Agronomic Performance, Genotype X Environment Interactions and Stability Analysis of Grain Sorghum (Sorghum bicolor L. Moench)

MATERIALS AND METHODS

The basic material consisted of 10 grain sorghum lines introduced from ICRISAT; 5 cytoplasmic male sterile lines (A-lines; A-73, A-93, A-604, A-613 and A-614) and 5 restorers (R-lines; R-210, R-272, R-273, R-295 and R-92010). These parents were sown for hybridization on three different dates 1st, 15th and 30th June Qena Agric. Res. Farm, South Valley Univ. during the two summer seasons of 2006 and 2007 to avoid differences in flowering time and to increase hybrid seed. Each genotype was sown in three rows, 6 m long, 75 cm apart and 20 cm between plants within a row. Twenty five hybrid seeds were obtained.

The 25 hybrids comprising their 10 parents (five B-lines and five R-lines) and two checks (Shandaweel-1 hybrid and Dorado variety) were tested in field trials at two locations (Assiut Univ. Exper. Farm and South Valley Univ. Exper. Farm at Qena). In each location, the materials were sown at early (15th June) and late (15th July) planting dates in 2007 and 2008 summer seasons. Some physical and chemical properties of a representative soil sample of the experimental sites are presented in Table 1. Minimum and maximum daily temperatures during the growing season were obtained from the metrological station. at each location (Table 2). The total growing degree days GDD (Table 3), (base = 10°C) was calculated according to Saeed and Francis (1984) as follows:

Table 1:	Some physical and chemical properties of a representative soil sample of the experimental sites
*Each value represents the mean of two seasons

Table 2:	Monthly high, low and mean temperatures for the 2007 and 2008 growing seasons in Qena and Assiut

Table 3:	Dates of planting and total growing degree days (GDD) in Qena and Assiut where grain sorghum trials were conducted

Genotypes were arranged in a randomized complete block design with three replications in each environment. Single row plot of 6 m length, 60 cm apart and 20 cm between hills within a row was used. After full emergence, three weeks after planting, seedlings were thinned to two plants per hill; 168700 plants ha^-1. All culture practices were applied as recommended for grain sorghum production.

The traits recorded for each plot were days to 50% blooming (when anthers dehiscing half-way down the panicle of 50% plants), plant height; cm, 1000-grain weight; g and grain yield of 3 m in middle portion of the plot (converted to Mg ha^-1 at 15% moisture content).

Statistical Analysis
The combined analysis of variance was done according to (Gomez and Gomez, 1984) after carrying out homogeneity test. Phenotypic stability parameters; regression coefficients (b_i) and mean square deviations from regression (S²d_i) were calculated for grain yield and for each genotype using the model described by Eberhart and Russell (1966).

RESULTS AND DISCUSSION

The planting dates at the two locations used to evaluate performance of the F₁ hybrids and their parents in this study provided a range of variation in seasonal climate (Table 2, 3). Seasonal climate, edaphic and planting dates are three important cultivation environmental factors influencing crop performance, in general. In terms of genetic effect, environment often causes instability in non additive genetic effect as also obtained by Pathak and Sanghi (1992) and Can et al. (1997).

The combined analysis of variance indicated that year and location effects were significant (p<0.01) for all the studied traits; days to 50% blooming, plant height, 1000-grain weight and grain yield (Table 4), reflecting the differences in climatic and edaphic factors prevailing at the two locations. Mean squares indicated that the effect of locations was more important than that of years for all traits. Planting Dates (PD) show significant (p<0.01) differences for all traits as it would be expected for optimum and late sowing dates. Highly significant differences among genotypes and their partitions; parents, crosses, females and males for all traits, which showed the presence of genetic variability in this material. Male x female interaction also showed highly significant differences for all traits, indicating specific combining ability.

Table 4:	Combined analysis of variance for days to 50% blooming, plant height, 1000- grain weight and grain yield Mg ha-1 of 25 F1,s and their 10 parents over all environments
*, **Significant at 0.05 and 0.01 probability levels, respectively

Moreover, the relative of mean squares due to parents vs crosses was high and significant (p>0.01) for all studied traits, emphasizing great heterotic effects for these traits. These results are in agreement with these reported by Patel et al. (1987), Mahmoud (1997), Amir (1999), Ali (2000) and Hovny et al. (2005).

Genotypes x year interaction effects were highly significant for all traits except 1000-grain weight (Table 4). Genotype x location and genotype x planting date interaction effects were significant (p<0.01) for all traits, indicating that these traits differed between locations and planting dates among genotypes (Table 4). Interaction of genotypes with locations and planting dates was more important than that with years for all studied traits. Therefore, testing at more locations or dates at given locations should be done rather than testing in more years. Several workers stated that genotype x location and genotype x location x year interactions was more important than genotype x year interaction for sorghum yields (Obilana and El-Rouby, 1980; Saeed et al., 1984; Ali, 2000). Moreover, genotypes x years x locations x dates interaction was highly significant for plant height and grain yield, this indicates that it is essential to evaluate genotypes for such traits under different environments.

Environmental conditions at Assiut were good for sorghum production in both seasons compared to Qena as observed in Table 1 and 2. Early planting was better during growth, pollination and grain filling compared to late planting.

