Evaluation of Drought Tolerance Selection Indices Using Grain Yield in Chickpea ( Cicer arietinum L . )

Among abiotic stresses, drought is undoubtedly one of the most important ones, that have great impact on crop growth and productivity worldwide. Therefore, identifying of plants' performance against drought stress and estimating drought tolerance become a necessary part of the breeding phase. The main purpose of the present study was to investigate the effect of several indices that combine drought tolerance and high yield potential in chickpea. The trials were conducted under both stressed and no-stressed environments for two growing seasons (2015/2016-2016/2017) in Southeast Anatolia Region of Turkey. Varyans analysis results showed that there were significant differences among genotypes regarding Yp, Ys, MP, MRP, GMP, REI, STI, MISTIk 1, MISTIk 2, HM, YI, PI, ATI, SNPI and RDY. The genotypes FLIP09-51C, FLIP97-503C and FLIP06-97C had high yield under non-stressed condition, while the genotypes FLIP09-51C, FLIP06-97C and ‘Aksu’ displayed high amount under stressed condition. Thus, the genotypes FLIP09-51C and FLIP06-97C were found as good candidates for commercial recommendation in both conditions. Spearman rank correlation matrix showed that drought indices were significantly related to each other. The yields in stress and no-stress conditions (Yp and Ys) showed a significant and positive correlation with MP, MRP, GMP, REI, STI, MSTIK1, MSTIK2 and HM and showed a negative correlation with PI and RDY. As a result, it has been found that MISTIK2, DI, HM, STI and YI can be used as optimal indicators for screening drought-tolerant genotypes, while FLIP09-51C, FLIP06-97C, EN934 and ‘Aksu’ varieties have been the most tolerant genotypes in terms of these indices examined in study.


Introduction
Chickpea (Cicer arietinum L.) is one of the most important edible legumes grown in marginal lands around the world.Its production is about 12.1 million tons from 12.7 million hectares cultivated land, with a mean seed yield of 95 kg ha -1 .Chickpea is produced on 351.687 hectares in Turkey and ranks fifth in the world after India, Australia, Myanmar and Pakistan (FAO, 2016).Among food legumes, chickpea is rich in nutritional compositions and does not contain significant quantities of any specific major anti-nutritional factors.Due to its high protein content, chickpea is used as a major protein source in most of developing countries.It is also an important pulses plant in cereal-legume planting rotation in many countries around the world (Zdemur et al., 2003).Biotic and abiotic stresses are important factors limiting the agricultural production in the world.From these abiotic stresses, drought causes severe yield losses in arid and semi-arid regions (Mollasadeghi et al., 2011).
Due to Ascochyta blight disease in chickpea, most of the chickpea production is done in marginal areas and as spring planting in many countries of the world.This situation causes terminal drought stresses during flowering, podding and seed filling period of plants in Turkey.Therefore, last week of June for stressed fields and in the first week of July for non-stressed experimental site.Di-Ammonium Phosphate  fertilizer was used at the rate of 130 kg per hectare as source of N and P before sowing.Weeds were controlled by hand during crop growth and development.

Calculation of indices
The seed yield of each plot was evaluated based on 6 m², and converted to the standard unit at metric kg per hectare.Drought indices were calculated using the consacrated formulas (Annex 1).

Statistical analysis
The correlation coefficients were calculated using the SPSS version 22.The experimental data were statistically analyzed for variance using JMP statistical software (SAS 2007).Differences were compared by Least Significant Difference Test (LSD) at alpha 0.05.

Analysis of variance
Analysis of variance for both potential yield (Yp) and stress yield (Ys) indicated significant differences among the fourteen chickpea genotypes.Variance analysis of both the potential yield (YP) and the stress yield (YS) indicated significant differences among the genotypes.Also, significant differences were observed for the examined indices MP, MRP, GMP, REI, STI, MISTIk1, MISTIk2, HM, YI, PI, ATI, SNPI and RDY (Table 2).
The ranks of the chickpea according to each one of the drought tolerance indices are given in Table 3.The average seed yield of the varieties under non-stressed conditions ranged from 3,954.4 ('Diyar-95') to 4,825.0 kg (FLIP09-51C), while ranged from 2,384.4 ('Arda') to 3,044 kg in the stressed conditions.Besides, the average grain yield in nonstressed condition (4,370.1 kg) was higher than in stressed condition (2,736.3kg).When the yields of the varieties were compared in two environments, the largest variation of yield was observed in 'Arda' variety, followed by FLIP97-503C and 'Azkan' respectively.Maximum reductions in grain yield were observed in 'Arda' and FLIP97-503C genotypes, while the least amount of reductions in yields was observed in FLIP03-131C and FLIP06-93C genotypes (Table 3).Fernandez (1992) classified plants according to their performance in stressful and stressless conditions to four groups: genotypes with high performance in both conditions (group A), genotypes with good performance selection for drought tolerant chickpea genotypes is the most important objective for breeding to drought stress.Drought tolerance selection is not easy due to the happening of strong interactions between genotypes and the environment and restricted knowledge about the function and role of tolerance mechanisms.Hence, researchers have used different drought resistance indices to evaluate genetic differences in drought tolerance.
Thus, the present study was conducted to (i) determine drought tolerance indices under stress and non-stress conditions, and (ii) identify some of chickpea genotypes reaction to drought stress.

