IntroductionDurum wheat (Triticum durum L.) is grown on 10% of the world’s wheat area. In spite of its low acreage, durum wheat is an economically important crop because of its unique characteristics and end products. Drought stress is one of the most devastating environmental stresses that depresses wheat yield productivity in many parts of the world. Breeding for drought tolerance is critical for sustainable wheat production in these areas. Different indices, including tolerance (TOL), mean productivity (MP), geometric mean productivity (GMP), stress tolerance index (STI), stress susceptibility index (SSI), harmonic mean (HAM), yield index (YI), and yield stability index (YSI) have been employed for screening the stress tolerant genotypes. The objectives of the study were to assess durum wheat genotypes under stress and non-stress conditions and to evaluate drought resistance indices in identifying genotypes adapted to the conditions.Materials and methodsThe experiment was carried out at the research farm of Faculty of Agriculture, Razi University, Kermanshah, during 2015-2016 cropping season. In this study, 23 durum wheat genotypes originally from Iran and ICARDA were evaluated using a randomized complete block design with three replications under stress (rain-fed) and non-stress (irrigated) conditions. Irrigated plots were watered three times at flowering and grain filling stages. Rain-fed plots received no water other than rainfall. Grain yield (g/m2) was measured. Tolerance indices were calculated for genotypes based on the grain yield. Combined analysis of variance appropriate to RCBD was carried out using SAS. Environments (rain-fed and irrigated) were considered as fixed effects. Least significant difference (LSD) test was used for the mean comparisons. Orthogonal comparisons and correlation analysis were performed by SAS software. Principal component analysis (PCA) and biplot diagram were carried out by MINITAB 17 and Stat graphics 18.1.01, respectively.Results and discussionThe results of combined analysis of variance showed significant differences between environments (rain-fed and irrigated) and genotypes for grain yield. Orthogonal comparisons showed that there was a significant difference between Iranian genotypes (contrast 1) in both conditions. Also, there were significant differences between ICARDA genotypes (contrast 2) and Iranian and ICARDA genotypes (contrast 3). Results showed that water stress reduced the grain yield of all genotypes and mean grain yield in rain-fed conditions was 32% lower than that in irrigated conditions (the stress intensity was 0.32). Based on all calculated drought indices, in most cases genotypes 15, 10, 18, 12 and 19 were tolerant and genotypes 2, 9, 17, 4 and 7 were susceptible to drought stress. The results of correlation analysis showed that TOL, MP, GMP, STI, YI and HAM had significant (P≤0.01) and positive correlations with grain yield under non-stressed condition. The MP, GMP, STI, YI and HAM revealed a significant (P≤0.01) and positive correlations with yield under stressed condition. Positive and significant correlation were observed between Ys and Yp and also with MP, GMP, STI, YI and HAM indicated that these indices are the most suitable indices to screen genotypes in drought stress conditions. Principal component analysis showed that the first component explained 71% of the variation with Ys, Yp, MP, YI, GMP, STI and HAM. First dimension can be considered as the yield in both environments and drought tolerance. Second component explained 28% of the total obtained variation and can be named drought susceptible dimension. Hence, selection of genotypes with high PCA1 and low PCA2 are suitable for both stress and non-stress environments. Thus, Genotypes 18, 22 and 23 with rather higher PCA1 and lower PCA2 are superior genotypes under both stressed and non-stressed conditions (Group A). Genotypes 19, 14, 3, 16, 21 and 20 could be known as Group B. These genotypes are suitable for non-stressed conditions. Genotypes 4, 7, 17 and 13 are drought susceptible and had low yield in both conditions (Group D). Genotypes 15, 10, 12, 11 and 6 with high amount of yield stability index (YSI) had a relatively low yield in both conditions, but they were more stable genotypes than the others (Group C).ConclusionWhat can be concluded from these results are: 1) Identifying the genotypes with high and stable yield in both conditions which are 18, 22 and 23 originated from ICARDA.2) Identifying genotypes with low yield in both conditions and susceptible to drought which are 4, 7, 17 and 13.3) Suggesting genotypes 19, 14, 3, 16, 21 and 20 for non-stress conditions.