PERFORMANCE OF SORGHUM RECOMBINANT INBRED LINES ( RIL ) DEVELOPED for RAIN-FED AREAS OF SUDAN

Sorghum (Sorghum bicolor (L) Moench) is the most widely produced and consumed cereal crop in Sudan. However, productivity is low since the crop is produced in favorable and unfavorable environments where the crop suffers from drought stresses at different growth stages. In the present study, six sorghum inbred lines developed by local breeding program and two commercial checks were evaluated for grain yield potential, yield stability, some important agronomic characters and grain quality properties. Series variety trials were conducted at Elobeid and Suki, Damazin areas of Sudan, during three consecutive rainy seasons. The selected locations represent low, medium and high rainfall areas of Sudan. The trial was laid out in a randomized complete block design. The results revealed that mean squares of genotypes, seasons, locations, location x season, location x genotypes, season x genotypes and season x location x genotype interactions were highly significant (P=0.01) for grain yield. The sorghum genotypes Edo 34-23-4, Edo 26-18 and Edo 16-dwarf produced substantially higher grain yields than commercial checks and the trial mean. Their percentage yield increase ranged from 5% to 75% over commercial checks. The stability analysis revealed that the above mentioned Edo-genotypes had high yield potentials and were stable across a wide range of agricultural conditions. Moreover, the same Edo-lines showed early maturing compared to selected commercial checks and also the Edo-lines possessed good food grains and were market preferred and acceptable for making quality kisra (fermented sorghum pancake-like flatbread). The Edo developed lines also possess the acceptable grain quality in addition to moderate physical grain characteristics such as protein content, fat acidity and moisture content.


Introduction
In Sudan, sorghum (Sorghum bicolor (L.) Moench) is the stable food crop utilized in various forms for humans and animal feed.The crop contributes with about 70% of total grain produced in the country.It ranks first in terms of total area cultivated as well as the total tonnage produced.However, the cultivated areas and production vary year after year due to many biotic, abiotic, technical and policy factors.The area is reported to be 4-8 million hectares with an average of 5.5 million/ha, about 90% of cultivated area is rain-fed; while total grain production varies between 3 million and 4.5 million tonnes with an average of 0.6 tonnes/ha (Mohamed et al., 2011;FAO, 2006;Ibrahim and Abbas, 2006).Soil moisture content stands out as one of the most important factors of biotic factors limiting sorghum productivity.Due to recent climate changes, the annual rainfall has become erratic and poorly distributed.Rainfall was reported to be decreasing in eastern Sudan from 1960 to 1990 and according to a high correlation between rainfall and sorghum yield (r =0.7) indicating that crop yields were dependent on rainfall (Larsson, 1996).The analysis of rainfall collected in twelve meteorological stations in Sudan located between latitudes 11º and 20º N showed a clear decrease in annual rainfall over the last 30-40 years (Osman and Ibrahim, 1998;Mohamed, 1998).This climate change seriously affected the traditional sorghum growing areas of northern Gedarif, Gezira, Sennar, White Nile States as well as northern parts of Kordofan and Darfur States.This area is estimated to be higher than 50% of the total sorghum cultivated area of the country.
In mentioned dry areas (250-400 mm), farmers used to grow their own local sorghums, which are low yielding and suffer drought stresses at almost any stage of crop growth.The outcome is either low yield or poor development of straw and chaff.In fact, farmer's sorghum is losing ground in these important areas.The national breeding program has developed medium maturing, high yielding varieties and hybrids such as Wad Ahmed, Tabat, and H. D-1.These varieties and hybrids were not suitable for these low rainfall regions because they required soil moisture between 550 mm to 650 mm which was not available in drought areas of the country.Accordingly, early maturing varieties were developed by national breeding program through conventional breeding.To achieve that, two crosses were made between Geshiesh x Wad Ahmed and Eriana x Dorado, and Gew and Edo lines were generated (respectively), followed by the pedigree method.This study is aimed to evaluate the performance of six sorghum genotypes developed by a national breeding program in three crop cultivated areas; Elobied, Suki and Damazin representing low, medium and high rainfall areas of Sudan.

