Maize nitrogen fertilization in two crop rotation systems under no-till 1

ABSTRACT


INTRODUCTION
Maize (Zea mays L.) is one of the main cereal crops grown in Brazil and worldwide (CONAB, 2012).The world maize production reached 905 million tons and is destined mainly for animal feed and ethanol production.In the United States of America (US), the annual production reaches 336 million tons (USDA, 2012), while Brazil produces about 68 million tons (CONAB, 2012).
In the 2011/2012 crop year, the mean grain yield of the maize crop in Brazil was around 4400 kg ha -1 (CONAB, 2012), which is considered low when compared with the US yield of 9240 kg ha -1 (USDA, 2012).Among the factors accounting for the high yield of the maize crop in the US is the expressive use of nitrogen fertilizers.The Brazilian low yield is related to soil fertility, climatic conditions (Chioderoli et al., 2012) and insufficient nitrogen (N) fertilization (Silva et al., 2006).Therefore, the rational and balanced use of fertilizers, which represent a large part of the crop establishment costs, has been gaining an increasing importance for financial gain (Amado et al., 2002).
The split and time of nitrogen application are alternatives to increase the efficiency of fertilizers and minimize losses, allowing for synchronization between the applications and the period of high nutrient demand (Yamada et al., 2006).Other factors that influence the recommended rate and split of fertilization are the chemical attributes, such as the content of soil organic matter (SOM) and the weather, primarily the rainfall (Chioderoli et al., 2012).
The no-till (NT) system with crop rotation has provided maize yields higher than other cropping systems.In this system, there is an increased amount of potentially mineralizable nitrogen (PMN) in the soil, because of the residues remaining on the ground (Silva et al., 2005), especially in areas where the use of this system is already consolidated (Gomes et al., 2007).However, this will depend on the degree of SOM decomposition, the time of NT adoption (adoption phase, transition, consolidation or of maintenance) (Anghinoni, 2007) and quality of straw decomposition (Cabezas et al., 2005).
The contribution of cover crops to N supply and the crop rotation of maize plus soybean are important practices that contribute to the increase of N content in the soil (Jantalia et al., 2006).Cover crops can meet the demands for N of maize by using rotation with legume crops.However, when using species with high C/N ratio, such as grasses, cereal yield can be reduced if there is insufficient N complementation, via fertilization (Lara- Cabezas et al., 2004).
For these reasons, in order to evaluate nitrogen management, this study evaluated the split N fertilization and the influence of the crop rotation system on agronomic characteristics, grain yield and N, P and K contents in the leaf tissue of maize.

