Stability Analysis And Correlations Among Different Stability Parameters For Grain Yield In Bread Wheat

Transcription

1 Scientia Agriculturae E-ISSN: X / P-ISSN: DOI: /PSCP.SA Sci. Agri. 6 (3), 2014: PSCI Publications Stability Analysis And Correlations Among Different Stability Parameters For Grain Yield In Bread Wheat M.A. Abd El-Shafi, E.M.S. Gheith, A.A. Abd El-Mohsen, H.S. Suleiman Agronomy Department, Faculty of Agriculture, Cairo University Corresponding Author: M.A. Abd El-Shafi Paper Information A B S T R A C T The present study aimed to evaluate ten bread wheat genotypes across 8 Received: 16 April, 2014 environments (the combinations of 2 years x 2 irrigation regimes x 2 sowing dates) during 2011/2012 and 2012/2013 seasons at Experimental Accepted: 26 May, 2014 and Research Station, Faculty of Agriculture, Cairo University, Giza Governorate, Egypt. Significant differences were observed among bread Published: 20 June, 2014 wheat genotypes for grain yield (ton/fed). Combined analysis of variance of grain yield across eight environments showed highly significant (p<0.01) mean squares due to genotypes (G), environments (E) and genotypes x environments interaction (GEI), suggesting differential response of genotypes across studied environments and the validity of stability analysis. Six parametric of stability statistics were performed ( X i, b i, S 2 di, R 2 i, W 2 i and S 2 i). Stability analyses for grain yield of wheat genotypes revealed that the genotypes Bohouth 8, Cham 8, L-R 40, Sids 1, Cham 10 and Sahel 1 were more stable than others, expressed in 4, 4, 3, 3, 3 and 3 out of all 6 stability statistics used, respectively. Thus, these genotypes could be suggested to be more stable than others for these statistics. They have a low contribution to the genotype by environment interaction. Therefore, the above mentioned genotypes could be recommended as commercially stable and high yielding cultivars and/or incorporated as breeding stocks in any future breeding programs aiming to produce high yielding lines of bread wheat PSCI Publisher All rights reserved. Key words: Genotype-environment interaction, Bread wheat, Stability, Rank correlations Introduction Bread wheat (Triticum aestivum L.) is one of the most important cereal crops in Egypt, where plurality of people depends upon in their food. Egypt imports about 45% of its wheat requirements (Gad, 2010). Therefore, increasing wheat production is an important goal to reduce the gap between production and consumption (Rizkalla et al., 2012). The production of wheat can be increased either by increasing cultivated area or by increasing yield per unit area. Currently, it is nearly impossible to increase area under wheat crop due to competition with other crops and because of restricted irrigation water supply, etc. Therefore, the only alternative left is to increase yield in unit area by better crop management techniques and introducing high yielding varieties of resistance against environmental stresses. However, cultivars often do not perform in a similar manner when tested in multiple environments. This phenomenon is due to the presence of genotype by environment interaction (GEI). GEI is differential genotypic expression across environments. GEI complicates identification of superior genotypes, pointing out the need for growing different cultivars in different areas of the target region. Thus detection of areas in which genotypes perform similarly becomes a priority for cultivar evaluation and recommendation (Gauch and Zobel 1997). Also, GEI is of major importance because it provides information about the effects of different environments on cultivar performance and plays a key role for assessment of performance stability of the breeding materials. So, new wheat varieties generally need to be evaluated at different environments for several years before being released. The new varieties with desired traits that add value to the product should be tested for the stability of these traits in the target environments (Kang 1998). Mustatea et al (2009) reported that high-yielding cultivar can differ in yield stability and suggested that yield stability and high grain are mutually exclusive. Also, Shah et al (2009) found highly significant variances for genotype x location, genotype x year and genotype x location x year interactions for all studied traits of ten wheat varieties. Many studies have been conducted to investigate stability of wheat genotypes under different environments (Amin et al 2005; Aycicek and Yildirim, 2006; Ülker et al 2006; Rasul et al 2006; Akcura et al 2009; Parveen et al 2010; Al-Otayk 2010; El- Ameen 2012; Mohamed et al 2013). Numerous methods have been used to identify the stability of a genotype. The first description was by Finlay and Wilkinson (1963), who defined stability as the linear relationship of the genotype yield over environments by the

