Science of the Total Environment 505 (2015) 724731

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Importance of soil and vineyard management in the determination of

grapevine mineral composition
M. Likar a,, K. Vogel-Miku a, M. Potisek a, K. Hanevi b, T. Radi b, M. Neemer c, M. Regvar a
a
b
c

Department of Biology, Biotechnical Faculty, University of Ljubljana, Vena pot 111, 1000 Ljubljana, Slovenia
Institute for Adriatic Crops and Karst Reclamation, Put Duilova 11, 21000 Split, Croatia
Joef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia

H I G H L I G H T S

Mineral compositions of grapevines in production vineyards were determined.

K, P and Mn explain 67% of total variance of the leaf mineral composition.
Soil K, Fe, Cu, organic matter and mycorrhiza have the strongest effect on plants.
Organic versus conventional management is an important factor.

a r t i c l e

i n f o

Article history:
Received 27 June 2014
Received in revised form 17 October 2014
Accepted 17 October 2014
Available online 5 November 2014
Editor: Charlotte Poschenrieder
Keywords:
Grapevine
Production vineyards
Mineral composition
Agricultural practice

a b s t r a c t
The spatial variability of the mineral composition of grapevines in production vineyards along the east Adriatic
coast was determined and compared between conventional and sustainable vineyard management. Cluster analysis
shows a high level of spatial variability even within the individual locations. Factor analysis reveals three factors with
strong loading for the macronutrients K and P and the micronutrient Mn, which explain 67% of the total variance in
the mineral composition. Here, 26% to 34% of the variance of these three elements can be explained by abiotic and
biotic soil parameters, with soil concentrations of K, Fe and Cu, organic matter content, and vesicular colonisation
showing the strongest effects on the mineral composition of the grapevines. In addition, analysis of the mineral composition data shows signicant differences between differently managed vineyards, with increased bioaccumulation
of P and K in sustainable vineyards, while Zn bioaccumulation was increased in conventional vineyards. Our data
conrm the importance of soil and vineyard management in the concept of terroir, and demonstrate the effects of
sustainable management practices on the mineral nutrition of grapevines that result from modied nutrient availability related to changes in the abiotic and biotic characteristics of the soil.
2014 Published by Elsevier B.V.

1. Introduction
In viticulture, the concept of terroir relates the sensory attributes of
wines to the environmental conditions of the grapes, and it therefore
represents an important descriptor of the connection between wines
and their origins. It encompasses both the natural factors of soil, climate
and topography, and the human role in vineyard management (Van
Leeuven and Seguin, 2006).
The soil is one of the most important factors of the terroir, which
makes it of special interest for the evaluation of the environmental
effects on the mineral composition of grapevines (Van Leeuven and
Seguin, 2006). This mineral composition reects the environment in
which the grapevines are cultivated (Bertoldi et al., 2011; Chopin
et al., 2008; Marengo and Aceto, 2003; Pessanha et al., 2010). As a
Corresponding author. Tel.: +386 1 3203418; fax: +386 1 2573390.
E-mail address: matevz.likar@bf.uni-lj.si (M. Likar).

http://dx.doi.org/10.1016/j.scitotenv.2014.10.057
0048-9697/ 2014 Published by Elsevier B.V.

consequence, the grapevine products (i.e., grapes, juice, wine) will be

inuenced by the composition of the soil (Rogiers et al., 2006; Pohl,
2007). The most important soil characteristics that affect the mineral
composition of grapevines are the soil pH (Jackson, 2008), the partitioning inside the soil (Chopin et al., 2008), and the Ca content
(Gruber and Kosegarten, 2001). In addition, biotic factors such as grapevine variety (Amors et al., 2011) and rootstock (Bavaresco et al., 2003;
Wooldridge et al., 2010), and soil microbiome, can also have profound
effects on the availability and uptake of minerals from the soil by the
grapevine. Vineyard soils also support indigenous arbuscular mycorrhizal
(AM) fungi (Likar et al., 2013; Oehl et al., 2005; Radi et al., 2014) that can
have positive effects on grapevine performance (Biricolti et al., 1997;
Linderman and Davis, 2001). Furthermore, low root density (Schreiner,
2005) and coarse root texture of grapevines suggest that they are highly
dependent on AM fungi (Eissenstat, 1992), making these an important
factor in the determination of the mineral composition of the grapevines.
This is especially true in organic vineyards that follow low-input

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

practices, as these support higher AM fungi spore abundance and diversity (Oehl et al., 2004).
In the ongoing effort to develop techniques for wine monitoring,
geochemical ngerprints would greatly improve the traceability of
wines to their origins, particularly as the mineral compositions of grapevines and their products are governed by the soil characteristics and the
cultural practices. As such, evaluation of the mineral composition of
grapevines from conventionally managed vineyards and vineyards
with organic practices under different environmental conditions should
greatly improve the monitoring system.
The main aim of the present study was to assess the inuence of the
soil on the mineral composition of grapevine leaves on the east Adriatic
coast and to evaluate the effects of conventional and sustainable agricultural practices on grapevine mineral composition. For this, grapevines of
the cultivar Mali Plavac (a red wine variety) were sampled in vineyards
along a 250-km-long transect, with the evaluation of: (i) the spatial
variability in the mineral composition of the grapevines in production
vineyards in the east Adriatic Karst region; and (ii) the effects of environment and vineyard management on the mineral composition of
the grapevines. Multivariate statistical techniques were applied to
evaluate the spatial variations in the mineral compositions of the grapevines. Furthermore, the effects of vineyard management practices
(i.e., conventional vs. sustainable) on the grapevine mineral compositions were tested, following the hypothesis that differences in cultivation
practices will inuence the mineral compositions of the grapevines
either directly (due to changes in soil characteristics) or indirectly (as
a result of improved microbiota diversity in organically managed
vineyards).
2. Materials and methods
2.1. Sampling locations
Eight production vineyards from four distinct localities (two
vineyards per wine region) in the Karst region of the east Adriatic area
were selected for sampling. The vineyards are located along a transect
of 250 km that stretches from north to south. The altitudes of the
vineyards range from 50 m to 350 m a.s.l. The climates of all of these
localities are of the Mediterranean type, with most of the annual precipitation in autumn and winter (8001300 mm), and with droughts very
often during the late spring and summer periods.
The bedrock is composed mainly of limestone or dolomite, and it is
covered with stony carbonate soils. The soil texture ranged from clay
loam (Ivan Dolac), to silty clay (Milna) and silt loam (Peljesac). At
Zadar conventionally managed vineyard was located on loam soil, whereas soil in sustainable vineyard had sandy loam texture (Table 1). The soils
have a pH of 7.9 to 8.2, with the exception of the Zadar location, where

