
Research Article


VarietyxEnvironment Interaction in Cotton Yield Trials


Volkan Sezener ,
Nedim Ozbek ,
Oktay Erdogan ,
Taner Bozbek ,
Ilkay Yavas
and
Aydin Unay



ABSTRACT

Aegean Region standard cotton varieties and some commercial varieties were evaluated at four environments. Varietyxenvironment interactions was found significant for lint yield and lint percentage. The stability parameters, regression coefficient and mean squares of deviation from regression coefficient were computed and the relations between regression coefficients and lint yield and lint percentage were figured. The standard variety, Nazilli 84 S was the most stabile variety for lint yield, while Carmen, Ozbek 142 S and Sahin 2000 were group of the midadaptation to all environments. Ozbek 142 S also was a group of the best adaptation to all environments because of its regression coefficient and the highest lint percentage.





Introduction Cotton is grown an area of about 698.000 hectares in Turkey, among which Aegean Region shares 176.000 hectares according to 2004/2005 data (Ozudogru and Cakaryildirim, 2005). The lint yield is much higher than the major cotton growing countries. However, Miller et al. (1959) also revealed that cotton is grown in an array of edaphic, climatic, disease and insect conditions. These factors vary from one location to another and from year to year in the same location and their effects are reflected in yields of cotton cultivars. In the result of cotton breeding, adaptation of improved cultivars is an important factor for increasing the productivity of cotton. But inconsistent yield performance of the cultivars under diverse environmental conditions due to high genotypeenvironment interaction is a major contributing factor to reduction in productivity (Misra and Panda, 1990).
The expression of complex characters, particularly yield, is heavily influenced
by environmental conditions. Furthermore, because all of the traits that contribute
to yield may not be positively correlated among themselves, material selected
for high yield in one environment may perform relatively poorly in another.
Genotypexenvironment interaction can be result of genotype rank changes from
one environment to another, difference in scale among environments, or a combination
of these phenomena (Cornelius et al., 1993). In the continuing debate
on whether or not modern highyielding varieties of cotton have contributed
to the observed increased variation in yield, there is often some misconception
about the aims of plant breeders, the methods they use and what they can achieve.
To solve the genotypeenvironment interaction has been made many attempts. Regression
techniques have often been used to estimate the stability of individual cotton
genotypes (Abebe et al., 1984a, b). The mean of genotypes in each environment
has been used as an index of that environment. The mean yield of each genotype
in each environment has also been linearly regressed on the mean yield off all
genotypes in each environment as suggested by Finlay and Wilkinson (1963). Eberhart
and Russel (1966) proposed that a stable genotype is considered to have the
highest yield over a broad range of environment, a regression coefficient value
of one and deviation mean square of zero.
The genotypexenvironment interaction was found to be significant for seed cotton and lint yield in many researches (Opondo and Ombakho, 1997; Palomo et al., 1998; Mert et al., 1999; Yuksekkaya and Unay, 2002; Unay et al., 2004). The stability analysis shown that genotype x environment interaction resulted from linear relationships (Sarma et al., 1994; Baloch et al., 1994; Opondo and Ombakho, 1997). The high yielding and stable cotton varieties for Aegean Region of Turkey were aimed to determine via regression techniques in this study. Materials and Methods Six cotton varieties, Carmen, Eksi 911, Gurelbey, Nazilli 84 S, Ozbek 142 S and Sahin 2000 (Gossypium hirsutum L.), were evaluated at four locations. These locations were NazilliAydin, SokeAydin, SaraykoyDenizli and MenemenIzmir in 2005. The experiment design was a Randomized Complete Block with four replications. The plots were 12 m length and consisted of six each row. The plant density for each experiment was 71420 plant ha^{1}. The harvested area was 28 m^{2}. The sowing dates of locations were between May 1 and May 15. The first and second pickings were realized at September and December, respectively. As preplanting fertilizer, 500 kg ha^{1} of 20200 was used. Then 130 kg ha^{1} of urea (4600) as preflowering fertilizer was used before first irrigation. All cultural practices (irrigation and plant protection) were conducted on commercial cotton farming operations across the region. The lint yield (kg ha^{1}) was determined by first and second picking the four center rows of each plot. Seed cotton was ginned at laboratory roller gin for lint percentage (%) in the Fiber Quality Laboratory at Nazilli Cotton Research Institute. In order to study the stability of lint yield and lint percentage performance of cotton varieties, the parameters viz. mean lint yield and lint percentage over all sites (mean), linear regression coefficient (b) and mean squares of deviation from regression coefficient (S^{2}d) were computed according to regression techniques proposed by Finlay and Wilkinson (1963) and Eberhart and Russell (1966). The relations between regression coefficients and lint yield and lint percentage were figure According to Arshad (1992). The confidence limits of the regression coefficients and lint yield and lint percentage on figure were estimated follows formula; Confidence limit = x (average)±tvalue. S_{x} Results and Discussion The results of variance analysis for lint yield and lint percentage are given in Table 1. There were significant differences between cotton varieties for both observed characteristics. The differences between environments (locations) were found significant for all characters. The significant genotypexenvironment interactions for all characters showed that the extent of such performance testing depended on the magnitude of genotypexenvironment interaction, which occurs when genotypes differ in their relative performance across environments (Bernardo, 2002). The genotypexenvironment interaction was also significant for seed cotton and lint yield in many studies (Opondo and Ombakho, 1997; Palomo et al., 1998; Mert et al., 1999; Yuksekkaya and Unay, 2002; Unay et al., 2004). These interactions must be significant to performing the stability analysis (Pederson, 1974). Table 1: 
ANOVA analysis for observed characters 

