Comparison of nitrification inhibitors to restrict nitrate leaching in a maize crop irrigated under mediterranean conditions

The aim of this paper was to compare dicyandiamide (DCD) and 3,4 dimethylpyrazole phosphate (DMPP) as inhibitors of ammonium oxidation and nitrate leaching after applying fertilizer to a maize (Zea mays L.) crop grown under Mediterranean conditions. The effects of nitrification inhibitors were also compared to those of N fertilization without inhibitors and with split N application. In plots fertilized with ammonium sulphate nitrate (ASN), either DCD or DMPP lengthened ammonium presence in soil and produced lower soil NO3 concentrations (30% lower than in plots with no inhibitor). The use of DCD or DMPP achieved significant reductions in nitrate leaching. DCD showed excellent properties for controlling nitrate leaching, taking into account the fact that grain yield and N accumulated by plant were similar for the ASN-DCD and ASN treatments applied at the same N doses. The split N treatment did not offer any advantages in terms of leached nitrate, either with the use of single ammonium sulphate nitrate (ASN) or with single application of nitrification inhibitors. The nitrification inhibitors did not increase the yield but did not reduce it either. The drainage rate was the most important component of nitrate leaching. The low drainage values of the first year resulted in a sharp decline of nitrate leaching. However, the experiment of the second year, showed clear differences in nitrate leaching between treatments due to the greater drainage. Additional key words: dicyandiamide, dimethylpyrazolephosphate, N fertilization, split N application.


Introduction
It is necessary to control nitrate leaching resulting from agricultural practices in order to protect or improve water quality, but effective management is difficult because of the complex interactions between soil, water and nitrogen (Zerulla et al., 2001).The declaration of vulnerable zones to nitrate pollution (CD 91/676/EEC; OJ, 1991) and the inclusion of protecting water from nitrate pollution as conditions for obtaining Common Agricultural Policy (CAP) support have made farmers more aware of the need to control this problem.Maize crops have a great demand for water and nitrogen, particularly under Mediterranean conditions, and irrigation is therefore a common practice.
Nitrification inhibitors have been used at low concentrations (Ashworth et al., 1982;Amberger, 1991) with the aim of reducing nitrate leaching.In such cases, the strategy was to hold ammonium in the soil thereby retarding its oxidation to nitrate.The use of nitrification inhibitors may offer an alternative to splitting N applications (Boswell et al., 1976) in order to improve the efficiency of applied N.
Dicyandiamide, H 2 NC (NH)NHCN (DCD), a dimmer of cyan-amide or cyan-guanidine, is an effective nitrification inhibitor (Ashworth et al., 1982).DCD contains about 67% N and is non-volatile, non-hygroscopic, relatively water soluble (23 g L -1 at 13°C) and chemically and physically stable in normal environment conditions (Prasad et al., 1971).These properties allow DCD to be effectively formulated with a wide variety of N fertilizers, including ammonium salts.McCarty and Bremner (1989) established that the inhibitor DCD declines in efficacy as soil temperature increases from 10°C to 30°C.DCD inhibits the cytochrome oxidase that is responsible for ammonium oxidation by Nitrosomonas.The decomposition products associated with DCD, guanilurea and guanidine have no effect on nitrification, which finally were converted in urea, a conventional fertilizer (Amberger, 1991).Commercial N fertilizers formulated with DCD contain between 5% and 15% DCD-N (Reeves et al., 1986).
On the other hand, DMPP (3,4-dimethylpyrazole phosphate), a nitrification inhibitor developed by BASF (Limburgerhof Research Centre, Germany), also inhibits only the first stage of nitrification, reducing the rate of conversion of NH 4 + to NO 2 - (Serna et al., 1994).DMPP can be added to either conventional fertilizers or to slurries; it is highly specific in its action, and only a small amount (0.8% of applied nitrogen) is needed to inhibit nitrification for several weeks.Zerulla et al. (2001) have referred to the physical and chemical properties of DMPP.These authors and Díez et al. (2008) have proved that DMPP reduces nitrate lixiviation, though the duration of its action depends on climatic conditions, and particularly temperature and humidity (Pasda et al., 2001).DMPP remains effective in the upper soil layer even after heavy rainfall due to its low solubility (Fettweis et al., 2001).DMPP has passed all of the toxicological and ecotoxicological tests to which it has been subjected (Roll, 1999) and has proved to be highly plant compatible (Zerulla et al., 2001).In general, as proved by Barth et al. (2001), nitrification inhibitors such as DMPP are more effective in soil with coarse texture.
The results obtained with nitrification inhibitors have, however, been contradictory, probably because there are many factors that affect their soil efficiency, such as organic carbon content, temperature, irrigation rate, and the possible toxicity of these compounds or associated ammonium accumulation (Reeves and Touchton, 1986).One factor that has contributed to these contradictory results has been the uses of unsuitable techniques for measuring the amount of nitrate leached (Starr and Paltineanu, 1998).Few recently published studies treat this subject with reference to Mediterranean agriculture (Serna et al., 1994).
In the case of lixiviation, drainage is considered one of the main factors determining nitrate leaching, which is difficult to measure with undisturbed soil conditions (Webster et al., 1993).However, indirect methods based on a detailed knowledge of local soil water dynamics can improve our ability to estimate drainage from cropping systems (Vázquez et al., 2005).
The main aim of the experiments reported in this paper was to compare DCD and DMPP as inhibitors of ammonium oxidation and nitrate leaching after applying fertilizer to a maize crop grown under Mediterranean conditions.A secondary aim was to compare the relative effects associated with different nitrification inhibitors and split applications of nitrogen on maize yield and nitrate leaching.

