The study was conducted at Burie campus of Debre-Markos University, Ethiopia located at latitude of 10˚42'N and longitude of 37˚4'E. The altitude of the area is 2091 meter above sea level, annual rainfall is 1500 mm and the mean temperature is 22˚C [14]. A total of 36 intact male sheep of three genotypes, Rutana, Gumuz and Washera of about 8 - 10 months of ages, 12 from each genotybe were purchased from local markets where each breed is found. The age was estimated by asking the owners and checking the pattern of eruption of the incisor teeth of the sheep [15]. Animals were quarantined for 3 weeks during which time they were vaccinated for ovine pasteurellosis and sheep pox. Animals were also injected with Ivermectin as a broad spectrum treatment against internal and external parasites, and de-warmed with Albendazole mainly against the adult stages of internal parasites. At the end of the quarantine, each sheep was weighed and placed in an individual pen and acclimated to the environment and experimental condition for 2 weeks, which was followed by 90 days of feeding trial and 10 days of digestibility trial.

Experimental treatments were arranged in a three by two factorial arrangement in a completely randomized block design. There were three genotypes (Rutana, Washera and Gumuz sheep) and two levels of concentrate supplementation (300 and 450 g DM/day/head of concentrate mix). Animals within genotype were blocked based on their initial body weight and randomly assigned to one of concentrate supplementation levels each consisting six animals. The experiment feeds consisted of mixed sward natural pasture hay as a basal diet and concentrate mix as a supplement. The hay was manually chopped to about 2.5 cm sizes and fed ad libitum ensuring a refusal of 20% with the amount offered being adjusted once every 3^rd day. The concentrate mix consists of 45% wheat bran, 26% maize grain and 29% Noug seed cake, and was formulated to have about 22.3% CP to fulfill the requirements of growing lambs [16]. The chemical composition of the concentrate mix and the hay is given in Table 1. The concentrate was offered twice a day in two equal portions at 0800 and 1600 hours. Clean water and salt lick were available all the time.

DM = Dry matter; OM = Organic matter; CP = Crude protein; NDF = Neutral detergent fiber; ADF = Acid detergent fiber; Concentrate Mix = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%).

At the end of the feeding trial, sheep were fasted for 12 hours overnight with free access to water. Slaughter body weight (SBW) was then measured and all sheep were slaughtered to determine carcass characteristics. The blood was drained into a bucket and weighed. After removal of digestive tract and non-carcass components, hot carcass weight (HCW) was recorded including tail fat. Edible (heart, liver, kidney, blood, digestive organs, total fat) and non-edible (skin, head, feet, spleen, lungs, trachea, genitalia) non-carcass components were weighed separately and recorded. Digestive tract was weighed while full and empty. The weight of digesta was computed as the difference between full and empty digestive tract. The empty body weight (EBW) was determined as SBW less gut contents. Dressing percentage (DP) was calculated as (HCW/SBW) × 100 and (HCW/EBW) × 100. Rib eye-muscle area (REM) was measured at the 12/13^th rib position of Longissimus dorsi muscle (LDM) using transparent paper. The left and right rib-eye muscle area were traced onto a square paper and the area of the squares that fell within the traced area was counted and those partially fell outside was estimated. The average of the two sides was taken as the REM. Kidney, scrotal, pelvic, and heart fat weight measurements were taken using a balance of sensitivity 0.01 kg and summed up as a non-carcass fat.

Feed and dietary ingredients were sampled weekly during the feeding trial and composited for chemical analysis. Samples were dried at 60˚C for 72 hours, and ground using laboratory mill to pass 1 mm screen. The samples were analyzed for dry matter (DM), nitrogen (N), ash [17], acid detergent fiber (ADF), and neutral detergent fiber (NDF) [18]. The CP was calculated as N*6.25.

Meat samples were taken from areas of LDM at 10^th to 13^th rib positions, chopped and thoroughly mixed and frozen at −20˚C until partially dried. The frozen fresh samples were partially dried at 55˚C for 72 hours, packed in polyethylene bags and stored in a refrigerator at 4˚C pending proximate chemical analysis. About 2 g of partially dried samples were weighed into a pre-weighed crucible dish, and dried in an oven at 102˚C overnight to determine DM. Ashing was carried out at 600˚C for 6 hours. The rest of the sample materials were oven-dried, ground and stored in a refrigerator at 4˚C until required for analysis of protein and fat percentages. Total fat was extracted using Soxhlet apparatus for 6 hour with diethyl ether (boiling point of 34.5˚C) and the dried residue was weighed for fat content. Crude protein was determined according the method outlined by AOAC [17].

