100 † Corresponding author © 2015 Conscientia Beam. All Rights Reserved. GENOTYPE X ENVIRONMENT INTERACTION AND STABILITY ANALYSIS FOR YIELD AND YIELD RELATED TRAITS OF DESI-TYPE CHICKPEA (CICER ARIETINUM L.) IN ETHIOPIA Getachew Tilahun1† --- Firew Mekbib2 --- Asnake Fikre3 --- Million Eshete4 1Amhara Regional Agricultural Research Institute, Gonder Agricultural Research Center, Ethiopia 2Plant Science Department, Haramaya University, Dire Dawa, Ethiopia 3Crop research Director, EIAR, Addis Ababa, Ethiopia. 4Ethiopian Institute of Agriculture Research, Debre Zeit Agriculture Research Center, Ethiopia ABSTRACT Chickpea is the major pulses grown in Ethiopia, mainly by subsistence farmers usually under rain-fed conditions. However, its production is constrained due to genotype instability, environmental variability and interaction of genotype with environment. This research was carried out to examine the magnitude of environmental effect on yield of chickpea genotypes and to investigate the stability and adaptability of the genotypes under different agro-ecological conditions. 17 genotypes each of were evaluated in RCBD with four replications in five environments. Various stability indices were used to assess stability and genotype by environment performances. The combined ANOVA for yield and yield related traits revealed highly significant (P≤0.01) differences for genotypes, environments and their interaction. The significant interaction showed that the genotypes respond differently across the various environments. At Akaki, Chefe Donsa, Debre Zeit, Dembia and Haramaya the top performing genotype were DZ-2012-CK-0040 (2229 kg/ha), DZ-2012-CK-0027 (3966 kg/ha), DZ-2012-CK-0040 (4060 kg/ha), DZ-2012-CK-0032 (1394 kg/ha) and Natoli (3247 kg/ha) respectively. The first two PCs explained 84.3% of the variance of original variables for the genotypes. There were remarkable inconsistencies with the univariate stability parameters to select stable genotypes. However, multivariate approach, the AMMI model was better for partitioning the G x E into the causes of variation. Based on ASV value, DZ-2012-CK-0035 was most stable genotype. As per AMMI biplot, Minjar and local variety were the most widely adapted genotypes. Dembia and Haramaya are the most discriminative environments. Environments Debre Zeit and Chefe Donsa were the favorable environment. Genotypes, DZ-2012-CK-0040, DZ-2012-CK-0036 and DZ- 2012-CK-0040, DZ-2012-CK-0032 and variety Natoli were recommended as specifically adapted to sites Akaki, Chefe Donsa, Debre Zeit, Dembia and Haramaya respectively. Keywords: AMMI, ASV, Clustering, Desi-type, Phonologic traits, Multivariate, Univariate statistics. Current Research in Agricultural Sciences 2015 Vol. 2, No. 3, pp. 100-113 ISSN(e): 2312-6418 ISSN(p): 2313-3716 DOI: 10.18488/journal.68/2015.2.3/68.3.100.113 © 2015 Conscientia Beam. All Rights Reserved. http://crossmark.crossref.org/dialog/?doi=10.18488/journal.68/2015.2.3/68.3.100.113 http://crossmark.crossref.org/dialog/?doi=10.18488/journal.68/2015.2.3/68.3.100.113 Current Research in Agricultural Sciences, 2015, 2(3):100-113 101 © 2015 Conscientia Beam. All Rights Reserved. 1. INTRODUCTION Chickpea is the third leading legume grain in the world and in Ethiopia as well. Its range of cultivation extends from the Mediterranean basin to the Indian sub-continent and south ward of Ethiopia and eastern Africa highlands [1]. About 92% of the area and 89% of production of the grain are concentrated in the semi-arid tropical countries [1]. In Ethiopia, it accounts for about one third of the area and production of pulses following faba bean(Vicia faba L.) and haricot bean (Phaseolus vulgaris L.) [2]. Desi-type chickpeas (Cicer arietinum L.) with colored and thick seed coat with small and angular in shape having an average 1-2 seeds per pod. The plants are short with small leaflets and purplish flowers, and contain anthocyanin. Desi-type chickpea seeds germinate at an optimum temperature (28-33°C) and moisture level 10 % in about 5-6 days [3]. The flowers are generally pink and the plants show various degrees of anthocyanin pigmentation. The Desi-types account for 80-85% of chickpea global production area [3]. The crop grows mainly concentrated in two into major administrative regional stats namely Amhara and Oromiya and it accounts 94% the total production in the country with altitude range of 1400-2300 masl, and annual rainfall of 700- 2000 mm [4]. Environmental factors such as soil moisture, sowing time, fertility, and temperature and day length have strong influence during various stages of plant growth [5]. The environment is changing day-by-day and this implies that it is necessary to evaluate crop genotypes at different environments to assess their performances. The performance of a genotype is not always the same in different environments as it is influenced by environmental factors. To assess yield stability among varieties, multi-location trials with appropriate stability analysis method is required. In Ethiopia, there is no sufficient information on the genotype by environment interaction effects on yield and yield related traits. Therefore the current research was undertaken To examine the magnitude of environmental effect on yield and yield related traits of Desi-type chickpea genotypes,to study the nature and extent of G x E Interaction on seed yield of Desi -chickpea genotypes and to investigate the stability and adaptability of the varieties under different agro- ecological condition. 