In this table, environment that defined as combination of year × condition and the genotypes showed high significant difference at 0.01 probability level for all studied traits except agronomic score (Table 1). Thus, indirect selection for a drought prone environment based on the results of optimum conditions will not be efficient. These results are in agreement with those of
Kanouni et al., (2012) and
Ghiabi et al., (2013). Genotype × environment (GE) interaction showed non-significant difference, also this GE interaction cannot use for determining genotypic stability. The estimates of stress tolerance attributes indicated that the identification of drought-tolerant genotypes based on a single criterion was contradictory. For example, according to STI and GMP and MP genotypes G1, G2, G3 and G4 were found drought tolerance with highest STI and grain yield under irrigation (non-stressed) condition, while genotypes G5, G6 and G7 were displayed the lowest amount of for these indices under irrigation condition (Table 2). Similar results were reported by
Mohammadi et al., (2012) in bread wheat and
Naghavi et al., (2013) in maize data.
Pireivatlou et al., (2010) was also noted that STI can be a reliable index for selecting high yielding genotypes. The greater the TOL value, the larger the yield production under supplementary irrigation conditions and the smaller the TOL value, the larger the yield production under rain-fed conditions. For TOL and SSI indices, the desirable drought tolerant genotypes for supplementary irrigation were G2 and G3. Suggest that selection based on TOL will result in reduced yield under well-watered conditions. Similar results were reported by
Farshadfar and Elyasi (2012) and
Talebi et al., (2009). YSI index selected the genotypes G4, G3 and G1 as the most relatively tolerant genotypes while for RDI the genotypes G4, G8 and G3 were the most relative tolerant. The smaller the SSPI value, the larger the yield production under rain-fed conditions, base on SSPI index, genotypes G8, G6 and G5 were relatively tolerant genotypes. According to K1STI and K2STI the genotypes G1, G2, G3 and G4 were the most relative tolerant genotypes (Table 2). DI selected the genotypes G4, G3, G1 and G2 as the best, while the genotypes G5, G6, G7 and G8 as the worst relatively tolerant genotypes.
Majidi et al., (2011) reported that GMP, STI and HM indices were similarly able to separate drought sensitive and tolerant genotypes of safflower in both mild and intense water stress environments.
Talebi et al., (2009) also reported that cultivars producing high yield in both drought and well watered conditions can be identified by STI, GMP and MP values. Selection based on a combination of indices may provide a more useful criterion for improving drought tolerance of chickpea but study of correlation coefficients are useful in finding the degree of overall linear association between any two attributes. Thus, a better approach than a correlation analysis such as biplot is needed to identify the superior genotypes for both rain-fed and supplementary irrigation environments. Principal component analysis (PCA) revealed that the first PCA explained 81.25% of the variation, thus, the first dimension can be named as the yield potential and drought tolerance. Considering the high and positive value of this biplot, genotypes that have high values of these indices will be high yielding under rain-fed and supplementary irrigation environments. The second PCA explained 18.68% of the total variability and correlated positively with TOL and SSI. Therefore, the second component can be named as a stress-tolerant dimension and it separates the stress-tolerant genotypes from supplementary irrigation tolerant ones. Thus, selection of genotypes that have high PC1 and low PC2 are suitable for both rain-fed and supplementary irrigation environments (Fig 3). Therefore, genotype G4 was superior genotypes for both environments with high PC1 and low PC2. Genotypes G1, G2 and G3 with high PC2 were more suitable for supplementary irrigation environment than for rain-fed environment.
Farshadfar and Elyasi (2012),
Sabaghnia et al., (2014), Hamayoon et al., (2011) in chickpea and
Karimizadeh and Mohammadi (2011),
Karimizadeh et al., (2012), Sio-Se
Mardeh et al., (2006) and
Golabadi et al., (2006) obtained similar results in multivariate analysis of drought tolerance in different crops.
