See

discussions, stats, and author profiles for this publication at: [Link]

Identification of superior walnut (Juglans regia)

genotypes with late leafing and high kernel
quality in Iran

Article in Scientia Horticulturae · September 2015

DOI: 10.1016/[Link].2015.06.049

CITATIONS READS

0 97

4 authors, including:

Aziz Ebrahimi Rouhollah Karimi

Purdue University Malayer University
50 PUBLICATIONS 54 CITATIONS 16 PUBLICATIONS 24 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Aziz Ebrahimi on 13 November 2015.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Scientia Horticulturae 193 (2015) 195–201

Contents lists available at ScienceDirect

Scientia Horticulturae
journal homepage: [Link]/locate/scihorti

Identification of superior walnut (Juglans regia) genotypes with late

leafing and high kernel quality in Iran
Aziz Ebrahimi a , Abdollah Khadivi-Khub b,∗ , Zia Nosrati c , Rouhollah Karimi d
a
Department of Forestry and Natural Resources, Agriculture College, Pfendler Hall, Purdue University, 715 West State Street, West Lafayette, IN 47907,
United States
b
Department of Horticultural Sciences, Faculty of Agriculture and Natural Resources, Arak University, 38156-8-8349 Arak, Iran
c
Iranian Academic Center for Education Culture and Research (ACECR), Department of Biotechnology and Physiology of Horticultural Plants, Karaj, Iran
d
Department of Landscape Engineering, Faculty of Agriculture, Malayer University, Malayer, Iran

a r t i c l e i n f o a b s t r a c t

Article history: In many areas of Iran, walnut production is limited by spring frosts that can damage the sensitive non-
Received 10 February 2015 dormant tissues. The present study was carried out within a seedling population in order to select superior
Received in revised form 25 June 2015 walnut genotypes in term of consistent late leafing time and high kernel quality. As a result of the pre-
Accepted 27 June 2015
selection according to leafing date, 61 trees were selected that were late leafing. Then, the selected
Available online 29 July 2015
trees were further evaluated according to nut and kernel characteristics. The variation was observed for
traits related to nut and kernel. Variability found in nut weight was between 7.70 and 22.33 g, in kernel
Keywords:
weight between 2.30 and 8.53 g, and in kernel percentage between 24.66 and 62.18%. Nut dimensions
Persian walnut
Late leafing
and weight were in significant positive correlation with kernel weight. The kernel percentage was in
Kernel weight positive correlation with kernel weight, kernel filled and kernel plumpness and in negative correlation
Kernel percentage with shell thickness. Cluster and principal component analyses confirmed considerable diversity in the
Promising selections studied germplasm. All of the 61 studied genotypes were late leafing and could be useful as a parent
to improve leafing and flowering season of cultivars. Furthermore, genotypes 6, 14, 15, 17, 34, 37, 38,
40, 45, 47, 53, 54, 56 and 57 were superior in terms of consistent high kernel percentages, large nut,
shell softness, large kernel, light kernel color, ease of kernel removing from nuts, and none blank nut
percentage and can be singled out for cultivation.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction genetic divergence present in seedling trees of walnut can help in

the selection of trees for further species improvement (Francesca
Persian walnut (Juglans regia L.) is one of the most important et al., 2010).
nut crops grown in temperate regions and produces edible nuts According to the FAO (2012), Persian walnut grows on 64,000 ha
having high nutritional value. Walnut is a high energy food, rich in Iran, producing 165,508 ton of nuts in shell, ranking second
in oil including omega-3 fatty acids, vitamins and minerals, and globally. In order to increase the walnut production and expor-
valued as healthy snack food and bakery ingredients (Rana et al., tation, new cultivars need to be introduced and old cultivars to
2007). It is an ancient species originating in areas of Central Asia, be conserved in gene banks (Francesca et al., 2010). To achieve
including Iran (Fjellstrom and Parfitt, 1994). Walnut trees grown this goal, the first step is identification of superior genotypes
are mainly raised from seeds and exhibit a tremendous variability in and subsequently characterization of their genetic variety. Per-
tree, foliage and floral characters in Iran. Also, the climatic diversity sian walnut was probably domesticated in the region of Iran and
and high heterozygosity because of continuing sexual propagation Afghanistan and subsequently introduced to China, Russia and East-
have formed a very rich genetic material among local walnut pop- ern Europe by ancient tribes (Bayazit et al., 2007). An ideal walnut
ulations in Iran. More variability within a species can give better cultivar must have late leafing, both terminal and lateral bearing,
chances for elite selection. Information on the nature and degree of low incidence of pistillate flower abscission, high yielding nuts
(>6 MT/ha) with large size, relatively smooth, >50% kernel per-
centage, plump and light colored kernel and at least moderately
resistant to pest and diseases (Botu et al., 2010; Cosmulescu et al.,
∗ Corresponding author. Fax: +98 8632762087. 2010).
E-mail address: a-khadivi@[Link] (A. Khadivi-Khub).

