Residual effects of poultry litter on silage maize (Zea mays L.) growth and soil properties derived from volcanic ash

Abstract

Poultry litter (PL) is a cheap alternative to conventional fertilizers. The use of PL in this way also reduces the environmental problems normally associated with its disposal. The residual effect of PL may reduce the amount of fertilizer (especially N fertilizer) required by subsequent crops. This study examines the residual effects of PL (with and without additional mineral fertilizer) on the properties of a volcanic ash soil and on silage maize (Zea mays) yields in central Chile. Poultry litter and mineral fertilizer were applied in 2002–2003 and their residual effects were determined in 2004–2006. The dry matter (DM) yield, nutrient balance and apparent nitrogen recovery efficiency (ANRE) of the silage maize were determined for each season, and the soil properties were analyzed at three depths (0–20, 20–40 and 40–60 cm) at the end of the third season. Crop yield showed a positive response to all fertilizer treatments. The residual effect, the nutrient balance, N uptake and ANRE also improved with fertilizer treatment, especially with the PL treatments. The average DM yield for the PL treatments was higher than that observed using mineral fertilizer by 2.8 and 1.2 Mg ha⁻¹ in the third and fourth years, respectively. The ANRE was generally higher in the PL treatments, although it decreased over time (12.4 and 1.7% for the last 2 years, respectively). The mean ANREs for the mineral fertilizer treatment were 4.1 and 1.6% for the same years. The results suggest that the PL treatments had an important positive residual effect in terms of N supply. This should be taken into account when planning the next crop. After two annual applications of PL, slight increases were observed in soil NO₃-N at a depth of 0–20 cm, and extractable P at depths of 20–40 cm and 40–60 cm. No other soil variables were significantly affected by any of the treatments. An additional source of K was found to be necessary to maintain an adequate soil K level.

Key words:

• apparent nitrogen recovery efficiency
• nutrient balance
• poultry litter
• residual effects
• silage maize
• soil properties

INTRODUCTION

The poultry industry in Chile has grown over the past decade. Approximately 193.4 million birds were produced in 2004 (Instituto Nacional de Estadística 2005), which was associated with an estimated production of 218,542–290,100 metric tons of bedding residue or poultry litter (PL) (Gascho et al. 2001; Wood et al. 1996). This PL must be disposed of carefully because it can adversely affect environmental quality via the emission of gases or runoff, and by nutrient leaching (Sharpe et al. 2004; Sharpley 1996; Sharpley et al. 1993; Sims and Wolf 1994). A safe, environmentally friendly way of disposing of this waste is to use it as a source of nutrients on agricultural land. This can greatly increase the production of cotton (Gossypium hirsutum L.), maize (Zea mays L.) and wheat (Triticum aestivum L.) (Cooperband et al. 2002; Eghball et al. 2004; Gascho et al. 2001; Hirzel et al. 2004; Sharpe et al. 2004; Wood et al. 1996;), and can reduce the fertilizer costs normally associated with these crops.

The application of organic wastes to soils increases their fertility (Brown et al. 1993; Cuevas et al. 2003; Gascho et al. 2001; Rivero and Carracedo 1999). In addition, organic waste can also generate a positive residual effect that should be taken into account when planning the next crop (Eghball et al. 2004). The residual N effect obtained with the organic residues mainly results from the fact that the N becomes part of the clay and organic soil fraction (Jensen et al. 2000) and is immobilized as part of the soil's microbial biomass (Jensen et al. 2000; Jokela and Randall 1997; Sainz et al. 2004). Eghball et al. (2004) found that, in maize monoculture, the application of beef cattle (Bos taurus) manure and compost over 4 years generated greater amounts of dry matter (DM) and increased the whole plant N concentration in the fifth and sixth years. The electrical conductivity of the soil, its pH, P availability, and the NO₃-N concentration is also increased above that achieved with mineral fertilizer at similar application rates. Sørensen et al. (1994) indicated that approximately 5–6% of the total N applied in the form of sheep manure is involved in this residual effect. Sørensen and Amato (2002) found that the residual N effect associated with the application of pig slurry in a barley (Hordeum vulgare L.)/ray grass (Lolium perenne L.) crop rotation involved 2.0–4.0% and 1.2–2.5% of the N applied as manure in the second and third year, respectively.

