Genetic effects of combining ability studies for yield and fibre quality traits in diallel crosses of upland cotton ( Gossypium hirsutum L . )

The present study was carried out to investigate the diallel analysis in upland cotton (Gossypium hirsutum L.) by involving seven parents and their 42 cross combinations. All the characters were predominantly controlled by additive gene action except number of sympodial branches, single plant yield and bundle strength. Non additive gene action was controlled for the characters sympodial branches, single plant yield and bundle strength. The parent BW4-1 had superior per se performance for single plant yield, boll weight, lint index, number of bolls per plant, seed index, and micronaire followed by MCU 13 which had recorded high per se for single plant yield, lint index, seed index and number of sympodial branches. The parent TCH 1726 exhibited positive and significant gca effect for single plant yield, ginning percent, number of sympodial branches, boll weight and plant height. Based on high per se performance and high gca effect, the parent MCU 13 was considered as best general combiner as it had significantly greater values for single plant yield, number of sympodial branches, lint index and seed index. The hybrids KC 2 × TCH 1726 and TCH 1705 × MCU 3 had recorded high per se for single plant yield, number of bolls per plant and number of sympodial branches. Based on sca effects, KC 2 × TCH 1726 showed significant sca effects for single plant yield and number of sympodial branches.


INTRODUCTION
Cotton is an important fiber, crop and plays a vital role as a cash crop in commerce of many countries.Cotton production, processing and trade in cotton goods provide employment to about 60 million people in India.It provides fibre for textile industry, cellulose from its lint, oil and protein rich meal from its seed (Ashokkumar and Ravikesavan, 2011).The development of new variety with high yield and fibre quality is the primary objective of all cotton breeders.The first step in a successful breeding program is to select appropriate parents.Diallel analysis provides a systematic approach for detection of appropriate parents and crosses in terms of investigated traits.Diallel analysis has been widely used by plant breeders in the selection of parents and crosses in the early generations (Marani, 1963;Green and Culp, 1990;Islam et al., 2001;Kiani et al., 2007;Karademir and Gencer, 2010).The analyses of diallel crosses contain further information on the nature of the predominant gene action in traits of major agronomic importance, besides estimates of general (GCA) and specific combining ability (SCA).
Information pertaining to the different types of gene action, relative magnitude of genetic variance and combining ability estimates are important and vital parameters to mold the genetic makeup of the cotton crop (Subhan et al., 2003).This important information could prove an essential strategy to the cotton breeder in the screening of better parental combinations for further enhancement.Several studies which reported variation in yield and fibre quality components were controlled by genes acting additively and non-additively.Studies of Shakeel et al. (2001), Ahuja and Dhayal (2007) and Ashokkumar et al. (2010) revealed that boll weight, number of bolls, and seed cotton yield were influenced by the genes acting non-additively and in contrast studies of Khan and Idris (1995) and Kumaresan et al. (1999) indicated that both additive and non-additive gene effects were important for controlling number of bolls and seed cotton yield.However, Lukange et al. (2007) discovered additive gene effects for microanire value and fibre strength and non-additive gene action for fibre length.Non-additive gene action for fibre length, fibre strength and micronaire value have been reported by Hassan et al. (1999Hassan et al. ( , 2000)), Ahuja and Dhayal (2007), Ashokkumar and Ravikesavan (2008) and Preetha and Raveendran (2008).Combining ability studies helps in selection of suitable parents for further exploitation.In order to achieve high production and productivity, the breeder should have thorough knowledge of the genetic potential of germplasm, the nature of gene action involved in promoting heterosis for yield and combining ability of parents.Information on general and specific combining ability for yield and its components will prove very useful in selection of appropriate parents for development of hybrids.Diallel analysis is a technique employed to gain information on hybrid vigour, combining ability and nature of gene action from the study of first generation itself.Among the several biometrical, methods developed to identify superior parents for heterosis breeding; the diallel analysis has received considerable attention.Therefore, present study objective was: 1) to understand the gene action for yield and fibre quality characters; 2) to estimate the GCA and SCA effects of parents and their 42 F 1 crosses, to obtain information combining ability potential as to develop hybrid with improved yield along with fibre quality traits through diallel analysis.

Genetic material
A field experiment was conducted to evaluate the growth, yield and fiber quality traits in upland cotton genotypes (Gossypium hirsutum L.).The genotypes consisting of Narasimha, TCH 1726, TCH 1705, KC2, MCU13, BW4-1, and MCU 3 were obtained from Department of Cotton, Tamil Nadu Agricultural University, Coimbatore, Tamil Nadu, India.

Experimental design and filed procedures
The cotton genotypes were evaluated in randomized block design (RBD) with three replications at Cotton Breeding Station, Tamil Nadu Agricultural University, Coimbatore, and Tamil Nadu in India.During winter 2006 to 2007, seven parents were raised in a crossing block.Each genotype was sown in 20 rows of 6 m length in crossing block with a spacing of 90 × 45 cm.Crosses were made between parents in a diallel mating design.The conventional hand emasculation and pollination were done, and crossed bolls were collected separately and ginned to obtain F 1 seeds.During winter 2007 to 2008, seven parents and forty two hybrids were raised along with the standard check with three replications.The diallel analysis was performed as model 1 and method 1 suggested by Griffing (1956).