2. MATERIALS AND METHODS The experiment was conducted during the 2012/13 main cropping season at five environments representing various chickpea agro-ecologies of Ethiopia. The sites were Akaki, Chefe Donsa Debre Zeit, Dembia and Haramaya. Fourteen pipelines and three released Desi-type chickpea genotypes were included in the study. The plant materials were obtained from Debre Zeit Agricultural Research Center. Planting of the genotypes was done in early mid August up to first week of September using Randomized Complete Block Design (RCBD) with four replications at each site under rain fed conditions. Each genotype was planted in six rows of 4 m row length and at 1.2 m width. A spacing of 30 cm between rows and 10 cm between plants were used on a plot size of 4.8 m2. Fertilizer was not applied. Weeding and other management practice were done as required for each site. Data were recorded on days to 50% flowering, 90% physiological Current Research in Agricultural Sciences, 2015, 2(3):100-113 102 © 2015 Conscientia Beam. All Rights Reserved. maturity, plant height, the number of pods per plant, the number of seeds per plant, the number of seeds per pod, 100-seed weight, biomass yield, grain yield, and harvest index. Data were computed by using SAS 9.1.3 for analysis of variance, Genstat13th for biplot graph and Agrobase20 for stability analysis. Table-1. Some characteristic features of the test sites Source: Debre Zeit Agricultural Research Center (2012) Table-2.List of Desi-type chickpea genotypes included in the experiment Entry no Entry name Entry no Entry name Entry no Entry name 1 DZ-2012-CK-0027 8 DZ-2012-CK-0027 15 Natoli 2 DZ-2012-CK-0027 9 DZ-2012-CK-0027 16 Minjar 3 DZ-2012-CK-0027 10 DZ-2012-CK-0027 17 Local check 4 DZ-2012-CK-0027 11 DZ-2012-CK-0027 5 DZ-2012-CK-0027 12 DZ-2012-CK-0027 6 DZ-2012-CK-0027 13 DZ-2012-CK-0027 7 DZ-2012-CK-0027 14 DZ-2012-CK-0027 3. RESULTS AND DISCUSSION The combined analysis of variance revealed significant (P≤0.01) differences for environments, genotypes and genotype by environment interaction in desi-type chickpea genotype. Except seed per pod, the variations were significant for all the characters studied viz. days to flowering, days to maturity, grain yield, number pod per plant, plant height, 100 seed weight , above ground dry biomass yield and harvest index when tested against pooled error variance. Yield and its components are poly genic traits and are strongly influenced by environment in chickpea. Highly significant variation was observed for grain yield in desi-type chickpea genotypes (Table 3). The mean squares due to G X E interactions were also highly significant exhibiting the differential response of genotype in various environments. Similar results were obtained by Khan, et al. [6], Khan, et al. [7] and Bains, et al. [8] in chickpea. Bartlett’s test showed that homogenous error variance for the grain yields these allowing proceeding further for pooled analysis of variance across environments. Indicating differences in environments and the presence of genetic variability among genotypes. The highly significant difference among the genotypes showed that they may carry genes with different additive effects. Variations for genotypes performances across environments were also reported by Malik, et al. [9], Bozoglu and Gulumser [10]. Current Research in Agricultural Sciences, 2015, 2(3):100-113 103 © 2015 Conscientia Beam. All Rights Reserved. Table-3. Mean sum of squares of yield and other traits from combined ANOVA of 17 desi type chickpea genotypes grown across five environments Table 4. Mean grain yield (kg/ha) 17 Desi-type chickpea genotypes grown at five environments Genotype Akaki Chefe Donsa Debre Zeit Dembia Haramaya Mean yield DZ-2012-CK-0027 2073 3966 2908 939 2273 2053 DZ-2012-CK-0028 1304 2374 2958 823 1710 1620 DZ-2012-CK-0029 1540 3575 2741 836 1523 1636 DZ-2012-CK-0030 1444 3488 3031 970 2028 1783 DZ-2012-CK-0031 1596 3501 3569 667 1502 1792 DZ-2012-CK-0032 1483 3446 2810 1394 1147 1663 DZ-2012-CK-0033 1785 2943 3398 745 1208 1784 DZ-2012-CK-0034 1556 3641 3208 833 1870 1805 DZ-2012-CK-0035 1556 2928 2997 667 1655 1686 DZ-2012-CK-0036 1821 3237 3717 1183 1991 2106 DZ-2012-CK-0037 1367 3150 3408 1117 1424 1737 DZ-2012-CK-0038 1693 3245 3367 766 1352 1774 DZ-2012-CK0039 1793 3347 3415 771 1538 1861 DZ-2012-CK-0040 2229 3622 4060 956 1641 2223 Natoli (SC) 1703 3655 3485 743 3247 2176 Minjar (SC) 1367 2945 3876 1010 1756 1875 Local check 1796 2837 2366 1329 1033 1664 Means 1653 3288 3250. 