Introduction
Information on propagation strategies of Mediterranean plant species is most relevant for ecosystem conservation, especially in the present context of climate change (Manso et al., Reference Manso, Pukkala, Pardos, Miina and Calama2014; Cochrane, Reference Cochrane2016). Plants living in Mediterranean ecosystems must cope with extreme summer droughts and frequent fires that play an important role in such ecosystems (Chaves et al., Reference Chaves, Pereira, Maroco, Rodrigues, Ricardo, Osorio, Carvalho, Faria and Pinheiro2002). Climate change might alter the hydrological cycle in the Mediterranean region (Mariotti et al., Reference Mariotti, Zeng, Yoon, Artale, Navarra, Alpert and Li2008), and models predict a reduction in total precipitation and drier summers (Christensen et al., Reference Christensen, Hewitson, Busuioc, Chen, Gao, Held, Jones, Kolli, Kwon, Laprise, Magaña Rueda, Mearns, Menéndez, Räisänen, Rinke, Sarr, Whetton and Solomon2007), as well as an increase in fire hazard (Moriondo et al., Reference Moriondo, Good, Durao, Bindi, Giannakopoulos and Corte-Real2006; Pausas et al., Reference Pausas, Llovet, Rodrigo and Vallejo2008; Moreno et al., Reference Moreno, Zavala, Martín, Millán, Settele, Georgiev, Grabaum, Grobelnik, Hammen, Klotz, Kotarac and Kuehn2010). Seed germination is subject to strong selection pressure and, consequently, is likely to be highly sensitive to climatic changes (Walck et al., Reference Walck, Hidayati, Dixon, Thompson and Poschlod2011). In light of these forecast changes, a better understanding of inter- and intra-population variation in seed germination is of high importance since plasticity in its response among populations and individuals will provide a buffer against climate change (Lacerda et al., Reference Lacerda, Filho, Goulart, Ribeiro and Lovato2004; Fernández-Pascual and Jiménez-Alfaro, Reference Fernández-Pascual and Jiménez-Alfaro2014; Hudson et al., Reference Hudson, Ayre and Ooi2015).
To produce seeds with different germination responses is the mechanism by which many plant species are able to adapt to changing environmental conditions, and it is a common strategy in Mediterranean wild species (Pérez-García, Reference Pérez-García1993; Kigel, Reference Kigel, Kigel and Galili1995; Pérez-García, Reference Pérez-García2009; Martínez-Fernández et al., Reference Martínez-Fernández, Martínez-García and Pérez-García2014; Cochrane et al., Reference Cochrane, Hoyle, Yates, Wood and Nicotra2014). Phenotypic variation in a trait can be the result of genetic and environmental influences. Populations from similar habitats may have differences in seed morphology or germination, and these differences can arise from local climatic factors during seed maturation and from the effect of maternal genotype (Fenner, Reference Fenner1992; Wulff, Reference Wulff, Kigel and Galili1995; Baskin and Baskin, Reference Baskin and Baskin2014). Furthermore, intra-population variation can be attributed to genetic differences among individual parent plants, even within a small geographic area (Pérez-García, Reference Pérez-García1993; Bewley and Black, Reference Bewley and Black1994; Qaderi and Cavers, Reference Qaderi and Cavers2002).
In this scenario, we examined the germination characteristics and seed morphology of wild populations and individuals of Phillyrea angustifolia L. (Oleaceae) – narrow leaf phillyrea, narrow-leaved mock privet, or evergreen privet – a small tree that grows in well-preserved Mediterranean habitats and plays an important role in the post-fire ecological dynamics (Herrera et al., Reference Herrera, Jordano, Lopezsoria and Amat1994; Vitale et al., Reference Vitale, Capogna and Manes2007). In recent years there has been a renewed interest in P. angustifolia plant production. It is increasingly used for landscaping purposes, for being thermophilic and low water demanding (De Marco et al., Reference De Marco, Gentile, Arena and De Santo2005), but also in restoration programmes, since many of its natural populations are under decline. Germination in P. angustifolia, both in nature and in nurseries, is poor, erratic, and too gradual, resulting in a set of plants at different physiological phases (Catalán, Reference Catalán and Catalán1991; Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Schupp, Green and Westcott2007). These trees produce blue-black drupes, containing a single seed enclosed by a lignified endocarp. Previous investigations in the Phillyrea genus indicate that the endocarp may mechanically interfere with the emergence of the radicle but that it is at the same time water permeable (Mira et al., Reference Mira, Veiga-Barbosa and Pérez-García2015b, Reference Mira, Arnal and Pérez-García2016). Therefore, and contrary to a previous suggestion (Takos and Efthimiou, Reference Takos and Efthimiou2003), seeds do not exhibit physical dormancy, defined as the result of a water-impermeable layer in the seed or fruit coat (Baskin and Baskin, Reference Baskin and Baskin2004).
