Introduction

Holm oak (Quercus ilex L.) is the most abundant and representative Quercus tree species in the Mediterranean forest. In Spain, its species is covering an approximate surface of 2,039,563 ha, with estimated values of over 120 million of euros in fruit acorn production per year (Gea-Izquierdo et al. 2006). The Holm oak acorns are a major component in the feeding systems of many Mediterranean wild and livestock species. Moreover, it is also the basic feed ingredient for domestically bred high-quality meat pigs (Gea-Izquierdo et al. 2006), which has a relevant economic and social impact on the Iberian Peninsula (Soto et al. 2008). Nowadays, forest restoration and reforestation are high-priority objectives, and Holm oak is acquiring more interest for Mediterranean forestry, together with other evergreen species, namely Quercus suber L., Ceratonia siliqua L., and Olea europaea subsp. sylvestris Brot. Spain Holm oak reforestation programs have heavily increased in the last 10 years, overtaking the Pinus species, which were widely used in the past (Pausas et al. 2004). Only in Andalusia (Spain), more than 75,000 ha were reforested from 1993 to 1999, with the introduction of millions of new seedlings.

As natural, non-domesticated plant species, with great plasticity and phenotypic variability, a key challenge prior to massive clonal propagation is the establishment of techniques for the cataloging and selection of genotypes among provenances with high survival percentage and productivity under specific environmental conditions. The existence of a high variability and polymorphism in Quercus spp. has been concluded from data obtained by using anatomical, chemical, and enzyme activity indicators, as well as DNA marker techniques (Finkeldey and Matyas 2003; Gandour et al. 2007; Gömöry et al. 2003; Jiménez et al. 1999; Lumaret et al. 2005; Lumaret and Jabbour-Zahab 2009; Martín et al. 2009; Michaud et al. 1995; Toumi and Lumaret 1998). In this context, Spain holm oak populations have been showed to present a prolonged isolation and complex geographic patterns of genetic variation, being native and distributed through diverse forest areas (Lumaret et al. 2002; Petit et al. 2002). However, several problems of typification particularly in the wide-ranging population are still unresolved, and modern techniques for the cataloging and selection of genotypes among provenances are needed to improve phylogenetic analysis. A proteomic research program is being carried out in order to study variability in Andalusia Holm oaks (Echevarría-Zomeño et al. 2009; Jorge et al. 2005, 2006; Martín et al. 2009; Valero et al. 2011). In our previous publications, we attempted to contribute toward understanding the diversity and phylogeny by analyzing leaf and seed proteome of several populations. The results generally indicated an appreciable difference between Andalusia Holm oak populations. In the present study, we continue our efforts to characterize the morphometry and chemical composition from the acorns of thirteen Andalusia Holm oak populations. Acorn chemical composition of oaks has shown a valuable tool for assessing variability (Dodd et al. 1993; Özcan 2007; Rafii et al. 1991, 1993), but few studies have been carried out on acorn chemical composition of Andalusia Holm oak acorns from a natural variability point of view (Rodríguez-Estévez et al. 2009; Valero et al. 2011). Chemosystematic differentiation based on differences in acorn fatty acid composition between Italian and Spanish populations of Q. ilex and Q. rotundifolia yielded partial separations of the individual populations (Rafii et al. 1991). Such separation using acorn fatty acids for native and hybrid populations of Q. agrifolia and Q. wislizenii was also achieved (Dodd et al. 1993). French Mediterranean evergreen oak populations were reported to be intermediate and heterogeneous for the fatty acid profiles between Spanish and Italian oak, suggesting a zone of hybridization (Rafii et al. 1993).

On the other hand, near-infrared reflectance spectroscopy (NIRS) is a highly valuable method that could be adapted for rapid measurements of acorn chemical composition traits. It is a non-destructive, simple, reliable, and accurate technique, allowing a complete chemical analysis in a single experiment, including fiber, starch, protein, and fatty acids (Jiang et al. 2007; Kim et al. 2007; Niewitetzki et al. 2010; Stuth et al. 2003; Takahashi et al. 2010). NIR spectroscopy is a rapid and cost-effective method for forestry-breeding programs and can be successfully applied to the selection of ‘high yield’ trees (Schimleck et al. 2000), differentiation of species (Adedipe et al. 2008), and oil content (Sousa-Correia et al. 2007).

