INTRODUCTION Coriander (Coriandrum sativum L.) is an important seed spice crop belongs to the Apiaceae family having somatic chromosome number 22 (2n=22). It grows well in tropical and sub-tropical regions with dry climate. The high temperature at maturity phase and insufficient and erratic distribution of rainfall affects seed yield as well oil content. Tolerance of plants against moisture stress is a complex quantitative trait with low heritability (Fereres and Soriano 2007) and thus it halts and complicates the breeding for resistance to drought or moisture stress. An ideal and effective approach for identifying stress tolerant genotypes is based on mean seed yield under drought stress and non-stress environments. Araus et al. (2002); White et al. (1998) has suggested selection of genotypes based on their yield response in two or more normal and stress environments. The various stress indices that screen or select the genotypes based on their resistant or susceptible (Fernandez, 1992) towards moisture stress have been used by many researcher and plant breeders (Mitra, 2001). Hall, (1993) defined the drought tolerance or resistance as response of genotype in terms of grain yield when they are subjected to same level of moisture stress. Blum, (1998) defined drought stress or susceptibility of a genotype a as function of reduction in yield under drought stress. The index MP (mean productivity) measures the average yield of genotype across the two environments, while TOL (Rosielle and Hamblin 1981) measures the difference in degree of yield response over the stress (Ys) and non-stress (Yp) environments. Another index SSI captures the changes in both potential and actual yields in variable environments (Fischer and Maurer 1978). The larger value of TOL and SSI indicates relatively more sensitivity towards stress, thus a smaller value of TOL and SSI are suitable and indicates tolerance for moisture stress. The value of stress susceptibility index more than one (SSI >1) indicates above average susceptibility, while a value less than one indicates (SSI<1) below-average susceptibility towards moisture stress (Guttieri et al., 2001). The index STI selects the genotypes that produce high yield under both stress and non-stress conditions (Fernandez, 1992). The GMP index measures relative yield response of a genotype over two environments since in field environments, the stress level can be vary in severity (Fernandez, 1992). Thus, two indices viz., STI and GMP screen genotypes having high yield potential and stress tolerance (Fernandez, 1992). Similarly, two indices as YI (Gavuzzi et al., 1997) and SSPI (Moosavi et al., 2008) evaluate the genotypes based on yield stability under two environments. MATERIALS AND METHODS The present investigation was carried out using thirty genotypes of coriander taken from germplasm center of AICRP on seed Spices, SKNCOA, Jobner. These genotypes were evaluated at Agronomy Farm of S.K.N. College of Agriculture, Jobner in RBD layout with 3 replications during rabi 2016-17 in two environments, namely, (i) normal condition (non-stress, E1) and (ii) Limited moisture condition (stress, E2). The required irrigations was supplied in normal condition (E1), while half of the irrigations was provided in limited moisture condition (E2) in staggered manner from sowing to maturity of the crop. The each genotype was sown in a single row plot of 3 m length by maintaining crop geometry of 30 x 10 cm (R x P) in each environment or replication. The Jobner is located in Jaipur district of Rajasthan with typically semi-arid climate and falls in agroclimatic zone III A (Semi-arid Eastern Plain Zone) of Rajasthan. Annually, the place receives 300-400 mm rainfall and temperature during summers goes high as 480 C, while extreme low in winter (-1.00 C) season. The five randomly plant from each plot/environment selected and average seed yield per plant (g) was worked out for each of the genotype under both the conditions as non-stress (Yp) and stress (Ys). Statistical Analysis: The analysis of variance was carried out on pooled basis to assess the interaction between genotypes and environments. The variation accounted to each source of variation was estimated as percentage of variance explained of total sum of squares. The ranking of genotypes was carried out according to each of the stress index. Based on indices, the genotype with the highest value for Ys, Yp, MP, GMP, STI, K1STI, K2STI and YI and the lowest value for SSI, TOL and SSPI received a rank one. The correlation among the moisture stress indices as well as Yp, Ys were also carried out to determine the most desirable combination of stress indices for screening the best tolerant genotypes. The pooled ANOVA for AMMI and biplot analysis were performed using