| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Plant & Soil Science, Hills Building, University of Vermont, Burlington, VT 05405 USA
fmagdoff{at}zoo.uvm.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Abbreviations: BK1, Bray and Kurtz 1 extractant • BK2, Bray and Kurtz 2 extractant • EPC0, equilibrium P concentration • ICP, inductively coupled plasma • M3, Mehlich 3 extractant • OL, Olsen extractant • pHs, pH measured in 0.01 M CaCl2 • pHNaF, pH measured in 1 M NaF • STR, iron oxide impregnated strip • VT1, Vermont 1 extractant • VT2, Vermont 2 extractant
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
A few studies have compared a number of tests on different soils, including those from various regions of the USA (Wolf et al., 1985; Sharpley and Smith, 1992). These studies have shown that soil texture and extractable Al and/or Fe are useful in explaining relationships between the quantity of extracted P and estimated P availability. In noncalcareous soils Al and Fe play important roles in governing the availability of recently applied P and native soil P (Lee and Bartlett, 1977; Sample et al., 1980; Sanchez and Uehara, 1980; Wolf et al., 1985; Sharpley and Smith, 1992; Beauchemin et al., 1996). While most of the research on the influence of reactive Al on P adsorption has used oxalate to extract Al (for example, Wolf et al., 1985; Beauchemin et al., 1996), routine soil test solutions containing either acetate or F may also extract Al from the fraction that reacts with P and influences P availability (Lee and Bartlett, 1977; Tran et al., 1990; Piha, 1993).
At this time, Vermont is the only state to use soil test–extractable Al to modify P fertilizer recommendations (Bartlett, 1982). In this system, when the VT1P soil test value is below the critical level, the amount of P recommended increases as the amount of VT1Al increases. A similar system was established recently in Quebec for modifying P recommendations based on the Mehlich 3 extractant (Conseil des productions vègètales, 1996). Beauchemin et al. (1996) found that factoring in an evaluation of reactive Al improved the relationship between soil tests and estimates of the potential of soil P pollution.
The Morgan and modified Morgan (pH 4.8 Na or NH4 acetate) extractants have been omitted from most of the recent studies that compare soil tests for evaluating plant availability or pollution potential (e.g., Wolf et al., 1985; Beegle and Oravec, 1990; Fixen and Grove, 1990; Sharpley and Smith, 1992), probably because these extractants are only used in few states. However, acetate-extractable P may represent an important parameter, the fraction of the P sorption capacity that already has adsorbed P (Kuo, 1990).
Four laboratory and greenhouse experiments were conducted over a period of 10 yr, using four different sets of soils. Although different extractants were sometimes used, all experiments were aimed at gaining a better understanding of the relationship of different extractants to plant available P and the potential for loss of P in runoff. As part of the overall objective, with individual experiments we evaluated the following: (i) the relationships of the amounts of P and Al extracted by various soil test solutions, especially those with and without F; (ii) changes in P adsorption and extractable-P and -Al following P addition to soils, and (iii) the relationship of P extracted by VT1- and F-containing extractants to plant P availability and potential P loss to field runoff waters.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Experiment I
Samples representing 31 (numbers 1–31) were collected from ten Vermont counties. Soils were selected to represent a variety of properties, such as soil texture and extractable P and Al, and were stored field moist in plastic bags in the dark at 4°C for approximately 1 mo before analysis.
Soil samples were passed through a 2-mm sieve, dried at 55°C, and extracted with the following solutions: (i) 1.25 M NH4OAc at pH 4.8 with a 1:5 soil:solution ratio (VT1, modified Morgan solution, McIntosh, 1969); (ii) 1.25 M NH4OAc and 0.03 M NH4F at pH 4.8 with a 1:10 soil:solution ratio (VT2, McIntosh, 1969); (iii) 0.1 M HCl and 0.03 M NH4F with a 1:10 soil: solution ratio (Bray and Kurtz 2, Bray and Kurtz, 1945); and (iii) 0.2 M CH3COOH, 0.25 M NH4NO3, 0.015 M NH4F, and 0.001 M ethylenediaminetetraacetic acid with a 1:10 soil:solution ratio by volume (M3, Mehlich, 1984).
