ENVIRONMENTAL TRIALS OF A MODEL POPULATION OF PHASEOLUS VULGARIS L. IN CONTRAST CLIMATES

The study purpose was to evaluate the potentials of bean genotypes in contrast climates, to identify the most informative environments for assessing the adaptability of genotypes, to select promising starting material for specific areas where this crop is supposed to be bred. To accomplish this purpose, environmental trials were conducted on a collection of bean ( Phaseolus vulgaris L.) accessions in four sites located in different climatic zones of Ukraine: the eastern For-est-Steppe – Kharkiv region; southern Forest-Steppe – Poltava region; Woodlands – Chernivtsi region; southern Steppe – Odesa region. Genotypes and environments were evaluated using the GGE biplot method, which allowed us to graphically visualize the environmental trial data. To develop cultivars for different zones of Ukraine, we recommend selecting parents, taking into account the weather conditions: drought- and highly heat-tolerant genotypes (Holberg) for the south of Ukraine; intensive genotypes (Pervomaicka, Mistseva Bomba 5) for Woodlands; genotypes with stable yields and high drought tolerance (Holberg, Nadiia, UD0300104) for Forest-Steppe. The GGE biplot analysis distinguished the most informative environments allowing full assessments of the adaptability of genotypes – the Forest-Steppe of Ukraine (Kharkiv region and Poltava region). The genotypes with stable yields of seeds were identified: Holberg, Otrada, UD0300104, Bogema, UD0300152, N 201-15 and Synelnykivska 8. Of these accessions, Holberg, Otrada, UD0300104 and Synelnykivska 8 gave high yields, and Holberg and UD0300104 approached the “ideal” genotype. The southern subzone of the Steppe is the most informative for identification of drought- and heat-tolerant bean accessions; the Woodlands – for selection of genotypes with high potential yields. Woodlands; Plant Breeding and Genetics Institute -National Investigation – Steppe. Results and discussion . Genotypes and test environments were assessed by GGE biplot method, which made it possible to graphically visualize the environmental trial results. To create varieties for different zones of Ukraine, we recommend to select parents with due account for weather conditions: South of Ukraine – drought-resistant genotypes with high heat resistance (Holberg); Woodlands – intensive genotypes (Pervomaiska, Mistseva Bomba 5); Forest-Steppe –stable genotypes with high drought resistance (Holberg, Nadiia, UD030104). Conclusions. From the GGE biplot results, the environmental trial locations, which allow one to fully assess the adaptability of a genotype, were identified: Plant Production Institute named after V.Ya. Yuriev NAAS and Ustymivka Experimental Station of Plant Production. In their conditions, stable genotypes were distinguished: Holberg, Otrada, UD0300104, Bogema, UD0300152, N 201-15 and Sinelnikivska 8. Of these accessions, Holberg, Otrada, UD0300104 and Sinelnikivska 8 combined stability and high performance, approaching the "ideal" genotypes Holberg and UD0300104.

To assess the response of a genotype to an environment, depending on the weather peculiarities during the bean vegetation period, the hydrothermal coefficient (HTC) was used. The HTC was calculated from the formula (1) : , (1) where ∑r -is precipitation amount during the vegetation period, mm; ∑t − sum of temperatures above 10°С during the vegetation period, °C; 0.1coefficient.
The PBGI is situated in the steppe in the South ean daily air temperature is 15.6-22.6 C; the precipitation amount is 130 mm. In 2015, the vegetation period was characterized by aboveaverage mean daily air temperature (20.4°C), and the precipitation amount was significantly below the multiyear average (83.8 mm). Thus, the HTC was 0.45, which is significantly lower than the optimum for the crop (1.00-1.10).
The UES and PPI are located in the Forest-Steppe of Ukraine; however, their multiyear average parameters during the bean vegetation period are different. The mean daily air temperature: UES -15.5-20.5°C, PPI -15.6-21.3°C. The precipitation amount: UES -210 mm, PPI -170 mm. Accordingly, in 2015, the weather conditions during the crop vegetation period were different: at UES the mean daily air temperature reached the upper limit of the multiyear average (20.5°C); the precipitation amount was above the multiyear average (226.7 mm); the HTC was 1.20, which is higher than the optimum. At the PPI, the mean daily air temperature was close to the average (19.6°C); the precipitation amount exceeded the multiyear average (211 mm); the HTC was not much higher than the optimum (1.17). It should be noted that precipitation was rain showers resulted in inopportune watering of plants.
