Rice (Oryza sativa L.) belonging to the family Poaceae; is life and princes among the cereals and is the staple food in areas of high population density and fast population growth. More than half of the world’s population depends on rice for calories and protein, especially in developing countries. The world’s population particularly in that of the rice consuming countries is increasing at a faster rate. So demand for rice is also increasing. In the year 2008 world’s rough rice production area was about 15.57 million ha and rough production was about 66.18 million ton (USDA, 2009).
Rice is the leading crop in our agro based country. It is grown as Boro, Aus, and Aman crops in three overlapping seasons with large number of varieties that suit various agro-ecological and climatic niches. Thus the varietal requirements for each type varied on respect of plant type, growth duration, and biotic and abiotic stress factors (Das, 2005).
Boro rice has become the leading rice crop through the last ten years. It was cultivated in 4.45 million ha of land and its record production was 17.76 million metric ton which was 61.4% of total rice production during 2007-2008 (BBS, 2009). But high yielding Boro rice varieties require a very long period of 150-165 days to be matured. On the other hand, Aus varieties require comparatively shorter (80-120 days) duration (Alam 1982). The long duration of Boro rice allows them to face more adverse conditions of high temperature, high humidity, frequent precipitation and other uncertain environmental conditions as a result of climate change, particularly at the early period of summer. Again, in most of the cases the Aman-Boro rice cropping pattern does not allow to grow a third crop in between them thus restricting only two crops in a year. In order to allow a third rabi crop between Boro and Aman rice crop and harvest Boro rice within a range of 130-140 days we need to develop short duration materials as quick as possible. The trait, earliness can be transferred from Aus to those of high yielding Boro varieties through crossing among the varieties. Intervarietal crosses between Aus and Boro varieties and evaluation of F1 need to be performed to select short duration materials for using them in generating short duration Boro rice varieties.
Crop improvement in rice depends on the magnitude of genetic variability and the extent to which the desirable genes are heritable. Again, one of the main problems of plant breeders for improving varieties with desired traits is to select good parents and crosses. A critical survey of the genetic variability, correct understanding of the gene effects and knowledge on the extent of heritability of these traits would help in planning an effective breeding program. Combining ability and heterosis are thus, excellent tools which help discern the goal and direction in a breeding program (Manonmani and Khan, 2003). Combining ability analysis also provides information on additive and dominance variance. Its role is important to decide parents, crosses and appropriate breeding procedure to be followed to select desirable segregants (Salgotra et al., 2009). Diallel analysis is one of the most powerful tools for estimating the general combining ability (GCA) of parents and selecting of desirable parents and crosses with high SCA for the exploitation of heterosis (Sarkar et al., 2002).
The research will facilitate to develop and select short duration Boro rice materials which would be used in further breeding programs to develop short duration Boro rice varieties. This would allow early harvest that will escape adversity more likely to occur at the onset of summer. The early material would allow a rabi crop to be grown between Aman and Boro rice to increase cropping intensity and to attain maximum annual economic return.
Some studies related to cross compatibility were reviewed here.
Niruntrayakul et al. (2009) determined the degree of cross compatibility between four cultivated rice (Oryza sativa) varieties, namely, two high yielding varieties, CNT1 and SPR1 and two pure lined traditional varieties, KDML105 and RD6, and two common wild rice biotypes (O. rufipogon) from Kanchanaburi (KC) and Nakorn Nayok (NY). Crosses were made between all varieties of cultivated rice and wild rice, resulting in eight hybrid combinations. Cultivated rice varieties were used as female parents. At maturity, the pollinated panicles were harvested and percentage seed set (the ratio of the number of seeds set to the number of spikelets fertilized) were determined. However, the percentage of seed set was significantly different among cultivated rice×wild rice combinations. For a given cultivated rice parent, the percentage of seed set was dependent on the wild rice parent, and vice versa for a given cultivated rice parent. For the crosses with KC wild rice, the highest percentage of seed set was found with the HYV’s, CNT1 (35%), and SPR1 (36%). For the crosses with NY wild rice, the highest percentage of seed set was again found in SPR1 (62%), and the lowest with CNT1 (10%).
A continuous degree of isolating barriers is observed at intra- and inter-specific levels in rice. Cross-incompatibility due to pre-zygotic barriers after fertilization is rare in rice. F1 inviability (or weekness) is one of the post-zygotic barriers. The products of intra and inter-specific hybridization may fail to reach maturity due to a failure of seed formation is found as cross-incompatibility after fertilization, which frequently results from an arrest of endosperm development in plants. The failure of endosperm development is known in hybrids between O. sativa and O. longistaminata (Hirano et al. 2008).
Hoque et al. (2007) carried out an experiment at the field laboratory of the Department of Genetics and Plant Breeding, Bangladesh Agricultural University, Mymensingh during the Aman and Boro season of 2001-2002 to study the crossability and genetic relationship in F1’s of two wild and three cultivated rice genotypes viz., Wild 4855, Jhora 4325, Dular, Hab. Aman 2 and Jagliboro. Seed set percentage in diallel progenies (without reciprocal) of five rice genotypes ranged from 4.91 to 65.77%. Crossability of Jhora 4325 was highest with Jagliboro followed by Hab. Aman 2, Dular and Wild 4855 that indicate the closeness of the genotypes with each other. Among the cultivated genotypes crossability of Dular with Jagliboro was highest followed by Hab. Aman 2, Jhora 4325, respectively. On the other hand Hab. Aman 2 showed highest crossability with Jagliboro followed by Jhora 4325. Jagliboro showed higher crossability in all the crosses, which could be count as a key genotype in the evolutionary process of the five rice genotypes.
Sah (2007) carried out a study with three wild species viz. O. latifolia (2n = 48, CCDD, IRGC Acc. 100914), O. minuta (2n = 48, BBCC, IRGC Acc. 10114), and O. officinalis (2n = 24, CC, NPGR Acc. 10602), and three cultivated rice viz. IR 64, Radha 4 and CMS IR 69618-A. The wild species were used as male parents while cultivated as female. Seed set was very poor. Among the combinations Radha 4×O. minuta expressed the highest seed set (29 seeds out of 206 crosses) and IR 64×O. officinalis gave the lowest number of seeds (5 seeds out of 190 crosses). All the resultant hybrid seeds were lack of endosperm and only filled with the watery fluid. Embryo degeneration phenomenon was observed across the combinations, and that was successfully overcome by applying the embryo rescue technique.
Crossing barriers are often recognized in rice when distantly related taxa are intercrossed. A range of variation in the rate of success is also found even in crosses between forms sharing the same gene pool (Chu et al. 1969, Sitch 1990). Recently, Okuno and Nakayama (1993) reported that Japanese varieties Akihikari and Nipponbare showed reduced seed setting when crossed with IR varieties (Ir28, IR 36 and IR58) as material parents. The reciprocal crosses, however, showed normal seed setting, indicating unidirectional cross incompatibility. They suspected that the cross incompatibility might be the result of wide hybridization during breeding procedures. If so, this implies that using wild rice creates a new reproductive barrier among rice cultivars (Khush et al. 2003).
Tao et al. (1999) found great variation in different season for interspecific crossability in rice using wild and cultivated rice.
Sano and Kobayashi (1995) detected unidirectional cross incompatibility when a segment of chromosome 6 was introduced from a common wild rice into a japonica-type (T65wx) rice. When pollinated by T65wx, the plant carrying the segment produced aborted and ungerminable seeds showing reduced seed setting while the reciprocal cross between the same parents showed normal seed setting. Cytohistological observations revealed that retardation of the endosperm took place 4-5 d after fertilization and it seemed to cause defective seeds with disturbed tissue differentiation. Furthermore, the japonica type tended to show reduced seed setting in crosses to the indica type and O. glaberrima female while the indica type and common wild rice tended to show normal seed setting.
