Maize Response to Blended Fertilizer Estimates through Nutrient Use Eciency and Economic Analysis from Assosa, Western Ethiopia

Context: Soil fertility depletion and soil acidity are critical problems for maize (Zea mays L) production in western Ethiopia. Objective: Therefore, a eld experiment was conducted on Nitisols of Assosa Agricultural Research Centre during 2016 to2017 cropping season to investigate the response of growth, yield and nutrient use eciency of maize (Zea mays L.) to different blended fertilizer rates and types. Methods: The treatments consists of: control, three rates of N and P combined (92/46, 115/57 and 138/69 N/P 2 O 5 kg ha − 1 and two formula of blended fertilizers with different rates, formula 2 consists of 100 kg NPSB + 73.9 N, 150 kg NPSB + 110.8 N and 200 kg NPSB + 147.8 N kg ha − 1 and formula 4 consists of 100 kg NPSZnB + 75.1 N, 150 kg NPSZnB + 112.6 N 1 and 200 kg NPSZnB + 150.2 N kg ha − 1 . The treatments were laid out as a Randomized Complete Block Design with three replications. Result and and conclusion: Application of blended fertilizers (NPSB, NPSZnB) hastened days to tasseling silking and maturity by 10, 7 and 15 days, respectively as compared to combined N and P rates. Application of blended fertilizer increases signicantly (p ≤ 0.01) the plant height, cob weight, ear length, 100 kernels weight, number of kernels per row and ear height as compared to combined N and P and the control. The analysis of variance revealed that fertilizer types and rates signicantly (P ≤ 0.01) affected on grain yield, straw yield and harvest index. However there was no signicant difference between the two blended fertilizer types. Maximum grain yield (7056.2 kg ha − 1 ) was recorded with 200 Kg NPSZnB + 150.2 N kg ha − 1 application, while minimum grain yield 2996.0 kg ha − 1 was recorded from control treatment. The application of 150 kg NPSB + 110.8 N kg ha − 1 had highest Marginal rate of return (MRR%) and net benet. Signicance: Therefore, we recommended the treatment (150 Kg NPSB + 110.8N kg ha − 1 ) since it produced high marginal rate of return, high net benet and relatively small total cost of production, for maize production in Asossa area. Furthermore, based on yield, net benet and relatively low total cost of production the farmer of Asossa area also can use 150 kg NPSZnB + 112.6 N in case of absence of NPSB in market. by sacricing yield, but that would not be economically effective or viable for the farmer, or the environment. The protability of the study showed that application of 150 kg NPSB kg + 110.8N and 150 kg NPSZnB + 112.6 N kg which provided relatively high net benet (32321.4 and 30,478.1ETB) was the best rates to apply. Marginal rate of analysis from undominated treatments indicated that for each one birr invested in purchase or production of fertilizers that was possible to recover one birr plus an extra of 4.36 birr ha − 1 and 53.29 birr ha − 1 as the fertilizer application changed from unfertilized plot to 100 kg NPSB + 73.9 N kg ha − 1 and 150 kg NPSB + 110.8N kg ha − 1 respectively.