The data presented in Table 5, revealed that early planting registered earlier blooming, taller plant height, heavier 1000-grain weight and higher grain yield than late planting date in both locations in the two years. This suggests that both of temperature and edaphic factors could be playing a significant role in different plant growth stages. Also, Assiut was higher than Qena for all the studied traits over years.

The overall mean of crosses (Table 5) indicates that the crosses were earlier at Assiut compared to Qena location at early and late plantings. Furthermore, Assiut was better than Qena location in plant height, 1000-grain weight and grain yield Mg ha^-1 in both of planting dates. Likewise, the crosses mean of days to 50% blooming and plant height was highly significant earlier and taller than the two cheks; Shandaweel 1 and Dorado. Respects to 1000-grain weight, the crosses mean (25.5 g) was highly significant heavier than Shandaweel 1, but not from Dorado. However, 10 crosses showed highly significant heavier grain weight than Dorado (24.6 g). The overall mean of crosses for grain yield (7.55 Mg ha^-1) was significantly higher than for the best check Shandaweel 1 (7.1 Mg ha^-1), and ten crosses highly significant outyielded the best check (Table 5).

Comparisons of F₁ hybrids with each of their parents and the two commercially checks (Shandaweel-1 hybrid and Dorado variety) exhibited that most of hybrids were significantly earlier blooming, taller plants, heavier 1000-grain weight and higher grain yield than both their parents and checks for all studied traits overall environments (Table 5). These data supports the results obtained by Borgonovi (1985), Tadesse and Debelo (1995), El-Menshawy (1996), Mahmoud (1997), Amir (1999), Ali (2000) and Hovny et al. (2005).

Stability Analysis
The joint regression analysis of variance showed highly significant yield differences among F1 hybrids as well as their parents (Table 6). The relative yield performance of crosses and their parents varied from environment to another as was indicated by significant hybrids x environments and parents x environments interactions.

Table 5:	Average performance of 25 F1, s and their 10 parents at the two locations and two sowing dates over years for days to 50% blooming, plant height, 1000-grain weight and grain yield

Table 6:	The joint regression analysis of variance for grain yield for F1 crosses and their parents
*, ** Significant at the 0.05 and 0.01 probability levels, respectively

However, both F₁ hybrids and their parents showed highly significant differences in deviations from regression. These results suggest that the magnitudes of genotype x environment interactions in this set of materials are largely due to differential nonlinear response of genotypes to varying environments.

Yield stability of genotype is evaluated from estimates of stability parameters; b_i and s²d_i (Table 7, Fig. 1). Although the grain yield of F₁ hybrids and their parents were influenced by year, location and dates, the fluctuations for a majority of them were considered stable according to Eberhart and Russell analysis. (A-73 x R-272), (A-604 x R-92010) and (A-613 x R-210) could be considered the best stable hybrids. It has a b- values of 1.05, 0.95 and 0.95, respectively and produced yield higher than average of all crosses. This indicates that these hybrids may be suitable for growing in a wide range of environments. R-273 also seemed to be the best stable parent since it has a b-value of 1.03 and produced yield higher than average yield of all parents. These genotypes would be promising cultivars. Bakheit (1990), Ahmed (1993) and Ali (2000) also demonstrated that some genotypes which were superior in grain yield showed average stability. In addition, grain yield was considered consistently better in favorable environments with the cross (A-614 x R-272) and three parents; (A-613, R-272 and R-92010) because of their high b-value (b_i>1.0). This emphasizes that these genotypes are less environmentally sensitive. Similar results were obtained by Muppidathi et al. (1995) and Ali (2000).

Table 7:	Genotypes average performance over eight environments and stability parameters of 25 F1 crosses and 10 parents for grain yield
*, **Significantly different from unity for (bi) and from zero for (S2di) at 0.05 and 0.01 probability levels, respectively

Fig. 1:	Distribution of stability parameters for grain yield

However, grain yield for the crosses (A-73 x R-273), (A-93 x R-295), (A-604 x R-210) and (A-604 x R-272) and one parent; (A-73) had low b-values, and thus, was considered relatively better in less favorable environments (Table 6, Fig. 1). Meanwhile, grain yield for the crosses No. 4, 6, 7, 8, 21 and 25 had high b-values and thus, were considered unstable. This suggested that these crosses are unstable particularly in low temperature environments.

The results indicated that stability analysis for grain yield demonstrated that most F₁ hybrids had higher yields than their parents, but the parents were relatively more stable.

CONCLUSION

Results from this study showed that among the 25 F₁ crosses and their parents under investigation, genotype x year interaction effect was significant for days to 50% blooming, plant height and grain yield, except 1000-grain weight. Significant genotype x location and genotype x planting date interaction effects was observed for all traits. Significant genotype x year x location x planting date interaction effects was observed for grain yield. This suggests that genotypes need to be tested under several environments before being recommended for given zone. Three crosses; crosses (A-73 x R-272), (A-604 x R-92010) and (A-613 x R-210) and produced yield higher than average of all crosses. These crosses are considered as promising cultivars, it can be tested in large scale and produce the high yielded crosses commercially. Only one parent; (R-273) was the best stable and gave yield higher than average of all parents. This parent can be used in a breeding program for growing in a wide range of environments.