Published April 28, 2015
RESEARCH
Field Screening of Salinity Tolerance
in Iranian Bread Wheat Lines
Somaye Sardouie-Nasab, Ghasem Mohammadi-Nejad,* and Babak Nakhoda
ABSTRACT
Soil and water salinity are major constraints for
wheat production in Iran. High genetic diversity for salinity tolerance has been observed
in Iranian bread wheat (Triticum aestivum L.)
genotypes. Hence, these genotypes can be
evaluated under field and controlled conditions
for identifying new genotypes with better performance under salt stress conditions. In a 2-yr
study, we evaluated 97 promising Iranian lines
with different levels of salinity tolerance along
with three tolerant cultivars as checks under
field conditions in Yazd province of Iran. Plant
materials were grown under saline and normal
field conditions (electrical conductivity [EC] of
10.0 and 3.0 dS m-1, respectively). Significant
genetic variation (P < 0.01) in salinity tolerance
was observed among wheat genotypes. In this
study, an equation was developed for estimating a stress tolerance score (STS). The results of
the equation were identical to those of multivariate analyses. The STS equation is much easier
to use than complicated multivariate analyses.
Therefore, it is suggested as a screening tool for
identification of salt-tolerant wheat genotypes.
Ultimately, ten lines with higher tolerance to salt
stress, in comparison with tolerant checks, were
detected. These lines can be used in wheat
breeding programs for salt-affected areas.
S. Sardouie-Nasab and G. Mohammadi-Nejad, Dep. of Agronomy and
Plant Breeding, College of Agriculture, Shahid-Bahonar Univ. of Kerman.
Kerman P.O.B. 76169-133- Iran; B. Nakhoda, Dep. of Molecular Physiology, Agricultural Biotechnology Research Institute of Iran. Received 2
June 2013. *Corresponding author: ([email protected]).
Abbreviations: FA1, first factor; FA2, second factor; FGN, number
of filled grain/spike; GMP, geometric mean productivity; HM, harmonic mean; MP, mean productivity; RCBD, randomized complete
block design; S, sensitive; SES, standard evaluation system; SSI, stress
susceptibility index; STI, stress tolerance index; STS, stress tolerance
score; T, tolerant; Tn, number of tillers per plant; TOL, stress tolerance;
TGW, 1000-grain weight; YS, the yield of lines under stress conditions;
YP, the yield of lines under normal conditions.
S
alinity stress represents a major constraint to food production (Yokoi et al., 2002) because it limits crop yields and
restricts the use of previously uncultivated lands. More than 800
million hectares of lands throughout the world are salt-affected
(including both saline and sodic soils), equating to more than 6%
of the world’s total land area (FAO, 2010). It has been estimated
that approximately 20% of agricultural lands and 50% of cropping lands in the world suffer from soil salinity (Flowers and Yeo,
1995). Due to the high costs of amelioration of saline soils, among
various strategies to overcome salinity problems, the possibility
of selection and breeding for enhanced salinity tolerance in crop
species has received considerable attention as it is an economic
and efficient alternative (Ashraf, 2009). Improving salinity tolerance of the world’s two staple crops, wheat and rice, are important
goals as the world’s population is increasing more quickly than
the agricultural lands to support it (FAO, 2010). Wheat (Triticum
Published in Crop Sci. 54:1489–1496 (2014).
doi: 10.2135/cropsci2013.06.0359
Freely available online through the author-supported open-access option.
© Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA
All rights reserved. No part of this periodical may be reproduced or transmitted in any
form or by any means, electronic or mechanical, including photocopying, recording,
or any information storage and retrieval system, without permission in writing from
the publisher. Permission for printing and for reprinting the material contained herein
has been obtained by the publisher.
crop science, vol. 54, july– august 2014 www.crops.org1489
aestivum L.) is one of the most important food crops of the
world, and its productivity directly affects human survival
and quality of life. Improving salt tolerance of bread wheat
and increasing its productivity are the major objectives
of our breeding programs. To differentiate stress-tolerant
cultivars, several selection indices described below have
been suggested on the basis of mathematical relationships
between stress and nonstress conditions (Huang, 2000).
Tolerance (TOL; Clarke et al., 1992), mean productivity
(MP; Mccaig and Clarke, 1982), stress susceptibility index
(SSI; Fischer and Maurer, 1978), geometric mean productivity (GMP), and stress tolerance index (STI; Fernandez,
1992) have all been employed under various conditions.
Fischer and Maurer (1978) explained that cultivars with an
SSI of less than a unit are stress tolerant, since their yield
reduction under stress conditions is smaller than the mean
yield reduction of all cultivars (Bruckner and Frohberg,
1987). Mean productivity, GMP, harmonic mean (HM),
and STI were reported as preferred criteria in selection of
drought-tolerant barley genotypes by Baheri et al. (2003).
To improve wheat yield for stress-prone environments,
it is necessary to identify selection indices able to distinguish high yielding wheat genotypes. Crop plants can be
divided into four groups based on their yield responses to
stress conditions: (i) cultivars producing high yield under
both stress and nonstress conditions (Group A), (ii) cultivars with high yield under nonstress (Group B) or (iii)
stress (Group C) conditions, and (iv) cultivars with poor
performance under both stress and nonstress conditions
(Group D) Fernandez (1992). It is reported that under
moderate drought stress, GMP, STI, and MP were the
most effective indices for identifying bread wheat cultivars
with high yields under both stress and nonstress conditions (Sio-Se Mardeh et al., 2006). The main objective
of this study was to identify the best wheat genotypes for
salt-affected areas of Iran and compare selection indices
for their relative effectiveness.
