ABSTRACT
The analysis of genotypes by environmental interactions in essential in plant breeding programme because it indicates the stability of genotypes before release. Highly stable genotypes are desirable. In the present study, the stability of performance of selected 14 grain maize genotypes were evaluated at four different locations in Malaysia for grain yield, number of kernels per ear row and 100-grain weight. For average main yield, number of kernels per ear rows and 100-grain weight ranged from 4541 to 6110 kg ha-1, 32.6 to 38.0, 23.1 to 26.4 g respectively. The different methods of stability analyses were in close agreement with each other in revealing the stability of the genotypes evaluated. Selected GxA was identified as the genotypes with highest grain yield and most stable performance. TWC-5 was the lowest yielding and the most unstable. For number of kernels per ear row, SC-2, TWC-2 and Selected GxA were the most stable. For 100-grain weight, Suwan 1 was the best performer, revealing average stability. The high stability of Selected GxA was probably due to its nature of having a broader genetic base, being a synthetic population, as opposed to its hybrid counterparts.
PDF Abstract XML References Citation
How to cite this article
DOI: 10.3923/ajps.2003.743.747
URL: https://scialert.net/abstract/?doi=ajps.2003.743.747
INTRODUCTION
The aim of a plant breeding programme is to improved genotypes for some predefined target population of environments. Population genotypes are usually evaluated in different environments before the desired ones are selected. Genotypes x environment (G x E) interaction, which is associated with the differential performance of materials tested at differential performance of materials tested at different locations and in different years and its influence on the selection and recommendation of cultivars has long been recognized (Lin et al., 1986; Geletu and Tsegayeaand, 1995; Annicchiarico, 1997a). Selection of genotypes is based on the assessment of their phenotypic performance at a number of locations provides useful information to determine their adaptability and stability (Crossa, 1990). Measuring G x E interactions helps to determine an optimum breeding strategy of either to breed for specific or wide adaption, which depends on the expression of stability under a limited or wide range of environments (Romagosa and Fox, 1993; Annicchiarico, 1997b; Yue et al., 1997) Moreover, genotype x location interaction allows the grouping of relatively similar sites in relation to genotypic performance (Eberhart and Russell, 1966 Romagossa and Fox 1993; Annicchiarrico, 1997b).
Stability refers to non-erratic performance with respect to agronomic traits and stable cultivars shows minimal or low interactions (Allard and Bradshaw, 1964). Stability also denotes consistency in rank relative to other cultivars in a given set of environments (Yue et al., 1997). Yield stability is the ability of a genotype to avoid substantial fluctuations in yield over a range of environments (Heinrich et al., 1983). Varieties with high and stable performance are desirable (Becker and Leon, 1988). Instability is caused by the different expression of traits across environments or G x E interactions, which decrease progress from selection, since they reduce the association between genotypic overcome the problem of interment very complicated (Romagosa and Fox, 1993).
The maize breeding programme conducted at Universiti Putra Malaysia since 1987 is directed towards the development of hybrid varieties (Saleh, 1998). The objectives of the present study were to investigate the G x E interment on selected grain maize genotypes with reference to grain yield, number of kernels per ear row and 100-grain weight and to determine the stability of the of the genotypes in plantings at four locations in Malaysia.
MATERIALS AND METHODS
Fourteen selected grain maize genotypes were grown and evaluated for performance at four locations, viz. Pandang Rengas (Peark), Rhu Tapai (Terengganu), Sungai Udang (Melaka) and Serdang (Selangor). This study was conducted from March, 2000 to December, 2001. The genotypes used in this study were single cross, three way cross and double cross hybrid as well as synthetic populations developed from maize breeding programme previously conducted at Universiti Putra Malaysia (Saleh and Sujiprihati, 1997), evaluated together with two check cultivars (Table 1).
Table 1: | Fourteen genotypes evaluated in the study and their pedigrees |
At each location, the experiment was conducted in a randomized complete block design (RCBD) with four replications. Similar agronomic practices were applied at each location. The middle three rows of 3m long were used as the harvest area of each plot. Ten plants from each plot were chosen randomly as samples for measurements of number of kernels per ear row and 100-grain weight.