928. 1670 1838 CV(%) 16 11 12 20 20 16 LSD (5%) 91 123 133 64 115 45 The environment’s mean yield varied from 928 to 3288 kg/ha (Table 4). Maximum mean grain yield was obtained from Chefe Donsa and Debre Zeit; while, the minimum was from Dembia. Genotypes means across the environments indicated that the maximum mean grain yield was obtained from DZ-2012-CK-0039 (2223 kg/ha) while Natoli (2176 kg/ha) was ranked second and DZ-2012-CK-0036 (2106 kg/ha) third and the minimum from genotype DZ-2012-CK-0028 (1620 kg/ha). Genotypes DZ-2012-CK-0040 (2229 kg/ha), DZ-2012-CK-0027 (2073 kg/ha) and DZ-2012-CK-0036 (1821kg/ha) were ranked first, second and third at environment Akaki based on mean grain yield. At Chefe Donsa genotype DZ-2012-CK-0027 (3966 kg/ha), Natoli (3655 kg/ha) and DZ-2012-CK-0034 (3641kg/ha) ranked first, second and third respectively. At Debre Zeit genotype DZ-2012-CK-0040 (4060 kg/ha), Minjar (3876 kg/ha) DZ-2012-CK-0036 (3717 kg/ha) performed first, second and third respectively. At Dembia genotype DZ-2012-CK- 0032(1394 kg/ha), local variety (1329 kg/ha) and DZ-2012-CK-0036 (1183 kg/ha) ranked first, second and third respectively. At Dembia, the performances of the genotypes were generally poor as compared to Akaki, Chefe Donsa, Debre Zeit, and Haramaya, because at Dembia there was Current Research in Agricultural Sciences, 2015, 2(3):100-113 104 © 2015 Conscientia Beam. All Rights Reserved. moisture stress during planting and vegetative stage. At Haramaya genotypes Natoli (3247 kg/ha) DZ-2012-CK-0040 (2273) and DZ-2012-CK-0030 (2028 kg/ha) were found first, second and third respectively. The majority of the genotypes were best performing at Debre Zeit. G X E interaction causes differences in yield rank of genotype in different environments hence it is important for chickpea breeders to enhance selection efficiency and development of specifically adapted genotypes for different environments. 3.1. Performance of Desi-Type Chickpea Genotypes for Yield Related Traits The mean square values due to genotypes, environments and genotype by environment interactions from the combined ANOVA were highly significantly for the traits: days to flowering, days to maturity, plant height, pods per plant, hundred seed weight, above ground dry biomass and harvest index. However, all of these were not significant for number of seeds per pod. Days to flowering and maturity (days): Days to flowering and to maturity were significantly affected not only by genotypes but also G X E interaction, reflecting genetic variability in experimental material as well as differential response across environments (Table 3). Averaged over all genotype, early flowering and maturity were observed at Debre Zeit (42 and 102 respectively) and late flowering and maturity occurred at Akaki (52 and 130 days respectively) (Table 5). The probable reason is due to high temperature and early cessation of rain at Debre Zeit and, relatively long rain season and low temperature at Akaki. In all research sites, mean value of days to flowering and days to maturity ranged from 42 days (Debre Zeit) to 52 days (Dembia) and 102 days (Debre Zeit) to 130 days (Haramaya), respectively (Table 6). Number of pods per plant: Number of pods per plant is an important selection criterion for the development of high yielding genotypes and it is strongly influenced by environment in chickpea [9]. Marked variation for number of pods per plant was observed in the performance of genotypes over the five environments (Table 3) and this indicated sensitiveness of number of pods for environmental variations. The highest mean number of pods per plant was recorded by genotypes local variety (57) followed by DZ-2012-CK-0036 (52) and DZ-2012-CK-0035 (42) (Table 5). Number of pods per plant was highest at Haramaya (50) and lowest at Akaki (31) (Table 6).The highest pod per plant was as result of extended vegetative and reproductive growth stages. These results are consistent with the findings of Malik, et al. [9] in Kabuli type chickpeas. Plant height: Significant effects were observed not only for genotypes but also for genotype environment interaction, reflecting genetic variability in experimental material as well as differences in genotype performances over location the (Table 3). The relative performance of genotypes for plant height was markedly inconsistent over the environments which is in line with finding of Malik, et al. [9] and Iliadis [11] in chickpea who found high magnitude of G X E interaction. Plant height was highest at Debre Zeit (40 cm) and lowest at Akaki (30 cm) (Table 5).the shortest was DZ-2012-ck-0036 (32 cm) and the longest was DZ-2012-ck-0031 (39 cm) (Table 6). Current Research in Agricultural Sciences, 2015, 2(3):100-113 105 © 2015 Conscientia Beam. All Rights Reserved. 