By collecting seeds from wild populations across a range of habitats and relating seed characteristics to climate parameters we intended to learn how germination strategies have maximized establishment probabilities for the species. The general aim of this work was to evaluate intraspecific variation of germination behaviour in P. angustifolia seeds. The specific aims were to investigate whether seed germination in P. angustifolia: (1) varied among populations, (2) varied among individuals of a population and (3) could be correlated with traits related either to seed morphology or to environmental parameters.
Materials and methods
Seed collection
The studied wild populations of P. angustifolia are all located in the Iberian Peninsula. Fruits from populations 1 and 5 (P1 and P5) were collected by the authors in September 2012, fruits from populations 2 and 3 (P2 and P3) were provided by Semillas Montaraz S.A., those of population 4 (P4) by Forestal Catalana S.A., and those of population 6 (P6) by Viveros Municipales Devesa-Albufera. Samples from populations 2, 3, 4 and 6 were also collected in September 2012. These six populations represent a broad range of the habitats of this species in the Iberian Peninsula (Table 1). Climatic data were obtained from the Spanish State Meteorological Agency (AEMET), from meteorological observatories close to the population sites. The climate data (Table 1) are mean values of 5 years (2008–2012). Fruit collection was performed according to the following standards: sampling randomly and evenly from as many plants as possible, in order to capture the widest possible genetic diversity from the population, and never collecting more than 20% of the available fruits from each individual on the day of collection (ENSCONET, 2009).
Table 1. Collection data and climatic parameters of six Phillyrea angustifolia populations
Abbreviations: total precipitation per year (P annual), precipitation June–August (P summer), number of drought months (D), annual mean temperature (T), mean temperature of the warmest month (Tw), mean temperature of the coolest month (Tc), annual mean of daily thermal oscillation (Osc), number of frost months (F), Universal Transverse Mercator (UTM). Climatic data from Spanish State Meteorological Agency (AEMET).
In parallel, in the case of P1, fruit collection was carried out separately for eight individual trees, in order to study intra-population differences in germination. For each of the eight individuals in P1, at least 250 fruits were collected.
In all cases, the fleshy exocarp and mesocarp of the fruit were manually removed. To carry out the assays with naked seeds (seeds without endocarp), endocarp was completely removed using pliers and a scalpel. Seeds were stored for 1 month under laboratory conditions (at about 23°C, 35% relative humidity, in darkness) until the start of trials in October 2012.
Image analysis
Representative images of 25 seeds with endocarp from each population were obtained with a digital camera (Nikon Coolpix 4500) mounted on a copy stand table (Kaiser 5410 RS2) equipped with four lighting units. Light conditions were stable throughout the acquisition procedure to achieve reproducibility and comparability of all images. Measurements on the digital images were determined using software ImageJ version 1.43 (Rasband, Reference Rasband1997–2015) for the following parameters: area, perimeter, diameter maximum, diameter minimum, Feret ratio (diameter minimum/diameter maximum), shape factor (4π area/perimeter2) and roundness factor (4 area/π diameter maximum2). Feret ratio indicates an aspect ratio close to an equiaxed particle when values are close to 1; shape factor indicates maximum shape homogeneity when values are close to 1; and roundness indicates maximum circularity when values are close to 1.
