Eco-geographic Variation of Common Wild Rice-Oryza rufipogon Griff . in Sri Lanka

Crossref, Google Scholar, Index Copernicus International (ICI), Directory of Research Journal Indexing (DRJI) Wild species of rice (Oryza) have superior agronomic characteristics to be incorporated in rice breeding programs worldwide. Population studies of wild relatives of rice in Sri Lanka have not been well documented despite a few attempts. In the present study, phenotypic diversity of Oryza rufipogon populations exist in Sri Lanka were characterized based on nine quantitative morphological traits. Populations (P1, P2, P3, P4 and P5) were established in a common-garden and were characterized. The results revealed moderate phenotypic diversity among O. rufipogon populations studied. However, flag leaf length and awn length were the most variable traits while plant height, flag leaf angle, flag leaf panicle neck length and spikelet angle were the least variable traits. O. rufipogon can be simply distinguished using flag leaf length and width, panicle branching type and distance from panicle base to lowest spikelet insertion. The dendrogram results indicated that four main clusters are at a similarity level of 98.73, showing the diversely related populations with a high identity based on higher similarity values. P1 and P2 populations grouped together by forming the first cluster. The second, third and fourth clusters consisted of P3, P5 and P4 populations, respectively. One population from the first cluster and P3, P5 and P4 populations can be used for conservation. This study highlights the phenotypic diversity of O. rufipogon populations existing in Sri Lanka across the geographic locations and Knowledge on such morphological diversity provides opportunities to design conservation strategies and the potentials of using particular population based on breeding objectives.

Wild species of rice (Oryza) have superior agronomic characteristics to be incorporated in rice breeding programs worldwide. Population studies of wild relatives of rice in Sri Lanka have not been well documented despite a few attempts. In the present study, phenotypic diversity of Oryza rufipogon populations exist in Sri Lanka were characterized based on nine quantitative morphological traits. Populations (P1, P2, P3, P4 and P5) were established in a common-garden and were characterized. The results revealed moderate phenotypic diversity among O. rufipogon populations studied. However, flag leaf length and awn length were the most variable traits while plant height, flag leaf angle, flag leaf panicle neck length and spikelet angle were the least variable traits. O. rufipogon can be simply distinguished using flag leaf length and width, panicle branching type and distance from panicle base to lowest spikelet insertion. The dendrogram results indicated that four main clusters are at a similarity level of 98.73, showing the diversely related populations with a high identity based on higher similarity values. P1 and P2 populations grouped together by forming the first cluster. The second, third and fourth clusters consisted of P3, P5 and P4 populations, respectively. One population from the first cluster and P3, P5 and P4 populations can be used for conservation. This study highlights the phenotypic diversity of O. rufipogon populations existing in Sri Lanka across the geographic locations and Knowledge on such morphological diversity provides opportunities to design conservation strategies and the potentials of using particular population based on breeding objectives.

INTRODUCTION
The rice genus Oryza contains approximately 23 species distributed in Asia, Africa, Australia and America. Of the species, only two are cultivated and the remaining 21 species are wild relatives of rice (Vaughan, 1989;Vaughan et al., 2003). Species in the Oryza genus are highly diverse morphologically (Clayton and Renvoize, 1986;Vaughan, 1994), and are represented genetically by different genome types, i.e. AA, BB, CC, BBCC, CCDD, EE, FF, GG, JJHH and JJKK genomes (Ge et al., 1997), providing valuable opportunities to explore variations for rice breeding programs. Of the wild relatives, species containing AAgenome are the most accessible genetic resources because of their high genetic compatibility with cultivated rice and therefore, transfer of alien genes from those wild relatives can be easily achieved through sexual hybridization (Ren et al., 2003).
Oryza rufipogon Griff. is known as the putative ancestor of Asian cultivated rice (Oryza sativa L.), is the most significant germplasm for rice genetic improvement (Oka, 1988). As an example, for the development of high yielding hybrid rice varieties, a male sterility (MS) gene found in O. rufipogon was first introduced to the cultivated rice. Other agronomically favourable characters, such as genes resistant to tungro virus and bacterial leaf blight in rice (Xa23 gene) and tolerant to acid sulfate soil found in common wild rice, are of great importance for rice genetic improvement (Song et al., 2005). O. rufipogon is broadly distributed in the regions of tropics and subtropics in Asia (Vaughan, 1994). This species is recorded to occur in 113 counties in the world. O. rufipogon has species-specific features compared with other Oryza species.
Besides its distribution in Sri Lanka (normally occurred in wet zone) O. rufipogon is found in Australia, south and southeast Asia, and Papua New Guinea. It is well grown in moist clay, loam and black soil. Generally, Oryza rufipogon is a perennial, cross-pollinator, photoperiod sensitive, found in permanently wet regions in terms of swampy areas like marshes, open ditches, swamps, grasslands, ponds, billabongs, pools along river banks, at the edges of lakes, and in or at the margins of rice fields. Some populations grow under submergence and saline conditions. The water depth preferences are between 0.2 -4m (Duistermaat, 1987). Flowering times vary from place to place (Field observation) with peak flowering time from December to January (Liyanage and Senanayake, 2010).
In Sri Lanka, our recent field observations showed that the current geographic range of this species has significantly reduced, and the extant of populations have gradually decreased in size owing to habitat destruction and degradation as a result of anthropogenic activities such as agriculture, industry and urbanization (Liyanage and Senanayake, 2010). Above drastic human activities have destroyed natural habitats, leading to habitat fragmentation that affects to reduced population sizes and isolation among populations, leading to 'isolation by distance' (IBD) (Zhao et al., 2013).
Most studies related to wild relatives of rice so far have been based on field surveys and collections of conservation neglecting population studies that could contribute to knowledge about genetic diversity and ecogeographic variation among natural wild rice populations. Effective conservation of these species is urgently needed to preserve the remaining populations. Our knowledge of the genetic diversity of the few natural populations on Oryza rufipogon species in Sri Lanka is rather limited, and a better understanding of its population genetics will be of great significance in the in-situ germplasm collection.
Understanding genetic diversity and ecogeographic variation of wild rice (Oryza rufipogon) in Sri Lanka is a timely requirement to implicate conservation & utilization of such endangered species. Comprehensive understanding of eco-genetic diversification and ecogeographic variation (in terms of morphologically) of wild rice (Oryza rufipogon) was focused in this study enabling their effective use in rice improvement.