2 regression coefficient (b i ); where a genotype with b i =1 was considered stable. Eberhart and Russell (1966) further improved the idea by implementing the regression deviation mean squares (S 2 di) as a measure of stability. They recommended that the genotype stability is expressed in terms of three empirical parameters: the mean performance, the slope of regression line (b i ), and the sum of squares of deviation from regression (S 2 di). Therefore, a genotype with high mean yield over the environment, unit regression coefficient (b=1) and deviation from regression equal to zero (S 2 di =0) will be a better choice as a stable genotype. Other indices proposed for measuring response of crop cultivars and stability of production in variable environments included the coefficient of determination (R i 2 ) (Pinthus 1973). This R 2 measures the proportion of a variety s production variation that is due to linear regression. Wricke (1962) suggested using genotype environment interactions (GEI) for each genotype as a stability measure, which he termed as ecovalance (W 2 i). Francis and Kannenberg (1978) used the environmental variance (S 2 i) and the coefficient of variation (CV i ) of each genotype as stability parameters. The level of association among stability estimates of different models is signal of whether one or more estimates should be obtained for prediction of cultivar behavior, and also helps the breeder to choose the best stability parameter(s) (Duarte and Zimmermann 1995). The objective of our study was to determine phenotypic stability of grain yield in wheat genotypes and evaluate the level of association among the stability parameters. Materials And Methods Ten bread wheat genotypes were evaluated in eight environments as follows: four environments in 2011/2012 at Experiment and Research Station, Faculty of Agriculture, Cairo University, (Giza Governorate) with two irrigation regimes and two different sowing dates. While the other four environments were conducted during 2012/2013 at the same location. Details of genotypes and the eight environments are given in Tables (1) and (2), respectively. Table 1. Name, pedigree and origin of the studied wheat genotypes. No. Name Pedigree Origin G 1 Sids 1 HD2172/Pavon ''S"// /Maya 74 "S" - SD46-4SD 1SD osd. Egypt G 2 Sahl 1 NS 732 / PIMA // VEERY "S" Egypt G 3 Cham 10 KAUZ//KAUZ/STAR CMBW90M4994-0T0PY-13M-015Y-4Y-0B-AP Syria G 4 Cham 8 JOPATICO-73/BLUE JAY//URES-81 CM67458 Syria G 5 Bohouth 8 KAUZ S /ABE ICW AP-0BR-5AP-0AP Syria G 6 Bohouth 6 CROW`S` CM Syria G 7 Bohouth 4 S-201 Syria G 8 L-R 40 KAUZ/STAR Sudan G 9 L-R 51 UP2338 Sudan G 10 L-R YSN Yemen Table 2. The environments used in this study. Environments Season Environments combinations Sowing date Irrigation regimes E1 2011/ November water-stress E2 2011/ November well-watering E3 2011/ December water-stress E4 2011/ December well-watering E5 2012/ November water-stress E6 2012/ November well-watering E7 2012/ December water-stress E8 2012/ December well-watering The trials were established in a split- split plot design with randomized complete blocks arrangement in 3 replications. Main plots were assigned to the two sowing dates (15 November as favorable date and 20 December as late date), sub-plots were assigned to the two irrigation regimes well-watering and water-stress (Well-irrigated plots were watered at planting, tillering, jointing, flowering and grain filling stages). Water stress plots received water only at planting) and sub-subplots were assigned to the 10 genotypes. Each sub-subplot consisted of 6 rows, 3 m long and 20 cm wide. At harvest the four middle area of each plot was taken to determine grain yield/plot (plot size = 2.4 m 2 ) and then converted to grain yield (ton/fed). All genotypes were sown by a seed drill at a seed rate of 60 kg feddan. The soil of the experimental site was clayey loam, containing 35% clay, 22% silt, 37% fine sand and 6% coarse sand with a ph of 7.8 (according to the analysis done by the Soil and Water Res. Inst., ARC, Egypt). All other agricultural practices were followed according to the recommendations of ARC, Egypt. Statistical analyses 136

