Impaginato 517 Adv. Hort. Sci., 2018 32(4): 517-524 DOi: 10.13128/ahs-20689 Stability analysis of fruit yield of some olive cultivars in semi-arid environmental condition I. Arji (*) Crop and Horticultural Science Research Department, Kermanshah Agricultural and Natural Resources Research and Education Center, AREEO, Kermanshah, Iran. Key words: adaptability, AMMi, Olea europaea L., olive, stability parameters, yield. Abstract: This study was conducted to evaluate yield stability of 12 Iranian and foreign olive cultivars in Dalaho Olive Research Station during 2006-2008. According to the variance analysis, significant variation (p<0.01) was observed between cultivars and years. Classification based on Duncan (p<0.05) showed that Konservolia was superior variety and Sevillano, Koroneiki and Zard were placed in the second group. Cultivars were divided into 3 groups based on clus- ter analysis using Ward method. The first principal component of the interac- tion between olive cultivars and the year’s show 69.25% of the variance and was statistically significant at 1% level based on AMMI analysis. According to regression coefficient (bi) deviation from regression (S2di), Wricke’s ecovalence (Wi), coefficient of determination (Ri 2) and Shukla’s stability variance (δi 2) methods ‘Mission’ and ‘Zard’ had the higher stability. According to the AMMI stability (ASV) ranking, the following cultivars were the most stable, Mission, Amigdalolia and Koroneiki, while the most unstable were ‘Konservolia’, ‘Sevillano’, ‘Roghani’, ‘Arbequina’ and ‘Abou-Satal’. ‘Konservolia’ even showed the lowest stability but its stability in all parameters was significant different in terms of performance. Generally ‘Konservolia’, ‘Sevillano’, ‘Koroneiki’ and ‘Zard’ were appropriate for fruit yield and will be introduced for breeding pro- grams in semi-warm climate. 1. Introduction O l i v e ( O l e a e u r o p a e a L . ) t r e e i s a n e v e r g r e e n n a t i v e t o t h e Mediterranean region. Some olive wild genotypes are present in different region of iran like Kermanshah province in the west of iran. There are m o r e t h a n 4 0 n a t i v e o l i v e g e n o t y p e s i n s u b t r o p i c a l r e g i o n s o f Kermanshah province like Sarpool-e-Zahab, Gilan-e-Gharb and Paveh. Marone and Fiorino (2012) reported that olive (Olea europaea L.) distrib- uted across three continents from South Africa to the central part of the Africa and Horn Africa, from Egypt and red Sea to the Mediterranean areas and Asia from Palestine, Syria, Mesopotamia and western and east- ern areas of Himalaya Chain to the Southwestern of China. This report (*) Corresponding author: issaarji@gmail.com Citation: Arji i., 2018 - Stability analysis of fruit yield of some olive cultivars in semi-arid environmental condition. - Adv. Hort. Sci., 32(4): 517-524 Copyright: © 2018 Arji i. This is an open access, peer reviewed article published by Firenze University Press (http://www.fupress.net/index.php/ahs/) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting information files. Competing Interests: The authors declare no competing interests. received for publication 24 May 2017 Accepted for publication 19 june 2018 AHS Advances in Horticultural Science Adv. Hort. Sci., 2018 32(4): 517-524 518 revealed that there are some olive genotypes in three continents. in recent years, due to higher olive oil demand, the cultivation of olive has been expand- ed in various regions of iran. However, the cultivation of olive tree is limited because of harsh environmen- tal conditions and water scarcity in most of the new olive plantation areas (Arji and Arzani, 2008). The limitation of water as well as long hot summers in the regions lead to poor fruit and oil quality (Saadati et al., 2013; Khaleghi et al., 2015). Cheng et al. (2017) stated that low temperatures would be improved olive oil quality by increasing unsaturated fatty acid amounts in the fruit. Temime et al. (2006) reported that more unsaturated fatty acid of Chetoui olive variety was recorded in cooler regions than dry and warm regions. Despite of good vegetative growth, some of the olive varieties do not show good perfor- mance as production in warm