For these eight genotypes, the crop cycle was reduced from 131 days (emergence to physiological maturity) in TOL group, to an average of 136 days in the SEN group in rain-fed condition. A comparison of performance on a per day basis between TOL and SEN groups showed that, under drought stress, total yield and yield per day at maturity had 37.7% and 32.6% differences, respectively. The crop cycle was reduced from 139 days (emergence to physiological maturity) in TOL group, to an average of 141 days in the SEN group in supplemental irrigation condition. A comparison of performance on a per day basis between TOL and SEN groups showed that, total yield and yield per day at maturity had 28.4% and 29.7% differences respectively. A comparison of other trait performance between TOL and SEN groups showed that, pods per plant, empty pods number and filled pods number had 25.4%, 22.7% and 48.9% differences respectively in rain-fed condition but these traits had 12.5%, 27.8% and 17.3% differences respectively in supplementary irrigation condition (Table 3).
Differences in yield and morpho-physiological traits for the TOL versus SEN genotypes was apparently associated with differing performance during grain filling, as reflected by the difference in grain filling rate, rather than parameters up until anthesis (Table 3).
The physiological data indicate similar contrasts between TOL and SEN genotypes for photosynthetic chlorophyll content and canopy temperature again with differences most pronounced from flowering onwards (Table 4). Result showed that chlorophyll content values in 50% flowering stage were highest value ratio grain filling stage in both conditions. The reducing of chlorophyll content in supplemental irrigation condition was clearer than rain-fed condition (Table 4). Differences between chlorophyll content values in 50% flowering stage in supplemental irrigation were smaller than rain-fed condition. One of the most important factors can be impact of irrigation water on creating of mild environment and therefore difference between chlorophyll content of leaves in TOL and SEN groups were decreases. This result is consistent with the reports of Talebi
et al, (2013) and
Karimizadeh et al., (2011).
The CTD measurements were taken at the stages of 50% flowering and watery ripe, clear liquid (grain filling stage). Genotypic differences were detected at the first stage for both supplementary irrigation and rain-fed condition on chickpea genotypes. CTD values changed between 4.6°C (SEN genotypes) and 6.2°C (TOL genotypes) and between 7.2°C (SEN genotypes) and 6.9°C (TOL genotypes) in rain-fed and supplemental irrigation conditions respectively. At grain filling stage, CTD changed between 3.1°C (SEN genotypes) and 3.9°C (TOL genotypes) and between 6.3°C (SEN genotypes) and 5.9°C (TOL genotypes) in rain-fed and supplemental irrigation conditions respectively (Fig 4 and 5).
High temperature after flower opening decreases chickpea seed yield by reducing the number of seeds per plant and weight per seed
(Wang et al., 2006). In chickpea,
Summerfield et al., (1984) suggested that the longer the exposure during reproductive development to a high day temperature of 35°C, the lower the yield. Most chickpea genotypes do not set pods when temperatures reach >35°C
(Basu et al., 2009). However, there is considerable variation among genotypes for response to high temperature. The period of anthesis and seed set are clearly critical stages for exposure to heat stress (
Gross and Kigel 1994).
Nayyar et al., (2005) suggested that the development of male (pollen, anthers) and female (stigma-style, ovary) parts are the most sensitive organs to abiotic stress in reproductive biology. Therefore, pollen viability, stigma receptivity and ovule viability are useful indicators of sensitivity to abiotic stress
(Nayyar et al., 2005). However, the effect of stress on either male or female organs depends upon the stage of sporogenesis (micro or mega). Due to heat stress, meiosis and pollen development are the most affected part in micro-sporogenesis. Megaspore formation in the ovule and fertilisation are the most important events in mega- sporogenesis under high temperature stress (Gross and Kigel,1994).
Devasirvatham et al., (2010) believes that although classification of heat responses of chickpea has been documented
(Upadhyaya et al., 2011), there has been little attempt to extrapolate these findings across the world’s chickpea production areas. The determination of a heat response phenotype through screening is vital if the genetic control of heat tolerance in chickpea is to be understood and significant progress made through plant breeding. Clearly, the research under high temperature stress shows that early phenology is the most important mechanism and pod set the primary yield component to be considered in heat tolerance breeding.