*, **; significant at probability level 0.05 and 0.01, respectively 
Table 2: 
Mean squares of deviation from regression (S^{2}d_{i})
and regression coefficients (b_{i}) for observed characters 

**; significant at probability level 0.05 and 0.01. respectively. 
The regression coefficient (bi) and deviation from regression (S^{2}d_{i}) as stability parameters and mean values for each variety are presented in Table 2. The lint yield was between 1768 kg ha^{1} (Gurelbey) and 2039 kg ha^{1} (Nazilli 84 S). For lint yield, bi values ranged between 0.580 (Eksi 911) and 1.484 (Carmen) and S^{2}d_{i} values ranged between 14.89 (Eksi 911) and 229.73 (Gurelbey). When both stability parameters were evaluated together, it was showed that the most stabile variety was Sahin 2000 because of its regression coefficient of one and deviation mean square of zero. In the mean of the lint percentage values, Ozbek 142 S had the highest values with 45.82%, while the lowest value was found for Sahin 2000 variety. The regression coefficient varied from 0.281 (Nazilli 84 S) to 1.724 (Carmen) for lint percentage. Deviation from regression were between 0.08 (Carmen) and 14.36 (Gurelbey). It was said that Eksi 911 was the stabile variety for lint percentage.
Eberhart and Russell (1966) evaluated that an ideal variety is one which has
the highest yield and a regression coefficient of one. Therefore, relationships
between regression and mean lint yield and lint percentage for six varieties
were evaluated. These results figured in Fig. 1 and 2.
Aegean Region standard variety and widely grown, Nazilli 84 S shown the best adaptation to all environments, while second commercial variety, Carmen and Ozbek 142 S and Sahin 2000 were group of the midadaptation to all environments. In study of Unay et al. (2004), Kurak 2, NAK 911 and Nazilli 84 S were found stabile genotypes. For Nazilli 84 S cultivar, this parallel result proved that Nazilli 84 S was the most stabile cultivar of Aegean Region. Gurelbey and Eksi 911 were outside of confidence limits for both lint yield and regression coefficient. When the varieties were evaluated for lint percentage, Ozbek 142 S was considered as the group of the best adaptation to all environments because of its 1.19 regression coefficient and the highest lint percentage (45.82%). Eksi 911 variety followed Ozbek 142 S with 0.944 regression coefficient and 44.43% mean value. This variety also considered with the lowest deviation from regression (0.18). Unfortunately, Aegean Region standard and commercial varieties, Carmen and Nazilli 84 S were put outside of this type classification as being outside the confidence limits.
In conclusion, genotypexenvironment interaction and stability analysis demonstrated
that there were significant differences among cultivars and environments. Therefore,
new cotton genotypes should evaluated for stability in yield and fiber properties
performance across the environments. Profitability of cotton production depends
more on lint percentage than other fiber properties in Turkey.
 Fig. 1: 
Lint yields and regression coefficients of six cotton varieties 
 Fig. 2: 
Lint percentages and regression coefficients of six cotton
varieties 
Therefore, Ozbek 142 S should be taken into consideration as suitable variety
for Aegean Region. Moreover, this variety can be easily used as a parent in
cotton breeding program.

