HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Ige, D. V. Articles by Flaten, D. N. Search for Related Content PubMed Articles by Ige, D. V. Articles by Flaten, D. N. Agricola Articles by Ige, D. V. Articles by Flaten, D. N. Related Collections Sorption/Exchange Nutrient Cycling Soil Chemistry

SOIL CHEMISTRY

Direct and Indirect Effects of Soil Properties on Phosphorus Retention Capacity

Dep. of Soil Science, Univ. of Manitoba, Winnipeg, MB, Canada R3T 2N2

^* Corresponding author (akinremi{at}ms.umanitoba.ca).

	ABSTRACT

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

 
Phosphorus retention ability of soil has been predicted from different combinations of soil properties as a result of significant correlations between these variables. However, a significant correlation between P retention capacity and a soil property does not necessarily imply a significant direct effect of the soil property on P retention. The objective of this study was, therefore, to evaluate the direct and indirect influence of soil properties on P retention capacity of neutral to calcareous soils of Manitoba (Canada). One hundred fifteen archived surface soil samples representing major soils of Manitoba were used for the study. The P retention index (PRI) of these soils was determined by the single-point adsorption method. The relationships between P retention and soil properties were evaluated by correlation analyses while the direct and indirect influences of these soil properties on P retention were evaluated using the path analysis procedures. Significant correlations (p = 0.05) were observed between PRI and pH; CO₃^2–; organic C; sand content; silt content; clay content; exchangeable Ca (Ca_Ex); exchangeable Mg (Mg_Ex); cation exchange capaciaty (CEC); Mehlich-3 extractable Ca (Ca_M3); Mehlich-3 extractable Mg (Mg_M3); and oxalate extractable Al (Al_Ox). However, path analysis showed that only Ca_M3 D = 0.62), Mg_M3 (D = 0.26), and Al_Ox (D = 0.21) had significant direct effects (p = 0.05) on PRI. The significant effect of sand (soil texture) was due to the indirect influence of Ca_M3. Our results show that the PRI for the neutral to calcareous soils of Manitoba is best predicted from Ca_M3, Mg_M3, and Al_Ox. Similar analysis conducted to relate the classical Langmuir adsorption maximum with the properties of a subset of soils also showed Ca_M3, Mg_M3, and Al_M3 as the soil variables that had significant direct effects on soil's P retention.

	INTRODUCTION

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

 
Phosphorus retention capacity, a measure of the ability of soil to retain P, is an important factor controlling the release of P from soil to water. It is often determined in the laboratory by equilibrating soil with a range of P concentrations for a set period of time. The amount of P sorbed is estimated as the difference between the amount of P added and the P remaining in solution at equilibrium. The data obtained are then fitted into different adsorption models and various indices of P adsorption capacity are determined (Barrow, 1978; Chien and Clayton, 1980; Kinniburgh, 1986). However, this is a time-consuming process that cannot be adapted to routine laboratory analysis. To overcome this problem, the use of a single standard P solution was suggested by Bache and Williams (1971) with the amount of P sorbed taken as a reasonably accurate index of the ability of the soil to retain P. A good correlation has been reported between the adsorption capacity determined from the multi-point isotherm and the retention index obtained from the single point adsorption (Bache and Williams, 1971; Indiati and Sharpley, 1997; Ige et al., 2005).

Various soil properties have been reported to be closely related to the P retention capacity of soils. Such properties include the extractable Fe and Al oxides (Toor et al., 1997; Freese et al., 1992), clay content (Johnston et al., 1991; Toor et al., 1997), organic C (Daly et al., 2001), pH (Barrow, 1984), calcium carbonate (Bertrand et al., 2003), and sand content (Yuan and Lucas, 1982; Leclerc et al., 2001). Because of these close relationships, efforts have been made to predict P retention capacity from these soil properties using various combinations (Lookman et al., 1996; Burt et al., 2002; Ige et al., 2005). Ige et al. (2003) predicted P retention capacity of tropical soils from Al_OX, soil pH, and the clay content. Borling et al. (2001) and Maguire et al. (2001) suggested the combination of Fe_OX and Al_OX for the prediction of soil P sorption capacity in non-calcareous soils. Calcium carbonate content was used by Bertrand et al. (2003) for the prediction of P sorption capacity for calcareous soils with CaCO₃ content > 1%.

Bertrand et al. (2003), however, stressed the need for caution in interpreting correlation relationships between P retention capacity and soil properties because of the inter-correlations among soil properties. Thus, a strong correlation of a soil property with P retention capacity may not necessarily imply the direct influence of the soil property on P retention capacity of the soil, but may be due to the indirect influence of some other soil properties. Inherent collinearity among the independent variables used to model P retention capacity may prevent obtaining a regression model of desired accuracy (Walker, 2004, Kutner et al., 2004). This was supported by Zhang et al. (2005) who reported that the significant correlation observed between P sorption capacity and clay content was due to the significant indirect influences of Al_OX and Fe_OX. As such, simple correlation analysis may not be sufficient in evaluating the direct influence of soil properties on P retention capacity.

To estimate the direct and indirect effects of soil properties on P retention, path analysis can be employed. While the direct effect represents the direct contribution of an independent variable to the dependent variable, the indirect effect represents the contribution of an independent variable to the dependent variable through the influence of another independent variable. Path analysis was developed by Wright (1921) to organize and present relationships between dependent and independent variables, thus decomposing the relationships into different pieces for interpretation of effects. Path analysis, which is an extension of regression model, is used to test the fit of the correlation matrix against two or more causal models. It decomposes the source of the correlation among variables and, thus, partitions the correlation between dependent and independent variables into direct and indirect effects. Therefore, correlation and path coefficient analysis lead us to a clearer understanding of the association between dependent and independent variables. Path coefficient is a numerical estimate of the causal relationship between two variables in the path analysis. While path analysis has been employed in soil studies to investigate the cause and effect of soil properties on heavy metal adsorption (Basta et al., 1993; Krishnasamy and Mathan, 2001), its application to P retention in relation to soil properties was a new development as employed by Zhang et al. (2005) for acidic soils.

