Growth Performance of Arsi, Borana, Harar and HF-Crossbred Bulls Finished under Similar Feeding Condition

The study was conducted to evaluate the fattening performance of Arsi, Borana, Harar and HF-Crossbred bulls nished under similar feeding condition at the beef farm in Haramaya University. A total of 24 bulls with age categories of 2-3 and 4-5 years were used in the complete block design for the experiment. Total mixed ration was provided at 3% of their BW during experimental periods. The average daily weight gain of the four breeds range from 0.49 to 0.71 kg. Feed conversion eciency also ranges from 0.11 – 0.15. Simple linear regression models were used to explore the relationship between live body weight change and change in BCS as well as seven linear body measurements for all age groups. An average change for a unite of body condition score was equivalent to 20.3, 20.61, 22.42 and 27.78kg for Borana, Arsi, Harar and HF-crossbred bulls respectively. BCS was signicantly (P<0.01) inuenced by breeds. There was a signicant (P<0.05) breed by age interaction effect on the initial body condition score of the four breeds. There was a signicant and positive strong association between change in BW and BCS. There was a signicant and strong correlation between change in BW and change in Total topline, neck length, heart girth, ank circumference and rump length having correlation coecient ranges (r = 0.57 to 0.97). Higher (P<0.01) net prot of 7,380.47 ETB per head was recorded by Borana bulls followed by Harar bulls, Arsi and HF Crossbred with net prot of 5,406.86, 5193.29 and 3,384.98 ETB per head respectively.


Introduction
The total cattle population for Ethiopia is estimated to be about 59.45 million. The majority (98.2%) of the total cattle in the country are local breeds. The remaining are hybrid and exotic breeds that accounted for about 1.62 percent and 0.18 percent, respectively. Regarding age groups, the majority of the cattle population (that is about 62.95%) was in 3 to 10 years and 16.75% is 1 to 3 years. The remaining 2.25 % was 10 and above years old (CSA, 2017).
In Ethiopia, cattle production plays an important role in the economies of the farmers and the country at large. FAO ( 2007) reported that cattle contribute 40% of the annual agricultural output and 15% of the total gross domestic product.
There are 33 recognized indigenous cattle breeds in Ethiopia (DAGRIS 2011). Because of its multi-purpose role, farmers has been used for milk, beef, draft power, farm yard manure and source of cash income in the country (FAO, 2018). So therefore, it has been evident that those diversi ed cattle breeds in all diversi ed agro-ecology of the country has good market and brings superior prices both at domestic and foreign markets. Such a scenario created an opportunity for small-scale cattle fattening systems in different parts of the country. Therefore, Cattle fattening has gained reputation as an important business project of the livestock industry in Ethiopia (Habtamu et al., 2008). This is special in eastern part of the country in general and in Hararghe highlands in particular (Dinku 2019).
Crossbreeding work in Ethiopia was initiated in the early 1950s. Following this initiation a number of governmental and non-governmental institutions have worked on the development of the dairy sector (Aynalem et al. 2011). This resulted improvement in milk production as well as surplus crossed bulls' availability in different parts of the country. However, lack of disposal mechanism for surplus male calves at dairy farm level has been reported as one of the main constraints in improving the dairy farm pro tability (Merera and Galmessa, 2013). The importance of dairy beef to minimize the problem of continuous supply of young bulls to the market was reported in different studies (Mummed 2015;Mummed and Webb 2014;Mummed and Webb 2015).
Animal feed both in terms of quantity and quality is a major bottleneck for livestock production in Ethiopia. Currently, with the rapid increase of human population and increasing demand for food, grazing lands are progressively shrinking by being converted to arable lands, and are restricted to areas that have little value (Mengistu et al., 2017). It is clear that feed resource utilization (Fikru 2015); fattening and marketing of cattle are undeveloped in different parts of the country (Ayalew et al., 2013). Furthermore, developing economic feeding system that enhance the existing traditional productivity of beef cattle (Negassa et al., 2011) and thereby achievement to growing demand for meat and also to the country's plan of increasing meat export, by encouraging small scale beef fatteners based on scienti c evidences. Most cattle supplied to market from pastoral, agro-pastoral and mixed crop livestock production were reported poor in their meat quality (Mummed and Webb 2014;Mummed and Webb 2015). Giving due emphasis on economically feasible concentrate supplementation is important in the fattening systems.
Environmental condition during fattening period, type and amount of feed, pre-fattening condition of cattle and feedlot management determine the length of fattening period. Therefore, scarcity of feed, animal in poor condition before fattening and improper management prolong the nishing period. Longer fattening periods reduces pro t realized from cattle nishing (Ebrahim et al., 2004).
Arsi, Harar and Borana bulls are also among the 33 recognized indigenous cattle breeds in the country. Their fattening performance has be studied and different results were reported by different scholars for example (Mohammed et al., 2008;Haile et al., 2009;Girma et al., 2015;Bedhane and Dadi, 2016) for Borana bulls, (O'Donovan et al., 1978;O'Donovan et al., 1980;Merera and Galmessa 2013) for HF-Crossbred with Borana, Barca and Horro breed respectively, (Tolla et al., 2002(Tolla et al., , 2003 for Arsi bulls, (Teklebrhan, 2019) for Harar bulls. As result of these works differ in time and space under different feeding condition; the performance of these breeds under similar fattening conditions were not studied before. Therefore, this study was conducted to evaluate the growth performance of Arsi, Borana, Harar and HF-Crossbred bulls nished under similar feeding condition.

