Introduction
Cotton, being the king of fibers in preparing human apparel has played a key role in civilization of mankind. Cotton is providing livelihood directly and indirectly to over 60 million people and accounting for about 16 per cent of India’s export earnings.

Hybrids have occupied nearly 90% area of cotton cultivated in India. There is a constant need to develop more potential hybrids and adopt noval approaches for improving hybrid performance. In cross pollinated crops like maize heterotic populations are developed and exploited through population improvement schemes meant for improving combining ability. Such programmes are integral part of hybrid breeding programme and these populations are shared among breeders and used further to obtain more potential hybrids. Studies have shown that even in cotton it is possible to adopt these concepts with suitable modifications in the procedure to suit the mating system of self pollinated crops (Patil and Patil, 2003; Patil et al., 2007).

Exploiting heterosis is one of the methods used to increase cotton yields that have stagnated in recent years. The success of the hybridization is largely dependent on the correct selection of parents. Estimates of genetic variation and combining ability are useful in determining the breeding value of some populations and the appropriate procedures to use in a breeding program. The general combining ability effects are important indicators of the value of genotypes in hybrid combinations. Differences in general combining ability effects have been attributed to additive, additive x additive, and higher-order additive interactions, whereas differences in specific combining ability have been attributed to non-additive genetic variance (Falconer, 1960).

The concept of combining ability is important in designing plant breeding programmes. It is especially useful in testing procedures, where it is desired to study and compare the performance of lines in hybrid combinations. Combining ability or productivity in crosses is defined as the ability of parents or cultivars to combine amongst each other during the process of hybridization so that favourable genes/characters are transmitted to their progenies. Two types of combining ability, general and specific, have been recognized in quantitative genetics. Specific combining ability is defined as the deviation in the performance of hybrids from the expected productivity based upon the average performance of lines involved in the hybrid combination, whereas general combining ability is defined as average performance of a line in a series of crosses. According to Sprague and Tatum, general combining ability is due to genes which are largely additive in their effects and specific combining ability is due to the genes with dominance or epistatic effect. Rawlings and Thompson used line x tester analysis to estimate GCA and SCA of inbred parents. Since the development of new cultivars through hybridization is a continuous process, information on combining ability of new cultivars remains important. Desphande and Baig noted that though GCA and SCA variances were important, the magnitude of SCA was higher than GCA indicating the preponderance of dominant genes controlling number of bolls, ginning outturn%, seed index, lint index and seed cotton yield. Contrary to the above findings, Rokaya et al. found significance of GCA and SCA suggesting the importance of additive as well as dominant genes, nevertheless the ratio of GCA/SCA was greater than the unity further indicating the preponderance of additive genes in the inheritance of seed cotton yield, seed index and lint%.

In hybrid research study on cotton, large number of crosses involving varietal lines are used for assessing combining ability status. On constantly observing the most potential crosses attempts are made to infer about the causes of high heterosis. What are the combinations that give potential crosses? What would be the probable cause for high potentiality revealed by the F₁? What is the genetic base or is there any physiological mechanism linked to high productivity of F₁ etc., are the questions which are examined and on the basis of the information available, heterotic groups are developed (Patil et al., 2011).

The most potential crosses observed in present study have been examined and based on this the combining ability behavior (Pattern) of the line involved is determined. With the help of this information diverse groups are formed which are capable of giving potential hybrids between them. A study of set of hybrids involving the line as a common parent gives an idea about the combining ability pattern of the concerned line. The higher or lower performance of the hybrids is itself taken as reflection of genetic distance existing between the parents. It has been possible to identify heterotic combinations (potential crosses) based on their percent superiority over the commonly used check. When these crosses show up to be consistently potential, they are considered while forming heterotic groups involving parents of such crosses.

The exercise of identifying diverse groups is a continuous process because the new breeding lines developed and stabilized and those lines obtained from other sources are included in developing crosses and these lines could be added in different heterotic groups after studying their combining ability behavior (pattern) by crossing with representative genotypes of different groups. Thus the grouping of genotypes is continuously revised and refined. In the recent years, the concept of developing heterotic groups is put to test in self pollinated crops like cotton. Segregating populations based on diverse pairs of genotypes can be the ideal base material required for implementing procedures like reciprocal selection for improving combining ability (Patil and Paltil, 2003; Patil et al., 2011).