Growth conditions and research materials
The research was conducted at two locations with different average rainfall amounts, namely Diyarbakır and Kızıltepe, during 2016 and 2017, both of which are located in the Southeast Anatolia Region of Turkey.The favorable site (Diyarbakır) is located at 37° 56' N latitude, 40° 15' E longitude, at an elevation of 670 m above sea level.The soil of the trial area carried out in this locality is a clay loam.The average annual maximum and minimum temperatures are 22.6 °C and 8.8 °C respectively, and average annual precipitation is 485 mm (Anonymous, 2017).Kızıltepe station represents the stressed semi-arid site with 300 mm as mean rainfall.It is located at 37° 19´ N latitude, 40° 58´ E longitude, at an elevation of 490 m.The soil is deep clay.The yearly average maximum and minimum temperatures are 26.4 °C and 12.1 °C respectively.
The field experiments were arranged in a randomized complete block design with three replications at both environments.The plots consisted of 4 rows that were 5 m long and spaced 0.30 m.Seeds were planted with parcel machine at density of 45 seeds per square meter.The sowings were performed on the last week of November in both environments and harvestings were carried out on the 440  According to this classification, the FLIP09-51C, FLIP97-503C and FLIP06-97C, FLIP03-104C, EN934 and 'Aksu' genotypes had higher yield performances than the mean yields of the genotypes in both stressed and nonstressed conditions, and were located on the group A. Whereas 'Diyar 95', FLIP03-112C, FLIP06-39C, 'Azkan' and FLIP06-133C genotypes had lower yields than the average yields of the genotypes for both environments and were located on the group C. On the other hand, FLIP03-131C and FLIP06-93C genotypes had higher yield performance than average yield for only stressed environments and were included in group D. However, the 'Arda' variety had higher yields than the average yield for stress-free environments and was included in group B. It was noteworthy that the FLIP 09-51 C and FLIP 06-97C genotypes performed well both under stressed and stressed conditions (Fig. 1).
Based on stress tolerance indices MP, MRP, GMP, REI, STI, MSTIk1, MSTIk2, HM and ranking method genotypes, FLIP09-51C, FLIP06-97C and 'Aksu' all with highly grain yield under stressed and non-stressed conditions were identified as drought tolerance genotypes.Other chickpea genotypes were identified as semi-tolerance or semi-sensitive to drought stress (Table 3).Genotypes 'Diyar 95', FLIP06-39C and FLIP03-112C displayed the lowest values for these indices and were noted as the most susceptible genotypes under both conditions.Farshadfar and Sutka (2002) and Gholiouri et al. (2009) considered it the best possible selection criteria where a genotype should be distinguished from the other genotypes.
The indices SSI, SDI, TOL, ATI and SSPI were highly correlated with yield under non-stressed condition.Thus, these indices can be used to identify cultivars in the tolerant group for non-stressed conditions.Therefore, the genotypes FLIP09-51C and FLIP97-503C for TOL, ATI and SSPI, and FLIP97-503C for SSI and SDI were found as the most drought tolerant genotypes under non-stressed condition when genotype was based on the ranking method.The genotypes FLIP06-93C and FLIP03-131C displayed the lower amount of SSI, SDI, TOL, ATI and SSPI and were as the most susceptible genotypes under non-stressed condition.The index SSI has been widely used by researchers to identify sensitive and resistant genotypes (Golabadi et al., 2006;Sio-Se Mardeh et al., 2006;Shirani Rad and Abbasian, 2011).The greater value of SSI indicated the larger drought tolerance under stress and the cultivars with greater SSI were higher drought sensitivity.Yp showed positive correlation with SSI, MP, TOL, STI, GMP, HM, PI and K1STI, while the Ys indicated positive correlation with the HM, YI, YSI, RR and K2STI indices (Sabaghnia et al., 2014).
The indices YI, DI, GM, SNPI, RDI and DTE showed significant and positive correlation with grain yield (Ys) under stress conditions.Based on these indices and ranking method, the genotypes FLIP03-131C, FLIP09-51C, FLIP06-97C and 'Aksu' for YI and DI, FLIP03-131C for GM, FLIP03-131C and FLIP06-97C for SNPI, RDI and DTE were as the most drought tolerant genotypes.The yields of these genotypes under stress conditions have also been high (Table 3).Shiranirad and Abbasian (2011) 442 reported that STI, GMP and YI were able to identify rapeseed cultivars producing high yield under both stress and non-stressed conditions.
In the present study, the GM, RDI, DI and DTE showed significant and positive correlation with yield under stress when stress is too severe (Table 4).FLIP03-131C genotype with high yield under stress produced a lower yield under stressless condition and showed the highest RDI, DI and DTE (Table 3).Positive correlations among the MP, GMP and Yp indices (Toorchi et al., 2012) and positive correlations among the GMP, MP and STI indices (Dehghani et al., 2009) were reported in canola.

