A total of 172 doubled haploid lines, were produced through another culture of the (F2-F3) generations derived from interspecific crosses among Korean Tongil-type (Semi-dwarfism, high-yielding potential with medium amylose and short grain) and Japonica (Earliness, low-medium amylose, soft and sticky endosperm) with African varieties and wild species (high biomass, disease resistance and low yield) were established at (Tissue culture laboratory Africa-Rice St. Louis). Of these doubled haploid lines were grown at Sahel Station in nursery test observation then after several series of selection and fixation, in advanced generation, 172 potential lines with the required adaptive traits were received in season 2018 and conducted in Observation Yield Trial; then again selected the superior 35 DH lines along with two checks namely Sahel177 and Arica3, were grown for two consecutive seasons (2019 and 2020) during the rainy season under irrigated condition; at White Nile Research Station Farm (latitude 14˚24'N and longitude 33˚22'E), Kosti of the Agricultural Research Corporation (ARC) Sudan. The soil of the experimental plots was classified as vertisol with high clay content (40% to 65%), less than 1% organic carbon, low in available nitrogen (0.03% total nitrogen) and medium in available P₂O₅ (406 to 700 ppm total phosphorus), pH values is slightly alkaline which ranging from 7 to 8.2. The climate was semi-arid (Table 1). The experimental plots were laid out in randomized complete block design (RCBD) with three replications. In both seasons deep plough, harrowing and leveling were practiced to prepare experimental area. The seeds were drilled on July 7^th and 5^th of 2019 and 2020 in turn, using a seed rate of 80 kg/ha. The plot size was (1 m × 6 m) with 0.2 m apart consisted 5 rows giving a total area of 6 m². Recommended fertilizer of Urea and Triple Super Phosphate (TSP) are at the rate of 129 kg of Nitrogen (N₂) per hectare and 43 kg of passphrase (P₂O₅) per hectare respectively. TSP was applied as basal dose during final land preparation. Nitrogen was top-dressed in two equal split doses at 21 days after sowing and the other before panicle initiation. Hand weeding was performed three times per season. All plots were irrigated at sowing and then at weekly intervals until it reached the flowering stage and then every 3 days. All other agronomic practices were applied as per the recommendation for rice production in the two seasons to raise a healthy rice crop.

Data were collected on days to flowering (days from sowing to time when the 50% of the plants were flowering) and days to maturity (days from sowing to time when 80% of the panicles reached full maturity). Then at harvest, plant height (cm), panicle length (cm), number of tiller/plant, number of filled grain/panicle, percent of unfilled grain/panicle, 1000 grain weight (g) and grain yield (t/ha) based on grain yield per plot were recorded. The data were collected according to standard evaluation systems for rice standard evaluation system (SES) [13] . Ten randomly selected plants in the net harvested area of each plot were used for data collection.

Analysis of variance (ANOVA) was carried out on the data to assess the genotypic effects and their interaction using the general linear model (GLM) procedure for randomized complete blocks design in [14] SAS (version 9.1). Estimates of variance components were generated and combined analysis of variance was done for the traits in which the mean squares were homogenous. The phenotypic and genotypic variances and their coefficients, heritability in the broad sense and genetic advance were computed according to the formula described by [15] . Combined over seasons were used to compute simple linear correlation coefficients., between grain yields and nine other traits.

The results showed that there is a presence of an acceptable amount of diversity

among the genotypes. This gives an opportunity for rice breeders to improve those traits through selection and hybridization to improve the desired traits. The range and mean of genotypes for all studied traits also indicated wide ranges of variation which also revealed a possible amount of variability among the genotypes (Table 2).

The broad-sense heritability is the relative magnitude of genotypic and phenotypic variances for the traits and it is used as a predictive role in selection procedures. This gives an idea of the total variation ascribable to genotypic effects, which are an exploitable portion of the variation. The characters that showed relatively high heritability estimates (≥60%) were days to flowering (90.3 and 64.1), grain yield t/ha (89.9 and 77.7), number of filled grain/panicle (84.0 and 85.8), panicle length (82.5 and 61.3) and number of tiller/plant (73.1 and 86.1) in both seasons respectively, indicating the optimizing of homogeneous of these lines and their improvement could be achieved by mass selection. Similar results were reported by [16] . This indicated that the selection of these traits would be more effective as compared to others. The moderate heritability estimate for grain yield was attributed to the fact that yield is a complex trait and is controlled by many genes. Since high heritability does not always indicate high genetic gain. Heritability with genetic advance considered together should be used in predicting the ultimate effect for selecting superior varieties [2] and [17] . High genetic advance and heritability were recorded for the number of tiller/plant number of filled grain/panicle 1000 grain weight and grain yield t/ha. These results suggested that these traits were primarily under genetic control and selection for these traits can be achieved through their phenotypic performance. High heritability estimates with low genetic advance observed for days to maturity, number of unfilled grain/panicle and panicle/m² indicated non-additive type of gene action and that genotype × environment interaction played a significant role in the expression of the traits. The genetic analysis of quantitative traits is a prerequisite for plant breeding programs, which can lead to a systemic method of design and to the appropriate planning of plant breeding strategies. The current study suggests that the PCV was higher than the GCV for all traits. This was also the case for all the traits observed in another study; [4] reported that the environmental effect on any trait is indicated by the magnitude of the differences between the genotypic and phenotypic coefficients of variation; large differences reflect a large environmental effect, whereas small differences reveal a high genetic influence. In this study, the small differences between the PCV and GCV for most of the traits, such as number of tiller/plant, number of filled grain per panicle, 1000 grain weight and grain yield t/ha, represented some degree of environmental influence on the phenotypic expression of these characters. It also suggests that selection based on these characters would be effective for future crossing programs. The other traits, which showed a higher difference between PCV and GCV, indicated that the environmental effect on the expression of those traits is higher and that selection based on these characters independently is not effective for further yield improvement.