Data for grain yield, number of kernels per ear row and 100-grain wight were subjected to combined analyses of variance. Location effects were considered as random variables while the genotype effects were treated as fixed. Different stability parameters were used and compared. The method of Eberhart and Russell (1996) employs the regression of individual means on the environmental index which is defined as the mean of all genotypes grown in that environment. The regression coefficient, b and the deviation from the regression s2d are the parameters of stability. Both b and s2d have been used widely in previous studies (Langer et al., 1979; Yue et al., 1997). Ecovalance (Wi) developed by Wricke (1962) and stability variance statistics) (θ2i) developed by Shukla (1972) and then other commonly used parameters of stability like coefficient of determination, r2; environmental variance, s2 (Becker and Leon, 1988; Nissila, 1992) and the grouping technique of mean against coefficient of variation (c.v.), were also considered.
RESULTS AND DISCUSSION
The genotypes evaluated revealed inconsistent performance for grain yield, as was reflected by their different rankings at the different locations.
Table 2: | Mean grain yields (kg ha-1) of the grain maize genotypes evaluated and their ranks (in parantheses), at four locations |
Table 3: | Mean number of kernels per ear row of the grain maize genotypes evaluated and their ranks (in paranthesis), at four locations |
This showed the varied performance of the genotypes and their superiority was dependent on environment. The highest yield was given by Selected GxA with mean over locations of 6110 kg ha-1 followed by GxA (location mean of 5848 kg ha-1) (Table 2). The lowest yield was observed on SC-4 (4541 kg ha-1). The check varieties. The check varieties, Putra J-58 and Suwan 1 produced mean over locations grain yields of 5840 and 4909 kg ha-1, respectively.
For number of kernels per ear row, DC-1 and SC-2 gave the highest mean values (39.5 and 38.1, respectively), while TWC-4 gave the lowest (33.9) (Table 3). These were comparable to that of hybrid check, Putra J-58 (37.7), but much higher than that of the open-pollinated check, Suwan 1 (34.1).
Table 4: | Mean 100-grain weight (g) of the grain maize genotypes evaluated and their ranks (in parantheses), at four locations |
Table 5: | Results of the combined analysis of variance for performance of grain genotypes at four locations |
*,** significant at p≤0.05 and 0.01, respectively |
For 100-grain weight, Suwan 1 had the highest mean value (26.4 g), while TWC-2 had the lowest (23.1 g) (Table 4). No genotype had average 100-grain weight over locations higher than Suwan 1.
Results of the combined analysis of variance for grain yield, number of kernels per ear row and 100-weight indicate that the genotype x location (GxL) interaction effects were statistically significant (Table 5). This demonstrates the presence of genotype and environmental differences governing the expression of these traits and the significant contribution of GxL interactions in influencing genotype performance. Partitioning of interaction effects using Eberhart and Russells (1966) regeneration method showed that Location (Linear) and GxL (Linear) effects were significant for all traits. This represents the heterogeneity of the regression coefficients.
Results of the different analyses of stability and adaption on the genotypes are shown in Table 6 for grain yield, Table 7 for number of kernels per ear row and Table 8 ranged from 0.40 to 1.62 (Table 6), while those for 100-grain weight ranged from 0.52 to 1.57 (Table 8). The deviations were low for all traits, showing the good fit of the regression model.
Eberhart and Russell (1966) proposed that an ideal population of genotype is one which has the highest yield over a broad range of environments, a regression coefficient (b) value of 1.0 and deviation mean square of zero. Based on result of the joint regression analysis, Selected GxA was classified as highly stable over environments because its regression coffined was close to unity (b=1), having lowest deviation from regression (s2d) and the highest grain yield (Table 6). Although TWC-4 showed b value of 1.0 s2d value was very high and therefore considered to be unstable. Generally, regression coefficient value of more than one (b>1) indicates that the genotype tested is of low stability, while value close to one indicates that the genotype is more stable and consistent in the performance across environments, DC-1 and Suwan 1 had regression coefficients significantly less than 1, which indicate their above-average adaptability and lack of response to environmental changes for grain yield, with relatively small fluctuation in performance between poor and good environments. This trend was also reported in sorghum (Heinrich et al., 1983) and tetraploid wheat landraces (Tesemma et al., 1998) to indicate their abilities to adopt to environmental changes.