REFERENCES 
Abebe, M., K. Yilma and G. Brhane, 1984. Genotypex environment interaction and yield stability in soghums of intermediate maturity. Ethiophian J. Agric. Sci., 6: 110.
Abebe, M., K. Yilma and G. Brhane, 1984. Yield potential and stability of early maturing soghum hybrids. Ethiophian J. Agric. Sci., 6: 5866.
Arshad, Y., 1992. The researches on some stability parameters used determining of general adaptation abilities of genotypes. M.Sc. Thesis, Aegean University, Institute of Scientific Science.
Baloch, M.J., B.A. Soomro, H. Bhutto and G.H. Tunio, 1994. Stability analysis for comparing cotton varieties. Pak. J. Sci. Ind. Res., 37: 528530.
Bernardo, R., 2002. Breeding for Quantitative Traits in Plants. Stemma Press, USA, ISBN 0972072403 pp: 369.
Cornelius, P.L., D.A. van Sanford and M.S. Seyedsadr, 1993. Clustering cultivars in to groups without rankchange interactions. Crop Sci., 33: 11931200. Direct Link 
Eberhart, S.A. and W.A. Russell, 1966. Stability parameters for comparing varieties. Crop Sci., 6: 3640. CrossRef  Direct Link 
Finlay, K.W. and G.N. Wilkinson, 1963. The analysis of adaptation in a plantbreeding programme. Aust. J. Agric. Res., 14: 742754. CrossRef  Direct Link 
Mert, M., M.E. Caliskan and E. Gunel, 1999. Genotype x environment interaction and stability analysis of seed cotton yield in upland cotton (Gossypium hirsutum L.). Turk. J. Field Crops, 4: 9195.
Miller, P.A., J.C. Williams and H.R. Roninson, 1959. Variety x environment interactions in cotton variety tests and their implications on testing methods. Agron. J., 51: 132134.
Misra, R.C. and B.S. Panda, 1990. Adaptability and phenotypic stability of improved soybean (Glycine max L.) varieties. Oil Crops Newslett., 8: 3637.
Opondo, R.M. and G.A. Ombakho, 1997. Yield evaluation and stability analysis in newly selected KSA cotton cultivars in Western Kenya. Afr. Crop Sci. J., 15: 119125. Direct Link 
Ozudogru, T. and N. Cakaryildirim, 2005. Cotton Situation and Outlook: 2005/2006. Agricultural Economics Research Institute, Ankara, Turkey.
Palomo, A., J. Santamaria and S. Godoy, 1998. Yield stability and fibre quality in cotton. Agric. Tecnica Mexico, 24: 145153.
Pederson, D.G., 1974. The stability of variety performance over years, II. Analyzing variety trials. Heredity, 33: 217228.
Samra, R.N., A. Roy, A. Roy and S.K. Samra, 1994. Phenotypic stability in upland cotton. Ann. Agric. Res., 15: 152155.
Unay, A., H. Basal, A. Erkul and Z. Yuksekkaya, 2004. Stability analysis of upland cotton genotypes to the aegean regon in Turkey. Asian J. Plant Sci., 3: 3638. CrossRef  Direct Link 
Yuksekkaya, Z. and A. Unay, 2002. The adaptation and stability analysis of some cotton (Gossypium hirsutum L.) genotypes under Aegean Region Conditions. Univ. Mustafa Kemal, J. Agric. Fac., 7: 3944.



