Introduction

The genus Juniperus L., consisting of 101 taxa (infraspecific incl.; Adams 2011), is the second largest of living conifers (Miller 1977; Farjon 2010), surpassed only by Pinus L.. Predominantly distributed in the Northern Hemisphere (the only exception is J. procera Hochst. ex Endl., in Africa), it presents 15 taxa exclusive to the western Mediterranean region (Gauquelin 2006). The phylogeny of the genus shows the Mediterranean taxa scattered into several groups in derived position within Asian lineages (Mao et al. 2010), suggesting their diversification through five to six independent colonization events from Asia. Junipers are xerophytes of a special ecological importance since they are among the few—sometimes the only—trees able to grow in arid and semi-arid climates (Ciesla 2002). This is the case in the Mediterranean basin of J. excelsa M. Bieb. (Douaihy et al. 2011) and J. thurifera L. (Montes et al. 2002; Romo and Boratyński 2005). These Juniperus species are key components of “resource islands” (Reynolds et al. 1999), i.e. patches of remnant woodlands of locally improved soil and microclimatic conditions, plant facilitative interactions and plant dispersal (Maestre and Cortina 2005, and references therein). The positive impact of trees and shrubs on community structure and dynamics of semi-arid areas promoted the formation of small biodiversity hot-spots (Pugnaire and Lázaro 2000). Additionally, junipers represent an essential wood resource for human populations living in semi-arid Mediterranean area (Gauquelin et al. 1999; Montes et al. 2002; Romo and Boratyński 2005), and are used in local pharmacopeia (González-Tejero et al. 2008). Considering the ecological and economic importance of Mediterranean junipers, the increasing fragmentation of their habitats is becoming a concern. The fact is that the fragmentation of semi-arid mountain woodlands was primarily natural and driven by an increased aridity and warmer climates during the Holocene (Quézel and Médail 2003; Carrión et al. 2004, 2010; Fady and Conord 2010). Nevertheless, it is currently being worryingly increased by human impact, such as wood removal and intensive pasture activity (Gauquelin et al. 1999; authors’ field observations). The distribution of Juniperus across the semi-arid Mediterranean regions results from a compromise between altitude, where competition is avoided, and resources, which are scarcer as altitude increases and limits the reproductive efficiency (Montesinos et al. 2010). Global warming is increasing summer aridity, which is likely to increment pressure acting on junipers (Gauquelin et al. 1999, and references therein, for J. excelsa and J. procera; De Soto Suárez 2010, for J. thurifera).

The impact of habitat loss and fragmentation on species genetics and survival has mainly been observed over a short timescale since most cases available for direct observations, driven by anthropic pressure, are rather recent (e.g. Araucaria nemorosa De Laub., since 3,000 years ago; Kettle et al. 2011). The lack of temporal distance is making it difficult to draw inferences about future consequences of these processes. There is a current awareness of the need to widen the knowledge of biodiversity-shaping processes to the evolutionary timescale (e.g. Stigall 2010; Escarguel et al. 2011). In this sense, organisms allowing a larger time perspective could provide valuable data (e.g. Stigall 2010), which may be the case for Juniperus. The genus, which arose during the Eocene, experienced major diversification rates during the climatic changes of the Miocene onwards (Mao et al. 2010). Although preferring cold environments, their wide temperature tolerance (Gauquelin et al. 1999) and longevity (up to several hundred years; García and Zamora 2003) permitted juniper taxa to persist through time, and especially to overcome the Holocene warming. During this period, distribution ranges alternated contractions and expansions in function of climatic fluctuations and land usage, which have at some point favoured juniper steppe against pine forest (Riera Mora 2006) before becoming a threat (Gauquelin et al. 1999).

