UPLAND RICE UNDER NO-TILLAGE PRECEDED BY CROPS FOR SOIL COVER AND NITROGEN FERTILIZATION ( 1 )

The grain yield of upland rice under no-tillage has been unsatisfactory and one reason could be the nitrate/ammonium balance in the soil. Cover crops and nitrogen fertilization can be used to change the nitrate/ammonium relation in the soil and improve conditions for the development of upland rice in the no-tillage (NT) system. The aim was to study the effect of cover crops and nitrogen sources on grain yield of upland rice under no tillage. The study was carried out on the Fazenda Experimental Lageado, in Botucatu, State of São Paulo, Brazil, in an Oxisol area under no-tillage for six years. The experiment was arranged in a randomized block split-plot design with four replications. The plots consisted of six cover crop species (Brachiaria brizantha, B. decumbens, B. humidicola, B. ruziziensis, Pennisetum americanum, and Crotalaria spectabilis) and the split-plots of seven forms of N fertilizer management. Millet is the best cover crop to precede upland rice under NT. The best form of N application, as nitrate, is in split rates or total rate at topdressing or an ammonium source with or without a nitrification inhibitor, in split doses. When the cover crops C. spectabilis, B. brizantha, B. decumbens, B. humidicola, and B. ruziziensis preceded rice, they induced the highest grain yield when rice was fertilized with N as ammonium sulfate source + nitrification inhibitor in split rates or total dose at topdressing.


INTRODUCTION
Cover crops are plant species used to cover and protect the soil against erosion and nutrient loss through leaching or runoff (Dabney et al., 2001).They also contribute to carbon sequestration and can improve soil quality and increase the soil organic matter, nutrient cycling and cation exchange capacity (Aranda et al., 2011).Cover crops are very important in the no-tillage (NT) system, a technique used on about 25 million hectares of agricultural production in Brazil (Febrapdp, 2012) and on around 120 million hectares worldwide (FAO, 2006).
In this context, we address rice, which is a staple grain in the diet of half the world's population, produced in both lowland and upland ecosystems (Santos et al., 2006).Upland rice, particularly, is cultivated in Asia, Africa and the Americas by mostly small or subsistence farmers in the poorest regions of the world (Oonyu, 2011).Its importance is increasing, since the water availability for rice irrigation by the flooding method is decreasing in Asia, mainly in China and India, due to the rapid growth of industry and urban centers (Feng et al., 2007).
However, producing upland rice under NT still requires improvements, in view of the low yields in this system (Kluthcouski et al., 2000).One reason for this low grain yield would be the low activity of the enzyme nitrate reductase (NR), inhibiting N nutrition of the rice seedlings in environments with high nitrate concentrations (Malavolta, 1980).Experiments performed by Sá (1999) andD'Andréa et al. (2004) showed that in a NT, the availability of N-NO 3 -in the soil is higher than under conventional tillage, due to the fact that the levels of soil moisture, nutrients and organic matter are higher under no-tillage, stimulating the microbial activity, especially nitrifying bacteria.Additionally, Holzschuh et al. (2009) reported that upland rice developed better when there were similar amounts of ammonium and nitrate in the early crop stages.
In this regard, according to Pacheco et al. (2011) and Nascente et al. (2013a), the evaluation of cover crops can be useful to improve conditions for NT upland rice.Cover crops can increase the nutrient availability (due to straw decomposition), reduce the possible release of allelopathic substances into the soil and decrease the rate of N immobilization by the soil microbial community (Nascente et al., 2013b).
Therefore, as the low performance of NT upland rice may be due to the predominance of nitrate in the soil, practices that limit nitrification temporarily and/ or increase the amount of ammonium in the soil could be a solution to improve the crop development.One of these practices could be crop rotation of cover crops preceding upland rice (Pacheco et al., 2011;Nascente et al., 2013a).
Another possibility would be the use of N sources such as ammonium at rice sowing and/or shortly after emergence.However, as the transformation from N-NH 4 + to N-NO 3 -in the soil occurs rapidly (Malavolta, 1980;Crusciol et al., 2011), some steps would be required, especially the use of nitrification inhibitors together with ammonium sources and the cultivation of rice after cover crop with potential for nitrification inhibition.
In the light of this information, the hypothesis of this study was: the low performance of upland rice in NT may be due to the predominance of nitrate in the soil; cover crops can inhibit nitrification and consequently balance the NO 3 -/NH 4 + ratio in the soil, improving the development of upland rice, raising yields to viable economic levels; and when growing upland rice in NT, the probability of success is greater with ammonium as N source.Therefore, the aim was to study the effect of the interaction of cover crops and N sources on the grain yield of upland rice under no tillage.