regions. This is due to lack of adapted and stable cultivars in such environ- mental conditions. Check-adapted varieties and opti- mal stability are essential for the fruit yield. it is assumed that the stability of a genotype is very important over time in each region (Finlay and Wilkinson, 1963). Homeostatic and agronomic are two genotypic stabilities. in homeostatic stability a certain genotype shows constant response under different conditions. But in agronomic stability, genotype yield is linked to p r o d u c t i v i t y p o t e n t i a l ( H a y w a r d e t a l . , 1 9 9 3 ) . Generally, the stability is defined as the actual perfor- mance of a genotype under changing environmental conditions. reliable stability of production efficiency under environment changing is very important (Kan et al., 2010). Stability analysis methods are catego- rized in two parametric and non-parametric groups (Sabaghnia et al., 2006). Several methods such as regression coefficient (Finlay and Wilkinson, 1963), sum of squared deviations from regression (Eberhart and russel, 1966), stability variance (Shukla, 1972) and additive main effects and multiplicative interac- tion (AMMi) (Gauch and Zobel, 1988) have been commonly used to parametric stability analysis. Environmental sustainability of individual geno- types can also be estimated by regression analysis and cultivar will be stable when the deviation of regression was zero or at least (Hayward et al., 1993). it is mentioned that regression analysis in bilinear models and analysis of variance in biadditive models have limitations in genotype and environment inter- action. This restriction reduced by multiplicative components for interactions in generalized linear models (GLM) such as additive effects and multiplica- tive interaction (AMMi) (Gauch, 1992). in this model the main additive effects was calculated by variance analysis and then genotypes and environment inter- action, which is known as multiplicative interaction, are analyzed by principal components analysis (romagosa and Fox, 1993). Olive is one of the fruit trees with alternate bear- ing tendency in which it not bear regularly (Lavee, 2007). This phenomenon is affected by different fac- t o r s l i k e g e n e t i c a n d p h y s i o l o g i c a l t r a i t s (Goldschmidt, 2005). The degree of alternate bearing in olive is highly dependent on environmental condi- tions (Lavee, 2007). Fruit production in olive is more irregular by climate change where adverse environ- mental conditions are frequent (Lodolini and Neri, 2012). For this purposes stability of olive production is very important in new olive growing region like Sarpool-e-Zehab environmental conditions. AMMi analysis was used to evaluate the stability of differ- ent crops (Esmaeilzadeh-Moghaddam et al., 2011), but there is lack of research in horticultural crops. Weather conditions are variable during different years in new olive cultivation regions so that we need to find out more stable olive cultivars. in the present work, the year was considered as environmental vari- able. Generally, the main goal of this study was the evaluation of yield stability of different olive cultivars in warm condition of Kermanshah province. 2. Materials and Methods Material, site characterization and experimental design This experiment was conducted in Dalahv Olive research Station of Sarpool-e-Zahab (longitude: 45° 51´ E, latitude: 34° 30´ N, altitude: 570 m asl) to verify the yield stability of 12 iranian and foreign olive culti- vars (Table 1). Two years old self-rooting plantlets were planted in the year 2000, with 6x6 m spacing Table 1 - Name and codes of genotypes Genotype Name 1 Amphisis 2 Konservolia 3 Zard 4 Amigdalolia 5 Koroneiki 6 roghani 7 Manzanillo 8 Abou-Satal 9 Mission 10 Arbequina 11 Sevillano 12 Shenge Arji - Olive fruit stability analysis 519 distance in a randomized complete block design with three replications. Each experimental unit consisted of 5 trees so that 15 trees of each cultivar were evalu- ated. Trees were pruned as vase shape and irrigated each three days with drip irrigation system. Climate of Sarpool-e-Zahab is warm with relatively low humidity during summer as shown