MATERIAL AND METHODS
The field experiment was conducted in no-till system, in place for six years (transition phase), in the municipality of Marechal Cândido Rondon, western Paraná, with geographic coordinates 24º 42' S and 54º 14', approximately 220 m altitude, average annual rainfall of 1642 mm, average temperature of 21.3 ºC and average relative humidity of 76.1% (IAPAR, 2011).The soil of the experimental area is classified as typical Eutroferric Red Latosol (LVef) (Embrapa, 2006b).
The experiment consisted of 36 treatments arranged in a 6 x 3 x 2 factorial randomized block design, with four replications.We evaluated the combination of six N rates (0, 20, 30, 40, 50 and 60 ha -1 ) at sowing, three N rates (40, 70 and 100 kg ha -1 ) in topdressing and two rotation systems in no-till.The rotation system at Site 1 consisted of five cropping sequences of maize and, at Site 2, of crop rotation (corn + soybean + oats + soybean + corn), totaling 144 experimental units.Before the establishment of the experiment, soil acidity was corrected by lime (TNP -total neutralizing power -85%) distribution on the surface of the two sites, without incorporation, to increase the saturation to 70%.
The experiment was conducted in the crop year 2006/ 2007 using maize seeds of the single early hybrid Dow AgroSciences 2B710.Furrows were opened for N fertilizer placement at sowing.Phosphorus (P) and potassium (K) fertilization rates were 75 and 50 kg ha -1 of P2O5 and K2O, respectively, in both sites using triple superphosphate (44% P 2 O 5 ) and potassium chloride (60% K 2 O).
Urea (45% N) was used as N source for band fertilization at sowing and topdressing.N topdressing was carried out when the maize plants had five expanded leaves (V5), by broadcasting over the entire area of the plot.Each experimental unit consisted of five rows (6 m long) spaced 0.70 m apart.Three central rows of each plot were harvested for evaluations, eliminating 0.5 m from each end.
The following variables were evaluated: grain yield, 1000-grain mass, N content in the grain, N, P and K in leaf tissue, plant height, height of ear insertion, stem diameter, row number per ear, grain number per row, grain number per ear, ear length and relative price between product (grain) and fertilizer.To obtain this ratio, we used the maize closing price on the Chicago Board of Trade (CBOT) of March 2013 for the region, which was R$ 24.50 per 60-kg bag (CBOT, 2013), and R$ 1.18 per kg of N for urea.The ratio was calculated using the following equation: RPR = [(PT x R$ kg) / (kg urea-N x R$ kg urea-N)], where: RPR = relative price in Real; PT = treatment production (kg); R$kg = Real kg -1 maize; kg urea-N = N amount used (kg); R$ kg N-urea = Real kg -1 N-urea.The increase in production of each treatment was calculated considering the production obtained with the lowest N rate, as follows: {[production per treatment / production obtained with the lowest N rate] x 100}.
Analysis of variance and regression analysis were performed using the software SAEG 8.0 (SAEG, 1999).Treatments with the same N rate were compared by contrasts using the Bonferroni test, at 1 and 5% significance.