One of the problems in using heterosis in cotton involves defining a strategy for the selection of parents that will ultimately produce productive hybrids. The present study evaluated parents and hybrids produced from line x tester mating. The objective of this study was to estimate parental general combining ability effects, to compare performance among F₁ hybrids, and to identify those superior for lint yield and yield components.

1 Results and Discussion
ANOVA for combining ability in respect of ten characters is presented in Table 1. Among the lines, the mean sum of squares was significant for all characters except monopodia per plant and mean boll weight. The testers differed significantly for most of the characters except monopodia per plant, mean boll weight and seed index. However, the line x tester interaction was significant for all the characters except monopodia per plant, mean boll weight, seed index and lint index.

 

Table 1 Combining ability analysis of variance as per Kempthorne (1957) in respect of kapas yield and its attributing characters of cotton hybrids (G .hirsutum L.)

The estimates of variance due to general combining ability (GCA), variance due to specific combining ability (SCA), GCA and SCA ratio, additive variance (VA) and dominance variance (VD) were worked out for different characters and presented in Table 2. It is evident from the table that, the estimates of SCA variance was predominant for all the characters. The GCA and SCA variance ratio was wider for kapas yield per plant (1:199.86) followed by monopodia per plant (1:150.00) and bolls per plant (1:81.60). From the estimates of additive and dominance variance, it is observed that dominance variance was predominant for all the characters and was maximum for kapas yield per plant (266.24) followed by plant height (12.81) and bolls per plant (12.37). However, both additive and dominance variance were found to be important in case of ginning per cent, monopodia per plant, mean boll weight, days to 50 per cent flowering, seed index and lint index. 

Table 2 Variance due to general and specific combining ability for kapas yield and its attributing characters of cotton hybrids (G .hirsutum L.)

SCA variances were higher in magnitude compared to GCA variances for almost all characters, which indicated predominance of non-additive gene action suggesting good scope for heterosis breeding.

Higher SCA variance than GCA variance was reported by Shanmugavalli and Vijendradas (1995) for days to 50% flowering; Bhatade et al. (1980); Singh and Singh (1980); Virk and Kalsy (1982); Shanti and Selvaraj (1995); Kalsy et al.(1981) ; Gill and Singh (1982) for plant height; Hapase et al. (1987) for Mononpodia per plant; Singh and Gupta (1970); Kalsy and Garg (1980); Singh and Raut (1983) for sympodia per plant, Singh and Singh (1985); Hapase et al. (1987); Raut et al. (1989) for bolls per plant; Bhandari (1980); Jagtap (1986); Shanti and Selvaraj (1995); Panchal et al. (1995) for boll weight; Singh et al. (1983) for Seed cotton yield (g/plant); Deshpande and Bhale (1984) for ginning per cent; Chakreshkumar et al. (1984); Hapase et al. (1987) for seed index; Grakh and Choudhary (1983) for lint index.

1.1 Estimation combining ability effects of parents and hybrids
The estimates of general combining ability (gca) effects of 14 parents (6 lines and 8 testers) and specific combining ability for 48 F₁ hybrids for different characters are presented in Table 3, Table 4 respectively and briefly discussed below.

Table 3 General combining ability effects of parents (lines and testers) in respect of kapas yield and its attributing characters of cotton (G .hirsutum L.)

Table 4 Specific combining ability effects of hybrids in respect of kapas yield and their attributing characters in cotton (G. hirsutum L.)

1.1.1 Days to 50 per cent flowering
None of the lines exhibited significant gca effect in desirable direction. However, line RAH 318 recorded maximum gca effect (-0.59) in desirable direction. On the other hand, the tester RGR 37 recorded positive and significant gca effect (1.39) in desirable direction among testers. Of 48 hybrids, none of the hybrids showed significant sca effects for this trait, indicating predominant role of additive gene action in the genetic control of this trait. Kolte and Thombre (1981); and Patil and Chopde (1983) reported similar findings in their study.

