Cotton is still a principal raw material for the world's textile industry, but its dominant position has been seriously eroded by synthetic fibers. Increased global production, emergence of synthetics as an alternative to cotton textiles and improved productivity are mainly contributing for world supply. World demand for cotton continued to be erratic, and some groups lobbied for increased price- supports, but an upward trend began in the 1980s.

India was the first country to introduce commercial cultivation of intra hirsutum hybrids in 1960s and then interspecific hybrids in 1970s. The success of Interspecific hybrids has led to overcome the acute shortage of ELS (Extra long stable) cotton which the country was experiencing during 1970s and the prominent hybrids from University of Agricultural Sciences, Dharwad namely Varalakshmi and DCH 32 were instrumental in saving very critical foreign exchange running to hundreds of crores of rupees. Later in 90s, due to increase of severity of pests and the inherent susceptibility to them which was contributed by barbadense parents, the performance of interspecific hybrids declined. This popularity of interspecific hybrids declined and the intra hirsutum hybrids have occupied major area under hybrids. In India research on G. barbadense has been neglected as compared to other cultivated species. Therefore no sufficient progress has been achieved in improving yielding ability of G. barbadense cottons. Therefore it is necessary to give much attention to improve the yielding ability of the high priced G. barbadense cotton looking to its superior fiber quality. Based on the factors such as changing consumption pattern, increasing population, competition from synthetic fibers, India needs to increase cotton production in various quality groups.

During Bt era, though hybrids have occupied over 90 per cent area of cotton cultivated in India, there is a very clear decline in diversity of cotton in terms of fiber length classes. There is a steep rise in production of intra hirsutum hybrids, while production of extra long staple and short staple cotton has come down drastically. The situation of ELS interspecific hybrid cultivation has worsened in Bt era that even though india boasts of exporting cotton, import of ELS cotton has become inevitable. The main reason for the decline in realized potentiality of interspecific hybrids has been the lack of prioritized research on improving potentiality of barbadense cotton and developing hybrid oriented populations based on them and utilizing them in deriving potential interspeciific hybrids. Genetic improvement of barbadense varietal lines for both productivity and fiber quality has been very limited. The standard Indian barbadense variety Suvin, is still regarded as a unique fiber class in international market, but the productivity of barbadense varieties like Suvin has reduced so much, that it is no more remunerative to cultivate. There is an urgent need to develop strong base material to serve the cause of developing new variety of barbadense and this improved barbadense varietal base, is essential for improving performance of interspecific hybrids.

Suvin, an ELS variety of Gossypium barbadense species was developed in 1974 by cross breeding Sujatha with St.Vincent Sea Island seeds. The release of ELS variety Suvin with a 2.5% span length of 33~40 mm; fiber strength of 29~33 g/tex and micronaire of 2.9 to 3.9 has been considered as a significant milestone in Indian cotton improvement. Apart from Suvin only a few ELS hybrids grown in India like DCH-32, TCH-213 meet the international ELS specifications. According to some reports, the share of India’s ELS production has declined from 32% of world ELS production in 1992-1993 to 10% in 2007-2008.

The main objective of this study was to identify suitable barbadense lines for developing heterotic sub populations against hirsutum testers to release new lines as an alternative for Suvin variety of barbadense.

1 Results
1.1 Analysis of variance (RBD)
The preliminary RBD analysis was carried out for fourteen characters under study for 53 barbadense lines (included Suvin as check). Mean sum of squares for fourteen characters are presented in Table 1. ‘F’ test indicated highly significant variation among the brbadense lines for all the 14 characters except number of bolls per plant. These results are in confirmity with the studies of Vande and Thombre (1983), Vaman et al. (1985), Mirkhmedov et al. (1987), Mehla et al. (1988), Simongulyan and Kim (1990), Munasov et al. (1990), Tagiev (1991), Virk et al. (1991), Dever and Gannaway (1992) and Akumurdov and Chapau (1992). These authors also observed substantial amount of variability for seed cotton yield among F2, F4 and F5 generations of different cotton crosses.