100-grain weight: Statistically, highly significant variance was observed for genotypes, genotype and environment (Table 3). In addition, the relative performance of the genotypes was quite inconsistent across the environments. Significant pooled deviation for 100-grain weight suggested that these genotypes differ considerably with respect to their suitability for this trait. The results obtained are in conformity with the findings of Singh and Singh [12] in chickpea and Sanghi and Kandalkar [13] in fodder cow pea. Hundred seed weight was highest at Akaki (25 g) and lowest at Chefe Donsa (22 g) (Table 5). Hundred seed weight varied from 13 g for farmers or local variety to 31 g for DZ-2012-CK-0031 (Table 6). Above ground dry biomass. Statistically, highly significant variance was observed for genotypes, environments and G X E interaction (Table 3). Averaged over all genotypes above ground dry biomass was highest at Debre Zeit (1469 g) and lowest at Akaki (765 g) (Table 5).On the other hand, where averaged over all environments above ground dry biomass ranged from 920 g for local check and 1320 g for Natoli (Table 6). Harvest index. Statistically, highly significant variance was observed for genotype and genotypes, environments and G X E interaction (Table 3). Over all genotypes harvest index was highest at Chefe Donsa (57%) and least at Dembia (28%) (Table 5). Pooled over all environments, harvest index ranged from 41.1% for DZ-2012-ck-0028 to 49.5% for DZ-2012-CK-0036 (Table 6). High harvest index is very important for increasing yield potential in crops because it is sensitive to environmental variations. However, a high vegetative growth is not always a symptom for high chickpea yield. 3.2. Stability Analysis 3.2.1. Wricke’s Ecovalence Analysis Wricke [14]; Wricke [15] defined the concept of ecovalence as the contribution of each genotype to the GEI sum of squares. Genotypes with the lowest eco valence contributed the least to the G X E interaction and are therefore, more stable. Accordingly, DZ-2012-CK-0035, DZ-2012-CK-0034, DZ-2012-CK-0036, DZ-2012-CK-0028, DZ-2012-CK-0039, DZ-2012-CK-0029 and DZ-2012-CK-0038 were the most stable genotypes. In that order these ranked 13th, 7th, 3th, 17th, 6th, 16th, and 11th for grain yield. Whereas most unstable genotypes were Natoli, DZ-2012-CK-0032, DZ-2012-CK-0040, DZ- 2012-CK-0027, DZ-2012-CK-0037, DZ-2012-CK-0030, DZ-2012-CK-0033, DZ-2012-CK-0031, Minjar and the local variety. These ranked 2nd 15th, 1st, 4th, 12th, 10th, 9th, 8th, 5th and 14th for grain yield, respectively. The results indicated that high yielders have higher ecovalence and vice versa, and because of this genotypes recommendation for general adaptability would be difficult (Table 7). According to Asrat, et al. [16] genotypes with high ecovalence mean and large estimated value are suitable for high input environments. Current Research in Agricultural Sciences, 2015, 2(3):100-113 106 © 2015 Conscientia Beam. All Rights Reserved. Table-5.Mean values yield related traits of desi-type chickpea genotype tested at five environments in Ethiopia Environment DF DM PPP SPP PHT HSW BM HI YLD Akaki 52 130 31 1.05 30.2 25.3 765.4 52.9 1653.3 Chefe Donsa 48 129 38 1.04 35.9 22.1 1408 57 3287.9 Debre Zeit 42 102 49 1.01 40.1 24.6 1469.1 53.3 3250 Dembia 52 109 38 1.01 37.2 22.7 1329.4 28 928.3 Haramaya 43 130 50 1.01 35.8 24 886.7 47.1 1699.5 Means 50 121 40 1.08 34.7 24.4 1043.2 44.9 1837.7 CV (%) 3.8 1.8 15.5 1.02 7.8 8.8 15.4 11.6 16.1 SE± 0.30 0.69 0.74 0.01 0.28 0.27 22.2 0.79 11.60 LSD (5%) 1.16 1.26 3.85 0.06 1.67 1.34 100.4 3.24 44.5 Where: YLD= grain yield, DF= days to flower, DM=days to mature, PPP= pod per plant, SPP=seed per pod, PHT=plant height, BM= biomass yield, HI= harvest index, HSW= hundred seed weight Table-6. Means for yield related traits of 17 Desi-type chickpea genotypes grown at five environments Genotype DF DM PPP SPP PHT HSW BM HI DZ-2012-CK-0027 48 122 38 1 36 29 1145 44 DZ-2012-CK-0028 58 119 38 1 36 21 975 41 DZ-2012-CK-0029 50 113 40 1 37 27 950 45 DZ-2012-CK-0030 46 122 29 1 35 29 970 47 DZ-2012-CK-0031 50 120 35 1 39 31 1030 45 DZ-2012-CK-0032 47 115 35 1 33 24 960 48 DZ-2012-CK-0033 50 122 40 1 35 23 1010 42 DZ-2012-CK-0034 46 120 33 1 32 28 1170 43 DZ-2012-CK-0035 49 120 41 1 33 22 945 43 DZ-2012-CK-0036 48 124 52 1 32 23 1015 50 DZ-2012-CK-0037 51 122 42 1 35 22 1095 46 DZ-2012-CK-0038 46 121 41 1 34 24 1060 45 DZ-2012-CK0039 48 118 40 1 35 24 1030 45 DZ-2012-CK-0040 49 124 41 1 34 27 1170 46 Natoli (SC) 54 123 36 1 35 28 1320 42 Minjar (SC) 48 119 38 1 35 20 970 47 Local check 47 118 57 2 34 13 920 46 Means 50 121 40 1 35 24 1043 45 CV (%) 4 2 16 10 8 9 15 12 SE± 0.3 0.7 0.7 0.01 0.3 0.3 22 1 LSD 1 1 4 0.1 2 1 100 3 Where: YLD= grain yield, DF= days to flower, DM= days to mature, PPP= pod per plant, SPP= seed per pod, PHT=plant height, BM= biomass yield, HI= harvest index, HSW=hundred seed weight; SC=standard Check. 