The aim of the present study was to determine the residual effect of PL on silage maize yield and soil properties over 2 years following its application during the two previous years. These effects were established by determining the DM yield, the nutrient balance and the apparent nitrogen recovery efficiency (ANRE) of the crop, and by measuring selected soil variables.

METHODS AND MATERIALS

This experiment was conducted at the Santa Rosa experimental farm (Agricultural Research Institute, Regional Center Quilamapu, Chillán, Chile; 36°36′#S, 71°54′#W). The soil at the site, a Typic Melanoxerands (United States Department of Agriculture 1994), had a silty loam texture and an average depth of 0.6 m. The climate of the area is Mediterranean, with high temperatures and low rainfall in the summer, and lower temperatures and high rainfall during the winter (Fig. 1). The trial area had been cropped with spring wheat and oats (Avena sativa L.) in previous years. Table 1 shows the initial physical and chemical properties of the soil.

During 2002–2005, silage maize was cultivated in a field trial that was designed according to a complete randomized block with six treatments and four replicates. The experimental plots measured 5 m × 3.5 m, which allowed for five rows of maize drilled at a spacing of 0.70 m.

Figure 1  Climatic characteristics of the experimental site during the study period. (a) Temperature, (b) evaporation and (c) rainfall.

In 2002 and 2003 the following treatments were applied: one control without fertilizer (T1), two applications of mineral fertilizer (urea: 50% applied by hand 1 day before sowing and 50% at the six leaf stage) at low (T2) and high (T3) N rates (equivalent to 300 and 400 kg N ha⁻¹), and three PL amendments (applied by hand 1 day before sowing) at low (T4), medium (T5) and high (T6) N rates (equivalent to 10, 15 and 20 Mg PL ha⁻¹, respectively). The low and medium PL rate plots also

Table 1 Initial physical and chemical properties of the soil

Property	Depth (cm)
	0–20	20–40	40–60
Bulk density (g cc^−1)	1.20	1.25	1.3
Total porosity (%)	54.72	52.83	50.94
Water retention to 0.33 bars (%)	30.02	25.42	20.04
Water retention to 15 bars (%)	15.17	13.44	12.15
pH	6.5	6.5	6.7
OM (%)	6.1	6.4	2.6
Inorganic N (mg kg^−1)	15	14	5
P Olsen (mg kg^−1)	11	11	6
Exchangeable K (cmol kg^−1)	0.22	0.25	0.12
Exchangeable Ca (cmol kg^−1)	5.81	5.31	2.12
Exchangeable Mg (cmol kg^−1)	0.44	0.42	0.32
Exchangeable Na (cmol kg^−1)	0.06	0.04	0.11
Exchangeable Al (cmol kg^−1)	0.04	0.04	0.03
Available S (mg kg^−1)	12.75	13.17	26.61
OM, organic matter.

Table 2 Nitrogen applied in the different treatments in the first 2 years (2002 and 2003)

Treatment	Year	N rate (kg ha^−1)
		N from PL	N inorganic	N Total
T1	2002	0	0	0
	2003	0	0	0
T2	2002	0	300	300
	2003	0	300	300
T3	2002	0	400	400
	2003	0	400	400
T4	2002	200	100	300
	2003	200	100	300
T5	2002	300	100	400
	2003	300	100	400
T6	2002	400	0	400
	2003	400	0	400
T1, control; T2, low rate mineral fertilizer; T3, high rate mineral fertilizer; T4, low rate poultry litter (PL); T5, medium rate poultry litter; T6, high rate poultry litter.

received 100 kg N ha⁻¹ of mineral fertilizer (urea) when the crop reached the six leaf stage. Table 2 shows the total N inputs for each plot. The PL used was produced at a poultry farm near the experimental area and the litter material was wood shavings. Table 3 shows its average composition, which was determined by analyzing five samples collected over the experimental period. Phosphorous (triple super phosphate) and potassium (potassium chloride) fertilizer were applied once, 1 day before sowing in T2 and T3. The P₂O₅ and K₂O rates used in the low and high mineral fertilizer treatments were similar to those inherent in the low and medium rate PL treatments. No fertilizers were applied in 2004 and 2005.

