Yield Stability Analysis of Elite Irish Potato (Solanum tuberosum L.) Varieties in Western Ethiopia

Potato is one of the most important horticultural crops widely grown in mid and high lands of Ethiopia. Several potato genotypes has been introduced and tested in different parts of western Ethiopia. However, the stability and performance of these genotypes under different parts of the regions were not yet assessed. Therefore, the objective of this study was to determine the effect of genotype, environment and their interaction for tuber yield and identify stable potato genotypes for possible recommendation. The study was conducted using nine potato genotypes during rainy seasons of 2016 and 2017 at three locations (Gedo, Shambu and Arjo) of western Ethiopia. The experiment was arranged in Randomized complete block design replicated three times. Among the testing locations, high yield (26.56 tha-1) was recorded at Arjo while, low (21.51 tha-1) at Shambu. Similarly, among the tested genotypes CIP39158.30 was showed high yield (36.41 tha-1) followed by CIP384321.30 (35.15 tha-1) while, CIP39264 showed low yield (13.3t/ha). Combined analysis of variance showed the main effect due to environments, genotype and genotype by environment interaction were highly significant (P≤0.01) for tuber yield. The genotype and genotype by environment interaction (GEI) was partitioned using GGE biplot model. The first two principal components obtained by singular value decomposition of the centred data of tuber yield explained 99.75% of the total interaction caused by genotype and genotype by environment interaction (GGE). Out of these variations PC1 and PC2 accounted 77.65% and 22.10%, respectively. Generally, the mean tuber yield, GGE biplot and regression slope identified CIP384321.30 as high yielding and stable genotype in the study area.


Introduction
Genotypes by environment interactions (GEI) is of the most importance to plant breeders [1]. Knowledge of the type and extent of GIE effects for a particular crop allows for efficient utilization of resources, accurate characterization of genotypes and to determine the selection gain over time [2]. GEI has also important implications for variety testing and variety recommendations [3,4]. Besides, multi-environment yield trials are crucial to identify adaptable high yielding varieties and discover sites that best represent the target environment [5].
The association between the environment and the phenotypic expression of a genotype constitute the GEI [6,7]. The GEI determines if a genotype is widely adapted for an entire range of environmental conditions or separate genotypes must be selected for different sub environments [8,9]. When GEI occurs, factors present in the environment (temperature, rainfall, etc.), as well as the genetic constitution of an individual (genotype), influence the phenotypic expression of a trait [10]. The impact of an environmental factor on different genotypes may vary implying that the productivity of crop may also vary among environments [11].
GEI is almost universally considered by plant breeders to be among the main factors limiting response to selection and in general, the efficacy of breeding programs. According to Ngeve [12] the presence of GEI effects is serious problem in comparing the performance of individual cultivars across the environments. It reduces the efficiency of genetic progress and leads to unreliable recommendation in terms of yield and adaptability of genotypes. The analysis of GEI, therefore, becomes an important statistical tool employed by plant breeders not only evaluating varietal adaptation but also in the selection of parents for base population, in classifying environments and improving genotypes with desired adaptability [13,14]. This study used the additive mean effects and multiplication interaction (AMMI) model and regression analysis to assess and rank potato genotypes for tuber yields, their stability, GEI effects and adaptability in three potato growing areas of western Ethiopia.
Among root and tuber crops in Ethiopia, potato ranks first in volume of production and consumption [14]. Potato has been considered as a strategic crop by the Ethiopian government aiming at enhancing food security and economic benefits to the country as, potato has a high potential to supply a cheap and quality food within a relatively short period [15,16]. But, a number of production problems that accounts for low regional as well as national yield have been identified. The major ones are lack of stable, well-adapted, disease and insect pests' tolerant varieties [16,17]. Farmers and researchers want successful potato varieties that show high performance for yield and other essential agronomic traits. Their superiority should be reliable over a wide range of environmental conditions also over years. The basic cause of differences between genotypes in their yield stability is the occurrence of GEI [1].
One of the most important characteristics of an ideal cultivar is stable and high yielder under inconsistent environmental conditions [18]. Moreover, improved crop cultivars need to be superior in yield as well as in other characteristics and this superiority should be expressed in the principal areas where the crop is grown [19]. However, so far less has been done to investigate the stability of superior elite potato genotypes tested in different growing environments in the western part of Ethiopia. Accordingly, a total of eight elite Irish potato genotypes were evaluated against the standard check to identify stable high yielding variety and understand the nature and magnitude of GEI.