in figure 1. Also soil and water analysis were reported (Tables 2 and 3). Data analysis methods Fruit yield was measured during 5 years from 2004 to 2008. As fruit yield was low in the years 2004 and 2005, therefore 3 years (2006, 2007 and 2008) were analyzed to determine yield stability. SPSS, irri- STAT and Excel were used for statistical analysis and the mean comparison was done by Duncan’s multiple range test at p<0.05. The model of AMMi analysis is presented in equation (1). Y ger = μ + α n + β e +Σ n λ n α gn γ en + ρ ge + ε ger (1) Where αn is the main effect of genotype; βe is the main effect of environment; n is the number of main components in AMMi model; λn is a single value related to the n remained main components in the model; αgn is thespecific vector for the g genotype from n main component; γen is the specific vector for the e environment from n main components; ρ ge is the noise and ger is the error (Clay et al., 1995). The following parameters were calculated to ana- lyze yield stability, coefficient of variability (CVi) (Francis and Kannenberg, 1978), Wricke’s (1962) eco- valance (Wi2), Shukla’s (1972) stability variance (σi2), Pinthus’s (1973) coefficients of determination (r2), and Finlay and Wilkinson (1963) regression coeffi- cient (bi). Alternate bearing index (ABi) was calculated dur- ing three successive years from 2006 till 2008, using t h e f o l l o w i n g e q u a t i o n ( 2 ) ( M o n s e l i s e a n d Goldschmidt, 1982): (2) Where n = number of years, and a1, a2... an = yields in the corresponding years. 3. Results and Discussion Fruit yield analysis of variance The results of variance analysis for yield of olive (kg/tree) show that the genotype, environment ( yea r) a n d i n t era c t i o n ef f ec t s w ere s i gn i f i c a n t (p<0.01) (Table 4). Specific response of the cultivars to ecological factors over a 3-year period were con- firmed by the results of Duncan Multiple range-Test, which proved that cultivar and year interaction effect was also significant (Table 5). it is evident from data in Table 5, for 3 study years, ‘Konservolia’ had the- highest mean yield, 24.69 kg/tree, while ‘roghani’ had the lowest mean yield, 4.87 kg/tree. Fruit yield variability was depending on the year but olive vari- eties show different responses (Table 5). So, this indi- cates that the genotypes present different behavior in that environment. This may be due to differences in genetic basis of cultivars (rakonjac and Živanovic, Fig. 1 - Precipitation, mean temperature and relative humidity during five years of experiment. Table 2 - Physical and chemical soil characteristics Soil depth (cm) Particle-size distribution (%) OC (%) pH TNV (%) Ava. K (mg/kg) Ava. P (mg/kg) Total N (%)Clay Silt Sand 0-30 34 52 14 2.25 7.7 41 520 6.2 0.18 31-60 40 37 23 0.78 7.7 45 275 2.6 0.06 Table 3 - irrigation water chemical characteristics EC (dS/m) TDS (mg/l) pH Meq/L S.S.P (%) S.A.r CO 3 -2 CO 3 H Cl- SO4 -2 Sum Anions Ca2++ Mg2+ Na+ Sum Cations 550 352 7.28 0 4.6 0.3 1.9 6.8 6.6 0.2 6.8 2.94 0.11 Table 4 - Analysis of variance for olive fruit yield S.O.S df SS MS replication 2 2083 0.115 NS Cultivar 11 3026.53 275.14 ** Error 22 199.59 9072 Year 4 1023.13 511.567 ** Cultivar x year 44 1423.3 64.696 ** Error 96 285115 5.94 CV% 19.18% Adv. Hort. Sci., 2018 32(4): 517-524 520 2008). Olive varieties with yield stability are impor- tant for sustainable production. Stable cultivars have high yield with lower variation during the years. B a s e d o n t h e r e s u l t s , ‘ K o n s e r v o l i a ’ , ‘ Z a r d ’ , ‘Koroneiki’, ‘Amigdalolia’, ‘Arbequina’ and ‘Sevillano’ have higher fruit yield with moderate yield fluctua- tion during the years. Analysis of variance is only able to express the presence or absence of interaction and is not possi- ble to interpret yield stability. For this reason, using univariate and multivariate nonparametric interpret better interaction of cultivars and years in the sus- tainability debate (Gauch, 1992; Falconer and McKay, 1996; Arciniegas-Alarcon et al., 2011; Gauch, 2013). AMMI analysis The ANOVA for fruit yield using the AMMi method is presented in Table 6. There were significant differ- ences among the