sanmartı´n, I. (2003) Evolucio´n biogeogra´fica de los Pachy- terrane´en (Afrique du Nord). Comptes Rendus Biologies, 330, deminae (Coleoptera, Scarabaeoidea) mediante ana´lisis de 589–605. dispersio´n-vicarianza. Graellsia: Revista de Zoologı´a, 59, Vos, P., Hogers, R., Bleeker, M., Reijans, M., Van De Lee, T., 427–441. Hornes, M., Frijters, A., Pot, J., Peleman, J., Kuiper, M. & Schlu¨ter, P.M. & Harris, S.A. (2006) Analysis of multilocus Zabeau, M. (1995) AFLP: a new technique for DNA fin- fingerprinting data sets containing missing data. Molecular gerprinting. Nucleic Acids Research, 23, 4407–4414. Ecology Notes, 6, 569–572. Weiss, H., Stuessy, T.F., Grau, J. & Baeza, C.M. (2003) Scho¨nswetter, P. & Tribsch, A. (2005) Vicariance and dispersal Chromosome reports from South American Hypochaeris in the alpine perennial Bupleurum stellatum L. (Apiaceae). (Asteraceae). Annals of the Missouri Botanical Garden, 90, Taxon, 54, 725–732. 56–63. Stehlik, I., Schneller, J.J. & Bachmann, K. (2002) Immigration Weiss-Schneeweiss, H., Stuessy, T.F., Siljak-Yakovlev, S., Baeza, and in situ glacial survival of the low-alpine Erinus alpinus C.M. & Parker, J. (2003) Karyotype evolution in South (Scrophulariaceae). Biological Journal of the Linnean Society, American species of Hypochaeris (Asteraceae, Lactuceae). 77, 87–103. Plant Systematics and Evolution, 241, 171–184. Stuessy, T.F., Tremetsberger, K., Mu¨llner, A.N., Jankowicz, J., Weiss-Schneeweiss, H., Stuessy, T.F., Tremetsberger, K., Guo, Y.-P., Baeza, C.M. & Samuel, R.M. (2003) The melding Urtubey, E., Valdebenito, H.A., Beck, S.G. & Baeza, C.M. of systematics and biogeography through investigations at (2007) Chromosome numbers and karyotypes of South the populational level: examples from the genus Hypochaeris American species and populations of Hypochaeris (Astera- (Asteraceae). Basic and Applied Ecology, 4, 287–296. ceae). Botanical Journal of the Linnean Society, 153, 49–60.

1396 Journal of Biogeography 36, 1384–1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd Phylogeographic patterns in Hypochaeris section Hypochaeris

Weiss-Schneeweiss, H., Tremetsberger, K., Schneeweiss, G.M., Please note: Wiley-Blackwell is not responsible for the Parker, J.S. & Stuessy, T.F. (2008) Karyotype diversification content or functionality of any supporting materials supplied and evolution in diploid and polyploid South American by the authors. Any queries (other than missing material) Hypochaeris (Asteraceae) inferred from rDNA localiza- should be directed to the corresponding author for the article. tion and genetic fingerprint data. Annals of Botany, 101, 909–918. Yokoyama, Y., Lambeck, K., Deckker, P.D., Johnston, P. & BIOSKETCH Fifield, L.K. (2000) Timing of the Last Glacial Maximum from observed sea-level minima. Nature, 406, 713–716. The focus of the research team is on the study of the Zhang, L.-B., Comes, H.P. & Kadereit, J.W. (2001) Phylogeny evolutionary history of the genus Hypochaeris in relation to and Quaternary history of the European montane/alpine other plant groups, combining molecular studies with studies of endemic Soldanella (Primulaceae) based on ITS and AFLP morphology and reproductive biology. S.T. and T.F.S. formu- variation. American Journal of Botany, 88, 2331–2345. lated the research questions; M.A´ .O., K.T. and A.T. collected the data and prepared them for publication; J.L.G.-C. performed ´ SUPPORTING INFORMATION the statistical analyses; and M.A.O. and K.T. led the writing.

Additional Supporting Information may be found in the online version of this article: Editor: Christine Maggs Appendix S1 Species, geographical–Bayesian groups, popu- This paper stems from a contribution initially presented at the lations, localities, geographical coordinates, collector numbers conference Origin and Evolution of Biota in Mediterranean and sample sizes for analysed populations of Hypochaeris Climate Zones: an Integrative Vision, held in Zurich on 14–15 arachnoidea, H. glabra, H. radicata and H. salzmanniana. July 2007.

Journal of Biogeography 36, 1384–1397 1397 ª 2009 The Authors. Journal compilation ª 2009 Blackwell Publishing Ltd