Seed germination
For all germination trials, four replicates of 25 seeds were incubated in glass Petri dishes (9 cm diameter) on top of two sheets of filter paper previously moistened with 4 ml of distilled water. To avoid contamination, seeds were disinfected with 10% HCl for 5 min before starting the assay. Filter papers were re-wetted regularly with distilled water as required. Incubation conditions were 15°C with a 16-h photoperiod provided by cool white fluorescent tubes with an irradiance of 35 μmol m2 s–1. These incubation conditions were chosen as the most suitable for P. angustifolia seed germination in a previous study (Mira et al., Reference Mira, Veiga-Barbosa and Pérez-García2015b). Samples were checked daily and germinated seeds were counted and removed. The incubation period was 45 days for naked seeds and 65 days for seeds with endocarp. Emergence of the radicle was the criterion for germination. In all germination trials, non-germinated seeds were checked for viability by the Tetrazolium test. Seeds were cut in half and submerged in a 1% solution of tetrazolium chloride for 24 h in the dark at 25°C. The number of non-viable seeds (equal or less than 5% of the total seeds) and empty seeds (equal or less than 2%) were excluded from calculation of final germination percentage.
Seedling emergence
To determine the effects of different substrates on seed emergence, naked seeds from P1 were sown in three substrates: an equal volume mix of sand and vermiculite, soil from the original habitat, and sterilized soil from the original habitat. Soil was sterilized by autoclaving at 120°C and 1 atm for 1 h. For each treatment, 52 seeds were individually sowed in the cells of a seedbed. Cells were partially filled with gravel (previously disinfected with HCl at 10%) and then completed with the assayed substrate. Seedbeds were incubated in a greenhouse, with a minimum prefixed temperature of 10°C. Seedling emergence was measured (cotyledons visible at substrate surface), and seed vigour was evaluated measuring plant height and number of leaves after 4 and 6 months.
Data analysis
The statistical analyses of seed germination data were done using the approach proposed by Ritz et al. (Reference Ritz, Pipper and Streibig2013) with the ‘drc’ package (Ritz and Streibig, Reference Ritz and Streibig2005) for the software environment R (R Core Team, 2015). We considered a non-linear log-logistic model to relate the cumulative germination and to monitor time after initialization of the test:
$$F\left( t \right){\rm} = d/\left( {1{\rm} + exp\left[ {b\left( {{\rm log} \left( t \right){\rm} -{\rm log} \left( {{\rm }MGT} \right)} \right)} \right]} \right),$$
where d is the maximum germination percentage; MGT (mean germination time) is the time where 50% of the seeds that germinated during the experiment have germinated; and b is proportional to the slope of F at time t. The estimation of non-linear regression parameters was based on treating data as an event time; that is, considering the monitoring interval when seeds were germinated or the time interval of the entire experiment if they did not germinate. Thus we have a multinomial distribution across these intervals and this distribution was used to obtain the parameter estimates by maximum likelihood. The time-event model implemented in the ‘drc’ package allows parameter comparison among germination curves for different treatment groups.
Also, one-way factorial ANOVA was used to test differences among populations in morphometric parameters of seeds and also differences on vigour parameters of seedlings in different substrates. Where ANOVA indicated a significant effect (P < 0.01), a multiple comparison test was carried out using the Tukey test.
Finally, data were subjected to principal component analysis (PCA) to assess associations between variables and to detect population separation trends based on the variables analysed. Twenty-one variables were included in the analysis of each population using a correlation matrix: climatic parameters of the population habitat, seed morphometric parameters, and germination and MGT of seeds with endocarp and naked seeds. Principal component analysis allowed us to condense as much of the original information contained in the 21 variables in orthogonal variables called principal components (PCs), which were weighed linear combinations of the original variables (Meglen, Reference Meglen1992; Wenning and Erickson, Reference Wenning and Erickson1994). After data were normalized, the contribution of each variable to each PC was assessed. The adequacy of this analysis is verified by the amount of the total information of the original variables retained by the PCs. Therefore, the initial set of 21 variables become a reduced set of two new variables (PCs), which enabled the two-dimensional projection of the six populations on the PCs. Germination percentages were arcsine square-root transformed for PCA analysis. PCA analysis was performed using SPSS 20.0 (IBM Corporation, 2013).
Results
Inter-population variation
The morphometric analysis of seed samples with endocarp allowed us to obtain extremely precise data on endocarp size and shape (Table 2). Endocarp shape was subspherical in all populations, as determined by: Feret ratio (0.77–0.86), shape factor (0.81–0.87), and roundness (0.79–0.89). Shape factor values showed little differences among populations. Endocarp area, perimeter, diameter maximum and minimum were also similar in all populations except for P3, in which values were higher. However, Feret ratio and roundness values were significantly different among populations. Mass of seeds with endocarp in P5 was significantly lower than in the rest of the populations. Therefore, seeds with endocarp showed little variation in size or mass, but did show differences in circularity and aspect ratio: the seeds with endocarp of P2, P4 and P6 were more rounded, and those of P1, P3 and P5 were more elongated.