Since 1986, IRRI has crossed various rice cultivars with nine wild species, using the cultivar as the female parent. Seed set of crosses with AA genome species ranged from 9% to 73%, depending on the rice cultivars and wild species accessions used. For example, IR64 crossed with five O. nivara accessions showed limited variation in seed set (13.9% 28.9%); crosses with IR36 gave wide variation (9.1%-62.2%). Similarly, crosses of IR54 with O. rufipogon IRGC Acc. 100907 gave lower seed sets (32.4%) than that with Acc. 103817 (73.0%). In crosses with IR64, Acc. 100907 gave the highest seed set (53.4%). A similar range in seed set was obtained with all AA genome species, suggesting no difference in crossability among species. These intragenomic crosses produced well-developed seeds; consequently, embryo rescue was not necessary. Whereas the intergenomic crosses were less successful, although hybrid seed was obtained with all species. Seed sets were low (0%-26.3%) and commonly less than 10%. Individual accessions of some species differed noticeably in crossability. For example, O. latifolia (CCDD genome) Acc. 100963 and Acc. 100966 gave good seed sets (4.3%-19.7%) with both IR36 and IR64, while Acc. 100964 and Acc. 100965 gave few seeds (0.1%) with only IR36. Similarly, crosses with O. ridleyi (genome not named) Acc. 100877 consistently gave higher seed sets than with Acc. 100821 and Acc. 101453. However, uniformly low seed sets were obtained from crosses with O. minuta (BBCC) (0-4.2%) and O. brachyantha (FF) (0-1.15), regardless of the accession used. Seed degeneration was also reported after two to three weeks in some pollinated spikelets in their experiment with wild rice and IRRI rice cultivars (Sitch et al. 1989).
Chu et al. (1969) found seed set rate 39 to 42% in iterspecific hybridization.
The concept of combining ability was put forward by Sprague and Tatum (1942) using single cross of maize. They defined the term general combining ability (GCA) as the average performance of a line in a series of hybrid combinations. Specific combining ability was used to designate dose effects in certain combinations, which significantly departed from what would be expected on the basis of average performance of the lines involved.
Chaturvedi et al. (2010) found that component analysis indicated importance of both additive and dominance gene action for the characters days to 50% flowering, 50% flowering to maturity, plant height, panicle length, grain filling percentage and length-breadth ratio, while only the dominance component was important for the characters like effective tillers per plant, panicle weight, grains panicle-1, 1000 grains weight and yield plant-1. Dominant genes were more frequent in the parents than the recessive genes for the characters days to 50% flowering, 50% flowering to maturity, plant height, panicle length and length-breadth ratio. Unequal gene frequencies for positive and negative alleles in the parents were observed for all the characters except grains panicle-1 for which equal gene frequencies were evident. The estimates of degree of dominance were more than unity for all the characters including yield plant-1, which indicated prevalence of over-dominance.
Rahimi et al. (2010) made quantitative valuations of observed heterosis for eleven traits of fifteen F1 hybrids generated by half diallel crosses of six diverse rice cultivars (Domsefid, Hashemi and Binam, three Iranian local cultivars; Dorfak, Kadous and IR30, three improved cultivars). The significance of specific combining ability (SCA) and general combining ability (GCA) for all studied traits revealed that both additive and non-additive gene effects contributed to the inheritance of the traits.
Sharifi et al. (2010) studied genetic components, the mode of inheritance and heterosis of various physiological traits by using seven-parent diallel mating design including reciprocals. Results revealed that both additive and non-additive gene effects were important for inheritance of days to flowering, plant height and panicle exsertion with preponderance of additive gene effects. Inheritance of harvest index was controlled only by additive effects. While, inheritance of grain yield, panicle length and flag leaf area were governed by non-additive gene effects. Wr-Vr graphic analysis indicated that increase of flag leaf area, plant height, panicle exsertion and panicle length was under the control of dominance alleles, but harvest index was increased by recessive alleles.
Alam (2009) found predominance of both additive and non-additive gene action and preponderance of non-additive gene action. He observed that the parents, BR7166-5B-1, BR7166-5B-1-Ran-1 and BRRI dhan28 were good general combiners for flag leaf area; filled spikelet and flag leaf area; dwarf plant and short growth duration respectively. The hybrids BRRI dhan28× BR7166-5B-1-Ran-1, BRRI dhan28×BG304, BR7166-5B-1×BR7166-5B-1-Ran-1, BRRI dhan28×BW328 and BG304×BRRI dhan45 were showed as good specific combiners for yield, short growth duration, lodging resistance dwarf rice varieties, early variety, plant with long panicle respectively.
District Wise Research:
Bagheri et al. (2009) evaluated five rice genotypes (Hasseni, Dailamani, Shastak-mohammadi, Sange-tarom and Daei-shastak) and their ten hybrids obtained through half diallel set for combining ability and gene effects for some rice traits. General combining ability (GCA) and specific combining ability (SCA) were assessed for all the agronomic traits. The results indicated significant differences between general (GCA) and specific combining ability (SCA) for the investigated characteristics. This indicates the role of additive and non-additive gene effects in inheritance of the traits. Variances due to GCA were greater than those due to SCA for plant height, flag leaf width, number of primary panicles, total grain and number of fertile grain per panicle, suggesting the role of additive gene effects as predominant in inheritance of the traits. These characteristics are of high heritability the selection for which may well succeed. But, low heritability of tiller number, days to 50% flowering, flag leaf length, panicle length, seed width, 1000-grain weight and plant characteristics, were emanated from a high degree of non-additive variance in the genetic variance. Thus these characteristics were suggested as important for production of hybrid variety and exploitation of heterosis. The GCA effects of each parent for the named traits showed that the Daei-shastak is a suitable general combiner for yield.
Chakraborty et al. (2009) carried out combining ability analysis for grain yield and its components in eight parental diallel crosses of bold grained rice excluding reciprocals. The general combining ability (GCA) and specific combining ability (SCA) effects were significant for all the characters indicating the importance of both additive and non-additive gene effects in their inheritance. The parents Ranjit followed by Matonga and Monohar Sali contributed significantly to high GCA effect towards high grain yield hill-1. The superior specific cross combinations for high grain yield per hill were Matonga×Bar Madhava, followed by Dhola Mula×Ranjit, Chandmoni×Hathi Sali, Dhola Mula×Mala, Matonga×Mala, Bar Madhava×Hathi Sali and Matonga×Hathi Sali. Further, these crosses exhibited high SCA effects for some other characters as well.
Roy (2009b) observed importance of both additive and non-additive gene actions and preponderances of non-additive gene action for all the characters except plant height in a 8×8 full diallel analysis. He found the parent BRRI dhan44 was the best general combiner and BRRI dhan32, BR11 and Pokkali were good general combiner for yield and most of yield related traits. Pokkali×BRRI dhan44 was the best specific cross combination followed by the cross Rajasail×BRRI dhan32 and Pokkali×BR11 for yield and most of yield related traits.
