1. Introduction
Sorghum is a diploid C4 cereal crop which was domesticized in Africa particularly Ethiopia and Sudan. It has 2n = 20 chromosome and genome size of 750 Mb [1]. Sorghum mainly reproduces through selfing with 15% - 30% out-cross depending on the nature of head compactness and shape [2]. It is the fifth major cereal crop in the world and third in Ethiopia which is the most important dry land crop grown for food, feed, fuel, fodder and other traditional uses [3]. Its production over the years is 1.49 tones ha−1, 1.74 tones ha−1 and 2.71 tones ha−1 in the World, Africa and Ethiopia respectively [4].
In the initial stages, knowledge on combining ability and heterosis of parental materials is essential for a productive breeding program dedicated in development of high yielding and dual purpose sorghum hybrids, especially in the areas where drought is one of the major factor for forage and grain production [5]. Though there has been a high level of genetic diversity, the potential of new locally developed inbred lines for hybrid cultivar development has no longer yet been exhaustively assessed. To satisfy the farmers' need, it is vital to maximize the production and productivity through developing hybrids with excessive grain and fodder yield through established, formal and continuous breeding programs. Currently many thousands of inbred lines are found in Ethiopia which are developed in Ethiopian sorghum research program via a non-stop crossing program. Those elite lines aren’t assessed exhaustively for hybrid production since they’re progenies of elite traces which are having exact tendencies such as higher yield. In general, information on heterotic performance and combing ability on Ethiopian elite sorghum lines is limited.
In this study hybrids and elite lines were evaluated to assess their performance, heterotic pattern and heterobeltiosis for yield, forage and yield components by identifying best heterotic parents and good combiner parents for sorghum hybrid breeding program under moisture stress areas in Ethiopia.
The specific objectives include:
1) To identify hybrids under moisture stress environments for grain yield and forage
2) To estimate and determine heritability, heterosis and combining ability (GCA and SCA) of the hybrids for important agronomic traits.
2. Materials and Methods
2.1. Description of the Study Area
The study was conducted at the dry lowland sorghum growing area at Miesso which is situated at 1394 m.a.s.l. and a coordination of 9˚14'N, 40˚45'E. Miesso has a dry lowland climatic condition with average maximum and minimum temperatures of 34˚C and 10˚C respectively and the average annual rainfall of the area is 790 mm with dominant Vertisol soil.
2.2. Genetic Materials
The experiment was conducted for a total of 70 F1 hybrids which was derived using 2 standard female A-lines viz. ETX623 and ICSA21 crossed with 35 inbred lines (pollinators). In this experiment both the hybrids and parents (male and female) including two hybrids (ESH-1 and ESH-4) chosen based on their yield performance and adaptation and one recently released better biomass producing sorghum OPV variety (Argiti) were used as a standard check. In total the experiment consisted of 110 genotypes. List of genetic material can be found on http://dx.doi.org/.
2.3. Experimental Design and Trial Management
The experiment was laid down in alpha lattice design with two replications. During planting, the seed was manually drilled into 5 meters long 2 row plots with a spacing of 0.75 m between rows in total each plot has of 7.5 m2 area. Three weeks later of sowing, the seedlings were thinned to 0.20 m distance between plants. Nitrogen and Phosphorus fertilizers were applied at the recommended rates of 46 kg/ha P2O5 and 54 kg/ha. Phosphorus was applied in the form of DAP during planting and urea when the seedling reached at 5 cm height.
2.4. Data Collection
Data was collected on growth and phenological parameters (DTE (Days to emergence), DTF (Days to flowering), DTM (Days to physiological maturity), PHT (Plant height), NGL (Number of Green leaves), NSL (Number of Senescence leave), CHL (Chlorophyll content)) and yield and yield components (PL (Panicle length), TGW (Thousand grain weight), GY (Grain yield), Biomass, HI (Harvest index)) using electronic data collection tools (tablets, barcode readers, and computer program to weigh grain yield automatically) to avoid error and for data precision.