Introduction
Among cereals, maize (Zea mays L.) ranks third after wheat and rice in the world 1 . It is one of the most important cereal crops used for human diet in large parts of the world, besides served as important feed component for livestock. In terms of total world production, maize out ranked paddy rice and wheat (Rasheed et al., 2004).
Cereals are the primary food crops in Ethiopia covering 79.88% (9,974,316 hectares) of total crop lands, while maize covered 16.91% (about 2, 111,518 hectares) of land allocated for cereal and gave 7,150,835.4 tons of grain yield (CAS 2015/2016).
Despite the large area under maize, the national average yield of maize is about 3.387 t ha − 1 . This is by far lower than the world's average yield which is about 5.21t ha − 1 and also lower than the average national research center based yields (8 t ha − 1 ) in Ethiopia (Hagos et al., 2016). These are attributed to many factors: lack of access to seeds of improved varieties, poor soil fertility, and diseases and insect pests.
Low soil fertility is one of the bottlenecks to sustain agricultural production and productivity in Ethiopia. The problem is aggravated by several factors which include among others, soil erosion, nutrient mining, soil acidity and low level of application of nitrogen and phosphorus. Furthermore, unbalanced application of plant nutrients may aggravate the depletion of other important nutrient elements in soils such as K, Mg, Ca, S and micro-nutrients (Getachew 2013). Thus, maize is one of the heaviest feeder of nutrients to produce high and quality yields among cereals. This is because of the fact that it produces higher grain and straw yields than other cereals. Hence application of balanced fertilizers is the basis to produce more crop output from the existing land under cultivation as nutrient needs of crops is according to their physiological requirements and expected yields (Ryan 2008). Most of the fertilizer experiment in Ethiopia focus on N and P requirements of crops, hence limited information is available on various sources of nutrients such as K, S, Zn and B and other micronutrients. However, recently blended fertilizers were introduced to Ethiopia to combat the limitations of various essential nutrients in crop production. Therefore, it is paramount important to assess maize response towards different blended fertilizers such as NPSB and NPSZnB, besides the conventional fertilizers N and P fertilizers application.
Nutrient use e ciency is the ability of a plant to utilize soil available nutrients to result in measurable yield or yield parameters such as plant height, leaf development, dry matter and fruit grain production (Hati et al., 2006). Partial factor productivity (PFP), agronomic e ciency (AE), uptake e ciency (RE), physiological e ciency (PE) and internal utilization e ciency (IE) constitute a set of simple indices and could be used in agronomic research to appraise the applied fertilizer e ciency particularly to assess the short-term response of crop to a nutrient ). The 'difference method' (calculation of nutrient use e ciencies using differences in crop yield and/or nutrient uptake between fertilized plots and an unfertilized control) is cost-e cient and simple making it particularly suitable for on-farm research . However, there is a lack of scienti c studies to examine the effect of blended fertilizer in improving the yield and nutrient uptake e ciency of maize in Asossa District. The amounts of nutrients exploited in the harvested portion of the crop will depend on the yield and the concentration of the nutrients in time and space, variety, soil and environmental factors (Fageria, 2011). To use fertilizer in a sustainable manner, management practices must aim at maximizing the amount of nutrients that are taken up by the crop and minimizing the amount of nutrients that are lost from the soil (Bationo et al., 2012). Therefore this experiment was designed to assess the role of blended and conventional fertilizer effects on yield, economic importance and nutrient use e ciency of maize at Asossa district

Description of the Study Sites
The experiment was conducted at Assosa Agricultural Research Center (AsARC) in Benishangul Gumuz Regional State, in 2016/17 main cropping season under rain fed eld condition. The region is located between 9°30' to 11°39'' N latitude and 34°20' to 36°30''E longitude covering a total land area of 50,000 square kilometer. The study site is located at 10°02' 05'' N latitude and 34 O 34' 09'' E longitudes. The study area is situated east of Assosa town and west of Addis Ababa about 4 km and 660 km distance, respectively. Assosa has unimodal rainfall pattern, which starts at the end of April and extends to mid-November, with maximum rainfall amount received in June to October. The total annual average rainfall of Assosa is 1275 mm. The minimum and maximum temperatures are 16.75°C and 27.92°C, respectively. The dominant soil type of Assosa area is Nitosols with the soil pH ranges from 5.0 to 6.0.

Experimental Design and Treatments
The experiment was laid out in randomized complete block design with three replications.