Material and Methods
Germplasm: The research plan was designed to develop early maturing and high yielding varieties suitable to different agricultural conditions.The original cross (F1) was made at Shambat, Khartoum North, research station in 1995.Then, the plant materials were advanced further using the pedigree breeding method (Table 1 Experimental procedure: multi-location sorghum variety trials were conducted for three seasons at three locations, namely: Suki, Damazin and Elobied representing medium, high and low rainfall areas of Sudan.Land was prepared by disc harrowing and ridging.Trials were laid out in a randomized complete block design with three replications.A plot size of 4 rows of 5-meter length with an interrow spacing of 0.8 m was used across testing sites.Planting was performed at the first two weeks of July each season depending on the effective onset of rains.About six to seven seeds were planted in each hole on the side of the ridge 20 cm apart.Plants were then thinned to two plants/hill three weeks after seedling emergence, and kept weed free using hand weeding.A fertilizer dose of 19g/m 2 urea was applied at the irrigated site after thinning.Then, the traits were measured as follows: 1. Agronomic traits: flowering and plant height; a. Days to 50% flowering: recorded as the number of days from the date of seedling emergence to the date when 50% of the plants in the plot reached anthesis at least halfway down the panicle.

b.
Plant height (cm): measured from 5 random plants in each plot, from ground surface to the tip of the main panicle just before the harvest time.2. Grain yield: weighted in kg/plot for the advanced yield trials and demonstration plots.Heads were cut, sun dried, threshed, weighed and adjusted to kg/ha.
3. Grain quality traits: grain physical and chemical properties were assessed at Food Research Center (FRC), Shambat, Sudan.Moisture, crude protein, fat acidity and total acidity were determined according to the methods of the AOAC (1990).The hardness of seeds was determined using a hardness tester No. 174886, Kiya Seisakusho Ltd, Tokyo, Japan.One hundred (100) seeds from each sample were used to assess the hardness.The kisra (fermented sorghum pancake-like flatbread) making quality was assessed by a panel of kisra makers and tasters who were asked to give their judgment/scores on 1 -kisra colour 2 -taste 3 -consistency and 4keeping quality.The entries included in the test were Wad Ahmed (control), Edo 26-18, Edo 16-dwarf and Edo 34-23-4.
Data analysis procedure: Each location in a particular year was treated as a distinct environment.Analysis of variance was performed on data collected from individual trials in each environment following the standard procedures for RCBD design.Combined analysis of variance was performed on data from each location over years, each year over locations and overall years and locations.The F-test for individual and combined analysis was made assuming random environments and fixed varieties (Finlay and Wilkinson, 1963;Lin and Binns, 1994;Osman andIbrahim, 1998, Steel andTorrie, 1980).Stability analysis of grain yield was made according to the model of Eberhart and Russel (1966).Then stable varieties across different environments were identified considering a regression coefficient equal to one (b i =1), deviation from regression equal to zero, (S 2 d i = 0) and the acceptable genotype should yield above the general mean.Analysis of variance was executed by using MSTAT computer package (MSTAT-C, 1991) for days to 50% flowering, while stability parameters for grain yield were obtained by using IRRISTAT model (IRRI, 2005).