The present work was undertaken to explore Holm oak morphometric and chemical composition variability. To accomplish this aim, morphometric and chemical composition in the mature acorns of 13 Andalusia Holm oak was performed. Data were subjected to clustering analysis and were correlated with geographical and climatic conditions. These data have provided biochemical markers that help establish phylogenic associations between populations and also reveal potentially account as an alternative source for dietary nutrition.

Materials and methods

Plant material

The present study was performed with thirteen populations distributed throughout the Andalusia (southern Spain) region (Fig. 1). Geographical coordinates, altitude, mean annual precipitation, and mean maximum and minimum temperatures corresponding to each surveyed area are presented in Table 1. The populations sampled in this study are distributed geographically in three major Andalusia regions: south (SAA, RG, RMA, SCA, and BCA), northeast (PCO, CCO, VJA, and TJA), and northwest (GSE, APS, CTH and CHU). The climate is diverse and characterized by a mean annual precipitation of 277.9 mm (SAA) to 1,263.6 mm (BCA), a mean winter precipitation of 78.9 mm (SAA) to 466.6 mm (SCA), a mean autumn precipitation of 103.4 mm (SAA) to 431.3 mm (RMA), a mean monthly maximum temperature of 23.1°C (TJA) to 26.8°C (PCO), and a mean monthly minimum temperature of 4.4–11.5°C (RG). Acorns were harvested during November 2007 from 15 trees per population. Trees 100 m apart from each other were randomly selected.

Fig. 1
figure 1

Spanish (inside square) and Andalusia maps with points indicating the locations where the 13 populations were surveyed. For more data see Table 1

Full size image
Table 1 Geographical and climatic dates of the 13 populations of Quercus ilex subsp. ballota included in this study
Full size table

Acorn morphometry

Healthy acorns (20 units per tree) were used for size and weight determinations. Length and maximum diameter were measured in acorn digitized images by using appropriate software (Win Seed V3.10b, Regent Instruments Inc., Canada 1996). Fresh weight was determined in an analytical balance (Mettler AJ150).

Acorn flour preparation

Acorns (20 units per tree) were scarified with a knife by making transversal and longitudinal cuts, thus permitting the pericarp to be rapidly removed. Once peeled, cotyledons (including embryos) were triturated in a blade mill (Moulinex AD56 42) until a powder was obtained. Powder was weighed and dried in a forced-air drier at 45°C for 48 h, homogenized in a Waring Blender (LB20E, Waring Products, New Hartford, CT, USA), and sieved (1 mm), to obtain a fine homogeneous flour. Samples were stored at 4°C in a desiccator until NIRS analysis. To determine the pulp dry weight of acorn, we weighted 2 g of powder and then we dried at 103°C in a ventilation oven for 24 h. Finally, we weighted them in order to obtain the dried weight of the powder (dry matter).

NIRS-derived chemical data

Ash, total protein, fat, starch, sugars, palmitic acid (C16:0), stearic acid (C18:0), oleic acid (C18:1), linoleic acid (C18:2), digestibility, and energy were analyzed at the University of Cordoba NIRS Service (http://www.uco.es/servicios/scai/nir.html). The instrument employed for spectra collection was a Foss-NIRSystems 6500 System II spectrophotometer (Foss-NIRSystems Inc., Silver Spring, MD, USA) equipped with a transport module and autogain detectors; one from 400 to 1,100 nm (known as the VIS region) and another from 1,100 to 2,500 nm (known as the NIR region). Samples were scanned by reflectance using a 1/4 rectangular cup, and spectra were collected every 2 nm using WinISI software 1.50 (Infrasoft International, Port Matilda, PA, USA). Before recording the spectra, the samples were thermostated at 24°C. The reflectance (log 1/R) spectra were collected in duplicate, and the mean spectrum of each sample was used for data analysis to determine relationship to reference characteristics. The scatter correction of standard normal variant and detrend (SNV-D) was applied to spectra, along with several different maths treatments for derivative order number, gap, and first smoothing. The second smoothing was set at 1 to indicate no second smoothing. Principal component analysis was used to identify and remove spectral outliers. Samples having spectra with Mahalanobis distance (H) values greater than 3.0 were considered to be outliers and were removed from the file. NIRS spectra were matched with reference data and analyzed using modified partial least squares (MPLS) regression, maths treatment 2, 5, 5, 1, and WinISI IIv. 1.04 (Infrasoft International, Port Matilda, PA) software. The standard error of calibration (SEC), the coefficient of determination in calibration (RSQ), the standard error of cross validation (SECV), coefficient of determination of the cross validation (1-VR), and RPD [standard deviation of laboratory reference data divided by the standard error of prediction (SEP)] were determined and used to find calibration equations that were acceptable for quantitative prediction (Deaville et al. 2009; Foss North America 2008; Shenk and Westerhaus 1990; Williams and Sobering 1996).