725

the vineyard soil has a pH of 6.0 to 6.1. Organic matter content of the
soil ranged around 1.53%, but reached above 5% at the locations of
Ivan Dolac (both vineyards) and Peljesac (sustainable vineyard).
The grapevine coverage of the vineyards is 5% to 10%, with total plant
cover reaching 587% (according to averages of Braun-Blanquet classes).
2.2. Sample collection and analysis
Root samples of ve randomly chosen grapevines and their rhizosphere soil were collected to a depth of 20 cm to 30 cm at each of the
vineyards, in the Summer of 2010. The element compositions of the
grapevines were assessed according to the European criteria for assessment of grapevine elements, which are based on whole-leaf analysis, in
contrast to the American criteria where only the leaf petiole is analysed
(oga et al., 2008).
The soil samples were air dried, sieved to 2 mm, and ground to a ne
powder in an agate mortar. Total organic matter was measured by wet
combustion, according to Kandeler (1995). The plant-available phosphorous was extracted using 0.5 M NaHCO3, and determined photometrically
according to Olsen and Sommers (1982). The soil pH was measured in
soil:water (1:2, v/v) extracts, using deionised water. The soil mineral
composition was measured by energy dispersive X-ray uorescence
spectrometry in soil pellets pressed from 0.5 g powdered soil material,
which were prepared using a pellet die and a hydraulic press. For uorescence excitation, a [109]-Cd annular radioisotope source was used
(300 MBq, 22.1 keV; Isotope Products Laboratories, USA). An X-ray spectrometer was used that was based on a Si (Li) detector (EG and G ORTEC,
USA), with a 25-mm-thick Be window. The energy resolution of the
spectrometer at count rates below 1000 c/s was 175 eV at 5.9 keV.
X-ray uorescence measurements were performed in air, and each
sample was irradiated for 3000 s. The analysis of the X-ray spectra
was performed using AXIL programme (Van Espen and Jansen, 1993),
as included in the QAES software package (Vekemans et al., 1994;
Kump et al., 2007). Quality assurance for the element analysis was performed using standard reference materials: NIST SRM 1573a (tomato
leaves, homogenised powder); CRM 129 (hay powder); and OU-10
(geological sample of longmyndian greywacke, GeoPT24).
The mineral compositions of the leaves were determined by total
reection X-ray spectrometry. For TXRF about 0.1 g of the powdered
material was digested in 3 mL HNO3 in Teon vessels, using the CEM
microwave sample digestion system (MarsX Press). The cooled digests
were diluted to 10 mL with Milli-Q water, and spiked with standard solution of Ga as an internal standard to a nal concentration of 10 g/L
(Golob et al., 2005; Kump et al., 1996; Vogel-Miku et al., 2006). Two
aliquots of 10 L of the resulting sample solutions were deposited
onto quartz sample carrier plates and dried in a dessicator overnight.
For excitation, a focused X-ray beam from a ne-focus X-ray tube

Table 1
Soil physico-chemical characteristics for selected vineyards from four locations under different agricultural practice (conventional, sustainable), with altitude and plant cover (according to
Braun-Blanquet 1964).
Ivan Dolac

Plant available Pa (mg/g)

Ka (mg/g)
Caa (mg/g)
Mna (mg/g)
Fea (mg/g)
Zna (mg/g)
Cua (mg/g)
OMb
pH
Soil texture
Altitude (m a.s.l.)
Plant cover (%)
a
b

Milna

Peljesac

Zadar

Conventional

Sustainable

Conventional

Sustainable

Conventional

Sustainable

Conventional

Sustainable

13.3 3.09
18.4 1.96
69.5 9.08
1090 140
41.3 0.85
103 25
170 28
5.76 0.33
8.02 0.06
Clay loam
45
5

3.77 1.67
21.5 1.39
42.4 10.1
1173 121
43.9 1.85
111 19
124 6
4.09 0.48
8.15 0.05
Clay loam
60
5

1.94 0.2
17.0 2.03
121 10.1
872 179
37.8 1.82
89 9
117 33
1.62 0.18
8.18 0.07
Silty clay
100
87

1.77 0.42
17.1 1.64
15.0 17.1
657 48
35.9 2.27
87 16
132 17
2.03 0.25
8.05 0.08
Silty clay
100
87

4.58 0.62
8.28 2.10
160 15.1
569 71
13.3 2.68
73 27
480 119
2.81 0.17
8.05 0.14
Silt loam
90
5

0.28 0.11
9.00 2.47
123 41.2
511 31
20.5 5.78
63 20
83 28
9.39 2.54
7.87 0.05
Silt loam
100
5

5.77 0.62
12.8 2.66
105 12.5
487 54
18.8 3.36
51 11
131 7
1.44 0.21
6.00 0.16
Loam
300
5

0.37 0.08
10.2 2.44
44.1 1.62
379 41
22.8 3.19
41 9
51 2
1.73 0.21
6.10 0.08
Sandy loam
300
5

Concentration of the element per soil dry weight.

OM, organic matter content (%).