Description of the Study Area
The study was conducted at Haramaya University beef fattening unit. It is located at 9.0°N latitude and 42.0°E longitude at an altitude of 1980 m above sea level and 515km east of Addis Ababa, Ethiopia. The area receives annual average rainfall of 790 mm and annual mean temperature of 16°C (Mishra et al., 2004).

Animal management and experimental design
A total of 24 intact bulls from Arsi, Borana, Harar breeds and Holstein Frisian cross were grouped into two age groups (2-3 years and 4-5 years) were used in this experiment at the Beef Farm of Haramaya University using completely randomized design as shown in Table 1.
The bulls were purchased from Kofele local market (Arsi breed), Dida Tuyura ranch (Boran breed) and Chefe Bante local markets (Harar breed) and Sululta local market (HF-Crossbred) and transported to Haramaya University using appropriate truck. The age of the animals was estimated using the dentition method as suggested by (Hammond et al., 1971;MLA, 2011).
They were quarantined for two weeks being vaccinated against blackleg disease and also dewormed by injecting with Ivermectin against internal and external parasites. After quarantine, each animal was kept in an individual pen and acclimatized for two weeks. The experiment lasted for 90 days

Experimental feed and chemical analysis
The experimental diet consists of roughage; natural grass hay, wheat straw and concentrate; maize grain, noug (gucia abysica) cake, wheat bran, limestone, salt, ruminant premix, and were analyzed for chemical composition at Haramaya university animal nutrition laboratory.
For chemical analysis 100 grams of samples of feeds were dried at 65 °C for 48 hr. Then dried samples were then ground (1 mm screen) and stored for subsequent analyses of dry matter (DM), crude protein (CP), ash, neutral detergent re (NDF) and acid detergent re (ADF). DM, N and total ash were determined according to the o cial methods of (AOAC, 1990) and NDF and ADF according to (Soest etal., 1991). Dry matter content of the feed was determined by drying the samples in an oven at 105 °C overnight while ash content was determined by burning the samples at 550 °C for 5 h in a muff furnace. Nitrogen (N) were determined by Kjeldahl method (CP = N × 6.25). The chemical composition of each dietary components is indicated in Table 2.

Feed formulation and feed intake measurement
The diet were formulated aiming to meet the maintenance requirements and to provide a weight gain of up to 1 kg per day as suggested by Hutcheson (2006) for beef fattening center in Ethiopia. The roughage (60% of total ration) component were natural grass hay (55% of roughage) and wheat straw (45% of roughage); the concentrate (40% of total ration) consisted of 34.78% wheat bran, 27.8% Noug cake, 33.14% maize grain, 1.7 % limestone and 1.7 % salt and 0.88% ruminant premix. It was formulated to contain CP and energy to meet the optimum recommendation of feedlot.
Three percent of their body weight per day for total mixed ration was given in two equal meals at 8:00 AM in the morning and 3:00 PM in afternoon of the day and the amount were adjusted based on body weight once per every week. Clean water was available all the time. The weight of concentrate and roughage offered and refused were recorded daily to derive feed intake.