NIRS analysis

The descriptive statistics including mean, standard deviation (SD), and range, for chemical composition of acorn flour samples used in the calibration and validation sets, are shown in Table 2. Mean values of the dry matter were 2.13% of ash, 5.80% of total protein, 9.52% of crude fat, 61.51% of starch, 6.09% of sugar, 15.45 of palmitic acid, 2.97% of stearic acid, 64.20% of oleic acid, 15.66% of linoleic acid, 72.68% of digestibility, and 4,706 kcal of energy.

Table 2 NIRS calibration statistics of acorn chemical composition used for predictions
Full size table

The standard error of calibration (SEC), the coefficient of determination in calibration (RSQ), the standard error of cross validation (SECV), coefficient of determination of the cross validation (1-VR), and RPD [standard deviation of laboratory reference data divided by the standard error of prediction (SEP)] were determined. The statistics of calibrations and cross-validations are shown in Table 2. For each trait, we selected the highest RSQ and 1-VR (closest to 1) and the lowest SEC and SECV. However, the RPD statistic is considered to be a useful indicator for evaluating the prediction capacity of an equation to predict the component of interest. The higher the RPD values, the greater the probability of the model to predict new samples accurately. If the RPD is >3, the calibration models are considered as being acceptable for analytical purposes (Williams and Sobering 1996; Williams 2001). According to Williams and Sobering (1996), calibration models with an RPD value of between 2 and 3 indicate approximate quantitative predictions. In this context, the equation models for total protein, crude, fat, sugar, and linoleic acid had high values of RSQ (0.96, 0.99, 0.98, and 0.94, respectively), 1-VR (0.95, 0.98, 0.97, and 0.88, respectively), and RPD (4.35, 7.77, 6.33, and 2.78, respectively), indicating a close relationship between reference values and NIRS estimated values. However, the models for ash, palmitic acid, oleic acid, and energy had relatively high values of RSQ (0.83, 0.89, 0.88, and 0.83, respectively), low 1-VR (0.79, 0.77, 0.77, and 0.81, respectively), and RPD (2.18, 2.08, 2.06 and 2.28, respectively), indicating that those were not strong models, but that they might be useful for screening purposes. Finally, the models for starch, stearic acid, and digestibility presented a relatively low RSQ (0.66, 0.81, and 0.74, respectively), relatively low 1-VR (0.79, 0.77, 0.77, and 0.81, respectively), and low RPD (1.65, 1.78, and 1.61, respectively), indicating that they were of a lesser quality, but that they might be useful for screening ends. Nevertheless, it has been considered to be convenient not to consider them capable of improving the quality of the corresponding models. Only the models for fat, sugar, total protein, linoleic acid, ash, energy, palmitic, and oleic acid were utilized for the characterization of Andalusia Holm oak populations.

Statistical analysis

Normality of frequency distributions was tested by the Kolmogorov–Smirnov test. A one-way ANOVA was performed to test for differences in acorn morphometry and chemical data. Pearson’s correlation was carried out to test for correlation between acorn size, chemical composition, geographical, and climate data. Ward’s clustering method using squared Euclidian distances was used with the aiming of classifying and establishing distances between acorn morphometry and chemical data. Statistical analysis was conducted using SPSS v.8.0 software (SPSS Inc. Chicago IL, USA).