1.1.2 Plant height (cm)
Among the lines, line RAH 318 registered highest negative and significant gca effect (-9.56) followed by line RAH 124 (-7.58) recorded negative and significant gca effect. The line RAH146 showed positive and significant gca effect (8.59). On the other hand, the tester RGR 37 (-11.27) also recorded highest negative and significant gca effect. For this trait, only one hybrid viz. Hybrid RAH128 x RGR 37 (L₃T₈) manifested significant negative sca effect (-23.25) indicating non additive gene action in this cross and additive gene action in the remain crosses in the genetic control of this trait. Shanti and Selvaraj (1995) also reported similar results.

1.1.3 Monopodia per plant
None of the lines recorded significant gca effect in positive direction, but the line RAH 124 recorded negative and significant effect (-0.18). On the other hand, among testers the tester RGR 32 recorded positive significant gca effect (0.19). Two hybrids viz., RAH 243 x RGR 37 (L₂T₈) and RAH 146 x SC 14 (L₄T₁) expressed significant negative sca effects of (-0.48) and (-0.51) respectively, indicating predominant role of non additive gene action in these crosses and additive gene action in the remain crosses. The same result respect of this trait reported by Katanalli et al. (2004), Ahuja and Dhayal (2007).

1.1.4 Sympodia per plant
Among the lines, line RAH 146 recorded significant positive gca effect (1.68), but the line RAH 318 showed negative and significant effect (-4.46). Among testers, positive significant gca effect (1.94) was recorded by tester SC68. Only one out of 48 hybrids viz, RAH 128 x RGR 37 (L₃T₈) showed significant negative sca effect (-5.07) for this trait. This according to the result of Katanalli et al. (2004).

1.1.5 Bolls per plant
Among six lines, the line RAH 128 manifested highest positive and significant gca effect (3.57) followed by lines RAH 146 and RAH 243 were also exhibited significant positive gca effect (3.49) and (2.00) respectively, while the lines RAH 97(-6.56) and RAH 124 (-2.60) showed highest negative significant gca effect. Among all testers the tester RGR 58 showed highest positive and significant gca effect (3.47) followed by testers SC 18 (-2.30), RGR 32 (-2.17) which registered negative and significant gca effect. Eight out of 48 hybrids exhibited significant sca effects. The crosses, RAH 318 x RGR 24 (L₁T₆), RAH 128 x SC 68 (L₃T₄), RAH 124 x SC 14 (L₆T₁) and RAH 124 x SC 18 (L₆T₂) manifested highest positive and significant sca effect of (9.07), (7.26), (6.03) and (5.30) respectively. On the other hand, the crosses RAH 318 x SC 18 (L₁T₂), RAH 243 x SC 68 (L₂T₄), RAH 146 x SC 14 (L₄T₁) and RAH 124 x RGR 24 (L₆T₆) showed significant negative sca effects of (-5.30), (-5.77), (-5.65) and (-6.13) respectively, which were on par with one another. Nageshwara and Shiva (2002) reported similar findings in their study.

1.1.6 Mean boll weight (g)
The line RAH 97 recorded positive significant gca effect (0.40), while the line RAH 243 had highest negative significant gca effect (-0.21). Among testers, the tester RGR 32 exhibited significant and positive gca effect (0.30) and the tester SC 68 exhibited highest significant and negative gca effects (-0.29). Among 48 hybrids, hybrid RAH 97 x RGR 37 (L₅T₈) and hybrid RAH 124 x RGR 37 (L₆T₈) exhibited significant negative and positive sca effects of (-0.56) and (0.63) respectively. Patel et al. (2004) found the same result in respect of this trait.