The present work focuses on Mediterranean junipers, with a special emphasis laid on J. thurifera, which belongs to Juniperus sect. Sabina, group IV (Mao et al. 2010). For comparative purposes, representatives of each section and infrasectional group reaching the western Mediterranean area were also considered. These are J. sabina L. (sect. Sabina, group III), J. phoenicea L. (sect. Sabina, group V), J. communis L. (sect. Juniperus, ‘blue seed cone’ group) and J. navicularis Gand. along with J. oxycedrus L. (sect. Juniperus, ‘red seed cone’ group) (Mao et al. 2010). Juniperus thurifera is a medium-sized, 8–15 (20) m tall, dioecious tree (Farjon 2005; Adams 2011), whose range covers western Mediterranean regions, mostly the Iberian Peninsula and Northwest Africa (Jalas and Suominen 1973; Quézel and Pesson 1980; Quézel and Barbero 1981; Greuter et al. 1984; Amaral Franco 1986; Charco 2001; Farjon 2005; Romo and Boratyński 2005; Adams 2011). The species is found from the Western Alps, Corsica and the Aurès mountains in Algeria, to the East, to the Pyrenees, Cantabrian range, Central Iberian range and the Moroccan Atlas Mountains to the West. These areas represent Pleistocene refugia of Tertiary floras (Carrión 2002; Benito Garzón et al. 2007; Thompson 2005; Médail and Diadema 2009), mostly as referred to the Ibero-Moroccan region, and to a lesser extent Corsica (Comes 2004; Médail and Diadema 2009).

Juniperus thurifera is a polymorphic species for which two subspecific entities are recognized: subsp. thurifera and subsp. africana (Maire) Gauquelin, Idr. Hass. & P. Lebreton ex Romo & Borat., as well as several chemovarieties of doubtful taxonomic value (Romo and Boratyński 2007; Adams 2011). The existence of the two subspecies has been confirmed through biometrical examinations (Gauquelin et al. 1988; Romo and Boratyński 2007), the correlation between biochemical diversity and number of seeds per cone (Adams et al. 2003) and genetic markers (Jiménez et al. 2003; Terrab et al. 2008).

While a pool of genetic data is being constituted for Mediterranean Juniperus (Jiménez et al. 2003; Meloni et al. 2006; Michalczyk et al. 2006; Michalczyk 2008; Boratyński et al. 2009; Douaihy et al. 2011; Dzialuk et al. 2011), comparatively very limited data are available regarding genomic aspects. These are mostly restricted to chromosome counts, all taxa being determined as diploids based on x = 11, with the exception of J. sabina, which presents both diploid and tetraploid cytotypes (Muratović et al. 2004, Siljak-Yakovlev et al. 2010). Genome size has been addressed in five Mediterranean species (J. excelsa, J. intermedia Schur, J. oxycedrus, J. phoenicea, J. sabina; Loureiro et al. 2007; Siljak-Yakovlev et al. 2010), and three other junipers (J. conferta Parl., J. rigida Siebold & Zucc. and J. virginiana L.; Hizume et al. 2001). Characterizing species from the genome size point of view is becoming essential as DNA content relates to morphology, ecology and phylogeny (Doležel et al. 2007), and its variation is thought to trigger taxonomic divergence (Kraaijeveld 2010). Additionally, genome size variants at intraspecific level have been shown to reflect, to a certain degree, the palaeovegetation history (e.g. Slovák et al. 2008) and also the more recent history (e.g. Pellicer et al. 2009).

In the present paper, we have established the karyological and genome size profiles of J. thurifera and other Mediterranean congenerics, with the aims of investigating whether: (1) interspecific genome size variation may respond to biological factors; (2) the genomic characteristics reflect the subspecific differentiation defined on the bases of genetic, biochemical and morphological studies; and (3) the genome size patterns may respond to habitat fragmentation within J. thurifera.

Material and methods

Species studied and collection of material

Efforts to germinate seeds collected in the field were mostly unfruitful, which led us to use root tips from cultivated plants of known origin from the Proyecto Forestal Ibérico (a forest tree nursery; www.proyectoforestaliberico.es) for chromosome number determination. Eight taxa were considered (Table 1). For genome size assessments, localities of the 43 studied populations—corresponding to six species and eight infraspecific entities—are indicated in Table 2. Juniperus thurifera populations were sampled across its currently fragmented distribution range. The species has not been extensively planted, and its seeds do not have a high dispersion potential (Santos et al. 1999), which guarantees a high preservation of the natural geographic pattern.