Ahangar et al. (2008) used a 5×5 half diallel cross genetic design in order to estimate combining ability and gene action of parents and their hybrids for yield and six yield components. Mean squares of genotypes for all traits were highly significant (p < 0.01). Significant mean squares of combining ability for all traits showed additive and non-additive effects in control of the related traits. Having non-significant MS(GCA)/MS(SCA) ratio for all traits but 1000-grain weight show higher importance of non-additive effects in comparison with additive effects of controlling genes. High relative importance of specific combining ability indicated that all traits but 1000-grain weight are highly affected by impacts of specific combining ability. Generally, Dasht, Neda and Binam were the best parents for general combining ability and Binam×IR62871-175-1-10 and Mashhad Domsiah×Binam were the best hybrids for grain yield and its components. Combining ability and gene action for yield and yield related traits were estimated in a diallel cross of rice involving five varieties (Dailamani, Sange-tarom, Hasani, Shastak-Mohammadi and Daei-shastak). The study indicated that both general and specific combining ability effects were significant and important for the more traits, except for primary branching of panicles and paddy width traits. This indicates the role of additive and non-additive gene effects in inheritance of the traits. Variances due to GCA were greater than those due to SCA for plant height, flag leaf width, primary and secondary branching of panicles and filled grain per panicle traits, suggesting the role of additive gene effects were predominant in inheritance of the traits. These characters have high heritability and selection for them may well succeed. But, low heritability of tiller number, flag leaf length, days to maturity, panicle length, paddy length and width, 1000-grain weight and grain yield characters, were because of a high amount of non-additive variance in the genetic variance. Thus these characters were important for production of hybrid variety and exploitation of heterosis. The GCA effects of each parent for these traits showed that the Shastak-Mohammadi and Daei-shastak are good general combiner for days to maturity and grain yield, respectively Bagheri et al. (2008). Rahimi et al. (2008) made crosses six rice cultivars in half diallel design and evaluated GCA and SCA for ten traits. In their study the analysis of variance showed significant differences (p≤0.01) between cultivars. Furthermore, general combining ability (GCA) and specific combining ability (SCA) for parents and hybrids were significant. They demonstrated that according to the analysis based on the second and fourth Griffing methods, additive gene effects were more than non-additive gene effects on controlling growth period, plant height, panicle length, number of panicles plant-1 and brown rice length, while other studied traits were more controlled by non-additive gene effects. Sasmal and Banerjee (2008) performed a diallel analysis to identify parental lines and specific crosses through combining ability estimation in a set of seven rice varieties. Data were collected on sixteen characters including a few root characters. GCA and SCA variances were significant for all the characters. Reciprocal effects were also significant except for rachilla per panicle and grain length. An early flowering NC 1281 Mutant and Taichung Native 1: bold grain mutant were good general combiners for majority of characters. The cross Nira×T-N-1: bold grain mutant recorded the best SCA effects for grain yield as well as all the important yield attributes and also for major root characters. In fact, in most of the high SCA registering crosses for various characters, Taichung Native 1: bold grain mutant or Nira was one of the parents commonly involved. Kumar, et al. (2007a) performed a nine parent full diallel analysis of nine rice genotypes. They found that the parents ADT 44 and CR 1009 were good general combiners for grain yield per plant and the best specific crosses involved parents with low GCA effects implying the need for heterosis breeding and recurrent selection for specific combining ability programme in the segregating generations for substantial improvement of grain yield per plant. Combining ability analysis on a 9×9 diallel design was performed by Kumar et al. (2007b). Results showed that GCA, SCA variances and reciprocal effects were significant for all the six characters of interest. The parent variety ADT 44 portrayed significant positive GCA effects for number of productive tillers per plant, number of filled grains per panicle, 100 grain weight, biomass per plant, grain yield per plant and harvest index. Combining ability analysis for grain yield, its components and some of the quality characters in a 9×9 diallel cross (excluding reciprocals) was carried out. Both general combining ability (GCA) and specific combining ability (SCA) variances were highly significant for all the characters indicating the importance of both additive and non-additive gene actions. However, preponderance of additive gene action was recorded for the traits viz., plant height, days to 50% flowering, dry matter, net assimilation rate, days to maturity, harvest index, 100 grain weight and grain length, while both additive and non-additive gene effects were almost equally important for grain yield plant-1, biological yield plant-1, leaf area index and grain length: breadth ratio. Preponderance of non-additive gene action was recorded for length of panicle and grain length. Parents HPR2047, VL93–3613 and JD8 were good general combiners for grain yield and other related characters. Parents HPR1164 and JD8 were good combiners for shorter plant height and earliness. These crosses involving the above mentioned parents were promising as revealed by their SCA effects. The cross combinations HPR2047xVL93–3613, HPR1164xIR57893-08, VL91-1754xJD8, VL93-3613xJD8 and VL91-1754xVL93–3613 showed significant positive specific effects for grain yield plant-1 and some associated characters Kumar et al. (2007c). Singh et al. (2007) carried out combining ability analysis for grain yield and its components in seven parental diallel crosses of rice (Oryza sativa L.) excluding reciprocal. The GCA and SCA were significant for all the seven characters, indicating the importance of both additive and non-additive genetic components for these traits. Among parents studied, Vaidehi and Rajshree were observed to be good general combiners for grain yield. The superior specific cross combinations Saket 4×Vaidehi, Rajshree×Kamini, Prabhat×Rajshree and Sita×Vaidehi appeared promising for further exploitation in rice breeding programme. Torres and Geraldi (2007) estimated some useful parameters which can be used to investigate the genetic control of agronomic characters in crosses combining cold tolerance and productivity. They used a partial diallel design in crosses between six tropical indica rice cold susceptible genotypes (group 1) and seven japonica or indica/japonica cold tolerant rice genotypes (group 2). Parents and crosses were evaluated for agronomic characters under field conditions in two different experiments in 2005. The predominant direction of dominance effects was negative for days to 50% flowering, and positive for all the other characters. General combining ability (GCA) and specific combining ability (SCA) were significant for all characters, although the GCA effects of the two groups were more important than the SCA effects. A 9×9 diallel mating design analysis of F2 generation (excluding reciprocals) of rice (Oryza sativa) for grain yield, various yield components and some grain quality traits was carried out in a field experiment conducted by Kumar et al. (2006). Difference in the general and specific combining abilities (GCA and SCA, respectively) for all the characters studied was recorded. The magnitude of GCA variance was relatively higher than the SCA variance and thus, the predominance of additive gene action (d) was observed for all the traits except for net assimilation rate (g/cm2/day), biological yield per plant (g), harvest index (%) and protein content (%). Based on GCA effects, the parental cultivars HPR 2047, VL 91-1754 and JD 8 were good general combiners for grain yield and most of its contributing characters. Based on highly significant SCA effects, the cross combinations HPR 1164×HPR 2047, HPR 1164×VLDhan 221, China 988×VL 91-1754 and HPR 2047×JD 8 are suggested for isolation of high yielding lines through pedigree method. In a genetic study conducted by Raju et al. (2006) the results indicated that the inheritance of leaf area index (LAI) at tillering stage, crop growth rate (CGR) from tillering to heading, net assimilation rate (NAR), harvest index (HI), days to 50% flowering, productive tillers plant-1, filled grains panicle-1 and grain yield plant-1 were predominantly under the control of non-additive gene action; whereas, 100-grain weight was largely governed by additive gene action. The parents IR 20 for LAI, RDR 763 for LAI, NAR and productive tillers plant-1 and Lunisree for CGR, HI, Biological yield and 100-grain weight were identified as good general combiners. The per se performance of the parents was found to be good indicator of their general combining ability. Mao et al. (2005) studied the combining ability for twelve traits of sixty early season hybrid rice derived from crossing six cytoplasmic male sterile lines with ten restorer lines. Additive effects were observed for plant height, panicle length, effective panicle number, maximum tiller number, seed-bearing panicle percentage, total number of seeds per panicle, seed set, 1000-grain weight, growth duration, and expected and observed yields. The general combining ability was not correlated with the phenotypic value in the parents nor the specific combining ability of the crosses. However, it was significantly and positively