Meat samples taken from LDM for proximate analysis were used for color and pH measurements. The pH measurements were made 1 and 24 (ultimate) hours after slaughter using portable pH-meter (Meat PH meter-HI99163, HANAN instruments) having sharp penetrating electrode. The probe was cleaned with distilled water and calibrated with pH 4.1 and 7.1 standard buffer solutions between sample measurements. For meat color measurements, the cut surface of frozen samples at 4˚C was freshly exposed on flat surface of white background and allowed to bloom for about 30 - 45 minutes at ambient temperature. Then, meat color parameters i.e., CIE L*a*b* values (L* = lightness, a* = redness and b* = yellowness) were obtained using the digital colorimeter (HunterLab MiniScan EZ, Serial No. MsEZ1547, 45/0˚ illumination/viewing system, D65 light source, and 10˚ observer angle) calibrated with black and white standardized calibration plates between sample measurements [19]. Three random readings at different locations per sample were taken and averaged.

Meat samples taken from LDM muscle were evaluated for tenderness, juiciness, flavor and overall acceptability by semi trained twelve panelists following the standard procedure [20]. The samples were cut into pieces and wrapped individually in aluminum foil and roasted in an oven at 125˚C for 45 minutes. Then, the roasted samples were cut into uniform bite-size pieces and kept in food warmers until evaluated by panelists. Pieces from each sample per breed were provided to panelists randomly in such a way that everyone got sample from each breed and muscle type. The panelists independently rated the tasted samples on a nine-point hedonic scale for Tenderness from Extremely Tender = 7 to Extremely Tough = 1; Juiciness from Extremely Juicy = 7 to Extremely Dry = 1; Flavor from Extremely Desirable = 7 to Extremely Poor = 1, and General Acceptability from Extremely Liked = 7 to Extremely Dislike = 1.

The steaks were prepared according to procedures develop earlier [21]. Each batch of aged samples was thawed for 24 hours at room temperature (24˚C - 25˚C) for steaks preparation. The Warner-Bratzler shear force method was used to determine objective tenderness [20]. The steak was allowed to cool down to room temperature prior for about an hour to evaluate instrumental tenderness using WBSF. After cooling, the steak was cut across the long axis putting the knife tip on heavy connective tissue side (dorsal) and handle of knife on ventral side in order to expose the fiber direction. Six cores were removed parallel with the muscle fibers. It was critical that the muscle fibers run parallel with the core so that the shear was across the grain. The WBSF device was used to shear each core. The shear was across the middle (center) on each core. The peak values of WBSF were recorded in N (Newton) for each core. The values for the seven cores were averaged for the determination of a single value for each steak.

Data were analyzed using the general linear model GLM procedure of SAS [22]. Adjusted Tukey test was used to locate the significant differences between means when F-test declare significance at p < 0.05. The statistical model used was: Y_ijkl = μ + B_i + S_j + C_k + (B × C)_jk + e_ijkl. Where: Y_ijkl = the response variable; μ = overall mean; B_i = effect of block; S_j = effect of genotype; C_k = effect of concentrate level; (B × C)_jk = interaction between genotype and concentrate level, and e_ijkl = random error. Since interaction between genotype and supplement level for all attributes evaluated were not statistically significant (p > 0.05), only means for main effects were presented and discussed.

Rutana sheep had higher (p < 0.05) SBW, EBW, HCW and REM area than the other two breeds (Table 2). Gumuz sheep had higher EBW (p < 0.05) but similar (p > 0.05) SBW, HCW and REM area as compared to Washera sheep. Dressing percentage both in SBW and EBW bass was higher (p < 0.05) for Rutana than Washera sheep, while the values for Gumuz sheep were comparable with both sheep breeds. All carcass parameters indicated in Table 2 were significantly increased (p < 0.05) by the higher level of concentrate supplementation.

Effects of genotype and concentrate level on edible and non-edible carcass were significant for some parameters only (Table 3 and Table 4). Liver, empty gut, heart fat and total edible offal were consistently higher (p < 0.05) for Rutana compared to Washera sheep. These same parameters and total non carcass fat were greater (p < 0.05) for the high level of concentrate supplemented group. Among

^a-cMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); SBW = Slaughter body weight; EBW = Empty body weight; HCW = Hot carcass weight; DP = Dressing percentage; REM = Rib eye muscle area; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

^a,bMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); GIT=gastro intestinal tract; TNCF = Total non carcass fat; TEO = Total edible offal; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

^a-cMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); TENO=Total none-edible offal; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

the non-edible offal spleen, penis, skin and legs, and total non-edible offal were significantly affected by genotype and concentrate level (p < 0.01). Total non-edible offal was lower for Washer than the other breeds, while the high level of concentrate supplementation resulted in higher non-edible offal components for those that showed significant effect.

