Ashalatha, K VChandran, K P2019-10-172019-10-172000No. of Reference 91http://krishikosh.egranth.ac.in/handle/1/5810131724"In the present study, an attempt was made to have am insight into the yield loss mechanism in groundnut [Arachis hypogaea L.) in presence of the fungal diseases viz., late leafspot and rust. Analysis of variance of yield and disease variables, revealed th a t there existed significant variations among the genotypes under study, both for yield potential and response to diseases. Moreover, disease severity was more in late sovra trial and th u s low yield, as compared to the early sown trial. Simple linear regression models of 3deld loss on disease variable at each stage, revealed that in early sown trial for pod, kernel and fodder loss, disease a t 85th day after sowing (DAS) was having more explanatory value compared to other stages, while it was 65th DAS for oil loss (based on r2). In late sown trial, explanatoiy value of the models were low, comparatively and better results were obtained with disease at 95th DAS for all yield loss variables. AUDPC models, polynomial and non linear models could not explain the variations in yield loss more precisely than the models with single point disease variables. Multiple point linear regression models of yield loss on disease a t different stages taken together, improved the explanatory value, b u t could not explain the variation precisely. Stepwise regression analysis on disease variable and physiological traits like leaf area index, harvest index, partitioning coefficient and growth rates improved the value of the models considerably, both in early and late sown trials. Grouping of genotypes with pod loss and disease variables using Mahalonobis D2 showed similar results in both trials, where variety TAG 24, which show moderate yield loss even at high disease severity, was grouped separately."ennullThesis