[Link]
0304-4238/© 2015 Elsevier B.V. All rights reserved.
196 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201

Table 1
Fruit traits utilized in the studied walnut genotypes.

No. Trait Abbr. Unit Min. Max. Mean SD CV (%)

1 Nut shape NuSh Code 1 9 3.90 2.24 57.44

2 Nut diameter NuDi mm 28.40 39.60 33.03 2.63 7.97
3 Nut length NuLe mm 31.60 43.70 36.45 2.64 7.25
4 Nut weight NuWe g 7.70 22.33 15.23 1.86 12.21
5 Shell texture SheTe Code 1 9 4.51 1.40 31.00
6 Shell color SheCo Code 1 7 4.08 1.57 38.58
7 Shell seal SheSe Code 1 7 4.34 1.74 40.12
8 Shell strength SheSt Code 1 7 4.70 1.40 29.79
9 Shell thickness SheTh mm 1.00 2.70 1.42 0.31 22.11
10 Integument thickness InTh mm 0.10 0.70 0.21 0.15 70.05
11 Kernel weight KeWe g 2.30 8.53 4.84 0.88 18.26
12 Kernel percentage KePe % 24.66 62.18 46.20 5.43 11.76
13 Kernel filled KeFi Code 1 5 3.43 1.35 39.39
14 Kernel plumpness KePl Code 1 5 3.81 1.43 37.43
15 Blank nut percentage BlNuPe % 0.00 4.00 0.19 0.58 302.63
16 Kernel color KeCo Code 1 7 3.19 1.43 44.73
17 Kernel shriveling KeShr Code 0 5 0.38 0.94 247.89
18 Kernel vein KeVe Code 0 5 2.47 1.31 53.12
19 Ease of kernel removing from nuts EKeNu Code 1 7 3.61 1.11 30.75

Table 2
Descriptor of the qualitative characters utilized for the studied walnuts.

Trait Code and state

1 3 5 7 9

Nut shape Round Broadly ovate Ovate Broad elliptic Elliptic

Shell texture Very smooth Smooth Intermediate Rough
Shell color Very light Light Dark Very dark
Shell seal Excellent seal Slightly open Intermediate Wide
Shell strength Soft Intermediate Hard Extremely hard
Kernel filled Low Intermediate Filled
Kernel plumpness Low Intermediate Very plump
Kernel color Very light Light Light amber Amber
Kernel shriveling None Slightly wrinkled Intermediate Wrinkled
Kernel vein None Slightly wrinkled Intermediate Wrinkled
Ease of kernel removing from nuts Very easy Easy Intermediate Difficult Very difficult

Table 3
The most important fruit traits for superior walnut genotypes in this investigation.

Genotype Nut length (mm) Nut diameter (mm) Nut weight (g) Kernel weight (g) Kernel percentage (%) Shell texture Kernel color

Khoramdareh-6 41.10 36.30 16.02 8.53 53.25 Smooth Light

Khoramdareh-14 35.60 35.50 17.32 8.44 48.73 Smooth Very light
Khoramdareh-15 37.90 34.00 16.44 8.50 51.70 Smooth Light
Khoramdareh-17 38.00 31.80 16.55 8.14 49.18 Smooth Very light
Khoramdareh-34 42.90 39.60 16.21 8.24 50.83 Smooth Light
Khoramdareh-37 39.30 38.90 15.23 8.02 52.66 Smooth Light
Khoramdareh-38 36.30 32.90 14.78 7.32 49.53 Smooth Light
Khoramdareh-40 40.60 34.20 14.32 7.44 51.96 Smooth Very light
Khoramdareh-45 37.40 36.20 13.31 7.89 59.28 Smooth Light
Khoramdareh-47 36.20 36.70 16.11 7.13 44.26 Smooth Very light
Khoramdareh-53 37.00 38.90 17.21 7.49 43.52 Smooth Very light
Khoramdareh-54 43.70 36.30 13.67 8.50 62.18 Very smooth Very light
Khoramdareh-56 38.40 33.50 13.32 8.35 62.69 Very smooth Light
Khoramdareh-57 37.40 31.10 16.78 8.47 50.48 Smooth Light

