The study was conducted at Shire Endaslassie, North-West Zone of Tigray, which is located 279 Km distance from Mekelle, the capital city of the Tigray region. Shire Endaslassie is located at 14⁰ 06` to 14<sup>0</sup> 07`N latitudes and 38⁰ 16' to 38⁰ 17'E longitudes and an altitude of 1953 meters above sea level with average annual temperature and rainfall of 25⁰C and 850 mm, respectively.

Twenty-four yearling intact male sheep with an average initial body weight of 18 ± 1.55 kg (mean ± SD) were used to conduct the study. Quarantine of sheep in the experimental area was done for 21 days. Moreover, sheep were sprayed and de-wormed against external and internal parasites, respectively; and vaccinated against the diseases prevailing in the area.

Hay from Natural pasture was used as a basal diet. Concentrate mixture (75% wheat bran and 25% sesame seed cake) was also purchased from Shire Endaslassie.. Atella was collected from traditionally “Tella” brewing individuals in Shire Endaslassie, which was prepared from finger millet and the wet Atella was dried for 3-4 consecutive days by the sun. The leaves of F. albida and S. sesban were collected by hand stripping and air-dried for 2-3 days.

The experiment was conducted using Randomized Complete Block Design (RCBD) and based on their Initial Body Weight (IBW), sheep were blocked into six blocks of four sheep each and sheep within a block were assigned randomly to one of the four treatment diets.

Based on the research findings of Zemichael and Solomon [12] for Arado sheep, the amount of concentrate mixture in this study was determined to be 300g on a Dry Matter (DM) basis. The supplement for the other three treatments was on an isonitrogenous basis. Samples of treatment feeds were analyzed for DM and Crude Protein (CP) content before the start of the experiment to determine the amount of the experimental rations, as given in Table 1.

Thus, the 300 g concentrate mixture supplied 73.6 g/day CP on a DM basis. For sheep in the other three treatments, 330g DM Atella, 360g DM F. albida leaf and 280g DM S. sesban leaf were offered daily to supply the same amount of CP on the isonitrogenous basis (i.e., 73.6 g/day). The layout of the experimental treatments was:

T₁ = Hay ad libitum + 300 g DM concentrate mixture

T₂ = Hay ad libitum + 330 g DM sun-dried Atella

T₃ = Hay ad libitum + 360 g DM Faidherbia albida leaf

T₄ = Hay ad libitum + 280 g DM Sesbania sesban leaf

The partial budget analysis was performed to evaluate the economic profitability of the treatment feeds using the procedures of Upton [16]. The analysis involved the calculation of the variable costs of sheep, feeds and benefits gained from the result. Purchasing and selling prices of sheep, purchasing price of the hay, concentrate mixture, Atella and the labor cost for F. albida and S. sesban leaves collection were recorded. Other expenses such as the cost of transport (sheep and feed), mineral licks, housing, labor (feeder and cleaner) and veterinary services, which was common for all treatments, were not considered in the calculation of partial budget analysis. The selling price of each sheep was determined by inviting well experienced three sheep dealers, who know the prevailing market price of sheep in the Shire Endaslassie market. Therefore, the average price determined by the dealers was used as the selling price of the sheep. The cost of feeds was computed by multiplying the actual DM intake of feed for the whole feeding period (90 days) with the current purchase price of each treatment feeds. The total cost associated with each sheep during the experimental period in each treatment was added and the average was taken as a Total Variable Cost (TVC). The Marginal Rate of Return (MRR) measures the increase in Net Return (∆NR) associated with each additional unit of expenditure (∆TVC). The change in a Total Variable Cost (∆TVC), Total Return (TR), change in Total Return (∆TR), Net Return (NR), change in Net Return (∆NR) and Marginal Rate of Return (MRR) were calculated as follows:

Where; SP = Selling price of sheep; PP = purchasing price of sheep

Data obtained from intake, digestibility and body weight change were subjected to analysis of variance using the General Linear Model procedure of SAS version 9.2 [17]. Tukey`s Studentized range test was used to separate treatment means. The model used for data analysis was:

Where; Y_ij = response variable, µ = overall mean, T_{i =} treatment effect, B_{j =} block effect, e_ij= random error

The chemical composition of treatment feeds is given in Table 2. Hay had lower CP and higher NDF and ADF content than the supplemental feeds. Although F. albida was slightly lower and that of S. sesban was a bit higher in CP content, the four supplemental diets are generally rich sources of CP. However, the NDF and ADF levels in S. sesban were lower than the other supplemental diets, and the ADL content was lower for the concentrate mixture.