GEA-R version 4.1 (CIMMYT, Mexico). RESULTS AND DISCUSSION A. Pooled ANOVA and mean Comparison The mean squares (Table 1) due to genotypes were found highly significant for seed yield per plant on pooled basis indicating significant differences among the genotypes. The variance due to environments as well as genotypes x environments interaction for seed yield per plant (g) were also found highly significant indicating differential yield response of genotypes towards the two different environments (E1 and E2) and their role in character expression. The per cent share of each source of variance towards per cent total sum of squares (% TSS) revealed that 69 per cent of total sum of squares was accounted for environments (E) effect indicating most important source of yield variation. Similarly, the per cent share of genotypes (G) and G × E interactions effects was 22.37 per cent and 2.44 per cent of total sum of squares, respectively. The mean seed yield (Yp) was ranged from 4.42 g (UD-554) to 6.15 g (UD-769) in non-stress environment (E1) with the overall mean of 5.01 g. While, mean seed yield (Ys) was ranged from 3.09 g (UD-461) to 4.63 g (UD-705) in stress environment (E2) with the overall mean of 3.52 g. The average seed yield of genotypes in the stress condition (Ys) was found to be 29.74 percent lower than in the non-stress condition. Thus, the genotypes experienced moisture stress during the crop growing period in stress environment (E2) and the stress intensity (Fischer and Maurer) [12] was equal to 0.29. The top five performing genotypes were UD-769, UD-705, RCr-684, UD-529 and RCr-20, whereas the genotypes UD-554, UD-704, UD-461, UD-169 and UD-489 showed lower seed yield in non-stress (E1) environment. Similarly, genotypes UD-705, UD-769, UD-529, RCr-20 and RCr-684 depicted higher seed yield and UD-461, UD-747, UD-751, UD-23 and UD-566 showed lower seed yield in stress environment (E2). B. Moisture stress indices According to stress tolerance index (TOL), higher value of TOL indicates susceptibility of the genotype to stress and fluctuation in response in terms of seed yield in stress and non-stress conditions (Table 3). The genotypes UD-554, UD-580, UD-723, UD-704 and UD-489 were found most tolerant as they occupied lower values TOL index, whereas genotypes UD-513, RCr-684, UD-783, UD-709 and UD-747 were found most sensitive as they displayed higher values of TOL index. In accordance with stress susceptibility index (SSI), genotypes UD-513, UD-783, UD-747, UD-709 and UD-23 had showed higher values of SSI index indicating their minimum tolerance to moisture stress. The genotypes UD-554, UD-580, UD-723, UD-705 and UD-704 were found more tolerance against moisture stress (lower values of SSI). It was found that the SSPI and TOL index resulted the same genotype ranking with different values of their respective index. Similarly, the geometric mean productivity (GMP) resulted the same genotype ranking as stress tolerance index (STI) with different values of their respective index. A relatively similar pattern of ranking of genotypes was observed for the three indices viz., MP, GMP and STI under the study. According to the STI index (high STI), genotypes UD-705, UD-769, UD-529, RCr-684 and RCr-20 were found stress tolerant with high seed yield under both the conditions E1 and E2, while UD-461, UD-169, UD-751, UD-717 and UD-704 were the genotypes with least rank. According to mean productivity (MP), higher value of MP for a genotype is directly proportional to its tolerance towards moisture stress and vice versa thus genotypes UD-705, UD-769, UD-529, RCr-684 and RCr-20 were found most tolerant, whereas, the genotypes UD-461, UD-169, UD-717, UD-751 and UD-704 were found least tolerant towards moisture stress. The yield index (YI) can also serve as a selection criterion as it only ranks cultivars based on their yield (Ys) under stress environment (E2). Thus, according to this index, genotypes UD-705, UD-769, UD-529, RCr-20 and RCr-684 had displayed highest values for YI and Ys, hence proved to be more resilient towards moisture stress. The genotypes UD-461, UD-747, UD-751, UD-566 and UD-23 depicted lower YI and Ys indicating their susceptibility for moisture stress. C. Correlation analysis The simple correlation was carried out among the various stress indices including Yp, Ys to know the most desirable combination of indices suitable for screening of tolerant genotypes (Table 2). Talebi et al. (2007) has suggested that those indices are best which have high correlation with seed yield in both non-stress (E1) and stress environment (E2) conditions and would be able to identify potential high yielding and moisture stress tolerant genotypes. A significant association (r= 0.820**) was found between seed yield under