3.3.2. Eberhart-Russell’s Joint Regression Stability Analysis Eberhart and Russell [17] joint regression model for stability of some agronomical and physiological data provide the estimate of the desired stability parameters. The GE (linear) interaction was not significant, indicating that the stability parameter ‘bi’ estimated by linear response to change in environment was the same for all genotypes or genotypes have the same slope (Table 8). This confirms that GE was not a linear function of environments indices. The variations among the genotypes and for G X E interactions were significant. It means that genotypes exhibited different performance in different environments, which is due to their different genetic makeup or the variation due to the environments or both. Current Research in Agricultural Sciences, 2015, 2(3):100-113 107 © 2015 Conscientia Beam. All Rights Reserved. Mean sum of squares due to pooled deviation from regression was significant (P≤0.01) for grain yield indicating the importance of the non- linear GE. Table-7. Wrickes ecovalence value for 17 desi-type chickpea genotypes at five environments Wi =wrickes ecovalence, SC= standared check Table-8. Analysis of variance for linear regressions of desi-type chickpea genotypes means on environmental index according to Eberhart and Russell’s joint regression model Source of variation Df SS MS Total 339 3437279 Genotype 16 166591 10412* Env + in Gen + Env 68 3270687 48098 Env. in linear 1 2859358 2859358** Gen x Env.(linear) 16 140037 8752 Pooled deviation 34 271293 5319** Residual 240 445758 1748 *, ** -significant at P ≤0.05 and 0.01; Grand mean = 441.04; R-squared = 0.91%; C.V. = 18.96% The stability parameters according to the model of Eberhart and Russell are given on Table 9. The most stable genotypes with the lowest S2di values were DZ-2012-CK-0028, DZ-2012-CK-0038, DZ-2012-CK-0039, DZ- 2012-CK-0034 in decreasing order. The most unstable genotypes with the highest S2di values were Natoli, local and DZ-2012- CK-0028. So, these genotypes would best fit for specific adaptation in favorable environments. If themean yield ( x ), regression coefficient value (bi) and the deviation from theregression (S2di) are considered together, then the most stable genotype would be DZ-2012-CK-0034 with a mean yield x = 1805 kg /ha ranked first, bi = 1.01 close to 1 and the S2di = 735 ranked fourth. On the other hand, genotypes DZ-2012-CK-0027, DZ-2012-CK-0028, DZ-2012-CK-0029, DZ-2012-CK- 0030, DZ-2012-CK-0032, DZ-2012-CK-0035 and local had regression coefficient values (bi) less than one (i.e. above average stability and significant deviation from regression). So these genotypes were specifically adapted to poor environments. Genotype DZ-2012-CK-0031, DZ- Genotypes Wi Rank Mean seed yield (Kg/ha) Rank DZ-2012-CK-0027 32724 5 2053 4 DZ-2012-CK-0028 6038 14 1620 17 DZ-2012-CK-0029 7232 12 1636 16 DZ-2012-CK-0030 13365 9 1783 10 DZ-2012-CK-0031 10795 10 1792 8 DZ-2012-CK-0032 36057 3 1663 15 DZ-2012-CK-0033 18199 7 1784 9 DZ-2012-CK-0034 3063 16 1805 7 DZ-2012-CK-0035 2338 17 1686 13 DZ-2012-CK-0036 3385 15 2106 3 DZ-2012-CK-0037 14406 8 1737 12 DZ-2012-CK-0038 8231 11 1774 11 DZ-2012-CK0039 6479 13 1861 6 DZ-2012-CK-0040 33034 4 2223 1 Natoli (SC) 110185 1 2176 2 Minjar (SC) 31917 6 1875 5 Local check 73873 2 1664 14 Current Research in Agricultural Sciences, 2015, 2(3):100-113 108 © 2015 Conscientia Beam. All Rights Reserved. 2012-CK-0033, DZ-2012-CK-0034, DZ-2012-CK-0036, DZ-2012-CK-0037, DZ-2012-CK-0038, DZ-2012-CK-0039, DZ-2012-CK-0040, Natoli, and Minjar had regression coefficients (bi) greater than one (i.e. below average stability and significant deviation from regression). Hence, these genotypes were specifically adapted to favorable environments. Similar results were obtained in common bean genotype tested [18] in different parts of Ethiopia and Ferreira, et al. [19] in Brazil. Table-9. Mean seed yield, regression coefficients (bi), coefficients of determination (r2i) and deviation from regression (S2di) Genotypes bi r2i S2di Seed Yield (kg/ha) Rank DZ-2012-CK-0027 0.74 1.01 5629 2053 4 DZ-2012-CK-0028 0.93 1.00 13 1620 17 DZ-2012-CK-0029 0.80 0.99 1535 1636 16 DZ-2012-CK-0030 0.89 1.00 2028 1783 10 DZ-2012-CK-0031 1.24 0.99 1405 1792 8 DZ-2012-CK-0032 0.68 1.01 4775 1663 15 DZ-2012-CK-0033 1.13 1.01 3318 1784 9 DZ-2012-CK-0034 1.01 0.99 735 1805 7 DZ-2012-CK-0035 0.97 0.99 1015 1686 13 DZ-2012-CK-0036 1.11 0.99 1302 2106 3 DZ-2012-CK-0037 1.06 1.00 2803 1737 12 DZ-2012-CK-0038 1.11 1.00 232 1774 11 DZ-2012-CK0039 1.11 0.99 357 1861 6 DZ-2012-CK-0040 1.31 1.01 3811 2223 1 Natoli (st.ck) 1.09 1.12 34519 2176 2 Minjar (st.ck) 1.30 1.01 3516 1875 5 local check 0.45 1.02 6418 1664 14 bi- regression coefficients, r2i-coefficients of determination, S2di-Deviation from regression. Table-10. AMMI analysis of variance for grain yield (kg/ha) of the 17 desi -type genotypes tested across five environments Source Df Sum of squares Mean of squares % Explained Total 339 22006783 64917 Environment (E) 4 17449755 4362439** 79.3% Genotype (G) 16 804949 50309** 3.7% G × E 64 1844379 28818** 8.4% IPCA1 19 779080 41004** 42.2% IPCA2 