The moisture content of the meat was higher and that of the CP content was lower (p < 0.05) for Washera compared to the other two sheep breeds that had similar values (Table 5). The fat and ash content of the meat was unaffected (p > 0.05) by genotype. The effect of concentrate level was significant (p < 0.05) only for fat content, the vale being higher for the high level of concentrate supplementation.

Average meat color parameters were 37.11 for lightness, 11.8 for redness and 11.14 for yellowness with mean ultimate pH of 5.55. Meat from Rutana sheep had higher lightness and lower redness and yellowness (p < 0.05) than Washera sheep. Gumuz sheep had similar (p > 0.05) meat color parameters with Rutana, but Gumuz sheep differs from Washera only in redness (Table 6). The pH taken one hour after slaughter was higher but ultimate pH was lower (p < 0.05) for Rutana sheep as compared to the other breeds. Level of concentrate supplementation did not impact (p > 0.05) the color and pH of meat.

There were significant (p < 0.01) variations among the genotypes and between the concentrate levels in all sensory and shear force parameters measured (Table 7). Washera had the lowest value in all parameters, while Gumuz had only lower value than Rutana in flavor and overall acceptability (p < 0.05). Sensory values with exception of shear force were higher (p < 0.05) for high level of concentrate supplemented group.

Results of this study highlighted a significant impact of both sheep genotype and

^a,bMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); CP = Crude protein; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

^a,bMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); L* = lightness index, a* = redness index; b* = yellowness index; pH1 = pH measure taken one hour after slaughter; pH₂₄ = pH measure taken 24 hours after slaughter; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

^a-cMeans with different superscript letter within genotype and concentrate level categories in a row differ (p < 0.05); SRF = shear force; OA = Overall acceptability; Concentrate mix (CM) = Wheat bran (45%), maize grain (26%) and Noug seed cake (29%); Low = 300 g/d CM; High = 450 g/d CM; SEM = Standard error of the mean.

level of supplementation on carcass yield and dressing percentage. Similar observation has been documented in a comparative study conducted in Ethiopian sheep [11] and goat genotypes [12]. Improvement in carcass yield and dressing percentage of indigenous sheep with increasing level of supplementation as in the case of the current study is well documented [7] [23] and is a function of increased supply of nutrients for muscle development. This is supported by greater rib-eye muscle area of the high concentrate level supplemented group in this study, as rib-eye muscle area is mostly used as a tool to indicate the proportion of carcass muscling [24]. Various studies showed that supplementation had a significant and positive effect on the rib-eye muscle area [11] [25], which is in line with the result of this study.

In the current study Rutana sheep showed a better potential in carcass yield and dressing percentages, although the latter was not significantly different from Gumuz sheep. Devendra et al. [26] described dressing percentage as the proportion of carcass weight to body weight, a parameter that helps to assess meat productivity of animals. Nutrition influence dressing percentage through variation in weight of gut contents and variation in actual carcass weights [27]. The reason for variation in dressing percentage between Rutana and Washera sheep in the present study appeared to be due greater differences in HCW as compared to differences in SBW or EBW between the two breeds. Conversely, similar proportional increase in HCW to that of increase in SBW or EBW in Gumuz and Rutana sheep might have resulted to comparable dressing percentage values of Gumuz with Rutana sheep. Dressing percentage expressed as percentage of shrunk and empty body weight in the current study averaged 41.7% and 52.1%, respectively, which was comparable to the average of 43% and 52% reported for Washera sheep [28].