MATERIALS AND METHODS
The genetic materials used in this study consisted of 97 bread
wheat lines (selected from 400 Iranian bread wheat genotypes
obtained from different wheat breeding programs of the Seed
and Plant Improvement Institute of Iran) varying in levels of
salinity tolerance, as well as three Iranian salt-tolerant bread
wheat cultivars, namely, Arg, Bam, and Kavir as tolerant
checks. (Table 1). Two field experiments were conducted in
two consecutive growing seasons (2008–2010). In the first year
(2008–2009), wheat genotypes were evaluated in a field experiment using an incomplete block design (LATTICE) with two
replications under salt stress and control conditions. Control
plots were irrigated with nonsaline water (EC of 3.0 dS m-1)
weekly, while the stress treatment plots were irrigated with
saline water (EC of 10.0 dS m-1) weekly. In the second year,
field screening for salt tolerance was done only under saline
field conditions (EC of 10.0 dS m-1), based on an augmented
1490
Table 1. Lines pedigree.
No.
lines
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
www.crops.org
Pedigree
Cham4/Tam200//Del 483/3/Mirtos
Cham4/Tam200//Del 483/3/Mirtos
Cham4/Tam200//Del 483/3/Mirtos
Cham4/Tam200//Del 483/3/Mirtos
Alamoot*2/Kavir
Zarrin/Shiroodi/6/Zarrin/5/Omid/4/Bb/Kal//Ald/3/Y50E/
Kal*3//Emu
Zarrin/Shiroodi/3/zarrin//Ombu1/Alamo
Zarrin*2/3/Alvand//Aldan”s”/IAS58,4072–48
Alvand*2//Opata*2/Wulp
Alvand*2//Opata*2/Wulp
Alvand*2/4/Kal/Bb//Cj“s”/3/Hork“s”
Owl*2/Kavir
Owl*2/Kavir
Owl*2/Shiroodi
Owl/Shiroodi/3/Owl//Opata*2/Wulp
Owl/Shiroodi/3/Owl//Opata*2/Wulp
Owl/Kavir/3/Owl//Vee/Nac
Owl/Kavir/3/Owl//Vee/Nac
1–68–120/1–68–22//Mirtos/3/1–68–120/1–68–22
Passarinho//Vee/Nac
Passarinho//Vee/Nac
Guadalop/Falat
494J6.LL/Roller//Mv17
Rsh/Tam200//Alvand
Marvdasht/Owl
TX62A4793/CB809/5/Gds/4/Anza/3/Pi/Nar//Hys/6/
Passarinho/7/Alvand
Marvdasht/3/Emu”s”/Tjb84–1543//1–27–7876/Cndr“s”
Pishtaz/3/Emu”s”/Tjb84–1543//1–27–7876/Cndr”s”
Pishtaz//Ald“s”/Snb“s”
Pishtaz/4/Bloudan/3/Bb/7C*2//Y50E/3*Kal
Shiraz/5/Nvd/4/Omid//H7/4P839/3/Omid/Tdo
4777//Fkn/Gb/3/Vee“s”/4/Buc“s”/5/1–66–44/6/Fertillo/
Vee#5/7/Pishtaz
Bloyka/7/T.Aest/5/Ti/4/La/3/Fr/Kad//Gb/6/F13471/crow”s”/8/
Mahdavi
1–65–55/5/Pewee“s”/Azd/4/Anza/3/Pi/Nar//Hys/6/Cocoraque
75/7/Bloyka
MV17/Alvd//Chamran/3/Pishtaz
Mrn/Catbird/4/Alvand//Aldan”s”/Ias 58
Kal/Bb//Cj “s”/3/Hork “s”/4/2*Alvd//Aldan/Ias 58
Mv17/Shiraz
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Tajan/7/
Milan/..
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Kauz*2/
Opata//….
Kauz/Sorkhtokhm//Hys//Drc*2/7C/3/2*Rsh/4/Bank“s”/Vee “s”
Kauz/Sorkhtokhm//Hys//Drc*2/7C/3/2*Rsh/4/Bank”s”/Vee “s”
Kauz/Sorkhtokhm//Mahooti/3/Bank“s”/Vee “s”
Marvdasht//Tui“s”/Star
Marvdasht//Tui“s”/Star
Bloudan/3/Bb/7C*2//Y50E/Kal*3/4/KRL.14
DH-line
Bloudan/3/Bb/7C*2//Y50E/Kal*3/4/Sholeh
Azd//Tob/Chb/3/Emu”s”/Tjb84/4/Bloudan/3/Bb/7c*2//Y50E/
Kal*3
Guadalop/Falat
Alamoot//Opata*2/Wulp
crop science, vol. 54, july– august 2014
Table 1. Continued.
No.
lines
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Pedigree
Alamoot/4/Bloudan/3/Bb/7c*2//Y50E/Kal*3
Alamoot/4/Bloudan/3/Bb/7c*2//Y50E/Kal*3
Azd//Tob/Chb/3/Emu”s”/Tjb84/4/Alvand//Aldan”s”/
IAS58,4072–48
1–72–92/Col.No.3617//Owl
Bow”s”/Vee”s”//1–60–3/7/T.Aest/5/Ti/4/La/3/Fr/Kad//Gb/6/
F13471/crow”s”/8/Flt/Attila
Alvd//Aldan/Ias 58/3/2*Rsh
Alvd//Aldan/Ias 58/3/1–60–3/5/Kal/Bb//Cj “s”/3/Hork“s”/4/
Alvd//Aldan/Ias58
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Tajan/7/
Milan/..