stress (Ys) and non-stress condition (Yp). These results are found in corroboration with the findings of Fernandez (1992), Mohammadi et al. (2010), Farshadfar et al. (2012) and Sahar et al. (2016). The seed yield under non-stress (Yp) had high significant positive association with the eight indices as TOL, STI, MP, GMP, YI, SSPI and K1STI and K2STI and positive non-significant association with SSI. Similarly, index Ys had high significant positive association with STI, MP, GMP, YI and K2STI, K1STI and significant negative association with SSI. It was observed that the seed yield in stress (Ys) and non-stress condition (Yp) was significantly and positively correlated with the stress indices STI, MP, GMP, YI, K1STI and K2STI indicating that these indices screen genotypes which have uniform superiority or stress tolerance (Fernandez, 1992). Siahsar et al. (2010) in lentil, Zare and Saeidi et al. (2012) in barley, Singh et al. (2015) and Mohammed and Kadhem (2017) in wheat obtained similar results. D. Ranking method The screening of tolerant genotypes according to single criteria (indices) was found contradictory and different indices introduced different or same genotypes as stress tolerant. Thus, the sum of rank of all indices including Ys and Yp of genotypes were used to calculate overall rank of genotypes. Thus, according to this criterion a genotype with least rank sum will be the best genotype. The overall ranks of all the genotypes based on the above criteria are presented in Table 5. The genotypes UD-705, UD-529, UD-769, RCr-20 and RCr-475 occupied rank 1,2,3,4,5 respectively and identified as the most moisture stress tolerant genotypes. The genotypes UD-747, UD-461, UD-751, UD-23 and UD-169 were found most sensitive for moisture stress as they occupied. higher over all rank sum. Farshadfar et al. (2012) and Mohammed and Kadhem (2017) used overall rank sum methodology to screen the tolerant genotypes in wheat. E. AMMI and biplot analysis The AMMI and biplot analysis was also performed to visualize G × E interactions and to minimize the noise from interaction to clearly examine the yield stability (Ajay et al., 2020) of genotypes. The ANOVA for AMMI (Table 4) analysis revealed that MSS due to genotypes were found highly significant for seed yield per plant indicating non-linear response of genotypes towards the two different conditions. The bulk of the total variation that explained (Percent) was attributed for by the environments as 73.55 per cent, indicating the environments were diverse. The genotypes shared 23.84 per cent of total variation, while share of interaction was 2.60 per cent. It was also found from the AMMI analysis that PCA-1 captured almost all the share of the interaction sum of squares. The AMMI biplot (Fig. 1) between PCA-1 and Main effects provides a visual expression of the relationships between the genotypes and interaction with the prevailing environmental condition. The stability or adaptability of the genotypes can be assessed by mean seed yield and PCA scores of genotypes in the AMMI analysis (Purchase, 1997; Martin and Alberts, 2004). The mean, PCA-1, PCA-2 values of all genotypes are presented in Table 5. Higher PCA scores of a genotype (High responsive) indicated it’s the specific adaptability or stability to a certain environment, while lower PCA scores as approximate to zero (low responsive), indicates stability or adaptability over all the environments sampled. The environment E1 located at the right side of the main axis indicating the superiority if the E1 environments or non-stress environment. Similarly, the genotypes 19 (UD-705), 25 (UD-769), 12 (UD-529), 29 (RCr-684), 27 (RCr-20), 7 (UD-488), 28 (RCr-436), 30 (RCr-475), 15 (UD-573) and 10 (UD-513) were found generally high yielding as they were placed on right-hand side of the mid-point representing grand mean. The genotypes 19 (UD-705) followed by 25 (UD-769) and 12 (UD-529) recorded best average seed yield of 5.35, 5.33 and 5.04 gram and attained relatively small values of PCA-1 (0.07, -0.26 and 0.15 respectively) indicating they were stable and widely adapted genotypes. Naroui et al. (2013) emphasized stability in addition to seed yield thus, genotypes UD-280, RCr-436, UD-472, UD-520, RCr-20 and UD-705 attained lowest PCA-1 values (0.01, 0.01,-0.04, -0.04, 0.05 and 0.07, respectively) and average seed yield (3.99, 4.49, 4.25, 4.16, 4.75 and 5.35 gram, respectively). The genotypes UD-461(3.76 g) and UD-751(3.86) yielded least seed and attained relatively small PCA-1 values (0.23 and 0.15, respectively) indicating their average adaptability. Similar findings of screening of genotypes using PCA values were also reported by Amir et al. (2018) in coriander, Fufa (2018) in cumin, Naik et al. (2022); Rao et al. (2022) in wheat.