Table 3 Characteristics of the poultry litter (PL) used in the experiment (data corrected to dry weight)

Variable	Value
Dry matter (g kg^−1)	700.0
pH, 1:2.5 PL : water	8.05
EC (dS m^−1)	7.36
OM (g kg^−1)	581.0
Organic C (g kg^−1)	336.98
C:N ratio	11.8
Total N (g kg^−1)	28.57
N-NO3 (mg kg^−1)	7.166
N-NH4 (mg kg^−1)	1.712
N inorganic (mg kg^−1)	8.878
Total P (g kg^−1)	9.73
Total K (g kg^−1)	17.14
Total Ca (g kg^−1)	16.071
Total Mg (g kg^−1)	4.790
Total Na (g kg^−1)	2.60
EC, electrical conductivity; OM, organic matter.

All plots were cultivated to optimize crop growth according to standard practices for silage maize in central Chile. Table 4 provides a summary of the crop husbandry practices used.

The trial site was ploughed in late winter each year before the fertilizers were applied. The soil was worked with conventional tilling equipment to form an acceptable seedbed. Seeds were planted using a disinfected standard drill. After emergence, weeds were controlled with a combination of herbicides depending on the year and weed pressure observed.

The crop was harvested at silage maturity (30–35% DM) (Millner et al. 2005; Plénet and Lemaire 2000); 10 contiguous plants in the center row were cut 10 cm above the soil surface.

Plant N was determined using the macro-Kjeldahl procedure. The P and K concentrations were determined by ashing a 2 g subsample in a ceramic crucible at 500°C for 7 h, dissolving the ash in 10 mL of boiling 2 mol L⁻¹ HCl for 5–10 min, and filtering through Whatman No.5 paper (Germany). The P concentration was measured in the extract by colorimetry using the molybdate ascorbic acid method; the K concentration was measured using atomic emission spectroscopy.

Nutrient uptake was calculated as the product of DM yield and the nutrient concentration obtained from each plot, and was used to establish the nutrient balance for each treatment. The accumulated nutrient balance for the first 3 years of the experiment was calculated at the end of the third year as the difference between the quantity applied and the uptake registered in each treatment. The residual N effect for the final 2 years (i.e. when no fertilizer was added) was calculated as the difference

Table 4 Crop husbandry information

	2002–2003	2003–2004	2004–2005	2005–2006
Corn variety	DK 567 (Dekalb)	P 527 (Pioneer)	TX75 (Tracy)	GK567 (Dekalb)
Plant population ha^−1	102,041	102,041	102,041	102,041
Sowing date	10 November	16 October	29 October	10 November
Harvest date	25 March	25 February	18 March	15 March
Days from sowing to harvest	135	132	140	125
Poultry litter application date	9 November	15 October	–	–
Mineral fertilizer application date	9 November	15 October	–	–
	6 January	14 December

between the N uptake in each fertilizer treatment and the control (Sørensen and Amato 2002).

The ANRE for the first 2 years was calculated as the percentage difference between the total N uptake in each fertilizer treatment and the control, divided by the total N applied (Dean et al. 2000; Ma et al. 1999; O’Neill et al. 2004; Rees and Castle 2002):

where Nt is the N uptake by the plants in the fertilizer treatment (first or second season), Nc is the N uptake by the control plants (first or second season), and Na is the N applied (first or second season).

The ANRE for the third season (no fertilizer added) was calculated using the following equation:

where Nt is the N uptake by the plants in the fertilizer treatment (third season), Nc is the N uptake by the control plants (third season), Na is the N applied, Δ is the difference in uptake (Nt − Nc), 1 is the first season, and 2 is the second season.