Plant Tissue Sampling and Analysis
Representative grain and straw samples were taken from each plot at crop physiological maturity. The samples were oven dried and ground for laboratory analysis of total N, P, S and K. The measurement of N was carried out according to the Kjeldahl procedure by transforming organic N into ammonium N by digesting with H 2 SO 4 and a catalyst (Chapman 1965). Potassiumwas measured using dry ashing, by ame Photometer as described by (Chapman 1965).Phosphorus was determined by spectrophotometer using the dry ash of maize samples. Total Swas analyzed using Calorimeter. The grain and straw concentrations of N, P, S and K were used to estimate the uptake of representative elements which was calculated by multiplying grain and straw yields on hectare basis with the respective N, P, S and K concentration.
Apparent fertilizer N and P recovery were calculated following the formula as: AP = (TU fertilized treatment -TU control) x 100; Amount of nutrient applied kg/ha Where AP = Apparent recovery, TU stands for total nutrient uptake at 'n' rate of fertilizer nutrient.
A Physiological e ciency is the yield obtained per unit of nutrient uptake (Fageria, 2009) Physiological e ciency = grain yield of fertilized kg/ha -grain yield unfertilized kg/ha Nutrient uptake of fertilized in kg-Nutrient uptake of control in kg Agronomic e ciency = grain yield of fertilized kg/ha -grain yield unfertilized kg/ha Amount of fertilizer applied (kg/ha)

Partial budget Analysis
Economic analysis was performed to investigate feasibility of fertilizer application for maize production in Assosa district (CIMMYT , 1998). Mean grain yield of maize was used for the study. A partial budget, dominance and marginal analysis were used. The average open market price (Birr kg − 1 ) for maize and the o cial prices of blended, Urea and TSP fertilizers were used for economic analysis. The dominance analysis procedure as detailed in (CIMMYT , 1998) was used to select potentially pro table treatments from the ranges that were tested. The selected and discarded treatments using this technique are referred to as undominated and dominated' treatments, respectively. The undominated treatments were ranked from the lowest (the farmers' practice) to the highest cost treatment. For each pair of ranked treatments, percentage marginal rate of return (MRR) was calculated. The MRR between any pair of undominated treatments denotes the return per unit of investment in fertilizer and expressed as a percentage.

Statistical Analysis
Analyses of variances for the recorded data were conducted using SAS GLM procedure (SAS, 1998). Least signi cant difference (LSD) test at 5% probability was used for mean separation when the analyses of variance indicate the presence of signi cant difference.

Result And Discussion
Effects of blended fertilizer rates and types on phonological, growth and yield components of maize.
Days to 50% silking, tasseling, maturity, Ear height, Plant height, Ear length, 100 kernels weight and Number of kernels per row except Number of kernel row per cob were highly signi cantly (P ≤ 0.01) affected by application of fertilizer types and rates (Table 1). Early silking, tasseling and maturity were recorded from plots which received blended fertilizers followed by recommended N and P. On the other hand the longest days to 50% silking, tasseling and maturity were recorded from the control (without fertilizer applied). This result is in agreement with the nding of (Dagne ,2016), who indicated early tasseling, silking and maturity days were recorded with the application of blended fertilizer and the longest days to 50% tasseling, silking and maturity In conformity with the results obtained from this study, Plant growth and development may be retarded signi cantly if any of nutrient elements is less than its threshold value in the soil or not adequately balanced with other nutrient elements (Landon, 1991).
Similarly application of treatments had increased the ear length, 100 kernels weight, number of kernels per row as compared to recommended N and P and the control. Blended fertilizer rates and types had highly signi cant (P ≤ 0.01) effects on ear length of maize (Table 1). However, there were no signi cant differences between the two formula types. Blended fertilizer which contains B improved cob weight. This results agree with the nding of Mozafar (1989), who reported that application of B fertilizer to maize production encourage good cob development. The largest ear length (16.10cm) was obtained under the application of 200 kg NPSZnB + 150.2 N (T10), while the shortest ear length of maize (11.57cm) was recorded under the control.
The two types of blended fertilizer formulas (NPSZnB and NPSB) gave similar response to these parameters. Comparing the 100 kernels weight showed that 150 kg NPSZnB + 112.6 N kg ha − 1 application resulted in 20.05% and 15.15% more 100 kernels weight as compared to the control treatment and recommended N and P respectively (Table 1). Both blended fertilizer types (NPSZnB and NPSZnB) gave more response to number of kernels per row than recommended N and P, and the control. The maximum number of kernels per row (37.10) was obtained under application of (T10), while minimum number of kernels per row (24.23) was recorded under the control plants.
The analyzed data of number of kernel row per cob indicated that non-signi cant difference was observed among fertilizer rates tasted. Application of blended fertilizer T10 increases the number of kernels per row by 53.11% over the control plot. As compared to the recommended N and P, the mean value of number of kernels per row increased by 20 Effects of blended fertilizer rates and types on maize grain and straw yields The analysis of variance for grain and straw yields revealed highly signi cant (P ≤ 0.01) difference among the treatments.
However the two blended fertilizer types was not signi cantly different in affecting most parameters ( Table 1). The two types of blended fertilizer had signi cantly improved grain yield, which might be attributed to the contribution of relatively more types of nutrients (N, P, S, B and Zn) supply in the blended fertilizer as compared to the recommended N and P, and control. The low yield of maize under application of recommended N and P might be due to the absence of other macro and micro nutrients(S, Zn and B). Similar trend has been reported by (Boorboori et al., 2012).
Application of 200 kg NPSZnB + 150.2 kg N ha − 1 increased grain yield by 135.5% and 111.1% over control and 100% recommended NP, respectively. Similar trend has been observed by 18 in wheat crop. Grain yield increment with application of blended fertilizers which contained S, B and Zn indicated that there is a need to supplement these elements for maize production.
The increase in grain yield could be attributed to bene cial in uence of yield contributing characters and positive interaction of nutrients in the blended fertilizer (Dagne 2016). The strong relationships were found between grain yield and ear length, grain yield and 100 kernels weight, and number of kernels per row.
The low yield in unfertilized plots might have been due to reduced leaf area development resulting in lesser radiation interception and, consequently, low e ciency in the conversion of solar radiation (Sallah et al.,1998).
The highest maize straw yield (7886.7 kg ha − 1 ) was recorded with 150 kg NPSB + 110.8Nkg ha − 1 application, while the lowest value (4400.9 kg ha − 1 ) was recorded with control treatment. Application of 150 kg NPSB + 110.8N kg ha − 1 resulted in 79.21% and 54.0% more straw yield as compared to the control and recommended N and P. Means followed by different letters within column showed signi cant differences at 5% while means followed by the same letter(s) with in a column are not signi cantly different at 5%. LSD = least signi cant difference at 5%. CV = coe cient variation.

