Page 1

Journal of Environmental Biology

©Triveni Enterprises, Lucknow (India)

Free paper downloaded from: www.jeb.co.in

Genotype X environment interaction and stability analysis of

cotton yield in Aegean region of Turkey

April 2006, 27(2) 427-430 (2006)

For personal use only

Commercial distribution of this copy is illegal

Fatih Killi1 and Eyup Harem2

1University of Sutcu Imam, Faculty of Agriculture, Department of Field Crops, Kahramanmaras 46060, Turkey

2Nazilli Cotton Research Institute, Aydin 09900, Turkey

(Received: 18 October, 2004 ; Accepted: 17 May, 2005)

Abstract: The objective of this study was to determine genotype X environment (GE) interaction and stability of cotton genotypes,

and effect of different environments on seed cotton yield to understand its adaptation to varying environments. Fourteen cotton

genotypes were evaluated at four locations across Aegean region of Turkey in 1997 and 1998. Genotypes were tested by two

stability parameters as linear regression coefficient (b) and deviations from regression (S2d). Significant differences were observed

for the mean yields in the 8 environments. Mean seed cotton yield ranged from 4.58 to 5.80 t ha-1. Genotypes showed significant

interaction with environments. Regression coefficients ranged from 0.23 to 1.46, and deviations from regression were significant for

only four genotypes. It was concluded that three high yielding cotton genotypes SG-1001, SG-125 and DLP-5409 were found to be

stable genotypes.

Key words: GE interaction, Stability analysis, Adaptation, Cotton yield.

Introduction

Genotype X environment (GE) interactions are of

major concern to plant breeders for developing improved

cultivars. A cultivar, to be commercially successful, it must

perform well across the range of environments in which allowed

to grow. The presence of GE interactions reduces the

correlation between phenotype and genotype, and makes it

difficult to judge the genetic potential of a genotype (Sharma et

al., 1987). Further, the stability of a cultivar refers to its

consistency in performance across environments and is

affected by the presence of GE interactions. In the presence of

significant GE interactions, stability parameters are estimated

to determine the superiority of individual genotypes across the

range of environments. Genotype X location, genotype X year

and genotype X location X year interaction components were

found to be significant for seed cotton yield in many researches

(Abou-El-Fittouh et al., 1960; Abd El Latif et al., 1975; Singh

and Gill, 1982; Killi and Gencer, 1995a; Unay et al., 2004)

In Turkey, cotton is grown in different environments

such as Aegean, Mediterranean and Southeastern Anatolian

regions, which differ in climate, soil, insect, disease and cultural

conditions spatially and temporally. Therefore, a search for best

suited cultivar requires evaluation of the comparative yield and

fiber properties of cotton varieties in replicated trails at different

locations over a period of years (Killi and Gencer, 1995b). The

present paper aims to determine the range of variability for

seed cotton yield of cotton cultivars and to estimate their

stability parameters for identifying superior cotton cultivars.

Materials and Methods

Fourteen cotton genotypes (SG-1001, SG-501, SG-

125, SG-404, Sicala 3/2, DLP-5690, DLP-5614, DLP-5409,

DLP-20, DLP-50, S-314, C-1518, N-84 and Cun S2) were

tested in 1997 and 1998 field trials at four locations: Manisa,

İzmir, Aydın and Denizli (Fig.1). The experimental design was a

randomized complete block with four replications at each

location. Plots were six rows of 12 m length spaced 0.7 m apart.

Each plot in an experiment had a plant population of 95 000

plants ha-1. Accepted cultural practices and fertilization based

on individual soil test were applied at each location. All plots

were seeded between 30 April and 14 May 1997 and between

24 April and 28 May 1998. The plots were harvested between 7

and 17 October 1997 and between 2 September and 26

October 1998. At maturity, seed cotton yield were obtained from

an area 2.8 m wide and 10 m long (28 m2) of the central four

rows of each plot. Stability parameters were estimated by the

method described by Eberhart and Russell (1966).

Analysis of variance procedure (Comstock and Moll,

1963) was adopted to test the significance of location, year,

genotype, and first and second order interactions assuming the

year and location effects as random and genotype effect as

fixed. The stability analysis of variance and stability parameters:

linear regression coefficient (b) and deviation from regression

(S2d) of genotype means over environment index were

computed as suggested by Eberhart and Russell (1966). For the

regression analysis of variance, the residuals from the

combined analysis of variance were used as a pooled error to

test the S2d values. A significant F value would indicate that the

S2d was significantly different from zero. The hypothesis that

each regression coefficient equaled unity was tested by the t

test using the standard error of the corresponding b value.