Results and Discussion
Plant height (cm): Highly significant (P = 0.01) differences were also observed between entries over years and locations, (Table 3).The result is similar to EL Naim et al ( 2012) who evaluated eight sorghum farmer varieties in two locations of North Kordofan state of Sudan, which found significant effect on plant height at the two locations.The trial means ranged from 124 cm for Edo 16-dwarf to 205 cm for Yarwasha.Significant differences (P = 0.01) in plant height were between Edo 16-dwarf (124 cm), Edo 26-18 (135 cm) and Edo 34-23-4 (163 cm).Edo 34-23-4 responds positively to high moisture level that it reaches > 180 cm which renders it uncombinable.Plant height in sorghum is positively correlated with grain yield, but shortness of the plants facilitate combine harvesting which is becoming essential for reducing production cost.Also the result correspond with Amare, et al (2015) who found plant height had positive direct effect through analysis of sixteen sorghum varieties planted in two locations.It stated that when there is positive association between the main yield components breeding would be very effective, however when these characters have negative association, it would be no easy to practices simultaneous selection for them while developing new variety (Nemati, et al. 2009).
Grain Yield: Rainfall amount (in mm) during the testing period (2005-2007) at three locations is presented in Table 2.The results showed superiority of three of the developed lines (Edo 34-23-4, Edo 26-18 and Edo 16-dwarf) over the commercial checks across nine environments including low, medium and high rainfall areas of the country over three years (Table 2).
Individual Yield Trials: Separate analysis of variance for each location in each year (environment) showed significant differences (P = 0.01) among entries in all of the nine environments.The grain yields for each entry in each environment are given in Table 4.As expected, the short maturing test entries produced higher grain yields in low rainfall environments compared to the commercial check Wad Ahmed which excelled them in high rainfall years at Suki and Damazin.The entry Edo 34-23-4 was superior to Wad Ahmed in four environments and comparable to it in five environments while it was superior to Yarwasha in eight environments and comparable to it in one environment.Edo 26-18 was superior to Wad Ahmed in three environments and comparable to it in six environments while the reverse is true in case of Yarwasha.Edo 16-dwarf was superior to Wad Ahmed in three environments, similar in four and inferior to it in two environments.It was superior to Yarwasha in five and comparable to it in four environments.When comparing the Edo lines; Edo 34-23-4 was ranked top followed by Edo 26-28 and Edo 16-dwarf as second and third, respectively.
Combined analysis of variance was made for grain yield in the nine rain-fed environments (Table 4).Evidence was presented suggesting that the years of testing could be reduced by increasing the number of test locations (Kambal and Mahmoud, 1978).The mean squares of genotypes, years, locations, year x location, location x genotype, genotype x year and year x location x genotype interactions were highly significant (P > 0.01).The location x year interaction indicated that locations ranked differently in different years.The significant (P = 0.01) genotype variance indicated that genotypes differed in their genetic potential for grain yield (Table 5).The significance of genotypes x location interaction is an indication that some genotypes were location specific.However, the genotype x year interaction was highly significant indicating the differential response of the genotypes in the years (Table 5).Similar results for significant effect due to variety × location (interaction effect) is reported by Ezzat et al. (2010), for days to 50% flowering, plant height, 1000 grain weight and grain yield.The three order interactions; genotype x location, genotype x year and genotype x location x year were showing that the genotypes responded differently certain locations in different years.Table 5. Mean squares obtained from combined analysis of variance for grain yield measured in three locations over three years.Adaptability and stability: An evaluation of varieties and hybrids aims at identifying genotypes that consistently produce stable yields over a range of diverse environments.The stability analysis was based on Eberhart and Russel (1966) model (b = regression coefficient) as a parameter of the genotype response or adaptation.They also used the deviation from regression (S2d) to estimate stability.The regression coefficient was employed by Finlay and Wilkinson (1963), who used the variety and the magnitude of mean yield to identify adaptability and stability.They both defined an ideal or an average stable variety as one with b = 1, S 2 d = 0 and the average mean is higher than the overall mean of grain yield of the trials.The stability analysis results showed clear differences in slopes of the regression lines (Table 6).Some regression coefficients (b i ) exceeded unity while others were less than one.According to Eberhart and Russel (1966), the genotypes Edo 34-23-4 and Edo 26-18 had high regression coefficients (b i ) of 0.934 and 0.809, and with minimum deviation from regression of 0.033 and 0.684, respectively, and also mean grain yield of these two genotypes was above the general mean of the trial, indicating that these two genotypes had better response in variable environments (Table 6).The aim of selection is to produce a population that has the mean value greater (or less) than the mean value of the parent population.This difference should be due to differences in genotype and not due to the environment (House, 1985).Regression coefficients greater than one with higher deviation (S 2 d i ) value were observed for Gew 22-15 and Gew 3-2 with mean grain yield below the grand mean grain yield, indicating that these two lines were not stable under adverse conditions, but may respond better in favorable environments (Table 6).The genotype Edo 16-dwarf had a regression coefficient b i = 0.522 and a high deviation (S 2 d = 10.977) and mean grain yield above the trial mean, so it could be considered adaptable and suitable for general cultivation.The commercial variety Wad Ahmed had the largest b i (2.927) and significantly different deviation (166.548)but outyielded the general trial mean, confirming that it was only responsive to good growing conditions (Table 6).

