Maize (Zea mays L.) is one of the important cereal crops of Iran and
the world after wheat and rice. Recent projections by the International Food
Policy Research Institute indicate that by 2020 the demand for maize in all
developing countries will overtake for wheat and rice (Gerpacio
and Pingali, 2007). In Iran, the average grain yield ha-1 in
2007 was 7.6 tons, whereas the soil and climatic condition of Iran are suitable
for maize production but the yield is low compared to the United States of America
with 9.5 t ha-1 in 2007. Thus, it is prerequisite to select promising
hybrids for different conditions in order to speed up economical crop production.
Maize breeders have successfully exploited heterosis for grain yield by crossing inbred lines to develop desirable hybrids. However, the nature of gene action involved in expression of heterosis for the grain yield of elite maize hybrids remains unresolved.
General Combining Ability (GCA) and Specific Combining Ability (SCA) are the
most important indicators for expressing the potential value of lines. The non
additive gene effect is distinguished by specific combining ability but additive
gene effect is distinguished by general combining ability (Nevado
and Cross, 1990; Choukan, 2008).
The choice of efficient breeding program depends on a large knowledge of type
gene action involved in expression of the character. Dominance gene action would
favor the production of hybrids, whereas additive gene action indicates that
standard selection procedures would be effective in breeding about changing
in character (Edwards et al., 1976).
Earlier studies have shown that both additive and non additive gene effects
were important for controlling grain yield (Malvar et
al., 1996; Iqbal et al., 2007). However,
Abdel-Moneam et al. (2008) have shown that grain
yield is governed by genes acting non additively.
Drought is one of the most important abiotic stress factors, which affects almost every aspect of plant growth.
Different type of gene action under drought and low nitrogen conditions were
reported by Betran et al. (2003). They concluded
that additive effects were more important under drought condition and dominance
effects were more important under low nitrogen condition.
Several indices have been utilized to evaluate genotypes for drought tolerance
based on grain yield such as mean production (Rosielle and
Hamblin, 1981), stress susceptibility index (Fischer
and Maurer, 1978), stress tolerance index (Fernandez,1993)
and tolerance (Rosielle and Hamblin, 1981). These indices
have been compared by some researchers (Fernandez,1993;
Shiri, 2005; Sanjari-Pirevatlou and
Yazdansepas, 2008 ), but there were a few studies on genetic properties
of these indices.
Based on the results obtained in previous studies, it seemed that STI and MP
were useful yield-based drought tolerance indices to select high yielding genotypes
in both non water stress and water stress conditions. On the other hand, SSI
and TOL were not useful indices to select for drought tolerant genotypes (Fernandez,1993;
Shiri, 2005; Sanjari-Pirevatlou and
Yazdansepas, 2008 ).
The objectives of this study were; (1) to estimate the general combining ability of lines and testers and specific combining ability of crosses for grain yield under water stress and non water stress conditions, (2) to estimate the gene action governing under non water stress and its changes under water stress condition, (3) to calculate the genetic parameters and especially the narrow sense heritability of the important yield-based drought tolerance indices by using linextester analysis and (4) to select the efficient drought tolerance indices.
MATERIALS AND METHODS
The seeds of twenty maize inbred lines were obtained from Seed and Plant Improvement Institute of Iran. In line x tester fashion, eighteen female inbreds and two male testers (K3653/2 and K3615/1) were crossed through controlled pollination to produce thirty six
hybrid progenies in field of Agricultural and Natural Resources Research Center of Ardebil Province (Moghan) in 2007. The parents were:
Thirty six generated maize hybrids were planted in two experiments with non water stress and water stress at grain filling stage in Pars Abad-e-Moghan (39° 41' N 47° 32' E, with 281.3 mm annual precipitation), Ardebil, Iran in 2008, using a RBCD design with three replications. The plot was made of four rows of 5 m length with the distance between rows and hills of 75 and 18 cm, respectively. Sowing was performed by three seeds per hill and thinning eighteen days after planting reduced the stand at one plant per hill. Thus, a planting density of 75000 plant ha-1 was achieved.