MATERIALS AND METHODS
Five O. rufipogon populations were sampled representing a wide distribution of natural habitats in the country based on GPS coordinates obtained from previously reported information (Liyanage and Senanayake, 2010), with each population of 25 individuals that were randomly collected at least 5 m distance between two individuals ( Figure 1, Table 1). Plants were established in cement pots at the size of 16 cm × 16 cm × 18 cm in a common garden at the Faculty of Agriculture, University of Ruhuna (Latitude 06.060337°N and Longitude 80.5681455°E) during January to December 2017 to characterize plant morphology. Phenotyping was conducted for nine quantitative parameters and was measured as described in the Bioversity International, International Rice Research Institute (IRRI), and the Africa Rice Center (Center, 2007). All measurements were taken after the time of flowering from three randomly selected tillers per plant, as flowering time is considered to be involved in the creation of various different characters of the Oryza species (Grillo et al., 2009).

DATA ANALYSIS
All the analyses were conducted using SAS 9.1 (SAS, 2002) and Minitab 17 (Minitab, 2014) statistical software packages. Analysis of variance was performed for each parameter separately to identify specific characters that are significantly different among populations. Descriptive statistics were used to obtain the coefficient of variance. Quantitative variables were subjected to Pearson's correlation analysis. A hierarchical cluster analysis was performed using the cluster variable function.