Experimental design and crop management
The experimental site was located at the La Poveda Field Station in Arganda del Rey (Madrid) (40°19'N, 3°19'W), in the Jarama River basin.The soil, a Typic Xerofluvent (Soil Survey Staff, 1993), was a sandyloam that became progressively sandier with depth and had a gravel layer at a depth of 1.5 to 2.2 m.Physicochemical characteristics in the top 0-50 cm of the soil profile are shown in Table 1.Soil samples were analyzed for pH, organic matter (Walkley and Black, 1934) and carbonate (ISO 10693, 1995).Nitrogen, P, K and Ca levels were estimated using the electroultrafiltration (EUF) technique (Nemeth, 1979).Total N was determined from EUF extracts (EUF-N) from soil samples by digestion with UV radiation and subsequent oxidation with potassium persulphate in an alkaline medium (Díez, 1988).The phosphorus level was determined also colourimetrically using ammonium molybdate as a reagent (AOAC, 1990).Potassium and Ca levels were determined by flame emission photometry (AOAC, 1990).
The basic fertilizer used in all the treatments was ammonium sulphate nitrate (ASN; 26% N, 19.5% N-NH 4 + and 6.5% N-NO 3 -) a traditional N source, when DCD or DMPP is used as a nitrif ication inhibitor.ASN-DCD contains 26% total N, of which 5% is DCD.For our study, this fertilizer was prepared by Fertiberia S.A. (Madrid, Spain) from a mixture of the two products, adding liquid vaseline (2% w/w) as an adhesive to improve its stability.It was prepared 10 days before application.In the case of ASN-DMPP, ENTEC (trade mark) was used, which is a commercial fertilizer manufactured by COMPO (0.8% N content) (Barcelona, Spain).
The experimental area included f ifteen 100-m2 plots.A completely randomized design (CRD), with 5 treatment and 3 replications, was used during the first and second year.The treatments included a control without fertilization (C), a single application of ASN (ASN), a split application of ASN (SP), a single N application with ASN-DCD 5% (DCD) formulated for us, and a single N application with ASN-DMPP (DMPP) commercially prepared.In the second year, the plots received the same treatments as in the first, but with modified doses of N fertilizer (see doses of nitrogen).
Maize (Zea mays cv.Helen) cycle 700 (Syngenta) was grown at the experimental site in 2006 and 2007.It was sown the beginning of April in both years.The rows were spaced 75 cm apart, and the plant density was 90,000 plants ha -1 .During seedbed preparation, super-phosphate (18% P 2 O 5 ) and K 2 SO 4 (50% K 2 O) were applied at 22 kg P ha -1 and 111 kg K ha -1 , respectively.The maize was grown using traditional farm practices (INFOAGRO, 2009) for the area and was harvested on 6 October 2006 and October 18, 2007, when the grain was mature.In May, the experimental plots were weeded manually.
Maize plants were harvested from the central 10 m of the rows in each plot, and aboveground biomass was determined.Ten harvested plants were selected randomly before their different parts (stalks, leaves, bracts, cobs and grain) were separated, weighed, oven-dried for 24 h at 60°C, and then kept for a further 2 h at 80°C before reweighing to determine their dry matter (DM) content.The harvest index (HI) was calculated as a percentage of grain weight over aboveground biomass.Grain yield was calculated by multiplying aboveground biomass by the harvest index and expressed on a 14% dry matter basis.Nitrogen concentration was determined in plant fractions using a Kjeldhal method (AOAC, 1990) and pretreated with a solution of salicylic acid and sulphuric acid (Bremner, 1965).Plant N accumulation was calculated by multiplying fraction yields by their respective N concentrations.