17 603319 35489** 32.7% IPCA3 15 318905 21260** 17.3% IPCA Residuals 13 143075 11006 7.8% **=significant at the P≤ 0.01 probability level 3.3.3. AMMI Analysis and Biplot Representation of 17 Desi-Type Chickpea Genotypes The main effects of E and G accounted for 79.3 % and only 3.7% respectively, and G X E interaction accounted for 8.4% of the total variation in GE data for grain yield (Table 10). The high percentage of the environment is an indication that environment is the major factor that influence yield performance of chickpea in Ethiopia. The variation due to GE is more than double than the variation due to genotypes as main effect. Tarakanovas and Ruzgus [20], reported significant G X E interaction for grain yield and stressed the usefulness of AMMI analysis for Current Research in Agricultural Sciences, 2015, 2(3):100-113 109 © 2015 Conscientia Beam. All Rights Reserved. selection of promising genotypes for specific environments or environmental conditions. The first two principal components (PC1 and PC2) which were used to create a two-dimensional biplot, explained 42.24% and 32.72% of AMMI sum of squares, respectively (Table 10). The first two IPCA captures more than 74.96% of the total interaction main effect. The large sum of squares for environments showed that the environments were diverse, with large differences among environmental means causing most of the variation in grain yield, which is in synchronization with the findings of Yan and Tinker [21] in chickpea. This result also indicated the great influence the test environments have had on the yield performance of Desi-type genotypes in Ethiopia. The AMMI I biplot for grain yield of 17 Desi-type genotypes at five environmental conditions is presented on Fig. 1. The main effects (genotypes and environments) accounted for 82.99 %, and IPCA 1 accounted for 42.24% and IPCA2 accounted for 32.71% of the total variation in genotype by environment interaction. The three IPCAs together accounted for 92.24% of the total interaction, the remaining 7.76% being the residual or noise, which is not interpretable and thus discarded. Environments showed high variation in both main effects and interactions (IPCA1) (Fig. 1).Chefe Donsa and Debre Zeit were most favorable environments; Haramaya and Dembia were least favorable environments as these two environments are far from the origin, while Akaki is the average environment. Environments are also classified into four main groups based on their IPCA1 scores, those of Haramaya and Chefe Donsa in quadrant I have large positive IPCA1 scores, which interact positively with genotypes that have positive IPCA1 scores and negatively those genotypes with negative IPCA1 scores. Debre Zeit in quadrant III has large negative IPCA1 scores, which interact positively with genotypes having negative IPCA1 scores and negatively with genotypes that have positive IPCA1 scores. Akaki and Dembia in quadrant IV have large negative IPCA1 scores, which interact positively with genotypes having negative IPCA1 scores and negatively with genotypes that have positive IPCA1 scores. The environments can be sub-grouped according to their average yield over the genotypes. The biplot analysis revealed that genotypes DZ-2012-CK-0029, DZ-2012-CK-0031, DZ- 2012-CK-0036, and Minjar exhibited IPCA scores close to zero and high mean yield; thus they are found to be stable. Variety Natoli and local are far from the origin, were sensitive to environmental interactive force, and are considered unstable. Genotypes DZ-2012-CK-0029, Minjar and DZ-2012-CK-0031 are close to the origin, were non-sensitive to environmental intercations and are stable. The biplot also revealed association between environment and genotypes. The local variety, DZ-2012-CK-0032, DZ-2012-CK-0033 is related to environment Akaki and Natoli are not associated to any particular environment. DZ-2012-CK-0027 and DZ- 2012-CK-0036 are related to Chefe Donsa and DZ-2012-CK-0040 is related to Debre Zeit. Genotypes DZ-2012-CK-0031was high yielder since its mean was greater than the grand mean; it had positive IPCA scores, thus adapted to Debre Zeit. Genotypes with negative IPCA score were not adapted to any of the environments. Among environments, Chefe Donsa and Haramaya are favorable environments as they exhibited the highest positive IPCA1 score, while Debr zeit, Akaki and Dembia are marginal environments showing the highest negative IPCA scores. Current Research in Agricultural Sciences, 2015, 2(3):100-113 110 © 2015 Conscientia Beam. All Rights Reserved. Genotypes DZ-2012-CK-0040, Natoli, DZ-2012-CK-0036 and DZ-2012-CK-0027 have higher average yields and adapted to favorable environments, while genotypes Minjar, DZ-2012- CK-0039 and DZ-2012-CK-0034 were adapted to poor environments. DZ-2012-CK-0027, DZ- 2012-CK-0030, DZ-2012-CK-0034, DZ-2012-CK-0036, Minjar were adapted to Chefe Donsa while DZ-2012-CK-0032, DZ-2012-CK-0039, and DZ-2012-CK-0040 were adapted to Debrezeit. Genotypes