726

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

(Seifert, Germany) was used, with a Mo anode operating at 40 kV and

30 mA and monochromatised to the energy of the Mo Ka line
(17.4 keV) using a carbontungsten (C/W) multilayer. Each sample deposited on the optically at quartz carrier plates was irradiated for 300 s
in air at an incident angle lower than the total reection critical angle of
the substrate material ( 1.8 mrad). The X-ray spectra were detected
using a Si (Li) detector (EG and G ORTEC) with a 25-mm-thick Be
window. The energy resolution of the spectrometer at count rates
below 1000 c/s was 140 eV at 5.9 keV. The quantication was validated
by BCR CRM 129 Hay powder, as a certied reference material. The bioaccumulation factors (BAFs) were calculated as the ratio between the
leaf and the total soil element concentrations.
For estimation of the fungal colonisation, the grapevine roots were
washed, cleared in KOH, acidied, stained with Trypan blue, and stored
in glycerol. In this way, prepared fragments of lateral roots were
mounted on slides and examined under light microscopy (Phillips and
Hayman, 1970). The hyphal colonisation, arbuscular colonisation, vesicular colonisation, and microsclerotia colonisation (calculated as for AM
fungi colonisation, as microsclerotia instead of arbuscules) were determined according to the magnied intersection method (McGonigle
et al., 1990; see Online Resource 2). Seven hundred and fty root intersections were examined per vineyard to estimate the colonisation
parameters (150 root intersections per grapevine).
2.3. Data treatment and multivariate statistical methods
Prior to statistical analysis, the experimental data were standardised
and scaled using the scale function in R, to avoid misclassication due to
wide differences in data dimensionality (Liu et al., 2003). All of the
mathematical and statistical computations were carried out using the
R 2.15.1 software.
2.4. Heatmap and cluster analysis
Using the function heatmap.2 in R package gplots 2.10.1, a heatmap
was generated for the median proles in clusters. Using the function
hclust in R package 2.15.1, the dendrograms of rows or columns were
added on the margins of the heatmap. Hierarchical clustering with the
complete linkage method was used for the dendrograms, based on the
Euclidean distances among clusters in rows or among elements in
columns.
2.5. Principal component and factor analysis
Principal component analysis (PCA) was used on the leaf and soil
datasets for the reduction of the initial number of parameters into a
smaller set of independent (uncorrelated) variables (the principal components), allowing data reduction with minimum loss of original information (Helena et al., 2000; Vega et al., 1998).
Factor analysis was performed on the correlation matrix of the
rearranged data, so that it explained the structure of the underlying
dataset.
2.6. Regression and correlation analysis
Multiple regression analysis was used to examine relationships
between elements selected through the PCA/factor analysis, for their
maximal impact on the spatial variability of the mineral composition
of the grapevine leaves with the soil parameters.
The possibility that soil structure could account for the dissimilarity
of mineral composition of grapevine leaves was investigated using
Mantel test and Pearson correlation coefcient. Mantel tests were performed with the mantel function of the R ecodist package (Spearman's
rank correlation, 999 permutations) (Goslee and Urban, 2007). Bray
Curtis dissimilarity matrices (Bray and Curtis, 1957) of percentages of

different size classes of soil particles (for raw data see Radi et al.,
2014) and mineral concentrations in grapevine leaves were used.
2.7. Nonparametric multidimensional scaling and perMANOVA analysis
The compositional dissimilarity was estimated using the BrayCurtis
dissimilarity matrix of the mineral compositions between the samples,
and pair-wise dissimilarities between the samples were analysed
using a non-metric multidimensional scaling plot. Nonparametric
multidimensional scaling was performed with the metaMDS function
of the R Vegan package (Oksanen et al., 2010). To explore the relative
importance of the measured environmental factors for the mineral
compositions of the grapevine leaves, unconstrained ordination was
performed, with the subsequent tting of the environmental variables.
The function envt from the Vegan library (Oksanen et al., 2010) was
used to identify the variables that were signicantly correlated with
the mineral compositions of the grapevine leaves. The signicance of
the correlations was assessed by comparing the R2 t to R2 values generated via 1000 random permutations of the environmental variables.
The BrayCurtis dissimilarity matrices obtained from the mineral
datasets were analysed in a hierarchical experimental design (with the
location of the vineyard or the agricultural practice as the grouping factor), using permutational multivariate analysis of variance (perMANOVA,
see Anderson, 2001). The analysis of variance was performed with the
adonis function of the R Vegan package 2.0-10 (Oksanen et al., 2010),
using 999 permutations. The possibility of signicant effects arising due
to differences in multivariate dispersion rather than compositional
change was excluded on the basis of the beta diversity measurements
within and between groups (using the betadisper function from the
Vegan package).
3. Results
The data for soil physico-chemical analyses, the grapevine leaf
element analyses, and the colonisation estimations are presented in
Table 1 and Online resources 12.
3.1. Spatial variability of mineral composition
Hierarchical clustering was performed for the similarity of the
vineyards based on the mineral compositions of the grapevine leaves.
The cluster analysis of vineyards provided a dendrogram (Fig. 1) in
which only the vineyards from Ivan Dolac were grouped together. The
vineyard pairs (conventional and sustainable) from all of the other locations were in separate groups. With two exceptions (Ivan Dolac conventional, Zadar sustainable), the locations of the conventional vineyards
showed above average z-scores, and those for the sustainable vineyards
were below average. The highest dissimilarity to average values (extreme
z values; Fig. 1, intensive colours) was observed with the z-scores obtained for the Zadar and Peljesac sustainable vineyards, and the Milna
conventional vineyards.
The cluster analysis of the minerals measured in the grapevine
leaves formed two larger clusters, with the rst cluster containing the
macronutrients K and P and the micronutrients Fe and Mn. The second
cluster was composed of micronutrient Zn and macronutrient Ca.
PCA was applied to the normalised grapevine mineral data from the
leaves, to compare the compositional patterns between the vineyard
samples and to identify potential factors that inuenced each of these.
As the scree plot did not have a clear drop-off (data not shown), we
experimented with a range of factor counts near the optimal coordinate
index (as calculated by nScree in the nFactors R-project library), with
three factors nalised, which corresponded to the number of PCs with
eigenvalues N1.0. A varimax rotation (raw) of the PCs to three different
variance factors (VFs) explained about 67% of the total variance
(Table 2). According to Liu et al. (2003), the factor loadings were classied
as strong, moderate, and weak, as those corresponding to absolute

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

727

Fig. 1. Heat map of the median elemental abundances (quantied as moderated z-scores) of the clusters generated using complete linkage hierarchical clustering based on Euclidean
distances for the mineral compositions of the grapevine leaves. The dendrogram represents the relationships between vineyards (rows) and minerals (columns), with the colour in the
intersections representing the magnitude of abundance: green and yellow if median for elements is b0; dark orange and red if median for elements is N0.

loading values of N0.75, 0.750.50 and 0.500.30, respectively. VF1

explained 26% of the total variance, and had strong loadings for Mn and
moderate positive loadings for Fe. VF2 explained an additional 21% of
the total variance, and showed strong negative loadings for K and moderate loadings for Ca. VF3 showed strong loadings for P, and explained 20%
of the total variance.
Multiple linear regression analysis was performed for the elements
in the leaves that were strongly correlating with rst three VFs (i.e., P,
K, Mn) according to the soil parameters (Table 3). None of the soil
parameters were signicant for the leaf-accumulated macronutrients
P and K. In contrast, pH and K, Ca and Fe concentrations were signicant
for the prediction of leaf-accumulated Mn. The variance ination factors
for pH, K and Ca were above 10, which indicates severe multicollinearity
between these variables (Neter et al. 1989).
To resolve the multicollinearity problem in the regression analysis,
PCA was applied to the normalised soil data. PCA of the entire soil
dataset (Table 1) resulted in ve PCs with eigenvalues N1 that explained
about 84% of the total variance in the mineral composition dataset. The
scree plot (data not shown) showed a clear drop-off after the fth PC.
We checked the number of used PCs with a range of factor counts
near the optimal coordinate index (as calculated by nScree in the
nFactors R-project library) and nalised on ve factors, which again
corresponded to the number of VFs with eigenvalues N 1.0.
Factor analysis with the varimax rotation variables of the soil produced ve different VFs with eigenvalues N1.0 that explained about
78% of the total variance (Table 4). VF1 explained 29% of the total variance and had strong loadings for K, Fe, Mn and Zn, and moderate