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Tajan/7/
Milan/..
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Tajan/7/
Milan/..
DH2–390–1563 F3Gds/4/Anza/3/Pi//Hys/5/1–6/6/Kauz*2/
Opata//….
Kauz/Sorkhtokhm//Hys//Drc*2/7C/3/2*Rsh/4/Bank”s”/Vee “s”
Kauz/Sorkhtokhm//Sakha 8/3/Bank“s”/Vee “s”
Kauz/Sorkhtokhm//Sakha 8/3/Bank“s”/Vee “s”
Kauz/Sorkhtokhm//Mahooti/3/DH-209–1557 F3,Vee “s”/
Nac//1–66–22
Kauz/Sorkhtokhm//Mahooti/3/DH-209–1557 F3,Vee “s”/
Nac//1–66–22
T.AestxTi(La(Frkal.xGb)) = (1–66–22)/5/KRL.1–4)/6/Bank “s”/
Vee“s”
Pishtaz//Karchia
Pishtaz//Karchia
Pishtaz//Karchia
DH-line
DH-line
Cereal research collection, Accession no:3101
DH-line
DH-line
Cereal research collection, Accession no: 2591
Cereal research collection, Accession no:2592
Cereal research collection, Accession no:2594
Cereal research collection, Accession no:2605
Cereal research collection, Accession no:2607
Cereal research collection, Accession no:2614
Cereal research collection, Accession no:2695
Cereal research collection, Accession no:2696
Cereal research collection, Accession no:2764
Cereal research collection, Accession no:2766
Cereal research collection, Accession no:2776
Cereal research collection, Accession no:2799
Cereal research collection, Accession no:2812
Cereal research collection, Accession no:2815
Cereal research collection, Accession no:2933
Cereal research collection, Accession no:2940
Cereal research collection, Accession no:2941
Cereal research collection, Accession no:2970
Cereal research collection, Accession no:2998
Cereal research collection, Accession no:3001
Cereal research collection, Accession no:3102
(Arg) tolerant check
(Bam) tolerant check
(Kavir) tolerant check
crop science, vol. 54, july– august 2014 design. In saline conditions, plant materials were grown in a
sandy loam soil with a pH of 7.4 and homogenized salinity level
of 12.5 and 10.0 dS m-1 for soil and irrigation water, respectively, in Yazd, located in the South-Eastern part of Iran (1200
m above sea level, 32°N, 54.4°E) with a hot and arid climate.
Sowing was done in mid-November in both experiments with
a density of 400 plants per square meter. Each experimental plot
consisted of four 1-m long rows with 25 cm distance between
rows. Phosphorous fertilizer was applied as 50 kg ha-1 P2O5
before sowing; nitrogen fertilizer was applied as 40 kg ha-1 N
urea at sowing and tillering. Weeds were manually controlled
during the growing season.
Ten plants were randomly selected from each plot to measure the number of tillers per plant (Tn), 1000-grain weight
(TGW), and number of filled grain/spike (FGN). After removing the borders, the whole plot was harvested to calculate the
grain yield (t ha-1). A standard evaluation system (SES; IRRI,
1996) was used to assign a tolerance score to each genotype,
where ‘‘tolerant’’ lines had an SES < 3 (no injury symptoms),
‘‘moderately tolerant’’ genotypes scored between 3 and 5, and
‘‘sensitive’’ ones scored > 5.
Stress resistance indices were calculated using the following formulas:
SSI = 1 - YS / YP S I:
SI
Stress Intensity = 1- Ys (Fischer and Maurer, 1978) [1]
YP
where YS is the yield of lines under stress, YP the yield of lines
under normal conditions (t ha-1), Ys and YP are the mean yields
of all lines under stress and nonstressed conditions, respectively,
and S I is the stress intensity.
TOL = YP – YS (Hossain et al., 1990) 0.5
GMP = (YP ´YS )
(Fernandez, 1992) [2]
[3]
STI = (YS + YP ) / (YP )2 (Fernandez, 1992) [4]
MP = (YS + YP ) / 2 (Hossain et al., 1990) [5]
HM = 2 (YP ´YS ) / (YP + YS ) (Kristin et al., 1997)
[6]
b: The coefficient of linear regression of grain
yield of a genotype in each environment on the
environmental index (mean yield of all genotypes
at any environment) (Bansal and Sinha, 1991). [7]
The analysis of traits was done based on a LATTICE design
using the Lattice procedure by SAS (SAS Institute, 2004), and
the efficiency of LATTICE was not higher than randomized
complete block design (RCBD); therefore, the analysis of variance was a combined analysis over the salinity levels (saline and
normal) from 2008 to 2009 according to RCBD.
www.crops.org1491
Spearman’s rank correlation coefficient between the ranks
of genotypes based on their grain yields under salinity stress in
both years was calculated using Microsoft Excel.