Materials and Method
The experiment was conducted for two years (2016 and 2017) at three locations (Arjo, Gedo and Shambu) in western parts of Ethiopia. These sites represent major potato growing areas of the country. Eight potato genotypes introduced from international potato center (CIP); namely CIP384321.30, CIP38502, CIP39016, CIP39158.30, CIP39261, CIP39264, CIP39328, CIP90142 were tested with Jallane potato variety as standard check. The experimental set-up was randomized complete block design with three replications at each site during all seasons. The plot size was four rows of 3 m long, 0.75 m spacing between row and 0.30 m between plants.
Weeding and hilling were carried out as recommended.
Dehaulming was done at 90 days after planting and harvesting 10-14 days later. At harvest, data were recorded for fresh marketable tuber weight for all genotypes across locations. Genotypes, Environments (year and location combinations) and GEI were considered during the analysis.
The Additive Mean effects and multiplicative interaction (AMMI) model was used for data analysis and interpretation of GEI effects on tuber yield. Genstat software version 16 th (http://www.genstat.co.uk/) was used to perform the AMMI calculation and to draw the biplot. The AMMI biplot was developed by placing both genotypes and environment means (main effects) on the x-axis or abscissa and the respective Eigen vectors or score of the first principal component (IPCA1) on the y-axis or ordinate [20]. Furthermore, Finlay and Wilkinson [21] suggested that the mean yield and regression coefficient (b) of genotypes over environments provide further insight into genotypes and environment stability. The locations were considered random and Genotype as a fixed effects and a mixed effect model analysis of variance (ANOVA) was used for statistical analysis. The treatment was broken down into three components: G, E and GEI effects in the following equation [22].
Where yijr, is the average value of the dependent variable of genotype i in environment j and block r, µ is a grand mean, αi is the effect of the i th genotype. Βj, is the effect of the j th environment, αβij is the effect of the i th genotype by the j th environment, bj is the block effect at the j th environment and, is the residual error term ijr

Results and Discussion
There were highly significant (P<0.001) difference among genotypes, year, environment, and the interactions (year x locations, year x genotype, locations x genotype, year x location x genotype) ( Table 1). Statistically, G x E interactions occurs if the performance of genotypes varies significantly across environments. The present study is in agreement with the work reported by Byarugaba et al. [6]. Results in the Table 2  Arjo (26.57 tha -1 ) while the lowest total yield recorded at Shambu (21.61 tha -1 ). Besides, maximum tuber yield was recorded in 2017 (25.28 tha -1 ) while the lowest was recorded in 2016 (23.02 tha -1 ). The highest yielding genotype was CIP39158.30 (36.41 tha -1 ) followed by CIP383421.30 (35.40 tha -1 ) while, CIP39264 (13.3 tha -1 ) was the least yielding genotype. Only four tested genotypes (CIP 90142, CIP 39261, CIP 39328 and CIP 39262) recorded lower total yield than the standard check, Jalane (24.81 tha -1 ). The result implied that the performance of Irish potato genotypes varied among genotypes, across locations and years. The significant differences in the performance of the genotypes across year and locations could therefore, be attributed to differences in the genetic potential of the tested materials, differences in agro-ecological conditions and also genotypes x environmental interactions. Similar results were reported by Nakitandwe et al. [22]. The G x E interaction was further studied using the additive mean effects and multiplicative interactions (AMMI) model. AMMI analysis for tuber yield across environments showed highly significant differences among genotypes performances (Table 3). This revealed that the genotypes response varied from one environment to another. The analysis of variance results partitioned the main effect treatments in to genotype (G), environment (E) and G x E interactions with highly significant differences among all the components. It also partitioned the G x E interaction effects into principal components, where IPCA-I and IPCA-II explained highly significant G x E interactions. The AMMI results are presented in table 5 which allows visualization of relationships between the Eigen values for the first principal component axis (PCA1) and the genotypes and the environment means (main effects). It also shows the variation in genotypes responsible to the environmental changes. Genotypes or environments which appear almost on a perpendicular line have similar means; those falling on horizontal line have similar interaction effects [23]. Genotypes or environments on the same parallel line relative to the Y-axis have similar yield and genotypes or environment on the right side of the midpoint of the axis has higher yield than those on the left side. According to Crossa et al. [24], the abscissa reflects the overall quality for environment and general improvement status for genotypes while the ordinate discriminate early (positive PCA scores) to late (negative PCA) maturing genotypes and correspondingly the length of growing season of locations. Basing on this argument, test genotypes CIP39158.30, CIP384321.30, CIP 38502, CIP 39016 and Jalane (the standard check) were displayed on the right hand side of the midpoint for the x-axis and were thus the high yielding than CIP 90142, CIP39264, CIP39262 and CIP39328 which were on the left hand side. CIP39158.30 at the extreme right and CIP39264 extreme left side were the best and least yielder genotypes. Test genotypes CIP39158.30, CIP384321.30, CIP 38502 and Jalane were categorized as early, but CIP39264 were considered as late maturing Genotypes or environments with large first PCA scores (either positive or negative) have large interaction; those with value close to zero have small interaction and are considered stable [25]. When the PCA values of genotypes and environments are close to zero, the entry has small interaction effects and its general response pattern across the environments parallel the mean of all genotypes in the trial and is thus considered stable [25]. According to Finlay and Wilkinson [20], mean yield of entries across environments and regression coefficents are important indecators of cultivar adaptation. This is in agreement with the ANOVA and AMMI results reported aelier.

Conclusion
Eight elite potato varieties were evaluated for their adaptability and stability with standard check in western Ethiopia at three locations for two consecutive years. AMMI and Regression coefficient showed genotype CIP384321.30 was high yielder and stable genotype implying general adaptation for same genotype. On the other hand, CIP38502 and CIP39158.30 showed specific adaptation to the environments with positive interaction. In conclusion genotype with general adaptation can be grown at all trial locations and similar environment in western Ethiopia while, genotypes with specific adaptation are better at specific.