SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (17)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Buehler, S.
Right arrow Articles by Frossard, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Buehler, S.
Right arrow Articles by Frossard, E.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Buehler, S.
Right arrow Articles by Frossard, E.
Related Collections
Right arrow Soil Kinetics
Soil Science Society of America Journal 66:868-877 (2002)
© 2002 Soil Science Society of America

DIVISION S-4—SOIL FERTILITY & PLANT NUTRITION

Sequential Phosphorus Extraction of a 33P-Labeled Oxisol under Contrasting Agricultural Systems

S. Buehlera, A. Oberson*,a, I. M. Raob, D. K. Friesenc and E. Frossarda

a Institute of Plant Sciences, Swiss Federal Institute of Technology (ETH), Eschikon 33, 8315 Lindau, Switzerland
b Centro Internacional de Agricultura Tropical, CIAT, A.a. 6713, Cali Colombia
c IFDC/CIMMYT Kenya, P.O. Box 25171, Nairobi, Kenya

* Corresponding author (astrid.oberson{at}ipw.agrl.ethz.ch)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Chemical sequential extraction is widely used to divide soil phosphorus (P) into different inorganic (Pi) and organic (Po) fractions, but the assignment of these fractions to pools of differing plant availability, especially for low P tropical soils, is still matter of discussion. To improve this assignment, the effect of land-use systems and related P fertilizer inputs on size of P fractions and their isotopic exchangeability was investigated. A Colombian Oxisol, sampled from a long-term field experiment with contrasting management treatments, was labeled with carrier free 33P and extracted after incubation times of 4 h, 1, and 2 wk. Phosphorus concentrations and 33P recovery in fractions sequentially extracted with anion exchange resin (Pi), 0.5 M NaHCO3 (Bic-Pi, Bic-Po), 0.1 M NaOH (Pi, Po), hot concentrated HCl (Pi, Po), and residual P were measured for each incubation time. Resin-Pi, Bic-Pi, NaOH-Pi, and hot HCl-Pi were increased with P fertilization, with the highest increase for NaOH-Pi. The recovery of 33P in the treatments with annual P fertilization clearly exceeding P outputs indicate that resin-Pi, Bic-Pi, and NaOH-Pi represented most of the exchangeable P. In these treatments, label P transformed with increasing incubation time from the resin to the Bic-Pi and NaOH-Pi fractions. The organic or recalcitrant inorganic fractions contained almost no exchangeable P. In contrast, in soils with low or no P fertilization, more than 14% of the 33P was recovered in NaOH-Po and HCl-Po fractions 2 wk after labeling, showing that organic P dynamics are important when soil Pi reserves are limited.

Abbreviations: Bic-P, 0.5 M HCO3-extractable P • CChl, microbially bound carbon released by chloroform • CIAT, Centro Internacional de Agricultura Tropical • CORPOICA, Corporacion Colombiana de Investigacion Agropecuario • Cp, P concentration in the soil solution • E1, quantity of P isotopically exchangeable in one minute (mg kg soil-1) • NChl, microbially bound nitrogen released by chloroform • PChl, microbially bound phosphorus released by chloroform • Pi, inorganic P • Po, organic P • Psum, sum of P extracted in all fractions of the sequential fractionation • Pt, total P in a defined P fraction • Ptot, total soil phosphorus extracted with H2O2 and H2SO4 • SA, specific activity (33P/31P)


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
PHOSPHORUS IS AN ESSENTIAL NUTRIENT for plants and often the first limiting element in acid tropical soils. Profound understanding of the P dynamics in the soil–plant system and especially of the short- and long-term fate of P fertilizer in relation to different management practices is essential for the sustainable management of tropical agroecosystems (Friesen et al., 1997). Chemical sequential extraction procedures have been and still are widely used to divide extractable soil P into different inorganic and organic fractions (Chang and Jackson, 1957; Bowman and Cole, 1978; Hedley et al., 1982; Cross and Schlesinger, 1995). The underlying assumption in these approaches is that readily available soil P is removed first with mild extractants, while less available or plant-unavailable P can only be extracted with stronger acids and alkali. In the fractionation procedure developed by Hedley et al. (1982) and modified by Tiessen and Moir (1993), the P fractions (in order of extraction) are interpreted as follows. Resin-Pi represents inorganic P (Pi) either from the soil solution or weakly adsorbed on (oxy)hydroxides or carbonates (Mattingly, 1975). Sodium bicarbonate 0.5 M at pH 8.5 also extracts weakly adsorbed Pi (Hedley et al., 1982) and easily hydrolyzable organic P (Po)-compounds like ribonucleic acids and glycerophosphate (Bowman and Cole, 1978). Sodium hydroxide 0.5 M extracts Pi associated with amorphous and crystalline Al and Fe (oxy)hydroxides and clay minerals and Po associated with organic compounds (fulvic and humic acids). Hydrochloric acid 1 M extracts Pi associated with apatite or octacalcium P (Frossard et al., 1995). Hot concentrated HCl extracts Pi and Po from more stable pools. Organic P extracted by concentrated HCl may also come from particulate organic matter (Tiessen and Moir, 1993). Residual P that remains after extracting the soil with the already cited extractants, represents very recalcitrant Pi and Po forms.