1.1.7 Kapas yield (g/plant)
Highest positive significant gca effect (14.00) for this trait was recorded by the line, RAH 146 and highest negative and significant gca effect (-13.45) was registered by the line RAH 124. Among testers, the tester SC 18 recorded positive significant gca effect (8.64) followed by tester RGR 37 (8.38), but the tester SC 68 showed highest significant and negative gca effect (-14.18) followed by tester SC 14 (-7.27). Ten out of 48 crosses exhibited significant sca effects of which 4 were positive and 6 were negative. The crosses RAH 97 x SC 18 (L₅ T₂) and RAH 318 x RGR 24 (L₁T₆), recorded highest positive and significant sca effects of (62.53) and (30.53) respectively. Contrary to this, the crosses RAH 97 x RGR 24 (L₅T₆) followed by the cross RAH 146 x SC 14 (L₄T₁) exhibited highest negative and significant sca effect of (-27.71) and (-23.65) respectively. Nageshwara Rao and Shiva Santha Reddy (2002) reported the same findings.

1.1.8 Ginning per cent (%)
Among the lines, the line RAH 146 manifested positive significant gca effect (2.19) while, the line RAH 243 recorded negative significant gca effect (-2.20). Among testers, the tester RGR 32 exhibited positive and significant gca effect (2.64) while tester SC 7 recorded negative significant gca effect (-2.42). Only one cross viz, RAH 318 x RGR 37 (L₁T₈) out of 48 crosses exhibited significant positive sca effect (5.88) indicates role of additive gene action. Kalsy et al. (1981); Bains et al. (1982); Singh et al. ( 1988); Panchal et al. (1995) reported the same findings.

1.1.9 Seed index (g)
Only the line RAH 97, among the lines recorded positive significant gca effect (1.16). On the other hand, only the tester RGR 24 manifested positive significant gca effect (0.75). Of 48 hybrids, none of the hybrids showed significant sca effects for this trait indicates predominant role of additive gene action. Singh and Singh (1980); Bhatade and Bhale (1983) reported the same results.

1.1.10 Lint index (g)
Among the lines also, two out of six lines viz, line RAH 97 and line RAH 243 expressed highest significant positive and negative gca effects of (1.07) and (-0.72) respectively. However, none of the testers exhibited significant gca effects for this trait. Of 48 hybrids, none of the hybrids showed significant sca effects for this trait reveals predominant role of additive gene action. Similar results was reported by Gururaja Rao et al. (1977).

1.2 Overall general combining ability status of lines and testers
Since the different traits are both positively and negatively correlated, it is usual to find, for a particular parent, gca in the desirable direction for some characters and in undesirable direction for the others. Hence, it becomes important to decide whether a parent is a good general combiner across all the characters or not. In this context, the method proposed by Arunachalam and Bandyopadhyay (1979) with slight modification as suggested by Mohan Rao (2001) was used in the present investigation to determine the overall general combining ability status of a line or a tester for its gca effects across the traits. The results of the same are presented in Table 5.

Table 5 Overall general combining ability status of parents in cotton hybrids

It is clear from the table that, the lines RAH 318, RAH 243 and RAH 124 among the lines recorded high (H) overall gca status, the line RAH 128, RAH 146 and RAH 97 exhibited low (L) overall gca status. Among the testers, SC 18, RGR 32, RGR 58 and RGR 37 recorded low (L) overall gca status, while the remaining four testers exhibited high (H) overall gca status. 

1.3 Overall specific combining ability status of crosses
It also becomes important to determine whether a cross is a good specific combination across all the traits or not for the same reason explained above. The overall sca status of all the 48 hybrids was determined and the same has been depicted in Table 6.

Table 6 Overall specific combining ability status of cotton hybrids

From the table, it is evident that 23 out of 48 hybrids had high (H) overall sca status, while remaining 25 crosses had low (L) overall sca status across all the traits studied. The highest total score over the final norm was recorded by the hybrid 149 followed by 150, 111 and 141. Hence, these crosses can be utilized for heterosis breeding. 