Feed conversion e ciency
Feed conversion e ciency (FCE) were calculated for each animal as proportion of weight gain to DM intake 2.6. Evaluation of daily weight gain and linear measurements Animals were weighed once every week before morning feeding. Initial weight (IW) and nal weight (FW) were recorded for different feeding periods.
Where; ADG= Average daily gain, n= number of days, n= 90 Body linear measurements such as total top line (the total length of the animal taken from front of the pool to back of the rump), neck length (the distance from front of the pool to the middle dip in vertebrate between the shoulder blades), hip height (height of the cattle on the vertical line passing through the hips), heart girth (the thoracic circumference), ank circumference (the total distance around the animal taken at the hips) , rump length (the distance taken from the hips to the pin bones), rump or pin bone width (the horizontal distance between the pin bones), shoulder width (the horizontal distance between shoulders) and chest depth (the distance taken with vertical caliper through the vertical transverse plane passing just to the rear of the point of the elbow) were taken immediately after acclimatization period and two days before slaughter.

Body condition scoring
Body condition scoring for each experimental animals (n=24) were recorded once in every week following body weight measurement during experimental feeding based on a 9 scale scoring method, with 1 being severely emaciated and 9 extremely obese, using the method proposed by Herd and Sprott, (1986)and modi ed by Jaymelynn et al. (2016) as brie y indicated in (Table 3) bellow. To avoid subjective biasness, an expert has scored the condition throughout the experimental periods.

Economic analysis
At the end of the experiment, partial budget analysis was performed to evaluate the economic advantage of feeding the four breeds of intact bulls with total mixed ration. It was done by considering the variable cost of bull price, concentrates and roughage prices, labor, medication and estimated bull sell prices at the end of the experiment. At the end of experimental feeding, four experienced animal dealers were estimate the sell price of each experimental animals and the average price determined by the dealers was used as the selling price of individual bull. The difference in sell and purchase price was considered as total return (TR) in the analysis.

TR = sell price -purchase price
The cost of feed was computed by multiplying the actual DM intake of feed for whole feeding period 90 days with the purchase price of the feed. The total cost associated with each animal during the experimental period were added and the total variable cost (TVC) were calculated for each animal. Net return (NR), was calculated by subtracting TVC from TR as follows.

Statistical analysis
The data were analyzed using General Linear Model (GLM) procedure of Statistical Analysis System (SAS, 2018) version 9.4. Simple linear regression was used for the determination of the most suitable model in the prediction of the live weight change using various body measurements change as continuous variables and simple regression was also used for change in body weight and body condition score to establish regression equations. The model included the main effects of age and breed to determine the in uence of the main effects on all dependent variables considered.
The difference among treatment means was tested using Tukeys' test at 0.05 level of signi cance.
The models used for the analysis were:

Results And Discussion
3.1. Bodyweight change, Average daily weight gain, body condition score and feed conversion e ciency Average initial and nal live weights and daily weight gain of bulls from the four breeds under two age groups are summarized in Table 4. The signi cant difference in nal and initial body weight of bulls from the four breeds and their respective two age categories cannot be explained as the bulls were different in their initial body weight at the start of the experiment. However, animals in higher age groups attain higher (P<0.001) body weight records in both initial and nal body weight. This is in agreement with the previous nding by (Bassa et al., 2016).
There was no statistically signi cant difference in average daily weight gain of the four breeds grouped to two age categories. However, the average daily weight gain in the current study ranges from 0.49 to 0.71 kg with an overall ADWG of 0.63kg. The average body weight gain in the present study was in line with the nding by Bedhane and Dadi (2016) who reported ADG 0.63 kg for Ethiopian Borana breeds with an estimated age of 4 years old.
The attained results in this study for Borana bulls were higher than the report of Haile et al. (2009)  (2013) which was 0.87kg for Horro-Friesian crossbred bulls in the age from 2 to 3 years; O' Donovan et al., (1978) which was 0.88 kg for crosses between Borana bulls and Friesian dam as well as 0.98kg for a cross between Barca sire with Friesian dam at Holota research center. However, higher than the report by O'Donovan et al., (1980) who reported 0.54kg ADWG for crossbreds at Bakko research center.
The ADWG of Afar bulls supplemented with different levels of concentrate under improved feeding management were between 0.54 and 0.62 kg (Weldegebrial, 2018). This is in agreement with the value of Harar bulls in the current study.
Moreover, Tolla et al. (2003) reported comparable ADG 0.44 to 0.57 kg for Arsi bulls. In contrast to the current nding Teklebrhan (2019) reported that Hararghe highland bulls feeding different concentrate feeds revealed a higher ADG of 1.21 to 1.33 kg.
The feed conversion e ciency of the experimental bulls in this study was not signi cantly different (P>0.05) between breed, age as well as the interaction between breed and age. This is in line with the nding of Adebabay et al. (2013) who revealed the absence of a signi cant difference between Fogera and Adet breed bulls. Tolla et al. (2002) also reported insigni cant (P>0.05) differences among Borana and Arsi breed bulls. Similarly, Merera and Galmessa (2013) reported that feed conversion e ciency was higher for the lower age group but the differences were not signi cant (P>0.05) for Horro-Friesian crossbred bulls. The overall mean value obtained in this study was comparable with the feed conversion e ciency of Fogera (0.10) oxen reported by Adebabay et al. (2013) but higher than the value for Adet (0.09) oxen. Whereas higher values were reported by Tolla et al. (2002) for Boran (0.14) and Arsi (0.13).
The pattern of body condition score of three indigenous breeds and HF crossbred bulls for the two age categories were presented in Table 5. The mean body condition score was the highest (P < 0.05) for Arsi bulls compared to FH Crossbred while Harar and Borana bulls show intermediate body condition scores obtained during the start of experimental feeding. This variation might be due to the difference in animal management at their respective source before experimental feeding. HF crossbred bulls scored lower (P<0.01) body condition compared to all Borana, Harar and Arsi bulls in the nal period of feeding.
Meanwhile, body condition score is an excellent indicator of the nutritional status of the animals and is an indirect re ection of the body reserve, the highest body condition score obtained for Arsi bulls, as well as the higher BCS change in Borana and Harar bulls, reveals that these animals were in a better nutritional status and had better body reserve at the initial score and BCS change respectively. Therefore, change in body condition score is an indirect estimation of live body weight change. This is con rmed by several scholars in their works (Nicholson and Sayers, 1987;Osuji and Capper, 1992;Berry et al., 2006).
The estimated body weight change based on the change in body weight to change in BCS ratio was signi cantly (P<0.01) in uenced by breeds as shown in condition score change for HF-Crossbred bull obtained in the current nding might be due to the heavier body size of HF-crossbred bulls than other local cattle bulls of similar age. Therefore, fattening HF-crossbred bulls was advantageous in beef production improvement. ***=P<0.001, **=P<0.01; *=P<0.05, N= sample number, BCS= body condition score, kg= kilo gram, SE= standard error of mean.
The two age groups are not different in the initial BCS. Similarly, the BCS of the bull was the same under the two-age group during nal feeding periods. The same holds true for change in BCS and BW change. On the other hand, there was a signi cant breed by age interaction effect on the initial body condition score of the four breeds. i.e., the difference in initial BCS between the four breed bulls depends on the age in which they are grouped. The Arsi bulls score higher body condition in the lower age group than the higher age group. However, the reverse was true in the Borana, Harar and HF crossbred (Figure 1).
The relationship between BW change and BCS change was presented in (Table 6) The regression equation predicting live weight change per body condition change for the two-age group and the entire experimental bulls were presented in Table 7. Accordingly, it means that change of one condition score is equivalent to 22.66, 20.87, 22.65 and 28.66 kg body weight in Arsi, Borana, Harar and HF-crossbred bulls respectively. When these experimental bulls were arranged in two age categories one body condition score change in 2-3 years and 4-5 years old bulls represent 23.65 and 26.53 kg bodyweight change, respectively. The average change of one condition score for all experimental bulls was equivalent to 25.99 kg body weight. Comparable with the current nding Ali and Muna (2013) reported one condition score was equivalent to 26 kg in males for Sudan cattle. The value of BW associated with the unit of BCS change in the current study was lower than previous report by Berry et al., (2006) who reported a change of 31 kg BW for every unit change in BCS for Holstein-Friesian cattle in New Zealand, Fox, et al., (1999) also reported the relationship between BW change and BCS in Holstein cows was 84.6 kg/BCS. However, the value of 25.99 kg from the present study agrees well with the range of 21 to 35 kg BW per unit BCS in Holstein-Friesian cattle reported by Enevoldsen and Kristensen (1997). Jaurena et al. (2005) reported higher values for the regression 32 to 47 kg/BCS. Differences between studies in the relationship between BCS and BW change were due to BW is affected by breed, age sex and statistical models used (Berry et al., 2011). Bodyweight values obtained as per change in one unit of body condition score was increasing as the age of animals increasing. This is in agreement with earlier ndings (Osuji and Capper, 1992;Berry et al., 2006).
Body condition scores of experimental bulls were improved over the feeding period where bulls in the nal period of feeding had the higher (P < 0.01) body condition score than initial BCS. This is corroborated by the works that con rm body condition scores were improved over the feeding period (Mekasha et al., 2011;Merera and Galmessa, 2013).