Several studies have related these different P fractions in tropical soils to plant growth (Crews, 1996; Guo and Yost, 1998) or showed the influence of land use and the fate of applied fertilizers (Iyamuremye et al., 1996; Linquist et al., 1997; Lilienfein et al., 1999; Oberson et al., 1999), and partly resulted in contrasting assignments of fractions to pools of different availability. By comparing the amounts of P extracted from the surface horizons of Brazilian Oxisols that had been under different land-use systems for 9 to 20 yr, either unfertilized or fertilized with mineral P fertilizer applications, Lilienfein et al. (1999) showed that most of the fertilizer was recovered in the Bic- and NaOH-Pi fractions, irrespective of the land-use system (resin-Pi was not measured). In a 4-yr field study conducted on a Hawaiian Ultisol, Linquist et al. (1997) recovered, 1 yr after fertilizer application, almost 40% of the applied triple super phosphate fertilizer in the hot HCl and H2SO4 fractions. Oberson et al. (1999) showed that in an Oxisol managed as a legume–grass pasture for 15 yr, resin-Pi, Bic- and NaOH-Pi, as well as NaOH-Po levels were maintained at a higher level over the whole year in comparison with the same soil with the same total P content but managed as a grass only pasture. Iyamuremye et al. (1996) found an increase in resin-Pi, Bic-Pi and -Po, as well as NaOH-Pi after addition of manure or alfalfa (Medicago sativa L.) residues to acid low-P soils from Rwanda. In the study of Guo and Yost (1998), resin-Pi, Bic-Pi, and NaOH-Pi were most depleted by plant uptake on highly weathered soils. NaOH-Pi was important in buffering available P supply while significant depletion of organic fractions could rarely be measured.

A possible method of gaining information about the availability of different P fractions is to label soil P, P fertilizers, or plant residues before applying the sequential fractionation scheme (MacKenzie, 1962; Weir and Soper, 1962; Dunbar and Baker, 1965). Two studies followed the movement of labeled P from plant residues to soil P fractions applying a modified Hedley (Daroub et al., 2000) or the Chang and Jackson (1957) fractionation procedures (Friesen and Blair, 1988). They found that at 6 or 11 d after plant residue addition, respectively, between 20 and 50% of the label was extractable as Pi with resin (Daroub et al., 2000) or with NH4Cl and NH4F (Friesen and Blair, 1988). For longer incubation periods up to 34 d, Daroub et al. (2000) showed a subsequent movement of the label from the resin-Pi fraction to the NaOH-Pi fraction. The results obtained in all these studies suggest that, in tropical soils, the amounts of P in the different pools measured by sequential P extraction procedures and the fluxes of P between pools are controlled both by physico chemical factors (sorption–desorption) and by biological reactions (immobilization–mineralization). However, the importance of these different reactions for different land-use systems, such as monocropping, pasture, or intercropping, remain largely unknown.

The objective of this study was to assess the effect of different land-use systems (native savanna, rice [Oryza sativa L.] monocropping, rice green manure rotation, grass legume pasture) on some physico chemical and biological reactions involved in P cycling in a Colombian Oxisol.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soils included in the study were sampled during the rainy season in September 1997 from a field experiment (Friesen et al., 1997) located at CORPOICA-CIAT (Corporacion Colombiana de Investigacion Agropecuario; Centro Internacional de Agricultura Tropical) research station, Carimagua, Meta, Colombia (4°30'N, 71°19'W). Mean annual temperature is 27°C and average rainfall is 2200 mm. The soils are well drained Oxisols (Kaolinitic, isohyperthermic Typic Haplustox) of clay loam texture (Table 1).


View this table:
[in this window]
[in a new window]
  		Table 1. Selected chemical and physical properties of the surface layer (0–20 cm) of the Colombian Oxisol under different land-use systems. Values are the average of four analytical replicates, except Fe- and Al-contents (three replicates).