Table 2. Combined analysis for traits under stress and nonstress conditions from 2008 to 2009.†
Mean square
Factor Analysis
Df
Tn
Grains/
spike
1
2
99
99
196
64.35*
1.71
1.19**
0.59**
0.07
2687.04ns
167.14
185.17**
46.017*
32.61
S.O.V
To use all stress-tolerance/sensitive indices simultaneously, a
factor analysis (FA) based on a correlation matrix of genotypes
with varimax rotation was conducted using the SPSS package
ver. 19.00 (Yaremko et al., 1986). Varimax rotation is an orthogonal rotation method that assumes the factors in the analysis are
uncorrelated (Brown, 2009; Mohammadi et al., 2011).
We used a biplot of the two first factors to show the distribution pattern of the genotypes in a graphical view.
Salinity level (SL)
Rep/SL
Genotype (G)
G ´ SL
Error
TGW
1068.57*
41.09
46.05**
5.09*
3.61
Yield
629.06**
2.67
4.61**
1.280ns
1.023
* Significant at the 0.05 probability level.
** Significant at the 0.01 probability level.
†
Tn: the number of tillers per plant, grains/spike: the number of grains per spike,
TGW: 1000-grain weight, S.O.V.: source of variation, ns: not significant.
Correlation Analysis
The most suitable index for selecting stress-tolerant genotypes
is an index that has strong correlation with grain yield under
stress and nonstress conditions (Farshadfar et al., 2001). Therefore, evaluating correlations between stress tolerance indices
and grain yield in both environments can lead to identification
of the most suitable index. To determine the most desirable salt
tolerance criterion, the correlation coefficient between YP, YS,
and other quantitative indices of salt tolerance was performed
using SPSS package ver. 19.00.
Discriminant Analysis Linear Equation
Discriminant function is a criterion for better selection of the
genotypes based on their phenotypic performance using several variables simultaneously (Smith, 1936). This analysis
is being used to determine which continuous variables discriminate between two or more naturally occurring groups.
To determine the most desirable salt tolerance criterion and
validation of detected tolerant and sensitive genotypes, discriminant function analysis was performed using SPSS
package ver. 19.00. Contrasting genotypes, out of 97 lines,
were selected based on the following tolerance index,
D = v1X 1 + v 2 X 2 + v 3 X 3 = ........vi X i , where D is the discriminant function, v is the discriminant coefficient or weight for
that variable, x is the salinity tolerance indices as variables, and
i is the number of predictor variables.
The correlation of each variable with each discriminant
function based on the structure matrix was used to create the
discriminant function (Eq. [8]). These Pearson coefficients are
structure coefficients or discriminant loadings and function
like factor loadings in factor analysis. By identifying the largest loadings for each discriminant function, researchers can
gain insight into how to name each function. Nowadays, many
researchers are using structure matrix correlations due to their
accuracy in comparison with the standard canonical discriminant function coefficients.
D = (0.88 ´ HM) + (0.83 ´ GMP) + (0.8 ´ STI)
+ (0.7 ´ MP) + (-0.56 ´ SSI) + (-0.26 ´ TOL )
[8]
The coefficients in the equation are the structure coefficients
or discriminant loadings which were extracted from the discriminant procedure in SPSS19.0 package. Wilks’ lambda also
1492
showed significance of the discriminant function and was calculated using SPSS package ver. 19.0.
Since multivariate techniques were too complicated, the
following equation was proposed (Abdolshahi et al., 2013):
STS = GMP + STI + HM + MP - TOL - SSI - b
[9]
Considering salinity-tolerance/sensitivity equations, large
values for MP, STI, GMP, and HM and small values for SSI,
TOL, and b, represent relatively more tolerance to salinity stress.
Thus, MP, STI, GMP, and HM have positive and SSI, TOL, and
b have negative coefficients. Equation [9] is not accurate for raw
data (Abdolshahi et al., 2013); hence, all indices in this equation
were standardized as follows:
Z ij =
X ij - X i
Si
Where Zij is the standard score for jth genotype in the ith index,
Xij is the raw data of jth genotype in the ith index, and Si is
the standard deviation of the ith index. After standardization of
indices, STS was calculated.
RESULTS
The analysis of variance was combined analysis over the
salt stress treatments (saline and control) from 2008 to
2009 based on RCBD. Salinity significantly affected all
of the measured traits except FGN. Salt stress dramatically
affected grain yield of all genotypes (Table 2). Significant
differences were observed among genotypes. The interaction between genotypes and salt stress were significant for
Tn, TGW, and grain/spike (Table 2). Some wheat genotypes showed much better performance than the tolerant cultivars under salt stress conditions (Supplementary
Table 1). Interestingly, these lines performed very well
under nonstress conditions too (Supplementary Table 1).
There was a strong Spearman’s rank correlation coefficient
between the rank of genotypes based on their grain yields
under salinity stress in both years (r = 0.99**). This can be
considered as an indicator of stability and low interaction
www.crops.org
crop science, vol. 54, july– august 2014
Figure 1. Biplot analysis graph compromised from two first factors of 100 bread wheat genotypes. The numbers in the figure show the
genotype position in the biplot. Highlighted genotypes by circle are the most productive ones under stress conditions in both years.
with years. Lines No. 49, 50, 54, 59, 65, 67, 72, 73, 75,
and 77 were the most productive ones under stress conditions in both years (Supplementary Table 1 and Fig. 1).
For better evaluation of 97 bread wheat lines for salinity
tolerance, six selection indices, including STI, MP, GMP,
TOL, SSI and HM, were used. Tolerance indices were calculated on the basis of grain yield of genotypes over the
salinity levels from 2008 to 2009 (Supplementary Table 1).