DZ-2012-CK-0037, DZ-2012-CK-0038, DZ-2012-CK-0032, DZ-2012-CK-0033 and local variety were adapted to Akaki. Genotypes DZ-2012-CK-0040, Natoli, DZ-2012-CK-0036, and DZ-2012-CK-0027 had higher average mean grain yield and had highest positive and negative IPCA1 scores which makes them unstable genotypes. Genotype DZ-2012-CK-0032and DZ-2012-CK-0033 had low yield and large IPCA1 scores which indicates that they unstable. Fig-1. AMMI biplot of IPCA1 vs. Main effects using yield data for Desi-type chickpea genotypes. Akk=akaki, dzt= Debre Zeit, hun= haramaya, dem= dembia and chf= Chefe Donsa; G1-G17 graph ID for the genotypes Table-11. Yield (kg/ha) and parametric stability statistics for grain yield on 17 genotypes grown in five environments Genotype IPCA1 IPCA2 ASV Rank Yield DZ-2012-CK-0027 4.81 -8.06 9.74 14 2053 DZ-2012-CK-0028 1.61 2.30 2.93 3 1620 DZ-2012-CK-0029 0.02 -6.21 6.21 10 1636 DZ-2012-CK-0030 3.91 -3.39 5.59 9 1783 DZ-2012-CK-0031 0.66 3.23 3.32 4 1792 DZ-2012-CK-0032 6.55 -6.32 9.76 15 1663 DZ-2012-CK-0033 -4.40 3.99 6.39 10 1784 DZ-2012-CK-0034 2.69 -1.89 3.59 5 1805 DZ-2012-CK-0035 1.44 0.13 1.64 1 1686 DZ-2012-CK-0036 0.17 3.80 3.81 7 2106 DZ-2012-CK-0037 -2.99 2.34 4.12 8 1737 DZ-2012-CK-0038 -2.65 2.12 3.68 6 1774 DZ-2012-CK0039 -1.26 1.82 2.31 2 1861 DZ-2012-CK-0040 -3.35 5.88 7.00 12 2223 Natoli (SC) 15.29 0.01 17.37 17 2176 Minjar(SC) 0.00 8.34 8.34 13 1875 Local check -8.07 -8.10 12.23 16 1664 SC=Standard Check Current Research in Agricultural Sciences, 2015, 2(3):100-113 111 © 2015 Conscientia Beam. All Rights Reserved. 3.3.4. AMMI Stability Value (ASV) According to the ASV ranking, the most stable genotypes were DZ-2012-CK-0035, DZ- 2012-CK-0039 and DZ-2012-CK-0028. DZ-2012-CK-0040 and Minjar were the first and second highest yielders based on the mean yield values (Table 11). However, DZ-2012-CK-0040 which was the highest for mean yield, ranked 12th for the ASV. The most unstable genotypes were Natoli, local, DZ-2012-CK-0032 and DZ-2012-CK-0027. These results are similar with most of the stability indices estimation. ASV was used similarly to identify the stability of common bean varieties in eastern Ethiopia by Nigussie [22]. 3.4. Cluster Analysis of Genotypes Cluster analysis was performed to study the patterns of groupings of genotypes. Dendrograms (Fig. 2) were generated from SAS clustering method of hierarchical algorism method based on Euclidean distances using AMMI adjusted mean yields of genotypes. Clustering of genotypes at a cut-off value of zero produced five clusters. Cluster one consisted of seven genotypes (DZ-2012-CK-0030, DZ-2012-CK-0031, DZ-2012-CK-0033, DZ-2012-CK-0037, DZ- 2012-CK-0038, DZ-2012-CK-0039 and Minjar). Cluster two consisted of five genotypes (DZ- 2012-CK-0028, DZ-2012-CK-0029, DZ-2012-CK-0032, DZ-2012-CK-0035 and local). Cluster three consists of two genotype (DZ-2012-CK-0027 and DZ-2012-CK-0036) and in this group both genotypes were characterized as high yielder and clearly shown in the AMMI biplot. Cluster four consisted of two genotypes (DZ-2012-CK-0040 and Natoli) which were characterized as high yielders. The last clustering group consists of only one genotype which is DZ-2012-CK-0034. Fig-2. Dendrogram depicting the clustering of 17 Desi-type chickpea genotypes Note; DZ-2012-CK-0027=1, DZ-2012-CK-0028=2, DZ-2012-CK-0029=3, DZ-2012-CK-0030=4, DZ-2012-CK-0031=5, DZ-2012-CK- 0032=6 DZ-2012-CK-0033=7, DZ-2012-CK-0034=8, DZ-2012-CK-0035=9, DZ-2012-CK-0036=10, DZ-2012-CK-0037=11,DZ-2012-CK- 0038=12 DZ-2012-CK-0039=13, DZ-2012-CK-0040=14,natoli=15,Minjar=16 and local variety=17. 4. CONCLUSIONS The combined analysis of variance revealed significant (P≤0.01) differences for environments, genotypes and genotype by environment interaction in Desi-type chickpea genotype. Except seed per pod, the variations were significant for all the characters studied In this study, univariate and multivariate stability parameters were used to select the most stable chickpea (Desi-type) genotypes for Ethiopia. The univariate stability parameters showed inconsistence to recommend the most stable genotypes. However, multivariate method, AMMI Current Research in Agricultural Sciences, 2015, 2(3):100-113 112 © 2015 Conscientia Beam. All Rights Reserved. model could be used to identify superior genotypes for specific adaptation and the AMMI model is one of most important package to investigate patterns, relationships and for predicting performance of genotype and environments. In this experiment Dembia and Haramaya are the most discriminative environments for Desi type chickpea genotypes and Debre Zeit and Chefe Donsa was the favorable environment and Akaki was moderately favorable environment for Desi- type chickpea genotype. Genotypes, DZ-2012-CK-0040, DZ-2012-CK-0036 and DZ-2012-CK- 0040, DZ-2012-CK-0032 and variety Natoli were recommended as specifically adapted to sites Akaki, Chefe Donsa, Debre Zeit, Dembia and Haramaya respectively for Desi-type chickpea. As the experiment was tested on five environments, it is imperative to use the recommended genotype for the given environment. 