Phosphorus was analyzed colorimetrically using the molybdate–stannous chloride method. Calcium and aluminum were determined by inductively coupled plasma (ICP). Soil pH was measured in 0.01 M CaCl2 (pHs) and 1 M NaF (pHNaF). The pHs was determined 15 min after adding 5 mL soil to 10 mL CaCl2 solution, and pHNaF was determined 1 h after adding 5 mL soil to 12 mL NaF solution.
Plastic greenhouse pots were filled with 0.6 kg (dry weight basis) of each moist soil. Potassium and magnesium were added per soil test recommendation. Phosphorus was applied to four replicate pots of each soil as 1 M KH2PO4 · K2HPO4 buffer at the rate of either 0 or 18 mg P kg-1. Soil from each pot was thoroughly mixed and 10 to 12 inoculated alfalfa seeds (Medicago sativa [L.], var LD 88) were planted in each pot. Plant population was thinned to five per pot 5 d after emergence. Pots were arranged in a randomized complete block design on a greenhouse bench. Plants were harvested at 56 and 84 d after planting, dried at 55°C, and weighed.
The soils had a range of properties including 60 to 680 mg clay kg-1 (by hydrometer, following organic matter oxidation by H2O2), pHs 5.0 to 7.0, and effective cation exchange capacity (sum of Ca + Mg + K in VT1) from 3.6 to 38.1 cmolc kg-1.
Experiment II
Samples representing 54 soils (numbers 32 to 85) were collected from the Lake Champlain valley of Vermont and New York. Soils were selected to represent a variety of textures and P levels and stored field moist in plastic bags in the dark at 4°C for 
1 mo. Samples were sieved, dried, extracted, and analyzed for P in the VT1 and VT2 solutions as described above, and the same was done for Al in the VT1.
Experiment III
Samples representing 24 soils (numbers 86 to 109) were collected from the Lake Champlain basin of Vermont and New York. The range of selected soil properties included 40 to 720 mg clay kg-1, pHs from 5.1 to 7.4, and effective cation exchange capacity from 1.0 to 35.5 cmolc kg-1, and VT1P 0.6 to 56.5 mg kg-1. The samples were stored moist in plastic bags in the dark at 4°C for 1 mo. Subsamples were subsequently passed through a 4-mm sieve and refrigerated in small plastic bags at field moisture. Soils were dried at 55°C, passed through a 2-mm sieve, extracted with the VT1 and VT2 buffers, and analyzed as described above. Phosphorus was also determined by the STR method (Sharpley, 1993). Soils were analyzed for M3P by the Soil Testing Laboratory of the North Carolina Department of Agriculture's Agronomic Division, and for BK1P (0.025 M in 0.03 M NH4F, 1:10 soil:solution ratio) and OLP (0.5 M NaHCO3, 1:25 soil:solution ratio) by the University of Nebraska Soil Testing Laboratory.
The texture of each soil was determined by the hydrometer method following organic matter destruction with hydrogen peroxide. All soils were analyzed for pHs, P in VT1 and VT2 using the molybdate–stannous chloride method, cations (K, Mg, Al, Ca) in VT1 by ICP, and organic matter by weight loss on ignition for 2 h at 375°C in a forced air oven.
Experiment IIIa
Three rates of P, 0, 20, or 40 mg P kg-1 as Ca(H2PO4)2, were added to 500 g soil on a dry basis. An additional treatment consisted of 40 mg P kg-1 as liquid manure added to 500 g (dry basis) soil. Samples were thoroughly mixed, placed in a plastic bag, moistened with distilled water to approximate field capacity, and stored in the dark at room temperature (26°C) with water added periodically to bring soils to close to field capacity. Soils were thoroughly mixed and subsamples removed for analysis at 7, 30, 90, 180, and 360 d. Each subsample was analyzed for P and Al in VT1, P in VT2, and pHs. Calcium chloride–extractable P was determined using a 1:25 soil:0.01M CaCl2 solution ratio, shaking for 16 h, followed by centrifugation, and P analysis by the molybdate–stannous chloride method. Phosphorus desorption or adsorption was determined at each sampling time. One gram (dry basis) of sample from each treatment was placed in each of six test tubes, 25 mL of P solution (either 0, 0.2, 0.5, 1, 5, or 10 mg P L-1 as KH2PO4) was added. The tubes were shaken for 16 h, centrifuged, filtered, and the supernatant analyzed for P using the molybdate blue–stannous chloride method. The amount of adsorbed P was calculated by subtracting the P recovered in solution from the amount added. The solution concentration at which P is neither adsorbed nor desorbed from soils, or EPC0, was estimated by using the equation P adsorbed = a + b(log P solution concentration).