Table 2. Morphometric parameters of seeds with endocarp from six Phillyrea angustifolia populations
Values are means ± standard error (SE). In each column, the mean values followed by the same letters are not significantly different (P > 0.01).
The germination of P. angustifolia seeds from six different populations was studied (Table 3, Fig. 1). The germination of seeds with endocarp was slow (34–54 days, MGT), and significant differences on final germination percentages were observed (Table 3, Fig. 1). P1 and P5 showed the highest germination (91–87%), and P4 the lowest (3%). Germination of naked seeds (without endocarp) was significantly faster in P1, P5 and P6 (15–16 days, MGT) than in P2, P3 and P4 (19–22 days), although there were no significant differences of germination percentages (78–88%) among populations (Table 3). To further study the germination behaviour of P. angustifolia, seedling emergence was tested. To do so, naked seeds from P1 were incubated in a greenhouse in three different substrates. No seedling emerged before 100 days (Fig. 2). After 200 days, seedling emergence was 50% in sand and vermiculite, 29% in soil, and 8% in the sterilized soil. Although the number of emerged seedlings varied among treatments, no significant differences (P > 0.01) were detected among the vigour parameters: average plant height was 4.0 ± 0.3 cm after 4 months, and 7.3 ± 0.4 cm after 6 months; and average number of leaves was 4.6 ± 0.4 after 4 months, and 12.2 ± 0.6 after 6 months.
Figure 1. Germination time courses of Phillyrea angustifolia seeds with endocarp (A) and naked seeds (B) from six populations: P1 (●), P2 (
), P3 (
), P4 (Δ), P5 (
) and P6 (□). Values are the averages of four replicates.
Figure 2. Germination time courses for Phillyrea angustifolia naked seeds sown in three types of substrate: an equal volume mix of sand and vermiculite (●), soil from the original habitat (), and sterilized soil ().
Table 3. Seed germination curve parameters of Phillyrea angustifolia seeds from six populations
Values are maximum germination percentage (mean ± standard error) and mean germination time (MGT) ± standard error. Parameters were estimated by maximum likelihood from a non-linear log-logistic model with a multinomial distribution. Within a column, values followed by the same letters are not significantly different (P > 0.01). For each treatment, four replicates of 25 seeds were tested.
PCA based on 21 climatic, morphometric and germination variables allowed us to separate the studied populations. The first three PCs explained 90% of the observed variation. Contribution of variables to PCs is presented in Table 4. The first principal component (PC1) accounted for 40% of the variation. In this component, the most relevant variables were germination of seeds with endocarp, seed morphometric parameters (such as diameter maximum, Feret ratio and roundness), and habitat characteristic (such as precipitation and temperature in summer) (Table 4). The second principal component (PC2) accounted for 29% of the observed variation. In this component, the most relevant variables were also seed morphometric parameters (such as area, diameter minimum and shape factor) and habitat characteristic (such as altitude, temperature in winter and daily thermal oscillation). The third principal component (PC3) explained 21% of the observed variation. The variables with most relevance in this third component were habitat characteristic (such as annual precipitation, temperature, and number of frost months). As PC3 is not clearly related to seed characteristics, like germination or morphometric parameters, discussion referred to this axis is not relevant for this study and was omitted.
Table 4. Contribution of variables, variance and cumulative variance explained by the three principal components for climatic and seed morphometric parameters of six populations of Phillyrea angustifolia
Bold values are used to highlight those variables with a high contribution to each principal component.
Two-dimensional projections of populations as defined by PC1 and PC2 are shown in Fig. 3. PC1 separated P1, P3 and P5 from P2, P4 and P6. PC1 clustered populations P1, P3 and P5 towards the region of greater germination of seeds with endocarp, seeds with endocarp with higher values of diameter maximum, lower Feret and roundness ratio, and regarding the habitat: higher temperature of the warmest month, lower precipitation in summer (June to August) and longer drought. PC2 separated population P3 from the remaining populations, especially from P5. PC2 situates P3 towards the region of greater area and diameter minimum, and regarding the habitat: lower altitude, greater temperature of the coolest month and lower annual mean of daily thermal oscillation.