Effects of blended fertilizer rates and types on maize harvest index
The physiological ability of maize to convert total dry matter in to grain yield is determined by its harvest index (HI). The analysis of variance revealed that fertilizer rates and types had highly signi cantly (P ≤ 0.01) in uenced harvest index of maize. However there were no signi cant differences between the two blended fertilizer types effect on HI of maize (Table 1). Both blended fertilizer types (NPSZnB and NPSZnB) gave more response to harvest index than recommended N and P and the control.
Nevertheless, non-signi cant difference between recommended N and P and control was observed with regard to harvest index.
The highest harvest index (0.49) was obtained at application of 200 kg NPSZnB + 150.2 kg N ha − 1 while the lowest harvest index (0.40) was recorded under the recommended N and P.
The increase in the harvest index due to application of micronutrients may be attributed to their role in enhancing the photosynthesis process and translocation of photosynthetic products to economic part. Generally application of blended fertilizers had signi cant effects on harvest index of maize crop than recommended N and P. This result agrees with the ndings of (Sallah et al.,1998) who reported that harvest index of tef was found to be highest in blended fertilizer treatments. This report was also slightly similar with (Tekle and Wassie 2018) those reported the harvest index of maize was found to be signi cantly higher in plots that received blended fertilizers at rate of 150 kg NPSZnB ha − 1 as contrasted to the control treatment but, it was signi cantly in par with 300 kg NPSZnB ha − 1 and recommended NP fertilizers.