correlated with over standard heterosis of the crosses. Akter (2004) studied combining ability in a six-parent full diallel analysis. Preponderances of additive genetic components for primary and secondary branches per panicle and predominance of non-additive gene action for duration, spikelet sterility, 1000 grain weight, harvest index and yield were observed in her study. She found BRRI dhan29 as a good general combiner for maximum number of morpho-physiological traits. For long panicles with more number of filled grains per panicle BR7166-4-5-3 was the best general combiner. The cross combination BRRI dhan28× BR7166-4-5-3 and BRRI dha29× BR7166-4-5-3-6-3-4 were the best specific crosses which could be exploited to produce prospective genotypes with desirable panicle architecture. On the other hand, the cross BRRI dhan29× BR7166-4-5-3 and BRRI dhan29× BR7166-4-5-3-6-3-4 having superior sca values for most of the physiological characters could be exploited to generate transgressive segregants with good yield potential and physiological superiority. Allah (2004) studied a half diallel cross including five rice varieties namely, Giza 177, Giza 178, Giza 182, IET 1444 and Moroberekan to estimate heterosis and combining ability effects of some root, agronomic, grain quality and grain yield characters under drought conditions. The results revealed that, both general and specific combining ability variances were significant for all studied characters except no. of days to heading and grain yield plant-1 indicating the importance of additive and non-additive genetic variance in the inheritance of these characters and that selection for these traits would be effective in early segregating generations. The highest estimates of GCA effects were found for the cultivar; Giza 178 for most of the studied characters. While IET 1444 and Moroberekan cultivars showed the highest GCA for all root characters studied. The most desirable specific combining ability effects were obtained from the crosses Giza 177×IET 1444 and Giza 182×IET 1444 for no. of days to heading and plant height; Giza 177×Moroberekan and IET 1444×Moroberekan for root characters, and Giza 177×Giza 178 and Giza 178×Giza 182 for grain quality characters. Iftekharuddaula et al. (2004) conducted an experiment in Bangladesh to investigate the per se performance, specific combining ability, heterosis and the interrelationships among them for yield and yield components in an eight-parent half diallel cross in rice. The parental genotypes used in the study were BRRI dhan29, BR4828-54-4-1-4-9, BRRI dhan28, IR8, Amol3, IR65610-38-2-4-2-6-3, Minikit and ZhongYu7, which were chosen for their genetic differences and diverse origins. Analysis of the per se performance of twenty eight hybrid combinations showed that the highest grain yield was from the cross BRRI dhan29×Amol3, due to a high number of panicles and a higher number of filled grains per panicle. The best specific combiners were Amol3×IR65610 for earliness, Amol3×ZhongYu7 for number of panicles plant-1 and number of filled grains panicle-1, BRRI dhan28×Amol3 for 1000-grain weight and BRRI dhan29×Amol3 for grain yield plant-1. There were highly significant positive correlations among per se performance, SCA for almost all the rice characteristics, which clearly suggested that SCA of a hybrid combination could reliably be predicted by per se performance of the cross combinations. Liu et al. (2004) used some fifty one japonica rice cultivars from nine ecotypes to develop crosses in a diallel set. They classified the genotypes into heterotic groups based on general (GCA) and specific (SCA) combining abilities for six traits (number of panicles per plant, number of filled spikelets per panicle, number of spikelets per panicle, seed set, grain weight per plant, and 1000-grain weight). The variance of GCA was higher than that of SCA. The characters were mainly controlled by additive gene effects. The ecotypes of northwest Taiwan and ITA were heterotypic. The relationship between GCA and hybrids was significant. Singh and Singh (2004) made crosses of rice cultivars Dihula, Niwari, Raimunuwa, JR 353, RWR 3-45, Poornima, Vanprabha and Kalinga 3 in all possible combinations in a field experiment. A predominance of additive gene action was observed for all the characters measured. Raimunuwa, JR 353 and RWR 3-45 were good general combiners for grain yield per plant. Dihula, Poornima and RWR 3-45 recorded significant negative general combining ability for number of days to 50% flowering. The cross Raimunuwa×Poornima and Poornima×Vanprabha were the best specific combiners for crop yield and some physiological characters. Verma and Srivastava (2004) made genetically designed analyses for gene governance components, combining abilities, heterosis, inbreeding depression and genetic gains utilizing a seven-parent half-diallel mating design having several locally adapted, traditional varieties or land races from three diverse rice ecosystems. Both additive and non-additive genes were found to control the expression of yield and its associated traits. The ratio of σ g2/σ s2 exhibited greater relevance of non-additive gene actions governing yield and its associated traits. Results revealed that per se performance of the parents is a reflection of their GCA effects in most of the crosses. Among the genotypes NDR 359 ranked as the top general combiner for yield and its associated traits followed by T 21 and IR 24. Higher yield was found to be greatly associated with morpho-economic traits such as harvest index, biological yield, 100-grain weight and productive tiller number per plant. Vanaja et al. (2003) produced twenty eight rice hybrids, from diallel crossing excluding reciprocals among eight parents (Mattatriveni, Hraswa, Mahsuri, Vyttila 3, Kachsiung Sen Yu 338, IR36, IR60133-184-3-2-2 and PK3355-5-1-4), and studied along with the parents for combining ability for yield and seventeen yield components. The study revealed the importance of both additive and non-additive gene effects in governing yield and most of the yield components with preponderance of non-additive gene action for most of the yield components. Additive gene action was found important for 1000-grain weight, second uppermost internodal length and height of plant at harvest. The parent Vyttila 3 was found to be a good general combiner. The hybrids PK3355-5-1-4×Hraswa, Vyttila 3×IR60133-184-3-2-2, Vyttila 3×IR36, Vyttila 3×Mattatriveni and IR36×Mattatriveni showed significant favourable specific combining ability effect for yield and different yield components. Basher (2002) observed importance of both additive and non-additive gene action for plant height, days to 50% flowering, harvest index, and grain yield and additive gene action for 1000 grain weight and preponderance of non-additive gene action. Generally, he found high×high general combiners produced the best specific crosses and he also found a good specific cross evolving from parents having poor general combining ability. Combining ability for eight agronomic important characters of F1 derived from different ecological types in japonica rice and combining ability of their parents were studied by an 8×8 diallel design. In the populations consisted of different ecological types of japonica rice, additive and non-additive genetic effects were equally important for the four characters of plant height, spikelets per panicle, filled grains per panicle and grain yield per plot. For the three characters of heading date, panicle length and effective panicles per plant, additive effects were more important than those of non-additive. And for the character of 1000-grain weight, non-additive effect was more important than that of additive. Xiushui 04 and 3726 were the parents with excellent general combining ability effects and specific combining ability variance for grain yield per plot and yield components. Koshihikari×Xiushui 04 was a superior combination with good specific combining ability effects and overall characters Hong et al. (2002). Mahmood et al. (2002) investigated the nature of gene action and the combining ability behavior of various genotypes in governing the various traits in an experiment involving 8×8 diallel crossing of rice. Twelve agro-physiological characters were included in the study in F1 generation. High additive effects were recorded for plant height, panicle length, productive tillers plant-1 and primary branches panicle-1. The non-additive effects were more pronounced for panicle fertility, days to maturity, shoot dry weight, paddy yield; Na, Ca and K contents of the shoot and K/Na ratio of the shoot. Out of the eight parental lines/varieties studied Jhona-349 and Bas-385, respectively, proved to be the best general and specific combiners in the experiment under salinized soil conditions Sah et al. (2002) performed combining ability analysis using ten fine-aromatic rice germplasms in a partial diallel crossing design to estimate General Combining Ability (GCA), Specific Combining Ability (SCA) and nature of gene action. They observed considerable variation among parents and crosses (F1S) for agronomic parameters. The parental mean for panicle weight (g) was 2.83 g while the F1s was 3.16 g with an average heterosis of 0.3 g (11.3%) and a maximum heterosis of 3.17 g (112.0%). Panical weight (g) being the important yield component, was analysed to estimate GCA, SCA and gene effects. The analysis of variance revealed significant effects (P=0.01) due to GCA and SCA. The GCA estimates of parents ARC10863 (0.77) and IR43850 (0.33) were significant and positive while those of ARC 10796 (-0.20) and Khumal-4 (-0.28) were significant (P=0.01) and negative. The superior crosses with high positive effect on panicle weight were from the above parents. The additive variance (σ2A) estimate was 0.30 and that of dominance variance (σ2D) was 3.79, thereby indicating low genetic heritability (0.064), and presence of both additive and non-additive