The experiment was conducted at Haramaya University goat farm (9.0˚N latitude and 42.0˚E longitude). A total of 24 Blackhead Ogaden lambs with age of 8 - 10 months, as estimated by dentition and information obtained from the owners, were purchased from Jijiga market. Experimental animals were quarantined for 3 weeks in isolated holding yards at the university and vaccinated against common infectious diseases in the area (pasteurellosis and anthrax), de-wormed against internal parasites by albendazol de-wormer, and sprayed with acaricides (Vetacidin 20%) against external parasites. Following quarantine, the experimental animals were placed in experimental house partitioned into individual pens (70 × 120 cm) that is equipped with feeding trough and watering bucket. Animals were acclimated to the experimental procedures and feeds for 14 days before the commencement of the trials. The experiment consisted growth and finishing phases each with three months of stall feeding Afterthe completion of 180 days stall feeding, all lambs were slaughtered for carcass linear measurements and compositions studies.

The lambs were blocked into six groups based on initial body weight. Lambs in a block were randomly as- signed to one of the four dietary groups making a total of six lambs in each group. Thus, a randomized com- plete block design (RCBD) was employed. The initial body weight of the experimental animals was determined by two consecutive weighing after overnight withdrawal of feed and the average of the two days were taken. After the completion of the growth experiment, the animals were used for finishing phase with no change of groups.

During the growth phase, the treatments were based on the amount of dried Acacia albida leaf and pod and Cactus pear cladodes mixture supplement at the ratio of 2:1, respectively. The level of supplement feeding was at the rate of 0ACM (the control groups did not get any supplement, only the basal diet), 0.9ACM (Supplementation with Acacia albida and Cactus pear cladodes mixture at a rate of 0.9% body weight), 1.2ACM and 1.5ACM. Hay was offered ad libitum at 20% refusal rate. Treatments during the finishing phase were different proportions of roughage (R) to concentrate (C) ratio, which included 60R:40C, 50R:50C, 40R:60C, and 30R: 70C. The proportion of roughage to concentrate was determined from animals’ ad libitum intake on dry matter basis, which was adjusted every four days. In both phases feed was offered in two equal meals.

After the completion of stall feeding, all experimental animals were fasted overnight with access to water and then slaughtered. The dressed carcass of each lamb was chilled at 4˚C in a refrigerator for 24 hrs. Cold carcass was suspended with hind legs on hock gambre of constant width (20 cm) and the following carcass linear measurements were recorded using tape in centimeter. Carcass length (from the basis of the tail to the basis of the neck), hind leg length (the smallest distance of the tarsal-metatarsal articular surface), anterior and posterior buttock circumference of thorax (measured horizontally around the buttock at the level of the maximum width of the trochantor and the caudal insertion, respectively), chest depth (from the sternum to the thoracic vertebra), and buttock and shoulder width (at the level of the scapula from one lateral surface to the other). Thereafter, the cold carcass was split in to two equal halves along the dorsal mid-line using a hand saw. The left half of each carcass was dissected into lean, fat, bone and used to estimate the whole composition.

Samples of feed offered, refusals, and feces were ground to pass a 1 mm sieve mesh kept in air tight plastic bags pending analysis. DM, OM, and Nitrogen (N) were determined according to the procedures of previous work [10] . The crude protein content was estimated as N × 6.25. Organic matter (OM) was calculated as the difference between 100 and ash content. Neutral detergent fiber (NDF), Acid detergent fiber (ADF) and Acid detergent lignin (ADL) were analyzed according to the procedures used in earlier studies [11] .

The data subjected to Analysis of Variance (ANOVA) using the general linear model procedure of [12] . Treatment means were separated using Tukey HSD (Tukey honestly significant difference) test. The model for data analysis was Y_ij = m + T_i + B_i + E_ij. Where, Y_ij is the response variable; m is the overall mean; T_i = the treatment effect; B_i is the block effect and E_ij is the random error.

The CP content of the natural pasture hay is at the borderline of minimum CP needed for maintenance requirements of ruminant animals while NDF and ADF were high (Table 1). Acacia albida has high CP content while cactus cladode is characterized by its low CP content. The hay contains the minimum (7%) CP needed to support acceptable ruminal microbial activity and the maintenance requirement of the host ruminant [13] . The CP content of A. albida was within the range of 12% - 18% reported by previous workers [14] and [15] . Cactus pear contains low CP than hay and the value obtained in the present study is comparable to that reported in earlier studies [16] [17] . Fiber fractions content was higher for the natural pasture hay than A. albida and Cactus cladodes. The fiber fractions content of A. albida is comparable to values of 560 g/kg NDF and 336 g/kg ADF reported earlier [18] . The CP content of hay used in the finishing phase is similar to that of growth phase, since the source is the same (Table 1). The CP composition of haricot bean haulm was relatively lower than hay. Low CP concentration in haricot bean haulm is attributed to late harvesting of the crop after much of the leaves have been shattered. As expected, the CP content of the concentrate mixture is well above required to satisfy the nitrogen demand for proper functioning of rumen microorganisms and it is a good complement to the low quality roughage used.