Results

Acorn morphometry and weight

Table 3 shows morphometry variations between seed collection sites. Statistic significant (P < 0.05) differences were observed in acorn weight, length, and maximum diameter between thirteen seed lots. Acorn weight ranged from 2.41 g (RMA population) to 6.12 g (PCO population). Acorn length ranged from 40.35 mm (GSE population) to 22.81 mm (RMA population), while acorn maximum diameter ranged from 17.06 mm (PCO population) to 12.35 mm (RMA population). Coefficient of variance (CV) showed that all three acorn sizes were not equally variable between the populations. Acorn weight showed a CV of 30%, and, depending on the population, it ranged from 22% (APS population) to 37% (CCO population), while maximum diameter showed a CV of 18%, and, depending on the population, it ranged from 35% (CHU population) to 13% (RMA population). However, length showed a CV of 9%, and, depending on the population, it ranged from 17% (CHU population) to 3% (RMA population). Acorn weight, length and diameter showed statistic significant (P < 0.01) inverse correlation with altitude, and positively with latitude range of seed source (Table 4; Fig. 2). When the acorn size was correlated with latitude, two geographical groups were obtained, with the southern populations (RMA, SAA, CCO, BCA, and SCA) showing lower values than the northern ones (TJA, GSE, CTH, CHU, APS, and PCO) (Fig. 2a–c). Similar results were observed when the acorn size was correlated with altitude data of population localities (Fig. 2d–f). In this direction, populations located geographically in the northern areas would be expected to have higher acorn morphometry and weight than southern populations. Moreover, it would decrease with the altitude location.

Table 3 Acorn size and weight
Full size table
Table 4 Correlationship between acorn size, weight, chemical determinations, and geographical (altitude, latitude, and longitude) and climatic (annual precipitation (1), mean monthly maximum (2) and minimum (3) temperature) data
Full size table
Fig. 2
figure 2

Correlation between acorn morphometry and geographical data. a Weight and latitude, b length and latitude, c diameter and latitude, d weight and altitude, e length and altitude, and f diameter and altitude. Pearson’s correlation coefficient is indicated with level of significance (P ≤ 0.05). Population names are indicated in Table 1

Full size image

Acorn chemical data

Table 5 shows chemical composition variations between seed collection sites. Statistic significant (P < 0.05) differences were observed in pulp dry weight, ash, total protein, sugar energy, fat, oleic acid, linoleic and palmitic acids between thirteen seed lots. Pulp dry weight ranged from 63.90% (RMA) to 83.64% (APS); ash ranged from 1.34% (CCO) to 2.02% (PCO); total protein ranged from 3.90% (RMA) to 5.94% (PCO); sugar ranged from 8.95% (PCO) to 12.47% (RG); energy ranged from 19,934.48 kJ/kg (SSA) to 20,398.66 kJ/kg (PCO); fat ranged from 9.14% (CCO) to 14.95% (GSE); oleic acid ranged from 64.98% (RMA) to 67.81% (GSE); while linoleic acid ranged from 12.15% (CCO) to 14.65% (CHU) and palmitic acid ranged from 15.62% (SCA) to 17.14% (CCO). Coefficient of variance (CV) showed that acorn composition did not vary equally between the populations. Crude protein presented a CV of 17%, and, depending on the population, it ranged from 10% (RMA) to 23% (GSE); sugar showed a CV of 16% and ranged from 7% (CCO) to 35% (PCO); fat showed a CV of 14% and ranged from 9% (GSE) to 23% (RMA). However, pulp dry weight, energy, palmitic and oleic acid were mostly homogenous between the populations (CVs < 10%). Table 4 shows the correlation between chemical composition and acorn size, climate conditions and geographical distribution. Pulp dry weight was positively correlated (P ≤ 0.05) with length (r = 0.54). Ash content was positively correlated (P < 0.05) with weight (r = 0.88), length (r = 0.84), diameter (r = 0.84), latitude (r = 0.60) and negatively correlated with altitude (r = −0.66) (Fig. 3a, b, c). Fat content was positively correlated (P < 0.05) with weight (r = 0.76), length (r = 0.87), diameter (r = 0.68) (Fig. 3d, e, f) and negatively correlated with altitude (r = −0.80).