Table 1 Analysis of variance for different quantitative characters (RBD) in 53 barbadense lines

1.2 Mean per se performance of barbadense lines
The mean performance of 53 barbadense lines included Suvin variety as check, are briefly presented in Table 2.

Table 2 Per se performance of 53 barbadense lines for different quantitative characters

1.2.1 Seed cotton yield (kg/ha)
The per se performance of 53 barbadense lines showed that 13 lines had higher values of seed cotton yield than Suvin line. Maximum seed cotton yield per plant was observed in the line DB 533×DB 534 F5 IPS 132 (2175.70) followed by the line DB 533×DB 534 F5 IPS 18 (2151.21), while the line DB 533×DB 534 F5 IPS 33 recorded minimum seed cotton yield per plant (441.81) followed by the line DB 533×DB 534 F5 IPS 24 (585.08).

1.2.2 Plant height (cm)
The barbadense line DB 533×DB 534 F5 IPS 62 recorded the tallest plant height (100.33) followed by the line DB 533×DB 534 F5 IPS 105 (97.83), while the line DB 533×DB 534 F5 IPS 34 exhibited the shortest plant height (65.33).

1.2.3 Number of monopodia per plant
Among 53 barbadense lines, the line DB 533×DB 534 F5 IPS 6 recorded maximum number of monopodia per plant (2.50) followed by DB 533×DB 534 F5 IPS 18 (2.17). On the other hand, the line DB 533×DB 534 F5 IPS 19 (0.17) recorded minimum number of monopodia per plant.

1.2.4 Number of sympodia per plant
Maximum number of sympodia per plant was observed in the line DB 533×DB 534 F5 IPS 37 (24.83) followed by DB 533×DB 534 F5 IPS 21 (22.50) and DB 533×DB 534 F5 IPS 20 (21.83) which are on par with one another. On the other hand, the lines DB 533×DB 534 F5 IPS 32 and DB 533×DB 534 F5 IPS 1 recorded lowest number of sympodia per plant (13.83) followed by the line DB 533×DB 534 F5 IPS 38 (14.17).

1.2.5 Number of bolls per plant
The per se performance of 53 barbadense lines showed that 17 lines had higher values of number of bolls per plant than Suvin line. The barbadense line DB 533×DB 534 F5 IPS 62 produced maximum number of bolls per plant (39.33) followed by the lines DB 533×DB 534 F5 IPS 48 (38.17), DB 533×DB 534 F5 IPS 44 (37.50), DB 534× DB 533 F5 IPS 22 (37.33) and DB 533×DB 534 F5 IPS 88 (35.33). On the other hand, the line DB 533×DB 534 F5 IPS 96 produced minimum number of bolls per plant (22.50) followed by DB 533×DB 534 F5 IPS 35 (24.33).

1.2.6 Mean boll weight (g)
Among all barbadense lines, 16 lines recorded higher value than Suvin line. The line DB 533×DB 534 F5 IPS 30 exhibited heaviest bolls (3.23) followed by the barbadense lines DB 533×DB 534 F5 IPS 132 (3.17), DB 533×DB 534 F5 IPS 37 (2.93) and DB 533×DB 534 F5 IPS 18 (2.93), while the line DB 533×DB 534 F5 IPS 33 had minimum mean boll weight (1.67) followed by the line DB 533×DB 534 F5 IPS 38 (1.73).

1.2.7 Reproductive points on sympodia
The line DB 533×DB 534 F5 IPS 62 exhibited maximum value of reproductive points on sympodia (5.25) followed by the barbadense lines DB 533×DB 534 F5 IPS 48 (5.08), DB 534 × DB 533 F5 IPS 22 (4.98) and DB 533×DB 534 F5 IPS 71 (4.93), while the line DB 533×DB 534 F5 IPS 101 had minimum value of reproductive points on sympodia (2.58) followed by the lines DB 533×DB 534 F5 IPS 65 (2.67), DB 533×DB 534 F5 IPS 20 (2.67) and DB 533×DB 534 F5 IPS 68 (2.67).