Physiological e ciency grain and apparent recovery of biological yield of maize
The highest mean of apparent recoveries of N and P recorded were 28.05% and 14.70%, respectively. The apparent N recovery decreased with increasing rate of blended fertilizer application (Table 3), however P recovery decreased with increasing rate of blended fertilizers were inconsistence. The maximum (28.05%) and minimum (7.27%) apparent recoveries of N were obtained at 100 kg NPSB + 73.9 N kg ha − 1 and 100 kg TSP & 200 kg Urea ha − 1 , respectively. There was a decrease in the apparent recovery of fertilizer N at each successive increment of fertilizer so that the highest recovery always occurred at lowest increment of fertilizer (Doyle et al., 1993). Similarly the maximum (14.70%) and minimum (0.92%) apparent recoveries of P were obtained at 100 kg NPSZnB + 75.1 N kg ha − 1 and 100 kg TSP & 200 kg Urea ha − 1 , respectively ( Table 4). The blended fertilizer had improved the N and P recovery over recommended N and P might be the contribution of macronutrient (S) and micronutrient (B and Zn) present in blended fertilizer increased the availability of macro nutrients. The N and P apparent recovery is in line with the ndings of (Sandana, 2016) which indicate that the level and types of nutrient fertilization affects the nutrient availability in soil and at high contents of soil nutrients and their availability more nutrients might be taken up by plants. In general, fertilizer N recovery by rice is never too high due to various types of losses including denitri cation, volatilization and leaching losses (Brady & Weil, 2002). Where, AR = Apparent recovery; PE = physiological e ciency; AUE = Agronomic fertilizer use e ciency The small number of grain and straw nutrient concentration and up take of P might be due to the P xation by acidity, and Al toxicity of the experimental site. Low available P of these soils could be one of the major soil fertility limiting factors in the study area, which limit the nutrient use e ciency of experimental site as was also con rmed by soil characterization of the studied area. This result is line with (Kirsten, 2014), who found phosphorus availability to plants is determined by the chemical characteristics of the soil and the P fertilizer source.
The physiological e ciency of N and P were in uenced by the application of blended fertilizer rates, types and recommended N and P (  (1984), explained physiological e ciency of crop and they found high physiological e ciency on N usage cereal achieved when high portion of N taken up is used for grain formation. Physiological N use e ciency  or N use e ciency for grain production (Borrell et al., 1998) refers to the additional yield produced for each additional kg of fertilizer N uptake and is determined as the ratio of net grain yield produced due to the applied fertilizer to the net uptake from applied fertilizer N.
The highest (847.02 kg kg − 1 ) physiological e ciency of P at a blended fertilizer rate of 100 kg NPSB kg + 73.9 N kg ha − 1 , mean while the lowest value of (431.92 kg kg − 1 ) was obtained at 150 kg TSP & 300 kg Urea ha − 1 (Table 18). According to  the physiological e ciency values should commonly range between 30 to 60 kg kg − 1 . If the obtained results are above these common values, it could be concluded that the farm was under well managed system and the reverse is true, if the results obtained are below the common values. The physiological e ciency of the experimental site was above the common values for both N and P physiological e ciency. Generally the physiological e ciency of P was high as compared to N this might be due to relatively higher yield produced with low uptake of P as compared with N of uptake.
Agronomic fertilizer use e ciency of maize grain Agronomic fertilizer use e ciencies of maize were in uenced by blended fertilizer rates and recommended N and P ( Table 3).
The agronomic fertilizer use e ciency of maize was varied from 2.5 to 19.64 kg ha − 1 at harvest stage of maize. The highest agronomic fertilizer use e ciency (19.64 kg kg − 1 ) was obtained under application of 100 kg NPSB + 73.9 N, while minimum value of agronomic fertilizer use e ciency (2.5 kg kg − 1 ) was recorded from 100% recommended N and P. Therefore, it seems that recommended N and P could not be an adequate application level regarding nourishing of this hybrid maize, perhaps due to limitation in the numbers of essential nutrients applied.
Karim & Ramasamy, (2000) Suggested that higher fertilizer use e ciency which is always associated with low fertilizer rate, cultural practices meant for promoting integrated nutrient management will help to effect saving in the amount of fertilizer applied to the crops and there to improve fertilizer use e ciency. Agronomic fertilizer use e ciency of any nutrient can be increased by increasing plant uptake and use of nutrient and by decreasing nutrient losses from the soil-plant system. The blended fertilizer applied improved agronomic fertilizer use e ciency by 682.47% as compared to recommended N and P fertilizers. Mengel et al. (2006) agronomic fertilizer use e ciency value for a nutrient should not be less than 5. This result therefore shows that the rates of recommended N and P of studied ranged from 2.51 to 4.7 kg kg − 1 which is less than the minimum standard AE according to (Mengel et al., 2006). Values of AE were lower than 5 for recommended N and P that may be due to nutrient imbalance of recommended N and P and this indicates that higher rate of N and P were not well utilized though a limiting nutrient. On the other hand, the agronomic e ciency for blended fertilizer types and rates of studied area were within the optimum rage (12.54 to 19.64 kg kg − 1 ). This result is similar with  who reported that agronomic fertilizer use e ciency should be within the ranges of 10 to 30 kg kg − 1 .