Table 5 Chemical composition of the different population of Quercus ilex subsp. ballota
Full size table
Fig. 3
figure 3

Correlation between acorn morphometry and acorn chemical composition. a ash and weight, b ash and length, c ash and diameter, d crude fat and weight, e crude fat and length, f crude fat and diameter. Pearson’s correlation coefficient is indicated with level of significance (P ≤ 0.05). Population names are indicated in Table 1

Full size image

Phylogenetic analysis

Morphometry and chemical composition data of thirteen seed lots were subjected to clustering analysis in order to establish groups of populations and distances among them. The tree groups obtained by Ward’s clustering method using squared Euclidean distances can be observed in Fig. 4. Two clusters can be distinguished, with an average distance value (ADV) of 25. The first cluster, however, was then divided into two small clusters at ADV of 3.8. CTH, TJA, SCA, CHU, and PCO populations fell in one cluster with APS and GSE populations. The second principal was divided into two small clusters with an ADV of 5.3. SSA, VJA, and CCO fell in one cluster with BCA and RG (at ADV of 2.5), whereas RMA was the most distant inside this group (at ADV of 5.3). This analysis showed that northern populations (CTH, TJA, SCA, CHU, PCO, GSE, and APS) would be expected to have similar values in acorn morphometry and chemical composition. This group of population showed a tendency to present values highest in acorn weight (5.42 g vs. 3.21 g); acorn length (37.21 mm vs. 28.73 mm); acorn diameter (16.12 mm vs. 13.79 mm); ash (1.88% vs. 1.55%); protein (4.82% vs. 4.38%); fat (12.55% vs. 9.92%); and oleic acid (66.93% vs. 66.4%). These populations grow in intermediate climate with an annual precipitation ranging between 635.7 and 1,000.5 mm and an altitude ranging between 184 to 643 m. On the contrary, southern populations (SSA, VJA, CCO, BCA, RG, and RMA) showed a tendency to present values lowest in sugar (10.28% vs. 11.25%); energy (20,076.71 kJ/kg vs. 20,149.26 kJ/kg); palmitic (16.11% vs. 16.39%) and linoleic acids (13.07% vs. 13.11%). Of these populations, SSA, VJA, CCO, and RG populations grow in xeric climates with an annual precipitation ranging between 277.9 and 593.3 mm and an altitude ranging between 891 and 1,241 m.

Fig. 4
figure 4

Graphical representation of the distances of acorn morphometry and chemical compositions, established after Ward’s clustering method analysis using squared Euclidian distance

Full size image

Discussion

Morphometry characteristics of Holm oak acorn exhibited variation at the population level (Table 3). These variations in acorn size were influenced by altitudinal and latitudinal variation (Table 4; Fig. 2). Populations from the southern sites had the smallest average seed size, while northern ones had the longest and broadest acorns. Populations growing in lowest altitudes presented the longest and broadest acorns; on the contrary, populations growing highest altitudes presented the smallest and thinnest acorns. Populations variability with respect to acorn morphometry have earlier been reported in some Quercus spp. Seed length and weight exhibited significant inverse correlation with altitude of seed origin of Quercus glauca (Singh et al. 2010). Furthermore, acorn morphometry also has showed a positive correlation with high latitudes of seed origin of Quercus suber (Ramírez-Valiente et al. 2009). Aizen and Woodcock (1992) also found a positive relationship between acorn size and latitude in American Quercus species. Díaz-Fernández et al. (2004) established that the percentage of biennial acorns produced and their ripening time differed with latitudes across Q. suber populations. Bigger acorns lead to increased seedling growth, higher root and shoot ratios, which can improve the seedling performance by developing and maintaining deep and extensive root system accessing water from deep within the soil profile (Ramírez-Valiente et al. 2009; Singh et al. 2010). Selection for big acorns as a consequence of selection for tall seedlings is likely to be stronger in water-limiting environments, indeed, a positive relationship among acorn size; seedling height and summer survival have been reported for Q. ilex (Gómez 2004).

The development of a rapid, precise, robust, non-destructive, and economical method to evaluate quality components is of major interest to growers, processors, and breeders. NIR reflectance spectroscopy (NIRS) is routinely used for the prediction of quality traits in many crops. In this study, we have used this approach for determining the acorn chemical quality using an ample number of samples and chemical parameters. The calibration equations employed for total protein, fat, sugar, and linoleic acid showed high values of RQS, low SEC and SECV, and high RPD, similar to those of other seeds and oilseed crops (Fontaine et al. 2002; Sousa-Correia et al. 2007; Baye and Becker 2004; Kim et al. 2007; Pazdernik et al. 1997), with higher RQS (>0.9) and lower SEC values and high RPD. However, the models for ash, palmitic acid, oleic acid, and energy had relatively high values indicating that these were not strong models, but that they might be useful for screening purposes. The models for starch, stearic acid, and digestibility did not have good statistics for RSQ, 1-VR, and RPD, indicating the threshold where an equation was not useful. However, CV values were similar or better than those presented in other crops using the Official Analytical Chemists Methods (AOAC 2000).