Experiment IIIb
Twelve of the original 24 soils used in Exp. IIIa were selected to represent a range of properties: pH 5.1 to 7.2, 40 to 720 mg clay kg-1, effective cation exchange capacity from 1.0 to 35.5 cmolc kg-1, and VT1P 0.6 to 56.5 mg kg-1. Samples were incubated with 0, 80, 160, 480, and 1920 mg P kg-1 and analyzed after 100 d for P and Al in the VT1 extract, and P in the VT2 extract. The incubation conditions and sampling procedures were the same as those described for Exp. IIIa.
Experiment IV
Fifty-four agricultural soil samples (numbers 110 to 163) submitted to the University of Vermont Agricultural and Environmental Testing Laboratory were selected at random from among those received over a 2-wk period during 1996. Aluminum and phosphorus were determined in VT1 extracts and P alone in VT2 extracts.
Two subsamples of 50 g each were placed in individual plastic bags and moistened with distilled water to approximate field capacity. Soil in one bag received no added P while the other received 40 mg P kg-1 soil as K2HPO4. All bags were stored in the dark at room temperature (26°C) with distilled water added periodically to maintain approximate field capacity. At 100 d, P and Al were determined using VT1 and VT2 extractants.
Calcium chloride (0.01 M)–extractable P and distilled water–extractable P were determined using a 1:25 soil:solution ratio, shaken for 16 h, followed by centrifugation, and P was analyzed by the molybdate–stannous chloride method.
Regression equations and correlation coefficients were calculated using Statview for the Macintosh (Abacus Concepts, 1991). The test of the equality of the regression slopes was performed using the analysis of covariance and the type III sum of squares in the generalized linear model in the Statistical Analysis System (SAS, 1990). To satisfy the assumptions of equality of variances we analyzed the data on the log scale.
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
0.93 (Table 1). Iron oxide on the strip apparently competes with or enters into an equilibrium relationship with Fe and Al on soil mineral surfaces and in organic chelates. Vermont 1 P and OLP are well correlated with the strip method, although the correlations are slightly lower than between P extracted with F and the STR method. The correlations between VT1P and OLP with most other tests were improved by omitting the three soils highest in available P (VT1P > 30 mg kg-1).
|
There are two independent indications that the VT1 acetate solution extracts Al from the same pool extracted by F. Vermont 1 Al is strongly related to the greater amounts of Al extracted by M3 or VT2 extractants that contain F (Fig. 1a and b)
. Aluminum extracted by Bray and Kurtz 2 (BK2) was even greater than that extracted by M3 or VT2, ranging from 1635 to 4148 mg kg-1, and it was related to VT1Al by the equation
. Thus, although solutions containing F extract more Al than similar extractants without F, Al extracted by acetate is related to the quantity extracted by F.
|
is related to the amount of VT1Al (Fig. 2)
. The high correlation between Al extracted by acetate and F (Fig. 1) indicates that a good relationship may also exist between 
pH and Al extracted by VT2P, M3, BK1, and BK2.
|
Relationship of VT1Al, VT1P, and VT2P
McIntosh (1969) noted that when soils of similar pH and texture were compared, there was a good correlation between VT1P and VT2P. Our data showed a similar relationship for soils separated by levels of reactive Al. Almost all soils fit into a linear relationship between VT1P and the amount of VT2P relative to VT1Al (Table 1 for the 24 soils used in Exp. III and Fig. 3
for all 163 soils). This relationship was selected for evaluation because the Vermont soil test system includes determination of all three parameters. With the highest VT1P soil of the 163 soils omitted from the regression, the equation becomes
,
. The r2 for the equation in Fig. 3 is higher than the r2 for the relationship between VT1P and VT2P; i.e.,
. The quantity of P extracted by acetate (VT1P), then, is an index that represents the extent of P saturation of the reactive Al extracted by VT1.
|
A similar relationship also exists between VT1P and the P and Al extracted in the M3 and BK2 extracts in Exp. I. The r2 for the relationships between VT1P and M3P/M3Al or BK2P/BK2Al (r2 of 0.94 and 0.43, respectively) are considerably greater than between VT1P and M3P or BK2P (r2 of 0.65 and 0.21, respectively).