Figure 3. Projection of six populations of Phillyrea angustifolia on a bidimensional plane defined by principal components 1 (PC1) and 2 (PC2). Population numbers are indicated in Table 1.
Correlations (coefficient of determination and P-value) among germination, morphometric variables, and habitat parameters of P. angustifolia seeds belonging to six populations were calculated (Table S1). Strong correlations between germination of seeds with endocarp and both Feret ratio and roundness (0.97 and 0.94, respectively, P < 0.01) and also with mean temperature of the warmest month (0.91, P < 0.05), were observed. Germination of naked seeds also correlated with mean temperature of the warmest month (0.84, P < 0.05), and with annual mean of daily thermal oscillation (0.90, P < 0.05).
Intra-population variation
The germination of P. angustifolia seeds from eight individuals of P1 was studied (Table 5, Fig. 4). Great differences were detected regarding germination of seeds with endocarp, with some individuals showing low germination (6–10%) and others showing high germination (81–88%) (Table 5). Differences were also detected among naked seeds, with most individuals showing high germination percentages (72–95%), but seeds from individual #3 achieved only 35% germination. Germination speed (MGT) was also significantly different among individuals, with values between 12 and 16 days for naked seed, and between 24 and 41 days for seed with endocarp.
Figure 4. Germination time courses of Phillyrea angustifolia seeds with endocarp (A) and naked seeds (B) from eight individuals of the same population (P1): 1 (●), 2 (), 3 (), 4 (Δ), 5 (■), 6 (□), 7 (
) and 8 (
). Values are the averages of four replicates.
Table 5. Germination parameters of Phillyrea angustifolia seeds from eight individuals of the same population (P1) after 45 days of incubation at 15°C
Values are expressed by final germination percentages (mean ± standard error) and mean germination time (MGT ± standard error). In each column, the mean values followed by the same letters are not significantly different (P > 0.01).
Discussion
Phenotypic plasticity of seed traits will play a major role in plant responses to climate change (Nicotra et al., Reference Nicotra, Atkin, Bonser, Davidson, Finnegan, Mathesius, Poot, Purugganan, Richards, Valladares and van Kleunen2010; Reed et al., Reference Reed, Schindler and Waples2011). Our data provide useful information on inter- and intra-population variability of seed traits of a wild species. Phillyrea angustifolia is difficult to propagate in nurseries, with low, delayed and too gradual seed germination, due to the hardness of the lignified endocarp impeding radicle emergence. Previous work indicated that Phillyrea endocarp could be water permeable (Mira et al., Reference Mira, Veiga-Barbosa and Pérez-García2015b, Reference Mira, Arnal and Pérez-García2016) and, therefore, seeds do not have physical dormancy sensu stricto (as defined by Baskin and Baskin, Reference Baskin and Baskin2004). Nonetheless, the hardness of the endocarp mechanically interferes with the emergence of the radicle, being a major impediment for germination. When the endocarp is removed, we found that seed germination showed an average of 84% among populations. These germination results are in agreement with previous data on Phyllirea (García-Fayos et al., Reference García-Fayos, Gulias, Martínez, Marzo, Melero and Traveset2001; Herranz et al., Reference Herranz, Ferrandis, Copete, Duro and Zalacain2006) and on seeds collected from P1 in 2011 previously studied by Mira et al. (Reference Mira, Veiga-Barbosa and Pérez-García2015b). Therefore, germination of naked seeds seems to be high, fast and homogeneous among populations and little influence of climate or year of collection could be inferred. However, great variability in germination percentage and speeds was found among populations when seed germination was assayed in the presence of the endocarp. Moreover, inter-population variability of morphometric parameters of seeds with endocarp was detected. Feret ratio and roundness, parameters that evaluate circularity, were significantly different among populations in seeds with endocarp. Also, while shape factor and seed mass were similar in most accessions, some populations were significantly different in these parameters. Ruprecht et al. (Reference Ruprecht, Fenesi, Fodor, Kuhn and Tökölyi2015) have reported that more rounded seeds are more tolerant to fire. Conversely, in our work, less rounded P. angustifolia seeds are related to climatic conditions (dry and hot summer) that favour fire occurrence in natural habitats.