Results and Discussion

According to the variance analysis, differences among

genotypes for seed cotton yield, and year, location and year X

location interaction were highly significant (Table 1). The first

Page 2

Fatih Killi and Eyup Harem

428

Fig. 1: Various locations (Manisa, İzmir, Aydın and Denizli) of the cotton field trials conducted in Aegean region of Turkey in 1997-98.

Table – 1 : Analysis of variance for seed cotton yield for 14

cotton genotypes based on 2 year data from four

locations in Aegean region of Turkey.

Source of variation df Mean squares

Year (Y) 1 10.18**

Location (L) 3 117.22**

Y x L 3 56.36**

Replication (Y, L) 24 3.04

Genotype (G) 13 20.30**

G x Y 13 5.57**

G x L 39 3.79**

G x Y x L 39 3.03**

Error 312 1.15

** Significant at the 0.01 probability levels.

order interactions (genotype X year and genotype X location)

and second order interaction (genotype X location X year)

components were also significant. Abou-El-Fittouh et al.,

(1960), Abd El Latif et al., (1975), Singh and Gill (1982), Killi

and Gencer (1995a) and Unay et al., (2004) evaluated cotton

genotypes for seed cotton yield over years and locations and

found all main effects and interactions to be significant.

Average seed cotton yield for the 14 genotypes

ranged from 4.42 (C-1518) to 5.53 (DLP-5614) t ha-1 (Table 2).

Seven of these genotypes (SG-1001, SG-501, SG-125, SG-

404, DLP-5690, DLP-5614 and DLP-5409) gave higher seed

cotton yield than the grand mean yield (4.97 t ha-1). According

to Eberhart and Russell (1966), an ideal cultivar would have

both a high average performance over a wide range of

environments plus stability. Becker et al., (1982) regarded

mean square for deviation from regression to be the most

appropriate criterion for measuring phenotypic stability in an

agronomic sense because this parameter measures the

predictability of genotypic reaction to environments. Langer et

al., (1979) suggested that the regression coefficient was a

measure of response to varying environments. The regression

coefficient (b) values of the 14 genotypes used in this study

ranged from 0.23 to 1.46 (Table 2). These variations in b values

suggested that the 14 cotton genotypes responded differently to

the different environments. Among the genotypes tested, the

regression coefficients of only 3 genotypes (DLP-5690, DLP-

5614 and C-1518) significantly deviated from 1. Deviations from

regression were significant for 4 genotypes (DLP-50, S-314, C-

1518 and N-84). Relationship between the regression

coefficients and mean seed cotton yields for 14 cotton

genotypes is shown graphically in Fig 2. The diagrammatic

presentation (Fig. 2) of stability parameters showed that three

genotypes (SG-1001, SG-125 and DLP-5409) excelled in yield

performance. Their regression coefficients and deviations from

regression were not significantly different from unity and zero,

respectively (Table 2). Therefore, these three genotypes formed

the group of the best adapted genotypes to all environments.

When the other genotypes were considered, SG-501, SG-404,

Sicala 3/2, DLP-50 and N-84 were defined as mid-adaptation to

all environments while DLP-20, S-314 and Cun S2 had poor

adaptation (Fig. 2). Three genotypes, DLP-50, S-314 and N-84,

had high deviation from regression indicating sensitivity to

environmental changes. Due to high values of S2d, these

genotypes are expected to give good yield under favorable

environmental conditions. Two genotypes, DLP-5690 and DLP-

5614, produced above - average yield with a regression

Page 3

G E interaction and salability of cotton yield 429

Fig. 2 : Mean seed cotton yields and regression coefficients of 14 cotton genotypes.

Table – 2 : Mean seed cotton yields and estimates of stability parameters in 14 cotton genotypes based on 8 environments (4 locations, 2

years) in Aegean region of Turkey.

Code number Genotypes

(t ha-1)

1 SG-1001

2 SG-501

3 SG-125

4 SG-404

5 Sicala 3-2

6 DLP-5690

7 DLP-5614

8 DLP-5409

9 DLP-20

10 DLP-50

11 S-314

12 C-1518

13 N-84

14 Cun S2

Mean

*,** Significantly different from 1.0 for regression coefficients and from 0.0 for the deviation mean squares at the 0.05 and 0.01 levels of probability, respectively.

