GENETIC DIVERGENCE OF JESSO-BALAM RICE ( Oryza sativa L . ) OF BANGLADESH

In total 27 genotypes of Jesso-Balam pure line rice germplasm were studied for fourteen morpho-physiological and four physico-chemical characters at Bangladesh Rice Research Institute, during T. Aman 2009 and T. Aman 2011 seasons. According to principal component scores, the two-dimensional scatter diagram (Z1-Z2) has apparently distributed the genotypes into seven clusters. According to Mahalanobis’ D2 statistic for the studied characters, the genotypes were clubbed into seven groups. The highest number of genotypes (6) were included in clusters I and III and the lowest two (2) in cluster V and VII. It was also revealed that no duplicate was existed among the genotypes. The intra-cluster distance was maximum (0.77) in cluster III and minimum (0.28) in cluster VII, while the maximum intercluster distance (28.37) was observed between the clusters III and IV and the minimum (2.49) between the clusters I and VII. The cluster V has produced the highest means for plant height (148.18 cm), straw yield per hill (39.97 g), panicle length (28.14 cm) and protein content (9.61%). But, the cluster IV has showed the highest means for seedling height (69.5 cm), flag and penultimate leaf area (1059.5 cm2), culm diameter (5.8 cm), filled grain number per panicle (3.64), grain length (8.63 mm) and 1000-grain weight (21.78 g), while the cluster VII had the highest means for grain yield per hill (31.27 g) and effective tiller number per hill (13.5) and the lowest mean for days to maturity (144). Therefore, the genotypes of cluster V need to be crossed with the genotypes of cluster IV or cluster VII, for developing new Balam rice with maximum other good characters. Finally, it could be concluded that the studied Jesso-Balam pure lines germplasm is a good source of valuable genes.


Introduction
Being rice as the staple food in this flooded basin, Bangladesh is one of the largest producers of rice in the world (Anonymous, 2013).Presently, Bangladesh is self sufficient for its production, but still the national average yield is quite low compare to the other countries like China, Japan and South Korea.Moreover, the increasing demand of rice needs to meet with less input and land in future.Besides, changing global climate has negative impact on crop production worldwide.Consequently, rice becomes more important for the resource-poor farmers in unfavourable ecosystems, where other food crops are no longer existed.As a result, high yield with high nutritional and indigenous or traditional values need to be considered for developing new rice varieties in future.
Rice genetic resource is the primary material for rice breeding and makes a concrete contribution to global wealth creation and food security (Zhang et al., 2011).Exploring diversity in a landrace collection is very important for identifying new genes and further improvement of the germplasm (Jayamani et al., 2007;Thomson et al., 2007).Genetic diversity is important for an efficient choice of parents for variety development programs.However, rice diversity in Bangladesh is now threatened due to extensive cultivation of modem varieties (MVs) all over the country along with various intervention of rice habitat (Ahmed et al., 2010).Therefore, it is important to relaunch the old famous local rice varieties like local Balam rice (a kind of Basmoti rice) of southern region by maximizing its yield with valuable traditional traits through hybridization programs for maintaining rice diversity in the field.As long as, limited work has been done on genetic divergence of local rice of Bangladesh.The present study was, therefore, undertaken to assess the extent of genetic diversity in 27 genotypes of Jesso-Balam rice germplasm of Bangladesh for selecting prospective parents for developing new promising varieties of Balam rice.

Materials and Methods
In total 27 genotypes of Jesso-Balam TAPL (Transplant Aman Pure Line) rice germplasm, developed from a single cross in the hybridization program by Plant Breeding Division, BRRI, through head to raw method, were selected and collected from Rice Gene bank, Bangladesh Rice Research Institute (BRRI), Gazipur, Bangladesh (Table 1) for study of fourteen morphophysiological and four physico-chemical characters.Thirty days old single seedling was transplanted per hill using spaces within and between rows of 20 and 25 cm, respectively for morpho-physiological study during T. Aman 2009 and T. Aman 2011 seasons.Fertilizers were applied @ 60:20:40 kg NPK/ha.Total quantity of TSP and MP were applied before final land preparation.Urea was top dressed at 15, 30 and 45 days after transplanting (DAT).Nine plants from each entry were randomly selected for recording data regarding seedling height, flag and penultimate leaf length and width, plant height, culm diameter, effective tiller number per hill, straw yield per hill, days to maturity, panicle length, primary and secondary branch number per panicle, filled grain number per panicle, grain yield per panicle, grain length, 1000-grain weight and grain yield per hill.Besides, milling outturn, cooking time, amylose and protein content were measured during 2011according to the laboratory manual, Grain Quality and Nutrition Division (GQND), BRRI.Genetic diversity was worked out following principal component analysis (Rao, 1964) and Mahalanobis' (1936) generalized distance (D 2 ) analysis extended by Rao (1952).Clustering of genotypes was done according to Tochers method (Rao, 1952).All multivariate analyses were performed in computer using GENSTAT 5.13 program.