Differences in seed germination among populations could be explained on the basis of inter-population variability of endocarp hardness, as previously suggested (Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Schupp, Green and Westcott2007). In wild Mediterranean plants, inter-population differences in hardness and permeability of the dispersion unit, whether it be seed or fruit, have been previously detected, as in the seed coat of Capsicum annuum (Hernández-Verdugo et al., Reference Hernández-Verdugo, Oyama and Vázquez-Yanes2001) and Vicia villosa (Renzi et al., Reference Renzi, Chantre and Cantamutto2016) or in the pericarp of Onopordum nervosum (Pérez-García and Pita, Reference Pérez-García and Pita1989) and Onopordum acanthium (Pérez-García, Reference Pérez-García1993; Qaderi and Cavers, Reference Qaderi and Cavers2002). The inter-population variability of the endocarp could explain the difficulties faced by nurseries to propagate this plant (Catalán, Reference Catalán and Catalán1991; Traveset et al., Reference Traveset, Robertson, Rodríguez-Pérez, Dennis, Schupp, Green and Westcott2007) and the contradictory results obtained with germination pretreatments such as scarification with sulphuric acid (Bacchetta et al., Reference Bacchetta, Bueno, Sánchez, Fenu, Jiménez-Alfaro, Mattana, Piotto and Virevaire2008; Mira et al., Reference Mira, Veiga-Barbosa and Pérez-García2015b, Reference Mira, Arnal and Pérez-García2016). Likewise, field experiments showed that germination increased when fruits had been eaten by goats (Grande et al., Reference Grande, Mancilla-Leyton, Delgado-Pertinez and Martin-Vicente2013) but decreased when eaten by birds (Traveset et al., Reference Traveset, Rodríguez-Pérez and Pias2008).
Our results indicate that the origin of the plant material used in Mediterranean reforestation and forest restoration must be strongly considered. Standardization of plant propagation protocols should take into account the P. angustifolia intraspecific variation, and caution is advisable when considering the results of studies based on one single wild population. After germination, seedling emergence and plant development were slow with any of the substrates tested in this study. Substrate had no effect on plant growth but had a strong effect on seedling emergence. Sterilizing soil was detrimental for seedling emergence, suggesting that the treatment might have killed mycorrhizal fungi in the soil. However, the artificial substrate of sand and vermiculite produced higher seedling emergence percentages (50%) than soil (29%) and would be, therefore, recommended for P. angustifolia plant propagation. The better performance of the artificial substrate is probably due to its higher water adsorption characteristics, since it is been stated by Lloret et al. (Reference Lloret, Penuelas and Ogaya2004) that drier conditions can limit seedling emergence by decreasing germination in the related species P. latifolia.
Phenotypic variation of the hardness of the endocarp can be the result of maternal genetic and environmental influences (as reviewed by Hudson et al., Reference Hudson, Ayre and Ooi2015) during embryo formation and fruit dispersal. Seed germination, the morphology of seed with endocarp and climatic variables allowed us to discriminate the six populations studied. We found that seeds with an endocarp that can be more easily broken by the radicle were more elongated in shape and collected originally from warmer and drier habitats: summer temperatures over 32.3°C, average summer rainfall under 51 mm, and over 3.8 drought months. This observation was supported by a significant correlation between germination of seeds with endocarp and Feret ratio, roundness, mean temperature of the warmest month and precipitation in summer. According to the correlation obtained among seed germination and climatic parameters, it could be hypothesized that less dormant seeds are able to germinate earlier in spring. Then, more developed seedlings are better able to withstand hot and dry summer conditions.