1.4 Proportional contribution of lines, testers and line x tester interaction to the total variance in hybrids
The per cent contribution of lines towards total variation in the hybrids (Table 7 and Figure 1) was higher from seed index (52.64%), followed by sympodia per plant (50.28%) and lint index (47.88%). On the other hand, per cent contribution of line x tester interaction to the total variance was higher for monopodia per plant (70.47%), kapas yield per plant (66.73%), plant height (51.98%), days to 50% flowering (48.82), bolls per plant (44.62%), mean boll weight (39.76%) and sympodia per plant (37.62%).

Table 7 Proportional contribution of lines, testers and their interactions to the total variance in Cotton (G. hirsutum L.)

Figure 1 Proportional contribution of lines, testers and their interactions to the variance in hybrids of cotton (G. hirsutum L.)

 
The per cent contribution of both tester and line and their interaction is dominant on par and important per cent contribution of line with respect to ginning per cent.

2 Conclusions and Recommendations
The term heterotic group refers to “a group of related or unrelated genotypes from the same or different populations, which display similar combining ability and heterotic response when crossed with genotypes from other genetically distinct germplasm groups” (Melchinger and Gumber, 1998). In the recent years the concept of developing heterotic populations is put to test in self pollinated crops like cotton, segregating populations based on diverse pairs of genotypes can be the ideal base material required for implementing procedures like reciprocal selection for improving combining ability (Patil and Paltil, 2003; Patil et al., 2011). In hybrid research study on cotton, a large number of crosses involving varietal lines are used for assessing combining ability status. On constantly observing the most potential crosses attempts are made to infer about the causes of high heterosis.

If more lines are found to be giving superior crosses with a tester then it is possible to initiate multiple crosses among such lines selected for combining ability and this can lead to creation broad gene pool of recombination variability for combining ability as the population developed in this manner based on number of components improved in ability to combine with the tester. This heterotic gene pool can be exploited for developing superior hybrid combinations with the tester concerned.

Utilization of heterosis depends on genetic diversity existing between the parents, magnitude of dominance at the yield influencing loci and the genetic distance between the chosen parental genotypes. It is possible to maximize heterosis by enhancing genetic distance between two chosen parental populations. Many population improvement schemes are followed in cross pollinated crops to increase genetic diversity, to create heterotic groups and exploit them. These schemes can be extended to self pollinated crops by introducing slight modifications in the procedures to suit the crossing system of self pollinated crops. In present study heterotic box was developed by involving barbadense and hirsutum varieties and it was exploited by creating recombinational variability for combining ability. If two lines A and B are found to give potential crosses with testers T₁ and T₂, it is possible to increase the genetic distance between these opposite pairs A, B vs, T₁ and T₂ by following population improvement scheme for improving combining ability defined in cross pollinated crops by introducing suitable changes to match the crossing system seen in self pollinated crops. The recombinational variability realized in a segregating generation like F₃/F₄ can be evaluated by crossing these lines with opposite testers representing opposite heterotic group.

Based on the principle of suggested by Patil and Patil (2003) predicted double cross performance and two pairs of crosses mentioned in table 8 and 9 are selected to form heterotic boxes. These opposite F₁s are advanced to F₂ and the material is ready for initiating reciprocal selection for combining ability. These crosses are confirmed for consistency of performance and based on this information one out of these heterotic boxes will be selected.

Table 8 Mean performance of best hybrids for different characters

Table 9 Identifying diverse pairs of crosses based on predicted performance of double cross combinations

 
3 Materials and Methods
The plant materials used in the present study were obtained by line x tester crossing. According to this method, RAH 318, RAH 243, RAH 128, RAH 146, RAH 97 and RAH 124 were crossed as the lines with SC 14, SC 18, SC 7, SC 68, RGR 32, RGR 24, RGR 58 and RGR 37 as the testers. The six hirsutum lines representing Robust plant type classes but differing in efficiency of physiological processes like photosynthesis, were selected and crossed to a set of eight testers representing compact types and faster growth rate. The seeds of the hybrids were supplied by Dr. S.S. Patil, Senior Cotton Breeder, Agricultural Research Station, Dharwad Farm, Karnataka, India, in 2008. The experiment comprising of 48 experimental hybrids along with 3 checks (BUNNY Bt, RCH2 Bt, RAHH 95) (one repeated two times) was laid out in Randomized Complete Block Design (RCBD) with two replications. Each entry was sown in 3 row plots of 6 m length spaced at 90 cm with recommended dose of fertilizer and treatment of seeds with Imidochloprid were sown on 10-7-2008, 2~3 seeds were dibbled per spot in each row and thinning was attended to retain one healthy plant per hill at 25 days after sowing. All the recommended package of practices were followed to rise healthy crop.