1.2.8 Sympodial length at 50 per cent plant height (cm)
The barbadense line DB 533×DB 534 F5 IPS 62 exhibited maximum value of sympodial length at 50 per cent plant height (39.75) followed by the barbadense lines DB 533×DB 534 F5 IPS 71 (35.92), DB 533×DB 534 F5 IPS 105 (35.83) and DB 533×DB 534 F5 IPS 14 (35.42), while the line DB 533×DB 534 F5 IPS 27 had minimum value of sympodial length at 50 per cent plant height (21.67) followed by the lines DB 533×DB 534 F5 IPS 11 (22.42) and DB 533×DB 534 F5 IPS 101 (22.83).

1.2.9 Inter branch distance (cm)
The barbadense line DB 533×DB 534 F5 IPS 27 exhibited maximum value of inter branch distance (9.83) followed by the lines DB 533×DB 534 F5 IPS 6 (9.33) and DB 533×DB 534 F5 IPS 112 (9.17), while the barbadense line DB 533×DB 534 F5 IPS 16 had minimum value of inter branch distance (6.17) followed by the lines DB 533×DB 534 F5 IPS 1 (6.50) and DB 533×DB 534 F5 IPS 25 (6.67).

1.2.10 Seed index (g)
Among all lines, the line DB 533×DB 534 F5 IPS 101 recorded maximum seed index (12.00) followed by the lines DB 533×DB 534 F5 IPS 132 (11.00), DB 533×DB 534 F5 IPS 34 (10.94) and the line DB 533×DB 534 F5 IPS 52 (10.06). On the other hand, the line DB 533×DB 534 F5 IPS 62 recorded minimum seed index (6.41) followed by the line DB 533×DB 534 F5 IPS 44 (6.49).

1.2.11 Ginning outturn (%)
The line DB 533×DB 534 F5 IPS 57 possessed maximum ginning out turn (35.23) followed by the lines DB 533×DB 534 F5 IPS 18 (35.13), DB 533×DB 534 F5 IPS 65 (34.76), DB 533×DB 534 F5 IPS 10 (34.64) and the line DB 534 × DB 533 F5 IPS 22 (34.00), while the line DB 533×DB 534 F5 IPS 49 recorded minimum ginning out turn (24.77).

1.2.12 Lint index (g)
The line DB 533×DB 534 F5 IPS 18 had maximum lint index (5.41) followed by the line DB 533×DB 534 F5 IPS 101 (5.28) and DB 533×DB 534 F5 IPS 132 (4.96) which are on par with one another, while the line DB 533×DB 534 F5 IPS 62 recorded minimum lint index (2.03) followed by DB 533×DB 534 F5 IPS 36 (2.86).

1.2.13 Photosynthetic rate (µmolCO₂m^-2s^-1)
The line DB 533×DB 534 F5 IPS 13 had maximum value of photosynthetic rate (27.15) followed by the line DB 533×DB 534 F5 IPS 49 (26.80) and DB 533× DB 534 F5 IPS 16 (26.69) which are on par with one another, while the line DB 533×DB 534 F5 IPS 132 recorded minimum value of photosynthetic rate (9.52) followed by DB 533×DB 534 F5 IPS 109 (11.96).

1.2.14 Stomatal conductance (µmolm^-2s^-1)
The line DB 533×DB 534 F5 IPS 55 had maximum value of stomatal conductance (2.38) followed by the line DB 533×DB 534 F5 IPS 10 (2.03) and DB 533×DB 534 F5 IPS 16 (1.29), while the line DB 533×DB 534 F5 IPS 132 (0.30) recorded minimum value of stomatal conductance followed by DB 533×DB 534 F5 IPS 19 (0.42).