In non water stress condition, the irrigation was performed nine times based
on crop water requirements during growth periods, but in water stress at grain
filling stage condition, the irrigation was done five times from planting time
till the end of flowering period and then, in order to apply water stress, irrigation
was withheld completely from the end of flowering till crop maturity (grain
filling stage). The environmental severity degree is estimated with SI (stress
intensity) and maximal rate of SI is one (Fischer and Maurer,
1978). In this study, SI was 0.31, so stress intensity was moderate. Grain
yield were determined under both non water stress and water stress experiments
and used as Yp and Ys, respectively.
For every genotype, the four drought tolerance indices were calculated based on their grain yield in non water stress and water stress conditions.
The drought tolerance indices were calculated as follows:
||Yield of a genotype in non water stress condition
||Yield of a genotype in water stress condition
||Mean yield in non water stress condition
||Mean yield in water stress condition
The recorded data were subjected to analysis of variance according to Steel
and Torrie (1980) to determine significant differences among crosses. The
significant differences among crosses were further partitioned by using linextester
analysis (Kempthorne, 1957). The estimation of General Combining
Ability (GCA) for lines, testers, Specific Combining Ability (SCA) for crosses,
dominance and additive variance were also estimated following function of Kempthorne
(1957). For testing of general combining ability for lines, testers and
specific combining ability for crosses were used T test method (Steel
and Torrie, 1980). The data were statistically analyzed by MSTATC, STATISTICA
and EXCEL computer programs.
RESULTS AND DISCUSSION
The analysis of variance (Table 1) showed that mean square
due to crosses for grain yield was significant in both water stress and non
water stress conditions, indicate the existence of variability among the cross
combinations for grain yield. The analysis of variance according to linextester
method revealed significant difference among lines, testers and linextester
interaction for grain yield in both conditions and for all of indices. This
indicated that both additive and non additive (dominance) gene effects were
important in genetic expression of all of indices and grain yield in both water
stress and non water stress conditions. The GCA/SCA ratio was less than unity
for all of indices and grain yield in both conditions; this means that these
characters were governed predominantly by non additive component. Also narrow
sense heritability estimates were generally lower than broad sense heritability,
indicating the presence of non additive gene action. These components can be
exploited by hetreotic breeding programme. The similar results have been reported
in maize under normal and high plant densities by Choukan
(1999). It has been frequently reported that grain yield is controlled by
both additive and non additive gene action in normal condition (Malvar
et al., 1996; Iqbal et al., 2007).
Grain yield recorded high genetic variance value under non water stress condition compared to those under water stress condition. Also narrow and broad sense heritability estimates in non water stress condition were higher than water stress condition (Table 1).
The low heritability in non water stress condition was related to remarkable
decrease in genetic variance than environmental variance. Ngaboyisonga
et al. (2009) concluded that the reduction in genetic variance was
the effect of drought on genetic variation of grain yield. Also, Hefny
(2007) stated that heritability and genetic variance component estimates
were high at optimal N fertilizer (normal environment) compared with low N fertilizer
||Mean square and variance components for grain yield and drought
tolerance indices under non water stress and water stress conditions according
to linextester analysis
|**:Significant at 1% level of probability..YP:
Grain yield under non water stress condition, YS: Grain yield under water
stress condition, TOL: Tolerance, SSI: Stress susceptibility index, MP:
Mean productivity, STI: Stress tolerance index, σ2A:
Additive variance, σ2D: Dominance variance, σ2gca/σ2sca:
The ratio of general combining ability variance and specific combining ability
variance, σ2g: Genotypic variance, σ2P:
Phenotypic variance, h2B: Broad sense heritability,
h2N: Narrow sense heritability
Analysis of General Combining Ability (GCA) indicated that the variation of combining ability of lines in non water stress condition was higher than water stress condition. The reactions of GCA of lines in both conditions were not similar. For example, the line 8 (L8) had significantly positive GCA effect in non water stress condition, whereas, this line had negative GCA effect in water stress condition. Conversely, in line 16 (L16), it had significantly negative and positive GCA effect in non water stress and water stress conditions, respectively. The Line 17 (L17) had significantly positive GCA effect in both non water stress and water stress conditions. Overall, the lines L8, L11 and L17 in non stress condition and the lines L15, L16 and L17 in stress condition showed better general combining abilities for grain yield (Table 2).