RESULTS AND DISCUSSION
A considerable variation of characteristics was found among populations for most of the traits measured. Morphological diversity was measured by 9 quantitative traits are presented in Table 2. ANOVA results indicated that evaluated quantitative traits showed moderate variations among populations of O. rufipogon. Such morphological variation shows significance in crop improvement on the basis of incorporate desirable traits in different aspects. The highest plant height observed in P4 and P5 populations belonged to the intermediate and wet zones respectively. Population P2 showed the significantly highest culm girth and P1 recorded the lowest (p<0.05). The highest length from flag leaf base to panicle neck (FLPNL) was recorded from plants in the P2 population.
In general morphological differences among O. rufipogon populations are directly correlated to their natural habitats, life cycle and breeding system (Banaticla-Hilario et al., 2013). Previous studies indicated that, culm related traits with wider culm diameter and less number of tillers were significantly associated with crop physiology and yield production by increasing plant lodging resistance (Chuanren et al., 2004). O. rufipogon showed markedly longer culm length and wider culm diameter, greater plant height and longer panicle length implying that higher lodging resistance had developed in O. rufipogon. The presence of such characteristics may help to survive in permanently inundated habitats (Banaticla-Hilario et al., 2013). In addition, longer plant height, longer flag leaf length and greater flag leaf width of O. rufipogon probably more appropriate for waterlogging environmental conditions. As reported by Maduakor and Lal (1989), morphological traits are vital as a tool for assessing the extent of genetic diversity of plant genetic resources. Therefore, the morphological variability observed among populations is indicative of the genetic divergence (Richards, 2000). Wild species of rice may have morphological, biochemical, anatomical adaptations for their survival in the particular environmental conditions.
Oryza rufipogon plants are found under different agro-ecological conditions of the island. This wild rice species have adapted to climatic conditions prevailing in the Hambantota, Matara, Kalutara, Puttalam and Colombo districts of Sri Lanka (Liyanage and Senanayake, 2010). For example, the P2 population from the wet zone in the Matara district showed vigorous plant growth with the highest CG and FLPNL. Moreover this population is grown on the river. The higher annual rainfall and favourable environmental conditions of this population compared to those in the other population would have led to the higher growth performance of O. rufipogon in the wet zone. In addition to that, plants in P3 may have a higher ability to pest attacks due to its AWL. But no significant difference with population P4. Highest FLL was observed in both P3 and P5, which are belonged to the wet zone. Hence, plants were morphologically adapted to wet environments together with high temperature and water availability to grow and survive by increasing plant height and culm girth when increasing water level (Banaticla-Hilario et al., 2013). Sri Lankan researchers have found several morphological differences among populations of O. rufipogon species collected from different locations of Sri Lanka. There are no documentary evidence to support these eco-morphological differences. Table 3 shows the within and among population variations of each parameter. Some traits showed higher variability among populations resulting in a high coefficient of variation (CV%) in PNSPL (44.65%) while some traits such as PH, CG and FLW showed a comparatively lower CV% (9-13%). Though, some traits (PH, CG, FLA and SPA) showed higher within-population variance than among population variance. Trait-specific variation patterns were observed within populations. For example, PH of P5, CG of P2, FLL and FLW in P5, AWL in P1, and PNSPL of P3 and P5, SPA of P4 showed high variations within the populations.
Quantitative traits are important for plant description (Kurlovich, 1998) and largely influenced by the socio-economic scenario and natural selection (Hien et al., 2007). The coefficient of variation is a valuable tool for gaining comparisons of variability in different traits. A reasonable amount of genetic variation was exhibited for the traits evaluated. Plant height was the only character with a coefficient of variation (CV) value less than 10%. However most traits have CV values above 10 % and as high as 44.65%. For the grain yield per plant representing that selection based on these characters is expected to be effective (Zafar et al., 2004).
The boxplots (Figure 2) clearly illustrated overall patterns of entire distribution of populations in relation to parameters, providing a convenient way to visualize the variation of nine morphological traits across populations.
Population P4, which belongs to the intermediate zone is morphologically distinct from each other as shown by a distinguishable cluster in the dendrogram (Figure 3). Hence, the results showed that O. rufipogon plants are adapted to survive in the intermediate zone. Because the current geographic range of O. rufipogon has significantly reduced, and the extant populations have gradually decreased in size owing to habitat destruction and degradation resulting from agriculture, industry and urbanization (Liyanage and Senanayake, 2010). Above drastic human activities have destroyed natural habitats, leading to habitat fragmentation that affects to reduced population sizes and isolation among populations, leading to 'isolation by distance' (IBD). Gradually decrease of population size experience bottlenecks, genetic drift and inbreeding which lead to the loss of alleles and to an increased homozygosity that affect to population extinction (Young et al. 1996). As reported by Banaticla-Hilario & Almazan (2010) natural populations of O. rufipogon can be simply distinguished using flag leaf length and width, panicle branching type and distance from panicle base to lowest spikelet insertion. The presence of stolons and long,

Relationship Among Morphological Traits
The correlation between parameters that was estimated using the Pearson correlation matrix is presented in Correlation is a measure of the degree to which variables differ together or a measure of the intensity of the relationship. The efficiency of selection for yield mostly depends on the direction and magnitude of the relationship between yield and its components and among themselves. In the present investigation, the correlation analysis indicated that the FLPNL was positively correlated with culm girth. Plant height showed positive and significant correlation with FLA, AWL and PNSPL. The FLL had a significant and negative association with FLPNL. This information can be used to recognize phenotypically different sources for traits of interest in breeding programs.

Identification of Closely Related Populations
The results indicated that the five main clusters are at a similarity level of 99.22, showing the diversely related populations with a high identity based on higher similarity values (Figure 3). In addition, P1 and P2 populations grouped together (similarity level = 98.73) forming the first cluster. Most interestingly P4 population was clustered separately with minimum similarity level. The results indicated that the four main clusters at a similarity level of 98.73, showing the most closely related populations with a high identity based on higher similarity values (Figure 3). The second, third and fourth clusters consisted of P3, P5 and P4 populations, respectively.
Populations from the different natural habitats (P1 and P2) clustered together (

CONCLUSION
The present study focused on the genotypic diversity of Oryza rufipogon in natural wild rice populations in Sri Lanka. Flag leaf length and awn length were the most variable traits while plant height, flag leaf angle, flag leaf panicle neck length and spikelet angle were the least variable traits. Increasing desirable traits are also possible using the trait from appropriate wild species. Likewise, some of the identified wild rice (O. rufipogon) populations with beneficial characters could be of direct interest for desirable traits. Some O. rufipogon populations grouped in the same cluster along with indicating close genetic relatedness. According to efficient conservation strategies, this study indicated that one population can be conserved instead of two populations that are representing in the same cluster. Information concerning the morphological variation of O. rufipogon that has been generated from this study reflects a baseline study in understanding its diversity in Sri Lanka. This result suggests a larger influence of ecological factors on the morphology of populations. All these results clearly indicate that the existing variations amongst the morphological traits of Sri Lankan wild rice (O. rufipogon) suggest that they can be used to distinguish it from weedy rice and cultivated rice. Moreover morphological variations are a reflection of highlighting genetic diversity which fulfils the knowledge gap regarding overall morphological variations of O. rufipogon populations in Sri Lanka with proper sampling strategies. Therefore, the genetic diversity information obtained will be useful in the efficient use of Sri Lankan wild rice germplasm collection. Moreover, this information will be useful in the management of O. rufipogon populations to design conservation strategies and the potentials of using particular population based on breeding objectives.