As shown in Supplementary Table 1, the greater the TOL
value, the larger yield reduction under stress conditions
and the higher salinity sensitivity. A selection based on
minimum yield reduction under stress conditions in comparison with nonstress conditions (TOL) failed to identify
the most tolerant genotypes (Rizza et al., 2004). Rosielle
and Hamblin (1981) reported that selection based on the
tolerance index often leads to selecting cultivars which
have low yields under nonstress conditions. The greater
TOL and SSI values, the greater sensitivity to stress, thus a
smaller value of these indices is favored. Lines with lower
SSI and TOL values than tolerant check varieties (lines
number 98, 99, 100) were identified as the most tolerant
lines. Line 18 had the highest SSI and TOL value was the
most sensitive line (Supplementary 1). The tolerance indices (GMP, STI, HM, and MP) measure the higher stress
tolerance and yield potential. Genotypes such as lines No.
26, 34, 48, 49, 51, 54, 55, 59, and some other lines that are
presented in Supplementary Table 1 were the most tolerant
crop science, vol. 54, july– august 2014 lines and had lower values of SSI and TOL, whereas some
lines, such as 2, 3, 9, 10, 12, 14, 18, 19, 21, 29, 35, 38, and
56, were the least relative tolerant genotypes. Lines 26, 34,
55, 59, 67, 69, 71, 73, 80, and 81 are the most tolerant
lines based on all quantitative indices. To determine the
most desirable salt tolerance criteria, the correlation coefficient between YP, YS, and other quantitative indices of
salt tolerance were calculated (Table 3). Number of filled
grain/spike, STI, HM, MP, and STS had a positive highly
significant correlation with both YS and YP, whereas, the
correlation between YS with SSI and TOL indices was
negative. Stress tolerance score had a positive highly significant correlation with GMP, STI, HM, and MP and a
negative correlation with TOL and SSI (Table 3). Mean
productivity, GMP, and STI were the better predictors of
YP and YS than other indices under both control and stress
conditions. Several reports have introduced these indices
as the most suitable criteria for selecting the best genotypes
for stress-prone areas (Fernandez, 1992; Sio-Se Mardeh
et al., 2006). As reported by other researchers, grain yield
under nonsaline conditions (YP) was positively correlated
with grain yield under saline conditions (YS), (Golabadi et
al., 2006; Mohammadi et al., 2011; Table 3). Nevertheless,
some reports showed a negative correlation between YS
and YP (Sio-Se Mardeh et al., 2006). The good responses
shown by some cultivars under stress conditions could be
ascribed to adaptation mechanisms (Clarke et al., 1992).
www.crops.org1493
Table 3. Simple correlation coefficients between tolerance and susceptibility indices of wheat lines from 2008 to 2009.†
YS
TOL
SSI
GMP
STI
HM
MP
b
STS
YP
YS
TOL
SSI
GMP
STI
HM
MP
b
0.572**
0.580**
0.081ns
0.816**
0.811**
0.731**
0.904**
-0.269**
0.606**
1
-0.336**
-0.753**
0.938**
0.932**
0.973**
0.868**
0.132ns
0.990**
1
0.841**
0.005ns
0.005ns
-0.127ns
0.175ns
-0.441**
-0.287**
1
-0.494**
-0.482**
-0.598**
-0.344**
-0.349**
-0.719**
1
0.991**
0.991**
0.983**
-0.023ns
0.949**
1
0.981**
0.977**
-0.014ns
0.940**
1
0.950**
0.033ns
0.978**
1
-0.094ns
0.883**
1
-0.003ns
** Significant at the 0.01 probability level.
†
YP: the yield of lines under normal conditions, YS: the yield of lines under stress, TOL: stress tolerance, SSI: stress susceptibility index, GMP: geometric mean productivity,
STI: stress tolerance index, HM: harmonic mean, MP: mean productivity, b: The coefficient of linear regression; STS: stress tolerance score.
Correlation analysis showed that TOL had a positive correlation (r = 0.58**) with YP and a negative correlation
(r = –0.33**) with YS (Clarke et al., 1992, Table 3). Since
MP is mean production under both salt stress and nonstress
conditions (Rosielle and Hamblin, 1981), it was highly correlated with YP and YS (Table 3). Hossain et al. (1990) used
MP as a resistance criterion for wheat cultivars under moderate stress conditions. Stress susceptibility index showed
a negative correlation (r = -0.75**, p < 0.01) with grain
yield under salt stress but no significant correlation was
found between this index and grain yield under nonstress
conditions (Table 3). Stress susceptibility index has been
widely used by researchers to identify sensitive and tolerant genotypes (Clarke and Towenley-Smith, 1984; Clarke
and Duncan, 1993; Winter et al., 1988). Different indices would not result in the same outcome. To employ all
indices simultaneously, multivariate statistics such as factor
analysis with Varimax rotation was performed. The first
two factors explained 99.4% of total variation between the
data (data not shown). Thus, a biplot was drawn based on
the first two factors (Fig. 1). The first factor (FA1), expressed
70.3% of total variation and had a high positive relationship with YS, YP, STI, GMP, HM, and MP and a negative
coefficient with SSI. Therefore, the first factor was named
as yield potential and salt tolerance. The higher scores for
FA1 were in accordance with the higher rank of salinity
tolerance, whereas low scores for FA1 showed salinity-sensitive genotypes (data not shown). The second factor (FA2)
accounted for 29% of total variation and had high communalities with TOL and SSI and a negative coefficient with
YS, GMP, STI, and HM that was named as salt sensitive
(data not shown). This factor was able to distinguish lowyielding genotypes under stress conditions with high SSI
and TOL values. Regarding the factor analysis, results for
the indices and biplot were displayed based on the first two
factors. The scores for the first two factors for all the genotypes are presented in Supplementary Table 1. The higher
scores for FA1 and lower scores for FA2 (part A from Fig.