In drought sensitive chickpea parent HC1 relative stress injury (RSI) was recorded 16.4% under irrigated whereas 31.3% under drought condition. Whereas, in drought tolerant ICC 4958, RSI was recorded 16.3% under irrigated and 20.9% under drought conditions
(Yadav et al., 2015). Also in tolerant parent ICC 4958, there is increase in CTD from -1.1 under irrigated condition to 0.9 under drought conditions is less as compared to HC 1 from -0.4 in irrigated condition to 2.6 in drought conditions. In best yielding progeny lines, CTD ranges from -1.97 to 0.50 under drought condition
(Yadav et al., 2015). Rees et al., (1993) reported that CTD values have been changed between 3.54 and 5.10°C before anthesis, 3.16 to 4.61°C after anthesis in bread wheat.
Reynolds et al., (1997) reported that CTD average values of heat stress tolerant genotypes in bread wheat were respectively 7.4, 9.0 and 6.5°C before heading, at heading and grain filling periods. These values were respectively 7.1, 7.9 and 5.7°C at the same periods in susceptible genotypes. In this study, it has been shown similar to this situation; for instance, CTD values have been observed such as 6.9, 9.1 and 5.3°C in G6 before heading, at heading and grain filling periods respectively in rain-fed condition. It has understood that this genotype have cooler plant canopy than the other cultivars. Also,
Barma et al., (1997) showed that CTD values could have been changed -2.4 and -5.5°C sometimes. The rankings of the indices based on the power of representation are shown in Fig 3. The results showed that the Ys Index showed the best representation power and the DI, STI and GMP indices had the lowest angle with the Ys Index and showed the next best rankings to evaluate diversity and power of representation (Fig 6).
In GGE biplot methodology, the yield and stability of the genotypes are examined by an average tester coordinate (ATC). The mean yield of the genotypes is estimated by their projections on the ATC × axis. The average location, as the virtual location, is shown by a circle and indicates the positive end of the ATC × axis. According to the ATC figure, the length of the average location vector was adequate to select genotypes based on mean yield. In the GGE model, G4 was the least stable genotype which had variable performance across test locations, while G6, G3 and G1 were the most stable genotypes. The performance of genotype G6 close to ATC axe was stable, whereas this genotype showed low mean yields (Fig 7).
In different crops, as well as in chickpea, differential genotypic response to drought stress, as a result of variation in physiological parameters has been reported
(Gunes et al., 2006; Gunes et al., 2008). In this study, we tried to explain the responses of the genotypes and discussed some physiological parameters that were affected by drought stress. These parameters were also evaluated as drought tolerance selection criteria. Drought stress can also alter the tissue concentrations of chlorophylls and carotenoids
(Jaleel et al., 2008; Kalefetoglu and Ekmekci, 2009). While increased chlorophyll and carotenoid content under drought stress may be related to a decrease in leaf area, it also can be a defensive response to reduce the harmful effects of drought stress
(Farooq et al., 2009). The photochemical efficiency was recorded in the range of 0.654 to 0.770 for highest yielding 20 progeny lines. Increase in CTD might have occurred due to decreased transpiration resulting from stomatal closure. The photosynthetic efficiency, transpiration and the values of relative stress injury declined in chickpea under drought conditions
(Kumar et al., 2012). The total chlorophyll content significantly decreased in all genotypes under drought stress, but the reductions were not as great in tolerant genotypes. Higher level of carotenoid concentration in drought-tolerant genotypes has also been reported
(Deng et al., 2003; Kalefetoglu and Ekmekci, 2009). Chlorophyll content was greater at 10 and 20 t. ha of biochar compared with control at all measurement dates in the winter sowing. This response could partly be attributed to the effect of biochar on plant nutrient status
(Lusiba et al., 2016). Drought-tolerant genotypes accumulated more carotenoids than susceptible genotypes. Accumulation of carotenoids for osmotic regulation in drought-stressed leaves in many crops has been reported
(Khan et al., 2001; Gunes
at al., 2008).