Choudhary, Ram KumarM, RANJAN L2024-06-142024-06-142023M/STAT/591/2021-22https://krishikosh.egranth.ac.in/handle/1/5810210462Sugarcane is one of the most important commercial crops grown in India, with vast varieties. India has second position in sugarcane production among the sugarcane growing countries. Primary industries such as sugar mills, jaggery producing units, and chemical industries are largely depend on this crop. The differential performance of genotypes over different environments is known as “Genotype × Environment Interaction” (GEI). Multi Location Trials (MLT) are being carried out for performance testing of genotypes over different locations. Identification of stable genotypes of sugarcane is important to increase the income of farmers as well meeting the input requirement of sugarcane based industries. In present investigation an attempt has been made to identify stable genotypes of sugarcane crop from secondary data of MLT data collected from G. B. Pant University of Agriculture and Technology, Pantnagar. Data comprises of 17 sugarcane genotypes grown in 6 locations with two replication using Randomizes Block Design. Nine characters were investigated for the assessment of GEI through regression models and stability analysis through 9 stability measures bi, Bi, S2di, Di, ri, Wi, ASI, ASTABi including mean yield. Results obtained by assessment of GEI were compared with results obtained by stability measures using rank correlation and cluster analysis. The analysis was carried out using R-studio, Ms- Excel, & GEA-R. The genotypes were coded from coded from G1 to G17 and locations from E1 to E6. Analysis of variance showed that genotypes were significant for all the 9 characters in all the locations. Pooled analysis confirms significant GEI for six characters under study namely cane yield, single cane weight, sugar content, germination%, number of tillers and number of millable canes. However for characters like brix%, juice extraction % and polarization% GEI was found non-significant. The contribution of variation in total variation due to G×E Interaction for characters under study varies from 11.37% to 25.75% and it was maximum for number of tillers (25.75%) and minimum for single cane weight (11.37%). Further, presence of significant genotype × environment interaction at 5% level of significance in sugarcane was confirmed by joint regression models namely Eberhart and Russell model and Perkins and Jink Model for the same 6 characters as confirmed by Pooled analysis. The contributions of variation in total variation by G×E (linear) Interaction for characters were varied from 4.2% to 17.82%. The minimum contribution by G×E (linear) Interaction in total variation was found in cane yield at harvest (4.2%) and maximum in number of millable canes at harvest (17.82%). Stability measures such as bi & Bi, S2di & Di and Wi & ri showed perfect positive correlation (r = 1) for all the traits. It was observed that mean was significantly correlated with only Eberhart’s regression co-efficient (bi) and Perkin’s co-efficient (Bi). The correlation were varied character to character from 0.67 to 0.92. Hence, they may be used for selecting stable as well high yielding genotypes. ASTABi highly positively correlated with shukla’s stability variance and Wricke’s ecovalence index and moderately correlated with S2di & Di with little variation character to character varied. The superior genotypes were identified for the characters namely cane yield, single cane weight, sugar content germination%, number of tillers character, number of millable as G5, G4, G5, G6, G14 and G5 respectively. While poorest genotypes identified for cane yield, single cane weight, sugar content germination%, number of tillers character, number of millable canes character were G16, G3, G10,G16, G10,and G10 respectively. From this study we can conclude that we can use either of Eberhart and Russell model or Perkins and Jinks model and Wricke’s ecovalence index or shukla’s stability variance for selecting the stable genotypes, since perfect positive correlation (r = 1) was found between them. All six character’s mean was significantly correlated with only regression co-efficients of Eberhart and Russell model (bi) & Perkins and Jinks model (Bi). So these regression co-efficient can be used for selection of stable genotypes with high or above average yield. These results were supported by dendrogram obtained through cluster analysis. Six clusters were identified. ASI performed similarly for the characters namely cane yield, cane weight and Sugar contents while for the traits namely germination %, number of tiller and number of millable canes performed similarly as stability measures shukla's stability value and Wricke’s ecovalence index and AMMI Stability measure (ASTABi).EnglishThesis