1a and Supplementary Table 1) were in accordance with
the higher rank of salinity tolerance. Whereas, low scores
1494
for FA1 and FA2 showed salt-sensitive genotypes (part D
from Fig. 1 and Supplementary Table 1). Genotypes with
lower FA1 and higher FA2 scores had low grain yields (part
C from Fig. 1 and Supplementary Table 1). Since the first
factor has more contribution in total variation (expressed
70.3% of total variation), genotypes in part B of the biplot
(having high scores of FA1 and FA2) had a moderately
tolerant reaction to saline conditions and high production
potential under control conditions (part B from Fig. 1 and
Supplementary Table 1). These results are confirmed by
the findings of other researchers on durum wheat genotypes (Golabadi et al., 2006; Talebi et al., 2009). The
results of factor analysis were also confirmed by discriminant function analysis. Wilks’ lambda indicated a highly
significant function (Wilks’ lambda = 0.108, p < 0.000);
therefore, the promising lines which were chosen as highly
tolerant lines showed better scores than the tolerant checks
(Supplementary Table 2). The results of Eq. [9] and ranking of genotypes based on this equation are presented in
Supplementary Table 1. This ranking was very similar to
the rankings of the other methods.
DISCUSSION
Wheat breeders have made significant improvements in
adaptation of wheat to stress-prone environments (Trethowan et al., 2002; Lantican et al., 2003). This success has
largely been achieved through field-based empirical selection for stress tolerance. Simultaneous analysis of multiple
parameters to increase the accuracy of the genotype ranking is the most important advantage of using a multivariate analysis in the evaluation of salt tolerance (Zeng et al.,
2002). These studies may lead us to a better understanding of the response of particular genotypes under particular environments. Field screening in salt affected areas
as a routine part of the plant breeding programs would
accelerate the identification of salinity tolerance in lines
that may be eventually released as new cultivars. Large
genotype ´ environment interaction for grain yield was
observed in the stress-prone areas. This interaction complicates selection of genotypes suitable for a wide range
www.crops.org
crop science, vol. 54, july– august 2014
of target environments; hence, it is essential that all the
experiments be conducted under appropriate field environments in target sites and repeated across seasons.
In this study, the interaction between genotype and
salinity was significant for all of the measured traits. This
variation can be explained, in part, by the fact that traits suitable for a given salt-affected site with its own salinity level
and its specific climatic conditions may be unsuitable in
another environment (Austin, 1987; Van Ginkel et al., 1998).
Since using high-yielding and salt-tolerant varieties as parents in crossing blocks for salinity can significantly increase the effectiveness of breeding programs in
developing high-yielding wheat cultivars for salt-affected
areas, it may be efficient to examine these genotypes first
for salinity tolerance and high-yield potential, as well
as for other important characters such as grain yield. In
this study, among all the genotypes and according to all
indices, lines No. 26, 34, 55, 59, 67, 69, 71, 73, 80, and
81were selected as the most salt-tolerant lines. Lines which
showed higher scores for the first factor and lower scores
for the second factor in factor analysis are best fitted for
both stress and nonstress environments. Hence, these
lines can be recommended to be used as donor parents
for salt tolerance genes in wheat breeding programs for
salt-affected areas of Iran (Supplementary Table 1 and
Fig. 1). Kaya et al. (2002) reported that genotypes with
larger PCA1 and lower PCA2 scores had higher grain
yields (stable genotypes) than the genotypes which had
lower PCA1 and larger PCA2 scores (unstable genotypes).
Based on present results, a positive and significant correlation was detected between GMP, STI, and MP (Table
3); therefore, these indices can produce similar results.
Thus, we concluded that selection based on GMP, STI,
and MP indices can result in high-yielding genotypes for
both normal and stress conditions. Based on the findings
of this study, the above mentioned indices are much better
predictors of YP and YS than other indices under both
normal and stress conditions. Therefore GMP, STI, and
MP favor genotypes with high-yield potential and TOL
and SSI favor genotypes with low-yield potential. Thus,
different indices would not result in the same outcome. To
employ all indices simultaneously, multivariate statistics,
such as factor analysis, was performed. Considering the
results of factor analysis and biplot display based on the
first two factors, lines with the higher scores for FA1 and
low scores for FA2 (part A from Fig. 1 and Supplementary Table 1) were in accordance with the higher rank of
salinity tolerance and can be considered as promising lines
that contain genes for stress tolerance and high production potential under saline conditions. Lines with grain
yield higher than tolerant checks under saline conditions
were identified as the most productive lines under saline
conditions and recommended as promising lines for saline
lands (Supplementary Table 1 and Fig. 1). The results of
crop science, vol. 54, july– august 2014 discriminant function analysis were also in concordance
with the explained findings (Supplementary Table 2).