Soil available nitrogen and doses of nitrogen
Topsoil samples (to a depth of 0.30 m) were taken in the 15 plots and in each season, which was done before planting the maize by use of a soil-core sampler.Available N was calculated by applying a EUF method  Mg m -3 ) 1.47 (Nemeth, 1979) (Vogel S-724) having previously calibrated organic EUF-N with respect to the amount of potentially mineralized nitrogen in the soil according to its N balance (Sánchez et al., 1998).Available N included mineral N plus potentially mineralized N (N available calculated was 86 and 125 kg N ha -1 in 2006 and 2007, respectively).The doses of N fertilizer were calculated according to nitrogen requirements and considering the nitrogen available in the soil.The treatments were applied at a dose of 220 kg N ha -1 in 2006 and 180 kg N ha -1 in 2007, for extraction of N provided by the cultivation of 240 kg N ha -1 .They were applied only once: on June 1, 2006 and on June 7, 2007, after sowing, as top-dressing when plants had six leaves.In the SP treatment, the ASN fertilizer was split into two applications: one third was applied during seedbed preparation (70 kg N ha -1 in each year) and the remaining two thirds (150 and 110 kg N ha -1 in 2006 and 2007, respectively) were applied on the same date as the rest of the fertilizers.

Monitoring soil mineral N
Soil samples were taken at depth of 0-0.30 m in each plot after the fertilization, once a week during two months, in 2006 and 2007.Twenty sub-samples were combined into one sample per plot at each sampling.The samples were air dried, extracted with 1 M KCl at a ratio of 1:5 parts by volume to weight-KCl solution to soil, centrifuged and decanted and a portion of supernatant were stored in a freezer for subsequent analysis.NO 3 --N concentration was determined colourimetrically using a Technicon AAII Auto analyzer with N1 naphtylethylenediamine (Strickland and Parsons, 1968).The ammonium content was analyzed using ionselective electrodes (Orion Research AG, USA).