 
The surface soil layer (0–20 cm) was sampled in the long-term "Culticore" field experiment, which was established in 1993 with the objective to test the effect of different farming systems on plant productivity and soil fertility (Friesen et al., 1997). The experiment had a split-plot design with four replicates with treatment subplots of 0.36-ha size. Soil samples used for this study were taken at random in two replicates of each treatment and the replicates were mixed for laboratory analysis. For our study, the following treatments were included.
  1. SAV (Native savanna): native grassland annually burned in February, not grazed; no fertilizer application.
  2. GL (Grass–legume pasture): rice (cv. Oryzica Sabana 6) in 1993, with undersown pasture, since 1993 grass–legume pasture with Brachiaria humidicola (Rendle) Scheickerdt CIAT 679, Centrosema acutifolium Bentham cv. Vichada CIAT 5277, Stylosanthes capitata J. Vogel CIAT 10280, and Arachis pintoi CIAT 17434. The pasture was partly resown for renovation in June 1996 with legumes (the same Arachis pintoi Krapovickas & Gregory, and Centrosema acutifolium and additionally Stylosanthes guianensis (Aubl.) Sw., CIAT 11833). Grazing intensity was on average 2.7 steers ha-1 during 15 d followed by a 15 d pasture regrowth phase.
  3. CR (Continuous rice): rice (cv. Oryzica Sabana 6, cv. Oryzica Sabana 10 since 1996) grown in monoculture; one crop per year followed by a weedy fallow incorporated with early land preparation at the beginning of the rainy season before sowing rice.
  4. RGM (Rice green manure rotation): Rice followed by cowpea Vigna unguiculata (L.) Walp. var. ICA Menegua] in the same year. The legume was incorporated at the maximum standing biomass level in the late rainy season before sowing rice in the following rainy season.

Before establishing the treatments GL, CR, and RGM on savanna, the soil was conventionally tilled after burning the native vegetation. At the beginning of the experiment all treatments except SAV received 500 kg dolomitic lime ha-1. Annual fertilization of rice consisted of 80 kg N ha-1 (urea, divided among three applications), 60 kg P ha-1 (triple superphosphate). In addition, 99 kg K as KCl, 15 kg Mg and 20 kg S (as MgSO4), and 10 kg Zn ha-1 were applied at establishment and at recommended rates thereafter. With cowpea, an additional 20 kg N and 40 kg P ha yr-1 and 60 kg K, 10 kg Mg, 13 kg S, and 10 kg Zn ha-1 were applied at establishment and at recommended rates thereafter. The introduced pasture (GL) received additional fertilization only in 1996 [per ha: 20 kg P, 20 kg Ca (lime), 10 kg Mg (lime), 10 kg S (elemental), and 50 kg K (KCl)]. Phosphorus input-output balances were estimated by subtracting the P removed from the system by grain and/or with animal live weight gains from the P applied in mineral fertilizers. Phosphorus exports in grain were calculated by multiplying weighed rice grain yields with measured P content in the grain. Phosphate exported in the animals was assumed to be 8 g per kg of live weight gain. Live weight gains in GL were on average 68 kg ha-1 yr-1 (Oberson et al., 2001). The systems differed in soil cultivation (frequent [CR and RGM], rare [GL], or no cultivation [SAV]) and in application of herbicides (frequent [CR and RGM], rare [GL], or no application of herbicides [SAV]). Cultivated soils were tilled to a maximum of 15-cm depth.

Topsoil samples (0–20 cm) were air dried and sieved to pass a 2-mm sieve before chemical analysis in the analytical service laboratory of CIAT or shipped to Switzerland where they were stored in an air-dried condition until use for the fractionation experiment in 2000.

Soil Characterization
Bray-II P (0.03 M NH4F, 0.1 M HCl) was extracted from 2-g subsamples of soil at a 1:7 soil solution ratio and 40 s shaking time (Bray and Kurtz, 1945). Total soil P (Ptot) was determined on subsamples of 0.25 mg soil with the addition of 5 mL concentrated H2SO4 and heating to 360°C on a digestion block with subsequent stepwise (0.5 mL) additions of H2O2 until the solution was clear (Thomas et al., 1967). Microbial nutrient pool sizes of P, C, and N (PChl, CChl and NChl) were estimated on the preincubated samples by extraction, of chloroform fumigated and unfumigated samples, with the Bray I extract (0.03 M NH4F, 0.025 M HCl) (PChl) (Oberson et al., 1997) or K2SO4 (CChl and NChl) (Vance et al., 1987). No k-factors (Brookes et al., 1982; Hedley and Stewart, 1982; McLaughlin et al., 1986) were used to calculate the total microbial nutrient pools from measured PChl, CChl and NChl as there exist no proper estimates for these in acid tropical soils (Gijsman et al., 1997). PChl was corrected for sorption of released P according to Oberson et al. (1997). Dithionite-citrate-bicarbonate extractable and oxalate extractable Fe and Al (Fed, Feox, Ald, Alox) were determined according to Mehra and Jackson (1960) and McKeague and Day (1966). The mineralogy of the soils was determined on total soil samples, pretreated with H2O2 to remove organic C, by X-ray diffraction analysis (XRD). The samples were ground under acetone in a tungsten carbide vessel of a vibratory disk mill (Retsch RS1) for 10 min. Longer grinding times were not applied because of the detrimental effect that further grinding can have on the crystallinity of minerals, especially Fe (hydr)oxides (Weidler et al., 1998). For the Cu K{alpha} radiation, the Bragg-Brentano geometry was chosen as an XRD routine setup. The measurements were carried out on a Scintag XDS 2000 (Scintag Inc., Ecublens, Switzerland) equipped with a solid state detector from 2 to 52° 2{theta} with steps of 0.05° 2{theta} and counting times of 16 s.