Samples containing 20 bolls were hand-harvested from each plot prior to picking. The days to 50 per cent flowering recorded by the number of days taken from the date of sowing to the date when the first flower opens in 50 per cent of the plants. The number of monopodia per plant are the number of branches on main stem which were lateral and axillary in position with vertical growth in acropetal succession was counted at maturity stage , avoiding small sprouts, but the number of sympodia per plant are branches which are extra-axillary in position and normally horizontal with zig -zag pattern of fruiting points were taken as sympodia. The number of such sympodia on main stem were counted at maturity stage. The boll samples were weighed to determine seed cotton weight per boll values, and ginned on a roller using laboratory gin for lint percentage (100 x lint weight/seed cotton weight) and 100-seed weight calculations (seed index). The ginned lint from each plot was weighed and divided by the number of plants within each plot to determine lint yield per plant. Five plants were selected randomly from each genotype to find the boll number per plant. The general combining ability variance of parents and the specific ability variance of hybrids were estimated via line x tester variance analysis according to Singh and Chaudhary (1977). The Microsoft Excel computer program was used to analyze the data. The line x tester mating design can provide information regarding the usefulness of male and female inbreds as parents for hybridization to generate segregating populations, which is expected to give prodigious selections. The general (GCA) and specific combining ability (SCA) effects were estimated using a 2-way table with the following formulae:

GCA_lines = (Xi../tr) - (XÉ/ltr)


GCA_testers = (X.j./lr) Ð (XÉ/ltr)

SCA = (Xij./r) - (Xi../tr) - (X.j./lr) + (XÉ/ltr)

where Xi.. is the sum of columns; X.j. is the sum of rows; Xij. is the total value of hybrids over replications; XÉ is the grand total; and l, t and r are the number of lines, testers and replications.

References
Ahuja S., and Dhayal L.S., 2007, Combining ability estimates for yield and fibre quality traits in 4 x 13 line x tester crosses of Gossypium hirsutum, Euphytica, 153(1~2): 87-98

Arunachalam V., and Bandyopadhyay A., 1980, Are ‘Multiple cross-multiple pollen hybrids’ an answer for productive populations in Brassica campestris variety brown sarson? Theor. Appl.Genet., 58(1): 5-10

Bains S.S., Murthy B.R., Joshi A.B., and Nanda G.S., 1982, Genetic of lint yield, boll weight, boll number and seed index in Gossypium hirsutum L., Indian J. Genet., 42: 134-139

Bhatade S.S., and Bhale N.L., 1983, Combining ability for seed and fibre characters and its interaction with locations in Gossypium arboreum L., Indian J. Agric. Sci., 53: 418-422

Bhatade S.S., Shobhane M.R., and Unchegoankar P.K., 1980, Heterosis and combining ability in Gossypium arboretum L., Indian J. Agric. Sci., 50: 310-316

Deshpande L.A., and Baig K.S., 2003, Combining ability analysis for yield, economic and morphological traits in American cotton (Gossypium hirsutum L.). J. Res. ANGRAU., 31: 28-34

Falconer D.S., 1960, Introduction of Quantitative Genetics, Ronald Press Cor., New York

Gururaja Rao M.R., Hiremath K.G., and Virupakshappa K., 1977, Genetic analysis of ginning and fibre properties in upland cotton-graphic analysis of six quantitative traits, Mysore J. Agric. Sci., 11: 444-452

Kalsy H.S., Garg H.R., Vithal B.M., and Singh T.H., 1981, Comining ability studies in a diallel cross of cotton. (Gossypium hirsutum L.), J. Indian Soci. Cotton Improv, 6: 73-79