Correlation analysis
A correlation analysis between grain yield and drought tolerance indices is a good criterion for screening the best genotypes.An appropriate index should have a significant correlation with grain yield under both conditions (Mitra, 2001).Correlation coefficients between Yp, Ys, and other quantitative indices of drought tolerance, were calculated to identify the most desirable drought tolerant criteria (Table 4).The results indicated that there were positive and significant correlations among Yp and Ys with MP, MRP, GMP, REI, STI, MSTIk1, MSTIk2 and HM under stressed and non-stressed conditions.However, the indices Pi and RDY showed negative and significant correlation with both Yp and Ys.Therefore, these indices can be used to select genotypes that are better adapted to both conditions.
The results concerning MRP, REI, GMP and STI are in agreement with Bennani et al. (2017).Naghavi et al. (2013) reported that yield in stress and non-stress conditions were significantly and positively correlated with STI, GMP, MP, MSTIk1 and MSTIk2 in corn cultivars.Toorchi et al. (2012) showed that correlation between MP, GMP, Ys and Yp was positive in spring canola.Moreover, significant repeatable correlations were found between yield (Ys) under stress conditions and the drought indices (YI, RDI, DI, GM, DTE, SSI, TOL, SDI, ATI and SSPI).Yield in stress condition (Ys) indicated significantly and positive correlation with the YI, RDI, DI, GM and DTE, while showed significantly and negative correlation with the indices SSI, TOL, SDI, ATI and SSPI.The yield (Yp) under non-stress condition had strong positive correlation with SSI, TOL, SDI, ATI and SSPI, where as it had significant and negative correlation with YI, RDI,DI, GM and DTE.These relationships were affected by the drought density and pointed that genotypes chosen based on these indices are characterized by drought tolerance criteria and will enhance yield for stress conditions.Significant and positive relationships were found between Ys and widely used drought indices MP, GMP, STI, SSI,TOL in many studies (Golabadi et al., 2006;Mohammadi et al., 2010;Farshadfar et al., 2012a;Rahmani et al., 2013).Also, Jafari et al. (2009) and Farshadfar et al. (2014) reported positive significant correlations between Ys and HM, YI, DI, MSTIk1 and MSTIk2 in their drought studies.Strong repeatable significant correlations were found between MP, MRP, GMP, REI and MSTIk1, and these indices have had greater corollation coefficient with grain yield (Yp and Ys) under both conditions.Similar findings were reported Golabadi et al. (2006), Mevlut andSait (2011), Drikvand et al. (2012), Moradi et al. (2012), Rahmani et al. (2013), Naghavi et al. (2013) and Farshadfar et al. (2014).
Given these results, the results of the hereby study showed that the indices MP, MRP, GMP, REI and MSTIk1 could be used as suitable indicators for screening drought tolerant chickpea genotypes.Farshadfar et al. (2001) stated that the most suitable indices for selection of drought-tolerant varieties are indicators showing a relatively high correlation with grain yield under both stressful and stress free environments.

Fig. 1 .
Fig. 1.The relationship between seed yield produced under nostress and drought stress environments

Table 2 .
Analysis of variance for yield performances of chickpea genotypes in non-stressed and stressed conditions and drought resistance indices DF: degrees of freedom; ns: non-significant; ** significant on 0.01 level.*significant on 0.05 level only in stressless conditions (group B) or stressful stressless conditions (group D), and genotypes with low performance in both conditions (group C).

Table 3 .
Mean comparison for yield performances of chickpea genotypes in non-stressed and stressed conditions and twenty-three drought resistance indices

Table 4 .
Simple correlation coefficients of stress indices with seed yield of 14 chickpea genotypes