Genetic Diversity Studies in Aromatic Rice ( Oryza Sativa L .) Germplasm

The genetic divergence study conducted with 50 rice genotypes comprising both Basmati and aromatic short grain types revealed significant differences among the genotypes for the yield, its components and grain quality characteristics. Based on the relative magnitude of D 2 values, the genotypes were grouped into eight clusters by both Tocher’s and Euclidian methods of divergence study. Among all the clusters, cluster IV was the largest one with 12 genotypes, whereas cluster VIII had only one genotype (Pusa 1121). The cluster VIII was separated by higher genetic distance from cluster III and II followed by cluster V. Cluster VIII with only one genotype (Pusa 1121) exhibited the highest mean value for kernel length, kernel length after cooking and 1000-grain weight. Cluster VII possessed the highest mean value for kernel length, L/B ratio and kernel length after cooking due to inclusion of the most promising genotypes like Pusa Sugandh-3, Sugandhamathi and Basmati-386, which are known for good quality traits. Cluster V containing genotypes, Gahansal, Chitti mutyalu and Godavari Isukalu exhibited good performance for most of the characters and registered more number of grains per panicle. Based on the inter cluster distances and other desirable attributes, crossing between Pusa 1121, JGL 15336, RNR 2465, Badsha bhog, NDR 6242, Gahansal and Chitti mutyalu is suggested for further improvement in grain quality and yield.

Aromatic rice constitute a special group of accessions well known for its aroma and superfine grain quality (Nene 1998;Singh et al. 2000a,b). Basmati rice has been the food of choice for rich people for centuries. Aromatic rice has occupied a prime position in Indian society, not only for high quality, but also considered auspicious. Modern methods of information technology and awareness about its unique palatability and easy digestibility have expanded the demand even to common man. Every year domestic, as well as international demand is on the rise. Basmati exports are the major source of addition to annual national exchequer. India, one of the major exporters of Basmati rice is well known for its immense diversity of aromatic rice varieties.
As per the growing demand of aromatic rice, the emphasis should be given for the development of high yielding, fine grain aromatic rice with outstanding quality traits like aroma, kernel elongation after cooking, fluffiness and taste. The diversity in crop genetic resources and its understanding is essential for crop improvement in terms of increasing food production. Study of genetic divergence among the plant materials is a vital tool to the plant breeder for an efficient choice of parents for crop improvement. Genetically diverse parents are likely to contribute desirable segregants and/or high heterotic crosses.
Grouping or classification of genotypes based on suitable scale is quite imperative to understand the usable variability existing among them. For the assessment of variation on multivariate scale, Mahalanobis' D 2 -statistic has been proved to be a powerful technique (Murty and Arunachalam, 1966). It provides a quantitative measure of association between geographic distribution and genetic diversity based on generalized distance (Mahalanobis, 1936). In the present investigation, an attempt was made to classify the extent of genetic diversity for certain yield and quality traits in scented rice genotypes for ultimate use in hybridization programme.

Materials and Methods
Fifty germplasm lines containing both Basmati and aromatic short grain types (Table 1)  Observations were recorded for days to 50 per cent flowering, plant height(cm), number of productive tillers per plant, panicle length(cm), number of grains per panicle, number of filled Table 1. Details of aromatic rice genotypes grains per panicle, 1000-grain weight (g), harvest index, grain yield per plant( g) and for grain quality characters viz; kernel length, kernel breadth, L/B ratio, kernel length after cooking, kernel elongation ratio and volume expansion ratio. Analysis of variance was computed as per standard statistical procedure (Panse and Sukhatme, 1978). A measure for group distance based on multiple characters was given following Mahalanobis (1936) using D 2 statistic to compute genetic divergence between genotypes. Each character was ranked on the basis of its combination towards divergence between two entries (di = r t i -r t j ). Rank 1 was given to the highest   (Rao, 1952) have been used.

Results and Discussion
Analysis of variance revealed highly significant differences among the genotypes for all the traits under study, indicating the variability among the genotypes (Table.2). Information about nature and degree of divergence would help the plant breeder in choosing the right type of parents for further breeding programme to improve the produce quality and yield traits. Hence, estimation of genetic diversity in yield and grain quality parameters among genotypes is important for planning the crossing programme. Fifty genotypes of aromatic rice were grouped into eight clusters (Table.3  The cluster IV was the largest one with 12 genotypes; cluster VI with nine genotypes, cluster I and VII with seven genotypes each, cluster III with 6 genotypes, cluster II and V with four genotypes each, while, cluster VIII had only one genotype. (2008), Ansari et al (2010), Bhadru et. al., (2012 and Praveen singh et al (2012). The clustering pattern of genotypes indicated existence of significant amount of variability, which was in conformation with the findings of Soni et al. (1999) and Ahmed et al (2010).

Fig. 1. Cluster diagram representing diversity for 50 genotypes of aromatic rice
The pattern of distribution of genotypes into various clusters was at random indicating that geographical and genetic diversity were not related. This suggested that forces other than geographical origin such as genetic drift, natural and artificial selection, exchange of breeding material might have played an important role in the evolution of diversity of genotypes. Variation in environment could also be responsible for this diversity. Similar conclusions have been drawn by several workers viz., Rao and Gomatinayagam (1997), Pandey et al. (1999), Hegde and Patil (2000), Rather et al. (2001), Vanaja et al. (2003), Nayak et al. (2004), Arun Sharma et al.