positive loadings for P. VF2 (15% of total variance) showed strong

loadings for Cu. VF3 (13% of total variance) showed strong loadings
for K and pH. VF4 (11% of total variance) had strong loadings for organic
matter content and vesicular colonisation, whereas VF5 (9% of total
variance) showed strong loadings for the biotic parameter of arbuscule
colonisation. The communality values showed that the model obtained
performed very well for K, Fe, Cu and pH, as the model explained almost
all of the variation for these variables.
The factor scores for the ve soil VFs were obtained through the factor
score coefcients given in Table 4 and were used as independent
variables in the regression analysis, to determine the signicant
factor(s) for the leaf-accumulated P, K and Mn. The results showed that
one each of the VFs was signicant for the P (VF3) and Mn (VF4) concentrations in the leaves, while two factors were signicant for the leaf K
(VF1 and VF2) (Table 5). Regression analysis showed that the VFs explained 26% to 34% of the variance in the P, K and Mn leaf concentrations.
The variance ination factors were b 10, which indicated that the
problem of multicollinearity presented in Table 3 was resolved.
Mantel test showed that soil structure affected the concentrations of
measured element in grapevine leaves (Mantel R = 0.40, p = 0.02).
More specically, a positive correlation between coarser fraction
(R = 0.73, p b 0.05 for 0.20.2 mm fraction) and leaf P concentration,
as well as negative correlation of leaf P with ner soil fractions (R =
0.74, p b 0.05 for 0.020.002 mm and R = 0.77, p b 0.05 for
b0.002 mm fractions) were observed. In addition, the coarsest soil fraction (N0.2 mm) correlated positively with Mn concentration in grapevine leaves (R = 0.81, p b 0.05).

Table 2
Rotated factor loadings of variables from the leaf mineral composition dataset of grapevines on the rst three rotated PCs.

3.2. Effects of environment and vineyard management

Variable
P
K
Ca
Fe
Mn
Zn
Eigenvalue
% total variance
Cumulative variance (%)

VF1

VF2

VF3
1.00

0.33
0.41
0.52
0.99
1.55
26
26

0.90
0.52
0.38
0.42
1.25
21
47

Communality
0.99
0.93
0.46
0.41
0.99
0.20

1.20
20
67

Bold values represent strong loadings. Loadings below 0.30 are not included.

Calculation of the BAFs for the measured elements showed that

those for P and K were increased in the sustainable vineyards at Peljesac
and Zadar (Table 6, Online Resource 3). Overall, the management practice was a signicant factor that affected the BAFs for P, K and Zn. Higher
BAFs for P and K were observed in the sustainable vineyards, whereas
the BAF for Zn was increased in the conventional vineyards. The BAFs
for Ca, Fe and Mn showed no effects of specic management practice,
with the BAFs for Ca were strongly increased for Zadar (regardless of
the management practice).
Comparisons of combined effects of location and management
practice on the mineral compositions of the grapevine leaves with

728

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

Table 3
Results of the multiple regressions for leaf P, K and Mn.
Leaf accumulated element
P

Constant
P
OM
pH
K
Ca
Fe
Cu
Zn
AC
VC
MSC

Mn

VIF

Estimate

SE

Estimate

SE

Estimate

SE

1.63E+06
1.66E+03
1.80E+04
1.53E+05
1.62E+01
5.57E01
3.00E+00
3.34E+02
2.79E+03
4.40E+03
3.17E+03
1.82E+04

5.67E+05
7.85E+03
1.27E+04
1.09E+05
1.89E+01
1.10E+00
9.90E+00
4.08E+02
2.13E+03
3.41E+03
4.34E+03
3.01E+04

2.876
0.211
1.418
1.408
0.855
0.505
0.303
0.818
1.311
1.288
0.730
0.605

0.008
0.834
0.169
0.172
0.401
0.618
0.764
0.422
0.202
0.210
0.472
0.551

1.52E+07
2.71E+03
5.22E+04
1.08E+06
3.40E+02
1.30E+01
1.27E+02
7.85E+03
9.92E+03
2.92E+04
8.30E+03
2.87E+05

6.05E+06
8.38E+04
1.35E+05
1.16E+06
2.02E+02
1.18E+01
1.06E+02
4.36E+03
2.27E+04
3.64E+04
4.63E+04
3.21E+05

2.505
0.032
0.386
0.931
1.686
1.100
1.201
1.803
0.437
0.803
0.179
0.895

0.019
0.975
0.703
0.361
0.105
0.282
0.242
0.084
0.666
0.430
0.859
0.380

8.07E+05
1.26E+03
5.91E+03
1.06E+05
1.68E+01
8.66E01
1.12E+01
1.73E+02
2.64E+02
1.17E+03
1.93E+03
7.03E+03

2.04E+05
2.82E+03
4.55E+03
3.90E+04
6.78E+00
3.96E01
3.55E+00
1.47E+02
7.63E+02
1.22E+03
1.56E+03
1.08E+04

3.965
0.447
1.299
2.723
2.474
2.186
3.162
1.179
0.345
0.952
1.238
0.652

b0.001
0.659
0.206
0.012
0.021
0.039
0.004
0.250
0.733
0.350
0.228
0.521

1.871
2.209
11.556
10.738
5.487
15.001
3.310
4.504
1.600
2.369
2.521

VIF, variance ination factor.

perMANOVA showed that the general locations of the vineyards from

which the samples were taken and the management practice affected
the mineral compositions of the grapevine leaves (Table 7). In addition,
the effects of management practice and the interaction between locations and management practice were statistically signicant. The beta
diversity tests were not signicant for any of the grouping variables.
Non-parametric multidimensional scaling of the grapevine mineral
composition dataset showed a small tendency towards sustainable
management of the vineyards towards the upper right part of the ordination plot (Fig. 2a). No clear separation from the conventional
vineyards was observed for either of the axes. Among the measured
environmental parameters, the soil concentrations of K, Fe and Cu, the
organic matter content, and the vesicular colonisation showed the
strongest ts to the ordination plot (Fig. 2b).
4. Discussion
4.1. Compositional elemental variability
Clustering did not resemble the grouping of the minerals in the
grapevine leaves according to the general needs of the grapevines
(i.e., macronutrients, micronutrients), but rather according to their interactions in the grapevines or in the soil, and their characteristics. Zn
was separated from two other micronutrients (Fe, Mn), as despite a
somewhat lithophile character, it is counted among the chalcophile
elements. Unlike other d-block micronutrients, Zn has only a single
oxidation state (a single valence of II), and thus it behaves differently
(Storey, 2007). In addition, Ca was separated from both of the other
Table 4
Rotated factor loadings of the variables from the soil characteristics and the colonisation
datasets of the grapevines.
VF1
P
K
Ca
Fe
Mn
Zn
Cu
pH
OM
AC
VC
MSC
Eigenvalue
% total variance
Cumulative variance (%)