Source of variation
Agronomic Traits: Results of analysis of variance of individual and combined trials over two years have revealed significant and very significant (P = 0.05, 0.01) differences for varieties and varieties x years, for 50% flowering, plant height and plant population.
Grain Quality: Results of assessment of some grain physical and chemical properties and kisra acceptability of entries under investigation are shown in Table 7.
Protein content (%): It ranged from 14.57 for Yarwasha to 9.42 for Edo26-18, however the protein contents of all the test entries are higher than the minimum levels specified by the codex (a minimum value is 8%) (Table 7).
Table7.Physical and chemical properties of sorghum grains.Grain hardness: Grain hardness influences kisra qualities.A good kisra is usually produced from soft, medium soft and medium hard grains (Bureng, 1995).In this investigation, grain hardness was measured using a hardness tester, the force power applied (kg) for each sample of seed was used as a measure for hardness (Table 7).The analysis indicated that Edo 34-23-4, Yarwasha and Wad Ahmed were soft, but not significantly different from Edo 26-18 and Gew 22-15.Others may be considered medium hard grains.
Fat acidity: A level of fat acidity above 50 mg KOH/100g causes rancid flavour in kisra product.The values obtained for all entries in this study ranged from 7.36 to 18.28.These values are far below the minimum level of 50 mg KOH/100g and therefore all are acceptable (Table 7).
Total acidity (titerable acidity): It is associated with volatile acid during baking.Excessive loss of acids causes poor kisra taste and could result from poor fermentation or grain variety.The optimum total acidity is found to be 1% lactic acid (Bureng, 1995).The values obtained in this study were significantly (P ≤ 0.05) below optimum and consequently very low acid losses are expected for all entries (Table 7).

Conclusion
Six advanced sorghum genotypes and two commercial check varieties have been evaluated for grain yield across nine environments under rain-fed conditions in the country.The experimental data demonstrated the superiority of Edo 34-23-4, Edo 26-18 and Edo -16-dwarf over Wad Ahmed and Yarwasha in terms of grain yield and some other agronomic traits.
Edo-34-23-4: It was superior to all tested lines in terms of grain yield.It was reasonably stable (b; 0.934), tan sorghum plant type, providing high protein content (11.13 %) and medium seed size, and clean, white grain flour.
Edo 26-18: Its grain yield potential, stability and adaptability were similar to Edo 34-23-4.It has also tan plant color and is short maturing.The grain is white, medium with brown testa, but provides lower protein content (9.42%) compared to other genotypes.It was medium soft, of low acidity and could provide acceptable kisra just similar to W. Ahmed.
Edo 16-dwarf: It has the good yield potential under rain-fed conditions but performs better under irrigation and high rainfall.It was moderately stable responding better to favourable environments.It is tan sorghum plant.The grain is white with brown testa, medium soft and medium in size.It had an average protein content (10.26%), with low fat acidity and was scored as having satisfactory kisra quality.

Table 1 .
).The plant material evaluated in this study consisted of eight entries, including six advanced breeding lines, namely: Gew 22-15, Gew 3-2, Edo 16dwarf, Edo 34-23-4 Edo 26-18 and Edo 27-3-25 and two commercial varieties: Wad Ahmed and Yarwasha.The six breeding lines were developed by the national breeding program by crossing local x exotic and local x local open pollinated varieties.Designation, source and pedigree of tested material.

Table 2 .
Rainfall amount (mm) during the testing period(2005)(2006)(2007)at three locations.The highest grain yield (2,722 kg/ha) was produced by genotype Edo 34-23-4 at Suki in 2006 crop season, while the lowest yield (557 kg/ha) was produced by genotype GEW 3-2 at Elobeid in 2007.The comparison level of superiority of the tested material across the nine rain-fed environments is shown in Table4.

Table 3 .
Mean number of days to 50% flowering and plant height (cm) over two seasons 2006/2007 at three locations.

Table 4 .
Mean grain yield (kg/ha) of sorghum genotypes for each environment.
*; ** P = 0.01 and SK = Suki, DM = Damazin and OB = Elobied Means having a letter within the same column are not significantly different according to DMRT.Entry rank is bracketed.

Table 6 .
Mean grain yield of each genotype, its regression b i (slope) and the deviation from regression (S 2 d i ).