Treatment groups did not differ in carcass linear measurements, except for anterior buttock circumference (ABC) and chest depth (CD) (Table 2). Lambs fed with 1.5ACM/30R:70C have wider ABC than the rest treatment groups, while the CD is longer in 1.5ACM/30R:70C than 0ACM/60R:40C. According to earlier work [19] , the linear carcass measurements are indices of skeletal development and indirectly help to determine carcass conformation. Effect of nutrition on carcass linear measures were reported earlier [1] [19] and generally confirmed that increased level of supplement improved most of the carcass linear measurements. In the present study, except ABC and chest depth values, most of the carcass linear measurements did not affected due to the different feeding regimes. Increasing level of supplement and proportion of concentrate during the growth and finishing

DM = Dry matter; OM = Organic matter; CP = Crude protein; NDF = Neutral detergent fiber; ADF = Acid detergent fiber; ADL = Acid detergent lignin; CSM = cotton seed meal; NSM = noug seed meal.

phases improved the circumference of anterior buttock and depth of the chest. This is in line with [20] who noted that almost all linear carcass measurements did not varied with the diet except buttock width which was higher for group of kids fed higher level of supplement diet. Contrary to this, [20] reported that lambs fed at high level of concentrate showed greater values for most of the linear measurement parameters. The average numerical value recorded for carcass length was comparable to that reported by [20] , but was greater than the value reported by [1] [21] , however, was lower when compared to 78 and 75 cm obtained for South African Merino and Durmer sheep [22] . The discrepancy in length might be due to differences in breeds of experimental animals and types feed used during the experimental periods. Chest depth was also increased with increased level of foliage supplements and concentrate proportion during the growth and finishing phases, respectively. Differences for these carcass linear measurements among the treatment groups showed a positive relationship with slaughter weight, since heavier lambs at slaughter attained greater carcass linear measurements [23] .

Percentage of lean, bone, and meat to bone ratio was not impacted (p > 0.05) by the different feeding regimes (Table 3). Percentage of fat content of the carcass was significantly (p < 0.05) higher in 1.5ACM/30R:70C groups than 0ACM/60R:40C. The major non genetic factor influencing carcass composition is nutrition. Accordingly, higher fat percentage of carcass of lambs fed higher level of concentrate is associated with higher muscular fat proportion due to intake and availability of essential nutrients. In concurrence with the present findings, other researchers also reported higher proportion of fat in lambs fed with higher concentrate ration than consumed lower concentrate containing diets [24] [25] . The result of the present findings, however, is in contrast

Means in a row with different superscript are significant at p < 0.05; PBC = Posterior buttock circumference; ABC = Anterior buttock circumference; ^*Growth phase treatments are based on Acacia albida and Cactus pear cladodes mixture (ACM) supplementation at a ratio of 2 to 1, respectively; 0ACM = no supplement, control; 0.9ACM = supplementation at a rate of 0.9% body weight; 1.2ACM = 1.2% body weight; 1.5ACM = 1.5% body weight, and finishing phase treatments are based on roughage (R) to concentrate (C) ratio: 60R:40C (60% roughage and 40% concentrate feeding); 50R:50C; 40R:60C; 30R:70C.

Means in a row with different superscript are significant at p < 0.05; ^*Growth phase treatments are based on Acacia albida and Cactus pear cladodes mixture (ACM) supplementation at a ratio of 2 to 1, respectively; 0ACM = no supplement, control; 0.9ACM = supplementation at a rate of 0.9% body weight; 1.2ACM = 1.2% body weight; 1.5ACM = 1.5% body weight, and finishing phase treatments are based on roughage (R) to concentrate (C) ratio: 60R:40C (60% roughage and 40% concentrate feeding); 50R:50C; 40R:60C; 30R:70C.

with earlier studies [26] [27] who did not found significant differences between dissected lean, bone, and inter muscular fat weight or percentage of growing ram lambs fed with different proportion of roughages and concentrate. This could be attributed to the relatively good diets used in all treatment groups during the studies. The lack of significant difference for the value of bone in this experiment is that it is a tissue which develops early in all animal species and does not depend on nutrition and sex at older ages [28] . Feeding regimes also did not impact value of meat to bone ratio. In line with this, several authors have been reported that levels of supplement and feeding regimes did not affect meat to bone ratio for different indigenous breeds of sheep and their crosses [20] [21] [29] [30] . Differences amongst these studies might be due to a consequence of environmental effects, probably related to diet compositions, the degree of maturity attained at the slaughter body weight and breed differences.