The ANRE for the fourth season (no fertilizer added) was calculated using the following equation:

where Nt is the N uptake by the plants in the fertilizer treatment (fourth season), Nc is the N uptake by the control plants (fourth season), Na is the N applied, Δ is the difference in uptake (Nt − Nc), 1 is the first season, 2 is the second season, and 3 is the third season.

In the autumn after the third growing season, 10 composite soil samples were collected manually from all plots at three depths: 0–20, 20–40 and 40–60 cm. The soil was mixed thoroughly in a bucket, air-dried and sieved through a 2 mm mesh before analysis. Soil pH was measured in water (1:2.5 soil : water ratio). Available soil P was determined by extracting 2.5 g of soil with 50 mL of 0.5 mol L⁻¹ NaHCO₃ for 30 min (Olsen et al. 1954). Available K was determined using 1 mol L⁻¹ NH₄OAc extraction followed by atomic absorption-emission spectrophotometry. Inorganic N (NO₃-N + NH₄-N) was extracted using 2 mol L⁻¹ KCl and determined by colorimetry using a Skalar autoanalyzer (segmented flux spectrophotometer). No soil samples were taken in the fourth year because the results obtained in the third year rendered this unnecessary.

The results were examined with anova and the least significant difference (LSD) test (P = 0.05) using the SAS software general model procedure (SAS Institute 1989). For the determination of ANRE, a contrast analysis of treatments with the same N rates was carried out.

RESULTS AND DISCUSSION

Crop dry matter yield

Figure 2 shows the DM crop yields for all treatments for the 4 years of the experiment: (a) first and second years, in which PL and mineral fertilizer were applied and (b) third and fourth years, when no fertilizer was applied. The highest DM production was obtained during the first 2 years, suggesting that neither the PL nor the mineral fertilizer had any detrimental effect on growth. Markedly increased maize yields (total aboveground maize) were observed in all fertilizer treatments, with ranges between 26.3 and 37.13 Mg ha⁻¹. For the two following years, when no fertilizer was applied, DM production was reduced. A reduction in the DM of the control was also observed during the third and fourth years, suggesting that the soil N supply had been reduced.

In the first year, DM production was similar among the fertilized treatments; the values obtained were all higher than in the control plots. A similar effect was seen in the second year; moreover, the difference between the control and the fertilized treatments was smaller than that seen in the first year. This could be because of the different variety of maize sown each year (Heckman et al. 2003). In addition, the sowing date differed by 20 days (Table 4).

In the third year, the difference in DM yield among the treatments showed a similar trend to that of the previous 2 years, but no significant difference was observed

Figure 2  Mean dry matter yields for the four growing seasons. (a) The first and second years when both the poultry litter (PL) and mineral fertilizers were applied and (b) the third and fourth years when no fertilizer was used. Error bars represent the standard error (n = 4). Letters above the bars indicate significant differences among treatments within a year. T1, control; T2, mineral fertilizer at the low rate; T3, mineral fertilizer at the high rate; T4, T5 and T6, poultry manure at the low, medium and high rates, respectively.

between the control and mineral fertilizer treatments T2 and T3. Significant differences were obtained, however, between the control and the PL treatments, suggesting that PL has a larger residual effect than mineral fertilizer. The smaller DM yield obtained in the control in the third year (compared to the previous years) suggests a smaller N supply from the soil, possibly because soil organic matter (SOM) turnover had reached an equilibrium in the absence of fertilizer (Johnston et al. 1989). In the fourth year, the differences among the treatments showed a similar trend to that observed in the third year. Dry matter production was generally less than that obtained in the previous years. Finally, the DM yield in the control was smaller than that observed in the third year, suggesting a reduction in the SOM turnover.