3. Statistical Analysis
3.1. Analysis of Variances
Analysis of variance for single location was done using the following model:
where;
• µ is the overall (grand) mean, is the overall (grand) mean,
• τi is the effect due to the ith treatment, (
)
• γj is the effect due to the jth replication, and, (
)
• ρl(j) is block within replicate effect
• εijl is the error term where the error terms, are independent observations from an approximately Normal distribution with mean = 0.
The analysis was performed using R: R core team (2018). Genotypes were considered as fixed effects, replications and blocks within replications as random effects.
3.2. Combining Ability Analysis and Estimation of Heterosis
Analysis of variance for combining ability was carried out using mean values across environments [6], to test the significance of differences among the genotypes including crosses and parents [7]. The sum of squares for hybrids was further partitioned into variation due to males, female and males * females interactions. The mean squares due to males and females were tested against the mean squares due to males * females, and the latter were tested against the pooled error. The mean squares due to environment * males and environment * females were tested against the mean squares due to environment * females * males, and the latter was tested against the pooled error. Estimate of GCA variances (σ2GCA) and SCA variances (σ2SCA) were obtained [8]. Mid-parent, better parent and better check heterosis were estimated and tested by working out the standard errors and tested by t test at 5% and 1% (Hays et al. 1955).
Proportional contribution of Females, Males and their interaction were found:
,
and
Correlation among variables was computed using R software (R Core Team, 2018).
3.3. Estimation of Genetic Components and Heritability
The phenotypic and genotypic variance components and coefficient of phenotypic (PCV %) and genotypic coefficients of variation (GCV %) was estimated based on the method suggested by [9].
Heritability in broad sense for all characters was computed using the formula given by [10].
4. Result and Discussion
4.1. Variability of Genotypes for Grain Yield and Biomass Component Traits
An understanding of grain and biomass yield with good quality for sorghum grain and yield is essential to breeding and cultivation of sorghum to produce sorghum grain and forage for livestock. Sorghum has recently been viewed as the ideal candidate feedstock crops for generation of both forage and fuel in the form of bioethanol in addition to its grain production. This crop has low input requirements and particularly well-adapted to marginal growth conditions such as water deficits, salinity, alkalinity, and other constraints which are came up of strange for other crops. The analysis of variance for yield and yield component traits revealed that the parents and their hybrids involved in this study differed significantly for all the characters. The mean square values of grain yield and yield component traits of parent (females and males) and their hybrids are presented in Table 1.
![]()
Table 1. Mean squares for yield and biomass component traits.
*significant at (P < 0.05), **significant at (P < 0.01).
All genotypes comprising parents, checks and hybrids are significantly different from each other for most traits except number of green leaves, total fresh weight and harvest index. Hybrids are significantly different for only for yield, head length, panicle width and number of productive tillers. Parents are significantly different for grain yield, plant height, head length, number of productive tiller and total dry biomass. Similarly, male lines are significantly different for grain yield, plant height, head length and number of productive tillers. The interaction of males and females is significantly different for grain yield, thousand seed weight (P < 0.05), panicle width and number of productive tiller (P < 0.01). Parents and Hybrids are significantly different for all traits except thousand grain weight, number of green leaves and panicle width. This revealed that hybrids can have better yield than OPVs.
4.2. Magnitude of Heterosis and Hybrid Performance for Biomass and Yield Related Traits
Magnitude of heterosis showed as for yield and biomass component traits was varied from traits to traits as well as from genotype to genotype (Table 2). For the case of MPH for grain yield, it was ranged from 162.1 to −57.3 (%). The highest heterosis (162.1%) was recorded for hybrid 106 × 94 and the lowest MPH (−57.6%) was recorded by 107 × 75. Among all hybrids 24 hybrids showed significant negative heterosis for grain yield and 14 hybrids were showed significantly positive heterosis. For better parent heterosis which was ranged from
*significant at (P < 0.05), **significant at (P < 0.01).