Table – 3: Mean seed cotton yield and values of environmental index of various locations of the cotton field trials conducted in

Aegean region of Turkey in 1997-98.

Locations Mean seed cotton yield (t ha-1)

Aydin-1997 5.08

Denizli-1997 4.69

Manisa-1997 5.19

İzmir-1997 5.12

Aydin-1998 4.58

Denizli-1998 4.60

Manisa-1998 5.80

İzmir-1998 4.70

Mean 4.97

Mean yield

b

0.97

1.06

1.20

1.00

1.03

0.23**

0.59*

0.86

0.91

1.13

1.23

1.46**

1.18

1.15

1.00

S2d

580

280

110

180

470

200

290

310

450

1490**

1590**

1170*

1020*

390

5.22

5.08

5.14

5.06

4.95

5.18

5.53

5.12

4.81

4.90

4.70

4.42

4.89

4.58

4.97

Values of environmental index

0.11

-0.28

0.22

0.15

-0.39

-0.37

0.83

-0.27

Page 4

Fatih Killi and Eyup Harem

430

coefficient significantly less than 1 (b= 0.23 and 0.59

respectively), but had small deviations from regression. The

genotype C-1518 showed significant deviation from regression

with a regression coefficient significantly higher than 1 (b=

1.46). Therefore, it was defined specifically adaptation to

favorable environments. The average seed cotton yield

performances of locations over varieties were different. Mean

seed cotton yield for the 8 environments (4 locations, 2 years)

ranged from 4.58 to 5.80 t ha-1 (Table 3). Location Manisa gave

the highest mean seed cotton yield in both years. It was the

best environment for cotton production in Aegean region.

It was concluded that the three high yielding cotton

genotypes SG-1001, SG-125 and DLP-5409 were found as

stable and thus, these would be recommended for

environmental conditions of Aegean region.

References

Abd El Latif, H.M., S.A. Nassar, I.M. Gomaa, N.S. Abd El Aziz and

A.M. El Solami: Effect of genotype X environment interactions on

seed cotton yield and lint percent. Agric. Res. Review, 53,1-7

(1975).

Abou-El-Fittouh, H.A., J.O. Rawlings and P.A. Miller: Genotype by

environment interactions in cotton, their nature and related

environmental variables. Crop Sci., 9, 377-381 (1960).

Correspondence to :

Dr. Fatih Killi, Assoc. Prof.

Kahramanmaras Sutcu Imam University

Agricultural Faculty, Field Crops Department

46060 Kahramanmaras, Turkey

E-mail: fakilli@ksu.edu.tr

Tel.: +90-344-223 76 66

Fax: +90-344-223 00 48

Becker, H.C., H.H. Geiger and K. Morganstern: Performance and

phenotypic stability of different hybrid types in winter rye. Crop

Sci., 22, 340-344 (1982).

Comstock, R.E. and R.H. Moll: Genotype-environment interactions. In:

Statistical genetics and plant breeding, (Ed: W.D. Hanson and

H.F. Robinson). Proc. Natl. Acad. Sci.-NRC Pub. no. 982, pp.

164-196 (1963).

Eberhart, S.A. and W.A. Russell: Stability parameters for comparing

varieties. Crop Sci., 6, 36-40 (1966).

Killi, F. and O. Gencer: Determination of adaptation abilities of some

cotton genotypes to environment using different stability

parameters. Tr. J. Agric. Forestry, 19, 361-365 (1995a).

Killi, F. and O. Gencer: Variety X environment interaction and

heritability for yield and fiber properties in cotton. Proceedings of

FAO Interregional Cooperative Research Network on Cotton, 18-

24 Septebmer, Adana-Turkey, pp. 85-88 (1995b).

Langer, S., K.J. Frey and T. Bailey: Associations among productivity,

production response, and stability indexes in oat varieties.

Euphytica, 28,17-24 (1979).

Sharma, R.C., E.L. Smith and R.W. McNew: Stability of harvest index

and grain yield in winter wheat. Crop Sci., 27, 104-108 (1987)

Singh, T.H. and S.S. Gill: Genotype X environment interaction in

upland cotton. Crop Improv., 13, 215-213 (1982).

Unay, A., H. Basal, A. Erkul and Z. Yuksekkaya: Stability analysis of

upland cotton genotypes to the Aegean region in Turkey. Asian J.

Plant Sci., 3, 36-38 (2004).