Acorn chemical composition of Holm oak has been little studied, and data in the literature are variable, especially when samples are small. In this direction, we have applied the NIRS technique to predict a complete chemical analysis of acorn flour in an attempt to compare, catalog, and characterize natural populations of Holm oak from the Andalusia region. The result indicated that there were statistic significant differences in acorn chemical composition between the different Holm oak populations (Table 5). Our results for ash, protein, and sugar content are in agreement with the findings on Quercus ilex. However, crude fat, oleic and palmitic acids contents showed relatively higher levels considering previous findings (Charef et al. 2008; León-Camacho et al. 2004; Rodríguez-Estévez et al. 2009; Vázquez 1998). Furthermore, linoleic acid percent showed relatively lower levels with respect to previous findings on Quercus ilex (Charef et al. 2008; León-Camacho et al. 2004; Rodríguez-Estévez et al. 2009; Tejerina et al. 2010). The energy content of the acorns ranged from 19,934.48 ± 417.34 to 20,398.66 ± 449.32 kJ/kg, while no other results are reported in the literature.

The relationship of geographical and climate data with acorn chemical composition of Quercus ilex has not been studied in much detail. However, in this study, the population located in the north (CHU, CTH, GSE, and APS) showed a tendency toward presenting high contents of ash, fat, oleic and linoleic acid, while the population located in the south (SAA, RG, BCA, and RMA) showed a tendency toward presenting low contents of sugar and palmitic acid. When the populations were classified according to climate data, those that grew in the driest site (VJA, SAA, CCO, and GR) gave high contents in total protein, sugar, and palmitic acid, the populations growing in mesic sites (RMA, BCA, and TJA) showed high contents in energy and linoleic acid, and, finally, the populations growing in an intermediate climate (CTH, CHU, APS, SCA, and GSE) presented a high content in ash, fat, and oleic acid. Similar geographical pattern differentiation has been observed when these populations were used to study the natural variability through acorn protein profiles (Valero et al. 2011). Variations in fatty acid levels were observed in different Quercus spp. taxas (Charef et al. 2008; Özcan 2007). Fatty acid profiles were affected by intra-annual and inter-annual time sampling (Tejerina et al. 2010). Differences in fatty acid were observed in different Quercus spp. populations (Dodd et al. 1993; Rafii et al. 1991, 1993). The results obtained did not allow one to decide whether the observed differences in total unsaturated acids are genetic or not; the data indicated that the differences were most likely related to the climatic or ecological variations in elevation (Rafii et al. 1991). Variations in the content of chemical composition in the acorns could be related to factors associated with (1) weather conditions, such as precipitation, temperature, and temperature oscillations and (2) the acorns physiological conditions such as ripening or germination and (3) sanitary conditions such as desiccation, dampening, rotting, and attacks from pests, such as Curculio spp., Cydia spp. or phytophages (Daza et al. 2007; Steele et al. 1996; Vázquez 1998). However, cross-pollination, that would enable gene flow, would prevent genetic divergence via drift and may account for the lack of spatial genetic structure of neutral markers (O’Brien et al. 2007; Soto et al. 2003). If selection is strong enough, local adaptation may evolve even in the presence of gene flow rates between populations situated at a geographical proximity, thus contributing to a stronger phenotypic similarity than what would be observed from the influence of environmental variations alone (Endler 1977), therefore representing a potential mechanism for explaining the structured distribution of phenotypic variation. The results of ongoing population genetic and phylogeographic studies of Q. ilex using nuclear and chloroplast DNA microsatellites may help to discern between these alternatives (Jiménez and Gil 2000).

Conclusions

In conclusion, based on averages, eastern populations of Q. ilex subsp. ballota are different from western populations in their acorn morphology and chemical composition, and the western populations display a more overall variation. Mother trees grown in native environments can be distinguished from each other using those variables. It is possible to restrict the material selection areas in accordance with the valuation parameter considered, although, in practically all the populations, trees with high–low values could be suitable for use in reforestation can be found. Because most populations are widely dispersed, the species is susceptible to a loss of genetic diversity, mainly among eastern populations. Additional studies on genetic diversity in Q. ilex at the molecular level would help to create efficient and effective conservation strategies for this species.