Nitrogen uptake, residual nitrogen effect and nutrient balance

Fertilizer application significantly increased plant N uptake (Table 5). In general, this followed a pattern similar to that for plant biomass. The average N uptake obtained with the different fertilizer treatments decreased over the years of the experiment. The uptake recorded in the control plots was similar for the first 2 years, suggesting that the N supply from the soil was stable over this period. The variation seen in N uptake between the fertilizer treatments could result from the different responses to the fertilizer by the two varieties used in the different years. In general, the highest N uptake obtained in the first 2 years of the study was seen in the low and high rate PL treatments.

In the third and fourth years (no fertilizer added) the N uptake in the PL treatments was greater than in the mineral fertilizer treatments, although the differences were not significant (Table 5). The N uptake recorded in the fourth year was smaller in all treatments than in the third year.

The residual N effect associated with the different sources of N was calculated as the difference in the N uptake in the treatment in question minus the N uptake seen in the control (Table 5). This effect was greater with the PL treatments (T4, T5 and T6) than with the mineral exclusive fertilizer (T2 and T3) over the 2 years in which no fertilizer was added. The average values obtained for the PL treatments were 40.7 and 17.6 kg ha⁻¹ for the third and fourth years, respectively, while for the mineral exclusive fertilizer treatments these were 19.7 and 10.9 kg ha⁻¹ respectively. These results suggest that PL is an important nutrient source and that its residual should be taken into account when planning for subsequent crops. Furthermore, the fertilization average cost with PL treatments was US$416 (US$323–481) compared with US$587 (US$446–728) for mineral fertilizer.

The results obtained at the end of the third year indicate that the plants in the high N rate treatments (T3, T5 and T6) showed no significant differences in their accumulation of N or P₂O₅ (Fig. 3). The T2 treatment led to a slightly negative plant N accumulation and positive P₂O₅ accumulation, although it would seem that this rate of application of the mineral fertilizer is enough to supply the nutritional needs of silage maize under the present experimental conditions. These results agreed with the DM results for the first 3 years of the study (Fig. 2). Treatments T1 and T4 showed a negative balance for these two nutrients. T1 (without fertilizer) showed the greatest differences with respect to the other treatments

Table 5 Nitrogen uptake, residual N effect and apparent nitrogen recovery efficiency (ANRE)

Variable	Season	T1	T2	T3	T4	T5	T6
N uptake (kg ha^−1)	First	138.1c	290.6b	300.8ab	389.6a	273.4b	330.2ab
	Second	138.7c	236.6b	217.0b	257.5ab	243.8ab	309.4a
	Third	81.48b	95.15ab	107.03ab	125.48a	127.5a	113.43ab
	Fourth	32.93b	39.63ab	48.08ab	45.63ab	51.33a	54.68a
ANRE (%)	First	–	50.83b	40.68	83.83a	33.83	48.03
	Second	–	32.66	19.58B	39.61	26.29AB	42.69A
	Third	–	3.7b	4.47	20.73a	8.66	7.91
	Fourth	–	1.65	1.49	3.6	0.88	0.59
	Average	–	22.21	16.56	36.94	17.42	24.81
N residual effect (kg ha^−1)	Third	–	13.7b	25.6A	44.0a	46.0A	32.0A
	Fourth	–	6.7b	15.2B	12.7a	18.4B	21.8A
For N uptake, mean values with different letters in the same line indicate significant differences (P < 0.05). For ANRE and N residual effect, mean values with different capital letters in the same line indicate a significant difference between treatments at the high N rate (400 kg ha^−1), and mean values with different small letters in the same line indicate a significant difference between treatments at the low N rate (300 kg ha^−1) (P < 0.05).

(Fig. 3). As expected, this led to a reduction in soil N, P and K concentrations over the years. The negative balance obtained for T4 (low rate PL plus mineral fertilizer) indicates that this treatment did not meet the nutritional needs of the growing maize. The T6 treatment, which received the highest amount of P, showed no more positive a balance than T5 or T3.

All the treatments showed a negative balance for K₂O, with T4 the most negative. This shows that the maize had a high K demand and that the supply of K from the fertilizer sources was insufficient to maintain soil fertility in terms of K.