VF2

VF3

VF4

VF5

0.56
0.88
0.86
0.86
0.77
0.82

0.40

0.93
0.79
0.81
0.89
0.68
3.53
29
29

0.68
1.75
15
44

1.61
13
57

1.35
11
69

Communality
0.53
0.94
0.84
0.96
0.69
0.74
0.91
0.99
0.69
0.83
0.59
0.58

1.05
9
78

Bold values represent strong loadings. Loadings below 0.30 are not included.

macronutrients measured (P, K), which is probably a result of the

good availability of Ca in calciferous soils found in the sampled
vineyards.
Factor analysis performed on the grapevine mineral compositions
showed that the overall variability can be reduced to three factors that
explain 67% of the total variability and that correspond to the two important macronutrients P and K, and the micronutrient Mn. Both of
these macronutrients can be decient if they are not supplied by
fertilising the soil, as they are steadily removed with the grapevine biomass (i.e., pruning, fruit). In vineyards, K is heavily removed with the
collected fruit (Mengel, 2007). Exchangeable and solution K equilibrate
rapidly, whereas non-exchangeable K equilibrates very slowly with the
exchangeable and solution forms. The transfer of K from the mineral
fraction to any of the other three forms is extremely slow in most
soils, and this K is considered essentially unavailable to crops during a
single growing season (Havlin et al., 1999). As such, K can be in deciency
if it is not supplied yearly with the application of mineral fertilisers. Linear
regression analysis isolated factors VF1 and VF2, which showed strong
loadings with the soil total K, Fe, Mn, Cu and Zn concentrations, as significant predictors for plant-accumulated K. The relationship between VF1
and K leaf concentrations was negative, which might be a result of the
high K-binding capacity of the soil (the total K concentrations were
higher, but the K was inaccessible for the grapevines) or antagonistic interactions between cations (especially Ca and Mg) (Marschner, 1995). As
all of these vineyards are on calcareous soils, high soil Ca concentrations
were observed, and these might reduce the K uptake by the grapevine
roots (Havlin et al., 1990).
Similar to K, the P concentrations in the leaves were below the optimal concentrations for grapevines (Sanchez, 2007). The soil pH is the
main factor that determines phosphorus solubility and availability to
grapevines. In the present study, multiple linear regression analysis isolated VF3, which showed strong loadings with soil Ca and pH as the
most signicant predictor for the measured plant P concentrations.
oga et al. (2008) observed a similar negative correlation between soil
pH and P concentrations in grapevine leaves. In addition, P inactivation
can occur in vineyards on calcareous soils, due to the high concentrations of Ca ions and the high soil pH (Marschner, 1995).
Manganese has an important role in the synthesis of chlorophyll and
in nitrogen metabolism, and it is present in soil as exchangeable Mn or
Mn oxide. Grapevines can develop Mn deciency when growing on
calcareous soils, as the Mn gets immobilised in the soil at higher pH
(Humphries et al., 2007). In agreement with this, we observed increased
BAFs for Mn at Zadar, which was characterised by the lowest pH in the
present study. Despite the soil pH, very high leaf Mn concentrations that
were in the range of, or above, optimal concentrations (30100 mg/kg)
were measured in the present study. Similar to our data, oga et al.
(2010) also observed supra-optimal concentrations of Mn in grapevine

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

729

Table 5
Results of multiple regressions for leaf P, K and Mn based on factor analysis.
Leaf accumulated element
P

Constant
VF1
VF2
VF3
VF4
VF5

Mn

VIF

Estimate

SE

Estimate

SE

Estimate

SE

588.72
49.47
36.89
93.89
59.53
20.43

27.16
28.28
28.8
28.72
30.8
30.15

21.679
1.75
1.281
3.269
1.933
0.677

b2.00E16
0.090
0.210
0.003
0.063
0.503

7263.9
872.9
799.8
172.6
134.7
390.9

284.6
296.3
301.8
300.9
322.8
316

25.526
2.946
2.65
0.574
0.417
1.237

b2.00E16
0.006
0.013
0.571
0.679
0.226

109.147
19.310
15.612
12.096
52.265
0.560

11.084
11.542
11.754
11.722
12.572
12.308

9.847
1.673
1.328
1.032
4.157
0.046

6.54E11
0.105
0.194
0.310
b0.001
0.964

1.005
1.003
1.014
1.011
1.002

VIF, variance ination factor.

leaves collected at veraison, which suggested improved uptake during

the owering/ fruiting period. Regression analysis showed a negative
relationship for leaf Mn concentrations with VF4, which suggested
that organic matter has an important inuence on the Mn uptake and
accumulation in grapevine leaves. The most probable reason for the
reduced Mn uptake at high organic matter levels is the formation of
organic-Mn complexes (Heintze, 1957; Geering et al., 1969; McBride,
1982).
In addition to other soil parameter, grapevine leaf P and Mn concentrations seem to correspond with different soil size fractions. Leaf P was
positively correlated to coarser fraction and negatively with ner soil
fractions. Due to higher adsorption capacity clays are able to restrict P
availability, whereas sand has fewer potential P binding sites than clay
and silt and does not impair the P uptake by the grapevine. In contrast
to P, Mn leaf concentrations were negatively correlated to the coarsest
fraction of the soil, which could be due to the reduced Mn availability
in the less water retentive sandy soils, as it has been reported to
decrease in availability under dry soil conditions (Humphries et al.,
2007).
4.2. Effects of environment and vineyard management practice on the
mineral compositions of grapevine leaves
Cluster analysis dened three clusters of locations based on the
mineral compositions of the grapevine leaves; these showed no distinct
connections to the geographical latitudes of the sample locations. Only
one of the clusters contained two vineyards from the same location,
which suggests high inter-vineyard variability in the elemental compositions. This effect was probably increased by the differences in the
management practices, with a tendency towards lower mineral concentrations in the grapevine leaves from the sustainable management
locations.
The perMANOVA results conrmed this, as both the general location
and the management practice signicantly affected the elemental compositions of the grapevine leaves. A deeper examination of the individual samples and the environmental parameters by non-parametric