Table 1 Chromosome numbers reported for Juniperus individuals from the Proyecto Forestal Ibérico forest tree nursery
Full size table
Table 2 Juniperus populations studied
Full size table

Chromosome counts

Root tips were pretreated with 0.05 % aqueous colchicine for 3 h at room temperature or with 0.02 M 8-hydroxiquinoleine for 7 h at 16 °C or with a mixture of both for 4–5 h at 17 °C. The root tips were fixed into absolute ethanol/chloroform/glacial acetic acid (6:3:1) for 48 h at 4 °C and transferred to 70 % ethanol and stored at 4 °C. Samples were hydrolyzed in 1 N HCl for 20 min at 60 °C and subsequently stained with Schiff’s reagent during 30 min. Meristems were squashed on slides in a drop of 45 % acetic acid/glycerol (9:1) and covered with a coverslip. The best metaphase plates were photographed with a digital camera (AxioCam MRc5 Zeiss) mounted on a Zeiss Axioplan microscope, and images were analysed with Axio Vision Ac software version 4.2.

Genome size assessments

Five male and five female individuals were analysed for each dioecious population, and five individuals for each monoecious one. Samples were independently processed and measured twice. Triticum aestivum L. ‘Chinese spring’ (2C = 30.9 pg; Marie and Brown 1993 ) was used as internal standard. Fresh leaf tissue was chopped in a Petri dish with a razor blade in 1,200 μl of LB01 lysis buffer (Doležel et al. 1989) supplemented with Triton X-100 (8 % v/v) and 100 μg/ml ribonuclease A (RNase A, Boehringer, Meylan, France), together with the chosen internal standard. The resulting nuclei suspension was filtered through a 70 μm pore-size nylon mesh and subsequently stained with 36 μl of propidium iodide (for a final concentration of 60 μg/ml; Sigma-Aldrich Química, Alcobendas, Spain), kept on ice for 5–10 min and measured in an Epics XL flow cytometer (Coulter Corporation), the instrument set up with the standard configuration described in Garnatje et al. (2004). Measurements were made at the Centres Científics i Tecnològics de la Universitat de Barcelona.

Demographic indicators for populations

The two parameters observed were the age distribution and sex ratio within the population, which have been roughly estimated during fieldwork. The health state of the populations was defined on the basis of whether they were (good) or were not (bad) showing plants of different ages, including young seedlings. Sex ratio estimates were based on the observation of at least 30 adult trees for each population.

Statistical analyses

Shapiro–Wilk was used to test the data normality. One-way ANOVA was carried out to test the genome size variation with sex (male and female) between taxa. The monoecious populations were excluded from the analysis. Because both subspecies and varieties are represented in the Juniperus studied, comparison of genome size at infraspecific level was restricted to the species whose infraspecific entities were within the same taxonomic category. However, for J. thurifera, there were too few populations of subsp. africana to allow comparison with subsp. thurifera. Differences in genome size among both specific and infraspecific taxa were tested and illustrated with box and whisker plots. A least significant difference (LSD) test was carried out to compare the means. A dendrogram, based on the Euclidean distance and the Ward agglomeration method, including all J. thurifera populations, was constructed. Statistical analyses and graphics were elaborated with STATA 10.0 (Stata Corp. Texas, USA) and XLSTAT (Addinsoft NY, USA).

Results

Chromosome numbers and ploidy level determination

Chromosome counts indicate diploid cytotypes with 2n = 22 for all taxa, with the exception of J. thurifera subsp. thurifera, which presents a tetraploid cytotype (2n = 44, Table 1, Fig. 1). These results are the first counts for J. oxycedrus subsp. badia (H. Gay) Debeaux, J. phoenicea subsp. phoenicea, J. phoenicea subsp. turbinata (Guss.) Nyman (2n = 2x = 22) and for J. thurifera subsp. thurifera (2n = 4x = 44). All J. thurifera populations considered for genome size measurements, including subsp. africana, have approximately twice as much DNA amount as analysed congeneric species (Table 2, Fig. 2), this suggesting that tetraploidy probably extends to the whole species.