MATERIAL AND METHODS
The trial was performed in the 2009/2010 growing season on the Fazenda Experimental Lageado of the College of Agricultural Sciences/UNESP, Botucatu, State of São Paulo, Brazil (22 o 51' S latitude, 48 o 26' W longitude, 740 m asl).According to the Köeppen classification, the prevailing climate in the region is Cwa, high altitude tropical, with dry winters and hot and rainy summers.During the trial, additional temperature and rainfall data were collected daily (Figure 1).
The soil was an Oxisol (Embrapa, 2006) managed under no tillage for six years.The crop sequence in this pre-experimental period was as follows: soybean/ oat; maize/Brachiaria brizantha; soybean/oat; common bean/oat; soybean/fallow and corn in the summer/ beginning of the experiment.
The experiment was arranged in randomized blocks, in a split plot design with four replications.The plots consisted of six species of cover crops (Brachiaria brizantha, B. decumbens, B. humidicola, B. ruziziensis, Pennisetum americanum, and Crotalaria spectabilis) and split plots (experimental unit) and seven forms of N fertilizer management, applied 0 and 30 days after emergence (DAE) of the rice seedlings [Control: no N fertilization; Calcium nitrate, NO (40+40 kg ha -1 ); NO (0+80 kg ha -1 ); Sulfate ammonium, NH (40+40 kg ha -1 ); NH (0+80 kg ha -1 ); Ammonium sulfate + dicyandiamide (DCD), NHI (40+40 kg ha -1 ); and NHI (0+80 kg h -1 )].Brachiaria species were used in view of their ability to inhibit nitrification.Millet was used because it is the species most commonly used as straw source in the Cerrado, and crotalaria is being increasingly used, mainly due to its ability to reduce nematodes.
Spontaneous vegetation was desiccated with Glyphosate (2,000 g ha -1 active ingredient, acid equivalent) and cover crops were sown with a planter tractor implement (model Personale DRILL-13, Semeato).The seeding rate and row spacing were according to the recommendations for each species.The cover crops were sown on April 2009 and desiccated in October 2009 (150 DAE) with Glyphosate (2,000 g ha -1 acid equivalent).Prior to fertilization, the cover crop dry matter was determined, as well as the accumulated amount of N and concentration of N-NH 4 + and NO 3 -in the straw (Table 1).
After 30 days, the concentration of N-NO 3 -, N-NH 4 + and total N in the soil was determined according to Silva (1999), in the 0-5 and 5-20 cm layers (Table 2), and rice cultivar IAC 202 was sown, on December 10, 2009, in rows spaced 0.34 m apart, at a density of 80 seeds per meter, with the same planter model for NT systems, in experimental units of 7 x 4.5 m.The fertilizer was applied at sowing according to the method proposed by Raij et al. (1997).
The eight central rows of each experimental unit were considered, disregarding 1.0 m at either end of each plant row for data collection.In the rice plants, N concentrations (ammonium, nitrate, total) sampled 42 DAE were evaluated according to Silva (1999).Rice leaves were sampled in the morning between 8:00 and 10:00 am.The number of panicles per square meter was obtained by counting panicles contained in 4 m in the assessed area of each split plot.From this result, the number of panicles m -2 was calculated.The total number of spikelets per panicle was obtained from 15 panicles per split plot.The spikelet fertility was determined by the ratio of the number of filled spikelets by the total number of spikelets per panicle x 100.The 1,000-grain-weight was obtained by weighing four samples of 1,000 grains per split plot.The grain yield was obtained by sampling of panicles of the usable area of each experimental unit and then the material was threshed and cleaned.After this operation, the weight of grain harvested was determined and extrapolated to Mg ha -1 at 130 g kg -1 wet basis.The data were subjected to analysis of variance, and means were compared by the LSD test (p 0.05).

RESULTS AND DISCUSSIONS
A higher amount of biomass from the cover crops ranging from 7,030 to 10,660 was observed (Table 1).According to Pacheco et al. (2011) and Nascente et al. (2013a,b), these cover crops are well-known for producing high amount of biomass, which is very important for regions such as the Cerrado, where soil fertility is low.Cover crops during straw degradation could release nutrients to the soil and increase the soil fertility.Corroborating this information, it was stated that all cover crop species contained significant foliar amounts of nitrate and ammonium.As a result, increased soil concentrations of ammonium and nitrate were observed for all cover crops, higher than those measured in the beginning of the experiment (Table 2).Ammonium levels in rice plants were influenced by cover crops and N management (Table 3).The split application of N-NH 4 + with nitrification inhibitor (DCD) resulted in the highest values, differing from split application of N-NO 3 -, and from NH 4 + with or without protection (DCD) in a single application.Therefore, it appears that the split N application was more suitable for rice, and results were best when using the N source ammonium.When fertilizing with a nitrate source, split applications were not efficient, possibly due to lower the "ability" of upland rice to metabolize greater amounts of nitrate in the early developmental stages (Malavolta, 1980).The results of N fertilization in a single dose 30 DAE suggest that in this stage, rice plants have a greater ability to metabolize N in nitrate form.Regarding the ammonium source, it seems clear that split applications obtained better results because it is likely that in early developmental stages upland rice plants "preferred" this N source (Soares, 2004).