In many areas of Iran, walnut production is limited by spring 2. Materials and methods
frosts that damage the more sensitive non-dormant tissues. Wal-
nut cultivars with late leafing can be cultivated in mountain areas, The current research was conducted in a vast walnut orchard
where the late frosts are frequent. The genetic variation of native (40 ha) in Khoramdareh region from Zanjan province, Iran. Kho-
walnut populations presents many opportunities for walnut breed- ramdareh is located at 36◦ 25 34 N latitude, 49◦ 25 13 E longitude
ing. Walnut germplasm has been extensively used in the selection and 1575 m above the sea level with an annual average temper-
studies for producing the superior walnut clones (Sharma and ature of 11.30 ◦ C and an annual average precipitation of 322 mm.
Sharma, 2001; Godeanu et al., 2004; Eskandari et al., 2005; Achim Additionally, minimum, mean and maximum of monthly tempera-
et al., 2007; Botu et al., 2010; Cosmulescu et al., 2010; Ebrahimi tures for March were 1.30, 6.70 and 13.10 ◦ C, respectively, and also
et al., 2011; Cosmulescu and Botu 2012; Khadivi-Khub, 2014; for April were 4.40, 11.50 and 18.40 ◦ C, respectively.
Khadivi-Khub and Ebrahimi, 2015; Khadivi-Khub et al., 2015a,b). In the first step, pre-selections were done according to leafing
The aim of the present work was to select new genotypes combining time within about 4000 walnut trees originated from seed. This
late leaf break with having high kernel quality and quantity. large walnut germplasm shows high diversity for most of traits.
 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201 197

Table 4
Simple correlations among fruit variables utilized in the studied walnut genotypes.

Trait NuSh NuDi NuLe NuWe SheTe SheCo SheSe SheSt SheTh InTh KeWe KePe KeFi KePl BlNuPe KeCo KeShr KeVe

NuSh 1.00
NuDi −0.02 1.00
NuLe 0.42b 0.43b 1.00
NuWe 0.17 0.71b 0.55b 1.00
SheTe 0.05 −0.06 0.15 0.09 1.00
SheCo −0.12 0.06 −0.10 0.09 0.25 1.00
SheSe 0.10 −0.04 0.13 0.06 0.22 0.14 1.00
SheSt 0.13 0.02 0.11 0.23 0.22 0.05 0.08 1.00
SheTh −0.02 0.13 −0.05 0.24 0.03 0.01 0.02 0.55b 1.00
InTh 0.18 0.04 0.05 0.08 −0.03 0.17 −0.07 0.34b 0.26a 1.00
KeWe 0.02 0.62b 0.37b 0.82b −0.05 −0.03 −0.10 −0.03 0.07 −0.14 1.00
KePe −0.24 −0.12 −0.27a −0.15 −0.26a −0.24 −0.26a −0.38b −0.22 −0.34b 0.43b 1.00
KeFi −0.02 −0.01 −0.21 0.10 0.17 0.05 −0.16 −0.11 −0.08 −0.27a 0.29a 0.35b 1.00
KePl −0.08 0.07 −0.15 0.24 −0.07 0.02 −0.10 −0.15 0.04 −0.33b 0.40b 0.31a 0.67b 1.00
BlNuPe 0.15 −0.16 0.08 −0.24 −0.11 −0.04 −0.02 0.42b 0.09 0.35b −0.37b −0.32a −0.32a −0.32a 1.00
KeCo 0.14 0.03 0.21 0.00 0.08 0.33b 0.13 −0.11 −0.14 0.14 −0.20 −0.36b −0.15 −0.24 0.06 1.00
KeShr 0.06 0.05 0.09 0.02 0.09 0.02 0.20 0.09 -0.03 -0.02 −0.11 −0.25 −0.02 0.02 0.14 0.17 1.00
KeVe 0.16 −0.16 −0.02 −0.25 −0.19 −0.19 0.07 −0.15 −0.18 −0.10 −0.29a −0.14 −0.19 −0.27a 0.14 0.21 0.00 1.00

For explanation of character symbols, see Table 1.

a
Correlation is significant at the 0.05 level.
b
Correlation is significant at the 0.01 level.