multidimensional scaling showed that both abiotic (e.g., organic matter,

soil K, Fe and Cu) and biotic (e.g., vesicular colonisation) factors had
strong effects on the mineral compositions of the grapevines. Potassium
as a macronutrient has an important direct effect on the growth and
physiology of the grapevines (Mengel, 2007). Organic matter, on the
other hand, is a source of minerals as well as a potential point of interaction with the elements, thus modifying their availability for the grapevines (Mackie et al., 2013; Marschner, 1995). In combination with pH,
organic matter can severely change the availability of P, which was
seen as a major contributor to the spatial variability in the elemental
compositions of the grapevines in the present study (Sanchez, 2007).
The majority of the vineyard soils in the present study were alkaline,
which increases the complexation of P (in the form of PO
4 ) to Ca oxides,
and decreases the availability of P for the grapevines (White, 2009).
Organic matter added as manure or litter and native organic matter
(humic materials) have the opposite effects, and can signicantly
change the subsurface retention of P (Sanchez, 2007). Organic matter
not only affects the absorption and precipitation of P, but often also
contains signicant amounts of this element, which is thereby added
to the soil and can be mobilised by the soil microbiota, like symbiotic
fungi. Although it is commonly believed that AM fungi cannot
decompose complex organic molecules, they have been observed to
accelerate decomposition and nitrogen acquisition of plants (Cheng
and Baumgartner, 2006; Hodge et al., 2001). In addition, the bioaccumulation of Zn and Fe was increased in the majority of the conventional
vineyards. This might be a result of the application of foliar fertilisers, as
increased Zn concentrations have been observed in grapevines treated
with Mg and/or Fe fertilisers (Brateevec et al., 2013). However, contrary to this, Daz et al. (2010) showed no effects, or even reduction of
Zn concentrations, in Fe-fertilised grapevines, which suggested the involvement of additional factors that were not monitored in the present
study.
Similar to other plant species with low root densities or with relatively coarse ne roots, grapevines appear to be reliant on colonisation
with AM fungi for their normal growth and development (Biricolti
et al., 1997; Karagiannidis and Nikolaou, 1999; Linderman and Davis,

Table 6
Bioaccumulation factors for the elements accumulated in grapevine leaves calculated as the ratios between leaf concentrations and soil concentrations.
Location

Management

Ivan Dolac

Conventional
Sustainable
Conventional
Sustainable
Conventional
Sustainable
Conventional
Sustainable

44.3
254
262
274
160
4575
117
2595

Milna
Peljesac
Zadar

Data are means SE (n = 5).

6.01
106
43.9
77.8
24.7
1880
23.6
593

0.33
0.25
0.43
0.47
0.72
1.13
0.51
0.95

Ca

0.03
0.03
0.06
0.05
0.04
0.21
0.11
0.10

0.32
0.62
0.26
0.13
0.14
0.12
4.32
4.21

Mn

0.02
0.14
0.01
0.01
0.01
0.02
1.32
0.34

0.05
0.05
0.26
0.21
0.13
0.04
0.31
0.55

Fe

0.004
0.004
0.04
0.04
0.003
0.004
0.15
0.12

0.003
0.002
0.003
0.002
0.007
0.004
0.005
0.009

Zn

0.000
0.000
0.000
0.000
0.001
0.001
0.001
0.001

0.14
0.16
0.64
0.21
1.12
0.27
1.75
0.56

0.02
0.01
0.39
0.03
0.23
0.05
0.31
0.02

730

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

Table 7
Results of the perMANOVA analysis and beta-diversity tests for the mineral compositions
of the grapevine leaves, using general location, management practice, and their interaction
as the source of diversity.
Factor

Location
Management practice
Location management practice

perMANOVA
2

Beta-diversity test

5.512
14.034
5.358

0.221
0.188
0.215

0.002
0.001
0.001

0.161
0.135
0.643

2001). Indeed, in some cases, grapevines are even fully dependent on

AM fungi (Menge et al., 1983). AM colonisation of grapevine coarse
roots in production vineyards along the east Adriatic Karst region has
been estimated to range from 64% to 82%, which was related to plantavailable P and location (Likar et al., 2013). We now demonstrate the
importance of the vineyard management practice within locations as a
factor that affects grapevine mineral nutrition. Grapevines form lowinput organic systems that harbour 30% to 60% higher root colonisation
levels compared to conventional systems, and up to 50% of variation in
root colonisation can be explained by the chemical properties of the soil
(pH, P, K, Mg) (Mdder et al., 2000). In part, this can be attributed to
differences in the AM fungi spore abundance and species diversity
that results from the lower fertiliser input levels (Oehl et al., 2004).
Additionally, soil management practices such as tilling, cover crops,
and grapevine understory have been demonstrated to signicantly
affect the microbial activity and AM fungi root colonisation in grapevines (Radi et al., 2012; Steenwerth and Belina, 2008). Appropriate
cover crops effectively enhance the soil organic matter content, with
effects including reduced run-off waters and improved soil water content, which is of particular importance in Mediterranean and semi-arid
agroecoystems (Steenwerth and Belina, 2008). Therefore, the consistent
relationship between the grapevine AM fungi colonisation parameters
(AM fungi and vesicular colonisation) with the organic matter content

(Likar et al., 2013; present study) indicates the need for a more sustainable approach in vineyard soil management, which should be directed towards supporting the complex intimate relationship between soil biotic
and abiotic properties, in order to achieve optimised grapevine functioning. Furthermore with promotion of healthy AM fungal communities in
agricultural ecosystems, increased concentration of benecial elements
in crop plants (Lehmann et al., 2014) (and/or its bioavailability to
humans) could be achieved in a sustainable way.
In conclusion, our data demonstrate the high impact of the location
characteristics on the mineral composition of the grapevine leaves, with
K, P and Mn explaining two-thirds of the variance between samples. All
of these three elements are important grapevine nutrients and they can
be decient in calcareous soils. Furthermore, not only abiotic, but also
biotic soil factors like AM fungi colonisation, have strong impact on the
mineral composition of the grapevines. In addition to the soil characteristics, the management practices are conrmed as important factors for the
terroir, as these inuenced the bioaccumulation of specic macronutrients and micronutrients.

Acknowledgements
The authors thank Peter Kump and Andreja Redek for the help with
the X-ray uorescence analyses. This study was supported by the
Slovenian Research Agency (ARRS) through the Plant Biology Programme group (P1-0212), and Unity through the Knowledge Fund,
Croatia (UKF, grant no 53). Dr Chris Berrie is acknowledged for English
revision of the manuscript.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.scitotenv.2014.10.057.