gene effects on panicle weight of fine-aromatic rice. Sugiono (2002) studied the genetic relationships between grain types and agronomic traits in rice (Oryza sativa L.) in 6×6 diallel cross (excluding reciprocals) involving long, medium, and short-grain cultivars. Combining ability analyses revealed that mean squares due to GCA were larger than those due to SCA, except for heading date. Vsca for heading date, plant height, panicle number, and grains per panicle were larger than Vgca for those traits. Panicle length, grain length, grain width, and 100-grain weight had larger Vgca than Vsca. Their study suggested that ‘Brazos’ was a good combiner for earliness. ‘Starbonnet’ was a good combiner for plant height, grains per panicle, and ‘Starbonnet’ and ‘Lebonnet’were good combiners for grain length. ‘Nortai’ combined well for high panicle number and ‘Nortai’ and ‘S201’ combined well for grain width. ‘M401’ and ‘S201’ were good combiners for 100-grain weight. Bansal et al. (2000) assessed heterosis and combining ability effects in an eleven-parent diallel cross involving eight scented (Dawag, Bindli, N750, Basmati 1A, Basmati 372, Karnal Local, Basmati 1 and Basmati 405) and three non-scented (Pusa 44-33, IET8585 and TNI) rice (O. sativa) stocks. The estimates of GCA and SCA indicated predominance of non-additive gene effects for days to flower, plant height, panicle length, tillers per plant, number of fertile tillers per plant, grain yield per plant, 1000-grain weight, and length to breadth (L:B) ratio. Cultivar Karnal Local showed good GCA for panicle length, tillers plant-1, number of fertile tillers plant-1, grain yield plant-1, and grain L:B ratio. Bindli semi-dwarf mutant Dawag and Basmati 372 showed good GCA for earliness and dwarfing. The study suggested that a greater number of favourable genes can be combined in one genotype to obtain maximum yield without loss in quality characteristics from crosses involving good combining scented parents like Karnal Local, Basmati 1, Basmati 372 selection, and Bindli semi-dwarf mutant. Kwak (1999) studied that the combining abilities of harvest index (HI) and productivity score (PS) from the partial six-parental diallel cross with indica and Tongil type rice varieties. The mean squares of GCA and SCA effects were highly significant for HI, PS, grain yield and straw yield. The variance of GCA effects was much larger than that of SCA effect in all the characters, showing predominance of additive gene actions for these characters. Rice varieties Milyang 42 Suweon 307 and IR356 showed high GCA effects to the direction of selection for harvest index and productivity score, and the cross IR747B2-6×Suweon307 did the highest SCA effects to the direction of selection of grain yield, harvest index and productivity score. Chen et al. (1997) observed that yield plant-1 appeared to be mainly controlled by non-additive effects, while plant height and heading data were mainly conditioned by additive gene effects. Ganesan et al. (1997) reported from a study conducted with twenty eight rice hybrids derived from four early (105-115 days) maturing varieties as lines and seven extra-early (70-90 days) varieties as testers. They observed the importance of both additive and non-additive gene action for yield and its component traits. However, predominance of non-additive gene actions observed for yield and its component traits except plant height. ASD16 among early parents and Heera and AS89011 among extra-early parents were identified as the best combiners. Ali and Khan (1995) estimated heterosis and combining ability in four rice (Oryza sativa) cultivars and their six F1 hybrids. Data was collected on nine yield components. A wide range of variation was noted for all of the characters for GCA and SCA. Non-additive gene action was observed for all of the traits except plant height and panicle length, for which additive gene effects were important. Mean varietal performance was linearly related to GCA values for grain yield and yield components. Basmati 385 proved to be a good general combiner for most of the traits. Based on mean performance coupled with SCA effects of the crosses and varietal GCA effects, crosses Basmati 370×Basmati 385, 4048×Basmati 198 and Basmati 370×Basmati 198 were recommended for pure line development. Lokaprakash et al. (1995) conducted an experiment to study heterosis and combining ability in rice. Information on heterosis and combining ability was derived from data on yield and seven yield components in seven lines and their twenty one F1 hybrids. HP19 and HP11 were the best general combiners for yield and its components. The crosses HP19×HP11, HP19×IET7575, HP11×HP8, HP15×IET7575 and HP8×HP32 were the best specific combinations. In a study with eight varieties and their fifteen hybrids Yuan et al. (1995) observed that the highest GCA was shown by 88122A. Among the restorer lines, R161 and R254 had high GCA. Geetha et al. (1994) information on genetic variance and combining ability is derived from data on yield and six of its components in a 4×4 half diallel involving cultivars ADT38, ADT39, Co45 and White Ponni. Variance due to GCA and SCA was significant. There was a preponderance of additive gene action for plant height, grain length, grain breadth and 100-grain weight, and non-additive gene action for panicles plant-1 and grain yield plant-1. ADT38 and White Ponni were the best combiners for grain yield due to increases in panicles plant-1 and number of grains panicle-1, respectively. White ponni and ADT39 were the best contributing parents for slender grain type. Zhou et al. (1982) in an uncompleted diallel cross of six male sterile lines and five restorers reported that the GCA was more important than SCA in most cases. In the hybrids, each character was influenced by the GCA of both male sterile line and restorer lines and also by the SCA of the combination together designated as “total combining ability”. 2.3 Heterosis Heterosis or hybrid vigor is manifested as improved performance for F1 hybrids generated by crossing two inbred parents. Heterosis can be defined quantitatively as an upward deviation of the mid-parent, based on the mean values of the two parents (Johnson and Hutchinson, 1993). The term heterosis was coined by Shull (1908) for quantitative measure of superiority of F1 over its parents and is usually referred to relative heterosis. The phenomenon of heterosis has been a powerful force in the evolution of plants and has been exploited extensively in crop production (Birchler et al. 2003). Heterosis may be positive or negative. Depending upon breeding objectives, both positive and negative heterosis are useful for crop improvement. In general, positive heterosis is desired for yield and negative heterosis for early maturity. Heterosis is expressed in three ways, depending on the criteria used to compare the performance of a hybrid. The three ways are: midparent, standard variety and better parent heterosis. However, from the plant breeder’s viewpoint, better parent and/or standard variety is more effective. The former is designated as heterobeltiosis and the latter as standard heterosis (Fanseco and Peterson, 1968). Although rice is a naturally self-pollinated crop, strong heterosis is observed in their F1 hybrids. In rice, heterosis was first reported by Jones (1926) who observed marked increase in clum number and grain yield in some F1 hybrids in comparison to their parents. Since then, several rice researchers have reported the occurrence of this phenomenon for various agronomic traits e.g. Yield, grain weight, grains per panicle, panicle per plant, plant height, days to flower etc. (Virmani et al. 1981). The literature reported on heterosis until 1987 is summarized by Virmani et al. (1981) and further updated by Virmani and Edwards (1983). These reports provided evidence for the occurrence of significant heterosis and heterobeltiosis for various agronomic characters. However, estimate of standard heterosis could not be made in most of the studies because the crosses were not made to develop F1 hybrid varieties. Majority of reports were based on a few crosses which were perhaps made for conventional rice breeding program and did not necessarily involved the selection of parents to manifest strong heterosis, despite this limitation, the extent of heterosis was high in some cross combination. The most comprehensive work on the study and exploitation of intrusion has been reported from the Peoples Republic of China were F1 hybrids have been developed, released and are cultivated on commercial scale (Lin and Yuan, 1980). The success of hybrid rice in China encouraged the IRRI to revive research on hybrid breeding in 1979 (Anonymous, 1989) to exploit its potentials and problems outside China. The literatures available so far on heterosis in rice after 1980 have been reviewed here as follows: Patil et al. (2011) studied heterosis in rice in a set of 10×10 diallel cross excluding reciprocals. The analysis of variance indicated highly significant differences among genotypes and hybrids for all the characters. The magnitude of heterosis varied from cross to cross for all the characters studied. High heterotic effects were observed for days to 50% flowering, productive tillers per plant, plant height, panicle length, grains per panicle, test weight, grain yield per plant and amylose content. They observed from the study that heterotic response for grain yield per plant was mainly due to high heterotic response observed for productive tillers per plant, panicle length, grains per panicle and test weight. Sen and Singh (2011) determined the extent of heterosis over better parent and standard checks using twenty five genotypes including check varieties in order to identify promising hybrids with high mean performance and high magnitude of heterosis for yield and yield components. Three boro rice cultivars were selected as female and single crosses were made randomly with six high yielding