Genotypes indicating low environmental variances (s2) and low coefficients of variation (c.v.) are considered stable (Lin et al., 1986). Low c.v. values and environmental variances for grain yield were shown by Selected GxA, DC-1 and Suwan 1, confirming their high stability. The unstable cultivars, TWC-5 and DC-1 had the highest s2 values and high c.v.s. for grain yield. Generally, genotypes with high b values tended to have high s2 and c.v.s. and vice versa. The r2 values f or grain yield were mostly high, indicating environmental effects as the main determinants of phenotypic performance. Ecovalance (Wi) measures the genotype environment effects for each environment sum of squares also increases. Lower values are indicators of genotype stability (Wricke, 1962). For grain yield, the lowest Wi was found on Selected GxA. The lowest θ2i value for grain yield was also revealed by Selected GxA (Table 6), thus further confirming that Selected GxA was the most stable and adaptable genotype.
For number of kernels per ear row, SC-2, TWC-2 and Selected GxA were the stable genotypes, with b<1 and s2d (Table 7). The b value for TWC-5 was significantly lower than 1, showing its lack of response to changes in environment and above average adaptability. Lower c.v.s. and low s2 were shown by SC-2, Selected GxA and Suwan 1, indicating their stability. The lowest Wi was found on TWC-5.
For 100-grain weight, SC-3 was the most stable genotype, with b<1 and low s2 (Table 8).
Table 6: | Stability parameters for grain yield of grain maize genotypes evaluated at four locations |
Table 7: | Stability parameters for number of kernels per ear row of grain maize genotypes evaluated at four locations |
Table 8: | Stability parameters for 100-grain weight of grain maize gernotypes evaluated at four locations |
TWC-4 and Putra J-58 had b values significantly less than 1, indicating the constant expression of the traits under the range of environments and better adaption to poor environments. SC-1, TWC3, TWC-5, DC-1, GxA, Selected GxA and Suwan 1 had b values > 1, which indicate their responsiveness to changes in environmental conditions and specific adaption to favourable ones. TWC-4 had the lowest c.v. value, while SC-4 the highest. The highest r2 and lowest Wi values were found on TWC-4. This indicates that environmental effects were mainly responsible for the expression of the traits and similar patterns of adaption and stability were reflected for all traits.
Selected GxA was found to be a highly stable genotype, while SC-3 and TWC-2 should also be noted as stable genotypes that have importance towards efforts in choosing suitable cultivars.
ACKNOWLEDGMENT
The authors wish to thank SEAMEO, SEARCA and the Malaysian of Science, Technology and the Environment (IRPA Grant No. 01-02-04-0353) for their financial support.
REFERENCES
- Allard, R.W. and A.D. Bradshaw, 1964. Implications of genotype-environmental interactions in applied plant breeding. Crop Sci., 4: 503-508.
Direct Link - Annicchiarico, P., 1997. Joint regression vs AMMI analysis of genotype-environment interactions for cereals in Italy. Euphytica, 94: 53-62.
Direct Link - Becker, H.C. and J. Leon, 1988. Stability analysis in plant breeding. Plant Breed, 101: 1-23.
CrossRef - Eberhart, S.A. and W.A. Russell, 1966. Stability parameters for comparing varieties. Crop Sci., 6: 36-40.
CrossRefDirect Link - Heinrich, G.M., C.A. Francis and J.D. Eastin, 1983. Stability of grain sorghum yield components across diverse environments. Crop Sci., 23: 209-212.
Direct Link - Langer, I., K.J. Frey and T. Bailey, 1979. Associations among productivity, production response and stability indexes in oat varieties. Euphytica, 28: 17-24.
CrossRef - Lin, C.S., M.R. Binns and L.P. Lefkovitch, 1986. Stability analysis: Where do we stand? Crop Sci., 26: 894-900.
CrossRefDirect Link - Nissila, E., 1992. Yield stability parameters of barley under finnish conditions. Acta Agric. Scand. Sect. B Soil Plant Sci., 42: 152-157.
CrossRefDirect Link - Shukla, G.K., 1972. Some statistical aspects of partitioning genotype-environmental components of variability. Heredity, 29: 237-245.
CrossRefDirect Link - Tesemma, T., S. Tsegaye, G. Belay, E. Bechere and D. Mitiku, 1998. Stability of performance of tetraploid wheat landraces in the Ethiopian highland. Euphytica, 102: 301-308.
CrossRef - Yue, G.L., K.L. Roozeboom, W.T. Schapaugh and W.T. Liang, 1997. Evaluation of soybean cultivars using parametric and non-parametric stability estimates. Plant Breed., 116: 271-275.
Direct Link