For instance, agronomic traits such as leafing, flowering and ripen- 2001). Besides, Ward method was used to calculate the distance
ing times among them are very early, early, middle and late. The coefficients between all pairs of genotypes. The representativeness
genotypes with early and middle leafing dates were eliminated and of the dendrogram was evaluated by estimating cophenetic cor-
finally 61 late leafing trees were selected the first year. Then, in relation for the dendrogram and comparing it with the distance
the second step, the late leafing selections were further evaluated matrix using Mantel’s matrix correspondence test (Mantel, 1967)
according to their nut and kernel characteristics for two consec- by the COPH and MYXCOMP programs of the NTSYS-pc version 2.02
utive years to determine superior types. The selected genotypes software (Rohlf, 2000). The result of this test is a cophenetic corre-
were named based on their location, and these names were supple- lation coefficient, r, indicating how well the dendrogram represents
mented with numerical characters. The selected trees were mature, similarity data.
healthy, and had a full crop. General orchard management, includ-
ing irrigation, nutrition, pest and disease control, was consistent
3. Results and discussion
with commercial practices.
Nineteen nut and kernel characters were evaluated for two
3.1. Characteristics of genotypes
consecutive years to detect promising genotypes (Table 1). Mea-
surements of each nut and kernel trait were based on 20 replicates
Late-leafing in Iranian walnuts, such as the Turkish cultivars
and the mean values were used. Some variables were measured by
(Akca and Ozongun, 2004) is not common (Khadivi-Khub et al.,
laboratory equipment. Nut dimensions (length and diameter) and
2015a). As a result of the pre-selections, the 61 selected walnut
shell thickness were measured using a digital caliper. The weight
trees in our study were selected based on possession of late leafing.
for nut and kernel was measured using electronic balance with
Then they were further assessed according to their nut and kernel
0.01 g precision. Also, some characteristics such as nut shape, kernel
quality and quantity. In climates with wet springs and dry sum-
traits (filled, plumpness, shriveling and color) and shell traits (color,
mers, late-leafing individuals escape the environmental conditions
seal, hardiness and texture), were determined based on rating and
conducive to blight infection. Flowering dates are similarly used for
coding (Table 2) according to walnut descriptor (IPGRI, 1994). Ker-
the selection of pollinizers (Forde, 1979). Together, later budbreak
nel percentage was estimated using formulas “kernel weight/nut
and flowering period create more opportunity to pistillate flowers
weight × 100”.
that survive the late spring frosts (Forde, 1979). The harvest dates
of the selected trees varied from the early to the end of September.
2.1. Data analysis Since leafing and harvest dates are characteristics with very high
and high heritability, respectively (Hansche et al., 1972), they are
Mean values were taken to analyze data and to draw the useful criteria to select promising genotypes (Khadivi-Khub et al.,
conclusions. The data were analyzed for mean, variances, corre- 2015a).
lations, regression and genetic diversity to deduce the genetic There were large variations between fruit traits in the stud-
similarity/dissimilarity. One-way analysis of variance (ANOVA) was ied walnut genotypes (Table 1). Nut length varied between 31.60
conducted to detect significant differences in the mean morpholog- and 43.70 mm, nut diameter between 28.40 and 39.60 mm, nut
ical traits between the studied genotypes using SAS (SAS Institute, weight between 7.70 and 22.33 g, shell thickness between 1.00 and
Cary, NC, USA). Coefficients of variation (CV) were determined as 2.70 mm, kernel weight between 2.30 and 8.53 g, kernel percent-
indicators of variability. To avoid effects due to scaling differences, age between 24.66 and 62.18% and blank nut percentage between
the mean of each character was normalised prior to cluster anal- 0.00 and 4.00 %. The highest nut weight among our evaluated trees
ysis using Z-scores. Correlations between traits were determined (22.33 g) was less than the corresponding data reported by Sen and
using the Pearson correlation coefficient with SPSS® software ver- Tekintas (1992) for Adilcevaz, Turkey (23.81 g), and more than that
sion 16 (SPSS Inc., Chicago, IL, USA, Norusis, 1998). Relationships by Atefi (1997) for Karaj, Iran (20 g), Sharma and Sharma (1998) for
among genotypes were investigated with principal components Himachal Pradesh, India (18.60 g), Yarilgac et al. (2001) for east Ana-
analysis (PCA) using SPSS® software. Scatter plot was created tolia, Turkey (17.04 g), and Aslantas (2006) for northeast Anatolia,
according the first two PCs using PAST software (Hammer et al., Turkey (16.01 g).
198 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201