The Specific Combining Ability (SCA) effects of 36 crosses were shown in Table 3. The crosses L1xT1, L4xT1 and L8xT1 showed significant and positive SCA effects in non water stress condition, whereas, the cross L9xT2 had highest value of SCA effects in water stress condition. The lines GCAs, both in direction and in magnitude changed with the change of conditions (Table 3).
Variation of specific combining ability of crosses in non water stress condition was higher than water stress condition. It means that the number of crosses with positive and significant SCA effects were more in non water stress condition than water stress condition (Table 3).
The results of heritability of indices showed that the STI index with 0.894
had the highest rate of broad sense heritability among indices. In this experiment
SSI exhibited negligible narrow sense heritability and STI was more heritable
than MP and TOL, as determined by narrow sense heritability estimates (Table
1). Genetic advances are directly related to the magnitude of narrow sense
heritability (Choukan, 2008). Thus, it seems that selection
for drought tolerance based on STI will be useful than based on SSI and TOL.
These results were in agreement with those of Saba et
To determine the most desirable drought tolerance index, the correlation coefficient
among YP (grain yield in non water stress), YS (grain
yield in water stress) and other quantitative indices of drought tolerance were
calculated (Table 4). The most desirable drought tolerance
index is the one which has significant and the same sign correlation with both
YP and YS. The correlation coefficient of Stress Susceptibility
Index (SSI) with YS and YP were -0.42 and 0.74, respectively.
||General Combining Ability (GCA) of lines and testers for grain
yield under non water stress and water stress conditions
|ns: Non significant, *, **: Significant at 5 and 1% levels
of probability, respectively
||Specific Combining Ability (SCA) of crosses for grain yield
under non water stress and water stress conditions
|ns: Non Significant.*, ** Significant at 5 and 1% levels of
Therefore, selection for SSI should give a positive yield response under water
||The correlation coefficients among grain yield under non water
stress (YP), grain yield under water stress (YS) and drought tolerance indices
|ns: Non significant. **: Significant at 1% level of probability.
YP: Grain yield under non water stress condition, YS: Grain yield under
water stress condition, TOL: Tolerance, SSI: Stress susceptibility index,
MP: Mean productivity, STI: Stress tolerance index
The correlation coefficient of Stress Tolerance Index (STI) with YP
and YS were 0.87 and 0.71, respectively. Thus, selection for STI
should give positive responses in both conditions. On the other hand, MP and
STI were highly correlated with each other as well as with YS and
YP. Thus, through these indices, it is possible to distinguish high
yielding genotypes in either condition. The non significant correlation between
SSI and STI (r = 0.32) would indicate that the combination of high STI with
a low to moderate SSI is biologically attainable in maize.
The observed relationship between YP and (MP and STI) and YS
and (MP and STI) were in consistence with those reported by Fernandez,1993,
Shiri (2005) and Sanjari-Pirevatlou
and Yazdansepas, 2008 .
Furthermore, A good drought tolerance index should be able to identify superior genotypes both in non water stress and water stress conditions from the genotypes that are favorable only in one condition.
According to Fernandez,1993, genotypes can be categorized into four groups based on their performance in stress and non stress environments: genotypes express uniform superiority in both environments (Group A); genotypes perform favorably only in non stress environments (Group B); genotypes yield relatively higher only in stress environments (Group C); and genotypes perform poorly in both environments (Group D).
The grain yield ranged from 6.09 to 11.2 tons per hectare in non water stress condition and from 4.7 to 6.6 tons per hectare in water stress condition (Table 5).