Sequential P Fractionation of Labeled Soils
Before starting the sequential P fractionation, the samples were preincubated in a climate chamber (24°C and 65% relative atmospheric humidity, no light) for 2 wk in portions of 100 g at 50% of their water holding capacity (300 g water kg-1 soil dry weight). Soil water content was controlled and adjusted every other day by weighing.

Subsamples of preincubated soils were labeled in portions of 15 g with 120 MBq 33P (half-life 25.4 d) kg-1 which were added with 10 µL deionized water per g soil. The mass of P introduced with the 33P label can be neglected (<2.5 x 10-3 µg P g-1 soil, Amersham product specification, July 2000). Therefore, the term ‘P concentration’ always refers to 31P and specific activities (SA) are calculated as:

[1]

Soil P was fractionated sequentially, after three different incubation times after labeling (4 h, 1 wk, and 2 wk), with three replicates per treatment following the modified method of Hedley et al. (1982), as described in Tiessen and Moir (1993), with HCO3-saturated resin strips (BDH #55164, 9 x 62 mm), followed by 0.5 M NaHCO3 (referred to as Bic-P), 0.1 M NaOH, (these first three steps each with an extracting time of 16 h) and concentrated hot HCl at 80°C for 10 min. The step using diluted cold HCl was omitted, as Ca-phosphates are only present at very low levels or are absent in highly weathered acidic soils (Agbenin and Tiessen, 1995), as shown for the soils used in this study by Friesen et al. (1997). Residual P was extracted as described previously for determination of Ptot.

The amount of soil extracted was doubled from 0.5 to 1 g with the original volumes of extractants (2 resin strips in 30 mL H2O, 30 mL NaHCO3, 30 mL NaOH, 15 mL concentrated HCl, 5 mL conc. H2SO4) used to get higher 33P-concentrations in the extracts. This was preferred to the alternative of higher label application as the radiation might affect microbial growth (Halpern and Stöcklin, 1977). After each extraction, the samples were centrifuged at 25 000 x g for 10 min before filtering the solutions of the Bic- and the NaOH-extracts through 0.45-µm pore size Millipore filters (Millipore Corporation, Bedford, MA, cellulose acetate), and the hot HCl and residual P extracts through Whatman No. 40 filter paper.

Phosphorus concentrations in all extracts were measured after neutralization by the Murphy and Riley (1962) method. This method was used directly for the P recovered from the resin strips and for Pi determination in the HCl extracts. Organic matter was first precipitated by acidification in the Bic- and the NaOH-extracts prior to Pi determination (Tiessen and Moir, 1993). Total P (Pt) in the Bic-, the NaOH- and the HCl-extracts was measured after digestion of Po with potassium persulfate (Bowman, 1989). Organic P was calculated as the difference between Pt and Pi in the Bic-, NaOH-, and hot HCl extracts.

To partition soluble 33Pi and 33Po in the Bic-, the NaOH-, and the hot HCl-extracts into separate solutions before counting, 5 mL of the extracts were shaken with acidified ammonium molybdate dissolved in isobutanol (Jayachandran et al., 1992). By this method, Pi is extracted into the isobutanol while Po remains in the aqueous phase. The complete recovery of Pi in the isobutanol phase was verified with the addition of a standard amount of 33P in 0.5 M HCO3, 0.1 M NaOH, and in 2.3 M HCl; recovery rates of added 33P in the isobutanol phase were between 97 and 103%, which was not significantly different from 100%. Counts in the aqueous phase were 1.1% (HCO3), 0.3% (NaOH), and 0.1% (HCl) of the original solutions showing that hardly any Pi went into this phase. Determination of total P in the aqueous phase is not possible because the presence of the molybdate interferes with the analysis (Jayachandran et al., 1992).

The radioactivity in each phase was determined with a liquid scintillation analyzer (Packard 2500 TR) using Packard Ultima Gold scintillation liquid in the ratio (extract to liquid) 1:5. The values were corrected for radioactive decay back to the day of soil labeling. All extracts were tested for possible quenching effects by adding defined 33P spikes. Quenching in the acid resin eluate could be prevented by dilution of 250 µL eluate with 750 µL deionized water for counting. The quench effect in the hot concentrated HCl extract could be avoided by counting solutions separated with acidified isobutanol because the separated phases were not affected by quenching. All other extracts were not affected by quench effects.

The recovery of the label calculated as the sum of all fractions, including residual P, was never complete. Therefore, subsamples of the soil residue after final acid digestion were dried and weighed into scintillation vials. These subsamples were then counted after addition of 1 mL water and 5 mL of scintillation cocktail.