Katanalli D.S., Patil S.S., and Pattanashetti S.K., eds., 2004, Combining ability study in compact and robust cotton genotypes, Patil, UAS, Dharwad, pp.255

Kolte T.B., and Thombre M.V., 1981, Diallel analysis of some quantitative characters in Gossypium L. Species, J. Maharashtra. Agric. Univ., 6: 8-9

Melchinger A.E., and R.K. Gumber, 1998, Overview of heterosis and heterotic groups in agronomic crops, In: K.R. Lamkey, and J.E. Staub, (eds.), Concepts and breeding of heterosis in crop plants, CSSA, Madison, WI., pp.29-44

Mohan Rao A., 2001, Heterosis as a function of genetic divergence in sunflower (Helianthus annuus L.), Ph. D. thesis submitted to Acharya N.G. Ranga Agricultural University, Hyderabad

Nageshwara Rao G., and Shiva Santha Reddy M., 2002, Combining ability studies of yield and yield contributing traits using diversified plant types in cotton, J. of Cotton Res Dev., 16: 19-23

Panchal S.S., Patel J.A., Saiyad M.R., Patel S.B., and Patil S.A., 1995, Line x Tester analysis of combining ability in Egyptian cotton (Gossypium barbadense L.), Gujarat Agric .Univ Res. J., 20: 58-63

Patil S.S., and Patil S.A., 2003, Role of improving combining ability in increasing performance of cotton hybrids, Third World Cotton Research Conference, Cape Town, South Africa, pp.234-238

Patil S.S., Ramakrishna V., Manjula S.M., Swati P., Ranganatha H.M., Kencharaddi H.C., and Deepakbabu H., 2011, Deploying reciprocal selection for combining ability for improving performance of hybrids in cotton (Gossypium hirsutum), Indian J. Genet., 71(2): 180-184

Patil S.S., S.A Patil, V. Ramakrishna, H.B. Deepak, and H.B. Malikarjun, 2007, Approaches of improving combining ability for enhancing performance of cotton hybrids, Proc Fourth World Cotton Research Conference, Lubbock, USA, pp.1818

Patil R.A., and Chopde P.R., 1983, Diallel cross analysis for combining ability in upland cotton (Gossypium hirsutum L.), J. Indian Soci. Cotton Improv., 8: 54-60

Rawlings J.O., and Thompson D.L., 1962, Performance level as criterion for the choice of maize testers, Crop Sci., 2: 217-220
http://dx.doi.org/10.2135/cropsci1962.0011183X000200030012x

Rokaya M.H., El-Marakby A.M., El-Agroudy M.H., and Seif G.M., 2005, Heterosis and combining ability for fiber-to-seed attachment force, earliness, yieldand yield components in a half diallel cross of cotton, Arab Univ. J. Agric. Sci., 13: 741-753

Shanmugavalli N., and Vijendra Das L.D., 1995, Line × Tester analysis of combining ability in cotton, Madras Agric. J., 82: 449-452

Shanti N., and Selvaraj U., 1995, Combining ability studies for quantitative characters in cotton (Gossypium hirsutum L.), Madras Agric. J., 82: 162-164

Singh P., and Singh H.G., 1980, Line × Tester analysis in upland cotton, Crop Improv., 7: 88-96

Singh R.K., and Choudhry B.D., eds., 1979, Biometrical methods in quantitative genetic analysis, Kalyani Publisher, New Delhi, pp.191-200
PMid:574497

Singh T.H., Randhwa L.S., and Singh M., 1988, Combining ability studies for yield and its components over environments in upland cotton, J. Indian Soci. Cotton Improv., 13: 11-15

Sprague G.F., and Tatum L.A., 1942, General versus specific combining ability in single crosses of corn, J. Amer. Soc. Agron., 34: 923
http://dx.doi.org/10.2134/agronj1942.00021962003400100008x

Virk P.S., and Kalsy H.S., 1982, Combining ability analysis for some metric traits in upland cotton (Gossypium hirsutum L.), Crop Improv., 9: 152-155