The CP content of the refusals of the basal diet observed in the present study was lower by 43.8, 37.4, 38.4 and 33.7% in T₁, T₂, T₃ and T₄, respectively compared to the CP content of hay offered. Conversely, hay refusals in all treatments had relatively higher NDF and ADF content than the offered. The lower CP and higher cell wall fiber content in hay refusals as compared to the offered might be associated with selective feeding of sheep for more palatable and nutritious parts of hay than the unpalatable and lignified parts of the feed.

The CP content of the hay used in this study was comparable with [18]; but lower [19, 20] and higher [21, 22] CP values as compared to the value noted in this study were also reported. The differences in the nutritive value of hay, including CP content, could be attributed to differences in species composition of the harvested hay, stage of maturity during harvesting period and the growing environment [23].

The CP content of concentrate mixture used in this study appeared to be comparable to 26.14% CP content noted by Zemichael and Solomon [12]. Similarly, the NDF, ADF and ADL values of the concentrate mixture used in this study were comparable to 36.84% NDF, 12.12% ADF and 2.67% ADL reported by the same author.

The CP content of Atella used in this study is comparable to 21%, 21.2% and 21.8% [7, 11, 24], respectively; and is higher than 10.2% [25]. Similarly, NDF and ADF contents of Atella are relatively in agreement with 60.2% and 22.5%, and 54% and 29% noted by Wondatir et al. [24] and Solomon [7], respectively; but higher than 32.7% NDF and 16.4% ADF reported by Guesh and Mengistu [25]. The ADL content of Atella is also comparable with 11% [7, 24]; but slightly higher than 5.9% and 7.3% [11, 25], respectively. Differences among results in Atella nutrient composition might be associated with differences in the ingredients used for Tela preparation and the methods of preparation followed by communities in different areas.

The ash content of F. albida was within the range of 5.7-7.2% reported by Solorio and Solorio [26], and Shayo and Uděn [27], but lower than 9.49% [9]. The CP content of F. albida in this study is in agreement with the findings of Takele and Getachew [8] and Gebreselassie et al. [9] who reported 19.5% and 20.8% CP, respectively; but lower than 25.3% CP [28]. The NDF, ADF and ADL content of F. albida in the current study are similar to the results of Gebreselassie et al. [9], who reported 51.3% NDF, 35.6% ADF and 15.6% ADL content. The differences in the nutritional composition of F. albida among studies may be associated with different factors including the age of the tree, maturity stage of F. albida leaves and twigs, the plant parts used for chemical analysis and feeding, the season of sample collection, soil fertility and other climatic factors. In line with this, Patricia [28] demonstrated considerable variations in NDF and ADF content of F. albida leaves in different seasons.

The CP content of S. sesban leaf in this study is within the range of 23.8-31.7% indicated by Mekoya [3]. The value was also comparable with 24.8% noted by Manaye et al. [29]; but slightly higher than the 23.9% and 22.1-23.1% reported by Sabra et al. [30] and Solomon et al. [31], respectively. On the other side, higher values of S. sesban leaf CP than the one in the present study were recorded by Debela et al. [32] who reported 29.7% CP. The NDF and ADF content of S. sesban leaf in this study is lower than 39.9% and 29.9% [32]. On the other hand, ADL content of S. sesban noted in this study was in agreement with 4% and 5.1-5.7% indicated by Mekoya [3] and Debela et al. [32], respectively; but lower than the 27.2-28.2% noted by Solomon et al. [1]. All the above differences in nutritional composition of S.sesban may be attributed to differences in accession, stage of plant growth, cutting frequency and harvesting regimen, soil type and fertility status [3], and parts of the plant (leafs, twigs, whole forage and green pods, etc.) included during feeding and chemical analysis [32].

The daily DM and nutrient intake of local sheep are given in Table 3. The DM intake of supplements were 300, 330, 360 and 280 g DM/day for concentrate mixture (T₁), Atella (T₂), F. albida (T₃) and S. sesban (T₄), respectively (each calculated to give 73.6 g CP on isonitrogenous basis) and the intake was 100% of the supplement offer for all treatments. This indicates the variation in intake among the supplements was because of differences in CP content. The hay DM intake was lower (P < 0.001) for T₃ as compared to the other three treatments that had similar values among each other. The lower hay DM intake for T₃ could be associated with the relatively greater amount of supplement DM fed to this group that might have resulted in a substitution effect on hay intake.

