535 AJCS 11(05):535-541 (2017) ISSN:1835-2707 doi: 10.21475/ajcs.17.11.05.p317 Genetic variability and traits association analyses on F 2 generations for determination of selection criteria in Indonesian inland swamp rice breeding Mohammad Chozin 1* , Sumardi 1 , Sigit Sudjamiko 1 , and Muhammad Faiz Barchia 2 1 Department of Agroecotechnology, University of Bengkulu, City of Bengkulu 38371A, Indonesia 2 Department of Soil Sciences, University of Bengkulu, Jl. W.R. Supratman, Kandang Limun, City of Bengkulu 38371A, Indonesia *Corresponding author: mchozin@unib.ac.id Abstract The progress of rice breeding program for better adaptation and extension on inland swamp ecosystem largely depends on the presence of genetic variability and relevant selection criteria. This study was undertaken to estimate the genetic variability parameters in segregating populations of rice and to determine the relationship of grain yield and yield contributing traits of rice to determine selection criteria for yield improvement under inland swamp condition. The F 2 population was generated from 11 crosses involving Bengkulu landraces of swamp rice. Then, the F 2 population was evaluated on the inland swamp for their yield and yield- related traits. The genetic parameter was estimated using variance component analysis. Simple correlation analysis was performed for all observed traits and the resulting matrix of correlation coefficients was subjected to factor analysis and path analysis. Analysis of variance indicated that significant variation present among and within F 2 families for most of the traits studied. Grain yield plant -1 exhibited the highest GCV (40.3 %) and PCV (60.11%). Moderate to high or moderate h 2 B with high GAM was obtained on plant height, tiller number, number of productive tiller, spikelet number panicle -1 , 100-grain weight, and grain yield plant -1 . Positive and significant correlation was found between grain yield plant -1 and plant height, tiller number, spikelet number panicle -1 , panicle length, and 100-grain weight. Factor analysis discerned the 9 observed traits into two-factor axes with overall explaining 79.28% of the total variation among traits. The first factor (59.12%) was strongly characterized by yield and yield components (plant height, tiller number, number of productive tillers, spikelet number panicle -1 , panicle length, and 100-grain weight), whereas the second factor (20.16%) was mainly associated with heading date. Path analysis revealed that among the yield components, tiller number and 100-grain weight (the grain size) had strong direct effects on grain yield and they can be taken into account as the selection criteria for rice grain yield improvement under inland swamp ecosystem. Keywords: rice, inland swamp, genetic parameters, factor analysis, path analysis, selection criteria. Abbreviations: GCV_Genotypic Coefficient of Variation; PCV_Phenotypic Coefficient of Variation, GAM_Genetic Advance as Percent of Population Mean; h 2 B _Broad Sense Heritability. Introduction Inland swamp is commonly considered as a marginal ecosystem for crop production due to the inherent physicochemical limiting factors in the soil. The seasonal flood resulted from impeded drainage in inland swamp make the cultural practices more laborious compared dry land or irrigated lowland. Similarly, high acidity and nutrients deficiency, that are main characteriztion of inland swamp, limit the number of crop species can be grown as well as the productivity of the crop (Fitzpatrick et al., 1993; Waluyo and Djamhari, 2011). Rice is probably the most swamp-adapted food crop and exploitation of swampy areas is commonly devoted to rice production (Aselmann and Crutzen, 1989; Zong et al., 2007; Verhoeven and Setter, 2010). However, all rice varieties cannot express their yield potential when grown on the swampy ecosystem (Nassir and Ariyo, 2011). Successful rice production under swampy ecosystem is majorly dependent to the ability of plant genotype to adapt to various environmental abiotic stresses. Aside from high yielding potential, rice varieties for inland swamp production should be devised with complete submergence tolerance (Sakagami and Kawano, 2011), stagnant flooding tolerance (Singh et al., 2011) and iron toxicity tolerance (Sahrawat, 2010). Inland swamps of Indonesia covers 13.3 million hectares and most of these areas are less developed and underutilized (Haryono, 2012). Similarly, the availability of a wide range of indigenous rice varieties that traditionally maintained and cultivated by farmers in swampy areas (Silitonga, 2004; Wahdah et al., 2012) can serve as valuable source of stress- tolerant prevalence to be incorporated with superior germplasms. Therefore, the breeding programs may aim at identifying high-yielding rice varieties with improved tolerance to abiotic stresses in swampy ecosystems. The available genetic resources of Indonesian rice can provide a significant contribution in attaining food self-sufficiency for Indonesia's growing population. Breeding for yield improvement in rice is often tackled by the genetic and physiological natures of rice yield. Like other cereal crops, grain yield of rice is known to be a complex agronomic trait, governed by polygenic inheritance which highly affected by environmental factors (Gravois and McNew, 1993; Wang et al., 2014). Furthermore, grain yield does not act independently and genetically related to yield