Isotopic Exchange Kinetics
The procedure of isotopic exchange kinetics was used to assess the exchangeability of Pi in the soils sampled in the different land-use systems. The method was conducted as described by Fardeau (1996). Suspensions of 10 g of soil and 99 mL deionized water were shaken for 16 h on an overhead shaker to reach a steady state equilibrium for Pi. Then, at t = 0, 1 mL of carrier free H333PO4 tracer solution containing 1.2 MBq was added to each continuously stirred soil water suspension. Three subsamples were taken from each suspension after 1, 10, and 100 min, immediately filtered through a 0.2- m pore size micropore filter, and the radioactivity in solution was measured by liquid scintillation as described previously. To determine the 31P concentration in the soil solution (Cp, mg P L-1) 10 mL of the solution were filtered through a 0.025-µm filter (Schleicher & Schuell GmbH, Dassel/Relliehausen, Germany, NC 03) at the end of the experiment. The smaller filter pore size was used to exclude any influence of suspended soil colloids on Cp determination (Sinaj et al., 1998). The P concentration in the filtrate was measured in a 1-cm cell by the Malachite green method (Ohno and Zibilske, 1991) with a Shimadzu UV-1601 spectrophotometer (Shimadzu Corp., Kyoto, Japan). Phosphorus concentrations in solutions from the SAV, GL and CR treatments were close to the detection limit and they were measured again in samples concentrated by evaporation (5:1). This procedure resulted in Cp values that were not significantly different from the non-concentrated solutions.

Assuming that at any given exchange time the specific activity (SA) of inorganic phosphate in the solution is equal to the SA of the total quantity of phosphate which has been isotopically exchanged, it is possible to calculate the amount of isotopically exchanged P (Et, mg P kg-1 soil). The amount of P exchangeable within one minute (E1), indicating the immediately available P, is expressed as (Fardeau, 1996):

[2]
where R is the introduced radioactivity and r1 is the radioactivity remaining in solution after 1 minute of isotopic exchange. The factor 10 results from the soil solution ratio of 1:10.

Statistical Analysis
The effects of land-use systems and incubation time after labeling on P fraction size were tested by SYSTAT (Systat, 1997) by two-way ANOVA and Tukey's multiple range test over all treatments and times of fractionation, with exception of the resin-Pi fraction where the interaction between treatment and time was highly significant. Therefore, for resin-Pi the ANOVA and Tukey's multiple range test for the treatment factor were calculated separately for each time. Percentage recovery data were log-transformed to meet the requirements of analysis of variance. Time and treatment influences on the SAs of each fraction were tested by a two-way ANOVA and, as the interaction time x treatment was significant for all fractions, the treatment influence was tested for each repetition in time of sequential fractionation, separately.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
The mineralogy and Fe and Al (oxy)hydroxides contents of the surface soil from the four treatments were normal for this type of soil (Gaviria, 1993). On average of all treatments, the soil contained 68% quartz, 23% kaolinite, 4% anatase, 3% gibbsite, 2% rutile, and <1% vermiculite. There were no significant differences among the different land-use systems (SAV, GL, CR, RGM). This implies that any difference seen in P dynamics among land-use systems was mainly due to the land-use system and not to differences in soil mineralogy.

Total Soil P and P Balance Induced by the Different Treatments
The amounts of total P directly extracted from the soil samples (Ptot) were not significantly different from the total sum of P (Psum) extracted in the different fractions of the sequential extraction for SAV and CR while the direct extraction led to significantly higher values (P = 0.05) for GL and RGM (Table 2). To evaluate whether differences in total P content were related to P fertilization, the increase in Ptot (calculated as the difference between Ptot extracted from fertilized GL, CR or RGM and Ptot extracted from non fertilized SAV) was compared with the estimated P balance of these treatments (significant correlation, r2 = 0.87; P < 0.001). The increases in Ptot were of the same order of magnitude as the calculated P balance. Given the imprecision of the methods used to determine total P (O'Halloran, 1993) and of the estimations made to calculate the P balance, these results suggest that most of the P added as fertilizer and not taken up by plants remained in the surface layers of the soils. Except for the CR treatment these results agree well with Oberson et al. (2001). In their study, for CR only, about half of the calculated positive P balance was recovered in total P. Their sampling depth of 0 to 10 cm might explain this difference: soil tillage may have mixed P in the 0- to 10-cm soil layer with soil in the 10- to 20-cm layer, resulting in incomplete recovery of P in the 0- to 10-cm sampling depth.


View this table:
[in this window]
[in a new window]
  		Table 2. Phosphorus status and calculated P balances of the studied Oxisol under different land-use systems. Total P was calculated as the sum of the sequential P fractionation (Psum) or extracted directly with H2O2 and H2SO4 (Ptot).