Correlation between change in body weight and linear measurement
Pearson's correlation coe cient (r) between body weight change and linear body measurement change was presented in Table 11. In both age categories and overall measurement, all of the changes in linear body measurements have a positive association with changes in body weight. Highly signi cant association were found between body weight change and total topline (r = 0.94, P<0.0001), heart girth (r = 0.85, P<0.0001), chest depth (r = 0.88, P<0.0001) and unk circumference (r = 0.81, P<0.01) change in the younger age category. There was a highly signi cant and positive correlation between BW change and total topline change (r = 0.80, P<0.01) and (r =0.89, P<0.0001) in the second age category and overall, respectively. Whereas change in neck length (r = 0.72, P<0.05) and (r = 0.57, P<0.05) moderately in uence body weight change in the rst (2-3years) and second age group (4-5years) respectively. Change in heart girth (r =0.68, P<0.05), ank circumference (r = 0.69, P<0.05) and shoulder width (r = 0.69, P<0.05) has signi cant positive association with BW change in the second age group. Considering the overall experimental bulls in the relationship also shows a signi cant correlation between change in body weight and some of the linear body measurements. Likewise, the highest relationship was found between body weight change and change in total topline (r = 0.89, P<0.0001), change in Heart girth (r = 0.75, P<0.0001), change in ank circumference (r = 0.73, P<0.0001). moreover, there was a moderately signi cant correlation between BW change and change in shoulder width (r = 0.56, P<0.05) and chest depth (r =0.55, P<0.05). Change in ramp width was signi cantly correlated with BW change. However, there was no signi cant correlation between body weight change and change in hip length and rump length in both age groups as well as overall experimental bulls. Meaning that they did not signi cantly affect the change in body weight, so they were not more important in the prediction of live body weight change of the animal because of weak association with body weight change. This study is in agreement with some of the previous works by Vanvanhossou et al.( 2018) that reported as correlation coe cients within age groups con rm the in uence of these parameters on the relationships between live body weight and linear body measurements. Tsegaye et al. (2013) also suggest that the best estimations of body weight from linear body measurements should be developed from those that showed a strong correlation.
Generally, as the result of correlation showed a change in total topline, neck length, heart girth, ank circumference, rump width and chest depth was the most important than other linear body measurement change for both age categories to estimate body weight change per unit change of linear measurement during fattening.
The prediction equations to estimate body weight change per unit change of linear body measurements were presented in Table 12. Individual linear body measurement parameters were used to establish the most appropriate models predicting the live body weight change of bulls during the fattening period. The different age categories as well as overall bulls were analyzed separately.
In the rst age category regression model reveals that about 91% of the change in BW was in uenced by a change in total topline and the remaining 9% was in uenced by other factors outside the model such as cattle condition when measured and weighed, measurements method, the accuracy of measuring instruments and others. According to the Model, increasing 1 cm of the total topline of bulls was equivalent to increasing 8.21 kg of BW. Whereas change in 1 cm of total topline represents 9.6kg and 8.0kg BW change in the second age group and when overall experimental bulls considered in the model respectively. In accordance with the present study, Bhagat et al. (2016) observed the highest R 2 value when body length alone was included in the regression model in 2-3 year Sahiwal male cattle. The variation in live body weight was also explained by the change in neck length, in this regression model, increment in 1cm neck length represent 16.96kg for bulls under 2-3years age, 8.24kg for 4-5years age and 12.01kg in overall experiment.
However, a maximum R 2 value of 0.52 and a minimum R 2 value of 0.33 were observed in the 2-3years and 4-5years age group respectively.
The coe cient of determination (R 2 ) value in simple linear regression using HG as the independent variable in both age groups was included in the high category ( increasing 1 cm of HG was followed by increasing 4.83kg, 3.27kg and 4.13kg of BW for 2-3year, 4-5years and overall estimation respectively. In line with the current nding Putra (2020) reported change in 1cm HG was equivalent to 4.55 kg for bull and 3.26 kg of BW for a heifer. Goe et al., ( 2001) also estimated a 1 cm change in heart girth would result in a weight change of 3.4 to 4.7 kg for Ethiopian highland oxen. Odadi (2018) reported a 1 cm change in heart girth would result in a weight change of 3.37 kg for Bos indicus cattle in Kenya.
Change in ank circumference in all age categories indicates that about 77% to 79% in uenced BW change. One unit change in ank circumference was equivalent to 3.18 to 4.88kg of BW in the current nding. 1cm change in chest depth was also equivalent to 2.35kg BW change in bulls under 2-3years, 3.2kg for 4-5years and 2.75kg BW change in overall estimation.
The higher R 2 value and smaller RMSE attained using a single predictor in each age categories showed that the linear body measurements used as independent variables were good estimators of body weight. Generally, rump width was the most determinant linear body measurement variables in establishing a regression model in predicting BW change per unit of rump width change during fattening in all age categories.