Based on grain yield in water stress and non water stress conditions, the studied crosses were divided into four groups as following: L1xT1, L4xT1, L5xT1, L8xT1, L11xT1, L12xT1, L14xT1, L17xT1, L18xT1, L15xT2 and L17xT2 crosses placed in group A. L9xT1, L10xT1, L6xT2, L11xT2, L12xT2 and L18xT2 crosses included in group B. L2xT1, L7xT1, L13xT1, L15xT1, L16xT1, L1xT2, L2xT2, L9xT2, L10xT2 and L16xT2 crosses were in group C and finally L3xT1, L6xT1, L3xT2, L4xT2, L5xT2, L7xT2, L8xT2, L13xT2 and L14xT2 crosses located in group D (Table 5, Fig. 1). Fernandez,1993 stated that an optimal selection criterion should be able to distinguish group A from the other three groups.
Genotypes with high values of TOL and SSI are sensitive to water stress and
therefore, selection must be done based on low rates of these indices. Based
on TOL and SSI indices, the crosses L16xT1, L2xT2, L4xT2 and L16xT2 had the
highest yield stability among the studied crosses. These crosses had low grain
yield under both stress and non water stress conditions and located in group
C and D (Table 5). Therefore, the use of TOL and SSI indices
lead the selection toward tolerant and low yielding genotypes. It is better
to use these indices for the omission of susceptible genotypes, but not for
the selection of both stress tolerant and high yielding genotypes. Moghaddam
and Hadizadeh (2001) have got similar results on this subject.
||Estimation of drought tolerance indices based on grain yield
of maize crosses under water stress and non water conditions (SI = 0.31)
|YP: Grain yield under non water stress condition, YS: Grain
yield under water stress condition, TOL: Tolerance, SSI: Stress susceptibility
index, MP: Mean productivity, STI: Stress tolerance index . +: Mean with
similar letters in each column are not significantly different at 1% probability
level by Duncans Multiple Range Test (DMART). ++: The crosses grouping
based on grian yield of maize crosses under water stress and non water conditions
The crosses of L1xT1, L4xT1, L5xT1 and L8xT1 had the highest rate of MP index
with 8.1, 7.5, 7.5 and 8.6, respectively (Table 5). So, these
crosses were selected based on MP index. These crosses had high grian yield
under both stress and non water stress conditions and located in group A.
According to Fernandez,1993, more stable genotypes have
higher rate of STI. The crosses L1xT1, L4xT1, L8xT1 and L17xT1 were selected
based on STI index. All of these crosses were located in the group A (Table
4). So, STI and MP indices could separate the group A crosses from other
group crosses. Fernandez,1993 compared effectiveness of
several stress tolerance criteria and concluded that MP, SSI and TOL failed
to identify genotypes with both high yield and stress tolerance potentials,
whereas through STI, genotypes with these attributes could be identified. Also,
Shiri (2005) and Sanjari-Pirevatlou
and Yazdansepas, 2008 noted the similar results.
To show the advantage of STI and separate genotypes into four groups based
on yield under stress and non water stress conditions, three-D plot among YS
(yield under water stress condition) ,YP (yield under non water stress
condition) and STI was used. As, it was shown in Fig. 1, most
of the group A crosses had high STI values.
||The 3-D plots among grain yield under non water stress (YP),
grain yield under water stress (YS) and Stress Tolerance Index (STI)
Thus, it seems that selection for drought stress tolerance based on STI index
will be more fruitful than based on TOL and SSI indices.
The results showed that the effect of gene action can be both non additive and additive in the expression of grain yield under both conditions. However, non additive genetic variance was more important in the expression of grain yield than additive variance. Genetic variance, narrow sense heritability and broad sense heritability estimates in non water stress condition were higher than water stress condition. Based on broad sense heritability, narrow sense heritability and correlation of indices, STI is a successful index to select high yield and tolerant genotypes than SSI and TOL indices. According to yield in both conditions, specific combining abilities and STI index, crosses L1xT1, L4xT1 and L8xT1 were the best crosses in this study.
Thanks to the Prof. Ramiz Tagi Aliyev and Dr. Rajab Choukan for his kind supports, help and suggestion during this study. This study was extracted from Ph.D Thesis of Mohammadreza Shiri.