Of the cover crops, C. spectabilis induced the greatest ammonium accumulation in upland rice plants, differing from millet, B. decumbens and B. humidicola (Table 3).This can be explained by the fact that C. spectabilis is a legume, with normally high foliar N levels (Table 1).
Regarding nitrate, it was found that N application without splitting (80 kg ha -1 ), 30 DAE resulted in the highest concentration in rice shoots (Table 3), when preceded by millet and Crotalaria spectabilis.Cazetta et al. (2008), Pacheco et al. (2011) and Nascente et al. (2013b) observed that millet recycles significant N amounts and emphasize the importance of growing cover crops in NT areas.In a comparison of the Brachiaria species, the highest nitrate concentration in rice shoots was observed when grown after B. ruziziensis.
The total N concentration in upland rice plants was influenced by cover crops and N fertilization management (Table 3).Thus, it appears that the levels were highest after the single application of 80 kg ha -1 , regardless of the source (NO, NH and NHI).The single application (30 DAE) a few days before sampling (42 DAE) explains these results.In cover crops, C. spectabilis provides the highest value, differing from B. decumbens, B. humidicola and B. ruziziensis (Table 3).These results can be explained because C. spectabilis is a legume cover crop, and it would have accumulated 220 kg ha -1 N at the time of chemical cover crop management (Table 1).
It is noteworthy that the highest levels of total N in rice plants with an application of 80 kg ha -1 30 DAE, may be due to the fact that the experiment was Pennisetum americanum 10,660 11.4 122 0.30 1.00 Crotalaria spectabilis (1)  8 ) and nitrate (N-NO 3 -) concentration in cover crops before fertilization (1) Management practices were applied to Crotalaria spectabilis in the early stage of grain filling.carried out after six years of a NT system, with soil ammonium and nitrate levels of 6.0 and 15.7 mg dm -3 , respectively (0-20 cm layer), where growing cover crops had been groan with high levels of N (Tables 1 and 2).These high nutrient levels may have possibly nurtured the plant prior to fertilization.In this sense, the history of the area, cover crops used and N levels in the soil must be taken into consideration, for adequate recommendations of N fertilization applied to the soil and consequently to the plant.
The number of panicles was influenced by cover crops and N fertilization management (Table 4).All sources of N fertilization produced higher values than the control.The increase in the number of panicles m -2 due to N fertilization was observed by Cazetta et al. (2008) and Nascente et al. (2011).Regarding the cover crops, millet and C. spectabilis induced the highest number of panicles m -2 , possibly because they produced the highest N levels in rice plants (Table 3).Among Brachiaria species, the highest number was observed in plots with B. ruziziensis, especially when N was supplied as ammonium.
Regarding the number of spikelets per panicle, all sources of N fertilization outperformed the control, with a slight advantage for management NHI-00+80.B. decumbens produced the largest number of spikelets (Table 4).This result can be partly explained by the lower number of panicles m -2 .Thus, due to less competition for space, it is likely that the plants produced larger panicles and consequently a larger number of spikelets (Santos et al., 2006).Cazetta et al. (2008) observed little influence of the cover crops on the number of spikelets.
Spikelet fertility was little influenced by the N management and cover crops (Table 4).Arf et al. (2003) and Farinelli et al. (2004) also found no effect of N fertilization on spikelet fertility.However, Bordin et al. (2003) observed higher spikelet fertility in rice grown after wild bean (Canavalia brasiliensis), C. spectabilis and millet.
In terms of 1,000-grain-weight, results were best in the control and N fertilized shortly after germination (Table 4).Cazetta et al. (2008) reported that this parameter is influenced by the cultivar, N fertilization distribution and level of available water.Of the cover crops, millet produced the highest 1,000grain-weight.The number of rice panicles was highest under millet straw.At a high number of rice panicles, the tendency is to reduce the panicle size, resulting in fewer spikelets, allowing greater grain filling and, consequently, higher grain weight (Santos et al., 2006), as observed in this trial.Bordin et al. (2003) also found superiority of millet for this characteristic in different cover crops.