non-boro rice cultivars. Significant difference among genotypes for all characters studied indicated that good amount of variation was present for effective selection. The hybrids in general recorded high mean values as compared to those of the parents for plant height, effective tillers per plant, panicle length, grains per panicle, thousand grain weight and grain yield. These crosses showed marked variation in the expression of heterobeltiosis and standard heterosis for yield and yield components and revealed the existence of considerable heterosis both in positive and negative direction for all the traits. Four hybrids IR64×HUBR2-1, IR64×JAYA, Krishna Hansa×Jaya and Krishna Hansa×BPT 5204 were identified for higher mean performance and high magnitude of heterosis for yield and yield components. The extent of heterosis was studied in a set of thirty six hybrids generated from a 9×9 diallel mating design excluding reciprocals involving nine promising genotypes HPR1164, HPR2047, China 988, VL91–1754, VL93–3613, VL93–6052, IR57893-08, VL Dhan221 and JD8 of diverse nature maintained in pure form. Out of thirty six hybrids, twelve showed significant heterosis for majority of the traits identified. In general, the estimates of heterosis values were low for quality traits when compared with yield and morpho-physiological traits. Nine hybrids exhibited positive and significant heterosis over standard check but seven over better parents for grain yield plant−1. Standard heterosis and heterobeltiosis for grain yield ranged from 14.12 to 65.32% and 22.59 to 65.32%, respectively. The hybrid HPR 2047×JD8 which recorded 65.32% higher grain yield over both better parent and standard check was identified as the best hybrid for exploiting hybrid vigor. In order of merit (standard heterosis) HPR 2047×JD8 (65.32%), VL93–3613×IR 57893-08 (60.05%) and VL93–6052×VL Dhan (54.94%) were recorded to be the three best performing hybrids for grain yield plant−1. The higher yield by these hybrids could be due to more panicle length, net assimilation rate, leaf area index, dry matter, harvest index and 100 grain weight. These crosses may be used in future breeding programmes for development of high yielding hybrids and varieties. The cross combinations VL91–1754×VL93–3613, VL91–1754×JD8 and VL933613×JD8 recorded desirable heterosis over both better parent and standard check for grain quality as well as grain yield plant−1 (Kumar et al. 2010). Rahimi et al. (2010) made quantitative valuations of observed heterosis for eleven traits of fifteen F1 hybrids generated by half diallel crosses of six diverse rice cultivars (Domsefid, Hashemi and Binam, three Iranian local cultivars; Dorfak, Kadous and IR30, three improved cultivars). Assessment of standard heterosis based on check variety Dorfak showed that there was significant heterosis for all the traits studied in the fifteen hybrids. For grain yield, the Dorfak×Domsefid cross had the highest heterosis. This hybrid had good heterosis values for many traits such as growth period, reproductive period and 1000-grain weight and was recommended as the most promising combination for developing high yielding hybrid rice varieties. Sharifi et al. (2010) studied genetic components, the mode of inheritance and heterosis of various physiological traits by using seven-parent diallel mating design including reciprocals. The highest value of relative heterosis was observed for panicle exsertion and grain yield. Alam (2009) found the highest desirable negative heterosis in BRRI dhan28× BR7166-5B-1 for development of the dwarf plant and the cross BRRI dhan29× BRRI dha47 for days to flowering. The highest positive heterosis for panicle length and yield were observed in BRRI dhan28× BR7166-5B-1-Ran-1. Both positive and negative heterosis was found for fourteen characters in eight F1s studied by Hassan (2009). The cross combinations IR58025A×KataribhogR and IR62829A×ChiniguraR showed the highest significant heterotic effects for earliness of flowering. For maturity of spikelet, plant height, grains per panicle, unfilled grains per panicle and grain yield per plant the cross IR58025A×KataribhogR showed the highest significant heterotic effect. The cross combination IR68888A×SakorkhoraR showed the highest significant heterotic effect for panicles per plant, thousand grain weight, straw yield per plant and panicle length. The cross IR62829A×chiniguraR showed the highest significant heterosis for harvest index over mid parent. In a study of estimation of heterosis with nine CMS and three restorer lines of rice for seventeen yield and yield contributing traits mid-parent and better parent heterosis of most of the crosses were significant for most of the characters studied either in positive or in negative direction. Desirable and significant mid-parent heterosis was observed in thirteen and seven cross combinations, respectively for grain yield and most of its related traits. Considering more than 20% mid-parent and better parent heterosis for grain yield along with most of its related traits, nine and five cross combinations respectively were identified as good heterotic combinations over mod parental and better parental value (Islam, 2009). Roy (2009a) found majority of hybrids exhibited significant heterosis positive heterosis over better parents for plant height, tiller number, yield per tiller, panicle length and 100-grain weight. On the other hand, the similar extents of heterosis for many hybrids were observed in negative but desirable direction for days to 100% flowering and days to first flowering. Saleem et al. (2008) reported that heterosis and heterobeltiosis ranged -21.06 to 60.13 percent and -33.34 to 42.99 percent respectively for leaf area, -3.94 to 12.98 percent and 3.25 to 32.21 percent for plant height, 17.88 to 53.10 percent and -25.90 to 23.97 percent for panicle density, -8.80 to 32.43 percent and -24.32 to 19.29 percent for harvest index, 7.54 to 58.77 percent and 12.88 to 104.37 percent for biological yield per plant. The results indicate that improvement of grain yield can be efficient through making some compromises within limits among morpho-physiological yield traits. Torres and Geraldi (2007) estimated some useful parameters which can be used to investigate the genetic control of agronomic characters in crosses combining cold tolerance and productivity. They used a partial diallel design in crosses between six tropical indica rice cold susceptible genotypes (group 1) and seven japonica or indica×japonica cold tolerant rice genotypes (group 2). Parents and crosses were evaluated for agronomic characters under field conditions in two different experiments in 2005. The results showed significant mid-parent heterosis for all characters (plant height, tiller number, days to 50% flowering, panicle length, grains per panicle, sterility, and one-hundred grain weight). Five cytoplasmic male sterile lines and ten testers of diverse origin were crossed in Line×Tester fashion to obtain fifty hybrid combinations. Many of the hybrids, IR 688886A×ADT 41, IR 688897A×ADT 41, IR 688886A×ASD 19, IR 68885A×ASD 16 and IR 58025A×ACK 03002 exhibited significantly negative heterobeltiosis as well as standard heterosis for days to flowering and maturity indicating the possibility of exploiting heterosis for earliness. The best hybrid was IR 68885A×White ponni, which showed standard heterosis and heterobeltiosis for panicle length, spikelets per panicle, grains per panicle, straw yield and grain yield. Relative heterosis for grain yield per plant ranged between -69.17 to 243.21 coupled with significant heterobeltiosis from -75.71 to 219.75. The standard heterosis ranged from -73.71 to 129.16 for grain yield per plant. Based on per se performance and standard heterosis the hybrids IR 68885A×White ponni, IR 688886A×White ponni, IR 688886A×ADT 41and IR58025A×CO 43 were identified as the superior hybrids for the characters of panicle length, grains per panicle and grain yield Malini et al. (2006). Studies performed by Raju et al. (2006) indicated that the amount of heterosis for leaf area index (LAI) at tillering stage, crop growth rate (CGR) from tillering to heading, harvest index, 100-grain weight and grain yield plant-1 was high. Kumar et al. (2005) made crosses of three cytoplasmic male sterile (CMS) lines, i.e. IR 58025A, IR 68886A and IR 68888A with seven scented traditional and improved cultivars, i.e. Indira Sugandhit Dhan, Madhuri-11, Loktimachi, Tarori Basmati, Govind Bhog, RP-2235-97-82-19-55 and JR-511. The resulting twenty one F1 hybrids and the parents, including the control cultivar Pusa Basmati, were evaluated to assess the extent of heterosis for number of productive tillers, panicle length and grain yield per plant. Among the twenty one hybrids, 70% gave higher yields than the respective mid-parents and 60% gave higher yields than the respective better parents. Approximately 75% of the hybrids were better than the control cultivar. Five high-yielding hybrids, i.e. IR 58025A/Madhuri-11, IR 68888A/JR-511, IR 58025A/RP-2235-97-82-19-55, IR 68888A/Indira Sugandhit Dhan and IR 68888A/Madhuri-11, showed more than 35% standard heterosis, indicating their potential for achieving high yields in scented rice. Liu et al. (2005) studied the heterotic ecotypes of japonica rice by analysis of heterosis of the F1 hybrids created in a diallel set of cross between nine ecotypes. The heterotic patterns of japonica rice were also established. The japonica ecotypes of North China, Taiwan, Japanese and Korea rice were considered as heterotic ecotypes. The heterotic patterns were: Korea ecotype×Japanese ecotype, Northwest ecotype×American ecotype and Taiwan ecotype×Japanese ecotype. The Northwest japonica ecotypes had heterotic genes for number of grains; Taiwan, Japanese and Korean ecotypes for number of panicle; and IRAT ecotypes for grain weight. Raju et al. (2005) made an investigation to study heterosis and inbreeding depression in rice for yield components