Fig. 1
figure 1

Somatic metaphase plates. a J. communis var. communis, 2n = 2x = 22. b J. communis var. saxatilis, 2n = 2x = 22. c J. oxycedrus subsp. badia, 2n = 2x = 22. d J. oxycedrus subsp. oxycedrus, 2n = 2x = 22. e J. phoenicea subsp. phoenicea, 2n = 2x = 22. f J. phoenicea subsp. turbinata, 2n = 2x = 22. g J. sabina, 2n = 2x = 22. h J. thurifera subsp. thurifera, 2n = 4x = 44. Scale bar = 10 μm

Full size image

Genome size quality and data distribution

The global mean in percentage of HPCV was 2.76 ± 1.05 for the target plant and 1.95 ± 0.92 for the standard, indicating methodological reliability. The Shapiro–Wilk test applied to diploid and tetraploid taxa separately showed that 2C values follow a normal distribution, allowing the use of parametric tests in statistical analyses. There were no significant differences in 1Cx between sexes, which led us to group the genome size measurements of male and female individuals from each population in further analyses.

Genome size and taxa

One-way ANOVA revealed a significant difference in 1Cx amongst the whole set of taxa, although no significant difference was found between J. communis and J. oxycedrus, and between J. phoenicea and J. sabina.

Within J. oxycedrus, one-way ANOVA shows significant 1Cx differences (P = 0.015), which discriminates after the LSD test J. oxycedrus subsp. badia from the other subspecies: subsp. macrocarpa (Sibth. et Sm.) Ball. (P = 0.046) and subsp. oxycedrus (P = 0.007). Within J. phoenicea, no significant differences were found between 1Cx of subsp. phoenicea and subsp. turbinata (P = 0.077). Except for J. thurifera, population sampling was not representative of the species’ whole distributions, and the results should be considered as preliminary.

Comparison of J. thurifera genome size throughout its geographical range

Genome size values assessed for the populations of J. thurifera are very homogenous, from 2C of 40.81 pg (J t15) to 43.2 pg (J t13), representing a 1.06-fold variation. However, significant differences (P < 0.0001) were found amongst them. Comparison of genome size and genetic structure patterns for populations represented in both Terrab et al. (2008) and the present study do not evidence overlap (Fig. 3a). However, in both cases, the maximum variability is found in Spain (Fig. 3). Genome size of J. thurifera populations does not follow a health-related (P = 0.155) or a sex ratio-related distribution (P = 0.442; Table 2). In turn, 87.5 % of age-biased populations (lacking young individuals) show an imbalanced sex ratio (lacking females), which is around ten times less frequent in populations of a normal age structure (9.09 %; Table 2).

Fig. 2
figure 2

Box and whisker plots of the genome size in studied juniper species and subspecies, with indication of sectional assignment. JCS J. communis var. saxatilis, JN J. navicularis, JOB J. oxycedrus subsp. badia, JOM J. oxycedrus subsp. macrocarpa, JOO J. oxycedrus subsp. oxycedrus, JPP J. phoenicea subsp. phoenicea, JPT J. phoenicea subsp. turbinata, JS J. sabina, JTA J. thurifera subsp. africana, JTT J. thurifera subsp. thurifera. At left, a branch of J. navicularis illustrates the acicular-leaved section Juniperus, at right, J. phoenicea for the scale-leaved section Sabina. Scale bars = 1 cm

Full size image
Fig. 3
figure 3

Distribution of populations of J. thurifera studied. a Populations from the present study (circles with grey line) and those of Terrab et al. (2008) (colour-filled circles) corresponding to genetic clusters: I (yellow), II (blue), III (green) and IV (red). b Populations from the present study. Genome size clusters resulting from Ward agglomeration method (see dendrogram) are represented by different grey intensities on the map. White: 2C > 41,343.88 Mbp, light grey: 41,343.88 ≥ 2C > 40,625.600 Mbp, dark grey: 40,625.600 ≥ 2C > 40,079.200 Mbp, black: 2C ≤ 40,079.200 Mbp

Full size image

Discussion

Genome size serves to adequately estimate ploidy level in Juniperus

In all the juniper species in which we determined the chromosome number, direct count was consistent with the one expected given the genome size value (Fig. 2). This suggests that estimation of the ploidy level through the flow cytometry technique could be an alternative to direct counting within the genus, which is limited by the availability of cultivated material of known wild origin, and the overall low germination rate of junipers.