The rice grain yield was influenced by the preceding cover crops and N fertilization management (Table 4).The highest yield was obtained with split application of N as ammonium (40 kg ha -1 applied twice) with inhibitor, which differed significantly from the others, with the exception of nitrate application (40+40 kg ha -1 ).
This indicates the importance of N fertilization for rice, since the average yield in the treatment where N was omitted was the lowest (4.0 Mg ha -1 ), i.e., 1.4 Mg ha -1 lower than the highest yield (Table 4).These results confirm those reported by Bordin et al. (2003), Farinelli et al. (2004), Crusciol et al. (2011) and Nascente et al. (2011), of significant increases in rice yield by N fertilization.Regarding the optimal distribution of N fertilization, it was observed that the total application at topdressing was not the most effective, whereas the split application, which resulted in higher yields, was more effective.In addition, later fertilization at higher N concentrations can increase susceptibility to blast (Freitas et al., 2010).
Millet was the cover crop that produced the highest upland rice grain yield (6.8 Mg ha -1 ), which differed from the other cover crops (Table 5).In the split application of N as ammonium, the yield was even higher (7.6 Mg ha -1 ) and statistically different from all other cover crops.Our data were consistent with Crusciol et al. (2011), Pacheco et al. (2011) and Nascente et al. (2013a, b), who also reported a higher grain yield of upland rice mulched with millet straw.These authors attributed the results to a rapid degradation of millet residue intensifying nutrient cycling.Moreover, this rapid straw decomposition decreases the possibility of the release of allelopathic substances after desiccation (Souza et al., 2006).
Regarding the interaction between sources and distribution of N application x cover crops it was observed that millet preceding rice induced the highest grain yield, with all sources and distribution of N application (Table 4).For all cover crops, the best temporary distribution of N fertilization was half the dose (40 kg ha -1 ) at sowing and half the dose (40 kg ha -1 ) at topdressing, when using the N source NHIammonium sulfate + nitrification inhibitor and applying the entire N rate (80 kg ha -1 ) with the same source at topdressing, with the only exception millet.In this sense, it can be inferred that no-tillage cover crops were more limiting to the success of upland rice than the N source in all treatments, since, on average, the highest yield was similar for all cover crops, under both N sources (5.2 Mg ha -1 for nitrate and 5.1 Mg ha -1 for ammonium).The only exception was C. spectabilis, which produced the highest rice grain yield in the control treatment (without N).On the other hand, the cover crops provided significant differences ranging from 3.5 Mg ha -1 (after C. spectabilis) to 7.6 Mg ha -1 (after millet).Our data indicate that high grain yields of no-tillage upland rice can be obtained when millet is the preceding cover crop and N applied as nitrate in split rates or the total dose at topdressing or ammonium source in split rates or ammonium source + nitrification inhibitor in split rates.
The grain yield of rice is determined by four components: number of panicles m -2 , number of spikelets per panicle, spikelet fertility and 1,000-grain- weight (Santos et al., 2006).In this context, the superiority of millet observed in this study was due to the larger number of panicles per m 2 and higher 1,000grain-weight (Table 4).
It is noteworthy that with the exception of B. ruziziensis (6.0 Mg ha -1 ) all other Brachiaria resulted in lower rice grain yields, differing from B. ruziziensis and also millet (Table 4).These results suggest that there may be some allelopathic effect of these plants that prevent rice development.Souza Filho et al. (1997), Martins et al. (2006) andSouza et al. (2006) reported allelopathic effects caused by Brachiaria species.These authors assume that these effects may reduce the crop germination and plant development, affecting yields.However, although rice poduced low grain yield when preceded by certain cover crops, as corroborated by other authors (Kluthcouski et al., 2000), our results are interesting, once millet seems to increase the rice grain yield and could be an alternative as a cover crop preceding upland rice in the no-tillage system.CONCLUSIONS 1. Millet is the best-suited cover crop to precede upland rice under no-tillage.The best distribution of N application, as nitrate, is in split rates or total rate at topdressing or an ammonium source with or without a nitrification inhibitor, in split doses; 2. The cover crops Crotalaria spectabilis, Brachiaria brizantha, B. decumbens, B. humidicola and B. ruziziensis cultivated in intercropping or in crop rotation preceding upland rice produced highest rice grain yields when N was applied as ammonium sulfate (source) + nitrification inhibitor, in split rates or total dose at topdressing.

Table 2 .
Ammonium (NH 4 + ), nitrate (NO 3 -), and total nitrogen (TN) concentration in the two soil layers as affected by cover crops, evaluated immediately before upland rice sowing

Table 4 .
Yield components of upland rice as affected by cover crops and nitrogen sources under no-tillage