Fig. 2. Clustering pattern (Ward's minimum variance dendrogram)
Though D 2 statistics using Tocher method for classifying the genotypes is useful non-hierarchical Euclidian cluster analysis (based on Wards minimum variance dendrogram) ( Figure 2) critically identifies sub clusters of the major groups at different levels. Maximum intra cluster distance was observed (Table 4) (Fig.2) in cluster V (634.21), followed by cluster IV (332.63) and cluster III (328.66). Thus, selection of genotypes based on high per se and other desirable traits from cluster IV, which had maximum number of genotypes (12) might be helpful to generate useful breeding materials. Minimum intra cluster distance in cluster VIII indicated the limited genetic diversity.
Based on inter cluster distance, cluster VIII, which has one genotype was separated by higher genetic distance from cluster III and II. Another cluster VII was divergent from the clusters III and II, followed by cluster V. The hybrids developed from the selected members on the basis of D 2 matrix value would produce highly variable population in the segregating generations.
A wide range of variation was registered in the cluster means for most of the characters studied (Table.5). Higher differences in the mean values were observed for plant height, number of productive tillers per plant, number of grains per panicle, number of filled grains per panicle, 1000-grain weight, grain yield per plant, harvest index and kernel length after cooking, whereas for the characters like days to 50 per cent flowering, harvest index, panicle length, kernel length, kernel breadth, L/B ratio, kernel elongation ratio and volume expansion ratio, the variation was low.
The genotypes included in cluster I were dwarf types with moderate kernel length. Cluster II exhibited the lowest mean value for plant height and highest mean values for number of grains per panicle, number of filled grains per panicle, harvest index, kernel elongation ratio and the lines included were RNR 2465 and JGL 15281. Cluster III with six genotypes exhibited the highest mean value for panicle length, number of grains per panicle and harvest index. Cluster IV exhibited the highest value for kernel length after cooking, kernel length, 1000grain weight and the lowest mean for days to 50 per cent flowering and plant height. It is observed that no single cluster contained unique genotype with all desirable traits, which ruled out the possibility of selecting directly for immediate use. Therefore, hybridization between the selected genotypes from divergent clusters is essential to judiciously combine all the targeted traits. Among 15 characters including yield that are considered for the estimation of genetic divergence, four characters were considered to be potential contributors for genetic divergence ( Table 6). The maximum genetic The contribution of PC1, PC2 and PC3, PC4 and PC5 was 33.28 per cent, 17.95 per cent 13.11 per cent, 7.52 per cent and 5.94 per cent, respectively. The first canonical root accounted for 33.28 per cent of total variance of uncorrelated variables indicating the differentiation of these traits in these genotypes was complete in three phases. The relative contribution of genotypes reflected in the existence of broad parallelism between groups obtained by D 2 analysis and vector analysis. For getting clear association of two dimensional representation of variation, the first three canonical roots should be more than 95 per cent. On the contrary, the two vectors as a whole contributed only 51.23 per cent towards genetic divergence because of which divergence was contributed by kernel length after cooking (61.31%),followed by plant height (10.94%),volume expansion ratio (8.08%), number of productive tillers per plant (7.18%) and these results are in conformity with the findings of Surender Raju (2002), Nayak et al. (2004) andShobha Rani et.al.(2012).
Canonical root analysis was used to confirm the clustering pattern obtained by D 2 statistics to plot the genotypes on two or three dimensional graphs. The canonical root analysis in the present study accounted for the total variance of 77.82 per cent (Table 7) by 5 principal components.   The characters like kernel elongation ratio, harvest index, number of grains per panicle, grain yield per plant, plant height contributed maximum towards genetic divergence in the first vector. In the second vector, plant height, number of filled grains per panicle, kernel breadth, panicle length, grain yield per plant, number of productive tillers per plant,1000-grain weight, days to 50 per cent flowering, number of grains per panicle, kernel elongation ratio and kernel length contributed maximum towards genetic diversity. Panicle length, number of productive tillers per plant, L/B ratio, days to 50 per cent flowering, number of grains per panicle, harvest index, kernel length after cooking, kernel elongation ratio, 1000-grain weight and kernel length contributed much towards genetic divergence in third vector. In the fourth vector, volume expansion ratio, days to 50 per cent flowering, number of productive tillers per plant, kernel elongation ratio,1000-grain weight, number of grains per panicle, kernel length after cooking, kernel breadth, L/B ratio and grain yield per plant contributed maximum towards genetic diversity. Days to 50 per cent flowering, grain yield per plant, kernel length after cooking, kernel elongation ratio, L/B ratio, kernel breadth and panicle length contributed maximum to the genetic diversity in the fifth vector. This was also in conformity with the relative contribution of characters through D 2 statistics and similar type of study was also carried out by Siddique et. al (2010), Maji and Saibu (2012) and Praveen Singh et. al (2012).
From the entire study, it can be concluded that kernel length, kernel breadth, days to 50 per cent flowering, kernel elongation ratio, number of grains per panicle and number of productive tillers per plant are the important traits contributing towards genetic divergence and for discriminating genotypes. Based on the inter cluster distances and high per se performance for the desirable attributes, crossing between the genotypes viz., Pusa 1121 of cluster VIII, JGL 15336, RNR 2465 of cluster II and Badsha bhog, NDR 6242 of cluster III, Gahansal, Chitti mutyalu and Godavari Isukalu of cluster V and Pusa Sugandh-3, Sugandamati and Basmati 386 in cluster VII can be used for improvement of grain yield and quality traits.