3 Normality distributions in each environment were checked out by the Wilk Shapiro test (Neter et al 1996). An analysis of variance (ANOVA) was done for each environment separately. A combined analysis of variance was done from the mean data of each environment, to create the means for the different statistical analyses methods. Homogeneity test of variances were performed according to procedures reported by Gomez and Gomez (1984). To evaluate the stability of tested genotypes across the eight environments, parametric stability statistics were used to estimate stability in this study. Six stability parameters were performed. The first and second were proposed by Eberhart and Russell s (1966), i.e. the slope value (b i ) and deviation from regression parameter (S 2 di). The third was coefficients of determination (R i 2 ) by Pinthus (1973). The fourth one was ecovalance (W 2 i) by Wricke (1962) and the fifth one was environmental variance (S 2 i) by Francis and Kannenberg s (1978), besides mean performance across environments ( X i ). Also, Spearman's rank correlation coefficients were computed for each pair of the possible pair-wise comparisons of the stability parameters and the significance of the rank correlation coefficient was tested according to Steel et al (1997). All statistical analyses were carried out using MSTAT-C software package (Freed et al 1989), GENES computer software (Cruz, 2013) and MS Excel. Results and discussion Analysis of variance Combined analysis of variance for grain yield is presented in Table (3). Results of combined analysis showed that differences among environments were highly significant for grain yield, indicating that the eight environments are different in their conditions. Significant (p<0.01) differences among genotypes were detected for grain yield. Significant (p<0.01) mean squares due to genotype x environments interaction were detected for grain yield, which indicated that genotypes performed differently at different environments. Table 3. Combined analysis of variance for grain yield of 10 bread wheat genotypes tested across different environments. S.O.V d.f Mean squares P-value Environments (E) ** Replicates/E Genotypes (G) ** G x E ** Error ** Significant at 0.01 probability level. It is clear from these results that the tested genotypes must be evaluated under different environments, especially for grain yield which is regarded as the most important trait. Mean performance The mean performance of the ten genotypes for grain yield at each environment and their combined means are presented in Table (4). Mean grain yield under the eight environments was 1.71, 2.02, 1.27, 2.72, 1.93, 2.30, 1.43 and 2.40 ton/fed, respectively. The overall mean for grain yield of the ten genotypes across the eight environments was 1.97 ton/fed, while, mean yield of the checks (Sahel 1 and Sids 1) were 1.95 and 2.05 ton/fed, respectively. Results of tested genotypes across environments presented two genotypes (Bohouth 4 and L-R 67) where, grain yield was significantly more than the check varieties. However the four genotypes (Bohouth 8, L-R 51, Bohouth 6 and L-R 40) were significantly less than the checks varieties for grain yield (ton/fed). Table 4. Mean grain yield (ton/fed) for ten bread wheat genotypes and their combined means across eight environments. Genotypes E1 E2 E3 E4 E5 E6 E7 E8 Combined Sahel e Cham d Sids c Bohouth a L-R h Cham e L-R b Bohouth f L-R f Bohouth g Mean Stability of tested genotypes Pooled analysis of variance for grain yield across the eight environments is presented in Table (5). The results revealed that there were significant differences among the tested genotypes for grain yield, which suggested that the genotypes differed considerably with respect to yield performance. Joint regression analysis of variance showed that the mean squares due to genotypes (G), environments (E) and GEI were highly significant for grain yield, indicating the 137