To elucidate the maternal genetic effect on seeds traits, intra-population variability of seed germination was studied. Seed germination of naked seeds differed only slightly among individuals. However, our data show greater differences in seed germination among individuals when the hard endocarp is not removed before planting. Great intra-population variability in germination has been detected in several wild Mediterranean species, such as Erica australis (Cruz et al., Reference Cruz, Pérez, Velasco and Moreno2003), Ceratonia siliqua (Pérez-García, Reference Pérez-García2009), Astragalus gines-lopezii (Martínez-Fernández et al., Reference Martínez-Fernández, Martínez-García and Pérez-García2014), Senecio coiincyi (Martínez-García et al., Reference Martínez-García, Guerrero-García and Pérez-García2012), Gentiana lutea (Pérez-García et al., Reference Pérez-García, Varela and González-Benito2012), and several species of Cistaceae (Pérez-García, Reference Pérez-García1997; Tavşanoğlu and Çatav, Reference Tavşanoğlu and Çatav2012). In the Mediterranean climate, with variable and unpredictable environmental conditions, an important survival strategy of wild species would be variation in either germination requirements (Kigel, Reference Kigel, Kigel and Galili1995; Cruz et al., Reference Cruz, Pérez, Velasco and Moreno2003; Pérez-García et al., Reference Pérez-García, Hornero and González-Benito2003; Pérez-García, Reference Pérez-García2005; Pérez-García et al., Reference Pérez-García, Huertas, Mora, Peña, Varela and González-Benito2006; Pérez-García and González-Benito, Reference Pérez-García and González-Benito2012; Pérez-García et al., Reference Pérez-García, Varela and González-Benito2012; Baskin and Baskin, Reference Baskin and Baskin2014; Copete et al., Reference Copete, Herranz, Copete and Ferrandis2014; Martínez-Fernández et al., Reference Martínez-Fernández, Martínez-García and Pérez-García2014), seed dormancy (Pérez-García et al., Reference Pérez-García, Varela and González-Benito2012), seed longevity (Mira et al., Reference Mira, Estrelles, González-Benito and Corbineau2011a; Lazar et al., Reference Lazar, Mira, Pamfil and Martínez-Laborde2014; Mira et al., Reference Mira, Estrelles and González-Benito2015a) or seed characteristics (Pérez-García et al., Reference Pérez-García, Huertas, Mora, Peña, Varela and González-Benito2006; Mira et al., Reference Mira, González-Benito, Ibars and Estrelles2011b; Pérez-García and González-Benito, Reference Pérez-García and González-Benito2012). In nature, P. angustifolia flowering takes place in March to June, fruits ripen in September to October and are dispersed from September to March. Seeds germinate with early spring temperature, from February to April (Herrera et al., Reference Herrera, Jordano, Lopezsoria and Amat1994; Andrés, Reference Andrés, Castroviejo, Aedo, Cirujano, Laínz, Montserrat, Morales, Muñoz, Navarro, Paiva and Soriano2011), after physiological dormancy of the embryo is broken and, therefore, the embryo becomes strong enough to break the endocarp. Our results suggest that endocarp phenotype has a high heritability, affecting intra-population variability in seed germination. The time when physiological dormancy of the embryo is broken and subsequent germination occurs will vary greatly among populations and also among individuals, seed germination being progressive along the season. Therefore, P. angustifolia plasticity would be an adaptative response to the enviromental variation of the Mediterranean climate.
Conclusions
Germination of P. angustifolia naked seeds was similar among populations and individuals, but great differences could be detected in the germination of seeds with an endocarp. Those diferences could be related to climatic and morphometric parameters of the population of origin, with higher germination associated with populations producing elongated seeds and habitats with a more severe summer. The phenotypic variation found in the endocarp is under both strong maternal genetic and environmental control, as indicated by the high variability among individuals within one population and the significant correlation between climate variables and seed germination. Selection pressures have shaped a high phenotypic plasticity in the species, which would allow it to adapt rapidly to the future scenario of climate change in which most Mediterranean areas are expected to become more stochastic (IPCC, Reference Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Girma, Kissel, Levy, MacCracken, Mastrandrea and White2014). Finally, the origin of the plant material used in Mediterranean reforestation must be strongly considered, while standardization of plant propagation protocols should take into account the intra-specific variation of Mediterranean species.
Acknowledgements
The authors would like to thank Semillas Montaraz S.A., Forestal Catalana S.A., and Viveros Municipales Devesa-Albufera for the seed material, the Soil Science Department (Universidad Politécnica de Madrid) for the advice on the soil analysis, Miguel Ibáñez for the advice on the statistical analysis and Juan B. Martínez-Laborde and Carol C. Baskin for assistance with revising the manuscript.
Financial support
This work was supported by a grant from the Spanish Ministerio de Educacion y Ciencia [CGL2006-10536]. During the work, Alberto Arnal was supported by a Beca de Colaboración 2012/2013, MEC.
Conflicts of interest
None.