*, **, *** Significant at 0.05, 0.01 and 0.001 probability levels, respectively.

The highest PCV was recorded for number of unfilled grain/panicle (123.62 - 132.23) and numbers of filled grains/panicle (98.02 - 122.4) in both seasons were also recorded by the following researchers ( [18] and [19] ).

The results further revealed that most of the traits exhibited wide range of variability (Table 3(a), Table 3(b)). The late flowering genotypes were KF170390, KF170502, KF170504 and KF170496 while, the early flowering genotypes were KF170399, KF170505 and KF170510. The late maturity genotypes were KF170390, KF170504 while the early maturity genotypes were KF170405 and KF170543. Generally, most of the genotypes showed early maturing period less than 120 days. This suggested the chance of selecting earliness genotypes which can escape terminal moisture and drought stress. The range for plant height was

74.5 to 103.0 cm with the local check (Arica3) as the tallest and genotype KF170489 as the shortest. According to IRRI irrigated rice plant height is classified as semi-dwarf (less than 110 cm), intermediate (110 - 130 cm) and tall (more than 130 cm). Based on this classification, in the present study 90.4% of the tested genotypes group is under the semi-dwarf class. This indicated that the tested genotypes had inherent variability in stature to develop lodging resistant varieties (semi-dwarf) that will have higher response to nitrogen application. Also reported variation in plant height by [19] [20] and [21] . The maximum panicle length value is 25.5 cm and the minimum is 19.4 cm for Arica3 and KF170495 respectively. Based on the IRRI irrigated rice classified argument, the present finding showed that there is enough medium variability for panicle length among the genotypes for improving panicle architecture and grain yield due to high association of this trait that determines the number of grains it can hold (Table 3(a)). The range for number of filled grain/panicle was 89 to 134 for genotypes KF170505, KF170498 and KF170506 respectively. The range of percentage of unfilled grain/panicle was 8 to 26 for genotypes KF170217, KF170510 and KF170390 respectively (Table 3(b)). Adequate number of fertile grains/ panicle and heavy grains are important traits, which should be considered in selection for high yield genotypes [4] and [19] . Relatively range performance revealed among the studied traits, grain yield (5.12 - 8.35 t/ha). The genotypes were significantly varied for grain yield and about 76.2% of the genotypes had higher grain yield. Among the genotypes, KF170506, KF170509, KF170542, KF170530, KF170543, KF170500 and KF170510 were the top yielders with grain yield of 8.35, 8.26, 7.58, 7.45, 7.40, 7.30 and 7.25 t/ha, respectively. Thus, the existence corresponding of enough variability among genotypes was highly significant. Hence, it offers a better scope for further selection of promising DH genotypes in rice breeding program particularly selection based on yield is reliable.

Yield is a complex product being influenced by several independent quantitative characters. Breeders always look for variation among traits to select desirable types. Some of these characters are highly associated among themselves and with grain yield.

The analysis of the relationships among these characters and their associations with grain yield is essential to establish selection criteria. When more characters are involved in correlation study it becomes difficult to ascertain the characters which really contribute toward yield. The relationships existing between 10 quantitative traits represented as simple correlation coefficients are presented in Table 4. Correlation of yield and other traits is important in indirect selection for high yield improvement in crop genotypes [22] . There was a highly significant and positive correlation of grain yield with number of filled grain per panicle (0.65), number of tillers per plant (0.64.), number of panicle/m² (0.54), panicle length (0.28), 1000 grain weight (0.23) (Table 4). Conversely, grain yield exhibited negative significant correlation with number of unfilled grain per panicle (−52) and days to 50% (−0.16). The significant and positive correlation with grain yield are a strongly indication that these traits are major factors in improving grain yield; also suggests that selection directed towards these characters will be effective in ensuring high grain yield in doubled haploid rice. These results collaborate with the finding of [23] who observed a positive and significant correlation between grain yield and number of filled gain per panicle. Also, these results were in agreement with that reported of [24] and [25] . The negative correlation obtained between number of unfilled grain per panicle and days to 50% heading indicate that grain yield can be improved by selecting early maturing genotypes.

*, ** and *** significant at 0.05, 0.01 and 0.001 probability respectively. See footnotes of Table 1 for traits explanations.

Among 37 elite rice genotypes, cluster analysis resulted in four clusters (Figure 1). Distribution pattern indicated that maximum genotypes (13) were aggregated in cluster B, while minimum (5) were grouped into cluster A.

Cluster-wise mean values of ten different traits were computed to assess superiority of cluster, which could be considered in the improvement of various characters [26] [27] . Mean performance of different clusters for the characters manifested that genotypes with maximum paddy yield were accumulated in cluster A, whereas genotypes with minimum were in grouped cluster D.