BSAES is located in the Woodlands of Ukraine, where the multiyear average parameters during the bean vegetation period are as follows: the mean daily air temperature is 13.8-18.3°C; the precipitation amount is 274 mm. However, the mean daily temperatures during the crop vegetation period in 2015 exceeded the multiyear average by 1°C (19.5°C), and the precipitation was much below the average (224.3 mm); therefore, the HTC did not significantly exceed that at UES (1.24) ( Table 1). The model population consisting of 24 collection bean accessions with various responses to fluctuations in temperature and humidity in the environment, which were taken from the core collection of the National Centre for Plant Genetic Resources of Ukraine (NCPGRU), served as the test material to investigate the "genotype × environment" interaction. The accessions were sown with manual planters within the timeframe that is optimal for bean. The collection accessions were assessed in accordance with the valid classifier [11] and methodical recommendations for studying collection accessions of grain legumes [12]. The experimental data were statistically processed by analysis of variance in Microsoft office Excel. The environmental assessment of the genotypes was performed by GGE biplot method [13,14,15].
Results and discussion. When studying the "genotype × environment" interactions, several researchers [16,17,18] noted a strong influence of the crop cultivation environment on the yield variability and emphasized that the "genotype" (G) and the "genotype × environment" (GE) effects should be analyzed. GGE biplot using in the researches allows to graphical visualise their results. It promotes to concretize analysis of this observing [13,19]. A clear "genotype × environment" relationship was established by the GGE biplot method in studies on wheat [20], sorghum [21,22,23], soybean [24,25] and bean [26,27].
In 2015, the environmental trials of the model population of bean were conducted in four environments (PPI, UES, BSAES, PBGI) differing in the temperature-water regime during the period of growth and development of the crop.
On average, in the model population, the highest yields were obtained in the woodlands of Ukraine (BSAES) -0.37 kg*m -2 . Moderate yields were registered in the forest-steppe of Ukraine: in the southern part (UES) -0.28 kg*m -2 and in the eastern part (PPI) -0.19 kg*m -2 . The lowest yields were in the southern steppe of Ukraine (PBGI) -0.07 kg*m -2 . There were accessions giving high yields in each environment. At the PPI, UD030076 (0.38 kg*m -2 ), Holberg (0.350 kg*m -2 ) and UD030104 (0.33 kg*m -2 ) gave the highest yields; at UES, such accessions were Dniprovska Bomba (0.48 kg*m -2 ), Nadiia (0.48 kg*m -2 ), Prelom (0.45 kg*m -2 ) and Holberg (0.44 kg*m -2 ); at BSAES Mistseva Bomba 5 (0.58 kg*m -2 ), Pervomaiska (0.58 kg*m -2 ), Sperantsa (0.54 kg*m -2 ) and Prelom (0.50 kg*m -2 ); at the PBGI Holberg (0.14 kg*m -2 ) and Bukovinka (0.10 kg*m -2 ) ( Table 2). The GGE biplot polygon ( Figure 1) shows patterns of the "genotype × environment" relationship. The polygon is divided into sectors. The genotype that is the most adapted to the growing conditions of an environmental trial site is located at the top of the polygon in the same sector with the test marker [14]. We identified genotypes that are the most adapted to these environments: to the forest-steppe Holberg (G6); to the woodlands Mistseva Bomba 5 (G10) and Pervomaiska (G1); to the southern Steppe Holberg (G6). In the environments where the limiting factor for the bean growth and development is water deficit during the crop vegetation period (PPI, UES and PBGI), drought-tolerant Holberg, giving a high yield of seeds, is advantageous.
Upon excess water (BES), genotypes with high potential yields, however with low stability of yields, which depend on the environment and respond positively to favourable for bean conditions of cultivation, are advantageous: Pervomaiska and Mistseva Bomba 5.
Thus, when selecting starting material for hybridization in the woodlands of Ukraine, one should give priority to genotypes with high potential yields, and in the forest-steppe and steppe of Ukraineto highly drought-tolerant genotypes with stable yields of seeds.
To analyse the genotypes" responses to various environmental conditions, the GGE biplot "mean vs. stability" (Figure 2), which is based on their average yields and stability in the environment trial sites, was used. The line passing through the biplot origin and the average environment is called the average environment coordinate (AEC) axis or the yield line. The arrow shows the direction of yield growth. Another axis, perpendicular to the AEC axis, is the stability line dividing the genotypes into highly and poorly stable ones. Accessions with the shortest projection lengths on this axis are considered the most stable . Stable genotypes and genotypes with high potential yields were identified. Among them, there are valuable genotypes with high yields of seeds and high stability of yields Nadiia (G17), Prelom (G2), UD0300104 (G4); genotypes with high potential yields Dniprovska Bomba (G12), Mistseva Bomba 5 (G 10), and Pervomaiska (G1).
The "ideal" genotype is a stable genotype with a high average yield, which has a wide range of adaptive characteristics. To find genotypes combining high yields and stability, the GGE biplot comparing the genotypes with the "ideal" one was used (Figure 3).  Holberg is the most valuable genotype, especially for the forest-steppe of Ukraine, since it is characterized by high drought tolerance. We recommend it as a reference accession for yield (Table 2).