Economics of Fattening
The partial budget analysis for the experimental feeding was reported in Table 13 which involved the evaluation of overall pro tability. The economic analysis shows the highest (P<0.001) net bene t obtained from Borana bulls with a net return of 7,380.47 ETB per head followed by Harar bulls, Arsi and HF Crossbred with a net pro t of 5,406.86, 5,193.29 and 3,384.98 ETB per head respectively. The variation of net bene t among different breeds was mainly similar to their trend in weight gain difference among breeds. BCS and physical appearance of individual bulls also have a direct in uence on the nal sell price of each bull. Similar to this result, Adebabay et al.(2013) revealed that the net bene t of Fogera old oxen increased with increasing weight gain that resulted from the increased quantity of supplementation and nal selling practice is based on body condition. According to the current result, feeding 3% of body weight of total mixed ration was highly (P<0.01) pro table for Borana bulls which may be affected by total weight gain, physical appearance and body condition score. HF-crossbred bull was economically less (P<0.01) pro table compared with other breed bulls but statistically similar net pro t was obtained from Arsi bulls. This indicates that HFcrossbred bull was less pro table under a limited feeding system because of their feed consumption is high as compared with other indigenous zebu bulls. This is concomitant with the report by Gojam et al. (2017) revealed that farmers around Holota agricultural research center were not even willing to use the crossbred cows, because of high feed consumption and di culties to handle the animals.
Even though it is obvious from this result that the bulls in both age groups are statistically similar in net pro t, one can go for the older groups which gives higher net return.
Generally, feeding 3% of body weight of total mixed ration for 90 days was economically pro table providing an overall mean of 5,341.40 ETB per head of bulls. However, further investigation on fattening HF-crossbred bulls under different levels of total mixed ration offering was recommended to evaluate economic pro tability based on their weight gain and condition score performance.

Conclusion And Recommendation
From the result, it was concluded that the average daily weight gain of the four breeds range from 0.49 to 0.71 kg. Change in body weight and body condition score was positively and strongly correlated in all breeds and age groups.
Fattening HF-crossbred bulls was advantageous in beef production because of higher body weight change per one condition score change. Changes in linear body measurements have a positive association with changes in body

Con ict of interest
The authors declared that there is no con ict of interest between authors and organizations regarding this paper.   Breed by age interaction effect on initial body condition score