In Ethiopia carcass offal components are categorized into edible and non-edible based on the tradition and eating habit of the people in different parts of the country. In general, most edible and non-edible offal components was not significantly impacted by genotype and/or level of concentrate supplementation, as such tissues are of early maturing being little impacted by differences in nutritional status and growth rate of the animals. Those that were significantly affected such as the liver, GIT and heart fat appeared to be the ones that can be affected by nutritional status of the animal and consequently by differences in growth rate of genotypes [29]. The increase in liver weight with high concentrate supplementation and differences among genotypes in liver weight might be related to the storage of reserve carbohydrates such as glycogen when animals are fed with energy dense diets [30], and might be associated with difference in body mass among genotypes that proportionately increase the size of the liver [29]. According to Wester et al. [31] large amounts of digesta present in the GIT would result to net tissue growth presumably to accommodate for the increased requirement for digestion, which might have been the reason for the observed differences among genotypes and between concentrate levels in GIT weight in the current study. Differences in the weight of penis, spleen, skin and leg among genotypes and between concentrate levels appeared to be a result of a proportional increase in such components as EBW increases [32].

Meat quality can be evaluated through its content of protein, fat, water and ash [33]. The high moisture and low protein content of Washera sheep as compared to the other genotypes noted in this study could be associated with variations in tissue development as a result of differences in stage of physiological maturity [34]. Despite the expected inverse relationship of moisture and fat content, the latter however fail to differ significantly among genotypes. On the other hand sheep on the high concentrate diet recorded higher meat fat content, as dietary levels of concentrate have a positive effect on fat content of the carcass [35]. Increased concentrate intake is associated with the production of more ruminal propionate and/or higher glucose in the small intestine from non-degradable starch [36] which stimulates insulin production and consequently lipogenesis [37]. Ash content in the current study was not affected by genotype and concentrate level, which is in line with the study of Dereje et al. [12] that reported similar ash content for Bati, Hararghe highland, and Short-eared Somali goats fed different concentrate levels.

Color is one of the important quality characteristics of fresh meat [38]. It can be measured numerically using a colorimeter or subjectively [39]. The higher meat lightness (L*) of Rutana than Washera sheep noted in this study could be a result of differences in muscle myoglobin content that could be observed among breeds [40]. The average meat lightness of Washera sheep of the present study was lower than the value of 37.2 reported for Black head Ogaden sheep of Ethiopia [41] and greater than value of 32.2 reported for Arsi-Bale and Afar sheep lambs [41]. In the current study the value of meat color coordinates (L* and a*) for all breeds were higher than the acceptable threshold values for lightness of ≥34 and redness (a*) of ≥9.5 established by consumer evaluations of fresh lamb meat [42].

The ultimate meat pH values obtained from this study were within the international acceptable range of meat pH (5.5 - 5.9) for export/import market, indicating the glycogen levels in the muscle of the meat was high enough to develop optimum level of lactic acid causing a fall in pH and thereby improving the shelf life of the meat [43]. Therefore, the current study showed that animals were in good physical condition with sufficient glycogen reserve at slaughter. The outcome of the current study are greater than previous finding [11] who reported the pH values within the range of 5.7 - 5.8 for comparative study conducted in Ethiopia sheep when the pH measurement was made 24 h after slaughter. In the present study they were a significant impact on genotype than concentrate level. The breed effect on meat pH in the present study might be attributed to changes in the muscle glycogen reserve at slaughter which is inversely related to the ultimate pH [44].

Quality of sheep meat is directly related to its sensory characteristics, such as tenderness, juiciness, taste, and odor [45], and the overall acceptability is highly correlated with flavor and tenderness [46]. The current finding showed a significant impact of both sheep genotype and level of concentrate supplementation on sensory characteristics. Previous studies also highlighted genotypes effect [40] and effect of level of concentrate feeding [47] on sensory eating quality of lamb meat. The average value of flavour and juiciness of the current study were higher than the value for flavour (3.33) and juiciness (3.47) noted for sheep in Zanzibar [48]. The differences in juiciness could be related primarily to the ability of muscles to hold water during cooking [49], and that of tenderness could be due to differences in the size of muscle fiber [50]. Differences in sensory quality of meat with level of concentrate supplementation noted in this study could be due to the fact that animal that grow and mature on a high energy diet can have improved tenderness, juiciness and flavour [39]. Generally, the eating quality scores given by the sensory panelist for all breeds and supplementation level were within the acceptable range of 6 to 9 set by AMSA [21].

The average shear force value of the current study was comparable to the finding of 29.83 N/cm² [51]. In contrast Hamdu [41] reported 26.71 N/cm² for Zanzibar sheep. Generally the sheer force value of less than 49 N/cm² is considered and acknowledged as tender [49]. Likewise Hopkins et al. [52] Shear force value of above 27 N/cm² has been recommended for Mutton to satisfy the consumers’ acceptance. The lower shear force value for higher supplement level is related with higher fat content of the meat [53].