Apparent nitrogen recovery efficiency

The ANRE is the fraction of N supplied by the fertilizer recovered in the harvested part of the plant (Cogger et al. 2001). The ANREs for the different treatments were, in all cases, lower in the second year than in the first. This might be attributable to different environmental factors or to the different maize varieties used. For the same rate of N, the ANRE was higher with the PL than with the mineral fertilizer in the first and second years of the study (Table 5). Furthermore, ANRE in the T6 treatment (PL fertilized exclusively) was high in 2002–2003, and the amount of mobilized N was approximately 10 g kg⁻¹ with respect to a content of 28.6 g kg⁻¹. This suggests that PL is an appropriate N supply source for silage maize under the experimental conditions of the present study. The ANRE in both the PL and mineral treatments was least efficient at the highest N rate. In the first year, a significant difference was observed between the PL and mineral fertilizer at the low N rate, while in the second year a significant difference was seen between the high rate treatments (T6 and T3).

The average ANREs for the 4 years of the experiment were 22.2, 16.6, 36.9, 17.4 and 24.8% for the T2, T3, T4, T5 and T6 treatments, respectively. These results indicate that the T4 treatment had the highest N efficiency. Furthermore, T4 showed the most negative N balance of all the fertilizer treatments. In addition, the average ANREs obtained over the 4 years for the treatments supplying 300 and 400 kg N ha⁻¹ were 29.6 and 19.6%, respectively. These results suggest that, for the growth of silage maize under the soil and the climatic conditions of the present study, the most appropriate N rate for both types of fertilizer was 300 kg ha⁻¹.

Soil chemical properties

The pH was not affected by the treatments applied. In general, pH increased with soil depth, as observed in the control treatment (Table 6). The evaluation period may, however, have been too short to generate differences in soil pH; the periods covered in other reports in which differences in pH were observed were much longer than the present study (Clark et al. 1998; Eghball 2002; Eghball et al. 2004). Consequently, the use of PL did not affect soil pH during the study period; this agrees with the results of Gascho et al. (2001).

The soil inorganic N concentrations (NH⁺ ₄ + NO⁻ ₃) and the amount of NO₃-N recorded (Table 6, Fig. 4, respectively) were only significantly different for the surface soil (20 cm soil depth). Similar results were reported by Eghball et al. (2004) in an experiment on maize supplemented with beef cattle manure and chemical fertilizers. The soil NO₃-N concentration decreased with soil depth in all treatments (Fig. 4), as reported by others authors (Cuevas et al. 2003; Eghball 2002). The average inorganic N concentration was similar at 20–40 cm and 40–60 cm, indicating that little leaching occurs and that the soil at these depths has less N than the topsoil. The topsoil inorganic N concentrations were similar in all the plots (Fig. 4, Table 6), suggesting that the sampling moment

Figure 3  Nitrogen, P2O5 and K2O accumulation over the first 3 years of the experiment. Error bars represent the standard error (n = 4). T1, control; T2, mineral fertilizer at the low rate; T3, mineral fertilizer at the high rate; T4, T5 and T6, poultry manure at the low, medium and high rates, respectively.

was either not the most appropriate or that the residual N effect might be better measured by analyzing plant N uptake. Similar results were obtained by Eghball et al. (2004). Moreover the average inorganic N concentrations obtained with the PL treatments were similar to those obtained with the mineral fertilizer, and only 5.13, 1.25 and 1.46 mg kg⁻¹ higher than the control for the first, second and third soil depths, respectively. It is important to note that the available soil N concentration obtained

Figure 4  Soil NO3-N concentration at the time of soil sample collection (end of the third growing season) at 0–20, 20–40 and 40–60 cm depths. T1, control; T2, mineral fertilizer at the low rate; T3, mineral fertilizer at the high rate; T4, T5 and T6, poultry manure at the low, medium and high rates, respectively.

under the conditions of this study suggest that N undergoes strong soil fixation and immobilization (Jensen et al. 2000; Jokela and Randall 1997; Sainz et al. 2004). Consequently, little leaching occurs in this type of soil; PL should, therefore, not contaminate groundwater.