Total DM and OM intakes appeared to be highly impacted by the supplemental DM intake in this study. As such total DM and OM intakes were the lowest for T₄, intermediate for T₁ and highest for the other two treatments (P < 0.001). Total NDF intake was in the order of T₃> T₂> T₁> T₄ (P < 0.001) and that of ADF intake was in the order of T₃> T₂> T₄ > T₁ (P < 0.001), both of which appeared to be associated with the level of NDF and ADF in the supplemental diets.

Generally, the total DM intake of sheep in this study was about 700 g/day, which was comparable with the results of many studies [19, 33-35] with different Ethiopian sheep breeds. Conversely, lower total DM intake [9, 25] and higher total DM intake [21, 36] by sheep as compared to the one noted in this study has been reported. Of course, variations in the type and amount of the basal diet as well as the supplement, breed of sheep, growth stage of the animal and other similar factors may contribute to differences in DM intake observed among the studies.

Although T₃ had slightly lower CP intake as compared to the other three treatments, the numerical difference is not as such too big. This slight variation in CP intake is associated with differences in the basal diet DM intake. The estimated Metabolizable Energy (ME) intake was similar among the treatments in this study. The ME requirement for a 20 kg lamb gaining 50-150 g/day is 3.7-6.4 MJ/day for diets with a metabolizability of 0.65 [15]; and according to ARC [37] the maintenance and growth (50-200 g gain) ME requirement for the same weight lamb is 4.5-7.9 MJ/day. Thus, based on these assumptions the estimated ME of the treatment diets (7.5-8.7 MJ/day) in the present study satisfies the energy requirement for maintenance and growth (38-50 g/day gain) of the sheep.

The mean apparent digestibility of DM and nutrients in local sheep are given in Table 4. Apparent digestibility of DM was similar (P > 0.05) among treatments. The apparent digestibility of CP was lowest for T₃ and values for T₁ were greater than those in T₄ but similar to T₂.

Digestibility of OM was lower for T₃ as compared to T₁ and T₂ (P < 0.01), but the value for T₄ was similar to all other treatments. Digestibility of NDF and ADF was lower (P < 0.001) for T₃ as compared to the other treatments. The lower apparent digestibility of CP, OM, NDF and ADF in F. albida supplemented sheep (T₃) may be associated with the higher fiber (ADF, NDF and ADL) content and the possible presence of anti-nutritional factors like tannins in the feed, which might have affected the availability of these nutrients for the sheep. The lack of significant differences in the apparent digestibility of NDF and ADF in most of the treatments (T₁, T₂ and T₄), except in T_3, are in agreement with the findings of Kaitho and Kariuki [38], who concluded that supplementation had little or no effect on fiber digestibility. Moreover, many researchers also found no significant effect of supplementation on the apparent digestibility of NDF and ADF [36, 39].

Improved digestibility values of DM, OM and CP with concentrate supplementation is in agreement with the findings of Kaitho and Kariuki [38]. The OM and CP apparent digestibilities of the group supplemented with Atella (T₂) are within the range of 73.6-79% OM and 69.6-76% CP digestibility reported by Mulu et al. [40] for Wogera sheep fed hay basal diet and supplemented with different proportions (100-300 g/day) of brewery dried grain. Apparent digestibilities of DM, OM, CP and ADF of the group supplemented with Atella in this study are somewhat comparable to the digestibilities of these nutrients reported [11], for lambs fed finger millet straw and supplemented with 100% Atella, and mixtures of 70% Atella and 30% Noug seed cake. Conversely, higher digestibility of CP (75.5-82.2%) than the result in T₂ was also reported by Guesh and Mengistu [25] for Black Head Ogaden sheep supplemented with urea-Atella block (30-50% ratio of Atella in the block).

The Initial (IBW) and Final Body Weight (FBW), Bodyweight Change (BWC), Average Daily Body Weight Gain (ADG) and Feed Conversion Efficiency (FCE) of local sheep are presented in Table 5. In the present study, Initial Body Weight (IBW) was similar (P > 0.05) among treatments. However, treatments differ in FBW, BWC, ADG and FCE (P < 0.05). Sheep that received a commercial concentrate mixture (T₁) had higher FBW and FCE (P < 0.01), BWC and ADG (P < 0.05) than those supplemented with F. albida leafs (T₃), but differences in BWC, ADG and FCE among T₁, T₂ and T₄ were not significant (P > 0.05).