 
Isotopic Exchange Characteristics
The effect of the four land-use systems on Pi exchangeability in the surface layer of the Colombian Oxisol is illustrated in Table 3. The ratio r1/R, which is inversely correlated to the P sorbing capacity of soils (Frossard et al., 1993), was below 0.05 for all treatments, suggesting that these soils have a high P sorbing capacity (Frossard et al., 1993). Furthermore, the r1/R-values of the four treatments were positively correlated to Ptot (r2 = 0.76; P < 0.001). This suggests that the different land-use systems have resulted, through P fertilization and cropping, in different sorption rates of Pi on soil minerals. Since in Oxisols P sorption is mainly governed by Al and Fe (oxy)hydroxides (Fontes and Weed, 1996), these treatments probably induced different degrees of Pi saturation on the (oxy)hydroxides such as gibbsite, which was identified in the soil from these treatments.


View this table:
[in this window]
[in a new window]
  		Table 3. Parameters of isotopic exchange kinetics influenced by the different treatments{dagger}.

 
The Pi concentration in the soil solution (Cp) was close to the detection limit in the SAV, GL and CR treatments (Table 3). Cp was significantly increased in all fertilized treatments (P < 0.001). In SAV, GL and CR, Cp was much lower than the critical concentration needed to sustain optimal growth for a large range of crops (Kamprath and Watson, 1980; Fox, 1981). The variation in the amount of Pi isotopically exchangeable in one minute (E1) followed the same trend as the variation in Cp.

P Concentrations in Different Fractions of the Sequential Extraction
The positive P balances of the fertilized GL, CR and RGM treatments resulted in significantly higher P concentrations (P < 0.001) compared to the savanna soil in all fractions except the organic fractions and residual P (Table 4). This agrees with the results of Friesen et al. (1997) and Oberson et al. (2001), who fractionated P forms according to the same method in the same field experiment, and studies conducted using other tropical soils (Beck and Sanchez, 1994; Linquist et al., 1997). Our results show that resin-Pi, Bic-Pi, and NaOH-Pi increased with P fertilizer input, with the NaOH-Pi fraction being the main sink for the applied P. The function of the NaOH-Pi fraction can be explained by the adsorption of Pi through ligand exchange with hydroxyl groups (Sposito, 1989) located on the surface of Fe and Al (oxy)hydroxides (Ainsworth et al., 1985; Parfitt, 1989; Torrent et al., 1992) and by the desorption of Pi from the surface of (oxy)hydroxides in the presence of 0.5 M NaOH (Houmane et al., 1986; Cross and Schlesinger, 1995).


View this table:
[in this window]
[in a new window]
  		Table 4. Distribution of P in various fractions according to the modified Hedley et al. (1982) fractionation procedure in the different treatments with and without P fertilization, at three times of incubation after soil labeling.

 
During the continuous 2-wk incubation of the soil samples, resin-Pi and Bic-Pi fractions increased significantly (P < 0.05) between the first and second fractionation date for all treatments (between 4 and 14 mg kg-1 for the sum of resin-Pi and Bic-Pi). There was no significant decrease in any fraction although total extractable Po tended to decline (between 8 and 18 mg kg-1) for all soils (Table 4). The absence of a significant movement of P out of Po fractions may be due to the high variability of the results, especially for the organic fractions where coefficients of variation for Bic-Po ranged from 13 to 70% and for NaOH-Po from 7 to 45%. Since Po is determined by the difference between Pt and Pi for a given extract, there are multiple sources of error. High variability of repeated measuring of Bic-Po and NaOH-Po were reported by Magid and Nielsen (1992). Problems in the determination of Pi are mentioned in Tiessen and Moir (1993), especially the possibility that Pi is precipitated along with the organic matter upon acidification and erroneously determined as Po (Pt-Pi). On the other hand, Po compounds could be hydrolyzed in the acidic solution during the measurement of P during the colorimetric measurement (Condron et al., 1990; Gerke and Jungk, 1991).

Increases in resin-Pi and Bic-Pi between 4 h and 1 wk of incubation suggest that mineralization of Po led to the release of labile Pi from Po fractions. As the first fractionation was started 4 h after labeling, the disturbance by mixing the soil with the label and the increased humidity may have stimulated microbial activity in spite of the preincubation. A temporary stimulation of the microbial activity by mixing when labeling the soil was indicated in microbial turnover studies conducted on soils from the same field experiment (Oberson et al., 2001). This assumption seems likely, as there were little changes in fraction sizes between the second and the third fractionation indicating a stabilization of the system.

Distribution of 33P Among P Fractions and Dynamics over Time
The fraction of 33P recovered in the resin-Pi fraction 4 h after labeling varied between 22% in SAV and 60% in RGM (Figure 1) . The 33P recovery in this fraction was positively correlated to Ptot (r2 = 0.87; P < 0.001, 4 h after labeling). The corresponding decrease with time of 33P in the resin fraction in RGM and CR coincided with an increase in label recovery in Bic-Pi and NaOH-Pi, while in SAV and GL the decline in resin 33P was accompanied by an increase in 33P in NaOH-Po (GL also NaOH-Pi), HCl-Pi and residual-P. For SAV and GL, the label recovered in resin-Pi decreased only slightly but significantly between the 1st and the 2nd wk, and the label recovery in Bic-Pi did not change significantly in this time. The amount of 33P in NaOH-Pi was almost stable over the entire incubation time with a small but significant increase between the first and the second week for GL. This shows that in SAV and GL the label was rapidly exchanged between these fractions and that equilibrium with the (labeled) soil solution was reached. In contrast, 33P in the Bic-Pi and the NaOH-Pi of CR and RGM was still increasing after 1 wk while resin-33Pi continued to decrease, showing that the exchange between these fractions was incomplete.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 1. Percentage of label recovery in the different fractions of the sequential P extraction and in the sum of all fractions at 4 h, 1, and 2 wk after labeling soil. Recovery in Bic-Po in all soils at all sampling times <1% (means of three replicates ±SD).