genotypes, environments (Years) and G × E interaction. in this experiment environ- ments were the years based on Citadin et al. (2013) method. Combined analysis of variance (ANOVA) for fruit yield of olive cultivars indicated that genotypes, year and genotype-by-year interactions (GEi) were the most important source of fruit yield variation (Table 6). The contribution of variation caused by the cultivar, year and GEi were 52.56%, 17.77%, and 24.72%, respectively. This result showed that olive cultivars had different yield performance across years. The high share of interaction in the total sum of squares is very important to use stability analysis for fruit yield of olive varieties. Similar results were reported in yellow passion fruit by Oliveira et al. (2014) and peanuts (Oliveira and Godoy, 2006). Maulión et al. (2014) stated that the significance of the environmental effect and GEi were used as a starting point to study yield stability among peach accessions. AMMi analysis indicated that two first iPCA were significant (P<0.01). The iPCA1 accounted for 69.25% of the GE interaction (Table 6). However, based on these results most information can be graphically dis- played using iPCA1. Biplot graph of the model (iPCA1 vs. yield) is presented in figure 2. According to figure 2, ‘Zard’ and ‘Mission’ showed greater yield stability by values near the origin of the iPCA1 axis. However, mean yield of ‘Mission’ was lower than total mean yield. ‘Konservolia’ with highest fruit yield and ‘roghani’ with the lowest fruit yield were unstable cultivars and the others were in the intermediate sta- bility. One of the most important parameter in olive sta- bility is alternate bearing. This index seems to be use- ful in determining the sustainability of production in f r u i t t r e e s . B a s e d o n b i p l o t A M M i 1 a n a l y s i s , ‘Konservolia’ was more productive (Fig. 2) in all years than the others and its alternate bearing index was low (Table 5). So it is recommended to use this para- meter in stability evaluation. in this experiment, vari- ability due to the year was greater than variability Table 5 - Fruit yield (Kg tree-1), mean yield (Kg tree-1) and Alternate Bearing index of olive cultivars during 2006- 2008 Cultivar 2006 2007 2008 Mean Alternate bearing index Amphissis 7.03 hij 3.5 j 16.37 cdef 8.97 efg 0.16 Konservolia 26.03 b 14.57 c-g 33.47 a 24.69 a 0.06 Zard 17.37 cd 8.23 hij 16.67 cde 14.09 cd 0.01 Amigdalolia 15.39 cdef 6.533 ij 10.43 fghi 10.78 def 0.1 Koroneiki 23.8 b 13.33 d-h 16.7 cde 17.94 bc 0.08 roghani 7.1 hij 4.5 ij 3 j 4.87 g 0.21 Manzanillo 24.53 b 7.86 hij 6.27 ij 12.89 de 0.31 Abou-Satl 7.78 hij 7.17 hij 9.1 ghij 8.02 fg 0.04 Mission 14.53 c-g 8.07 hij 14.33 c-g 12.31 def 0.003 Arbequina 7.36 hij 10.18 fghi 16.08 cdef 11.21 def 0.19 Sevillano 25.13 b 10.8 e-i 20.48 bc 18.81 b 0.04 Shenge 10.34 fghi 6.4 ij 7.13 hij 7.96 fg 0.09 Mean 15.53 8.43 14.17 12.71 Table 6 - Analysis of variance for fruit yield of 12 olive cultivars by AMMi during 2006 -2008 S. O. V df SS SS% MS Genotypes 11 3026.52 52.56 275.14 ** Year 2 1023135 17.77 511.57 ** Cultivar x year 22 1423305 24.72 64.7 ** iPC1 12 985614 69.25 82.13 ** Noise 10 437688 30.75 43.77 NS Error 96 285115 4.95 2.97 Total 35 5758075 Fig. 2 - Biplot AMMi1 (means vs PC1) for the data on the yield of olive (Ton ha-1) with 12 cultivars () and five years (Δ). Arji - Olive fruit stability analysis 521 caused by varietal effects based on scattered effect (Fig. 2). AMMi analysis method is highlighted to study G x E i n t era ct i o n wh i ch co mb i n es a u n i va ri a t e method for the additive effects of genotypes and years with a method for the multiplicative effects of the G x E interaction (Zobel et al., 1988; Citadin et al., 2013). Gauch and Zobel (1996) stated that this method can contribute to the identification of widely adapted genotypes with high yields, as to the agro- nomic zoning for regional cultivar recommendation. A genotype will be ideal with high yields and iPCA1 values near zero. in general, according to the results of AMMi analysis Zard was the most stable cultivar w i t h h i g h y i e l d a n d i P C A 1 v a l u e s n e a r z e r o . ‘Konservolia’ and ‘Sevillano’ had high yield but higher iPC1 values than zero, therefore we recommend them as superior cultivars for pickling purpose. Ferreira et al. (2006) reported that an undesirable genotype has low stability as well as low yields. Cluster analysis According to the obtained dendrogram from clus- ter analysis using Ward method, genotypes were divided in three groups (Fig. 3). This result is con- firmed by Biplot AMMi1 (Fig. 2). Stability analysis results Eberhart and russell’s (1966) stated that a stable cultivar is considered to be the one that has regres- sion coefficient approximating 1.0 and standard error of regression as low as possible. According to this model a genotype with the higher mean fruit yield has general adaptability. in the present research, regression coefficients ranged from 0.02 to 2.11 for fruit yield (Table 7). This variation in regression coef- ficients indicates that cultivars had different respons- es to year’s fluctuations. A genotype would be adapt- ed to favorable conditions when regression coeffi- cient is higher than one and other would be adapted to unfavorable conditions when regression coeffi- cient is less than one. A genotype with regression coefficient equal to one would have an average adap- tation to all environments. According to Table 7, ‘Amphissis’, ‘Mission’ and ‘Amigdalolia’ with regression coefficients near to one are most stable all the years. ‘Koroneiki’, ‘Zard’, ‘Manzanillo’, ‘Sevillano’ and ‘Konservolia’ with regression coefficients higher than one were stable (Table 7, Fig. 2), while other cultivars like Abequina, Table 7 - Mean yields (kg/tree) and various stability measurements and their ranking orders of 12 olive cultivars evaluated during five years 2006-2008 Cultivar Fruit yield (Kg/tree) rank bi rank S2di rank Wi rank δi 2 rank CVi rank ri 2 rank ASV rank Amphissis 8.97 9 1.01 6 59.32 11 59.32 9 33.44 9 74.14 11 0.33 3 0.291 4 Konservolia 24.69 1 2.11 12 54.93 10 89.87 11 51.77 11 38.57 6 0.7 7 1951 12 Zard 14.09 4 1.34 9 0.66 2 3.94 2 0.21 1 36.08 5 0.99 10 0.546 6 Amigdalolia 10.78 8 1.08 7 6.27 6 6.44 3 1.71 2 41.14 9 0.84 9 0.032 2 Koroneiki 17.94 3 1.21 8 15.34 8 16.62 5 7.81 5 29.78 4 0.73 8 0.184 3 roghani 4.87 12 0.18 3 7.67 7 26.73 8 13.88 8 42.63 10 0.11 2 1345 10 Manzanillo 12.89 5 1.57 10 134.4 12 143.68 12 84.05 12 78.49 12 0.34 4 0.5 5 Abou-Satal 8.02 10 0.16 2 1.22 3 21.27 6 10.61 6 12.33 1 0.37 5 1239 8 Mission 12.31 6 0.96 5 0.65 1 0.69 1 1.74 3 29.86 3 0.98 11 0.027 1 Arbequina 11.21 7 0.02 1 39.64 9 66.81 10 37.93 10 39.73 8 0.0004 1 1265 9 Sevillano 18.81 2 1.92 11 2.13 4 26.21 7 13.57 7 38.88 7 0.98 11 1349 11 Shenge 7.96 11 0.43 4 3.57 5 12.85 4 5.56 4 26.36 2 0.59 6 0.937 7 bi = Finlay and Wilkinson’s (1963) regression coefficient; Sdi2 = Eberhart and russell’s (1966) deviation from regression parameter; Wi = Wricke’s (1962) ecovalence; δi2 = Shukla’s (1972) stability variance; CV% = Francis and Kannenberg’s (1978) Coefficient of variability; ri2= Coefficient of determination; ASV = AMMi stability value Fig. 3 - Dendrogram from cluster analysis based on Ward method. 522 Adv. Hort. Sci., 2018 32(4): 517-524 Shenge, Abou-Satal and roghani with regression coefficients less than one were unstable (Fig. 2). ‘Konservolia’ (bi=2.11) was productive during 2006 and 2008 than the others. High yielding varieties were not found stable with regression coefficients (bi). Similar results were found by Maulión et al. (2014) in peach stability evaluation. As olives have alternate bearing, ‘Konservolia’ had the highest fruit yield in non-bearing year (2007) in compare to the others (Table 5). The most stable cultivars with the lowest S2di val- ues were Mission and Zard. The most unstable culti- varswith the highest S2di values were Manzanillo, Amphissis and Konservolia. According to the Eberhart and russell’s (1966) model, regression coefficients (bi) approximating 1.0 coupled with S2di of zero indi- cate an average stability. ‘Mission’ and ‘Zard’ with regression coefficients near to 1 and S2di near to zero were most stable than the others. Zard cultivar had higher mean yield so it has general adaptability all the years. Concept of ecovalence was defined by Wricke (1962), where the genotypes