4 presence of wide variability among the genotypes as well as environments under which the experiments were conducted. The significant estimates of GEI indicated that grain yield was unstable and may considerably fluctuate with change in environments. These findings are in close agreement with those of Amin et al (2005), Aycicek and Yildirim (2006), Ülker et al (2006), Rasul et al (2006), Akcura et al (2009), Parveen et al (2010), Al-Otayk (2010), El-Ameen (2012) and Mohamed et al (2013). The GEI was further partitioned into linear and non-linear components and mean squares for both of them were highly significant (P < 0.01), suggesting that predictable as well as un-predictable components were involved in the differential response of stability for grain yield. Similar results were reported by Amin et al (2005), Aycicek and Yildirim (2006), Ülker et al (2006), Rasul et al (2006), Akcura et al (2009), Parveen et al (2010), Al-Otayk (2010), El- Ameen (2012) and Mohamed et al (2013). Table 5. Joint regression analysis of variance for grain yield of the 10 genotypes tested in eight environments. S.O.V d.f Mean squares P-value Total Environments (E) ** Genotypes (G) ** G E ** E + (G E) ** Environment(Linear) ** G E (Linear) ** Pooled deviation ** Pooled Error ** Significant at 0.01 probability level. Significant environment (linear) variance implies linear variation among environments for grain yield. The G x E (linear) interaction was significant against pooled deviation, suggesting the possibility of the variation for grain yield and indicated the presence of genetic differences among genotypes for their regression on the environmental index (Table 5). The linear component of GEI was found to be more than the non-linear component (pooled deviation). These results are in consistent with those of Mohamed et al (2013) who have reported predominance of linear component of GEI for grain yield per plant. The estimates of six parametric stability statistics for 10 bread wheat genotypes grain yield (ton/fed) and their ranks tested across eight environments are presented in Table (6). The genotypes showed significant differences in grain yield. Taking mean yield as the first parameter for evaluating the genotypes, Bohouth 4, L-R 67, Sids 1 and Cham 10 gave the best mean yields while Bohouth 6 and L-R 40 had the lowest mean yields across environments (Table 6). Mean grain yield across eight environments showed substantial changes in ranks among the genotypes, reflecting the presence of high G-E interactions (Baker 1998). Table 6. Mean values of grain yield (ton/fed) and 5 parametric stability statistics and their rank for 10 bread wheat genotypes tested across 8 environments. Genotypes X Rank b i Rank S 2 di Rank R 2 i Rank W 2 id Rank S 2 i Rank Fr Sahel ** Cham ** ** Sids ** ** Bohouth ** ** L-R ** Cham ** L-R ** ** Bohouth ** L-R ** * Bohouth ** ** Mean *,** Significantly different from 1.0 for the regression coefficients and from 0.0 for the deviation mean squares at the 0.05 and 0.01 levels of probability, respectively. Fr. =frequency of the number of stability parameters showing stability for each genotype, if a genotype had seven values of Fr., it could be considered most stable. According to Finlay and Wilkinson (1963), who defined varieties with general adaptability as those with average stability (b i = 1.0) when associated with high mean yield across tested environment. Eberhart and Russell (1966) proposed that an ideal genotype is the one which has the highest yield across a broad range of environments, a regression coefficient (b i ) value of 1.0 and deviation mean squares of zero. Then, regression coefficient value near indicates less response to environmental changes, and hence showing more adaptiveness. Thus, a genotype with unit regression coefficient (b i =1) and deviation not significantly different form zero (S 2 di=0) is said to be the most stable genotype. The regression coefficients (b i ), ranged from to for grain yield (Table 6). This wide range of regression coefficients indicates that the ten genotypes had different responses to environmental changes. Three of studied genotypes (30%) had regression slopes for grain yield that did not differ from 1.0, indicating good potential for yield response under environmental 138

5 conditions for these genotypes. Based on results of the regression analysis, the genotypes Bouhouth 8, Cham 8 and Sahel 1 were classified as highly stable across environments because the regression coefficients of these genotypes did not differ significantly from 1.0. Moreover, the S 2 di values (Table 6) of L-R 40 was not significantly different from zero, and therefore it can be considered as of good adaptability. The genotype Bouhouth 4 had larger b i value, indicating greater sensitivity to environmental change and was relatively suitable in favorable environments. All genotypes except L-R 40 were unstable genotypes according to the S 2 di values, indicating high sensitivity to environmental changes. Regarding the fourth parameter coefficient of determination (R i 2 ), the results in Table 6 revealed that R i 2 ranged from to 0.991, which indicated that 89% to 99% of the mean grain yield variation was explained by genotype response across environments and indicating stability differences among genotypes. The coefficient of determination is often considered a better index for measuring the validity of the linear regression than S 2 di, because its value ranges between zero and one. Bilbro and Ray (1976) suggested that R i 2 could be useful in measuring dispersion around the regression line and therefore related to the predictability and repeatability of the performance within environments. Data in Table (6) revealed that the coefficient of determination (R i 2 ) values for Cham 8, Bohouth 4, Sids 1, L-R 40 and Cham 10 were 99%, 99%, 99%, 98% and 98%, respectively indicating the reliability of the linear response of these genotypes. Ecovalence indicates the contribution of each genotype to the GEI (Wricke 1962). The cultivars with the lowest ecovalence contributed the least to the GEI and are therefore more stable. Using Wricke (1962) stability parameter, W 2 i, the genotype Cham 8 followed by Bohouth 8 and Sahel 1 with the lowest ecovalence were considered to be stable, whereas the Bohouth 4, Bohouth 6, L-R 40 and L-R 51 with the highest W 2 i were unstable and had the highest contribution to GEI. Francis and Kannenberg (1978) reported that the environmental variance (S 2 i) is used as the stability parameter. Ortiz et al (2001) suggested that it may be possible to select simultaneously for high and stable grain yield by selecting outyielders that exhibit a low S 2 i. According to Francis and Kannenberg (1978), genotypes exhibiting low environmental variance (S 2 i) are considered as stable (Lin et al.,1986). Table (6) shows that genotypes L-R 40, L-R 51, Bohouth 6, Bohouth 8 and Cham 8 had smaller environmental variance (S 2 i) than those of the rest for grain yield, confirming their high stability. In summary, parametric stability analysis for grain yield of wheat genotypes revealed that genotypes Bohouth 8, Cham 8, L-R 40, Sids 1, Cham 10 and Sahel 1 were more stable, expressed in 4, 4, 3, 3, 3 and 3 out of all 6 stability statistics used, respectively. Thus, these genotypes would be suggested to be more stable than others for these statistics. They have a low contribution to the genotype by environment interaction. Therefore, the above mentioned genotypes could be recommended as commercially stable and high yielding cultivars and/or incorporated as breeding stocks in any future breeding programs aiming at producing high yielding lines of bread wheat. Interrelationships among stability parameters Correlation analysis was used to study the relationships between mean yield and stability parameters, as well as between studied stability parameters. Table (7) shows the Sperman`s correlation coefficient by using ranking of the ten bread wheat genotypes, after applying the methods of stability analysis. The ranks of genotypes and 8 environments after applying the method of stability analysis were used for rank correlation. Table 7. Estimates of rank correlation coefficients among grain yield and stability parameters. Variable Mean yield b i S 2 di R 2 i W 2 i S 2 i Mean yield b i S 2 di R 2 i 2 W i CV i% ** ** ** ** Correlation coefficients are significantly different from zero at 0.01 level of probability. The results of Spearman s coefficient of rank correlations showed that mean yield was statistically significant (P<0.05) and negatively correlated with environmental variance (S 2 i) parameter (r = -0.95**). Furthermore, the correlation was positive between mean yield and R 2, but this correlation was statistically non-significant. The results in Table (7) showed that b i tended to be independent of the other stability statistics except with W i 2. These results were in harmony with those obtained by Shah et al (2009). Deviations from regression (S 2 di) exhibited a positive and highly significant correlation (r=0.73**) with R 2. These findings agree with other researchers (Letta 2007 and Shah et a, 2009). Conclusion These results revealed that significant GEI affects the grain yield of studied genotypes and there is a necessity for multiple testing through different environments. It is advisable to test new genotypes in the environments of intended use before release to farmers and it is essential to identify genotypes, which manifest relatively low G E interactions with stable yields in test environments. Genotypes Bohouth 8, Cham 8, L-R 40, Sids 1, Cham 10 and Sahel 1 are likely to be 139