Recently, Khadivi-Khub and Ebrahimi (2015) and also Khadivi-

Khub et al. (2015a,b) have studied several walnut germplasms
from different parts of Iran and reported range of 3.60–23 g for
nut weight and also 1.32–14 g for kernel weight. Furthermore, they
observed range of 17.44–83.88% for kernel percentage among the
evaluated genotypes. Desirable nut and kernel weights should be
12–18 g and 6–10 g, respectively, or kernel weight should be at least
50% of the entire nut weight, and the kernel should have a light color
(McGranahan and Leslie, 1990). In the present study, 21 of the 61
studied trees were genotypes possessing these desirable traits.
Kernel percentage in 27 genotypes was >50%. In breeding pro-
grams, genotypes with a kernel ratio above 50% are more desirable
(Germain, 1997; Korac et al., 1997). The range of shell thickness
(1.00–2.70 mm) in our evaluated genotypes was lower than values
reported for selections from east Anatolia (Yarilgac et al., 2001),
Fig. 1. Variance percentage accounted by seven principal components (PCs) in PCA.
the middle Black Sea region (Akca and Ozongun, 2004) and north-
east Anatolia (Aslantas, 2006) in Turkey. The most desirable range
of shell thickness in walnut is 0.70–1.50 mm (Zhadan and Strukov, Knowledge of the relationship between nut and kernel character-
1977). The shell thickness of 46 selected genotypes varied between istics can guide appropriate selection schemes for walnut breeding
1.00 and 1.50 mm; thus, it is clear that 75% of the genotypes are programs (Khadivi-Khub and Ebrahimi, 2015).
promising in terms of shell thickness.
Ten genotypes had very light-colored kernels and kernels of
39 genotypes were light. In addition, removal of the kernel from 3.3. Multivariate analyses
nuts in 19, 21, 13 and 8 genotypes was very easy, easy, moderate
and difficult, respectively. In general, the most interesting geno- Principal component analysis (PCA) was used to study patterns
types are those whose kernels can be easily removed from the of variation in a set of interrelated traits through the identification
shell and that have kernels of a light color. These characters have of subsets of these traits, called factors, which are substantially cor-
been used to select promising genotypes for walnut breeding pro- related with each other, simultaneously affecting these traits to a
grams (Eskandari et al., 2005; Ebrahimi et al., 2011; Cosmulescu large extent. As a criterion to extract the main principal compo-
and Botu, 2012; Khadivi-Khub and Ebrahimi, 2015; Khadivi-Khub nents, eigenvalue greater than 1 was taken, and to determine which
et al., 2015a,b). Nut shape varied from oval to ellipsoidal elongated. of the PC scores accounted for the greatest amount of variation, the
Among genotypes, the shell seal was weak in 30 genotypes and eigenvalues of these components were compared for each trait.
strong in 31 genotypes. Shell texture was very smooth (two geno- For each factor, loading value above 0.54 was considered as sig-
types), smooth (17 genotypes), intermediate (37 genotypes), rough nificant that indicated seven components with explaining 74.20%
(four genotypes) and very rough (one genotype). Shell strength of the total variance (Table 5). This result revealed a great fruit
was low in 10 genotypes (soft shell), intermediate in 37 (moderate variation, indicating a genetic diversity between the genotypes,
shell), high in 12 (hard shell) and very high in 2 (very hard shell). In and suggested that the evaluation of different fruit characteristics
breeding programs, nuts with clean, strong, thin shells, with a tight remains necessary for a meaningful characterization.
seal and light kernels are most desirable (McGranahan and Leslie, A plot of the variance extracted by seven PCs showed that the
1990). The most important fruit traits for superior genotypes are variance explained by a single PC decreased strongly between PC1
presented in Table 3. and PC4, but from PC4 onward, the decrease became very slow
(Fig. 1). The first three components explained 41.29% of the total
3.2. Simple correlations between traits variability observed which 15.99% of the variance was accounted
for by the first principal component (PC1), followed by 12.97% for
Study of the relationships among characters can be beneficial the second (PC2) and 12.32% for the third principal component
for breeders in their breeding programs. Significant correlations (PC3). It is suggested that the use of first three principal components
were found among the studied characters (Table 4). Nut weight was will save considerable time (since a few characteristics are stud-
positively correlated with nut diameter (r = 0.71) and nut length ied) for characterization and identification of walnut genotypes in
(r = 0.55). A moderate positive coefficient of correlation was also breeding programs (Iezzoni and Pritts, 1991).
recorded among nut length and nut diameter (r = 0.43). Kernel The characters nut diameter, nut length, nut weight and ker-
weight was positively correlated with nut weight (r = 0.82), nut nel weight were found influential for the first component (PC1). In
diameter (r = 0.62) and nut length (r = 0.37), in agreement with this regard, this component could be called nut traits. The charac-
findings of others (Sharma and Sharma, 2001; Eskandari et al., ters that significantly loaded on PC2 were kernel filled and kernel
2005; Ebrahimi et al., 2011; Cosmulescu and Botu, 2012; Khadivi- plumpness. Therefore, PC2 represents the second most important
Khub and Ebrahimi, 2015; Khadivi-Khub et al., 2015a,b). Kernel factor, called kernel flesh. For the third component (PC3), the char-
percentage showed low negative correlations with shell thickness acters which showed a high loading included shell strength, shell
(r = −0.22), shell texture (r = −0.26), and shell strength (r = −0.38), thickness and integument thickness expresses another important
suggesting the thicker the nut, the lower the kernel percentage. factor, called shell traits. PC4 with 9.08% of total variance included
Negative correlation was found between kernel percentage and can be consisted as a factor of color traits, because this compo-
shell thickness by others (Sharma and Sharma, 2001; Cosmulescu nent was strong positively correlated with shell color and kernel
and Botu, 2012; Khadivi-Khub and Ebrahimi, 2015; Khadivi-Khub color. These results in some cases were in agreement with the
et al., 2015a,b). Furthermore, kernel percentage showed positive results reported in several Iranian walnuts (Ebrahimi et al., 2011;
correlations with kernel weight (r = 0.43), kernel filled (r = 0.35) Khadivi-Khub, 2014; Khadivi-Khub and Ebrahimi, 2015; Khadivi-
and kernel plumpness (r = 0.31), in agreement with findings of Khub et al., 2015a,b). Previously, PCA had been used to establish
others (Sharma and Sharma, 2001; Cosmulescu and Botu, 2012; genetic relationships among cultivars and genotypes, to study cor-
Khadivi-Khub and Ebrahimi, 2015; Khadivi-Khub et al., 2015a,b). relations among morphological traits and to evaluate germplasm
 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201 199

Table 5
Eigenvalues of the principal component axes from PCA of fruit characters in the studied walnut genotypes.

Component

Trait 1 2 3 4 5 6 7

Nut shape 0.05 0.01 0.05 −0.01 0.88a 0.05 0.01

Nut diameter 0.86a −0.07 0.04 0.09 −0.07 −0.11 0.09
Nut length 0.61a −0.25 −0.03 0.01 0.46 0.18 0.05
Nut weight 0.91a 0.13 0.20 0.06 0.14 0.09 0.01
Shell texture −0.06 0.16 0.15 0.23 0.14 0.79a −0.13
Shell color 0.03 0.11 0.07 0.81a −0.24 0.19 −0.04
Shell seal 0.03 −0.23 −0.03 −0.03 −0.03 0.67a 0.35
Shell strength 0.02 −0.06 0.85a −0.03 0.15 0.15 0.08
Shell thickness 0.15 −0.01 0.75a −0.11 −0.17 0.03 0.02
Integument thickness −0.02 −0.27 0.54a 0.42 0.18 −0.28 −0.16
Kernel weight 0.85a 0.34 −0.06 −0.17 −0.02 −0.05 −0.18
Kernel percentage 0.01 0.38 −0.37 −0.42 −0.27 −0.25 −0.35
Kernel filled −0.01 0.90a −0.11 0.00 0.05 0.04 −0.04
Kernel plumpness 0.17 0.84a −0.07 −0.12 −0.10 −0.07 0.13
Blank nut percentage −0.31 −0.31 0.45 0.05 0.30 −0.30 0.23
Kernel color 0.01 −0.23 −0.26 0.70a 0.22 0.00 0.23
Kernel shriveling 0.00 0.08 0.06 0.08 0.03 0.07 0.87a
Kernel vein -0.25 −0.39 −0.35 −0.13 0.28 −0.14 0.21
Eigenvalue 2.88 2.34 2.22 1.63 1.58 1.44 1.27
% of Variance 15.99 12.97 12.32 9.08 8.78 8.00 7.06
Cumulative% 15.99 28.97 41.29 50.36 59.14 67.14 74.20
a
Eigenvalues are significant ≥0.54.

Fig. 2. Scatter plot for the studied walnut genotypes based on the first two principal components (PC1/PC2). Numbers (1–61) represent genotypes. The superior genotypes
are determined by arrows.

of different plant species (Khadivi-Khub and Anjam, 2014; Khadivi- cophenetic matrix of dendrogram, indicating the goodness of fit of
Khub and Ebrahimi, 2015; Khadivi-Khub et al., 2015a,b). the cluster analysis. The cophenetic correlation coefficient is con-
PCA was used to construct two-dimensional scatter plot for sidered to be a very good representative of the data matrix in the
graphical overview of the relationships among genotypes. The pro- dendrogram if it is 0.90 or greater (Romesburg, 1990). The first clus-
jection of walnut genotypes on the PC1/PC2 plane is shown in Fig. 2, ter (I) included 20 genotypes, while the second cluster (II) consisted
based on the regression factor score values. The scatter plot showed of six genotypes. The third cluster (III) contained 35 genotypes and
geometrical distances among the genotypes reflecting phenotypic was split off into two distinct subclusters, defined as subclusters
and in this regard presumable genetic dissimilarity between them, III-A and III-B. Subcluster III-A consisted of 15 genotypes, while
so that genotypes were distributed in four sides of plot. 20 genotypes formed subcluster III-B. Study of walnut germplasm
The dendrogram obtained based on all of the fruit traits by revealed considerable phenotypic diversity among the genotypes
the cluster procedure showed numerous hierarchical levels of for most of the characters studied. Similar level of diversity and
the studied genotypes separated in three major clusters (Fig. 3). similar intervals of variation for fruit traits were found with differ-
The cophenetic correlation coefficient indicated high correlation ent walnuts evaluated (Sharma and Sharma, 2001; Eskandari et al.,
(r = 0.91) between the morphological distance matrix and the
200 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201