with low eco valence have smaller fluctuations across environments and therefore are stable. The most stable cultivars according to the ecovalence method of Wricke (1962) were Mission and Zard. These cultivars were in the ranked 6 and 4 for mean yield, respectively. The most unstable cultivars according the eco valence method were Manzanillo and Konservolia with the mean yield rank of 5 and 1 respectively (Table 7). This method would not be suitable to select high-yielding cultivars but it is useful to select cultivars with the same yield of the mean yield (Table 5). For this reason, genotypes with a low Wi value have smaller deviations from the mean across years and are thus more stable. Shukla’s (1972) stability variance (δi2) revealed that ‘Zard’, ‘Amigdalolia’ and ‘Mission’ had the small- est variance across the years and were stable, while Manzanillo and Konservolia cultivars had the largest δi2 and were unstable. The ‘Konservolia’, ranked first for mean yield, showed insteadpoor stability based on Shukla’s stability variance. The mean CV analysis was proposed by Francis (1977) to study the physiological basis of yield stabili- ty. The stable cultivar is the one that provides a high yield performance and consistent low CV (Crossa et al., 1990). According to this method, ‘Abou-Satal’, ‘Shenge’, ‘Mission’ and ‘Koroneiki’ were the most s t a b l e ; ‘ Z a r d ’ , ‘ K o n s e r v o l i a ’ , ‘ S e v i l l a n o ’ a n d ‘ A r b e q u i n a ’ w e r e i n t e r m e d i a t e s t a b l e , w h i l e Amigdalolia, roghani, Amphissis and Manzanillo w e r e t h e m o s t u n s t a b l e c u l t i v a r s ( T a b l e 7 ) . Moghaddam and Dehghanpour (2001) stated that the main problem with this method is that low-yield- ing cultivars are placed into the category of stable cultivars. in this experiment high yielding varieties were in intermediate parts of classification. A greater coefficient of determination (ri2) value is desired because higher ri2 values indicate favor- able responses to environmental changes (Sayar et al., 2013). in our study, Zard, Mission and Sevillano cultivars had higher ri2 values for fruit yield and ‘Amigdalolia’, ‘Koroneiki’, ‘Konservolia’ and ‘Shenge’ with medium ri2 values have high and medium stabil- ity in yield, respectively while others with low ri2 val- ues were unstable cultivars (Table 7). According to the ASV ranking, the following culti- vars were the most stable, Mission, Amigdalolia and K o r o n e i k i , w h i l e t h e m o s t u n s t a b l e w e r e ‘Konservolia’, ‘Sevillano’, ‘roghani’, ‘Arbequina’ and ‘Abou-Satal’ . Based on yield cluster analysis olive cultivars were classified into three categories. Category 1 was culti- vars having high yield and medium alternate bearing (‘Konservolia’, ‘Sevillano’ and ‘Koroneiki’) (Fig. 3). These cultivars are widely adapted around the world (Barranco et al., 2000; Therios, 2009). Barranco et al. (2000) reported that ‘Konservolia’ has a high produc- tivity and alternate bearing but ‘Sevillano’ is produc- tive with constant production in Mediterranean regions. Also, Therios (2009) stated that ‘Sevillano’ is cultivated in warmer regions in Spain and italy with- o u t a n y p r o b l e m s . O u r r e s u l t s r e v e a l e d t h a t ‘Sevillano’ had relatively constant production during the experiment. Koroneiki is one of the most impor- tant olive oil cultivar in the Greece with high fruit yield and good oil quality (Barranco et al., 2000). Our results indicated that its productivity was relatively high and constant but oil content (data not present- ed) was low. Category 2 was cultivars having medium yield and m e d i u m o r h i g h a l t e r n a t e b e a r i n g ( ‘ Z a r d ’ , ‘Manzanillo’, ‘Mission’, ‘Arbequina’ and ‘Amigda- lolia’) (Fig. 3). results showed that ‘Arbequina’ had medium productivity with medium alternate bearing. Our result was not confirmed by Therios (2009) and Barranco et al. (2000) findings, where ‘Arbequina’ has a high productivity with constant yield and high oil content in the italy. Therios (2009) stated that Manzanillo is categorized as a good performance o l i v e c u l t i v a r i n t h e w o r l d . i n o u r r e s e a r c h , ‘Manzanillo’ had medium productivity with high Arji - Olive fruit stability analysis 523 alternate bearing. Mission is a dual-purpose commer- cial olive cultivars in the world (Therios, 2009). Mission’s productivity was medium and alternate in our research. Amigdalolia is an olive cultivar originat- ed from Greece with medium productivity and alter- nate bearing (Barranco et al., 2000). Our result repre- sent that this cultivar show medium productivity and alternate bearing. Category 3 was cultivars having low yield and low, medium or high alternate bearing (‘Abou-Satl’, ‘Shengeh’, ‘roghani’ and ‘Amfissis’) (Fig. 3). We do not recommend these cultivars for planting in warm environmental condition. 4. Conclusions in conclusion, one of major purpose of yield-trial research is to select the best cultivar for a growing region. An ideal cultivar should have the highest mean performance and be highly stable. Such an ideal cultivar would have the greatest vector length of the high-yielding genotypes and zero (G × E). in this study, Zard cultivar performed as the ideal culti- v a r b a s e d o n a l m o s t m e n t i o n e d m e t h o d s . Konservolia, Sevillano and Koroneiki were the highest yielding cultivars in the regional trials. Generally, ‘Konservolia’ and ‘Sevillano’ are introduced for pick- ling use; while ‘Koroneiki’ is not suitable for cultiva- tion in hot and dry regions due to low oil content (data not presented). Acknowledgements We are grateful to the Seed and Plant improve- ment institute for financial support (Project number: 4-100-120000-04-0000-83054). References ArCiNiEGAS-ALArCON S., GArCiA-PENA M., DiAS C.T.S., 2011 - Data imputation in trials with genotype × envi- ronment interaction. - interciencia, 36: 444-449. Arji i., ArZANi K., 2008 - Effect of water stress on some biochemical changes in leaf of five olive (Olea europaea L.) cultivars. - Acta Horticulturae, 791: 523-526. BArrANCO D., CiMATO A., FiOriNO P., rALLO L., TOUZANi A., CASTANEDA C., SErAFiNi F., TrUjiLLO i., 2000 - World catalogue of olive varieties. - international Olive Oil Council, Madrid, Spain, pp. 360. CHENG Z., ZHAN M., YANG Z., ZUMSTEiNA K., CHEN H., HUANG Q., 2017 - The major qualitative characteristics of olive (Olea europaea L.) cultivated in Southwest China. - Front. Plant Sci., 8: 559 CiT AD iN i. , S CAriOT T O S . , S ACHE T M. r. , rOS A F . j. , rASEirA M.C.B., jUNiOr A.W., 2013 - Adaptability and stability of fruit set and production of peach trees in a subtropical climate. - Sci Agric., 71(2): 133-138. CLAY H., SNELLEr C.H., DOMBEK D., 1995 - Comparing Soybean cultivar ranking and selection for yield with AMMI and Full-Data performance estimates. - Crop Sci., 35: 1536-1541. CrOSSA j., GAUCH H.G., ZOBEL r.W., 1990 - Additive main effect and multiplicative interaction analysis of two international maize cultivar trials. - Crop Sci., 30: 493- 500. EBErHArT S.A., rUSSELL W.A., 1966 - Stability parameters for comparing varieties. - Crop Sci., 6: 36-40. E S M A E i L Z A D E H - M O G H A D D A M M . , Z A K i Z A D E H M . , AKBAri-MOGHADDAM H., ABEDiNi ESFAHANi M., SAYAHFAr M., NiKZAD A.r., TABiB GHAFFAri S.M., LOFTALi AEiNEH G., 2011 - Genotype × environment interaction and stability of grain yield of bread wheat genotypes in dry and warm areas of Iran. - Seed and Plant improvement journal, 27-1(2): 257-273. FALCONEr D.S., McKAY T.F.C., 1996 - Introduction to quan- titative genetics. - Longman group Ltd, Essex, UK, pp. 464. FErrEirA D.F., DEMETriO C.G.B., MANLY B.F.j., MACHADO A.A., VENCOVSKY r., 2006 - Statistical models in agri- culture: biometrical methods for evaluating phenotypic stability in plant breeding. - Cerne Lavras, 12: 373-388. FiNLAY K.W., WiLKiNSON G.N., 1963 - The analysis of adaptation in a plant breeding program. - Aust. j. Agric. res., 14: 742-754. FrANCiS T.r., 1977 - Yield stability studies in short-season m a i z e ( Z ea m a y s L . ) . - Ph D T h es i s , C ro p S c i en c e Department, University of Guelph, Guelph Ont. FrANCiS T.r., KANNENBErG L.W., 1978 - Yield stability studies in short season maize. I. A descriptive methods for grouping genotypes. - Can. j. Plant Sci., 58: 1029- 1034. GAUCH H.G., 1992 - Statistical analysis of