111.1 to −66.1 (%) for hybrids 107 × 75 and 106 × 94 respectively, 20 hybrids showed that significant positive heterosis and 45 of hybrid exhibited negatively significant heterosis. 29 hybrids have negative significant heterosis from the standard check one (ESH-1) and 29 hybrids exhibited significant positive heterosis from the same standard check. Standard heterosis in comparison to standard check two (ESH-4), 68 of hybrids showed positive significant heterosis and the rest 2 hybrids were showed insignificant positive heterosis. In the case of standard heterosis in consider of standard check three (2005MI5064) only 2 hybrids showed none significant negative heterosis and 65 of them exhibited significant positive heterosis. In this case, since standard check 3 is an OPV we can look at the yield advantage of hybrids over a OPVs is much better.
In the case of total fresh weight of biomass, the magnitude of mid parent heterosis was ranged from 233.2 for 107 × 93 to −37.2% for hybrid combinations of 107 × 82. Similarly, better parent heterosis was ranged from 206.8% to −57.9% for hybrid combination of 107 × 93 and 107 × 82 respectively. The hybrid combination 107 × 84 (134.7%) showed the higher magnitude of significant and positive heterobeltiosis (Better parent heterosis) for total dry biomass weight. Hybrid of 107 × 105 showed highest significant positive standard heterosis over checks 109.3 (ESH-1), 181.1 (ESH-4) and 106.5 (2005MI5064).
4.3. Combining Ability Analysis for Grain Yield and Biomass Related Traits
From Table 3 analysis of variances for combining ability analysis showed that except days to flowering and total fresh biomass weight are significantly different for all traits of genotypes.
![]()
Table 3. ANOVA for combining ability of the studied traits.
*significant at (P < 0.05), **significant at (P < 0.01); GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
Mean squares due to Females were significant different for only DTF, HL, GY, NH and HI. Mean square due to Male lines is also significantly different for traits only PHT, HL and NH. Whereas Hybrids (interaction of Males × Females) PW, NPT, GY and HI were also significant (P < 0.01 or P < 0.05). Similarly, Parents and Hybrids are significantly different from each other for all measured traits except HI.
The total variance is a contribution of genotype as in their interaction. In this case, Females and Males including their interaction have its own contribution for the total variance in combining ability analysis. So, the highest contribution is due to males for GY and the contribution of Females for the total variance due to TDBM is null and 0.4 for TFW. Means that Male lines are responsible for the increment of variations for all traits and female lines were less in their biomass. Similarly, the second higher contribution is due to the interaction of males and female lines for TFW. The variance for TFW is almost equally contributed by both Male lines and the interaction of Males and Females (Table 4). We can conclude that Male parents are more likely important to increase grain yield and total dry biomass simultaneously.
4.4. General and Specific Combining Ability Effects of Yield and Biomass Component Traits
Both female lines showed that highly significant general combining ability (GCA) effect for traits of grain yield, head length and total fresh biomass weight at probability level of (P < 0.01). Since, female lines are two the magnitude of GCA effect is equal and the only difference is direction of magnitude. That is one female line will be positive and the other female line will be negative (Table 5 and Table 6). Both female lines exhibited significant GCA effect at P < 0.05 probability level and female line 106 was negative and female 107 showed positive GCA effects. This means, female line 106 was flowered earlier than female 107.
All male lines exhibited highly significant GCA effects except male line 79 for Grain yield. Among these male lines 18 of them showed highly significant negative GCA effects and 16 of them showed positive significant GCA effects for GY. Male line 81, 87, 92 and 94 exhibited positive highly significant magnitude GCA effects of 1.31, 1.53, 1.46 and 1.78 respectively. Male lines 71, 74, 75, 77 and 83
![]()
Table 4. Proportional contribution of females, males and their interactions to total variance.
GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
![]()
Table 5. GCA effects of GY and biomass components for male and female lines.
*significant at (P < 0.05), **significant at (P < 0.01). GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
![]()
![]()
Table 6. SCA effects of GY and biomass related traits.