 
The data for 33Po were, because of the determination after the separation from Pi with the isobutanol method, not affected by the inherent problems in determination of the Po fractions in the Hedley fractionation scheme as described previously. Only small amounts of the label were found in organic fractions after 4 h, but there were already significant differences in NaOH-33Po (P < 0.001) in the order:

This might be due to differences in microbial activity as observed by Oberson et al. (2001) in the same field experiment. Microbial biomass in the incubated soils, indicated by measured PChl, CChl, and NChl values, was significantly different between the soils (Table 5), despite the fact that the samples had been stored in an air-dried condition for more than 3 yr before being used in this study. The assumption that recovery of the label in organic fractions was actually due to active processes and not to any analytical artifact is supported by the observed increases of NaOH-33Po and HCl-33Po for all soils over time. The total recovery of 20% (SAV) or 14% (GL), respectively, of the label in organic fractions 2 wk after labeling shows that these compartments have to be taken into account to understand the fate of P in these very low-P soils (Tiessen et al., 1984; Beck and Sanchez, 1994; Linquist et al., 1997). Oberson et al. (2001) suggested that Po mineralization significantly contributes to P availability in low input pastoral systems on these soils but that methods for quantification of mineralization remain to be developed.


View this table:
[in this window]
[in a new window]
  		Table 5. Size of soil microbial biomass nutrient pool in different agricultural systems after 20 days of incubation of the formerly air-dried soils. Values are the averages of three replicates.

 
The proportion of label in the hot HCl and residual P fractions increased significantly with incubation time in all soils. This contradicts the prevailing opinion of recalcitrance of P in these fractions (Guo and Yost, 1998; Neufeldt et al., 2000). While the total P content in the residual fraction varied significantly with time (Table 4), this was not the case for hot HCl extractable Pi, while hot HCl extractable Po tended to decrease. This suggests that the movement of the label to these fractions was not due to net P movement but to exchange processes.

Total 33P Label Recovery
At all sampling times during the incubation study, in total between 67 and 94% of the applied 33P label could be recovered in the sum of all P fractions (Fig. 1). This total of label recovery was generally in the order SAV < GL < CR < RGM. Incomplete recoveries can be explained by the fact that the method used to assess total P or residual P was not efficient enough to extract all of the soil P. Comparative studies have shown that total P can only be extracted reliably by alkali fusion (Syers et al., 1967; Bowman, 1988), which could not be used in this work. The analysis of soil residues after the acid extraction of residual P (Table 6) indicated that significant amounts of the label remained unextracted, these being higher for SAV and GL than CR and RGM. Although counting of 33P bound to solid phases is generally possible, problems of phase, impurity, self absorption of scintillations by the soil particles, or color quenching effects (Gibson, 1980) are difficult to correct, as these influences might be highly variable between samples. However, the recovery of standard additions of 33P to our soil residues was complete and the correlation of the measured radioactivity in the different soil treatment residues with the sample weight was linear (data not shown), thus confirming the qualitative information obtained from the counting of the soil residues.


View this table:
[in this window]
[in a new window]
  		Table 6. Radioactivity measured in soil solid residues by scintillation counting after extraction of residual P by sequential P fractionation starting 1 week after soil labeling.

 
Altogether the results suggest that the transfer of 33P among the different fractions determined by the sequential extraction was strongly dependent on the degree of saturation of soil Al and Fe (oxy)hydroxides with Pi, and therefore on the bonding energy of Pi to the soil minerals. It is indeed known that a high Pi saturation of metal oxide surfaces causes a more negative charge on the surface and prevents the specific sorption of further Pi ions (Ryden et al., 1977; Bowden et al., 1980). In the low P treatments (SAV and GL), most Pi would be sorbed with such a high energy that its exchangeability would be very limited. Specific sorption of 33P to the surface of Al and Fe (oxy)hydroxides of these soils, although unlikely (Frossard et al., 1994), cannot be excluded (Barrow, 1991). In contrast, in the P rich soils (CR and RGM), annual P additions may have resulted in the build up of larger quantities of Pi that was exchangeable with 33P.