Prelom is the most adapted to mechanized harvesting. Its plants are bushy with indeterminate growth and above-average resistance to lodging. Sperantsa has seeds of an above-average size (1000-seed weight is 0.30 kg). UD0300104 is the earliest ripening (the vegetation period is 73 days) ( Table 3). 0.30 1) -1bushy, 3bushy with nutating top; 2) -3high, 5moderate, 6above-moderate Fan et al. [28] used GGE biplot analysis to assess the discriminating and representative capacities of environments. Due to this method, the testing environments that are more informative for assessing the adaptability of the collection accessions were identified (Figure 4). The vector length is directly proportional to the standard deviation of the genotype yield in this environment. If the testing environment marker is close to the biplot center, i.e. has a short vector, all the genotypes have similar yields, so this environment is not informative for the differentiation of genotypes. A small angle between the environment vector indicates a strong correlation between the genotype yields in these environments [29,30]. Thus, the GGE biplot categorized the testing sites of the environment trials into three groups.
Group I includes the PBGI site (southern Steppe of Ukraine) with a short vector, which indicates a low informativity of this environment about the genotypes under investigation. The accessions grown in this site gave similar yields ( Table 2). This is a region with high summer temperatures and water deficit. Consequently, all the genotypes were suppressed by drought and heat, which made it impossible to determine the whole range of their adaptive features. However, we identified genotypes with high drought-and heat-tolerance: Holberg (G6) and Bukovinka (G20), which gave the highest yields in this environment.
Group II includes the site with a long vector and a large angle with the AEC axis BSAES (Woodlands of Ukraine). This environment has the highest discriminating capacity, which does not allow evaluating genotypes in terms of the whole range of adaptive features, either. Excess water prevents from determining the level of drought tolerance a trait that limits the bean cultivation in Ukraine. In this environment, the intensive genotypes were chosen: Pervomaiska, Mistseva Bomba 5, Sperantsa and Prelom, which gave seed yields of over 0.50 kg*m -2 ( Table 2).
Group III includes the environments that have long vectors and moderate yields (eastern and southern forest-steppe of Ukraine PPI and UES, respectively) and are the most informative for assessing the adaptability of genotypes. These sites of the environment trials are located in the moderately moistened zone and allow evaluating genotypes for stability of seed yields.
Thus, the GGE biplot analysis of the environments singled out the most informative sites of the environment trials, where the study of the genotypes was continued in 2015-2017: PPI eastern Forest-Steppe of Ukraine and UES southern Forest-Steppe.
The weather during the bean vegetation period in 2015-2017 differed both between the sites of the environment trials (eastern Forest-Steppe -PPI and southern Forest-Steppe -UES) and between the study years ( Figure 5).  As a result of the three-year study, we identified the most productive genotypes: for the eastern Forest-Steppe of Ukraine (PPI) -UD0300076 (G3) (0.36 kg*m -2 ), Holberg (G6) (0.31 kg*m -2 ) and UD0300104 (G4) (0.36 kg*m -2 ); for the southern Forest-Steppe (UES) -Holberg (G6) (0.36 kg*m -2 , Dniprovska Bomba (G12) (0.34 kg*m -2 ), UD0300104 (G4) (0.31 kg*m -2 ) and Nespodivanka (G21) (0.31 kg*m -2 ) ( Table 4). The GGE biplot analysis showed that the most stable seed yields were obtained from Holberg (G6), Otrada (G19), UD0300104 (G4), Bogema (G23), UD0300152 (G5), N 201-15 (G25) and Synelnykivska 8 ( G7) (Figure 6). Holberg and UD0300104 were close to the "ideal" genotype (Figure 7). Conclusions. The environment trials of the bean genotypes at the sites differing in the climatic conditions ensured comparison of their informativity. It was established that the southern Steppe (PBGI) and Woodlands of Ukraine (BSAES) did not allow a full assessment of the adaptive potentials of the genotypes due to the extreme for bean weather during its vegetation period: at the PBGI site, it is high temperatures and water deficit in summer; at the BES site excess water. However, these environments are necessary for evaluating drought-and heat-tolerance (PBGI) as well as for identification of genotypes with high potential yields (BSAES). The weather conditions in the eastern (PPI) and southern (UES) Forest-Steppe of Ukraine, which make it possible to assess genotypes in terms of stability of their seed yields, are the most informative for assessing the adaptability of the genotypes.
When developing cultivars for different zones of Ukraine, one should select parents, taking into account their responses to the weather: for the South of Ukraine (PBGI)droughttolerant genotypes with high heat tolerance (Holberg); for the Woodlands (BSAES)intensive genotypes (Mistseva Bomba 5, Pervomaiska); for the Forest-Steppe (PPI and UES)stable genotypes with high drought tolerance (Holberg, Nadiia, UD0300104).
The study of the bean genotypes in the Forest-Steppe of Ukraine (PPI and UES) revealed that the following genotypes gave the most stable yields of seeds: Holberg, Otrada, UD0300104, Bogema, UD0300152, N 201-15 and Synelnykivska 8; of them Holberg, Otrada, UD0300104 and Synelnykivska 8 are noticeable for high yields. Holberg and UD0300104 were close to the "ideal" genotype.