Fig. 2. Non-metric multidimensional scaling plots of grapevine leaf mineral compositions according to the individual vineyards (a) and with overlaid soil variables (b). Abbreviations in the
middle of the centroids represent the locations of the vineyards: ID, Ivan Dolac (triangles); M, Milna (upside-down triangles); P, Peljesac (squares); Z, Zadar (circles) and agricultural practice used: C, conventional vineyard (lled symbols); E, sustainable vineyard (empty symbols). (b) The vectors point in the direction of the most rapid change of the parameter with the
length scaled according to the strength of the correlation (only vectors with signicance p b 0.1 are shown). OM, organic matter; V%, vesicular colonisation.

M. Likar et al. / Science of the Total Environment 505 (2015) 724731

References
Amors JA, Garcia FJ, Sanchez CJ, Prez de los Reyes C, Garca R, Jimenez R. Trace element
distribution in red soils under semiarid Mediterranean environment. Int J Geosci
2011;2:8497.
Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral
Ecol 2001;26:3246.
Bavaresco L, Giachino E, Pezzutto S. Grapevine rootstock effects on lime-induced chlorosis,
nutrient uptake and sourcesink relationships. J Plant Nutr 2003;26:145165.
Bertoldi D, Larcher R, Bertamini M, Otto S, Concheri G, Nicolini G. Accumulation and
distribution pattern of macro and microelements and trace elements in Vitis vinifera
L. cv. Chardonnay berries. J Agric Food Chem 2011;59:722436.
Biricolti S, Ferrini F, Rinaldelli E, Tamantini I, Vignozzi N. VAM fungi and soil lime content
inuence rootstock growth and nutrient content. Am J Enol Vitic 1997;48:939.
Brateevec K, Sivilotti P, Vodopivec BM. Soil and foliar fertilization affects mineral contents in Vitis vinifera L. cv. Rebula leaves. J Soil Sci Plant Nutr 2013;13:65063.
Braun-Blanquet J. Panzensoziologie: Grundzge der Vegetationskunde (Phytosociology:
Fundamentals of Vegetation Science). Berlin: Springer; 1964.
Bray JR, Curtis JT. An ordination of the upland forest communities of southern Wisconsin.
Ecol Monogr 1957;27:32549.
Cheng X, Baumgartner K. Effects of mycorrhizal roots and extraradical hyphae on 15N uptake from vineyard cover crop litter and the soil microbial community. Soil Biol
Biochem 2006;38:266575.
Chopin EIB, Marin B, Mkoungafoko R, Rigaux A, Hopgood MJ, Delannoy E, et al. Factors affecting distribution and mobility of trace elements (Cu, Pb, Zn) in a perennial grapevine (Vitis vinifera L.) in the Champagne region of France. Environ Pollut 2008;156:
10928.
oga L, Slunjski S, usti M, Gunjaa J, osi T. Phosphorous dynamics in grapevine on
acid and calcareous soils. Alps-Adria Scientic Workshop, Stara Lesna, Slovakia; 2008.
oga L, Slunjski S, Herak usti M, Horvat T, Petek M, Gunjaa J. Effect of soil pH reaction
on manganese content and dynamics in grapevine (Vitis vinifera L.). Acta Horticult
2010;868:2038.
Daz I, Barrn V, del Campillo MC, Torrent J. Testing the ability of vivianite to prevent iron
deciency in pot-grown grapevine. Sci Hortic 2010;123:4648.
Eissenstat DM. Costs and benets of constructing roots of small diameter. J Plant Nutr
1992;15:76382.
Geering HR, Hodgson JF, Solans C. Micronutrient cation complexes in soil solution. Soil Sci
Soc Am J 1969;38:815.
Golob T, Doberek U, Kump P, Neemer M. Determination of trace and minor elements in
Slovenian honey by total reection X-ray uorescence spectroscopy. Food Chem
2005;91:593600.
Goslee SC, Urban DL. The ecodist package for dissimilarity based analysis of ecological
data. J Stat Softw 2007;22:119.
Gruber B, Kosegarten H. Depressed growth of non-chlorotic vine grown in calcareous soil is
an iron deciency symptom prior to leaf chlorosis. J Plant Nutr Soil Sci 2001;165:1117.
Havlin JL, Kissel DE, Maddux LD, Claassen MM, Long JH. Crop rotation and tillage effects
on soil organic carbon and nitrogen. Soil Sci Soc Am J 1990;54:44852.
Havlin JL, Beaton JD, Tisdale SL, Nelson WL. Soil fertility and fertilizers. London: Prentice-Hall
International Ltd; 1999.
Heintze SG. Studies on soil manganese. J Soil Sci 1957;8:287300.
Helena B, Pardo R, Vega M, Barrado E, Fernandez J, Fernandez L. Temporal evolution of
groundwater composition in an alluvial aquifer (Pisuerga River, Spain) by principal
component analysis. Water Res 2000;34:80716.
Hodge A, Campbell CD, Fitter AH. An arbuscular mycorrhizal fungus accelerates decomposition and acquires nitrogen directly from organic material. Nature 2001;413:2979.
Humphries J, Stangoulis J, Graham R. Manganese. In: Barker A, Pilbeam D, editors. Handbook of plant nutrition. New York: Taylor and Francis; 2007. p. 35166.
Jackson RS. Wine science. Principles and applications. London: Elsevier; 2008.
Kandeler E. Organic matter by wet combustion. In: Schinner F, hlinger R, Kandeler E,
Margesin R, editors. Methods in soil biology. Berlin: Springer; 1995. p. 3978.
Karagiannidis N, Nikolaou N. Arbuscular mycorrhizal root infection as an important factor
of grapevine nutrition status. Multivariate analysis application for evaluation and
characterization of the leaf and soil parameters. Agrochimica 1999;43:15165.
Kump P, Neemer M, Smodi B, Jaimovi R. Multielement analysis of rubber samples by
X-ray uorescence. Appl Spectrosc 1996;50:13737.
Kump P, Neemer M, Vogel-Miku K, Rupnik Z, Ponikvar D, Pelicon P. Improvement of the
XRF quantication and enhancement of the combined applications by EDXRF and
Micro PIXE. First research co-ordination meeting under coordinated research project
on Unication of nuclear spectrometries: integrated techniques as a new tool for
materials research. Vienna: IAEA; 2007. p. 915.
Lehmann A, Veresoglou S, Leifheit EF, Rillig MC. Arbuscular mycorrhizal inuence on zinc
nutrition in crop plants a meta-analysis. Soil Biol Biochem 2014;69:12331.
Likar M, Hanevi K, Radi T, Regvar M. Distribution and diversity of arbuscular mycorrhizal
fungi in grapevines from production vineyards along the eastern Adriatic coast. Mycorrhiza 2013;23:20919.
Linderman RG, Davis EA. Comparative response of selected grapevine rootstocks and cultivars to inoculation with different mycorrhizal fungi. Am J Enol Vitic 2001;52:811.