Monitoring soil water content and drainage
The water used throughout the experiment was taken from an irrigation channel fed by the River Jarama.This water was sampled 18 times in the course of the experiment to determine the quality components of the irrigation water.An overhead mobile-line sprinkler system was used to irrigate the maize.Irrigation started on 19 June, 2006 and on 9 June, 2007 and continued until the end of August.The maize was watered every 7 to 10 days following the schedule traditionally (INFOAGRO, 2009) used by most growers in the area (although we applied lower water rate than their).The depth of the water table fluctuated from 4 to 4.5 m below the soil surface, depending on rainfall and river discharge.The average rainfall in this area is 460 mm yr -1 .
Drainage at a depth of 1.5 m was calculated by applying the following water balance equation for each measurement: where: D is the drainage (mm), R is the rainfall (mm), I is the irrigation (mm), ET c is the crop evapotranspiration (mm) and ∆S is the observed change in soil water reserves (mm) from depths of 0 to 150 cm.The water storage in each layer was obtained as the product of the θv (m 3 m -3 ), which was measured in each layer using capacitance probes and multiplied by the thickness of the layer (m) in question.
Crop evapotranspiration (ET c , Allen et al., 1998) was estimated from the equation: where ET o = reference crop evapotranspiration (mm d -1 ), estimated from meteorological data according to the FAO Penman-Monteith equation; K c = crop coefficient (Allen et al., 1998) (dimensionless) during the different phenological stages of the crop.Meteorological data were collected with a Vantage Pro Plus weather station (Davis Instruments, Hayward, California, USA) placed next to the plots.A data logger recorded data on an hourly basis (Fig. 1).Four (50 mm inside diameter) EnviroSCAN probes (Sentek Pty Ltd, South Australia) were positioned at a depth of 1.5 m. in 4 plots corresponding to the Control, DCD, DMPP, and SP treatments to monitor volumetric soil water content (θ v ).Five capacitance sensors using frequency domain reflectometry (FDR) (Fares and Alva, 2000) were installed in each probe to measure θ v at depths of 10, 40, 70, 120, and 150 cm.The capacitance sensors, which had previously been mounted inside the probes on pieces of plastic in order to place them at the specified soil depths, were then inserted into previously installed PVC tubes, and connected by wire to the data logger.This installation method prevented the formation of air pockets around the tube and caused minimal disturbance to the soil (Starr and Paltineanu, 1998).A comparison between Enviro-SCAN and other measuring devices is reported by Paltineanu and Starr (1997).The frequency signal (FS) from the device was converted into a percentage of volumetric water (θv) using a normalization equation based on frequency readings from the sensor when exposed to air and water.The equipment was specifically calibrated for the soil in question, using the calibration equation proposed by Paltineanu and Starr (1997).The device was programmed to take one reading every hour throughout the cultivation periods in both years.A data logger recorded the data.
Drainage was calculated as the mean drainage for each of the four plots (20 measurements: 4 probes × 5 depths).Figure 2 shows the data for cumulative drainage obtained in 2006 and 2007.

Nitrate leaching
A ceramic candle extraction system was used to obtain samples of the soil solution (interstitial water).This involved installing two tubes at a depth of 1.4 m in each plot (Díez et al., 1997).These depths were determined after previously estimating the particle size distribution within the soil profile and the heterogeneity with depth of the gravel layers in the different plots (Díez et al., 2000).We considered that any water reaching this level, near the gravel layer, had been leached to the groundwater (at an average depth of 4 m) because of the high hydraulic conductivity (Smith and Mullins, 1991).Consequently, the amount of drainage water at a soil depth of 1.4 m was the same as that at greater depths due the textural characteristics of the soil profile.Water samples extracted using the ceramic candle were assumed to represent the nitrate concentration of the drainage water.
The ceramic candle was fitted with (63 mm inside diameter, 7 cm long) porous ceramic cups (Nardeux Humisol, Les Ulis, France).The soil solution was collected, on a monthly basis, by means of an electric vacuum pump connected to a nylon tube and was then transferred to a storage bottle.A -80 kPa vacuum was applied  to the tubes and maintained for a period of 7 to 10 days at each sampling.After this period, water samples were extracted from inside of tube, using air pressure.Samples of the soil solution were extracted 9 times in both 2006 and 2007, in both cases during the crop periods and NO 3 -, NH 4 + concentration, and EC were determined.During drainage periods, NO 3 -leaching was calculated on a weekly basis by multiplying the weekly drainage time by the corresponding NO 3 -con-centration at 1.4 m for each sampling event (Díez et al., 1997).Nitrate and ammonium concentrations were determined following the same procedures as previously mentioned.

Statistical analysis
Two-factor analyses of variance (ANOVAs) [Statgraphics Plus 5.1 (Manugistics, 2000)] were used to examine differences between treatments (with five levels: C, DCD, DMPP, ASN and SP) and the cropping periods (with two levels: years 2006 and 2007) with respect to the variables: nitrate concentration in the soil solution, dry matter, grain yield and plant N accumulated.ANOVAs were performed with a 0.01 α-level.Duncan's multiple range tests (Duncan, 1955) were used to compare differences between treatments.