Results and Discussion
Analysis of variance has shown highly significant differences among the 27 genotypes for all the 18 morpho-physicochemical characters.The first seven components in the principle component analysis with eigen values >1, contributed 85.72% of the total variations among the genotypes (Table 2).Siddique et al. (2011) andChakravorty et al. (2013) found the contribution of 77.91 and 75.9% of the first two and four components to the total variation in rice, respectively.However, Hossain (2008) observed that the first eight axes accounted about 90% of the total variations by PCA in 78 aromatic and fine grain rice landraces.
Based on the values of principal component score 1 and 2 obtained from the principal component analysis, a two-dimensional scatter diagram (Z1-Z2) using component score 1 as X-axis and component score 2 as Y-axis was constructed (Fig. 1).The genotypes were clubbed apparently into seven clusters indicating the existence of considerable diversity among the genotypes.The inter-genotypic distance (D 2 ) was obtained from principal coordinate analysis.The highest inter-genotype distance (1.3940) was observed between the genotypes JBPL13 and JBPL25, followed by 1.3393 between JBPL16 and JBPL25, 1.2416 between JBPL23 and JBPL25.The lowest distance (0.2388) was observed between the genotypes JBPL2 and JBPL6, followed by 0.2801 between JBPL22 and JBPL27, 0.2921 between JBPL1 and JBPL18, respectively.
The distribution pattern of the cluster indicated that the highest number of genotypes (6) included in clusters I and III, followed by five in Therefore, if the genotypes of cluster V are crossed with the genotypes of cluster IV or cluster VII, would exhibit a high heterosis as well as transgressive segregation and would be clubbed the maximum good characters.The inter-cluster distance between the clusters IV and V was 16.22 and between the clusters V and VII was 10.56 indicating moderate distance among the genotypes of these clusters.Besides, moderately diverse genotypes are more suitable in producing desirable variability in the segregating population (Chauhan and Singh, 1982).As a result, the genotypes under clusters V, IV and VII may be selected as parents for future breeding program for developing high yielding Jesso-Balam rice variety along with good grain attributes.Hosan et al. (2010), Mahalingam et al. (2012), Sohrabi et al. (2012) and Medhabati et al. (2013) earlier reported similar trend of conclusions using Mahalanobis' D 2 statistics on rice.
In the present study, it also appeared from the canonical analysis that 85.67% of the total variation was accounted for canonical root 1 and 6.33% by canonical root 2 (Table 6).The coefficients pertaining to the different characters in the first two canonical roots are presented in Table 7.It was revealed from canonical variate analysis that grain yield per panicle, 1000-grain weight, grain yield per hill, milling outturn, straw yield per hill and plant height were responsible the most for both the primary and secondary differentiations and contributed maximum to the genetic divergence for the characters studied.Hossain (2008) studied the genetic divergence in 78 aromatic and fine grain landraces of rice genotypes for 18 morpho-physicochemical characters and found that head rice recovery percentage had the highest contribution followed by milling outturn, stem length, elongation ratio, days to flowering and days to maturity in descending order of contribution towards the total divergence among the characters studied.Similarly, Medhabati et al. (2013) by studying the genetic divergence on 37 wild and cultivated rice's for 12 agro morphological characters and observed that the grain yield per plant, spikelet per panicle, 100grain weight, grain length, days to 50% first flowering, ear bearing tiller/plant and flag leaf length contributed maximum towards genetic divergence.

Fig. 1 .
Fig. 1.Two dimensional scatter diagram based on PCA scores for 18 morpho-physicochemical characters of 27 Jesso-Balam PL rice germplasmThe 27 genotypes of Balam rice germplasm grouped into seven clusters according to nonhierarchal clustering method using Mahalanobis' D 2 statistic for 18 morphophysico-chemical characters (Table3).Sohrabi et al. (2012) reported six clusters in 50 Malaysian upland rice accessions for 12 growth characters, yield and yield components.However,Roy et al. (2004) found five clusters from 35 Aman rice cultivars for 10 traits (characters) during Kharif season.However, the result was in conformity with the cluster pattern of the genotypes obtained though Principal Component Analysis.

Table 1 .
The list of 27 genotypes of Jesso-Balam TAPL with BRRI Genebank accession number

Table 6 .
Values of canonical roots and percentage of variation of 18 morpho-physicochemical characters of 27 Jesso-Balam PL rice germplasm