1.2.15 Transpiration rate (mmol H₂Om^-2s^-1)
The barbadense line DB 533×DB 534 F5 IPS 105 had maximum value of transpiration rate (16.66) followed by the line DB 533×DB 534 F5 IPS 10 (13.62) and DB 533×DB 534 F5 IPS 112 (11.72), while the line DB 533×DB 534 F5 IPS 132 recorded minimum value of transpiration rate (7.13) followed by DB 533×DB 534 F5 IPS 19 (7.23).

1.2.16 2.5% S L (mm)
The line DB 533×DB 534 F5 IPS 34 had maximum value of 2.5% S L (38.16) followed by the lines DB 533×DB 534 F5 IPS 112 (38.08), DB 533×DB 534 F5 IPS 80 (37.42) and DB 533×DB 534 F5 IPS 21 (37.42), while the line DB 533×DB 534 F5 IPS 4 recorded minimum value of 2.5% S L (28.78) followed by DB 533×DB 534 F5 IPS 25 (30.54). The per se performance of 53 barbadense lines showed that 27 lines had higher values of 2.5% SL than Suvin line.

1.2.17 Fiber uniformity ratio (%)
The barbadense line DB 533×DB 534 F5 IPS 19 had maximum value of fiber uniformity ratio (47.54) followed by the lines DB 533×DB 534 F5 IPS 4 (47.48) and DB 533×DB 534 F5 IPS 11 (46.91), while the line DB 533×DB 534 F5 IPS 34 recorded minimum value of fiber uniformity ratio (42.64) followed by DB 533×DB 534 F5 IPS 23 (42.69).

1.2.18 Fiber micronaire value (g/inch)
The line DB 533×DB 534 F5 IPS 101 had maximum value of fiber micronaire value (4.47) followed by the lines DB 533×DB 534 F5 IPS 19 (4.39) and DB 533×DB 534 F5 IPS 18 (4.14), while the line DB 533×DB 534 F5 IPS 44 (2.42) recorded minimum value of fiber micronaire followed by DB 533×DB 534 F5 IPS 38 (2.49). Among all barbadense lines, 10 lines exhibited higher value than Suvin line for fiber micronaire.

1.2.19 Fiber maturity ratio (%)
The barbadense line DB 533×DB 534 F5 IPS 19 had maximum value of fiber maturity ratio (0.70) followed by the line DB 533×DB 534 F5 IPS 101 (0.69), while the lines DB 533×DB 534 F5 IPS 44, DB 533×DB 534 F5 IPS 38 and DB 533×DB 534 F5 IPS 32 recorded minimum value of fiber maturity ratio (0.55) followed by DB 533×DB 534 F5 IPS 34 (0.56) and DB 533×DB 534 F5 IPS 6 (0.56).

1.2.20 Fiber strength (Tenacity) (g/tex)
The barbadense line DB 533×DB 534 F5 IPS 88 had maximum value of fiber strength (31.67) followed by the lines DB 533×DB 534 F5 IPS 57 (30.80) and DB 533×DB 534 F5 IPS 10 (30.75), while the line DB 533×DB 534 F5 IPS 52 recorded minimum value of fiber strength (26.52) followed by DB 533×DB 534 F5 IPS 2 (27.08). The per se performance of 53 barbadense lines showed that 29 lines had higher values of fiber strength than Suvin line.

1.2 21 Fiber elongation (%)
The lines DB 533×DB 534 F5 IPS 101 and DB 533×DB 534 F5 IPS 57 had maximum values of fiber elongation (6.94) followed by the lines DB 533×DB 534 F5 IPS 79 (6.91), DB 533×DB 534 F5 IPS 19 (6.91) and DB 533×DB 534 F5 IPS 11 (6.91), while the line DB 533×DB 534 F5 IPS 2 (6.18) recorded minimum value of fiber elongation followed by DB 533×DB 534 F5 IPS 27 (6.24).

1.2 22 S/L ratio (%)
The line DB 533×DB 534 F5 IPS 4 had maximum value of S/L ratio (1.01) followed by the lines DB 533×DB 534 F5 IPS 96 (0.96) and DB 533×DB 534 F5 IPS 88 (0.96), while the line DB 533×DB 534 F5 IPS 52 (0.75) recorded minimum value of S/L. Among all barbadense lines, 27 lines showed higher value than Suvin barbadense line.