Sampling trait measurements and analysis
In each genotypes and their cross combinations, data were recorded on five randomly selected plants per replication for twelve characters namely, days to boll bursting, number of sympodia per plant, plant height at maturity, number of bolls per plant, boll weight, lint index, seed index, ginning percent, single plant yield, 2.5% span length, elongation percent and fiber fineness.Quality parameters were analyzed by high volume instrument (HVI).Statistical analysis was carried out by using the mean values over five sample plants through INDOSTAT package.

Analysis of variance
Analysis of variance showed highly significant differences due to genotypes for all the traits indicating the presence of sufficient variability in the experimental materials (Table 1).Significance of variance in parents versus hybrids interaction provided adequacy for comparing the heterotic expression for all the characters except boll weight and micronaire.Parents and hybrids showed significant differences between all the characters studied except single plant yield and bundle strength.Ashokkumar and Ravikesavan (2013) observation for all the characters were significantly differenced with parents and hybrids in upland cotton, and it confirmed our results.GCA and SCA variances showed significant for all the characters studied except number of sympodia per plant and boll weight for GCA and SCA, respectively.Reciprocals were significantly differenced for all the traits (Table 2).

Effects of gene action
The higher SCA than GCA of a character indicates the preponderance of non-additive gene action.Additive gene action provides fixable effects, and the non-additive gene action results are non-fixable.If GCA variance was greater than SCA for the particular character indicates preponderance of additive gene action.Present study analysis of combining ability revealed that the variances for the GCA were larger than those for all the traits except number of sympodial branches, single plant yield, bundle strength and micronaireindicating the preponderance   cotton yield, and Miller and Marani (1963) for fiber length.
In this present study, non-additive gene action was determined for the characters number of sympodial branches, single plant yield, bundle strength and micronaire.Studies of Shakeel et al. (2001), Ahuja and Dhayal (2007), Preetha and Raveendran (2008) and Ashokkumar et al. (2010) revealed that seed cotton yield were influenced by the genes acting non-additively, and contrast studies of Khan and Idris (1995) and Kumaresan et al. (1999) indicated that both additive and nonadditive gene effects were important for controlling seed cotton yield.Non-additive gene action for fibre strength was reported in earlier studies of Ashokkumar and Ravikesavan (2008), and it is also confirmed by our results.

Per se performance and GCA effects
The 7 parents and 42 F 1 hybrids were used in this study, and varied significantly for each yield components and fibre quality parameters.Information on the per se performance, and nature of general combining ability of characters is necessary for selection of suitable parents for developing hybrids.The objectives appear to be realizable only when the parents are evaluated for the potentialities like per se performance and combining ability attributes.Therefore, the present study was aimed for their mean per se performance and general combining ability effects.Parental genotypes had significant variation for all the traits.Mean per se performance of parents were presented (Table 3).Genotype BW4-1 had superior per se performance for single plant yield (119 g/plant), boll weight (4.7 g/boll), number of sympodial branches per plant (17), lint index (6.76g), number of bolls per plant (27), seed index (10.27g), and micronaire (4.27 ug/inch) followed by MCU 13, which recorded high per se for single plant yield, lint index, seed index and number of sympodial branches.The parent TCH 1705 recorded high per se performance for ginning percent, number of sympodial branches, number of bolls per plant and micronaire value.The parent TCH 1726 recorded high per se for bundle strength and boll weight.Narasimha recorded the highest per se performance for 2.5% span length and number of sympodial branches.High mean values remain as a selection index in the choice of parents and the parents possessing high per se performance will result in superior hybrids.Therefore, these parents can be exploited in hybridization for improving this character through pedigree breeding.
General combining effects for plant height, number of sympodia per plant, boll weight, number of bolls per plant, single plant yield, seed index, 2.5% span length and bundle strength were presented in Tables 4 to 8).For gca effects, the parent MCU 13 had recorded high gca effect for single plant yield, number of sympodial branches, number of bolls per plant, lint index, seed index      Manickam and Gururajan (2004), and high gca effect in desirable direction for a particular character indicates the presence of additive genes for that character in the parent, it could be expected that when the parents possessing high gca effects were combined; larger proportion of progenies would have high per se value for the character concerned facilitating easy selection for the character.Based on the high per se performance and high gca effect, the parent MCU 13 was considered as best general combiner as it had significantly higher values for single plant yield, number of sympodial branches, lint index and seed index.The parent BW4-1 had recorded high per se performance and gca effects for lint index and boll weight.For 2.5% span length, MCU 3 showed better expression for high per se and gca effects.Earlier findings like Arumugampillai and Amirthadevarathinam (1998) reported that identification of parents for breeding programme based on either per se performance or gca effects alone was misleading in the selection programme.
In the present study, considering gca effects and per se performance together, the parents MCU 13, TCH 1726, MCU 3, BW4-1 were selected as the best, since these were having high mean values for four, two, two, two traits and also good combining ability for six, five, four and seven yield component traits, respectively.None of the hybrids were found to be excellent general combiner for all the traits.Hence, it would be desirable to have multiple crosses and subject them to selection in segregating generations to detect superior genotypes with high yield and quality traits.The best criteria for evaluating the hybrids are based on per se performance.However, sca effects, and hybrid vigour of the crosses are also considered frequently in cases where a nonadditive component of genetic variance predominate the inheritance.The superior hybrids were selected based on high per se performance, sca effects and heterosis for each of a trait.In the present investigation, hybrids KC 2 × TCH 1726 and TCH 1705 × MCU 3 recorded high per se for single plant yield, number of bolls per plant and number of sympodial branches.Hybrid BW4-1 × TCH 1726 recorded high per se effects for ginning percentage.The hybrid KC 2 × TCH 1726 recorded high per se effects for single plant yield and number of bolls per plant.The parent BW4-1 × TCH 1705 recorded high per se for seed index.The hybrid BW4-1 × TCH 1726 had shown per se for ginning percent.