Partial Budget Analysis
The net bene t curve is allows to mark out an e cient set of technologies for recommendation. ETB net bene t respectively. The lowest net bene t was obtained by the application of the 100% recommended N and P with total of 14,891 ETB followed by 125% recommended nitrogen and phosphorous and control with net bene t of 15,528.5 and 16,080 ETB the respectively. Furthermore 150% recommended nitrogen and phosphorous also had lower net bene t of 16,595.5ETB. The increased production of the crop due to the application of inputs might or might not be bene ciary to farmers (CIMMYT, 1988). Therefore, partial budget analysis (CIMMYT, 1988) was employed to estimate the net bene t, dominance analysis and marginal rate of return that could be obtained from various alternative treatments (CIMMYT, 1988). The pro tability of the study showed that application of 150 kg NPSB + 110.8N kg ha − 1 and 150 kg NPSZnB + 112.6 N kg ha − 1 which provided the relatively high net bene t (32,321.4 and 30,478.1ETB) respectively, was the peak to apply fertilizers. The total costs that vary increased over the optimum level, the net bene t obtained reduced as the result of higher variable costs associated with lower earnings. Grain &straw yield) = price /kg* yield in kg and Net return = gross return -Total cost, VC = variable cost, GR = growth return, TGR = total growth return from straw and grain, NB = net bene t.

Dominance analysis
The highest net bene ts from the application of inputs for the production of the crop might not be su cient for the farmers to accept as good practices. In most cases, farmers prefer the highest pro t (with low cost and high income). For this purpose it is necessary to conduct dominated treatment analysis (CIMMYT, 1988). The % MRR between any pair of undominated treatments denotes the return per unit of investment in fertilizer expressed as a percentage. A dominated treatment is any treatment that has net bene ts that are less than those of a treatment with lower costs that vary (Stephen and Nicky, 2007). Where VC = variable cost, NB = net bene t, MRR% = marginal rate of return, D = dominated, B:C ratio = bene t cost ratio.
The dominance analysis showed that the net bene t of all treatments were dominated except application of 100 kg NPSB + 73.9 N kg ha − 1 and 150 kg NPSB + 110.8N kg ha − 1 ( Table 5). This result indicated that the net bene t was decreased as the total cost that varies increased beyond undominated fertilizer treatments application. Therefore, no farmer may choose other dominated treatments in comparison with the undominated treatments. This also helps to avoid the dominated treatment in further estimate of marginal rates of return.

Marginal rate of return
Economic analysis revealed that maximum marginal rate of return was recorded with application of 150 kg NPSB + 110.8 N kg ha − 1 (5329.4%), followed by 100 kg NPSB + 73.9 N kg ha − 1 (436.0%). The marginal rates of those treatments were well above the 100% minimum (CIMMYT, 1988). According to (CIMMYT, 1988) experience and empirical evidence, for the majority of situations indicated that the minimum rate of return acceptable to farmers would be between 50 and 100%. In the present study the treatments that had above 100% marginal rate return was recommended for the farmers, with treatments that had small number of variable cost. This treatment was 150 kg NPSB + 110.8 N kg ha − 1 . The % MRR between any pair of undominated treatments denotes the return per unit of investment in fertilizer expressed as a percentage. The results of undominated treatments indicated that for each one birr invested in purchase or production of fertilizers that was possible to recover one birr plus an extra of 4.36 birr ha − 1 and 53.29 birr ha − 1 as the fertilizer application changed from unfertilized plot to 100 kg NPSB + 73.9 N kg ha − 1 and 150 kg NPSB + 110.8N kg ha − 1 respectively. Passing from the rst treatment that had the lowest costs that vary to the end treatment which had the highest cost that vary, the marginal rate of return obtained was above the minimum acceptable marginal rate of return.
In this study, 100% was considered as minimum acceptable rate of return for farmers recommendation. Accordingly, the study revealed that application of 150 kg NPSB + 110.8N kg ha − 1 was considered as the best for recommendation. The best recommendation for treatments subjected to marginal rate of return is not necessarily based on the highest marginal rate of return, rather based on the minimum acceptable marginal rate of return and the treatment with the highest net bene t, relatively low variable cost together with an acceptable MRR becomes the tentative recommendation (CIMMYT, 1988).