Adding one more case of polyploidy in conifers

Our result concerning J. thurifera represents the first report of an exclusively polyploid species in the genus. Natural tetraploid cytotypes were previously found in rare occurrences in the mostly diploid species J. chinensis L. (J. chinensis ‘Pfitzeriana’; Sax and Sax 1933) and J. sabina (Muratović et al. 2004; Siljak-Yakovlev et al. 2010). Sporadically-occurring triploid and tetraploid cytotypes of J. chinensis, J. sabina, J. squamata Buch.-Ham. (J. squamata var. meyeri Rehder; Jensen and Levan 1941) and J. virginiana were preserved in nursery culture (Hall et al. 1973). Polyploidy, which is frequent in ferns, allied plants (95 % of species; Grant 1981) and angiosperms (70–80 % of species; Masterson 1994; Soltis et al. 2009, and references therein), is in turn rare or even absent from gymnosperm lineages (around 5 % of species; Khoshoo 1959; reviewed by Ahuja 2005). Apart from the three junipers mentioned above, natural polyploids are indeed only known otherwise in gymnosperms from Encephalartos hildebrandtii A.Br. & Bouché, triploid (Abraham and Mathew 1966), Ephedra L. (in 50 % of the species), Fitzroya cupressoides I.M.Johsnt., tetraploid, and Sequoia sempervirens Endl., hexaploid (Ahuja 2005; Ahuja and Neale 2005; and references therein). Therefore, Juniperus stands out for grouping three cases of natural polyploids, the most in any gymnosperm genus after Ephedra.

Due to the extreme rarity of polyploidy in gymnosperms, and the phylogenetic relevance of this plant group as the most ancient seed plants (Ohri and Khoshoo 1986), much attention has been paid to deciphering the origin of genome duplication in the few species concerned (Ahuja 2005, and references therein). In this sense, the J. chinensis tetraploid cytotype was shown to arise from the hybridization between J. chinensis and J. sabina (De Luc et al. 1999). No such data are available for J. sabina and J. thurifera. The only statement that may be made so far is the independent occurrence of the polyploidization events in junipers. Juniperus chinensis and J. sabina comprise both tetraploid and diploid cytotypes, this meaning that the genome duplication events happened within both species. Since the 19 populations of J. thurifera, measured throughout its total distribution, are tetraploid, polyploidy certainly happened early in the history of the species. The presence of diploid cytotypes in early- and late-diverging members of sect. Sabina group IV, to which J. thurifera belongs (respectively in J. procera, and J. chinensis plus J. procumbens Siebold), indicates that diploid cytotypes should be inferred at inner branches of this group (Mehra and Khoshoo 1956; Hall et al. 1973; Nagano et al. 2000 for chromosome counts; Mao et al. 2010, for phylogenetic framework). Taken together, this locates the genome duplication as concomitant with J. thurifera speciation, making the species the only one known in the genus to have undergone speciation through polyploidisation. Karyotype homogeneity—common in gymnosperms—impedes to determinate whether autopolyploidy or hybridization is involved in the formation of this tetraploid. The use of molecular cytogenetic techniques could shed light on this subject as they did in genera Cedrus Mill. and Pinus (Bou Dagher-Kharrat et al. 2001; Bogunic et al. 2011).

Strong homogeneity of genome size within Juniperus

This study contributes genome size values for six Juniperus species representing eight subspecific entities, and providing the first data for J. navicularis, J. thurifera subsp. africana and J. thurifera subsp. thurifera. Along with the previous reports (Hizume et al. 2001; Loureiro et al. 2007; Siljak-Yakovlev et al. 2010), our data account for a 1.7-fold variation of 2C within Juniperus, with the lowest value found in J. navicularis (2C = 19.18 pg) and the highest in J. thurifera (2C = 41.23 pg), and a 1.27-fold variation of 1Cx, with the lowest value found in J. navicularis (1Cx = 9.59 pg) and the highest in J. rigida (1Cx = 12.15 pg).