1.3 Intergeneration (F₃-F₅) correlation and regression analysis for fiber quality
The simple correlation was worked out between parent F3 and progeny F₅ mean for all the quality characters by considering same characters in both the generations and the results are presented in Table 3.
Intergeneration correlation values revealed that, the barbadense lines in both generations exhibited highly significant positive correlation for 2.5% SL (0.48) and significant positive correlation for fiber uniformity ratio (0.37) and fiber micronaire value (0.29). None of barbadense in two generations exhibited significant correlation for fiber maturity ratio, fiber tenacity and fiber elongation.

Table 3 Intergeneration correlation coefficient between F3 and F5 generation for fiber qualit

1.4 Narrow sense heritability estimates for fiber quality
The evaluation of parent and progeny provided an opportunity to compute heritability values following parent offspring regression method can be utilized for predicting the response to selection in early and advanced generations. The results obtained from the study are presented in Table 4.

Table 4 Estimates of heritability (Narrow sense) for fiber quality characters in barbadense lines (F3/F5)

The moderate heritability was noticed for fiber maturity ratio (32%), fiber tenacity (41%) and fiber elongation (34%). The high heritability was noticed for 2.5% SL (89%), fiber uniformity ratio (74%) and fiber micronaire value (63%).

2 Discussion
There is a steep rise in production of long staple cotton, while production of extra long staple, short staple and even medium staple cotton has come down drastically. Because of this, India is importing extra long staple cotton by causing heavy birth on exchequer and there is a need of improving performance of interspecific hybrids. This is possible through genetic improvement of barbadense lines. So that both productivity and fiber quality of barbadense improved. An improved barbadense varietal base is essential for improving performance of interspecific hybrids. This study also focused attention on determining whether these barbadense lines can straight away be utilized as varieties for producing ELS cottons. Hence, 53 F₅ barbadense lines were evaluated for productivity traits and fiber quality traits and compared with Suvin known to be standard barbadense variety in international market.

Among 53 barbadense lines (included Suvin variety as check), thirteen lines recorded higher per se performance for seed cotton yield than Suvin. These lines also recorded high values for other characters like number of bolls per plant, mean boll weight, seed index, ginning outturn, lint index, photosynthetic rate and stomatal conductance. These parents exhibited good fiber quality traits like, 2.5% S/L, fiber micronaire value, fiber tenacity and S/L ratio. The per se performance of these potential barbadense lines showed higher values than suvin for seed cotton yield and fiber strength.

The potential barbadense line DB 533×DB 534 F5 IPS 18 (Figure 1) showed superiority over Suvin line for seed cotton yield, mean boll weight, seed index, ginning outturn, lint index, fiber micronaire value, fiber strength and S/L ratio and can be used for developing separate trait based populations with other barbadense lines for yield components and fiber quality. Confirmation of their superiority will enable release of these new lines as an alternative for Suvin variety of barbadense.

Figure 1 Most productive barbadense line showed superior than Suvin line

3 Materials and Methods
To create recombinational variability for combining ability, the elite barbadense lines DB 533 and DB 534 were crossed during 2007-2008. During two seasons 2008-2009 and 2009-2010 these barbadense crosses were advanced to F₂ and F₃ generations, respectively. The F₃ lines were evaluated for productivity and fiber quality parameters realizing the emphasis laid on developing ELS (Extra Long Stable) cotton hybrids out of 171 F₃ lines, only 53 F₃ lines with acceptable fiber strength were utilized in this study. The crossing programme was taken up during 2010. The 28 out of 53 F₄ barbadense lines were crossed with four common hirsutum testers to obtain derived F₁ crosses seeds. The barbadense population of F₄ lines was selfed and material was advanced to F₅ generation during the same year.