6 stable and may be recommended for cultivation in different locations in Egypt as they had high relative yield performance and revealed high stability. References Akcura M, Kaya Y, Taner S Evaluation of durum wheat genotypes using parametric and nonparametric stability statistics. Turk. J. Field Crops 14: Al-Otayk SM Performance of yield and stability of wheat genotypes under high stress environments of the Central Region of Saudi Arabia. Met. Env. Arid Land Agric. Sci. 21: Amin M, Mohammad T, Khan AJ, Irfaq M, Ali A, Tahir GR Yield stability of spring wheat (Triticum aestivum L.) in North West Frontir province, Pakistan. Songklanakarin J. Sci. Technol. 27: Aycicek M, Yildirim T Adaptability performances of some bread wheat (Triticum aestivum L.) genotypes in the Eastern Region of Turkey. Int. J. Sci. Technol. 1: Baker RJ Tests for cross over genotype-environment interactions. Can. J. Plant Sci. 68: Bilbro JD, Ray IL Environmental stability and adaptation of several cotton cultivars. Crop Sci. 16: Cruz CD GENES - a software package for analysis in experimental statistics and quantitative genetics. Acta Scientiarum 35(3): Duarte JB, Zimmermann MJ Correlation among yield stability parameters in common bean. Crop Sci. 35: Eberhart SA, Russell WA Stability parameters for comparing varieties. Crop Sci. 6: El-Ameen T Stability analysis of selected wheat genotypes under different environment conditions in upper Egypt. Afr. J. Agric. Res. 7: Finlay KW, Wilkinson GN The analyses of adaptation in a plant-breeding programme. J. Agric. 14: Francis TR, Kannenberg LW Yield stability studies in short-season maize: I. A descriptive method for grouping genotypes. Canadian J. Plant Sci. 58: Freed R, Einensmith SP, Gutez S, Reicosky D, Smail VW, Wolberg P Guide to MSTAT-C Analysis of Agronomic Research Experiments. Michigan State University, East Lansing, U.S.A. Gad KIM Genetic studies on earliness in wheat. Ph.D. Thesis, Faculty of Agriculture, Genetics Department, Cairo University, Egypt. Gauch HG, Zobel RW Identifying genotypes and targeting genotypes. Crop Sci. 37: Gomez AK, Gomez AA Statistical Procedures for Agricultural Research. John Wiley and Sons. New York, USA. Kang MS Using genotypes by environment interaction for crop cultivar development. Adv. Agron. 35: Letta T Genotype-environment interactions and correlation among some stability parameters of yield in durum wheat (Triticum durum Desf) genotypes grown in South East Ethiopia. African Crop Sci. Conference Proceeding, 8: Lin CS, Binns MR, Lefkovitch LP Stability analysis: Where do we stand? Crop Sci. 26: Mohamed SH, Mohamed GIA, El-Said RAR Stability analysis for grain yield and its components of some durum wheat genotypes (Triticum durum) under different environments. Asian J. of Crop Sci. 5(2): Mustatea P, Saulescu NN, Ittu G, Paunescu G, Voinea L Grain yield and yield stability of winter wheat cultivars in contrasting weather conditions. Rom. Agric. Res. 26: 1-8. Neter J, Khutner M, Nachtsheim C, Wasserman W Applied Linear Statistical Models. 4 th Ed. Chicago Irwin Series. Time Mirror. Education Group, pp Ortiz R, Wagoire WW, Hill J, Chandra S, Madsen S, Stolen O Heritability and correlations among genotype-by-environment stability statistics for grain yield in bread wheat. Theor. Appl. Genet. 103: Parveen L, Khalil IH, Khalil SK Stability parameters for tillers grain weight and yield of wheat cultivars in North-West of Pakistan. Pak. J. Bot. 42: Pinthus JM Estimate of genotype value: a proposed method. Euphytica. 22: Rasul I, Zulkiffal M, Anwar J, Khan SB, Hussain M, Din R Grain yield stability of wheat genotypes under different environments in Punjab. J. Agric. Soc. Sci. 2: Rizkalla AA, Hassien BA, Al-Ansary AMF, Nasseef JE, Hussein MHA Combining ability and heterosis relative to RAPD marker in cultivated and newly hexaploid wheat varieties. Aust. J. Basic Applied Sci. 6: Shah SIH, Sahito MA, Tunio S, Pirzado AJ Genotype-environment interactions and stability analysis of yield and yield attributes of ten contemporary wheat varieties of Pakistan. Sindh Univ. Res. J. 41: Steel RGD, Torrie GH, Dickey DA Principles and Procedures of Statistics: A Biometrical Approach. 3 rd ed. McGraw-Hill, New York. Ülker M, Sönmez F, Ciftci V, Yilmaz N, Apak R Adaptation and stability analysis in the selected lines of tir wheat. Pak. J. Bot. 38: Wricke G Über eine methods zur Erfassung der Ökologischen streubreite in Feldversuchen. Pflanzenzuchtung. 47: