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 ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bünemann, E. K.
Right arrow Articles by Oberson, A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bünemann, E. K.
Right arrow Articles by Oberson, A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Bünemann, E. K.
Right arrow Articles by Oberson, A.
Related Collections
Right arrow Tropical Soils
Right arrow Microbial Processes
Right arrow Nutrient Cycling
Right arrow Isotopes
Right arrow Phosphorus
Published in Soil Sci. Soc. Am. J. 68:1645-1655 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-4—SOIL FERTILITY & PLANT NUTRITION

Phosphorus Dynamics in a Highly Weathered Soil as Revealed by Isotopic Labeling Techniques

E. K. Bünemanna, F. Steinebrunnerb, P. C. Smithsonc, E. Frossardb and A. Obersonb,*

a Univ. of Adelaide, Soil and Land Systems, Glen Osmond, SA 5064, Australia
b Inst. of Plant Sciences, Swiss Federal Inst. of Technology Zurich (ETH), Eschikon 33, CH-8315 Lindau, Switzerland
c Berea College, CPO 2064, Berea, KY 40404, USA

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


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Isotopic labeling techniques have the potential to elucidate soil P dynamics and the fate of P sources added to the soil, but they have rarely been applied to highly weathered tropical soils. We collected soils from two crop rotations [continuous maize (COM; Zea mays L.) and maize-crotalaria (MCF; Crotalaria grahamiana Wight & Arn.) fallow rotation] in a field experiment in Kenya and incubated them for 9 wk after addition of a plant residue or inorganic phosphorus (Pi), both labeled with 33P and added at 6 mg P kg–1 soil, or after carrier-free labeling of isotopically exchangeable soil phosphorus (soil IEP). The amount of P and recovery of 33P were determined in resin-extractable Pi (Presin), microbial P (Phex), and in a 0.1 M NaOH extract of samples from which Presin and Phex had been removed. The Presin increased after addition of Pi, while Phex increased after plant residue amendment, involving considerable microbial uptake of soil P. The recovery of 33P in Presin followed the order added Pi > soil IEP > plant residue, and decreased steadily from 7 to 22% after 1 d to 3 to 5% after 9 wk. The recovery of 33P in Phex remained constant throughout the incubation, being greater after plant residue amendment (15%) than in the other two treatments (4–7%). An additional 66 to 76% of 33P was recovered in the NaOH extract, as much as 27% of which was in organic phosphorus (Po) after plant residue amendment and 2 to 8% in the other two treatments. Similar to P dynamics after plant residue amendment, the comparison of the two rotations indicated a shift toward Phex and Po with increasing microbial activity due to previous fallow biomass incorporation.

Abbreviations: COM, continuous maize • ddH2O, double-distilled water • MCF, maize-crotalaria fallow rotation • NaOH-Pi, inorganic phosphorus extracted with 0.1 M NaOH • NaOH-Po, organic phosphorus extracted with 0.1 M NaOH • NaOH-Presin, phosphorus of 0.1 M NaOH extracts recovered on resin membranes • NaOH-Pt, total phosphorus extracted with 0.1 M NaOH • Pfum, phosphorus extracted with resin membranes in the presence of hexanol • Phex, hexanol-labile microbial phosphorus (difference between Pfum and Presin corrected for sorption of microbial P) • Pi, inorganic phosphorus • Po, organic phosphorus • Presin, inorganic phosphorus extractable with anion-exchange resin membranes • Pt, total P • SA, specific activity • soil IEP, isotopically exchangeable soil phosphorus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
LIMITED AVAILABILITY of P is often the main constraint for plant growth in highly weathered soils of the tropics. A better understanding of soil P dynamics is required to improve management practices in tropical agroecosystems. Soil P dynamics are characterized by physicochemical (sorption–desorption) and biological processes (immobilization–mineralization). Walker and Syers (1976) showed that the proportion of Po increases at later stages of soil development, especially in relation to labile Pi. Thus, the relative contribution of soil biological processes to deliver plant-available P may become more important when the availability of Pi is low. The decomposition of soil organic matter and plant residues is indeed often the main source of plant nutrients in low-input small-scale farming systems of the tropics (Gijsman et al., 2002).

On highly weathered soils in western Kenya, production of maize (Zea mays L.) is doubled by annual applications of 50 kg P ha–1 yr–1 as triple superphosphate (Bünemann et al., 2004b), but mineral P fertilizers are often not economical or not available to small-scale farmers. Without any P fertilizer added, maize yield and P uptake can also be twice as great after a one-season planted fallow with legumes such as Crotalaria grahamiana (subsequently referred to as crotalaria) than after maize (Niang et al., 2002; Smestad et al., 2002; Bünemann et al., 2004b). This suggests improved P availability after incorporation of fallow biomass, which consists of planted legume plus weed biomass. Measurable changes in the topsoil 1 yr after incorporation include an increase in Po and Phex under maize-fallow rotations compared with COM, while the availability of Pi does not differ between crop rotations (Bünemann et al., 2004b). In an incubation experiment with the same soils, incorporation of crotalaria residues that added 14.2 mg P kg–1 soil increased Phex by 5 to 8 mg P kg–1 soil within the first week, while it decreased Presin by 0.3 to 0.4 mg P kg–1 soil (Bünemann et al., 2004a). This suggests microbial immobilization of P from the plant residue as well as from soil P pools, which in the absence of isotopic labeling cannot be distinguished.

Labeling of plant residues with the radioisotopes 32P or 33P has been used to investigate the recovery of P from plant residues in soil P pools and in growing plants (Dalal, 1979; McLaughlin and Alston, 1986; Friesen and Blair, 1988; McLaughlin et al., 1988a, 1988b; Daroub et al., 2000). In the pot experiment of McLaughlin and Alston (1986), 65% of 33P added with a labeled medic (Medicago truncatula Gaertn.) residue was recovered in the microbial biomass, compared with 8% in the aboveground wheat biomass after 34 d. Even the recovery of simultaneously added calcium phosphate labeled with 32P was greater in the microbial biomass (29%) than in the shoots of wheat (10%). While such results point to the importance of the microbial biomass in soil P dynamics, these studies were restricted to temperate soils. The partitioning of labeled P additions among soil P pools may differ substantially in highly weathered tropical soils with greater P sorption on sesquioxides.

The radioisotopes 32P or 33P can also be added to soil without simultaneous application of 31P (i.e., carrier-free). In this case, the amount of P introduced with the radioisotope is negligible in relation to native P in the soil solution and other soil P pools, but the radioisotope can easily be detected because of the sensitivity of ß counting. Carrier-free addition of 32P or 33P is commonly used to determine isotopically exchangeable P in soil-solution mixtures and in pot experiments (Fardeau, 1996; Bühler et al., 2003). Recently, incubation experiments using carrier-free labeling of soil IEP were conducted on highly weathered soils from Colombia to elucidate soil P dynamics. The recovery of the tracer was determined either in the microbial biomass (Oberson et al., 2001) or in sequentially extracted P fractions (Bühler et al., 2002). Depending on the agricultural system, as much as 25% of added 33P was recovered in chloroform-labile (microbial) P within 2 d (Oberson et al., 2001). The recovery of as much as 20% of added 33P in Po fractions after 2 wk of incubation (Bühler et al., 2002) also points to the importance of biological soil P transformations. In the extraction scheme used by Bühler et al., however, microbial P was not removed by fumigation-extraction before the Po fractions were extracted, and a large proportion of the 33P recovered in Po may have originated from the living microbial biomass. Inclusion of a fumigation step as proposed in the original sequential extraction scheme (Hedley et al., 1982) could allow a better distinction between P contained in the living microbial biomass and Po in nonliving soil organic matter.

Our main objective was to compare the fate of P when added to a highly weathered soil as plant residues or as water-soluble Pi. To this end, we followed the incorporation of 33P added with these P sources into various soil P pools. Carrier-free labeling of soil IEP was used as a control, elucidating soil P dynamics in the absence of fresh P additions. To better characterize the role of the soil microbial biomass in soil P dynamics, we used two soils with contrasting microbial activity but otherwise similar characteristics.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Design
The incubation experiment had a two-factorial design, with three different 33P-labeled P sources and soils from two crop rotations (COM and MCF). These soils differ in their microbial properties (Bünemann et al., 2004a, 2004b) and were therefore used as a model to study the implications of size and activity of the soil microbial biomass on P dynamics. The P sources were two freshly added sources (either plant residues or water-soluble Pi in the form of phosphoric acid, both added at 6 mg P kg–1 soil) and soil IEP (Table 1). In carrier-free labeling of soil IEP, the recovery of 33P in soil P pools allows to monitor ongoing exchange processes and P transformations at steady state, that is, in the absence of net changes due to P additions. It can also be used as a reference for the interpretation of isotope recovery from labeled P additions.


View this table:
[in this window]
[in a new window]
  		Table 1. Description of the experimental factor labeled P source.

 
We followed the distribution of 33P added with the labeled P sources or by carrier-free labeling of soil IEP into Presin, microbial P, and NaOH-extractable P. The size of these pools, the recovery of 33P in them, and their isotopic composition [specific activity (SA), i.e., the ratio between the 33P radioisotope and the stable 31P] were determined seven times during 9 wk of incubation. The Presin includes Pi in the soil solution and weakly adsorbed Pi (Amer et al., 1955) and can be used as an indicator of plant available P (Tiessen and Moir, 1993). It thus permits to study the interaction between microbial P and available Pi. Microbial P (Phex) was determined as P rendered extractable by hexanol fumigation and corrected for sorption, without applying a conversion factor to account for incomplete lysis of microbial cells. To investigate the fate of 33P beyond these two labile P pools, samples from which Presin and Phex had been removed were extracted with 0.1 M NaOH, and different methods were tested to separate 33Pi and 33Po in NaOH extracts. Sodium hydroxide extracts aluminum and iron associated Pi (NaOH-Pi), as shown by McDowell et al. (2003). The NaOH-extracted Po (NaOH-Po) is a less well-defined fraction that may comprise easily mineralizable and more stable Po forms (Tiessen and Moir, 1993).

Soil Sampling and Soil Properties
Soil samples were taken in January 2002 in a field experiment in western Kenya (0° 09' N, 34° 33' E) on a kaolinitic, isohyperthermic Kandiudalfic Eutrudox (USDA classification) with 390 g kg–1 clay and 370 g kg–1 sand in the top 15 cm. Continuous maize and various maize-fallow rotations combined with different P fertilization rates have been compared in this experiment since 1997 in a randomized block design with four replications (Bünemann et al., 2004b). For the present study, two crop rotations were selected which did not receive any Pi fertilization. Continuous maize represents the traditional system with two maize crops per year, whereas the rotation of maize with a crotalaria fallow is tested as an option for soil fertility improvement (Table 2). The plant biomass produced by the crotalaria fallow during 6 to 8 mo is incorporated to the 15-cm depth by manual tillage 2 to 5 wk before planting the next maize crop. Further details of the field experiment are given in Bünemann et al. (2004b).


View this table:
[in this window]
[in a new window]
  		Table 2. Crop rotations selected for the study, and their effects on total, organic, inorganic, and microbial nutrients in the 0- to 15-cm soil layer.{dagger}

 
Composite soil samples consisting of 15 random cores (0–15 cm) were collected from each plot (6 by 6 m) in each of the four field replications. At sampling in January 2002, maize in COM had just been harvested while the fallows in MCF were still standing. Field-moist samples were sieved at 4 mm to remove coarse plant debris, and stored for two months at 4°C until the setup of the labeling experiment. A small portion of each sample was air-dried and sieved at 2 mm before chemical analysis. The selected field treatments did not significantly affect soil pH (4.9, measured in H2O), CEC (6.4 cmolc kg–1), and base saturation (64%). Total C and N as well as microbial C, N, and P were significantly greater in MCF than in COM, while Presin was less and total P (Ptot) was similar in both crop rotations (Table 2).

Production of Phosphorus-33 Labeled Crotalaria Residues
Presoaked crotalaria seeds (provenance: Maseno, Kenya) germinated on filter paper at 33°C in the dark were transferred to a container with 22 L of a 25% nutrient solution after Hoagland and Arnon (1938) but without P. The container was then placed in a climate chamber (day/night 28/18°C, 14/10 h, 65% relative humidity, light intensity {approx} 500 µmol s–1 m–2). Six days after germination, 48 seedlings were transferred to a container with aerated 25% P-free Hoagland solution, where they received a total of 68.1 mg P and 185 MBq applied with a 33P-labeled 1 M KH2PO4 solution and were grown for 22 d. At harvest, the shoots were cut and the roots washed twice in double distilled water (ddH2O) before drying for 48 h at 70°C. Shoots and roots were crushed together through a 1-mm sieve. The stalks remaining on the sieve (approx. 8 g) were ball-milled for 1 min and pooled with the crushed parts.

The concentration of Ptot and the radioactivity in the plant residue were determined after dry combustion at 550°C and dissolution of the ash in 2 mL of 20% HCl (n = 6). Similar to the determination of Presin and Phex in soil (see below), the amount of P and radioactivity extracted from the plant residue by anion-exchange resin membranes (no. 55164, 31 by 20 mm, BDH Limited, Poole, UK) were determined in the presence (Phex) and absence (Presin) of 1 mL of hexanol, respectively (n = 4). The concentration of Pi in all extracts was determined colorimetrically (Murphy and Riley, 1962), while the radioactivity was determined using a liquid scintillation counter (2500 TR, Packard Bioscience Company, Meriden, CT, USA) with 5-mL Packard Ultima Gold scintillation liquid per 1 mL of sample. To achieve complete recovery of standard additions of 33P, neutralization with NaOH was required in the case of acid extracts. Ahead of the 33P-labeled crotalaria residues, unlabeled residues had been produced in the same way to determine nutrient concentrations for the setup of the experiment. Ball-milled subsamples of unlabeled residues were used to determine total C and N contents with a CN analyzer (Carlo Erba Instruments, NA 1500, Rodano-Milano, Italy) and the concentration of K, S, Ca, Mg, Cl, and micronutrients with an x-ray fluorescence spectrometer (X-Lab2000, Spectro Analytical Instruments, Kleve, Germany). Although all growth conditions were kept identical, dry matter production was higher for the unlabeled (1.0 g plant–1) than for the labeled residues (0.9 g plant–1), resulting in lower P concentrations in the unlabeled (1.56 mg P g–1) than in the labeled residues (1.81 mg P g–1). Total N concentration in unlabeled residues was 30 mg g–1, but could not be determined for the labeled residues, as all labeled material was used in the experiment.

Table 3 shows some properties of the labeled crotalaria residues. The SA was greater in Presin than in Ptot. Apparently, the 33P taken up by the plants moved preferentially into Presin, suggesting that homogenous labeling was not achieved. Neither the P concentration nor the SA in Presin were affected by the presence of hexanol during extraction. Thus, the addition of the plant residue did not represent a source of error for the determination of microbial P by hexanol fumigation of soils.


View this table:
[in this window]
[in a new window]
  		Table 3. Concentration and specific activity of total and resin-extractable P (in the absence or presence of hexanol) in 33P-labeled crotalaria residues.{dagger}

 
Soil Amendment and Conditions of Incubation
Equal amounts of soil from each of the four field replications of the same crop rotation were thoroughly mixed to give a composite sample and preconditioned at 25% water content (60% water holding capacity) at 25°C for 12 d in the dark. At the start of the experiment, soils were amended and labeled in portions of 2.1 kg soil (dry weight equivalent) by manual mixing for 15 min. The amount of plant residue added (3.3 g DM kg–1 soil) corresponds to a biomass of 5 Mg DM ha–1 incorporated into the 0- to 15-cm soil layer. This constitutes a realistic amount of fallow biomass and the common tillage depth in western Kenya. The Pi was added in 7 mL kg–1 soil of a 33P-labeled solution of phosphoric acid (Titrisol). Soil IEP was labeled by mixing 7 mL of a solution containing 1.47 MBq 33P mL–1 into each kg of soil without the addition of a 31P carrier (Fardeau, 1996).

The added Pi treatment was adjusted to add the same amount of P as with the plant residue (Table 1). Both nonresidue-amended treatments (added Pi, soil IEP) received a nutrient solution containing the nutrients added with the crotalaria residue. This resulted in the following nutrient additions (in mg kg–1 soil): 49.5 N, 71.7 K, 6.3 S, 40.2 Ca, 9.8 Mg, 17.5 Cl, 0.1 Mn, 0.07 Cu, 0.3 Zn, 0.01 Mo, and 0.25 Bo. As only a part of the total N (Ntot) of the plant residue was expected to be mineralized during the experiment, half the amount of Ntot added with the residue (99.0 mg N kg–1 soil) was applied to the nonresidue-amended treatments (as NH+4 and NO–3 at a ratio of 1:4).

Of the amended soils, portions equivalent to 300 g dry weight per treatment were filled in 1-L plastic containers with a perforated lid. Subsamples were removed at each sampling date. Containers were placed in a dark box in the climate chamber where the same conditions were maintained as for the production of crotalaria residues. Water loss from incubated samples was adjusted gravimetrically weekly. The remaining portions of the amended soils were used for the determination of soil respiration and in a pot experiment (Bünemann, 2003).

General Calculations for the Phosphorus-33 Data
As the total applied radioactivity differed between the P sources (Table 1), the recovery of 33P (in %) in a given pool (Presin, Phex, NaOH-P) is calculated as

[1]
where r and R are the radioactivity (in kBq kg–1 soil) recovered in the pool and the total applied radioactivity, respectively (Fardeau, 1996). Likewise, SAs [33P/31P, in (r/R)/mg P kg–1] are presented as relative specific activities (rel. SA) according to

[2]
where QP is the amount of P (in mg P kg–1 soil) in a given pool.

The amount of P in a given pool derived from a labeled P addition (q, in mg P kg–1 soil), i.e., from added Pi or plant residue P, is calculated as

[3]
where SApool and SAPadd are the specific activities (in kBq mg P–1) of the pool and the labeled P addition, respectively (Fardeau, 1996). As soil IEP was labeled carrier-free, no SAadd was available in this treatment to calculate q.

Determination of Phosphorus-31 and Phosphorus-33 in Resin-Extractable and Microbial Phosphorus
On Days 1, 10, 21, 30, 42, 51, and 63 after soil amendment, the amount of 31P and the recovery of 33P in Presin and Phex were determined. The method by Kouno et al. (1995) with simultaneous liquid fumigation and extraction with anion-exchange resin membranes (BDH no. 55164, 3.1 cm x 2 cm) in bicarbonate form was followed, except for using hexanol as the liquid fumigant instead of chloroform which was found to dissolve these anion-exchange membranes. Fumigation with hexanol has been shown to be as effective as chloroform fumigation to release microbial P (McLaughlin et al., 1986). All analyses were performed with four replications. Briefly, moist soil equivalent to 2 g dry weight was shaken horizontally with 30 mL of ddH2O and two resin membranes with (fumigated) or without (nonfumigated) 1 mL of hexanol for 16 h at 170 reciprocations min–1. The resin membranes were then rinsed carefully with ddH2O and eluted with 20 mL 0.5 M HCl. The amount of 31P extracted from nonfumigated samples equals Presin. To calculate Phex, the amount of 31P extracted from fumigated samples in addition to that extracted from nonfumigated samples has to be corrected for sorption of microbial P released during the extraction period, as determined with additional nonfumigated subsamples receiving a known addition of inorganic 31P. The recovery of added 31P is described by a linear function in these soils, at least for additions of as much as 50 mg P kg–1 soil (data not shown). A single 31P spike is therefore sufficient to correct for P sorption.

Microbial 31P (31Phex, in mg P kg–1 soil) is then calculated as

[4]
where 31Pfum and 31Presin are the amounts of 31P (in mg P kg–1 soil) extracted from fumigated and nonfumigated subsamples, respectively, and 31Prec is the fraction of the 31P spike that is recovered. Averaged across all dates, 31Prec amounted to 0.60 and 0.64 for COM and MCF, respectively (significantly different between soils at P < 0.001, but showing no significant effect of the P sources).

Similar to 31P, the difference in 33P recovered in Pfum and Presin has to be corrected for sorption of 33P, as determined with a 33P spike. However, the amount of 31P released from the microbial biomass has to be taken into account as well, as it affects the recovery of 33P from the previously added labeled P sources due to sorption or desorption and isotopic exchange reactions (McLaughlin et al., 1988a). In our study, the addition of a 31P spike to 33P-labeled soils increased the recovery of 33P in Presin, but in contrast to the results of Oehl et al. (2001), pretests did not reveal a strictly linear relationship. The added amount of 31P was therefore kept close to the amount of 31Phex (SD ± 0.5 mg P kg–1 soil) as determined in a preliminary incubation. At the same amount of 31P added, the recovery of 33P was then linear across a large range of SAs in the 33P spike, as was observed by Oehl et al. (2001). Thus, only one 33P spike was required which contained the same amount of 31P as the 31P spike.

The recovery of 33P in Phex (33Phex, in %) is calculated as

[5]
where 33Pfum and 33Presin+P represent the recovery of 33P (in % of R, Eq. [1]) in fumigated and 31P-spiked (nonfumigated) subsamples, respectively, and 33Prec is the fraction of the 33P spike that is recovered. Averaged across all dates, 33Prec amounted to 0.54 and 0.60 for COM and MCF, respectively (significantly different between soils at P < 0.001), with 33Prec being significantly less than 31Prec on four out of seven sampling dates. The P sources had a significant effect on 33Prec (P = 0.039), which averaged 0.58, 0.57, and 0.55 for the treatments plant residue amendment, added Pi, and soil IEP, respectively.

In total, 15 subsamples of soil (equal to 2 g of dry weight) were prepared to determine 31P and 33P in Phex, with four replicates each for Pfum, Presin, and the 31P spike, and three that received a 33P spike, respectively.

Sequential Phosphorus Fractionation
On Day 10, 30, and 51, a separate set of samples (n = 4) was prepared to determine the fate of 33P beyond Pfum by extraction with NaOH. The first extraction was similar to the treatment of samples for Pfum in the method for Phex, but using half the amount of soil, ddH2O, hexanol, and one resin membrane in a 50-mL centrifuge tube. After shaking, the resin membrane was removed and the 10 mL ddH2O used to rinse it were added to the sample. The volume of water in the tube was adjusted to 29 mL by weight and 1 mL 3 M NaOH was added, resulting in 30 mL of 0.1 M NaOH. The next steps followed the protocol of Tiessen and Moir (1993): After shaking overnight, the samples were centrifuged for 10 min at 25000 x g and the supernatant vacuum-filtered through a millipore filter (cellulose acetate, pore size 0.2 µm, Sartorius AG, Göttingen, Germany). For Pi determination, an aliquot of 5 mL was acidified to pH 1.5 by adding 0.6 mL of 1 M H2SO4, kept at 4°C for 30 min and centrifuged to separate the precipitated organic matter from the clear supernatant. For Pt determination, another 2 mL were digested on a hot plate with 0.7 g K2S2O8, 1 mL of 11 M H2SO4, and ddH2O as required until the solution was clear. The concentration of Pi in the supernatant (NaOH-Pi) and in the digested NaOH extract (NaOH-Pt) were measured colorimetrically after neutralization (Murphy and Riley, 1962), and NaOH-Po was calculated as the difference between NaOH-Pt and NaOH-Pi.

The total radioactivity extracted with NaOH was determined by liquid scintillation as described above, but after dilution of the colored extracts with ddH2O (1:4). A small color quenching effect (4–6%) as revealed by standard additions of 33P to the extract was corrected for. We tested three methods for the separation of 33Pi and 33Po in the NaOH-extract: use of acidified molybdate and isobutanol (Jayachandran et al., 1992), counting of 33P in the supernatant obtained by acidification-centrifugation for colorimetric Pi determination (Tiessen and Moir, 1993), with the difference to 33P in NaOH-Pt representing 33Po, and recovery of 33Pi with resin membranes. Using the isobutanol method (Jayachandran et al., 1992), brown streaks of humic material in the extract were observed to move into both the aqueous and the organic phase, suggesting that the separation of Pi and Po was not complete. This method was therefore not used in our study. During acidification-centrifugation, Pi may precipitate along with organic matter, while on the other hand some organic materials (e.g., fulvic acids) remain in the supernatant (Tiessen and Moir, 1993). Therefore, the third method was tried, using the abovementioned anion-exchange resin membranes to extract 33Pi from the NaOH extract. Ten resin membranes (saturated with HCO–3) were added to 5 mL NaOH extract diluted with 40 mL ddH2O and placed on an overhead shaker for 24 h. The resin membranes were rinsed briefly with ddH2O and eluted with 0.5 M HCl. On each date, 83 to 84% of 31P and 33P from standard additions to selected samples (n = 3, CV 1–3%) were recovered on the resin membranes. The percentage of 33P in the NaOH extract recovered on resin membranes (NaOH-Presin) is presented after correction for this incomplete recovery.

Statistical Analysis
Statistical analysis was performed with SYSTAT (2000, SPSS Inc., Chicago). A t test was performed to check for differences in soil properties between crop rotations in the field samples. For each sampling date of the incubation experiment, data were tested by two-way ANOVA with the factors P source, crop rotation, and the P source x crop rotation interaction. Data were also subjected to three-way ANOVA with the factors P source, crop rotation, sampling date, and all possible interactions. Multiple comparisons using Tukey's test were performed whenever the ANOVA indicated significant differences (P ≤ 0.05).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Resin-Extractable Phosphorus
On each sampling date, the amount of Presin was significantly greater after addition of Pi than after plant residue amendment or carrier-free labeling of soil IEP (Fig. 1a) . The greatest difference was observed on Day 1, when the amount of Presin was 1.6 to 1.8 mg P kg–1 soil greater after addition of Pi than in the other two treatments (1.0–1.5 mg P kg–1 soil). Thereafter, Presin decreased in all cases. Throughout the incubation, Presin was significantly greater in COM than in MCF, on average by 0.3 mg P kg–1 soil.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 1. Amount of P, recovery of 33P and relative specific activity (rel. SA) in (a–c) resin-extractable phosphorus (Presin) and (d–f) microbial phosphorus (Phex) after addition of 6 mg P kg–1 soil as plant residue or inorganic phosphorus (added Pi), both labeled with 33P, or after carrier-free labeling of isotopically exchangeable phosphorus (soil IEP) in soils with continuous maize (COM) and maize-crotalaria fallow rotation (MCF); bars show Tukey's honestly significant difference (0.05).

 
On Day 1, the recovery of 33P in Presin (Fig. 1b) was significantly greater after addition of Pi (22%) than after carrier-free labeling of soil IEP (12%) and plant residue amendment (7%). Thereafter, the recovery of 33P in Presin declined steadily in all cases. On Day 63, it still followed the treatment order added Pi (3.8%) > soil IEP (2.6%) > plant residue (1.8%). Crop rotation had a minor effect on the recovery of 33P in Presin, which was significantly greater in COM than in MCF from Day 21 onwards. Figure 1c shows the relative SAs in Presin, which were generally lower after plant residue amendment than for the other two P sources, did not differ between crop rotations, and decreased throughout the incubation.

Microbial Phosphorus
Amounts of Phex were always greater after plant residue amendment than for the other two P sources, with a difference of 1.4 to 2.9 mg P kg–1 soil (Fig. 1d). The Phex was on average 2.7 mg P kg–1 soil greater in MCF than in COM, and no clear time trend was observed. As much as 18% of applied 33P was recovered in Phex (Fig. 1e), with the recovery generally ranging in the order plant residue (15%) > soil IEP (7%) ≥ added Pi (4%). The percentage of 33P in Phex was on average 2.4% greater in MCF than in COM, and it did not change significantly during the course of the experiment.

Relative SAs in Phex were generally greater after residue amendment than for the other two P sources, showing no significant effect of crop rotation or time (Fig. 1f). Figure 2 shows the calculated amount of Phex derived from added Pi or from the plant residue. In soil from both rotations, on average 0.2 mg P kg–1 soil (5.5%) of Phex was derived from added Pi, compared with 0.9 mg P kg–1 soil (15.8%) of Phex from the added plant residue (Eq. [3]). As this was less than the absolute increase in Phex after plant residue amendment, additional uptake of soil P (0.6–2.1 mg P kg–1 soil) was implied, with a tendency to decrease during the course of the incubation.



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 2. Source of microbial phosphorus (Phex) after plant residue amendment or addition of inorganic phosphorus (added Pi) in (a) soil from crop rotation continuous maize (COM) and (b) soil from crop rotation maize-crotalaria fallow rotation (MCF); error bars show SD of Phex in the respective treatment; dotted line indicates average level of Phex after carrier-free labeling of isotopically exchangeable soil P.

 
Sequential Phosphorus Fractionation
In the separate set of samples for sequential P fractionation, similar amounts of P and percentages of 33P were extracted in Pfum as during the determination of Phex. The total amount of P extracted in the subsequent extraction with NaOH averaged 284 mg P kg–1 soil and was not affected by P source or crop rotation. The distribution of NaOH-Pt into Pi and Po, however, differed slightly between P sources and more clearly between crop rotations, with a shift toward Po observed in MCF (Table 4).


View this table:
[in this window]
[in a new window]
  		Table 4. Main treatment effects on the amount of P extracted during the sequential fractionation and significance of treatment interactions.{dagger}

 
Figure 3 summarizes the distribution of 33P across the extracted pools at the three sampling dates, additionally showing the recovery in Presin as determined during the measurement of Phex. The total recovery of 33P in the two sequential extraction steps was greater after plant residue amendment (90%) than after addition of Pi (76%) or after carrier-free labeling of soil IEP (80%), but was not affected by crop rotation or time of incubation. The recovery of 33P in NaOH-Pi (64-68%) did not differ between P sources. The greater total recovery of 33P after plant residue amendment was because of the greater recovery in Pfum and especially in NaOH-Po.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 3. Recovery of 33P during sequential extraction of soils continuous maize (COM) and maize-crotalaria fallow rotation (MCF) after addition of 6 mg P kg–1 soil as plant residue or inorganic phosphorus (added Pi), both labeled with 33P, or after carrier-free labeling of isotopically exchangeable phosphorus (soil IEP). Resin-extractable phosphorus (Presin) determined separately; phosphorus extracted with resin membranes in the presence of hexanol (Pfum) and inorganic and organic phosphorus extracted with 0.1 M NaOH separated by acidification-centrifugation (NaOH-Pi and -Po) extracted sequentially. Bars show SD of the sum of 33P recovered in Pfum, NaOH-Pi, and Na-OH-Po.

 
The recovery of 33P in Pfum tended to decrease with time, mainly because of the decrease in Presin. Except for residue-amended soils, an increase in 33P recovered in NaOH-Po was observed during the incubation, although the recovery of 33P in NaOH-Pt was similar on all dates (Table 5). The two methods to separate 33Pi and 33Po in the NaOH extract yielded similar results in the case of carrier-free labeling of soil IEP, while after addition of Pi, an additional 2 to 3% of 33P was recovered on the resins than in NaOH-Pi. The greatest difference between both methods was observed after plant residue amendment, where 10% more 33P was recovered in NaOH-Pi than in NaOH-Presin.


View this table:
[in this window]
[in a new window]
  		Table 5. Main treatment effects on the total recovery of 33P in the NaOH extract (NaOH-Pt), and percentage thereof recovered in inorganic P as determined by acidification-centrifugation (NaOH-Pi) or by extraction with anion-exchange resin membranes (NaOH-Presin).{dagger}

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Phosphorus Dynamics after Plant Residue Amendment
Net changes in soil P pools after addition of a small amount of P are often not detectable against the amount of P already present in soil because of sorption processes. With isotopic labeling techniques, the distribution of added P across the various soil P pools can be followed with greater sensitivity.

In the present study, plant residue amendment resulted in a net increase in Phex, which was accompanied by the greatest recovery of 33P from a given P source in Phex (15%). In the study of McLaughlin and Alston (1986), the amount of P in the microbial biomass was also greater after the addition of a plant residue than after the addition of Pi. At the same time, they recovered 68% of the label from plant residue amendment in the microbial biomass, as compared with 23% after addition of Pi. As they used a conversion factor of 0.4 to correct for incomplete recovery of microbial P, this translates into 27 and 9% of isotope rendered extractable by fumigation. Oehl et al. (2001) observed that 2 d after carrier-free labeling of soil IEP, 66% of the tracer was recovered in chloroform-labile P when soils were amended with glucose and ammonium nitrate, compared with 8% in the absence of easily available sources of C and N. Thus, all three studies agree that the stimulation of the microbial biomass with a plant residue or a C substrate also boosts the recovery of label in this pool.

When the plant residue alone was extracted with resin membranes, 69% of 31P and 87% of 33P were recovered in Presin. Thus, the information obtained from the 33P tracer slightly overrepresents the most readily available P fraction of the plant residue. Daroub et al. (2000) also reported that the majority of 33P in labeled soybean residues (70%) was resin-extractable. For the comparison of P added with a plant residue or as Pi, it is therefore mainly the presence of C in the plant residue that affects the recovery of 33P not only in Phex, but also in Presin. This is additionally supported by the fact that soil respiration was not affected by addition of Pi, whereas an additional 800 mg C kg–1 soil were released as CO2 after plant residue amendment during the incubation period (Bünemann, 2003).

In our study, the calculated amount of Phex derived from the plant residue was less than the net increase in Phex, suggesting microbial uptake of as much as 2.1 mg P kg–1 soil or 39.5% of Phex from soil. Microbial uptake of P from soil after plant residue amendment was not reflected in a decrease in Presin, as the amount of Presin after plant residue amendment was similar to that after carrier-free labeling of soil IEP (1.0 mg P kg–1 soil). The recovery of 33P and the relative SA in Presin, however, were lower after plant residue amendment than after labeling of soil IEP. Thus, microbial depletion of Presin after plant residue amendment may have been buffered from other soil P pools with a lower SA than Presin. The relative SA in NaOH-Pi of 0.009 (SD ± 0.001) r/R (mg P kg–1)–1, for example, was always lower than that in Presin (Fig. 1c). This agrees with the decrease of SAs in Pi fractions in the order of extraction observed by Bühler et al. (2002). In our study, as much as 90% of the total applied radioactivity was recovered during the sequential extraction (Pfum and NaOH-P), while at the same time only 40% of Ptot in these soils (720 mg kg–1) was extracted, suggesting that the SA in the nonextracted P would be even lower than in P extracted with NaOH.

Phosphorus Dynamics after Addition of Inorganic Phosphorus
Addition of Pi increased Presin and NaOH-Pi without affecting the amount of Phex or microbial activity as indicated by soil respiration (data not shown). The recovery of 33P in Presin was also greatest after addition of Pi, while for NaOH-Pi it was similar for all three P sources. Annual additions of triple superphosphate of 50 kg P ha–1 in the field experiment also increased available inorganic, but not Phex and Po (Bünemann et al., 2004b). Bühler et al. (2002) also observed that P added with mineral fertilizer moved into inorganic pools, with NaOH-Pi being the main sink. A review of studies from tropical agroecosystems found considerable transformation of fertilizer P into Po only for systems with great availability of carbon substrates, such as pastures and agroforestry systems (Nziguheba and Bünemann, 2004).

In the present study, the recovery of 33P in Phex was even less from added Pi than after carrier-free labeling of soil IEP (4 vs. 7%, P < 0.001), although the amount of Phex and soil respiration were similar in both treatments. A possible explanation could be that 1 d after soil amendment, the relative SA of Presin was lower after addition of Pi than after labeling of soil IEP (0.07 vs. 0.10, P = 0.002). If the SA of Presin is similar (or directly related) to the SA of the source of microbial P uptake, this may explain the lower recovery of 33P in Phex after addition of Pi. From Day 10 onward, however, the relative SA in Presin was similar in both treatments.

Comparison of Soils from the Two Crop Rotations
In agreement with the shift from Pi to Po observed under maize-fallow rotations compared with COM (Bünemann et al., 2004b), the amounts of Phex and NaOH-Po were greater in MCF than in COM. The SA in Phex was generally similar for both crop rotations, indicating that the recovery of 33P in Phex was proportional to pool size. This differs from the observations made by Oberson et al. (2001). In their study, the recovery of label in the microbial biomass was also greater in soils from land-use systems that increased the microbial biomass, but the resulting SAs in microbial P differed by a factor of three to eight, with higher SA in systems with higher microbial biomass. In contrast to a previous study (Bünemann et al., 2004a), neither the amount of CO2 released from residue-amended in addition to nonamended soils (data not shown) nor the net change in Phex after plant residue amendment differed between COM and MCF. Apparently, the microbial biomass in soils from both crop rotations was able to respond similarly to the smaller amounts of plant residue added in the present study.

Both methods used to separate 33P-labeled Pi and Po in the NaOH extract indicated that for nonresidue-amended soils, a smaller proportion of 33P in the NaOH extract was recovered in the inorganic fraction in MCF than in COM (Table 5). In the absence of plant residue addition, 33P-labeled Po can only be of microbial origin, and the greater proportion of 33Po in MCF is in agreement with the greater microbial activity in MCF. The trend for a greater proportion of 33P in the organic fraction after labeling of soil IEP than after addition of Pi also corresponds to the greater recovery of 33P in Phex from soil IEP than from added Pi. The recovery of microbial P during the fumigation-extraction step may be incomplete not only because of P sorption, but also because of incomplete lysis and extraction of microbial cells as well as incomplete hydrolysis of microbial Po. Alternatively, 33P-labeled Po excreted from microorganisms may already have become stabilized on soil surfaces. In any case, the sequential extraction after fumigation represents a conservative estimate of label recovery in Po, as phosphatases released from lysed cells may mineralize labile Po during the extraction of Pfum, thus reducing the amount of Po in the subsequent NaOH extract.

Temporal Dynamics in the Recovery of Phosphorus-33 in Soil Phosphorus Pools
In accordance with the principles of isotopic exchange (Fardeau, 1996), the recovery of 33P in Presin diminished steadily, with the greatest decrease occurring between Days 1 and 10 (Fig. 1a). The recovery of label in Presin decreased also during incubation of 33P-labeled Oxisols from Colombia (Bühler et al., 2002) and of temperate soils amended with 33P-labeled soybean residues (Daroub et al., 2000). In these studies, a simultaneous increase in the recovery of 33P in other sequentially extracted pools (NaHCO3, NaOH) was generally observed. In contrast, the recovery of 33P in the NaOH extract in our study remained unchanged at 70 to 72% at all sampling dates. This could be because of the fact that our first sequential extraction took place after 10 d, compared with extractions performed on the day of labeling in the other two studies, and therefore did not capture the initial movement of 33P from Presin into other pools. In addition, NaOH-Pt in our study was not extracted after only Presin but after Pfum, in which the recovery of 33P decreased only from 12 to 9% between Days 10 and 51. This would not have been detectable against the variation in the much higher recovery of 33P in NaOH-Pt. The predominance of fast exchange reactions was also observed in the nonfertilized soil studied by Bühler et al. (2002), whereas in P-fertilized soils of the same mineralogy, isotopic exchange between fractions proceeded more slowly.

Throughout the study, the recovery of 33P in Phex did not change significantly for any of the P sources. In the field experiment of McLaughlin et al. (1988a), the percentage of label from the applied plant residue recovered in microbial P also remained stable during 95 d, while Kouno et al. (2002) observed a decrease in labeled microbial P between 10 and 60 d after addition of labeled ryegrass. In the absence of plant residue amendment, Oehl et al. (2001) generally observed a steady increase in 33P recovered in the microbial biomass during 9 wk after carrier-free soil labeling. The only exception was a nonfertilized soil in which the recovery remained constant after the first 5 d. Under conditions of low P availability, the SA in microbial and available P may soon be too similar to detect fluxes between the two pools (Oehl et al., 2001). Alternatively, efficient internal cycling of 33P within the microbial population may be indicated; that is, fresh microbial biomass taking up P from dying microorganisms with little P flux into other soil P pools. In the absence of plant residue amendment, it is also possible that microorganisms returned to a resting state soon after the initial activation by small amounts of substrates becoming available during stirring and mixing of soils at the onset of the experiment (de Nobili et al., 2001). In fact, daily respiration rates in the nonresidue-amended treatments were two to three times greater during the first 24 h than during the following 48 h (data not shown). Even in the case of plant residue amendment, maximum respiration rates occurred during the first 24 h, when most 33P was taken up.

Methodological Considerations
The correction of 31P and 33P extracted from fumigated samples in addition to that extracted from nonfumigated samples for P sorption by using 31P and 33P spikes is a laborious approach. At the same amount of 31P added, the average recovered fraction of the 33P spike (0.57) was less than that of the 31P spike (0.62), but the difference was nonsignificant on three out of seven dates. In addition, none of the recoveries (31P or 33P spike) showed a clear time trend. Thus, it might be sufficient to determine the recovery of a 31P spike which adds a similar amount of 31P as contained in Phex, and to use the recovery of the 31P spike also to correct for incomplete recovery of 33P. The validity of this approach needs to be tested for every specific experimental situation.

The use of anion-exchange resin membranes to separate 33Pi and 33Po in the NaOH extract is not satisfying because of the incomplete recovery of standard additions of Pi to the extract. In addition, the same type of resin membranes has also been used to extract organic anions from soil (Szmigielska et al., 1996; George et al., 2002), and sorption of Po on the membranes and subsequent elution cannot be excluded. Nevertheless, treatment effects on the recovery of 33P in NaOH-Pi (by acidification-centrifugation) and NaOH-Presin were generally similar (Table 5), except for residue-amended soils, where a smaller proportion of 33P was recovered in NaOH-Presin than in NaOH-Pi. Apparently, some 33Po remained in the supernatant during the acidification and centrifugation of the NaOH-extract, as suggested by Tiessen and Moir (1993). Other methods to separate 33P-labeled Pi and Po should be tested in future experiments, as for example, ultrafiltration after addition of a flocculant (Gerke and Jungk, 1991).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
In this incubation experiment using isotopic labeling with 33P to investigate P dynamics and the fate of added P sources in a highly weathered soil, biological processes of soil P dynamics were most important after plant residue amendment, where a shift toward microbial P and Po was observed. This shift was also confirmed by the greater importance of biological processes in soil MCF as a result of previous fallow biomass incorporation. Isotopic labeling indicated that the increase in microbial P after plant residue amendment was partly derived from available soil P, the depletion of which was buffered from other soil P pools. While the recovery of 33P in resin-extractable P decreased steadily throughout the incubation, the recovery of 33P in the microbial biomass did not change after the rapid initial uptake of as much as 17% of applied 33P within one day. Possible explanations are: that the similar SAs in available and microbial P due to the predominance of fast exchange reactions did not allow to detect fluxes between the biomass and the soil for more than a few days; that P was cycled efficiently within the microbial biomass; or that the microbial biomass returned to dormancy soon after the initial stimulation during soil amendment. The recovery of >60% of added 33P in Pi extracted with 0.1 M NaOH reflects the importance of P sorption on sesquioxides in these soils.


   ACKNOWLEDGMENTS
 
We are grateful for the help of R. Ruh with production of crotalaria residues and for the assistance of I. Jansova and T. Rösch in chemical analyses. Nutrient contents of crotalaria residues were determined with x-ray fluorescence spectroscopy by K. Barmettler at the Institute of Terrestrial Ecology (ETH Zurich). This publication results from a joint project between the International Centre for Research in Agroforestry (ICRAF) and the ETH Zurich, funded by the Swiss Centre for International Agriculture (ZIL).

Received for publication December 8, 2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 





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 ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bünemann, E. K.
Right arrow Articles by Oberson, A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bünemann, E. K.
Right arrow Articles by Oberson, A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Bünemann, E. K.