Conclusion
The fertilizer rates and types on grain yield and straw yield highly signi cant difference (P ≤ 0.01), however there was no signi cant differences between the two blended fertilizer types. Maximum grain yield 7056.2 kg ha − 1 was recorded with T10 (200 kg NPSZnB + 150.2 N), while minimum grain yield 2996.0 kg ha − 1 was recorded from control treatment. This maximum grain yield was followed by T6, T7 and T9 with corresponding grain yield of 6863.4, 6563.8 and 6538.7 kg ha − 1 respectively, where these treatments were statistically at par with each other. The maximum maize straw was recorded with T6 (7886.7 kg ha − 1 ), while minimum value (4400.9 kg ha − 1 ) was recorded with control treatment. Accordingly, the study revealed that application of 150 kg NPSB + 110.8N kg ha − 1 and 150 kg NPSZnB + 112.6 N kg ha − 1 as the best rates recommended for maize production at Assosa area.
Blended fertilizer had improved nutrient concentration, uptake, agronomic e ciency, physiological e ciency and apparent recovery of maize as compared to recommended N and P. The improvements of uptake and nutrient use e ciency of maize by blended fertilizer might be due to the contribution of macro and micro nutrients present in blended fertilizer. Improving nutrient e ciency is an appropriate goal for all involved in agriculture, and the fertilizer industry, with the help of agronomic studies at different agro-ecologies. However, effectiveness cannot be sacri ced for the sake of e ciency. Much higher nutrient e ciencies could be achieved simply by sacri cing yield, but that would not be economically effective or viable for the farmer, or the environment. The pro tability of the study showed that application of 150 kg NPSB kg + 110.8N and 150 kg NPSZnB + 112.6 N kg which provided relatively high net bene t (32321.4 and 30,478.1ETB) was the best rates to apply. Marginal rate of analysis from undominated treatments indicated that for each one birr invested in purchase or production of fertilizers that was possible to recover one birr plus an extra of 4.36 birr ha − 1 and 53.29 birr ha − 1 as the fertilizer application changed from unfertilized plot to 100 kg NPSB + 73.9 N kg ha − 1 and 150 kg NPSB + 110.8N kg ha − 1 respectively.
The best recommendation for treatments subjected to marginal rate of return is not necessarily based on the highest marginal rate of return, rather based on the minimum acceptable marginal rate of return and the treatment with the high net bene t, relatively low variable cost together with an acceptable MRR becomes the tentative recommendation. Therefore we recommend the treatments (150 kg NPSB + 110.8N kg ha − 1 ) that have high marginal rate of return, high net bene t and relatively small total cost of production for maize production in Asossa zone. But based on yield data, net bene t and relatively small total cost of production the farmer of Asossa district can also use 150 kg NPSZnB + 112.6 N when NPSB formula were not available on the market.

Declarations
Funding Ethiopia Institutes of Agricultural Research had funded only the study design, data collection and analysis budget. That means the do no supporting the charge of the publication of the manuscript.

Availability of data and materials
Supporting data for the results have reported in this article can be found in progress and annual research review reports in Ethiopian Institute of Agricultural Research.
Ethics approval and consent to participate  Net bene t curve of maize as in uenced by blended fertilizer types and rates in Asossa district