Although our results extend the range of 2C values of Juniperus at both lower and upper limits, Juniperus still fits well with the overall homogeneity of gymnosperms’ genome size (Ohri and Khoshoo 1986; Leitch et al. 2001), as its range is still below the 2.17-fold variation found between Pinus species, all of them being diploid (Grotkopp et al. 2004). Adaptive response to habitats has been evoked to explain high genome size values and its variation in Pinus (Wakamiya et al. 1993). Bogunic et al. (2007) reported a very narrow C-value range amongst five subspecies of Pinus nigra Arnold s.l., and concluded that infraspecific diversification occurred without change in genome size. Beside differences in habitats, Juniperus displays a reduced genome size range suggesting that environmental factors such as temperature, precipitation, etc. did not impact C-value in the taxa considered.

Previous analyses of gymnosperm chromosomes have revealed a number of unusual features that differentiate them from the remaining seed plants (reviewed by Murray 1998). Amongst them, large and relatively few-in-number chromosomes (producing high linkage disequilibrium, and, as a consequence, genetic hitchhiking), karyotypes of remarkably uniform structure, and, with few exceptions, no polyploidy, and a constant basic chromosome number for families. It has been argued that this low karyological dynamism could in part explain why gymnosperms were unable to diversify as much as angiosperms did (Gorelick and Olson 2011).

Genome size and leaf type

Mediterranean species with squamiform (scaled) leaves from our study were found to have significantly higher 1Cx values than species with needle-like (acicular) leaves (P = 0.006; Fig. 2). However, this result should be taken with caution because only six species were considered.

Deciphering whether genome size difference might be reflecting phylogeny rather than leave shape is necessary since Juniperus sections are based on leaf form character. Accordingly, acicular leaves are found in sect. Juniperus and its sister group sect. Caryocedrus, whereas sect. Sabina representatives are scale-leaved, with very few exceptions (J. coxii A.B.Jacks., J. morrisonicola Hayata, J. recurva Buch.-Ham. ex D.Don, J. saxicola Britton & P.Wilson, J. squamata, and sporadic individuals of J. barbadensis L. and other juniper species; Adams et al. 2008; Adams 2011). Acicular-leaved juvenile configuration is the rule in Cupressaceae, while its persistence in the adult plant, restricted to some junipers, is exceptional within the family (Little et al. 2004) and may illustrate cases of neoteny (e.g. J. saxicola; Adams 2011). In animals, neoteny has been related to increased genome size (Gregory 2002), but no such research has been done for plants, although it is well recognized that neoteny paid an important role in their evolution (Li and Johnston 2000, and references therein).

The fact that tetraploids—natural or not—exclusively occur in section Sabina (J. chinensis, J. sabina, J. squamata, J. thurifera, J. virginiana; see above), that is to say, are restricted to scale-leaved species, would also deserve more in-depth research.

Genome size and minimum seed-bearing age

A positive correlation between age of sexual maturity and genome size has been observed in Pinus (Wakamiya et al. 1993). The very few data available for Juniperus species show the minimum seed-bearing age not related to genome size, leaf type/section, but more likely to climatic conditions. The seed-bearing age of the Mediterranean species with acicular leaves, J. communis, is 17–25 years (Iszkulo and Boratyńsky 2011) and that of the also acicular J. oxycedrus is 17–21 years (Baldoni et al. 2004), whereas scale-leaved J. thurifera reaches sexual maturity at a mean age of 31, according to Montesinos et al. (2006). Since non-Mediterranean members of section Sabina reach sexual maturity much earlier, from the age of 10 (J. virginiana) or from the age of 12 years (J. monosperma Sarg., J. pinchotii Sudw., J. scopulorum Sarg.; Coder 2008), a possible interpretation could be that late offspring might be climate-constrained (delayed in unfavourable xeric conditions). Following this assumption, late offspring would have been reached independently by J. thurifera (together with polyploidy and increased genome size?) and by J. communis and J. oxycedrus (through a slower developmental rate that has, inter alia, given rise to the neotenic acicular leaves?). The pattern of low genome sizes found in Mediterranean J. communis and J. oxycedrus compared to non-Mediterranean species reminds the negative correlation between aridity and genome size previously established for pines, all diploids (Wakamiya et al. 1993).