Owing to complications in using multivariate analyses,
Eq. [9] was developed and proposed. The results of this
equation and other used statistical methods were almost
the same with respect to the order of genotypes according
to their salinity stress tolerance. Equation [9] is much easier
to use than multivariate techniques, such as factor analysis
and discriminant function. Thus, to have all indices at the
same time, Eq. [9] is proposed as an efficient screening
tool for identification of salinity-tolerant genotypes.
CONCLUSIONS
This research is a part of a comprehensive breeding program for salinity tolerance of bread wheat in Iran. Based
on the findings of this study, it may be concluded that the
application of all tolerant/sensitive indices simultaneously
is a good approach for screening salinity-tolerant genotypes. In this study, an equation (Eq. [9]) was developed
to estimate STS using all salinity tolerant/sensitive indices
concomitantly. Equation [9] gave the same results as statistical multivariate analyses. Since Eq. [9] was much easier
to utilize than complicated multivariate analyses, such as
factor analysis and linear discriminant function, it is suggested as an efficient screening tool for identification of
salinity-tolerant genotypes. We identified 10 lines as the
most salinity-tolerant lines with the highest grain yield in
both years. These lines can be recommended as promising
lines for salt-affected areas of Iran. These lines can be utilized through appropriate selection and as donor parents
in wheat breeding programs for further improvement of
wheat germplasm for salinity tolerance.
Supplemental Information Available
Supplemental information is included with this manuscript.
Supplementary Table 1. Mean of grain yield in normal
(Yp) and salinity stress (YS) conditions and tolerance/sensitive indices for the 100 bread wheat lines in 2008–2009
(averaged over two replications).
Supplementary Table 2. Discriminant function analysis on selected wheat genotypes.
Acknowledgments
This work was supported through a joint project between Agricultural Biotechnology Research Institute of Iran (ABRII),
Seed and Plant Improvement Institute (SPII), and ShahidBahonar University of Kerman-Iran. Our appreciations also go
to Mr. Ashkboos Amini and Mr. Mohammad-Taghi Tabatabaei
from Cereal Research Department of SPII for their technical
assistance. The authors would also like to acknowledge contributions made by Iranian Center of Excellence for Abiotic Stress
in Cereal Crops located in Shahid-Bahonar University of Kerman. The authors also wish to thank the Plant Stress Center of
Excellence (Psce) at the University of Isfahan.
www.crops.org1495
References
Abdolshahi, R., A. Safarian, M. Nazari, S. Pourseyedi, and G.
Mohamadi-Nejad. 2013. Screening drought-tolerant genotypes
in bread wheat (Triticum aestivum L.) using different multivariate
methods. Arch. Agron. Soil Sci. 59:685–704. doi:10.1080/03650
340.2012.667080
Ashraf, M. 2009. Biotechnological approach of improving plant salt
tolerance using antioxidants as markers. Biotechnol. Adv. 27:84–
93. doi:10.1016/j.biotechadv.2008.09.003
Austin, R.B. 1987. The climatic vulnerability of wheat. In: D.
Burns, editor, Proceedings of International Symposium on
Climatic Variability and Food Security. Indian National Science
Academy, New Delhi, India. 5–9 Feb. 1987. International Rice
Research Institute, New Delhi, India.
Baheri, S.F., A. Javanshir, H.A. Kazemi, and S. Aharizad. 2003.
Evaluation of different drought tolerance indices in some spring
barley genotypes. J. Agric. Sci. 13:95–100.
Bansal, K.C., and S.K. Sinha. 1991. Assessment of drought resistance
in 20 accessions of Triticum aestivum and related species. Part I:
Total dry matter and grain yield stability. Euphytica 56:7–14.
Brown, J.D. 2009. Choosing the right type of rotation in PCA and
EFA. Shiken: JALT Test. Eval. SIG Newsl. 13:20–25.
Bruckner, P.L., and R.C. Frohberg. 1987. Stress tolerance and
adaptation in spring wheat. Crop Sci. 27:31–36. doi:10.2135/cro
psci1987.0011183X002700010008x
Clarke, J.M., R.M. Depauw, and T.M. Townley-Smith. 1992.
Evaluation of methods for quantification of drought tolerance in
wheat. Crop Sci. 32:728–732. doi:10.2135/cropsci1992.0011183
X003200030030x
Clarke, J.M., T.M. Towenley-Smith, T.N. Mccaig, and D.G. Green.
1984. Growth analysis of spring wheat cultivars of varying
drought resistance. Crop Sci. 24:537–541.
Clarke, R.B., and R.R. Duncan. 1993. Selection of plants to
tolerate soil salinity, acidity and mineral deficiencies. Int. Crop
Sci. 1:371– 379. doi:10.2135/1993.internationalcropscience.c57
FAO. 2010. FAO land and plant nutrition management service.
www.fao.org. (accessed 10 Dec. 2010).
Farshadfar, E., M. Zamani, M. Motallebi, and A. Imamjomeh. 2001.
Selection for drought resistance in chickpea lines. (In Persian.)
Iran. J. Agric. Sci. 32:65–77.
Fernandez, G.C.J. 1992. Effective selection criteria for assessing
plant stress tolerance. In: C.G. Kuo, editor, Proceedings of an
International Symposium on Adaptation of Vegetables and
Other Food Crops to Temperature Water Stress, Taiwan. 13–16
Aug. 1992.Asian Vegetable Research and Development Center,
Tainan Taiwan. p. 257–270.