5. ACKNOWLEDGMENTS We would like to thank express our appreciation to TL-I Project for its financial support. We are also indebted for the assistances provided by the research and technical staff of the highland pulse improvement program of Debre Zeit Agricultural Research Center (DZARC) and Gonder Agricultural Research Center (GARC). REFERENCES [1] Q. Ali, H. A. Sadaqat, S. Arshad, J. Farooq, M. Ahsan, M. Waseem, and A. Iqbal, "Genetic variability and correlation analysis for quantitative traits in chickpea genotypes (Cicer Arietinum L)," J. Bacteri. Res., vol. 3, pp. 6-9, 2011. [2] CSA, "Central statistical agency area, production and productivity of agriculture. Addis Ababa, Ethiopia," vol. 1, p. 532, 2012. [3] P. M. Gaur, S. Tripathi, C. L. L. Gowda, G. V. Ranga Rao, H. C. Sharma, M. S. Pande, and Sharma, Chickpea seed production manual. India: Patancheru, Andhra Pradesh, ICRISAT, 2010. [4] K. Menale, S. Bekele, A. Solomon, A. Tsedeke, M. Geoffrey, F. Setotaw, E. Million, and A. Kebebew, "Current situation and future outlooks of the chickpea sub-sector in Ethiopia. ICRISAT and EIAR," 2009. [5] J. K. Bull, M. Cooper, I. H. Delacy, K. E. Basford, and D. R. Woodruff, "Utility of repeated checks for hierarchical classification of data from plant breeding trials," Field Crop. Res., vol. 30, pp. 79-95, 1992. [6] I. A. Khan, B. A. Malik, and M. Tahir, "Phenotypic stability for yield in chickpea," Pak. J. Sci., vol. 30, pp. 455-456, 1987. [7] I. A. Khan, B. A. Malik, and M. Bashir, "Investigation of genotype x environment interaction for seed yield in chickpea (Cicer Arietinum L)," Pak J. Bot., vol. 20, pp. 201-204, 1988. [8] M. S. Bains, S. M. Sharma, S. K. Rao, and S. P. Singh, "Genotype x environment interaction for three selection methods in segregating populations of chickpea," Legume Res., vol. 11, pp. 117-122, 1988. [9] B. A. Malik, S. A. Hussain, and M. A. Zahid, "Grain legume status in agriculture," Progressive Farming, vol. 3, pp. 23-26, 1988. Current Research in Agricultural Sciences, 2015, 2(3):100-113 113 © 2015 Conscientia Beam. All Rights Reserved. [10] H. Bozoglu and A. Gulumser, "Determination of genotype by environment interaction of some agronomic characteristics in dry bean (Phaseolus Vulgaris L)," Turk. J. Agric., vol. 24, pp. 211-222, 2000. [11] C. Iliadis, "Evaluation of six chickpea varieties for seed yield under autumn and spring sowing," J. Agric. Sci. (Cambridge), vol. 137, pp. 439-444, 2001. [12] V. Singh and F. Singh, "Genetic diversity and stability in chickpea," Indian J. Genet. Pl. Breed., vol. 49, pp. 349-353, 1974. [13] A. K. Sanghi and V. S. Kandalkar, "Phenotypic stability of yield and its components in fodder cowpea," Indian J. Genet., vol. 43, pp. 164-167, 2001. [14] G. Wricke, "Ber eine methode zur erfassung der ökologischen streubreite in feldversuchen," Z. Pflanzenzüchtg, vol. 47, pp. 92-96, 1962. [15] G. Wricke, "Zur berechnung der ıkovalenz bei sommerweizen und hafer," Z. Pflanzenzuchtg, vol. 52, pp. 127-138, 1964. [16] A. Asrat, T. Assefa, A. Birhanu, K. Negash, and F. Fisum Alemayehu, "Adaptation and yield stability of small red beans elite lines in Ethiopia," Inter. J. of Pl. Breed. and Genet., vol. 2, pp. 51-63, 2008. [17] S. A. Eberhart and W. A. Russell, "Stability parameters for comparing varieties," Crop. Sci., vol. 6, pp. 36-40, 1966. [18] M. Firew, "Yield stability in common bean (Phaseolus Vulgaris) genotypes," Euphytica, vol. 130, pp. 147-153, 2003. [19] D. F. Ferreira, C. G. B. Demetrio, B. F. J. Manly, A. A. Machado, and R. Vencovsky, "Statistical model in agriculture: Biometrical methods for evaluating phenotypic stability in plant breeding," Cerne Lavras, vol. 12, pp. 373-388, 2006. [20] P. Tarakanovas and V. Ruzgus, "Additive main effect and multiplicative interaction analysis of grain yield of wheat varieties in Lithuania," Agron. Res., vol. 4, pp. 91 - 98, 2006. [21] W. Yan and N. A. Tinker, "Biplot analysis of multi-environment trial data: Principles and applications," Can. J. Pl. Sci., vol. 86, pp. 623-645, 2006. [22] K. Nigussie, "Genotype x environment interaction of released common bean (Phaseolus Vulgaris L) varieties in Eastern Amhara," M.Sc. Thesis, Presented to the School of Graduate Studies of Haramaya University, Ethiopia, 2012. BIBLIOGRAPHY [1] B. Geletu and E. Million. Chickpea in Ethiopia. In: Adaptation of Chickpea in the West Asia and North Africa Region. Saxena, N.P., Saxena, M.C., Johansen, C., Virmani, S.M. and Harris, H., Ed. India: ICRISAT, 1996. [2] M. J. Pinthus, "Estimate of genotypic value: A proposed method," Euphytica, vol. 22, pp. 121-123, 1973. Views and opinions expressed in this article are the views and opinions of the author(s), Current Research in Agricultural Sciences Research shall not be responsible or answerable for any loss, damage or liability etc. caused in relation to/arising out of the use of the content.