Soil nitrogen
The changes in mineral N over time determined from soil extracted with 1M KCl in the 2006 and 2007 seasons are shown in Figure 3 (ammonium) and Figure 4 (nitrate).In both years, higher ammonium values were observed after fertilization in treatments including a nitrification inhibitor.concentrations reached their highest values in treatments with nitrification inhibitors: 29 and 22 mg NH 4 + -N kg -1 , in the DCD and DMPP treatments, respectively, in 2006; and 23 and 21 mg NH 4 + -N kg -1 in the DCD and DMPP treatments, respectively, in 2007.ANOVA showed significant differences between the control no fertilized and the rest of the treatments (P < 0.01) and among days after fertilization (P < 0.01), but not between replications.
In contrast, nitrate levels (Fig. 4) increased significantly in the first 15 and 27 days after fertilizer application in treatments without a nitrification inhibitor up 29 and 28 mg NO 3 --N kg -1 in 2006 and 2007, respectively.The treatments with DCD or DMPP reached their highest soil nitrate concentration of 19 (DCD) and 18 (DMPP) mg NO 3 --N kg -1 in 2006, and 13 (DCD) and 14 (DMPP) mg NO 3 --N kg -1 in 2007.The ANOVA for nitrate showed significant differences between the control no fertilized and the rest of the treatments (P < 0.01) in both years.However, no significant differences were observed between the ASN and SP and between the DMPP and DCD treatments, in either season.Significant differences were observed between days after fertilization (P < 0.01), but not between replications.
Under our experimental conditions, 80% of the water applied during the two year study was used by the crop, although drainage was greater in 2007 (161 mm) than in 2006 (71 mm), due to different irrigation frequencies and particularly the rainfall pattern.Similar ET values were obtained in 2006 (782 mm) and 2007 (810 mm).The amounts of irrigation water applied to the maize crops in 2006 and 2007 were 788 and 778 mm, respectively.On the other hand, the mean nitrate concentration in irrigation water was 12.7 mg NO 3 --N L -1 in 2006 and 5.7 mg NO 3 --N L -1 in 2007.Consequently, the N contributions through irrigation water were 94 and 45 kg N ha -1 in 2006 and 2007, respectively.The water balances during the crop-growing periods of both years have been included in Table 2.

Drainage and nitrate leaching
The water lost due to drainage represented an average equivalent to 10 to 20% of the total irrigation water  applied.In 2007, the drainage loss was great due to intense rain (293 mm), whereas in 2006 loss was only 100 mm, favouring the tests to establish the effect of the nitrification inhibitors on NO 3 -leached under conditions of higher drainage.The differences observed between water inputs and losses in the system were attributable to the water reserves present in the soil before previous to the crop period.Significant differences were obtained between seasons (P < 0.01) and among treatments (P < 0.01) (Table 3).The nitrate concentrations at a depth of 1.4 m obtained in 2007 were generally greater than those for 2006.Duncan's test shows significant differences between treatments within each year (Table 3).Although in 2007 there were no significant differences between the DCD and DMPP treatments, in 2006, the DCD treatment showed significant differences from DMPP.The higher nitrate concentrations were obtained in the without inhibitor treatments (ASN and SP).
Due to low drainage in 2006, the nitrate leached in all treatments was very low, as Figure 5 shows, and the differences observed between treatments were very small.However, when N losses are high due to high drainage, as occurred in our 2007 experiment, a greater response to DMPP and DCD would be expected.Figure 5 shows the greatest losses due to NO 3 -leaching occurred in 2007 and also that there were appreciable differences among treatments, especially between those involving nitrification inhibitors (which exhibited smaller losses) and those involving ASN alone.DCD and DMPP treatments, obtained in 2007, showed similar performances (54.7 and 53.1 kg N ha -1 with DCD and DMPP, respectively).The treatments C and ASN showed clear differences with values of nitrate leaching of 12 and 78 kg N ha -1 , respectively.

Plant N accumulation and grain yield
In general, grain productions were higher in 2007 than in 2006 no doubt because of more moderate climatic conditions.Figure 1 shows maximum temperatures of 36°, 37°and 41°C respectively observed in May, June and July, 2006, while in 2007, the maximum temperatures were 32°, 33°and 40°C in the same months.But the lowest maximum temperatures in 2006 were 20°, 26°, and 34°C, registered in May, June, and July, respectively; whereas in 2007, the equivalent registers  were 15°, 22°and 29°C.Also, it must be regarded that May rainfall was higher in 2007 than 2006.
Significant differences in dry matter, plant N accumulation and grain yield were observed between seasons and treatments (P < 0.01).The Duncan multiple range tests showed only significant differences between treatment C (no fertilized) and the rest of the treatments in 2007.However, in 2006, significant differences were observed not only with respect to the treatment C but also between fertilized treatments, and especially between DMPP and SP (Table 4).

Discussion
Soil analysis carried out after fertilization showed an increase of ammonium and a reduction of nitrate content in the treatments with nitrification inhibitors.The ammonium content in 2006 was greater than in 2007 (Fig. 3); this was probably due to higher air temperatures during May and June 2006 and also to lower level of rainfall (31 mm in 2006 as opposed to 134 mm in 2007).The higher rainfall during May and June 2007 could explain the lower values in soil ammonium and nitrate contents observed in June.
Soil solution nitrate concentrations were affected by the different treatments (Table 3).In general, drainage water from fertilized plots contained high NO 3 -concentrations (between 2.2 and 435 mg NO 3 -L -1 in 2006 and between 10 and 950 mg NO 3 -L -1 in 2007) and very low NH 4 + concentrations (between 0 and 0.37 mg NH 4 + L -1 in 2006 and between 0 and 0.26 mg NH 4 + L -1 ).The average concentrations of nitrate in drainage water were lower in treatments with inhibitor (DMPP and DCD) (Table 3).No significant differences occurred between DMPP and DCD in 2007.Either DCD or DMPP lengthened ammonium presence in soil in a similar manner and showed lower NO 3 -concentrations (30%) than in the control plots fertilized with ASN.
Data on nitrate concentrations for the soil solution at a depth of 1.4 m were used to study the possibility of groundwater pollution.Cumulative NO 3 -discharge at a depth of 1.4 m depended mainly on the irrigation water and fertilizer treatment applied.The poor results obtained in 2006 in terms of leached nitrate were due to low drainage (71 mm).These results are similar to obtained by Díez et al. (2000) who observed, total leaching depended mainly on drainage and to a lesser extent on variations in NO 3 -concentration at the percolation depth.
In 2007, the drainage was greater than in 2006 because the rainfall was higher, and moreover, the frequency of irrigation was modified (leaving only one day between water applications).These changes resulted in a greater value of drainage (161 mm) with similar doses of irrigation and, consequently, the differences between treatments were clearer.Figure 5 shows the marked difference in nitrate leaching between the two seasons.
The results obtained relating on leaching, which show that treatments involving nitrification inhibitors The nitrate concentrations and levels of nitrate leaching observed with the SP split were higher than if ASN was applied in a single top dress; this was possibly due to the rainfall in May, particularly in 2007, which dragged a part of fertilizer.The May rainfall obviously did not affect the rest of the treatments in which fertilizers were applied in June.
Previously published results relating to the use of DCD are somewhat contradictory.Some authors, such as Williamson et al. (1998), attribute only very minor effects to DCD in terms of reducing nitrate leaching, but the doses used in their experiments were very low (1.1% DCD).Similarly, Davies and Williams (1995), working with a soil column, did not observe any significant effects with DCD.However, other authors have reported similar results to those reported by us.For example, Francis et al. (1995) and Cookson and Cornforth (2002) reported that DCD was effective for reducing nitrification and that it consequently reduced leaching.Serna et al. (1994) observed that if they applied DCD at 2%, nitrate leaching was reduced, with the loss of only 20% of the N added with this treatment as opposed to 68% without DCD.These authors (Serna et al., 2000) concluded that nitrate concentrations in drainage waters were reduced with DMPP (68% and 53% of the applied N was leached to below 0.60 m in the ASN and ASN + DMPP treatments, respectively).Furthermore, Chaves et al. (2006) concluded that, under favourable conditions, DCD is able to inhibit nitrification from cauliflower crop residues for 50 days and DMPP is able to do the same for at least 95 days.Our f inding are consistent with those obtained by Irigoyen et al. (2003) who established that either DCD or DMPP extended the presence of ammonium in soil.