RESULTS AND DISCUSSION
Maize yield was significantly different between the sites with a positive interaction between N band fertilization at sowing and topdressing (Table 1).The area with crop rotation (Site 2) had higher mean yield, greater ear length and higher 1000-grain mass than the area with successive maize (Site 1).There was significant effect of N fertilization at sowing for almost all variables, except for height of ear insertion and row number per ear.
The interaction between N rate at sowing and N rate in topdressing (with 40 and 70 kg ha -1 N) showed an increase in yield of 24.82 and 26.76 kg ha -1 grain, respectively, for each kg of N applied at planting (Figure 1a).The rates 40 and 70 kg ha -1 in topdressing with N at sowing above 54 and 39 kg ha -1 , respectively, resulted in production larger than the mean production obtained with 100 kg ha -1 in topdressing, which was 8864 kg ha -1 (Figure 1a).
The increase in production with the use of higher N rates at sowing, in comparison with those commonly used (20 to 30 kg N ha -1 ), indicate less need for N in topdressing, allowing high yields (Embrapa , 2006a).However, care must be taken with N rates above 60 kg ha -1 at sowing, which may favor salinity and/or alkalinization of the rhizosphere, damaging the plant root system and reducing the rate of nutrient absorption (Fancelli & Dourado Neto, 2004).There were differences between N rates at sowing for plant height, but not for height of ear insertion, for any of the rates evaluated at sowing and topdressing (Table 1).Lana et al. (2009) observed increased plant height and height of ear insertion, 2.1cm and 1.8 cm, respectively, per 30 kg of N added.
Figure 1b shows the mean yield of the two sites, with total rates of N, independent of the application form.It was possible to fit a quadratic equation with the maximum point of 132.4 kg N ha -1 and grain yield of 9038 kg ha -1 , which was also reported by Gomes et al. (2007), with a maximum point of 151 kg ha -1 N.
This total N rate is lower than that recommended by Embrapa (2006a), which for an expected yield above 8000 kg ha -1 , recommends 20 to 30 kg ha -1 at sowing and 140 kg ha -1 in topdressing.This may be explained by the significant amount of crop residues in both areas, which may be releasing N through SOM mineralization, especially in the system of rotation with soybean.Araújo et al. (2004) have also reported linear responses of increased yield as a function of maize nitrogen fertilization in an Oxisol and Costa et al. (2012) in an intercropping with brachiaria.
For most N rates evaluated, with exception of the treatments 40 kg ha -1 N at sowing and 100 kg ha -1 N in topdressing, the site with crop rotation had higher yields than the site with successive maize crops.The differences between the systems ranged from 38 kg ha -1 to 1468 kg ha -1 of grain produced in the treatments with 50 kg ha -1 N at sowing and 70 kg ha -1 N in topdressing and 60 kg ha -1 N at sowing and 70 kg ha -1 N in topdressing, respectively (Table 2).
The importance of the N fertilization at sowing for corn yield is confirmed when comparing the results of the application of 30 kg N ha -1 at sowing plus 40 kg ha -1 N in topdressing with the application of 70 kg N ha -1 only in topdressing (Table 2).In this case, there is an increase in yield of 590 and 428 kg ha -1 for the sites 1 and 2, respectively.The same happens with the rates 50-70 kg ha -1 compared with 20-100 kg N ha -1 , and 50-40 kg ha -1 compared with 20-70 kg ha -1 N at sowing and in topdressing, respectively.These results may be due to the greater efficiency of N utilization, with small losses by volatilization compared to losses with surface applications at higher rates (Yamada et al., 2006).Silva et al. (2006) obtained better results with up to half of the rates applied at sowing (60 kg ha -1 ) and the remaining in topdressing, from stages V4 to V6. Cabezas et al. (2005) agree with the rates 30 to 40 kg ha -1 of N at sowing in the early years of NT adoption, reducing and/ or eliminating the initial N deficiency, because of the immobilization caused by the decomposition of antecedent crop residues.The highest yield among treatments was recorded for the rates 60 kg ha -1 at sowing and 70 kg ha -1 in topdressing for site 2, reaching 9970 kg ha -1 (Table 2).Araújo et al. (2004) reported yield of 11,203 kg ha -1 by applying 240 kg N ha -1 in topdressing split four times, at the stages V4, V8, V12 and bolting.
Maize nitrogen fertilization in two crop rotation systems under no-till Rev. Ceres, Viçosa, v. 60, n.6, p. 852-862, nov/dez, 2013 Table 1.Summary of analysis of variance and mean comparisons for yield, N content in leaf and grain, leaf P, leaf K, height of ear insertion, plant height, stem diameter, ear length, row number per ear, grain number per row, grain number per ear and 1000 grain mass in response to nitrogen fertilization at sowing and topdressing in two sites The mean yield of Site 2 was 577 kg ha -1 higher than the mean yield of Site 1, i.e., increasing from 8388 kg ha -1 to 8965 kg ha -1 grain (Table 2).There were yield increases in relation to the lowest N rate, ranging from 6.8% for the treatment 0 kg ha -1 at sowing and 70 kg ha -1 in topdressing to 25.8% for the treatment 60 kg ha -1 at sowing and 70 kg ha -1 in topdressing.Similar increases in production were also reported by Araújo et al. (2004) in monoculture systems (maize+ maize + maize and maize + soybean + maize).