The entire experimental material was planted on a medium black soil at College of Agriculture, Dharwad under irrigated condition. All the 53 F₅ (included Suvin variety as check) lines, four hirsutum testers and derived F₁ crosses along with the straight crosses (Bench Mark Crosses (BMC)) and ruling commercial checks (MRC 6918 Bt check and DCH 32 non Bt check) were sown during kharif 2011 in a Randomized Block Design with two replications and a spacing of 90 cm between rows and 60 cm between the plants within a row. Recommended fertilizer doses were applied and other cultural practices were carried out at regular interval. Plant protection measures were taken at appropriate time to control pests and diseases. To estimate fiber quality parameters, one sample was taken from pool of kapas derived from all the replications. Hence, these characters were not subjected to RBD analysis.

Correlation coefficient of each character (fiber quality characters) between F₃ and F₅ generation was found out by calculating the phenotypic correlation coefficient. Intergeneration correlation coefficient analysis was done for entire populations of selected F₃ and F₅ progenies for all the fiber quality characters.

Narrow sense heritability estimates were made based on the regression of F₅ on F₃ using the following (Cahaner and Hillet, 1980).

                     Covariance of F₃ F₅ for the character
b (F₅F₃) = ----------------------------------------------------
                                     Variance of F₃
Regression of each character between F₅ and F₃ was found out by calculating the regression coefficient taking the same character in both the generations. Regression coefficient between F₅ and F₃ generations was estimated as heritability value using multiplicative factor of 2/3 or 0.67. Heritability estimates were computed for entire population of selected progenies for all the characters. Based on breeding, coefficients of the generations were involved in regression analysis.

The heritability estimates were categorized according to Robinson et al. (1949) and given below: 0%~30% - Low; 30%~60% - Moderate; 60% and above – High.

References
Akumuradov S., and Chapau A., 1992, Response of varieties to irradiation, Khlopok., 1: 36-38

Cahaner A., and Hillet J., 1980, Estimating heritability and genetic correlation between traits from generations F₂ and F₃ of self fertilizing species: A comparison of three methods. Theor. Appl. Genet., 58: 33-38
http://dx.doi.org/10.1007/BF00264666

Dever J.K., and Gannaway J.R., 1992, Relative fibre uniformity between parents and F₁ and F₂ generations in cotton, Crop Sci., 32:1402-1408
http://dx.doi.org/10.2135/cropsci1992.0011183X003200060019x

Mehla A.S., Mor B.R., and Naidu M.R., 1988, Genetic analysis of earliness characters in upland cotton (G. hirsutum L.), Crop Improv., 15: 156-159

Mirakhamedov S., Stakov V., and Rakhimov M., 1987, Transgressive selection for fibre length, Khlopkovodstvo., 7: 30-33

Munasov K., Alikho Dzhaeva S.S., and Lemeshev N.R., 1990, Variability and heritability of fibre length F₁-F₃ hybrids and their parents, Refretivinya Zhurnal., 9: 3378

Robinson H.F., Comstock R.E., and Harvey .H., 1949, Estimates of heritability and degree of dominance in corn, Agron. J., 41: 353-359
http://dx.doi.org/10.2134/agronj1949.00021962004100080005x 
 
Simongulyan N.G., and Kim V.L., 1990, Variation in quantitative characters I an irradiated cotton population, Genetika., 26(10): 1815-1821

Tagiev A.A., 1991, Variation in cotton induced by chemical mutagens. In: Khimicheskii Mutagene Selecletsii., pp. 214-217

Vaman B.M., Nadarajan N., and Pater S.D., 1985, A note on the evaluation of genetic stock of Gossypium barbadense L. Madras. Agri. J., 72: 116-117

Vande N.N., and Thombre M.V., 1983, Variability studies in Egyptian cotton, Cotton Improv., 13(1): 3

Virk P.S., Kalsy H.S., Virk D.S., and Pooni H.S., 1991, An assessment of the breeding potential of a Gossypium hirsutum L., Crop Improv., 18: 7-13