regional trials. AMMI analysis of factorial designs. - First edition Elsevier Pub. Amsterdam, The Netherlands. GAUCH H.G., 2013 - A simple protocol for AMMI analysis of yield trials. - Crop Sci., 53: 1860-1869. GAUCH H.G., ZOBEL r.W., 1988 - Predictive and postdictive success of statistical analyses of yield trials. - Theor. Appl. Genet.,76: 1-10. GAUCH H.G., ZOBEL r.W., 1996 - AMMI analysis of yield trials, pp. 85-122. - in: KANG M.S., and H.G. GAUCH (eds.) Genotype by environment interaction. CrC Press, Boca raton, CA, USA, pp. 416. GOLDSCHMiDT E.E., 2005 - Regulatory aspects of alter- nate-bearing in fruit trees. - italus Hortus, 12: 11-17. HAYWArD M.D., BOSEMArK N.O., rOAGOSA i., 1993 - Plant breeding, principle and prospects. - Chapman and Adv. Hort. Sci., 2018 32(4): 517-524 524 Hall, London, UK, pp. 576. KAN A., KAYA M., GUrBUZ A., ANLi A., OZCAN K., CiFTCi C.Y., 2010 - A study on genotype x environment interac- tion in chickpea cultivars (Cicer arietinum L.) grown in arid and semi-arid conditions. - Sci. res. Essays, 5(10): 1164-1171. KHALEGHi E., ArZANi K., MOALLEMi N., BArZEGAr M., 2015 - The efficacy of Kaolin particle film on oil quality indices of olive trees (Olea europaea L.) cv ‘Zard’ grown under warm and semi-arid region of Iran. - Food Chem., 166: 35-41. LAVEE S., 2007 - Biennial bearing in olive (Olea europaea). - Annales Ser. His. Nat., 17: 101-112. LODOLiNi E.M., NEri D., 2012 - How growth and reproduc- tion cycles affect alternate bearing in olive. - Acta Horticulturae, 949: 191-198. MArONE E., FiOriNO P., 2012 - Oleiculture in Progress. - Adv Hort. Sci., 26(3-4): 163-175. MAULióN E., VALENTiNi G., OrNELLA L., PAirOBA C.F., DAOrDEN M.E., CErViGNi G.D.L., 2014 - Study of statis- tic stability to select high-yielding and stable peach genotypes. - Sci. Hort., 175: 258-268. M O G H A D D A M A . , D E H G H A N P O U r Z . , 2 0 0 1 - Interrelationship among several stability statistics esti- mated in maize yield trials. - Seed and Plant journal, 17: 329-338. MONSELiSE S.P., GOLDSCHMiDT E.E., 1982 - Alternate bearing in fruit trees. - Hortic. rev., 4: 128-173. OLiVEirA E.j., FrEiTAS j.P.X., jESUS O.N., 2014 - AMMI analysis of the adaptability and yield stability of yellow passion fruit varieties. - Sci. Agric., 71(2): 139-145. OLiVEirA E.j., GODOY j., 2006 - Pod yield stability analysis of runner peanut lines using AMMI. - Crop Breed. Appl. Biotechnol., 6: 311-317. PiNTHUS M.j., 1973 - Estimate of genotype-value: A pro- posed method. - Euphytica, 22: 121-123. rAKONjAC V., ŽiVANOViC T., 2008 - Stability of yield and fruit quality in promising peach cultivars. - jCEA, 9(1): 161-168. rOMAGOSA i., FOX P.N., 1993 - Genotype x environment interaction and adaptation, pp. 373-390. - in: HAY- WArD M.D., N.O. BOSENMArK, and i. rOMAGOSA (eds.) Plant breeding: Principles and prospects. - Springer Verlag, Berlin, Germany, pp. 550. SAADATi S., MOALLEMi N., MOrTAZAVi S.M.H., SEYYEDNE- jAD S.M., 2013 - Effects of zinc and boron foliar appli- cation on soluble carbohydrate and oil contents of three olive cultivars during fruit ripening. - Sci. Hort., 164: 30-34. SABAGHNiA N., DEHGHANi H., SABAGHPOUr S.H., 2006 - Nonparametric methods for interpreting genotype x environment interaction of lentil genotypes. - Crop Sci., 46: 1100-1106. SAYAr M.S., ANLArSAL A.E., BASBAG M., 2013 - Genotype- environment interactions and stability analysis for dry- matter yield and seed yield in Hungarian Vetch (Vicia pannonica CrANTZ). - Turk. j. Field Crops, 18(2): 238- 246. SHUKLA G.K., 1972 - Some statistical aspects of partition- ing genotype-environmental components of variability. - Heredity, 29: 237-245. TEMiME S.B., WAEL T., BECHir B., LEiLA A., DOUjA D., MOKHTAr Z., 2006 - Changes in olive oil quality of Chetoui variety according to origin of plantation. - j. Food Lipids, 13: 88-99. THEriOS i.N., 2009 - Olives. Crop production science in hor- ticulture series. - CABi, Walllinford, Oxfordshire, UK, pp. 409. WriCKE G., 1962 - Über eine methode zür Erfassung der Okoiogischen streubreite in Feidversuchen. - Zeitschrift für Pflanzenzüchtung, 47: 92-96. ZOBEL r.W., WriGHT A.j., GAUCH H.G., 1988 - Statistical analysis of a yield trial. - Agron. j., 80: 388-393.