Soil Chemical Changes Following P Additions
The increase in NH4OAc-extractable P (VT1P) per unit of P added decreased as VT1-extractable Al increased (Fig. 4)
. Thus, with increasing reactive Al, more P must be added to increase VT1P by a given amount. Because VT1P represents the fraction of reactive Al saturated by P (Fig. 3), an increase in VT1P implies an increase in the amount of P held by Al. Although researchers such as Cox (1994) have found clay content to be related to the change in P soil test as P is applied, for our soils clay content was not significantly related to soil test change with added P. The explanation for the different results may involve the relatively high organic matter levels and predominantly 2:1 clays in our soils compared with low organic matter soils with predominantly 1:1 and sesquioxide clays of soils in the Southeast. Reactive Al in the soils of Vermont appears to be mainly complexed with organic matter (Bartlett, 1982) rather than associated with mineral surfaces.
|
3% of the VT2P increase for the soil with the highest Al. This is another indication of the importance of reactive Al, as estimated by the amount extracted by VT1, to soil P dynamics.
|
The quantity of P adsorbed by a soil from a solution is an indicator of its capacity to fix P added as fertilizer. The range in P adsorption is indicated by a selection of soils from Exp. IIIa (Fig. 6)
. Although the relationship between P adsorption and P solution concentration is not usually linear, the slopes are proportional to a soil's adsorption capacity. If a soil such as no. 102, with a high VT1Al (194 mg kg-1) and low VT1P (1.2 mg kg-1) adsorbs close to 100% of the P in solution, even when the initial solution P is as high as 10 mg P L-1, then the steep slope would indicate that the soil is a strong P adsorber. Soil no. 90, with a low VT1Al (15 mg kg-1) and high VT1P (56.5 mg kg-1), desorbed P until the solution was >3 mg P L-1 and adsorbed only 20% of the P from the 10 mg P L-1 solution.
|
Phosphorus Extracted by CaCl2 and Water and Potential Losses to Runoff
Vermont-1 P is strongly related to the amount of CaCl2P for the individual treatments of the 24 soils of Exp. IIIa (Fig. 7) . However, the relationship between VT2P/VT1Al and CaCl2P (Fig. 8)
resulted in a higher r2 than for VT1P vs. CaCl2P.
|
|
. Thus, while BK1 and M3 extractions were not performed following P applications, these F-containing extractants are probably not as good as VT1 or VT2P/VT1Al as predictors of CaCl2P.
More P was extracted by distilled water than by CaCl2. The regression equations for the relationships of extracted-P vs. VT1P were as follows: water-P = 1.62 + 0.56
, r2 = 0.78; and CaCl2P = -0.50 + 0.28, r2 = 0.87. The slopes of the two lines are significantly different
. The larger amount of P extracted by distilled water than by weak salt solutions agrees with the findings of Olsen and Watanabe (1970) and Soltanpour et al. (1974).
The dissolved P limit of 1 mg P L-1 for point source discharge has also been proposed for agricultural runoff (US EPA, 1986). Using the relationship from Exp. IV for distilled water extractable P with VT1P (equation given in paragraph above), and converting water extractable P from mg kg-1 to mg L-1, leads to an estimate of 42 mg VT1P kg-1 at 1 mg P L-1 in the water extract. It may take 3 mg P L-1 in a distilled water extract to be equivalent to 1 mg P L-1 in actual runoff (estimated from data in Pote et al., 1996). Thus, it apparently takes extremely high soil test P levels, perhaps >120 mg VT1P kg-1, to result in runoff P concentrations >1 mg P L-1. To help put these estimates in perspective, of the approximately 2600 agricultural soils submitted to the Vermont Agricultural and Environmental Testing Laboratory in 1996, the highest individual sample level was 111 mg VT1P kg-1 and the highest 10% averaged 32.8 mg VT1P kg-1. The remaining 90% of samples were below 17 mg VT1P kg-1. This suggests that runoff water with dissolved P > 1 mg L-1 would be rare under normal conditions, unless runoff occurred soon after surface P application.