Specific Activities in the Fractions Determined by the Sequential Extraction
The highest specific activities observed in the incubation experiment were obtained in the resin extract after 4 h of incubation (Table 7). This is consistent with the assumption that the amount of P desorbed from the soil by resin is in very rapid exchange with Pi in the soil solution, as suggested by other studies (Amer et al., 1955; Bowman and Olsen, 1979; Tran et al., 1992; Schneider and Morel, 2000). The subsequent decrease in SA of resin-Pi reflected the process of isotopic exchange between 33P and stable Pi located on the soil's solid phase (Fardeau, 1996). The order of the SAs in the Pi fractions after 4 h of incubation followed the extraction sequence (resin-Pi > Bic-Pi > NaOH-Pi > HCl-Pi > residual P), showing that the strongest reactants extracted either large quantities of slowly exchangeable P or a large quantity of P in which only a small part was rapidly exchangeable. After 2 wk the SAs of resin-Pi, Bic-Pi and NaOH-Pi became closer, suggesting that equilibrium with respect to P transfer between these fractions was being approached. The SAs of resin-Pi, Bic-Pi and NaOH-Pi were not significantly different in SAV while the SA of resin-Pi was still significantly higher than the SA of Bic-Pi in GL and higher than in Bic-Pi and NaOH-Pi in CR and RGM. These observations show that it is not possible to discuss the exchangeability of a certain P fraction without reference to a defined time of exchange (Fardeau, 1996).


View this table:
[in this window]
[in a new window]
  		Table 7. Specific activities (33P/31P) in extracts of the Hedley sequential fractionation in the different treatments of the labeled Oxisol at different times after labeling.{dagger}

 
Although the SAs of the NaOH-Po and HCl-Po fractions were relatively low, they showed that, depending on land use, these fractions were connected through active processes with the soil solution, most probably through microbial activity (Oehl et al., 2001). This indicates that the determination of plant available P with short-term isotopic exchange experiments might lead to errors since the dynamics of organic P forms are excluded.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The effect of contrasting land-use systems on soil P fractions extracted by a modified Hedley et al. (1982) P sequential fractionation procedure was assessed in an Oxisol during a 2-wk incubation on soils labeled with carrier free 33P. The quantities of 31P and 33P recovered in the different fractions were strongly dependent on the total P content of the soil, which was affected by the amount of P added as fertilizer and removed by plant P uptake.

In the two treatments fertilized annually with P and with a large positive P input-output balance, most of the Pi was stored in the resin-Pi, Bic-Pi, and NaOH-Pi fractions. The use of carrier free 33P confirmed that, under all land-use systems studied, these soil P fractions contained most of the exchangeable P and that 33P was transferred from the soil solution first to the resin fraction and then to the Bic-Pi and NaOH-Pi fraction. This suggests that, when this Oxisol is regularly fertilized, P is stored in these three fractions while plants might take up P from the same fractions. In the two other treatments, which had either never been fertilized or had been fertilized only once at the beginning of the field trial, the transfer of 33P in these three fractions (i.e. resin-Pi, Bic-Pi, and NaOH-Pi) was less clear, suggesting that the soil Pi was much less exchangeable. In these soils, however, the transfer of 33P into organic P fractions was more important (up to 20% of the label was found in the organic P fractions 2 wk after labeling). As the pool sizes of these organic fractions did not change significantly over time of incubation, the label recovery indicates relatively quick cycling processes, probably as a result of microbial activity. In low P Oxisols, these processes are relevant and should be considered when estimating soil P availability for plants.


   ACKNOWLEDGMENTS
 
We thank Dr. P.G. Weidler (ETH Zürich, ITÖ) for the XRD measurements, Mrs Roesch (ETH Institute for Plant Science) for measuring the Al and Fe concentrations and the field staff at CORPOICA-CIAT Carimagua research station for taking soil samples. This research was funded by Swiss Centre for International Agriculture (ZIL) and Swiss Development Cooperation (SDC).

Received for publication June 5, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
Z. He, C. W. Honeycutt, B. J. Cade-Menun, Z. N. Senwo, and I. A. Tazisong
Phosphorus in Poultry Litter and Soil: Enzymatic and Nuclear Magnetic Resonance Characterization
Soil Sci. Soc. Am. J., August 20, 2008; 72(5): 1425 - 1433.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
R. G. Myers, A. N. Sharpley, S. J. Thien, and G. M. Pierzynski
Ion-Sink Phosphorus Extraction Methods Applied on 24 Soils from the Continental USA
Soil Sci. Soc. Am. J., March 1, 2005; 69(2): 511 - 521.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
E. K. Bunemann, F. Steinebrunner, P. C. Smithson, E. Frossard, and A. Oberson
Phosphorus Dynamics in a Highly Weathered Soil as Revealed by Isotopic Labeling Techniques
Soil Sci. Soc. Am. J., September 1, 2004; 68(5): 1645 - 1655.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (17)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Buehler, S.
Right arrow Articles by Frossard, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Buehler, S.
Right arrow Articles by Frossard, E.
GeoRef