The on-farm feeding experiment was conducted between April and June, 2020 for 60 days to demonstrate and evaluate oats-vetch mixed fodder-based feeding system on feed intake, milk yield, milk quality and cost benefit analysis using crossbred dairy cows under farmers’ management condition. The study was conducted in two selected milk shed districts (Ejere and Wolmera), West Shewa Zone of the Oromia Regional State in the central highlands of Ethiopia ( Figure 1 ). The two districts were typical representatives of market-oriented smallholder crop livestock production systems.

Ejere district is located 40 km west of Addis Ababa between 8°51’16”N to 9°14’53”N and 38°15’2”E to 38°28’45”E. The altitude of the area ranges from 2060 to 3185 masl with an average rainfall of 1100 mm. The minimum and maximum temperature of the district is 9 and 28°C, respectively (Ejere Agricultural Development Office, 2020). The district is characterized by a crop-livestock mixed farming system where cereal crops such as Teff, Wheat, Barley and Maize are commonly grown. Cattle, sheep and goat are the dominant ruminant species that support the livelihood of the farming community in the district. This district is one of the milks shed areas and well linked to fluid milk market. Moreover, from the various crops grown in the area, different crop residues are produced and conserved for use as feed during periods of feed shortage. The area is closer to Holetta Agricultural Research Center and easily accessible for monitoring and sampling of crop residues for laboratory assessment.

Welmera district is located 30 km West of Addis Ababa along the main road to Ambo. Geographically, the area is situated at 8°50’ - 9°15’N and 38°25’ - 38°45’E with an altitude ranging from 2060 to 3380 masl. The district has a bimodal rainfall patterns. The minimum and maximum annual temperature is 0.1 and 27°C, respectively. Crop-livestock mixed farming is a typical agricultural practice in the area. Livestock production consists of cattle, sheep, goats, equines and poultry. Dairy production is a common practice in the district and because of its proximity to Addis Ababa with better market access, fresh milk sale is the major source of income for producers.

As part of the adoption and demonstration of forage technologies under the fodder component of the EQUIP project, improved forage materials were distributed and evaluated for their agronomic performance, yield and quality parameters under the farms of more than 140 dairy farmers in different locations around the Addis Ababa milk shed. Among the different improved forage materials demonstrated, further study was conducted to evaluate the feeding value of oats and vetch mixture at a ratio of 3:1 by seed weight. Twelve volunteer dairy farmers who were willing to allocate an average of 0.25 ha land for oats-vetch production, who own two crossbred dairy (Boran-Frisian cross) cows and who were willing to keep record and provide data were selected among the participants. The oats-vetch mixture was planted during the main rainy season (last week of June 2019) and fertilizer (diammonium phosphate) was applied at the rate of 100 kg/ha at the time of planting. The oats-vetch mixture was harvested 3 months after planting when the oat was at soft dough stage and when the vetch attained about 50% flowering stage. Before the total harvest, representative samples were collected using a 1m-by-1m quadrat, from three random areas per plot for yield and quality estimation. Oats-vetch mixed forage was then harvested, properly dried as hay and stored under shade until the feeding trial commenced.

Twelve farmers who own at least two lactating cows and who own sufficient amount of feed resources were considered for this study. From each of the 12 farmers, two lactating cows were selected, one of which was subjected to the traditional feeding practice and the other one to the oats-vetch based intervention diet. The intervention diet, T1, based on oats-vetch mixture as a basal diet supplemented with a concentrate mixture. The control group, T2, (farmer’s practice which consisted of native pasture hay, crop residue and locally available supplemental feed resources). The cows were fed and watered individually. For the intervention diet, the basal feed was offered ad libitum and the concentrate mixture was offered at the rate of 0.5 kg per liter of milk produced. A concentrate mixture was prepared from wheat bran, noug (Gizotia abyssinica) seed cake and salt at a ratio of 60:39:1, respectively. The cows were exposed to the diet for an adaptation period of 14 days, followed by 60 days of actual data collection. Samples of offered feed ingredients were collected to determine chemical composition and nutritive value. Milking was conducted by hand twice per day at 6:00 am in the morning and at 6:00 pm in the evening. The data on feed intake, milk yield, milk quality, variable cost and milk sale price were collected and subjected to analysis using the independent t-test.