Genome size and sex allocation

Contrariwise to those dioecious species in which small differences in DNA amount were found between males and females [Manoao colensoi (Hook) Molloy; Davies et al. 1997; possibly Ginkgo biloba L.; Murray 1998], no such trend is detected for dioecious junipers (P = 0.840). Genome size difference between sexes may be indicative of sex chromosomes (e.g. in Ginkgo biloba) that have been reported in a number of gymnosperms (Murray 1998 and references therein), but not in Juniperus.

Genome size and subspecific differentiation

No difference of genome size has been found between the well-differentiated J. thurifera subsp. africana and subsp. thurifera, from which the former underwent a morphological (Maire 1926; Gauquelin 1988; Romo and Boratyński 2007), chemical (Adams et al. 2003) and genetic (Jiménez et al. 2003; Terrab et al. 2008) divergence. Subspecific differentiation within J. thurifera resulted from the last opening of the Gibraltar Strait c. 5.33 Mya ago (Terrab et al. 2008). The Gibraltar Strait acted as an effective barrier against gene flow between Iberian and Moroccan populations for many species, allowing new species and even new lineages to arise in a number of plant groups (reviewed by Rodríguez-Sánchez et al. 2008; Jaramillo-Correa et al. 2010). Only subspecies were formed in the case of J. thurifera, although vicariant populations were genetically isolated since the last deglaciation started (Terrab et al. 2008), stating for a comparatively lower taxonomic divergence in J. thurifera than for other plant groups. Furthermore, a certain degree of genome size heterogeneity indicative of ongoing incipient speciation would have been expectable, but it was not found. This recalls the above-discussed low diversification rate of the gymnosperms, associated to their low karyological dynamism. A gene flow increase has been proposed to counterbalance the effects of population fragmentation in wind-pollinated plants with limited seed dispersal such as Juniperus, but testing this hypothesis has shown controversial results (Provan et al. 2008).

However, it has been evidenced that plants having experienced recent polyploidy diversify at a lower rate (Mayrose et al. 2011), and the fact that J. thurifera experienced a genome duplication concomitant to its arisal might also account for the comparatively low taxonomic divergence tendency observed within this species.

Genome size and habitat fragmentation

The knowledge of ecological traits and population dynamics of a species is considered to be crucial for predicting the effects of habitat fragmentation (Lindenmayer and Fischer 2006, and references therein). This first study of genome size in a species affected by habitat fragmentation, J. thurifera, reveals that C-values are not influenced by this factor, at least for the moment, in the same way that genetic diversity within populations remains high (Terrab et al. 2008). Juniperus thurifera populations are unfortunately good candidates to pay an extinction “debt” (Lindenmayer and Fischer 2006, and references therein), and their response to the landscape changes, although having possibly been delayed (due to the huge longevity of Juniperus trees), is already well visible through a very low regeneration rate and a deeply imbalanced sex ratio. Habitat fragmentation was shown to have altered the genetic range-wide structure of J. thurifera, with inter-population divergence higher than many other conifers, indicating a low connectivity between surveyed populations (Terrab et al. 2008). Equally worrying are the effects on population demography, especially with regard to the age and sex structure. Many populations are lacking seedlings and young individuals, pointing to an extremely low fertility (Montesinos et al. 2010). Furthermore, since most age-biased populations also show an imbalanced sex ratio and the scarcity of females reduces the overall seed production, the two factors probably aggravate each other. A critical point for regeneration might be reached, a situation that adult longevity may buffer temporarily (Montesinos et al. 2010). However, the fact that populations are not yet affected at genetic and genomic levels gives hope that recovery might be possible if measures for encouraging a successful reproduction are undertaken soon.