Specific combining ability effects
Specific combining ability effects for plant height, number of sympodial branch per plant, number of bolls per plant, boll weight, single plant yield, seed index, 2.5% span length and bundle strength were presented (Tables 4 to  8).Based on sca effects, KC 2 × TCH 1726 showed significant sca effects for single plant yield and number of sympodial branches.The crosses, KC 2 × TCH 1726 for single plant yield, with low × high gca resulted in dominance effects.Parents with positive significant gca effects involved additive type of gene action that would be easily fixable.The crosses, KC 2 × TCH 1726 number of sympo-dial branches, low × high gca effects.In this situation, we could go for recombination breeding, whereas the parents with low negative gca effects and parents with positive significant gca effects.For yield contributing characters such as number of bolls per plant, hybrid MCU 13 × Narasimha had possesed significant positive sca effects.Earlier studies were reported positive sca effects for plant height (Waldi et al., 1980;Nirania et al., 1992), sympodial branches per plant (Bhatade and Bhale, 1983), number of boll per plant (Waldi et al., 1980), lint yield per plant (Chaudari et al., 1990) and lint index (Bhatade and Bhale, 1983;Avtar et al., 1992).For yield contributing characters, MCU 13 × TCH 1726 showed significant sca effects for number of bolls per plant, number of sympodial branches and ginning percentage.Hybrid MCU 13 × TCH 1726 showed high × high for the number of bolls per plant, low × high for number of sympodial branches and ginning percentage.In this situation, we could go for recombination breeding.

Conclusion
The analysis of combining ability revealed that the variances for the GCA were larger than those for all the traits except number of sympodial branches, single plant yield and bundle strength indicating the preponderance of additive action, which could be explotied for the improvement of these traits by mass selection.In the present study, parent BW4-1 had superior per se performance for single plant yield, boll weight, lint index, number of bolls per plant, seed index and micronaire.Based on high per se performance and high gca effect, the parent MCU 13 was considered as best general combiner as it had significantly higher values for single plant yield, number of sympodial branches, lint index and seed index.Based on sca effects, KC 2 × TCH 1726 showed significant sca effects for single plant yield and number of sympodial branches.The hybrids MCU 3 × Narasimha possessed the significant sca effects for three characters such as plant height, number of bolls per plant and 2.5% span length.An increasing in the seed cotton yield along with other traits will be a valuable addition to cotton cultivars.

Table 1 .
Analysis of variance showing means square for yield and fibre quality traits.

Table 2 .
Analysis of variance and genetic contribution of parents, hybrids and reciprocals.

Day to first boll bursting Number of sympodia/ plant Plant height (cm) Number of bolls/plant Boll weight (g) Lint index Seed index Ginning percent Single plant yield (g) 2.5% span length (mm) Bundle strength (g/tex)
Kalsy et al. (1981)))98)(1995)on, which could be exploited for the improvement of these traits by mass selection.Similar findings were reported bySubrahmanyam et al. (1991)andSubrahmanyam and Bhalod (1995)for fiber fineness in diploid cottons.Regarding ginning percentage and 2.5% span length, additive component of variance was reported byModi et al. (1999)andPatel and Pethani (1998)in desi cotton.Mandloi et al. (1998)also reported predominance of additive gene effects for fiber fineness, ginning percentage and halo length, andSurana et al. (1997)for the byMarani (1963)andKalsy et al. (1981)for seed

Table 3 .
Mean per se performance of parents for yield and fibre quality traits.

Table 4 .
General and specific combining ability effects for plant height and number of sympodia per plant.

Table 5 .
General and specific combining ability effects for boll weight and seed index.

Table 6 .
General and specific combining ability effects for 2.5% span length and bundle strength.

Table 7 .
General and specific combining ability effects for number of bolls per plant and single plant yield.

Table 8 .
General and specific combining ability effects for 2.5% span length and bundle strength.