Availability to Plants
The solution P concentration when no adsorption or desorption from the soil occurs has been termed the equilibrium P concentration, or EPC0. This is the soil solution concentration, or intensity, that will initially be present as P uptake by roots begins. The EPC0 is strongly related to P extracted by CaCl2 (Fig. 9)
, as was found with distilled water extracts by Sharpley et al. (1994b). The correlation for EPC0 with VT1P and VT2P/VT1Al was as great or greater than with P extracted with VT2, BK1, and M3 solutions (Table 2)
. With the highest VT1P soil omitted, the correlation of EPC0 with VT1P and VT2P/VT1Al increased to 0.92 and 0.90, respectively, while the relationship between EPC0 with VT2P decreased to 0.36.
|
|
, Table 1) suggests that it probably would not have been highly related to P response. The critical value, the point above which there is a low probability of response, was estimated to be 4 mg VT1P kg-1 by the Cate-Nelson procedure (Cate and Nelson, 1971). This is the same critical value found in field trials in Vermont (Jokela et al., 1998).
|
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Interpretation of P availability extracted by the various F-containing solutions is enhanced by knowing the amount of reactive Al in the soil. Soil solution P appears to be governed by the extent of P saturation of the reactive Al in the soil as estimated by VT1P or VT2P/VT1Al. Compared with other extractants, VT1P is better correlated with P availability to alfalfa. Compared with solutions containing F, NH4OAc-extractable P (VT1P) was also more highly correlated with water-extractable P and the EPC0. Thus, VT1P appears to be a good predictor of plant availability, as well as a sound parameter to be included as part of an index that ranks soils according to their potentials for contribution of P to runoff.Conseil des productions vègètales du Quèbec Inc. 1996; Olsen Sommers 1982; SAS Institute 1990; EPA 1986
Received for publication May 13, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. R. Haden, Q. M. Ketterings, and J. E. Kahabka Factors Affecting Change in Soil Test Phosphorus Following Manure and Fertilizer Application Soil Sci. Soc. Am. J., June 8, 2007; 71(4): 1225 - 1232. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Heckman, W. Jokela, T. Morris, D. B. Beegle, J. T. Sims, F. J. Coale, S. Herbert, T. Griffin, B. Hoskins, J. Jemison, et al. Soil Test Calibration for Predicting Corn Response to Phosphorus in the Northeast USA Agron. J., February 7, 2006; 98(2): 280 - 288. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. H. Anderson and F. R. Magdoff Relative Movement and Soil Fixation of Soluble Organic and Inorganic Phosphorus J. Environ. Qual., November 7, 2005; 34(6): 2228 - 2233. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Page, P. M. Haygarth, K. J. Beven, A. Joynes, T. Butler, C. Keeler, J. Freer, P. N. Owens, and G. A. Wood Spatial Variability of Soil Phosphorus in Relation to the Topographic Index and Critical Source Areas: Sampling for Assessing Risk to Water Quality J. Environ. Qual., November 7, 2005; 34(6): 2263 - 2277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. H. Anderson and F. R. Magdoff Autoclaving Soil Samples Affects Algal-Available Phosphorus J. Environ. Qual., October 12, 2005; 34(6): 1958 - 1963. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Gaston, C. M. Drapcho, S. Tapadar, and J. L. Kovar Phosphorus Runoff Relationships for Louisiana Coastal Plain Soils Amended with Poultry Litter J. Environ. Qual., July 1, 2003; 32(4): 1422 - 1429. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Guillard and W. M. Dest Extractable Soil Phosphorus Concentrations and Creeping Bentgrass Response on Sand Greens Crop Sci., January 1, 2003; 43(1): 272 - 281. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Atia and A. P. Mallarino Agronomic and Environmental Soil Phosphorus Testing in Soils Receiving Liquid Swine Manure Soil Sci. Soc. Am. J., September 1, 2002; 66(5): 1696 - 1705. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