Both theoretical and practical training was provided to participant farmers (two members per household), development agents (three per district) and district experts (one per district) on improved forage production, management, ration formulation and utilization at Holetta Research Center. Concentrate mixture was provided to participants in the intervention group that is sufficient to supplement one lactating cow for the experimental period. Before feeding, the oats-vetch mixture was properly chopped, weighed and stored in clean sacks.

The chemical composition and digestibility of the feed samples were analyzed at the national animal nutrition laboratory of Holetta Agricultural Research Center and International Livestock Research Institute. Dried samples were ground to pass through a 1mm sieve. Near Infrared Spectroscopy (NIRS) was used to analyze the samples. The NIRS machine used equations calibrated and validated by the conventional wet chemistry analysis (AOAC, 1990). The NIRS instrument used was a FOSS Forage Analyzer 5000 with software package WinISI II. Predicted nutritional variables were nitrogen (N) (Crude protein = N*6.25), neutral detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin (ADL), dry matter (DM), total ash and in vitro organic matter digestibility (IVOMD). Metabolizable energy (ME) was estimated from digestible energy (DE) which in turn was estimated from the IVOMD (NRC, 2001): DE = (0.01*(OM/100) *(IVOMD+12.9) *4.4) - 0.3; ME (MCal/kg) = 0.82*DE.

Milk samples were collected in a clean plastic cup at both morning and evening milking times during the last seven days of the experimental period and immediately placed inside an icebox and taken to laboratory for milk composition analysis. The milk samples were analyzed for fat, protein, total solids, and lactose contents. NIR-Infrared milk product analyzer (user manual ver. 1.1, 2000) was used for analyzing milk composition.

The partial budget analysis was performed to evaluate the economic advantage of the intervention diet over the traditional feeding practices. The analysis involved the calculation of the variable costs of experimental feeds and other inputs and milk sale. The feed cost for both treatments were computed by multiplying the average daily feed intake per cow with the current feed price at the local market. The total variable cost of improved forage, native pasture, crop residue and concentrate feeds was estimated using the local market price. Other input costs were calculated as 30% of the feed cost because 70% of the total cost of production for dairy production is assumed to be feed cost. The average daily revenue was calculated based on the amount of average milk yield/cow/day multiplied by farm gate milk selling price. Based on the price given by each dealer, the average selling price for each treatment was calculated. In the analysis, the average daily return (TR) was determined by calculating the difference between the average daily milk sale per cow per day and the average cost of variables per cow per day. The partial budget analysis measures profit or losses which were the net benefits or differences between gains and losses for the proposed change and includes calculating net return (NR), i.e., the amount of money left when total variable costs (TVC) are subtracted from the total returns (TR).

Total variable costs include the costs of all inputs that change due to the change in production technology. The change in net return (ΔNR) was calculated as the difference between the change in total return (ΔTR) and the change in total variable cost (ΔTVC), and this is to be used as a reference criterion for decisions on the adoption of new technology.

The marginal rate of return (MRR) measures the increase in net income (ΔNI) associated with each additional unit of expenditure (ΔTVC). This is expressed in percentages as:

The data was subjected to independent T-test analysis using the Stata procedure (StataCorp, 2017) to compare feed and nutrient intake, milk yield, milk quality and cost benefit analysis between intervention and control groups. The following statistical model was used for the analysis. Yij = μ + Xi + Eij. Where, Yij = is the response variable, μ = over all mean, Xi = the treatment effect and Eij = random error.

The feed type and proportion of ingredients in the intervention and traditional diet are indicated in Table 1 . For the intervention group, the feed type grown were oats-vetch mixed forage in the proportion of three to one respectively. In this group, the basal diet and the concentrate mixture consisted of 56 and 44% of the diet respectively. While in the traditional feeding system, the ration was formulated from basal diet and concentrate mixture at the rate of 62% and 38% respectively. Within the traditional system, the basal diet used were crop residue (55%) and natural pasture hay (45%). When the two groups are compared, the proportion of the basal diet was higher in the traditional group while the proportion of concentrate mixture was higher in the intervention group ( Table 1 ).

The chemical composition and in vitro dry matter digestibility of feed ingredients and dietary treatments are indicated in Table 2 . Based on the laboratory evaluation, the CP content and IVDMD were observed to be higher for the intervention diet by 30% and 16% respectively. On the other hand, the NDF and ADF content was lower for the intervention diet by 13% and 14% respectively. Dry matter and nutrient intake of the intervention and traditional group are indicated in Table 3 . There was no variation in total dry matter intake between the two feeding systems. However significant difference (P<0.05) was observed in basal diet and concentrate intake between the two groups.