Khoram-12
Khoram-53
Khoram-26
Khoram-45
Khoram-7
I Khoram-59
Khoram-46
Khoram-5
Khoram-15
Khoram-41
Khoram-36
Khoram-55
Khoram-23
Khoram-28
Khoram-8
Khoram-13
Khoram-32
Khoram-61
Khoram-27
Khoram-35
Khoram-34
II Khoram-54
Khoram-37
Khoram-43
Khoram-6
Khoram-50
Khoram-1
Khoram-14
Khoram-2
Khoram-9
III-A Khoram-21
Khoram-40
Khoram-56
Khoram-20
Khoram-24
Khoram-44
Khoram-60
Khoram-19
Khoram-22
III Khoram-30
Khoram-31
Khoram-51
Khoram-57
Khoram-33
Khoram-52
Khoram-3
Khoram-48
Khoram-47
III-B Khoram-17
Khoram-38
Khoram-18
Khoram-58
Khoram-10
Khoram-42
Khoram-16
Khoram-39
Khoram-4
Khoram-29
Khoram-11
Khoram-25
Khoram-49

Fig. 3. Dendrogram of grouping for the studied walnut genotypes based on fruit traits. The superior genotypes are determined by arrows.

2005; Cosmulescu and Botu, 2012; Khadivi-Khub and Ebrahimi, ing programs. The present research was conducted in one of the
2015; Khadivi-Khub et al., 2015a,b). important walnut growing regions of Iran. We selected promising
genotypes through the use of criteria with high heritability, such
as leafing date and then nut and kernel characteristics. Thus, after
4. Conclusion pre-selections among many genotypes, all of the 61 studied geno-
types were late leafing and flowering and could be useful as a parent
In the areas with frequent late spring frost, such as many regions to improve leafing and flowering season of cultivars. Furthermore,
of Iran, late budbreak and flowering are the most important selec- genotypes 6, 14, 15, 17, 34, 37, 38, 40, 45, 47, 53, 54, 56 and 57 were
tion criteria. The knowledge of leafing date and fruit attributes of superior in terms of consistent high kernel percentages, large nut,
the walnut genotypes studied here could be useful to choose the shell softness, large kernel, light kernel color, ease of kernel remov-
appropriate ones to be grown or used as parents in future breed-
 A. Ebrahimi et al. / Scientia Horticulturae 193 (2015) 195–201 201