*significant at (P < 0.05), **significant at (P < 0.01). GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
exhibited negative significant GCA effects with magnitude of −0.97, −0.8, −1.4, −1.05 and −0.97 respectively. Based on this situation male lines 81, 87, 92 and 94 can be selected for good positive combiner for varietal development. Generally, 16 male parents are highly significant (P < 0.01) positive combiner and the rest 18 males are highly significant negative combiner for Grain yield. In this case positive GCA effects are selectable to increase grain yield and those which exhibited highly significant positive GCA effect can go for further varietal development based on the magnitude of their GCA effects.
For Plant height male lines 73 (−33.7), 75 (−48.6), 80 (−33.4), 86 (−24.5), 93 (−25.9), 96 (−33.1) and 102 (−27.6) showed that highly significant negative GCA effect. Contrarily, 12 male lines exhibited highly significant positive GCA effects. These are male line 74 (15), 81 (15.9), 82 (23.4), 84 (27.4), 89 (17.2), 90 (16.5), 92 (21.7), 94 (16.9), 95 (21.0), 98 (15.8), 99 (16.8) and 105 (47.6). Male line 94 exhibited positive highly significant GCA effect for Plant height and also it has highly significant positive GCA effect for GY. So, male line 94 can be selected for both high plant height and GY as a good combiner male parent. In general, 17 male lines are good combiner based on their GCA magnitude for increasing plant height (Table 5).
In the case of total fresh biomass weight, among all male lines 8 male lines exhibited positive significant GCA effect and 7 male lines showed significant negative GCA effects. The rest 20 males have explored none significant GCA effects. Male line 94 exhibited highest significant positive GCA effect valued 380.1 (P < 0.01) followed by male lines 99 (306.2), 93 (270.4), 73 (267.5), 79 (220.2), 100 (183.5), 81 (178.30 and 102 (170.8).
For the trait total dry biomass weight (TDBM), 8 males exhibited positive significant GCA effect and 9 male lines showed negative significant GCA effect. Male line 94 exhibited highest GCA effect (192.9) followed by male lines 73 (159.3), 100 (134.2), 79 (126.9) and 93 (116.2).
Male line 94 exhibited positive highly significant GCA effect for PHT, GY, total dry biomass weight (TDBM) and total fresh biomass weight (TFW). This implies male line 94 can be select to improve Biomass contest and GY as the same time and also male line 94 can be select to develop dual purpose varieties. All male and female parents showed none significant general combining ability (GCA) effects.
Estimates of specific combining ability (SCA) effects for GY and other agronomic traits for all hybrids computed are presented in Table 6. SCA is used to designate deviations of certain crosses from expectations on the basis of the average performance (GCA effects) of the parents involved. In the current study, among seventy single cross hybrids that demonstrated significant and positive SCA effects for GY, cross combination of female line and male line of 106 × 78 (1.3), 107 × 99 (1.4), 107 × 104 (1.5), 107 × 105 (1.2) and 106 × 94 (1.0) had the highest SCA effect. Among all Crosses Female line 106 × Male line 94, which exhibited the highest GY mean, was among the top five crosses with highly significant and positive SCA effect of 1.0. These crosses contain parents (female line 106 and male line 94) with high GCA effects for GY, indicating the increased concentration of favorable alleles. On the other hand, Vasal et al. (1992) argued that positive SCA effects indicate that lines are in opposite heterotic groups while negative SCA effects indicate that lines are in the same heterotic group. For TSW, DTF, PHT, NGL, HL, HI and PW almost all except few crosses showed none significant SCA effects, indicating the ability of the crosses to produce single cross hybrids having increased performance of these traits is failed (Table 6).
A cross of female line 107 and male line 105 exhibited highest significant positive SCA effect valued as 458.8 for total fresh biomass weight (TFW). a cross of these parents showed higher GY mean performance, these parents can be selected for dual purpose hybrid production. That means, female line 107 and male line 105 exhibited reasonable GY mean and highest positive significant SCA effect for TFW and this implies by crossing these two parents we can get reasonable GY and high Biomass product for feed and forage use. Across of these two parents also exhibited highly significant positive SCA effect for total dry biomass (TDBM) weight (Table 6).