Table 6 shows that the only variation in extractable P levels among plots was at the second and third soil depths. The surface soil P concentration was slightly higher in T2, T3, T4 and T5 than in the control, but not significantly so. This contrasts with the results of Sharpley et al. (1993) and Sharpley (1996), who applied PL to non-volcanic soils. As large amounts of P were applied in the present experiment, a greater increase in available P was expected. This may reflect the high P retention capacity of this volcanic soil, as indicated by same authors (Barreal et al. 2001; Beck et al. 1998; Curtin and Syers 2001; Eghball and Power 1999; Johnson et al. 2004; Mazzarino et al. 1997; Plénet et al. 2000). The P concentration fell with soil depth, as indicated by Eghball et al. (2004), although the levels obtained in the present study were well below the values reported by other authors (Cogger et al. 2001; Cuevas et al. 2003; Parkinson et al. 1999; Preusch et al. 2002; Singer et al. 2004) when a number of organic amendments were applied. The positive P balance obtained in the T3, T5 and T6 treatments (Fig. 3) had no residual effect on the soil. The average soil P concentration obtained in the PL treatments was 1.43, 2.01 and 1.29 mg kg⁻¹ higher than in the control plot at the first, second and third depths, respectively.

The available soil K was very low, and did not increase significantly with fertilizer treatment (Table 6). In addition, it became lower with soil depth. A negative balance was

Table 6 Soil chemical properties at three depths (0–20, 20–40 and 40–60 cm) in the different treatment plots

Property	Depth (cm)	T1	T2	T3	T4	T5	T6
pH (1:2.5 soil : water)	0–20	6.65	6.68	6.66	6.63	6.75	6.61
	20–40	6.62	6.58	6.69	6.65	6.70	6.65
	40–60	6.73	6.74	6.73	6.69	6.78	6.68
Ni (NH^+ 4 + NO^− 3) (mg kg^−1)	0–20	5.19c	11.1ab	6.25c	11.1ab	7.78bc	12.1a
	20–40	5.70	5.52	6.28	9.08	5.88	5.88
	40–60	4.37	4.27	6.96	7.96	4.83	4.70
Available P (mg kg^−1)	0–20	14.1	16.7	15.5	17.0	15.6	13.8
	20–40	5.99b	5.86b	9.54ab	10.1a	7.70ab	6.17ab
	40–60	5.01b	5.45b	6.22ab	5.95ab	6.98a	6.02ab
Available K (mg kg^−1)	0–20	101	117	94	109	86	78
	20–40	35	43	51	60	47	35
	40–60	27	55	43	55	51	35
Ni, inorganic N. Mean values with different letters in the same line indicate significant differences (P < 0.05). Analysis carried out in March 2005.

seen for this element in all treatments (Fig. 3), suggesting that the crop has a high K demand. Therefore, an appropriate K source should also be applied.

CONCLUSIONS

The results of this study show that PL is a good alternative to mineral fertilizer for the growth of silage maize. Its benefits include an increased DM yield, a better nutrient balance, an improved residual N effect, and an improved ANRE. The improved residual N effect has important implications for subsequent crops and could lead to savings in the use of fertilizer.

Compared with the mineral fertilizer, the application of PL did not result in higher soil NO₃-N levels, nor did it result in a higher residual soil extractable P. Consequently, its use poses no potential ground or surface water contamination hazard.

Maize has a high K demand and sufficient amounts of K could not be supplied by either of the fertilizers used in the present study. A negative K balance was, therefore, obtained; its soil concentration was very low in all treatments. This indicates that other sources of K need to be applied if adequate soil K levels are to be maintained.

Significant benefits were obtained when PL was used as a fertilizer to encourage the growth of silage maize. Agricultural soils may, therefore, provide an important PL recycling route for poultry producers.