The performance of ruminants is influenced by the level of nutrients in the feed [41]. The higher performances (FBW, BWC and ADG) of sheep in T₁ as compared to sheep in T₃ might be due to the higher energy density, lower cell wall fiber contents, higher total and digestible CP intake and higher nutrient digestibility in the concentrate mixture than the F. albida containing diet. Moreover, the lower growth rates in this study in sheep that received F. albida leaves than sheep supplemented with concentrate mixture may be in part attributed to anti-nutritional factors like tannins found in the leaf of F. albida. This is supported by the research findings of Ibrahim and Tibin [42] who reported that as the level of F. albida pod supplementation increased from15% (157.5 g/day) to 30% (306 g/day) and 45% (454.5 g/day), the ADG was decreased from 55 to 44 and 39 g/day, respectively, which could be due to the decreased efficiency of feed utilization associated with an increase in the intake of antinutritional factors. However, the results of BWC, ADG and FCE in T₁, T₂ and T₄ were not significantly different (P > 0.05) from each other. There were also similarities among T₂, T₃ and T₄ in BWC, ADG and FCE, which reflected that the supplements were comparable in their potential to supply nutrients to improve the weight gains of sheep. Several studies on small ruminants in Ethiopia [8, 9, 11, 12, 19, 25, 33, 40] showed a significant change in body weight gain when sheep on poor quality roughages are supplemented with diets rich in energy or protein or both.

The ADG of sheep in T₂ was within the range of 31-55.5 g/day gain indicated by Guesh and Mengistu [25] for Black Head Ogaden sheep fed hay and supplemented with urea-Atella block (30-50% Atella ratio). Similarly, Wogera sheep fed grass hay and supplemented with 100 g brewers dried grain gain 44.4 g/day [21], which is also very close to the present study (45 g/day) in T₂. Moreover, a comparable result (51 g/day) to the current study was also recorded by Almaz et al. [11] for sheep fed finger millet straw and supplemented with sole Atella. In agreement with the current study in T₂, Emebet [43] also reported 41.8 and 44.6 g/day gain for Black Head Ogaden sheep fed haricot bean haulms and supplemented with wheat bran and brewers dried grain mixture (2:1) and sole brewers dried grain, respectively. However, higher gains of 56.4 g/day; and 56.7, 63 and 60 g/day than the current result in T₂ were recorded by Emebet [43] for sheep supplemented with wheat bran and brewers dried grain mixture (1:2); and Almaz et al. [11] for sheep groups supplemented with mixtures of 70% Atella and 30% NSC, 70% NSC and 30% Atella and sole NSC, respectively.

The mean daily gain of sheep in T₃ (38 g/day) is also within the range of 28-42 g/day stated by Takele and Getachew [8] for Horro lambs fed vetch haulm and supplemented with wheat bran and A. albida leaf mixtures, and A. albida leaf alone. Similarly, a gain of 36 g/day reported by Gebreselassie et al. [9] for sheep fed barley straw and supplemented with F. albida leaf alone is very close to the current results in T₃.

The mean daily gain (42 g/day) of sheep in T₄ in the present study is slightly higher than the findings of Solomon et al. [31], who reported 33.4-35.7 g/day gain for Menz sheep fed teff straw and supplemented with sole S. sesban; but it is comparable to 38.4 and 40.7 g/day gain for sheep groups supplemented with mixtures of S. sesban and A. angustissima, and mixtures of S. sesban and Leucaena pallida, respectively in the same experiment. Contrary to the ADG in T₄, Manaye et al. [29] reported higher ADG (83.3-99.8 g/day) for local sheep fed mixtures of 70-90% Napier grass and 10-30% S. sesban.

The partial budget analysis for the feeding trial is presented in Table 6. Sheep received Atella (T₂) supplement returned higher net income followed by sheep fed S. sesban (T₄) and F. albida (T₃); and the lowest net return was scored for sheep supplemented with concentrate mixture (T₁)_. The differences in total return, net return and marginal rate of return (MRR) among the treatments are mainly associated with the differences in the selling price of the sheep and differences in intake and cost of the supplement feed among the treatments.

The MRR Table 5 implies that each additional unit of one US$ per sheep cost increment resulted in one US$ and an additional 0.937, 0.561 and 0.504 US$ benefit for T₂, T₃ and T₄, respectively as compared to the sheep in T₁. Therefore, from this study, it can be said that, when available, supplementation of sheep with sun-dried Atella (T₂) is recommended as economically profitable considering the net return and MRR. However, for areas where availability of Atella is scarce, supplementation of either S. sesban or F. albida or both, especially for smallholder farmers, who are involved in integrated agroforestry farming systems, can be recommended because of their multipurpose uses.