Milk yield, milk quality and feed conversion ratio data of the intervention and traditional groups are indicated in Figure 2 . The use of oats-vetch mixed fodder adlibtum and supplementation with 0.5 kg of concentrate mixture per liter of milk, increased daily milk yield by 4 liters per cow per day compared to the traditional practice.

The average milk yield in the intervention group increased from 8.5 to 11.6 liters per cow per day for the first three weeks and then remained stable afterwards while there was no change observed (7.5 liter per cow per day) in the traditional feeding practice ( Figure 3 ). Similarly, the dry matter intake increased with increasing level of milk and remained constant after the third week, which indicates that the milk yield potential of the cow does not go beyond that level and feeding higher than that is not economical ( Figure 4 ). In the intervention group, feed conversion efficiency decreased for the first three weeks and become stable for the following 5 weeks ( Figure 5 ). While in the traditional diet it remained constant but higher than the conventional diet. This implies that the amount of feed required per liter of milk was higher for the traditional feeding practice.

Total variable cost, gross return, net benefit and marginal rate of return are indicated in Table 4 . Based on the partial budget analysis, the highest net benefit was obtained from the use of intervention diet (ETB 95/cow/day) compared to the traditional group (ETB 45/cow/day). It was observed that the intervention diet had the highest variable cost, however the net benefit obtained was much higher than the variable cost. As indicated in Figure 6 , income increased in the intervention group for the first three weeks and remained stable for the following weeks whereas in the traditional practice there was no change observed throughout the study period. This shows that the income in the intervention group remained higher compared to the traditional group. This confirms that the adoption and further scaling of such promising improved forage technologies bring change in milk productivity and economic benefit to small scale farmers.

Soaking of crop residue with local brewery and water is commonly practiced in the traditional feeding system and such practices observed to improve intake. In the traditional feeding practice, feed provision was not based on the performance of the dairy cow and most farmers provided adlibtum amount of crop residue supplemented mostly with energy sources (molasses, wheat bran, pulse haulms or maize grain). On the other hand, some farmers used native pasture hay or crop residue as a basal diet and brewery by product as a supplement. This shows that the ration was not properly formulated in terms of protein and energy balance because of poor awareness on balanced ration formulation and higher cost of commercial supplements. It was also observed that ration formulated varied among smallholder farmers because of differences in available feed resources in each household and locality. Farmers usually purchase concentrate feeds that are cheaper and available in their localities without due consideration on quality parameters.

The dry matter and crude protein yield of oats-vetch mixed fodder was in the range of 6.2 to 6.7 t/ha and 0.5 to 0.54 t/ha respectively. This was in line with the report by (Shimelis et al., 2021) who indicated that dry matter (DM) yields of the oats-vetch mixed fodder ranged from 6.4 to 7.1 t/ha. Mixtures containing 25-50% legume produce more quality and quantity forage per unit area than those of pure sowings (Alemu et al., 2007). This report was in agreement with our study where we considered 25% of the basal diet from vetch. The current study revealed that the quality of oats-vetch mixed fodder is higher than pure oats hay, native pasture hay or crop residues. This is in agreement with previous report by (Negash et al., 2017) who stated that mixed cropping of cereals with forage legumes improved both quality and quantity of fodder over a pure cereal crop.

Smallholder farmers who participated in oats-vetch mixed fodder production and utilization indicated that they are well aware of the quality and higher dry matter yield compared to the existing traditional forage production practices. In support of our findings, previous study demonstrated that the use of grass legume mixture provides attractive opportunities for sustainable animal production as they contribute to increases in forage yield, substitute inorganic nitrogen inputs for the symbiotic fixation and also increase the nutritive value of the forage (Luscher et al., 2014).

The basal feed intake in the intervention and traditional group was 57 and 63% of the total dry matter intake respectively. High values of fibrous materials in the traditional feeding practice are likely to depress digestibility of crop residue and native pasture that constituted large segment of the diet. On the other hand, the concentrate intake was 48 and 37% of the total dry matter intake in the intervention and traditional group respectively. This shows that the basal diet intake was significantly higher in the traditional group whereas the concentrate intake was higher in the intervention group.