731

Liu C-W, Lin K-H, Kuo Y-M. Application of factor analysis in the assessment of groundwater
quality in a blackfoot disease area in Taiwan. Sci Total Environ 2003;313:7789.
Mackie KA, Mller T, Zikeli S, Kandeler E. Long-term copper application in an organic
vineyard modies spatial distribution of soil micro-organisms. Soil Biol Biochem
2013;65:24553.
Mdder P, Edenhofer S, Boller T, Wiemken A, Niggli U. Arbuscular mycorrhizae in a longterm eld trial comparing low-input (organic, biological) and high-input (conventional)
farming systems in a crop rotation. Biol Fertil Soils 2000;31:1506.
Marengo E, Aceto M. Statistical investigation of the differences in the distribution of
metals in Nebbiolo-based wines. Food Chem 2003;81:62130.
Marschner H. Mineral nutrition of higher plants. London: Academic Press; 1995.
McBride MB. Electron spin resonance investigation of Mn2+ complexation in natural and
synthetic organics. Soil Sci Soc Am J 1982;46:113743.
McGonigle TP, Miller MH, Evans DG, Fairchild GL, Swan JA. A new method which gives
and objective measure of colonization of roots by vesiculararbuscular mycorrhizal
fungi. New Phytol 1990;115:495501.
Menge JA, Raski DJ, Lider LA, Johnson ELV, Jones NO, Kissler JJ, et al. Interaction between
mycorrhizal fungi, soil fumigation and growth of grapes in California. Am J Enol Vitic
1983;34:11721.
Mengel K. Potassium. In: Barker AV, Pilbeam DJ, editors. Handbook of plant nutrition. Boca
Ranton: CRC Press; 2007. p. 91120.
Neter J, Wasserman W, Kutner MH. Applied Linear Regression Models. Homewood, IL:
Irwin; 1989.
Oehl F, Sieverding E, Mder P, Dubois D, Ineichen K, Boller T, et al. Impact of long-term
conventional and organic farming on the diversity of arbuscular mycorrhizal fungi.
Oecologia 2004;138:57483.
Oehl F, Sieverding E, Ineichen K, Ris E-A, Boller T, Wiemken A. Community structure of
arbuscular mycorrhizal fungi at different depths in extensively and intensively
managed agroecosystems. New Phytol 2005;165:27383.
Oksanen JF, Blanchet G, Kindt R, Legendre P, O'Hara RB, Simpson GL, et al. Vegan: community ecology package. http://CRAN.R-project.org/package=vegan, 2010. [R package
version 1.17-4].
Olsen SR, Sommers LE. Phosphorus. In: Page AL, Miller RH, Keeney DR, editors.
Methods of soil analysis, part 2. Am Soc AgronMadison, Wisconsin: Soil Sci Soc AM;
1982. p. 40330.
Pessanha S, Carvalho ML, Becker M, Von Bohlen A. Quantitative determination on heavy
metals in different stages of wine production by total reection X-ray uorescence
and energy dispersive X-ray uorescence: comparison on two vineyards. Food
Chem 2010;112:2634.
Phillips JM, Hayman DS. Improved procedure of clearing roots and staining parasitic and
vesiculararbuscular mycorrhizal fungi for rapid assessment of infection. Trans Br
Mycol Soc 1970;55:15961.
Pohl P. What do metals tell us about wine? Trends Anal Chem 2007;26:9419.
Radi T, Hanevi K, Likar M, Protega I, Jug-Dujakovi M, Bogdanovi I. Neighbouring
weeds inuence the formation of arbuscular mycorrhiza in grapevine. Symbiosis
2012;56:11120.
Radi T, Likar M, Hanevi K, Bogdanovi I, Paskovi I. Occurrence of root endophytic fungi
in organic versus conventional vineyards on the Croatian coast. Agric Ecosyst Environ
2014;192:11521.
Rogiers S, Greer D, Hateld J, Orchards B, Keller M. Mineral sinks within ripening grape
berries (Vitis vinifera L.). Vitis 2006;45:11523.
Sanchez CA. Phosphorus. In: Barker AV, Pilbeam DJ, editors. Handbook of plant nutrition.
Boca Ranton: CRC Press; 2007. p. 5190.
Schreiner RP. Mycorrhizas and mineral acquisition in grapevines. In: Christensen LP,
Smart DR, editors. Proceedings of the Soil Environment and Vine Mineral Nutrition
Symposium; 2005. p. 4960.
Steenwerth K, Belina KM. Cover crops enhance soil organic matter, carbon dynamics and microbiological function in a vineyard agroecosystem. Appl Soil Ecol 2008;40:35969.
Storey JB. Zinc. In: Barker AV, Pilbeam DJ, editors. Handbook of plant nutrition. Boca
Ranton: CRC Press; 2007. p. 91120.
Van Espen PJM, Jansen KHA. Spectrum evaluation. In: Van Grieken RE, Markowicz AA, editors. Handbook of X-ray spectrometry. New York: M. Decker Inc.; 1993. p. 181293.
Van Leeuven C, Seguin G. The concept of terroir in viticulture. J Wine Res 2006;17:110.
Vega M, Pardo R, Barrado E, Deban L. Assessment of seasonal and polluting effects on the
quality of river water by exploratory data analysis. Water Res 1998;32:358192.
Vekemans B, Janssens K, van Espen P, Adams F. Quantitative X-ray Analysis System
(QXAS). Vienna, Austria: Distributed by the International Atomic Energy Agency;
1994.
Vogel-Miku K, Pongrac P, Kump P, Neemer M, Regvar M. Colonisation of a Zn, Cd and Pb
hyperaccumulator Thlaspi praecox Wulfen with indigenous arbuscular mycorrhizal
fungal mixture induces changes in heavy metal and nutrient uptake. Environ Pollut
2006;139:36271.
White RE. Understanding vineyard soils. Oxford Press; 2009.
Wooldridge J, Louw PJE, Conradie WJ. Effects of rootstock on grapevine performance, petiole and must composition, and overall wine score of Vitis vinifera cv. Chardonnay and
Pinot Noir. 31. S Afr J Enol Viticult; 2010. p. 458.