The SP split treatment did not offer any advantages with respect to nitrate leaching, either with the use of single ASN or compared to the use nitrification inhibitors.SP was associated with greater losses due to leaching (106 kg N ha -1 in 2007) than ASN (78 kg N ha -1 ).Authors have compared the effects of nitrification inhibitors vs split N applications (Boswel et al., 1976) with different results.Arregui and Quemada (2006) concluded that applying N fertilizer at rates that were not excessive, neither splitting N fertilizer application nor the use of a nitrification inhibitor, consistently re-duced nitrate leaching.However, Molina and Ortega (2006) established that the larger N-NO 3 -leaching losses associated with the use of fertilizers without a nitrification inhibitor were restricted by split N applications.
Significant differences in ANOVAs, were obtained between years and between fertilized treatments with respect to dry matter, grain yield, and N accumulation (P < 0.01).In 2006, the Duncan test (Table 4) shows signif icant differences between the DMPP and SP treatments.In 2007, only significant differences were obtained between C no fertilized and the fertilized treatments.The results obtained in this paper show that the treatments with nitrif ication inhibitors did not increase the grain yield but neither did they reduce maize yields.However, some authors have reported positive effects on yield.Molina and Ortega (2006), working with Chilean soils in a ryegrass experiment, established that ASN + DMPP increased dry matter production and the efficiency of N use, and that leaching losses were reduced.Leaf N levels were also higher in plants fertilized with ASN + DMPP (Serna et al., 2000).Other authors, such as Reddy (1964), established that plant toxicity of the nitrification inhibitors was expressed by a reduction in the number of chloroplasts per cell and observed toxicity symptoms in plant at doses of 6 to 7 ppm N-DCD.However, the same author emphasizes that maize, wheat, or oats moderately tolerate DCD at rates of between 6 and 17 ppm of N-DCD.Our results were similar to those obtained by Reeves and Touchton (1986) who reported that commercial N fertilizers formulated with DCD contain between 5 and 15% DCD-N produced no observable signs of toxicity.
The drainage rate was the most important component of nitrate leaching.The experiment of 2006 showed that the low drainage rate resulted in a sharp decline of nitrate leaching.However, the experiment of 2007 showed clear differences in nitrate leaching between treatments due to the greater drainage.The applying of 5% DCD or 0.8% DMPP had the positive effect of reducing nitrate pollution.Either nitrification inhibitors lengthened ammonium presence in soil in a similar manner.In 2007, soil NO 3 -concentrations in treatments involving nitrification inhibitors were 30% lower than in the plots fertilized with ASN.In consequence, the use of N-containing fertilizers plus DMPP or DCD reduced nitrate leaching losses of up to 29% (P < 0.05) with respect ASN treatment.
Freshly prepared ASN-DCD (see fertilizer in Material and methods), exhibited excellent properties for controlling nitrate leaching.This was particularly evident if we take into account the fact that grain yields and N accumulation were similar for the DCD and ASN treatments for the same N doses.The use of DCD was associated with higher grain yields than DMPP, no significant differences were apparent.There were no significant differences among fertilizer treatments in terms of maize yield, either with or without nitrification inhibitors; this demonstrates that the DCD or DMPP treatments used in this experiment had no toxicity effects on plants.These results are consistent with those reported by Roll (1999) and Zerulla et al. (2001).

Figure 2 .Figure 3 .
Figure 2. Cumulative drainage in 2006 and 2007.The vertical lines indicate error bars for each measurement.

Figure 4 .
Figure 4. NO 3 -extracted from top soil by 1M KCl in 2006 and 2007, shown in days after N fertilization with treatments C, DCD, DMPP, ASN and SP.Values are means of three replicates.The vertical lines indicate error bars.

Figure 5 .
Figure 5. Nitrate leaching in 2006 and 2007 during the maize crop, according to treatment (C, DCD, DMPP, ASN and SP).Values are means of four drainage replicates and six nitrate concentration replicates.The vertical lines indicate error bars.

Table 1 .
Physicochemical properties of the top soil before sowing

Table 3 .
Mean nitrate concentration (in mg NO 3 -L -1 ) in the soil solution for each treatment, at 1.40 m depth during 2006 and 2007.Duncan's tests show signif icant differences between treatments within each year

Table 4 .
Mean of dry matter (kg ha -1 ), grain yield (kg ha -1 ) and plant N accumulation (kg ha -1 ) during 2006 and 2007.Duncan's tests show significant differences between treatments within each year