Fischer, R.A., and R. Maurer. 1978. Drought resistance in spring
wheat cultivars. Part 1: Grain yield response. Aust. J. Agric. Res.
29:897–912. doi:10.1071/AR9780897
Flowers, T.J., and A.R. Yeo. 1995. Breeding for salinity resistance
in crop plants: Where next? Aust. J. Plant Physiol. 22:875–884.
doi:10.1071/PP9950875
Golabadi, M., A. Arzani, and S.A.M. Maibody. 2006. Assessment
of drought tolerance in segregating populations in durum wheat.
African J. Agric. Res. 5:162–171.
Hossain, A.B.S., A.G. Sears, T.S. Cox, and G.M. Paulsen. 1990.
Desiccation tolerance and its relationship to assimilate partitioning
in winter wheat. Crop Sci. 30:622–627. doi:10.2135/cropsci1990
.0011183X003000030030x
Huang, B. 2000. Role of root morphological and physiological
characteristics in drought resistance of plants. In: R.E. Wilkinson,
editor, Plant– environment interactions. Marcel Dekker Inc.,
New York. p. 39–64.
1496
IRRI. 1996. Standard evaluation system for rice. 4th ed. International
Rice Research Institute, Manila. p. 52.
Kaya, Y., C. Palta, and S. Taner. 2002. Additive main effects and
multiplicative interactions analysis of yield performances in
bread wheat genotypes across environments. Turk. J. Agric. For.
26:275–279.
Kristin, A.S., R.R. Senra, F.I. Perez, B.C. Enriquez, J.A.A. Gallegos,
P.R. Vallego, et al. 1997. Improving common bean performance
under drought stress. Crop Sci. 37:43–50. doi:10.2135/cropsci19
97.0011183X003700010007x
Lantican, M.A., P.L. Pingali, and S. Rajaram. 2003. Is research
on marginal lands catching up? The case of unfavorable
wheat growing environments. Agric. Econ. 29:353–361.
doi:10.1111/j.1574-0862.2003.tb00171.x
Mccaig, T.N., and J.M. Clarke. 1982. Seasonal changes in
nonstructural carbohydrate levels of wheat and oats grown in
semiarid environment. Crop Sci. 22:963–970. doi:10.2135/crops
ci1982.0011183X002200050016x
Mohammadi, M., R. Karimzade, and M. Abdipour. 2011.
Evaluation of drought tolerance in bread wheat genotypes under
dryland and supplemental irrigation conditions. Aust. J. Crop
Sci. 5:487–493.
Rizza, F., F.W. Badeckb, L. Cattivellia, O. Lidestric, N. Di Fonzoc,
and A.M. Stancaa. 2004. Use of a water stress index to identify
barley genotypes adapted to rainfed and irrigated conditions.
Crop Sci. 44:2127–2137. doi:10.2135/cropsci2004.2127
Rosielle, A.A., and J. Hamblin. 1981. Theoretical aspects of selection
for yield in stress and non-stress environments. Crop Sci. 21:43–
46. doi:10.2135/cropsci1981.0011183X002100010013x
SAS Institute. 2004. Base SAS 9.1 procedures guide. SAS Institute
Inc., Cary, NC. p. 36.
Sio-Se Mardeh, A., A. Ahmadi, K. Poustini, and V. Mohammadi.
2006. Evaluation of drought resistance indices under various
environmental conditions. Field Crops Res. 98:222–229.
doi:10.1016/j.fcr.2006.02.001
Smith, H.F. 1936. A discriminant function for plant selection. Ann.
Eugen. 7:240–250. doi:10.1111/j.1469-1809.1936.tb02143.x
Talebi, R., F. Fayaz, and A.M. Naji. 2009. Effective selection criteria
for assessing drought stress tolerance in durum wheat (Triticum
durum Desf.). Gen. Appl. Plant Physiol. 35:64–74.
Trethowan, R.M., M.V. Ginkel, and S. Rajaram. 2002. Progress
in breeding for yield and adaptation in global drought
affected environments. Crop Sci. 42:1441–1446. doi:10.2135/
cropsci2002.1441
Van Ginkel, M., D.S. Calhoun, G. Gebeyehu, A. Miranda, C. Tianyou, R. Pargas Lara, et al. 1998. Plant traits related to yield of
wheat in early, late or continuous drought conditions. Euphytica
100:109–121. doi:10.1023/A:1018364208370
Winter, S.R., J.T. Musick, and K.B. Porter. 1988. Evaluation of
screening techniques for breeding drought-resistance winter
wheat. Crop Sci. 28:512–516. doi:10.2135/cropsci1988.0011183
X002800030018x
Yaremko, R.M., and H. Harari, R.C. Harrison, and E. Lynn. 1986.
Handbook of research and quantitative methods in psychology:
For students and professionals. Lawrence Erlbaum Associates,
Hillsdale, NJ.
Yokoi, S., R.B. Bressan, and P.M. Hasegawa. 2002. Salt stress
tolerance of plants In: M. Iwanaga, editor, Genetic engineering
of crop plants for abiotic stress. JIRCAS Working Report No.
23. MAFF, Tsukuba, Ibaraki, Japan. p. 25–33.
Zeng, L., M.C. Shannon, and C.M. Grieve. 2002. Evaluation of salt
tolerance in rice genotypes by multiple agronomic parameters.
Euphytica 127:235–245. doi:10.1023/A:1020262932277
www.crops.org
crop science, vol. 54, july– august 2014