DIVERSITY ANALYSIS IN BORO RICE ( Oryza sativa L . ) ACCESSIONS

An experiment was conducted with 36 accessions of traditional (local) Boro rice germplasm accessions of three different groups (20 accessions as Kaliboro, 12 as Jagliboro and 4 as Tepiboro) during Boro season 2004 at BRRI farm to identify the duplicates with the help of morph-agronomic characters. On the basis of D values, the 36 genotypes were grouped into six clusters with a range of intra (0.00 for cluster II to 1.78 for cluster I) and inter cluster (1.99 between cluster I and III to 21.20 between cluster II and III) distances. Cluster I comprised the highest number of genotypes (10) and cluster II the lowest (1), while cluster III, IV, V, and VI included 6, 7, 5, and 7 genotypes, respectively. Differences in cluster means existed for almost all the characters. The highest mean value for seedling height (21.68 cm), 1000-grain weight (20.97 g) and grain yield/hill (6.87 g) were observed in cluster I, II for days to 50% flowering (116), panicle length (22.80 cm), grains/panicle (74), and grain length (8.35 mm), cluster IV for tillers/hill (16.44) and panicles/hill (14.17), cluster V for harvest index (0.32) along with cluster I and VI, and cluster VI for plant height (117.17 cm) and flag leaf area (30.68 cm). None of the 12 characters had the highest mean value under cluster III. The canonical variate analysis showed in general that the important characters for the differentiation in the descending order of importance were grain length, days to 50% flowering, grains/panicle, grain yield/hill, panicle length, flag leaf area, plant height, seedling height, 1000-grain weight, panicles/hill, harvest index, and tillers/hill, but the similar name group accessions are not duplicate mainly due to dissimilarity of grain length, days to 50% flowering, grains/panicle and grain yield/hill characters. It is apparent from the results that the same name group accessions were quite different from each other.