Crude protein intake for the intervention group was 1.57 kg per cow per day that was 13% of the total dry matter intake. On the other hand, the CP intake for the traditional feeding practice was 1.21 kg per cow per day that accounts for 10% of their total dry matter intake. Fiber (NDF) intake was higher for the traditional practice (64% of the total DM intake) than the intervention group (58% of the diet). In this study even though there was similar total dry matter intake between the two feeding systems, there was variation in nutrient intake. For instance, crude protein intake was observed to be higher in the intervention group by 23% compared to the traditional practice. Higher CP intake associated with higher digestibility has contributed to higher milk yield in the intervention group (11.67 kg per cow per day) compared to the traditional feeding practice (7.52 kg per cow per day). This indicates that promotion of improved forage development and supplementation with balanced concentrate mixture has enhanced supply of good quality feed resource.

The basal diets in the traditional feeding practices (wheat, barley and oat straw) were identified to be poor in quality that has low CP (< 2.8%) and digestibility (<48%). Such kind of traditional feeding practice based on poor quality fodder has limited the production potential of the crossbred dairy cows. In agreement with our findings, previous report indicated that traditional feeding of crossbred dairy animals suffer from under feeding in terms of nutrient requirement that caused the animals to produce below their potential (Mohamed et al., 2004).

In the current study, milk yield was significantly higher by 36% in the intervention group compared to the traditional practice. This difference in milk yield between the two groups is associated to the quality of the feed and proper formulation of balanced ration. In agreement to our findings, a study conducted in Kenya indicated that feeding oats and vetch fodder increased milk yield by up to 20% over the traditional practice and to realize adoption of such promising forage technologies, awareness creation and functioning forage seed systems need to happen concurrently (Mwendia et al., 2018). Similarly, a report by (Hintsa, 2016) indicated that the use of the oats-vetch mixed fodder as a supplement to traditional feeding resulted in daily additional milk yield of 2.33 liters per cow per day for crossbreed cattle, and 1.0 liter for local cows. Another study by Marishet et al., 2019 indicated that crossbred cows fed a basal diet of oats-vetch mixture supplemented with cottonseed cake-based concentrate mixture provided 7.97 kg of milk per cow per day which was lower than the present findings. These differences in milk yield might be associated to the variation in the level, type of ingredients, feeding management of animals among smallholder farmers and also the difference in the performance of dairy cows.

On the other hand, there was no difference observed in milk fat, milk protein, solid not fat as well as lactose content between the intervention and control group. The milk fat and protein content in the intervention diet were 4.43 and 3.15% respectively, which was in agreement with the previous report by Marishet et al., 2019, who indicated that the milk fat and protein content of dairy cows fed oats-vetch fodder were 4.44% and 2.9% respectively. The variation in milk protein in the present study can be attributed to the usual inverse relationship between milk yield and the content of milk solid components. In general, the feed conversion ratio was significantly (P<0.05) lower for the intervention group implying that less dry matter intake was required to produce a unit of milk compared to the traditional group.

Economic visibility study is the main driving factor for adoption of feed technologies by smallholder farmers (Shimelis et al., 2021). Under the current price scenario for feed and milk, the higher benefit was obtained from the intervention diet. The marginal rate of return obtained from the intervention diet implies that for 1.0 birr invested in milk production, the producers obtained 2.66 Birr, an achievement that is above the minimum acceptable rate of return (CIMMYT, 1988). This study demonstrates that the adoption of improved feeding practices with sufficient training on formulation of balanced ration can result in higher economic benefit for smallholder dairy farmers. The cost benefit analysis of using oats-vetch mixed fodder for milk production resulted in a positive response. The variation in net return between treatments occurred due to the difference between the quality of feed and total milk yield per treatment.

Most of the participant smallholder farmers described that farm gate milk price is lower and it is decided by milk collectors despite the continuous increase in feed cost. This has been discouraging farmers from sustainable involvement in dairy production. Unless and otherwise feeding management is improved, the efficiency of livestock in terms of production and productivity as well as economic benefits are critically limited. This study demonstrated that changes in dairy feeding based on improved forage have shown significant improvement on farmer’s livelihood through increased dairy productivity and economic return. In agreement with our study, Shimelis et al. (2021) reported that the use of good quality home grown forage can result in higher economic benefit for smallholder farmers.

At the end of this study, participant smallholder farmers visualized that, production of improved forage at their own farm land enabled them to obtain good quality fodder with higher dry matter yield at a cheaper price and that significantly increased milk yield performance and economic return. Thus, it was taken as a lesson learnt for further adoption and wider application of the technology.