ing from nuts, and blank nut percentage and can be singled out for Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics
cultivation. software package for education and data analysis. Palaeontol. Electron.,
[Link] org/2001 1/past/issue1 [Link].
Hansche, P.E., Beres, V., Forde, H.I., 1972. Estimates of quantitative genetic
References properties of walnut and their implications for cultivar improvement. J. Am.
Soc. Hortic. Sci. 97, 279–285.
Achim, G., Botu, M., Botu, I., 2007. ‘Portval’—a new walnut rootstock. Acta Hortic. Iezzoni, A.F., Pritts, M.P., 1991. Applications of principal components analysis to
760, 549–553. horticultural research. HortScience 26, 334–338.
Akca, Y., Ozongun, S., 2004. Selection of late leafing, late flowering, laterally fruitful IPGRI, 1994. Descriptors for Walnut (Juglans spp.). International Plant Genetic
walnut (Juglans regia) types in Turkey. N. Z. J. Crop Hortic. Sci., 337–342. Resources Institute, Rome, Italy.
Aslantas, R., 2006. Identification of superior walnut (Juglans regia) genotypes in Khadivi-Khub, A., 2014. Genetic divergence in seedling trees of Persian walnut for
north-eastern Anatolia, Turkey. N. Z. J. Crop Hortic. Sci. 34, 231–237. morphological characters in Markazi province from Iran. Braz. J. Bot. 37,
Atefi, J., 1997. Study on phenological and pomological characters on walnut 273–281.
promising clones in Iran. Acta Hortic. 442, 101–108. Khadivi-Khub, A., Anjam, K., 2014. Morphological characterization of Prunus
Bayazit, S., Kazan, V., Gulbitti, C., Ayanoglu, H., Ergul, A., 2007. AFLP analysis of scoparia using multivariate analysis. Plant Syst. Evol. 300, 1361–1372.
genetic diversity in low chilling requiring walnut (Juglans regia L.) genotypes Khadivi-Khub, A., Ebrahimi, A., 2015. The variability in walnut (Juglans regia L.)
from Hatay, Turkey. Sci. Hortic. 111, 394–398. germplasm from different regions in Iran. Acta Physiol. Plant. 37, 57.
Botu, M., Tudor, M., Botu, I., Cosmulescu, S., and Papachatzis, 2010. Evaluation of Khadivi-Khub, A., Ebrahimi, A., Mohammadi, A., Kari, A., 2015a. Characterization
walnut cultivars in the conditions of the Oltenia’s hill area regarding and selection of walnut (Juglans regia L.) genotypes from seedling origin trees.
functioning potential. Analele Universitatii din Craiova, Biologie, Horticultura, Tree Genet. Genomes 11, 54.
Tehnologia prelucrarii produselor agricole, Ingineria Mediului 15: 94–103. Khadivi-Khub, A., Ebrahimi, A., Sheibani, F., Esmaeili, A., 2015b. Phenological and
Cosmulescu, S., Botu, M., 2012. Walnut biodiversity in South-Western pomological characterization of Persian walnut to select promising trees.
Romania—resource for perspective cultivars. Pak. J. Bot. 44, 307–311. Euphytica, [Link]
Cosmulescu, S., Botu, M., Trandafir, I., 2010. Mineral composition and physical Korac, M., Cerovic, S., Golosin, B., Miletic, R., 1997. Collecting, evaluation and
characteristics of walnut (Juglans regia L.) cultivars originating in Romania. utilization of walnut (Juglans regia L.) in Yugoslavia. Plant Genet. Resour.
Selçuk Tarım Gıda Bilimleri Dergisi 24, 33–37. Newsl. 111, 72–74.
Ebrahimi, A., Fatahi, R., Zamani, Z., 2011. Analysis of genetic diversity among some Mantel, N., 1967. The detection of disease clustering and generalized regression
Persian walnut genotypes (Juglans regia L.) using morphological traits and SSRs approach. Cancer Res. 27, 209–220.
markers. Sci. Hortic. 130, 146–151. McGranahan, G.H., Leslie, C., 1990. Walnuts (Juglans). Acta Hortic. 290, 905–951.
Eskandari, S., Hassani, D., Abdi, A., 2005. Investigation on genetic diversity of Norusis, M.J., 1998. SPSS/PC Advanced Statistics. SPSS Inc., Chicago, United States.
Persian walnut and evaluation of promising genotypes. Acta Hortic. 705, Rana, J.C., Singh, D., Yadav, S.K., Verma, M.K., Kumar, K., Predheep, K., 2007. Genetic
159–166. diversity collected and observed in Persian walnut (Juglas regia L.) in the
FAO, 2012. FAO Production Yearbook. Food and Agriculture Organization, Rome, western himalayan region of India. Plant Genet. Resour. Newsl. 151, 68–73.
Italy. Rohlf, F.J., 2000. NTSYS-pc Numerical Taxonomy and Multivariate Analysis System.
Fjellstrom, R.G., Parfitt, D.E., 1994. Walnut (Juglans spp.) genetic diversity Version 2. 1. Exeter Software, Setauket, NY.
determined by restriction fragment length polymorphisms. Genome 37, Romesburg, H.C., 1990. Cluster Analysis for Researchers. Krieger Publishing,
690–700. Malabar, FL, USA.
Forde, H.I., 1979. Walnuts. In: Janick, J., Moore, J.N. (Eds.), Advances in Fruit Sen, S.M., Tekintas, F.E., 1992. A study on the selection of Adilcevaz walnuts. Acta
Breeding. , 2nd ed. Purdue University Press, West Lafayette, Indiana, United Hortic. 317, 171–174.
States, pp. 439–455. Sharma, O.C., Sharma, S.D., 2001. Genetic divergence in seedling trees of Persian
Francesca, P.I., Pamfil, D., Raica, P., Petricele, I.V., 2010. Assessment of the genetic walnut (Juglans regia L.) for various metric nut and kernel characters in
variability among some Juglans cultivars from the Romanian National Himachal Pradesh. Sci. Hortic. 88, 163–171.
Collection at SCDP valcea using RAPD markers. Rom. Biotechnol. Lett. 15, Sharma, S.D., Sharma, O.C., 1998. Studies on the variability in nuts of seedling
41–49. walnut (Juglans regia L.) in relation to the tree age. Fruit Variety J. 52, 20–23.
Germain, E., 1997. Genetic improvement of the Persian walnut (Juglans regia L.). Yarilgac, T., Koyuncu, F., Koyuncu, M.A., Kazankaya, A., Sen, S.M., 2001. Some
Acta Hortic. 442, 21–31. promising walnut selections (Juglans regia L.). Acta Hortic. 544, 93–100.
Godeanu, I., Baciu, A., Achim, G., Cosmulescu, S., 2004. Achievements and Zhadan, V.M., Strukov, M.V., 1977. Breeding walnut for fruit size. Plant Breed. 47,
perspectives in creating rootstocks and in the technology for obtaining grafted 918–925.
walnut trees in Romania. Acta Hortic. 658, 479–486.