Analysis of variance
The economic analysis also has shown that grain yield per kilogram of applied N was satisfactory, providing a good return on capital invested in the form of nitrogen fertilizer.Investment in nitrogen fertilizer was positive for all treatments, with average return of R$ 30.88 and R$ 33.21 for every R$ 1.00 invested in nitrogen fertilizer, for sites 1 and 2, respectively, representing an average gain of R$ 32.04 in grains per Real of N invested (Table 2).The rate 70 kg ha -1 in topdressing combined with 30 and 50 kg ha -1 at sowing also provided larger increases in production.Farinelli & Lemos (2012) reported that the nitrogen fertilization provided significant increases in agronomic and nutritional traits of maize, with the maximum grain yield obtained with 151 kg ha -1 N.
Table 3 shows the comparisons between the rates of N in topdressing (C1, C2 and C3) and comparisons involving the split of nitrogen fertilization at the rates 70 (C4), 90 (C5), 100 (C6, C7 and C8), 120 (C9) and 130 (C10) ha -1 of total applied N. Maize yield was higher for the topdressing N rates 70 and 100 kg ha -1 N than the yield obtained with 40 kg ha -1 N.However, the yields at 70 and 100 kg ha -1 were not significantly different; therefore the best topdressing rate was 70 kg ha -1 N.
The fourth contrast shows that there were differences by splitting the rate 70 kg N ha -1 , indicating that splitting the rate into 30 kg ha -1 N at sowing and 40 kg ha -1 N in topdressing, the maize yield increased in 509 kg ha -1 (Table 3).The same trend was observed for the rate 90 kg ha -1 , with significant difference for the splitting form in contrast 5, showing that the increase in the N rate at sowing to 50 kg ha -1 resulted in an increase of 426 kg ha -1 of grains compared with that obtained by the addition of 20 kg ha -1 of N at sowing .The contrasts 6, 7 and 8 provided comparisons between the splitting forms of the rate 100 kg ha -1 N. Estimates of the contrasts 6 and 7 emphasize again the importance of N fertilization at sowing, because when the fertilizer was applied in topdressing only, there was a   Site 1: area with five cropping sequences of maize; Site 2: area with crop rotation (maize+soybean+oat+soybean+maize). 1 Maize price, quotation from January 2013 = R$ 24.50 bag; Urea price (kg ha -1 ) = R$ 1.18 kg.reduction of 605 and 607 kg ha -1 in the maize production, respectively, i.e., a significant reduction in yield.

Treatment
The contrast established for maize production with the rate 120 kg N ha -1 (C9), indicated that the best splitting form was the application of 50 kg ha -1 of N at sowing and 70 kg ha -1 N in topdressing.There was an increase of 300 kg ha -1 in grain production compared with that obtained from the management commonly used in various maize production regions, which is 20 kg ha -1 N at sowing and 100 kg ha -1 N in topdressing.No yield differences were found for the split of the rate 130 kg N ha -1 (C10).Fancelli & Dourado Neto (2002) stated that a high input of N at sowing provides larger number of grains per ear and allow greater flexibility in the period of topdress application of nitrogen, which can be extended to the stages V7 and V8.
Site 1 had, on average, larger 1000-grain mass than Site 2, which reflected in the yield (Figure 2a).There was an increase of 0.045 g per kg ha -1 of N applied at sowing (linear increase) in the Site 1, but no differences were found for the rates of N applied at sowing in Site 2. This trend of grain mass increase was also reported by Lana et al. (2009), but different from that reported by Gomes et al. (2007).
The conflicting results in the literature show that the grain mass is a trait not only influenced by nutrient availability, but is also influenced by genotype and climatic conditions of the region.This behavior is probably a result of soil management, which in case of Site 2, there may be a better balance between mineralization and immobilization of N. Thus, over the years, when the no-till system is consolidated, it will be possible to reduce fertilizer rates (Anghinoni, 2007), as the effect of increased biomass production and, consequently, the increase of carbon stocks and N in the soil.
There was an increase in N content of grains with increasing N rates at sowing for both sites.In Site 1, the N content increased from the rate 15 kg ha -1 N at sowing   2b), a trend also observed by Fornasieri Filho (2007).This increase may be related to the higher content of P in the soil of this site and the nutritional balance of the plants, which emphasizes the importance of good management of soil fertility for maize cultivation.The presence of adequate levels of P stimulates N uptake, in the same way that N has a positive effect on the absorption of P by stimulating root growth at the site where N is in the highest concentration (Novais et al., 2007).The agronomic traits such as plant height, ear length, grain number per row and grain number per ear tended to increase as a function of nitrogen fertilization at sowing (Figure 3), which was also reported by Fancelli & Dourado Neto (2002).Difference for plant height occurred only as a function of N rates at sowing (Table 1), with increase in height up to the rate 45.2 kg N ha -1 (Figure 3a).Increase in plant height was also observed by Lana et al. (2009), with increasing applications of N. The largest grain number per row as a function of N rates at sowing was obtained with the rate 38 kg ha -1 N.
There was difference for the row number per ear between the sites, with Site 1 having a higher mean (Table 1).This fact can be explained by the higher contents of P and K in the soil of Site 1.There was no effect of N at sowing for this variable.However, even with the Site 1 having larger row number per ear, there was no difference in the grain number per ear between the sites (Figure 3c) and this is probably due to the smaller ear length of Site 1 (Figure 3d).Costa et al. (2012) found that in the intercropping (maize + brachiaria), the row number, grain number per row and grain number per ear were directly associated and had linear increase with topdress N rates, showing the key role of this nutrient for the maize grain formation.The same fact was reported by Lana et al. (2009), emphasizing the important role of N for the agronomic performance of crops and production components.
There was a significant interaction between N rates at sowing and sites for ear length (Table 2).At Site 1, a linear increase of 0.027 cm was recorded for each kg ha -1 of N applied at sowing (Figure 3d).However, no difference was found for Site 2 as a function of N rates at sowing, with a mean of 16.53 cm for ear length.
The leaf N content increased from the N rate 16.5 kg ha -1 at sowing (Figure 4b).However, even with the highest rate of N at sowing (60 kg ha -1 N) the leaf N content was below the adequate levels recommended by Oliveira (2004).This low leaf N may be due to the dilution effect in the plant, since there was increase in plant height with rates up to 45.2 kg ha - 1 N at sowing (Figure 3a).This result differs from that obtained by Costa et al. (2012), who found that the leaf N content increased linearly with increasing levels of N.
The leaf K + content decreased with N rates at sowing up to 41.8 kg ha -1 N, but increased from that rate (Figure 4c), i.e., without N fertilization at sowing, K + concentration was higher than that found with increasing N rates at sowing, probably because of the lower plant growth, which resulted in higher K + concentrations.
For leaf P content, there was difference between the sites and significant interaction between N rates at sowing and topdressing (Table 1).Site 1 had mean P content greater than Site 2, possibly due to the higher initial concentration of P in the soil.However, the contents were within the range considered adequate for the crop (Oliveira 2004).
In the interaction of leaf P content with the N rates, at sowing and in topdressing, there was a tendency to equate the P contents at the highest N rate at sowing (60 kg ha -1 ).These results show a positive interaction between N and P, that is, increased N at sowing resulted in increased P absorption, especially at the rate 40 kg ha -1 N in topdressing, even for soils with high P available, in which there is usually little response to fertilization (Figure 4a).

