ANALISIS GGE BIPLOT PADA HASIL KLON-KLON UBI KAYU MENGGUNAKAN METODE RESTRICTED MAXIMUM LIKELIHOOD

Abstract

The study was conducted in five locations i.e Kediri, Ponorogo, Probolinggo, Malang, and Mojokerto, from November 2010 until August 2011. The planting materials used were 15 cassava clones. The research objective was to compare analysis methods of the genotype × environment interaction, namely: a) GGE technique using REML without A matrix by assuming homogeneous residual error variance, b) GGE technique using REML with A matrix by assuming homogeneous residual error variance, c) GGE technique using REML without A matrix by assuming heterogeneous residual error variance, and d) GGE technique using REML with A matrix by assuming heterogeneous residual error variance. The results showed that GGE technique using REML without A matrix by assuming heterogeneous residual error variance was more appropriate. Clones CMM 03038-7 (G8) had a wide adaptability and high yield potential, and its clone was closest to the ideal criteria for genotype compared with other genotypes. Clones CMM 03094-4 (G10) had specific adaptability in the environments S2 (Malang) and S5 (Mojokerto), and it had higher yield potential than the control varieties UJ5, Malang 6, and Adira 4. Environment Kediri (S1) had the highest yield among other environments and Kediri was a suitable environment for the growth and selection of cassava.