and kernel characteristics involving seven parents, twenty one F1s and corresponding F2s. High degree of heterosis was observed by them for days to 50% flowering, productive tillers plant-1 and grain yield plant-1; whereas, in case of filled grains per panicle and 100-grain weight, it was low. For kernel length heterosis was observed only on mid-value and it was in negative direction for kernel shape. Suh et al. (2005) grown six hybrid combinations between the Dian type cytoplasmic male sterile (CMS) lines, Hwayeongbyeo A and Nonghobyeo A, and the three restorer lines, YA3414-3-1-1-1-1, YA3416-1-1-1-wx3-1 and YA3428-2-1-1-2-1 at the Experimental Farm of Yeungnam University in 2004 in order to investigate the heterosis for yield in Dian-type hybrid rice with the backgrounds of the Korean rice varieties. The number of panicles per hill and spikelet fertility of the hybrids in Hwayeongbyeo A combinations, and the number panicles per hill and spikelet fertility of the hybrids in Nonghobyeo A combinations were increased significantly than those of their parents respectively. However, one thousand grain weight of the hybrids was almost same as their parents in the six hybrids. The heterobeltiosis of milled rice yield of the hybrids ranged 114-128% in Hwayeongbyeo A combinations, and 175-208% in Nonghobyeo A combinations. The standard heterosis over the standard rice variety Hwayeongbyeo ranged 114-128% in the Hwayeongbyeo A combinations and 107-127% in the Nonghobyeo A combinations. Alam et al. (2004) studied the genetic basis of heterosis through mid-parent, standard variety and better parent for eleven quantitative traits in seventeen parental lines and their ten selected hybrids in rice (Oryza sativa L.). In general the hybrids performed significantly better than the respective parents. Significant heterosis was observed for most of the studied characters. Among the ten hybrids, four hybrids viz., 17A×45R, 25A×37R, 27A×39R, 31A×47R, and 35A×47R showed highest heterosis in 10-hill grain yield per 10-hill. Allah (2004) studied a half diallel cross including five rice varieties namely, Giza 177, Giza 178, Giza 182, IET 1444 and Moroberekan to estimate heterosis and combining ability effects of some root, agronomic, grain quality and grain yield characters under drought conditions. The most desirable heterosis as deviation from better-parent were obtained from the crosses Giza 177×IET 1444 and Giza 182×IET 1444 for no. of days to heading and plant height; Giza 177×Moroberekan and IET 1444×Moroberekan for root characters, and Giza 177×Giza 178 and Giza 178×Giza 182 for grain quality characters. Iftekharuddaula et al. (2004) conducted an experiment in Bangladesh to investigate the per se performance, specific combining ability, heterosis and the interrelationships among them for yield and yield components in an eight-parent half diallel cross in rice. The parental genotypes used in the study were BRRI dhan29, BR4828-54-4-1-4-9, BRRI dhan28, IR8, Amol3, IR65610-38-2-4-2-6-3, Minikit and ZhongYu7, which were chosen for their genetic differences and diverse origins. Analysis of the per se performance of twenty eight hybrid combinations showed that the highest grain yield was from the cross BRRI dhan29×Amol3, due to a high number of panicles and a higher number of filled grains per panicle. The nature and extent of heterosis in both types (over mid parent and better parent) were found to vary depending on cross combinations and the characters under study. Based on mean values of mid- and better parental heterosis in twenty eight crosses, the highest magnitude of heterosis was obtained in grain yield plant-1 followed by number of panicles plant-1 and number of filled grains panicle-1. Most of the cross combinations produced significantly positive heterosis for grain yield plant-1. The maximum heterosis for grain yield plant-1 was displayed by the hybrid BR4828×Amol3, followed by IR65610×Minikit. Lin and Yang (2004) investigated the heterosis and correlation of main yield characters of indica hybrid rice by using a 4×8 incomplete diallel cross design with four sterile lines and eight restorer lines in an experiment carried out in 1999. Results showed that the main yield characters had significant heterosis in indica hybrid rice. Vanaja and Babu (2004) made crosses among eight genetically diverse high yielding rice cultivars selected from clusters formulated through Mahalanobis D2 statistics, in a diallel fashion. The parents and crosses were evaluated and heterosis for yield and its principal components was estimated in experiments. Results suggest that yield increase was largely due to significant and favourable heterosis in yield components, i.e. number of spikelets panicle-1, panicle length, leaf area plant-1 (at maximum tillering stage) and number of panicles m-2. Five top heterotic crosses over their mid and better parents for each trait were identified. Satish and Ramaiah (2003) evaluated fifteen rice crosses and their six parents to estimate heterosis for yield and yield components, certain physiological and grain quality characters. The extent of heterobeltiosis and heterosis ranged from -25.63 to 5.27% and -4.56 to 17.89% for grain yield, respectively. Higher heterosis for grain yield was manifested through positive heterosis for number of productive tillers per hill, test weight, number of grains per panicle and kernel weight. The top heterotic combinations identified for yield evaluation were: ARC 5780×BPT 1768, ARC 5780×NLR 33641, BPT 5204×BPT 1768, BPT 5204×BPT 4358, MTU 2067×BPT 4358 and NLR 33641×BPT 4358. Bashar (2002) found significant positive desirable heterosis over mid-parent for eight yield related characters including panicle length and harvest index. Hong et al. (2002) studied heterosis of eight agronomic important characters of F1 derived from different ecological types in japonica rice and combining ability of their parents by an 8×8 diallel design. Results indicated that both positive and negative heterosis existed in all the eight characters. On the average of fifty six F1 combinations, grain yield per plot showed 19.5% heterosis over high parents, 87.5% of the combinations showed higher grain yield per plot than that of the high parents, and for the other seven characters, F1 values were between mid-parent and high parent. Verma et al. (2002a) intensively assessed twenty-one F1 and F2 lines obtained from half-diallel (7×7) cross combinations among seven ecotypes from two rice ecosystems (i.e. upland/irrigated ecotypes, and rainfed lowland ecotypes) for five physiological traits (grain filling period, plant height, flag-leaf area, biological yield, harvest index) along with grain yield for their contributions to heterosis and inbreeding depression (ID). Significant economic heterosis was recorded in fifteen hybrids, e.g. three for harvest index; three for biological yield; three for flag leaf area; three for grain filling period; and three for plant height. Heterosis estimates were attributable to genetic interactions arising from both additive and high degree of dominance or epistatic interactions or both for one or more contributing physiological traits. Two physiological traits, biological yield and harvest index, followed by flag leaf area remained as major contributors to heterosis. High magnitude of heterosis accompanied by high ID for biological yield, harvest index and grain filling period further revealed the preponderance of epistasis or non-additive interallelic genes. Verma et al. (2002b) evaluated seven diverse rice ecotypes, along with their F1 and F2 populations, in 1994-95 for heterosis and inbreeding depression for five important quantitative traits (days to 50% flowering, number of productive tillers per plant, number of spikelets per panicle, 100-grain weight and grain yield per plant). Analysis of variance showed that variance due to genotypes was highly significant for all the traits studied. The majority of the crosses showed significant heterobeltiosis and standard heterosis over for grain yield and number of productive tillers per plant, and 100-grain weight. Only a few crosses showed heterobeltiosis for days to 50% flowering and spikelets per panicle. Bansal et al. (2000) assessed heterosis and combining ability effects in an eleven-parent diallel cross involving eight scented (Dawag, Bindli, N750, Basmati 1A, Basmati 372, Karnal Local, Basmati 1 and Basmati 405) and three non-scented (Pusa 44-33, IET8585 and TNI) rice (O. sativa) stocks. Heterosis was significant for grain yield and its component characters in most of the hybrids. Maximum heterosis for grain yield per plant was obtained in the cross between Karnal Local, a high quality scented rice and a semi-dwarf scented parent Dawag. Dwivedi et al. (1998) studied moderate to high heterosis for yield and ten related characters in forty five crosses involving six indica (I) and four tropical japonica (J) varieties of rice in three (E1 – optimum sowing and high fertility, E2 – optimum sowing and optimum fertility and E3 – late sowing and high fertility) environments. Trends in magnitude of heterosis for grain yield and plant height were I×J > I×I > J×J and for days to 50% flowering I×J > J×J > I×I. Estimates of standard heterosis for grain yield were -64.5 to 146.1% in E1, -70.4 to 82.2% in E2 and -67.2 to 63.8% in E3. Environment E1 seemed to be more favourable for higher heterosis expression than others. Higher heterosis in yield also accompanied heterosis in panicle number, dry matter and spikelet and grain number per panicle. Most estimates for days to flowering were negative. Heterotic I×J hybrids also recorded maximum heterosis for earliness. Moderate to low standard heterosis for plant height across environments (2.0-13.7%) was recorded. Hybrids were identified in specific environments for direct exploitation in hybrid breeding. Hybrids B4116×Sarjoo 52, B4122×Pant Dhan 4 and B4122×Narendera 359 were more stable than others over the three environments.