CONCLUSIONS
Split nitrogen applications, with the highest rate applied at sowing, for the same final amount of fertilizer, resulted in higher grain yield.
The rate 70 kg ha -1 nitrogen in topdressing provided the highest yield the lowest cost, compared with the yield obtained with the rates 40 and 100 kg ha -1 .
The crop rotation (maize + soybean + oat + soybean + maize) provided significant increase (7%) in maize yield compared with the sequential maize cultivation.
Increased nitrogen at sowing increased the 1000 grain mass, plant height, ear length, grain number per row and grain number per ear.

Figure 1 .
Figure 1.Mean yield of maize from two sites as a function of N rate at sowing (A) and total N rate applied at sowing and topdressing (B).

Figure 2 .
Figure 2. Thousand grain mass (A) and N content in maize grain (B) in response to nitrogen fertilization at sowing for areas without (Site 1) and with crop rotation (Site 2).

Figure 3 .
Figure 3. Plant height (A), grain number per row (B), grain number per ear (C) and ear length for areas without (Site 1) and with crop rotation (Site 2) (d) in response to nitrogen fertilization at sowing.

Figure 4 .
Figure 4. Leaf phosphorus content as a function of N applications at sowing and topdressing (A), leaf nitrogen content (B) and leaf potassium content (C) in response to nitrogen at sowing.

Table 3 .
Contrast estimates and significance tests for maize yield as a function of nitrogen fertilization at sowing (S) and topdressing (C)

Table 2 .
N rates at sowing (S) and topdressing (C) and respective maize yields, relative price between maize production and nitrogen fertilizer and the average increase in yield in relation to the lowest fertilizer rate