Seleksi Simultan Karakter Daun Mengering dan Produktivitas pada Galur-galur Padi

Abstract

Drought is an important constraint for rice production in rainfed lowland and shallow freshwater swamp. The area often experiences drought stress at the generative stage of the plants. This study aimed at selecting adaptive lines to terminal drought and formulating a multiple regression model to estimate the productivity under drought stress conditions at the generative stage. The experiment was conducted in a greenhouse of the Indonesian Center for Rice Research, Subang, West Java, using an augmented design with five blocks. The genetic material used was ninety-nine lines and four checks varieties, namely Inpari 30, Limboto, Salumpikit, and IR 20. The model was formulated using stepwise regression analysis. Based on this study, ten lines were adapted to drought stress at the generative stage, namely B13983E-KA-12-2, B13926E-KA-13, B13507E-MR-19, B14366E-KY-50, B14366E-KY-37, IR86384- 46-3-1-B, BP20452e-PWK-0-SKI-1-1, BP20452e-PWK-0-SKI-2-4, BP20452e-PWK-0-SKI-3-3, and BP29790d-PWK-3 -SKI-1-5. The B13507E-MR-19 had the highest productivity (4.02 ton ha-1) under drought stress conditions. Yield under drought stress in the greenhouse could be predicted using a linear regression model involving plant height at early vegetative stage, plant height up to the panicle, tiller number at early vegetative stage, tiller number at late vegetative stage, tiller number at flowering, heading time, number of filled grain, and panicle exsertion length. This model was able to explain 75.92% of yield variation. Potential rice lines and the regression model obtained are expected to contribute to the development of rice varieties adaptive to drought.

Keywords: drought tolerant, freshwater swamp, rainfed, regression model