Similarly, for trait of Number of productive tiller (NPT) which can contribute to improve both GY and Biomass contents, across of 107 × 71 (2.4) exhibited high positive significant SCA effect followed by 106 × 87 (1.6). Productive tillers have positive contribution to increase grain yield as well as biomass content simultaneously (Table 6).
4.5. Genetic and Phenotypic Correlation for GY and Biomass Traits
Genetic correlation for many traits of GY and Biomass indicates that as there was a significant genetic correlation between traits and improving one trait can help to improve the other correlated traits as the same time. The correlation matrix showed that, GY was perfectly correlated with TFW and this may be disproved the fact that many works define as biomass was not positively correlate with GY. In similar way GY showed negatively strong correlation with DTF that indicates the earlier flowered Genotypes could exhibit higher GY. HI also strongly correlated with GY and this indicates that to improve HI of the given genotype can improve GY as the same time. GY was correlated strongly with NH, TSW, NGL, PW and PHT. Some findings indicate that GY was not positively correlated with PHT and other biomass componential traits. But the present study finds that GY was strongly correlated with PHT and Biomass component traits (Table 7). TFW and TDBM strongly correlated with PW and improving genotypes for higher PW can improve genotypes for Biomass contents (TFW and TDBM) simultaneously.
Phenotypic correlation indicates that GY was Significant negatively correlated with DTF and this implies that genotypes that can flower earlier could have good yield as compared to genotypes that flower lately. In the other way GY was correlated positively with biomass component traits (PHT, TFW, TDBM and HI) and where strongly positively correlated with TFW and TDBM. PHT also showed positive significant correlation with TFW and TDBM. TFW shows positive strong correlation with TDBM. Panicle width also exhibited positive strong correlation with total fresh biomass weight and total dry biomass weight as well (Table 8).
![]()
Table 7. Genetic correlations for GY and Biomass component traits.
*significant at (P < 0.05), **significant at (P < 0.01). GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
![]()
Table 8. Phenotypic correlations of GY and biomass component traits.
*significant at (P < 0.05), **significant at (P < 0.01). GY = Grain yield, TSW = Thousand seed weight, DTF = Days to flowering, PHT = Plant height, NGL = Number of green leaf, HL = Head length, PW = Panicle width, NPT = Number of productive tillers, TFW = Total fresh weight, TDBM = Total dry biomass, HI = Harvest index.
5. Summary and Conclusions
Analysis of variance for all genotypes showed that highly significant difference for all measured traits except number of green leaves, total fresh biomass and harvest index. This revealed that there was genetic variability between the tested genotypes. The F1 hybrids are significantly different only for yield, head length, panicle width and number of productive tillers whereas hybrid parents are significantly different for grain yield, plant height, head length, number of productive tiller and total dry biomass. There was a difference in magnitude of heterosis which showed for yield and biomass component traits that varied from traits to traits as well as from genotype to genotype. Similarly, there was yield advantage over the OPV check in yield and biomass production.
For total dry biomass weight, male line 94 exhibited highest GCA effect (192.9) followed by male lines 73 (159.3), 100 (134.2), 79 (126.9) and 93 (116.2). Male line 94 exhibited highest significant positive GCA effect valued 380.1 (P < 0.01) followed by male lines 99 (306.2), 93 (270.4), 73 (267.5), 79 (220.2), 100 (183.5), 81 (178.30 and 102 (170.8) for total fresh biomass weight. Male line 94 exhibited highly significant positive GCA effect for PHT, GY, total dry biomass weight (TDBM) and total fresh biomass weight (TFW). This implies male line 94 can be used as a gene source to improve forage and grain yield. That means, it could be used to develop dual purpose sorghum varieties.
Information stating the degree of association between traits could serve for the simultaneous improvement of those traits. In specific in the improvement of quantitative traits such as drought tolerance, it is suggested to use secondary traits that have higher heritability. The correlation between and among the various yield and other agronomic traits and the biomass components was strong and significant while some others have weak association. Among the studied traits, grain yield and biomass component traits were significantly and positively correlated each other. That means GY, PHT, DTF, PW and HL were found significantly and positively correlated with biomass yield related traits (TFW, TDBM and HI).