Information on heterosis is derived from data on twelve yield and quality related traits in six rice parents and their thirty hybrids grown during kharif 1993. IR50×ADT41, ADT41×IR50, ADT41×ADT39 and ADT39×IR50 were found to be promising combinations for heterosis breeding since they registered high standard heterosis for grain yield in addition to one or two other yield component characters Geetha et al. (1998).
Singh and Zaman (1998) studied fifteen F1s along with six parents from a 6×6 half diallel involving diverse parents for physiological efficiency and reported that the cross, Jaya×Swarnaprabha, which exhibited a high heterobeltiosis for biological yield (54.9%), also showed a significant positive heterosis for grain yield, HI, CGR, LAI and leaf area/plant.
Chen et al. (1997) made reciprocal crosses between five indica thermo-sensitive genic male-sterile lines (TGMS) and six japonica wide compatibility varieties (WCVs). Grain yield per plant, plant height, heading date and other yield components were investigated in sixty F1 hybrids and their eleven parents, using indica hybrid Shanyou 63 as control. TGMS×WCVs and WCVs×TGMS hybrids showed 13.9 and 20.4% heterosis for grain yield plant-1, and 26.8 and 29.9% heterosis for spikelets panicle-1, respectively, but had negative heterosis for fertility, plant height and days to heading.
The study on the selected seven crosses revealed that heterosis over the mid and better parent were negative for days to panicle emergence and positive for panicles plant-1, grain yield plant-1, dry matter production and harvest index. The superiority of ASD16 and Heera among parents and ADT36×Kalyani II, ADT36×AS89011 and IR50×Heera among hybrids was confirmed based on the per se performance and heterosis over better parent in the subsequent season (Ganeasan et al. 1997).
Ali and Khan (1995) estimated heterosis and combining ability in four rice (Oryza sativa) cultivars and their six F1 hybrids. Data was collected on nine yield components. Heterosis was positive and significant for all of the characters except plant height, percentage filled spikelets and spikelet density. Heterobeltiosis was significant and positive over the better parental value in most of the crosses for number of tillers plant-1, panicle weight and grain yield plant-1.
Lokaprakash et al. (1995) conducted an experiment to study heterosis and combining ability in rice. Information on heterosis and combining ability was derived from data on yield and seven yield components in seven lines and their twenty one F1 hybrids the crosses HP19×HP11, HP19×IET7575, HP11×HP8, HP15×IET7575 and HP8×HP32 showed high overall better-parent heterosis.
Thirty-six hybrids derived from nine high yielding and widely adapted cultivars with parental lines and the standard variety, Jaya, were evaluated to determine the nature and extent of heterosis for seven characters including grain yield per plant. Most of the crosses manifested significant heterosis for grain yield and panicle number per plant, panicle length, 1000-grain weight, grains per panicle, and days to maturity. Very few hybrids manifested significant heterosis for plant height. The range of heterosis for seed yield per plant was -96.7 to 258.2% over the better parent, -96.1 to 268.2% over midparent and -96.3 to 301.6% over the standard variety. The highest heterotic effects for grain yield were observed in the crosses Prasad×PP72, Govind×PP72, Govind×Sita, Prasad×Mahsuri and Govind×Jaya, which were, respectively, 258.2, 216.8, 177.8, 161.6 and 108.8% over better parent, 268.2, 250.8, 178.1, 204.2 and 119.4% over midparent, and 301.6, 255.2, 151.6, 285.0 and 108.8% over the best control variety Jaya. A comparison between heterotic and non-heterotic hybrids for yield and its components indicated that heterosis for grain yield was due to two or more direct yield contributing characters Pandey et al. (1995).
Reddy and Nerkar (1995) studied F1 heterosis over midparent (MP) and better parent (BP) and F2 inbreeding depression (ID) in eight crosses of rice for grain yield plant-1 and its four component traits, i.e., plant height, effective tillers plant-1, filled grains panicle-1 and 1000-grain weight. Highly significant and positive heterosis for grain yield over MP and BP was expressed by four hybrids. Such high grain yield heterosis was due to additive heterotic effect of one or more component traits. In all the cases, heterosis for number of effective tillers was observed to be the major contributor and number of filled grains panicle-1 to a lesser extent, to grain yield heterosis. High heterosis accompanied by high ID for effective tillers plant-1 and filled grains panicle-1 indicate that non-additive gene effects govern the inheritance of these traits.
Shen and Xue (1995) used six indica cytoplasmically male sterile lines and seven japonica widely compatible varieties to make a set of incomplete diallel crosses. Information on heterosis was determined from data on yield-related characters in parents and hybrids in field experiments carried out at three different sowing dates (20 May, 10 June and 30 June) in 1993. Significant differences in heterosis were observed under the three sowing dates. Heterosis over both the mid-parent and better-parent based on population means decreased with the postponement of planting for most of the characters. In the hybrids, plant height, flag leaf length and heading time increased at under early sowing but total number of grains per main panicle decreased at later sowing dates.
Yuan et al. (1995) observed significant heterosis for grain weight and dry plant weight in the F1 in a study of eight varieties and their fifteen hybrids. Among the yield-related traits, grain number per panicle and number of well-filled grains showed the greatest heterosis.
A study was made to assess the nature and extent of heterosis and heterobeltiosis for yield and its components in a 7×7 diallel excluding reciprocals (Lokaprakash et al. 1992). Heterosis for all the charcters was evident in most of the hybrids studied. Heterosis for yield was mostly due to simultaneous heterosis for number of productive tillers, panicle weight, panicle length, fertile spikelets per panicle, 1000-grain weight and harvest index.
For harvest index in rice both positive and negative heterosis was reported by several workers, Peng and Virmani (1991), Virmani et al. (1981), and Nijaguna and Mahadevappa (1983).
Saini et al. (1974) studied heterosis for yield and four yield components, viz. panicle number per plant, panicle length, spikelet per panicle and 1000-grain weight, in fifteen crosses involving six varieties of rice, Oryza sativa L. Positive and significant heterosis for yield was observed in eleven and eight crosses, over the mid-parent and the better parent, respectively. Hamsa×Hybrid 27 gave the highest heterosis both over the mid-parent (156.23%) and the better parent (136.38%) followed by Jaya×Norin 18 with respective values of 155.18 and 56.29%. The heterosis for yield was due to simultaneous heterosis for a number of yield components.