In view of the use of different combinations of soil properties to estimate P retention capacity of soil and the use of these estimates to evaluate the risk of P loss to the environment, it is important to determine the soil properties that have significant direct influence on P retention. This information will allow P retention capacity to be estimated using appropriate soil properties contained in (i) routine soil survey database (e.g., pH, CEC, %Sand, %Silt, %Clay, and CaCO₃); and (ii) routine soil test extracts such as P_M3, Ca_M3, Mg_M3, and Al_M3. Therefore, the objective of this study was to evaluate the direct and indirect effect of soil properties on P retention capacity of neutral to calcareous soils of Manitoba (Canada).

	MATERIALS AND METHODS

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

 
One hundred fifteen (115) archived surface (0–15 cm) soil samples from distinct soil groups across the province of Manitoba (Canada) were obtained from the Manitoba Soil Survey Department. The samples had been collected for various Soil Survey research studies, air dried and sieved using 2-mm pore size sieve, and stored. The physical and chemical properties of the soils were reported by Ige et al. (2005). Various extracting agents (e.g., Mehlich-3, sodium bicarbonate, ammonium oxalate, ammonium acetate, and sodium nitrilotriacetate [NTA]) were used to extract P, Ca, Mg, Fe, Al, and Mn; their values were reported by Ige et al. (2005). The soils studied had a wide range of pHs (5.3–8.1) but most of the soils had neutral to high pHs. Fifty-one of the soils have pH values > 7.5, 48 soils have pH values that ranged between 6.5 and 7.5 while 16 soils have pH values < 6.5. The carbonate content ranged between 0 and 382 g kg^–1 with a mean of 32 g kg^–1. Tables 1 and 2 present the basic physical and chemical properties of the soils.

Since most statistical models for multivariate data analysis were developed with the assumption that the variables were normally distributed (Legendre and Legendre, 1998), all the variables were standardized to assume a mean of 0 and a standard deviation of 1 before statistical analysis. The standardization was achieved by subtracting the mean of a variable from all values in that variable, then dividing the centered values by the standard deviations as:

where y_i is the value for ith dependent variable y, is the mean of dependent variable y, s_y is the standard deviation of dependent variable y, z_yi is the standardized ith dependent variable y; x_i is the ith value of independent variable x, is the mean of independent variable x, s_x is the standard deviation of independent variable x, z_xi is the standardized ith independent variable x.

Backward regression analysis was conducted using the SAS statistical package (SAS, 2000) with PRI as the dependent variable and the various soil properties (pH, organic C, sand, clay, silt, carbonate, Ca_M3, Mg_M3, Al_M3, Fe_M3, and Mn_M3, NTA-extractable Ca, Mg, Al and Fe, Al_OX, Fe_OX and Mn_OX, and ammonium acetate-extractable Ca and Mg referred to as Ca_EX and Mg_EX) as independent variables with the option to test collinearity among the independent variables. Backward regression is a multiple regression procedure in which all the independent variables are entered into the regression equation and variables that do not contribute significantly to the fit of the regression model are stepwisely eliminated until only statistically significant variables remained. Thus, soil properties that did not make significant contributions to PRI determination at p = 0.10 were stepwisely eliminated from the regression equation. Following this exercise, seven variables were selected for path coefficient analysis: sand, carbonate, Ca_M3, Mg_M3, and Mn_M3, and Al_Ox and Fe_Ox (Fig. 1 ). The fitness of the prediction model was measured by the coefficient of determination (R²) of the regression model.

View larger version (24K): [in this window] [in a new window]   Fig. 1. Path diagram illustrating the relationship between soil properties and soil P retention capacity. Dij denotes the direct path; rij denotes the correlation coefficient; P retention index (PRI) is the soil P retention capacity; sand is the percentage of sand; CaM3, MgM3 and MnM3 denote Mehlich-3 extractable Ca, Mg, and Mn, respectively; AlOx and FeOx are the ammonium oxalate-extractable Al and Fe, respectively; and U is the residual estimated as where R2 is the coefficient of determination for the regression model (R2 = 0.88).

where r_YXi is the simple correlation coefficient between an independent variable Xi and a dependent variable Y; D_YXi is the path coefficient between an independent variable Xi and a dependent variable Y and is the direct effect of Xi on Y; and r_XiXj is the simple correlation coefficient between independent variables Xi and Xj.

The residual, U, is an unmeasured variable in the path model that represents the unexplained part of an observed variable. It is calculated as:

where R² is the coefficient of determination of the regression model.

	RESULTS AND DISCUSSION

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

 
Correlation between Soil Properties and Phosphorus Retention Index
The correlation matrix between soil properties and P retention index is presented in Table 3. Significant correlations (p < 0.05) were observed between the PRI and several of the soil properties. Those soil properties that were significantly related with retention capacity included pH, CO₃^2–, organic carbon, sand content, silt, clay content, exchangeable Ca, exchangeable Mg, CEC, Mehlich-3 extractable Ca, Mehlich-3 extractable Mg, and oxalate extractable Al. Such significant relationships between P retention capacity and different soil properties have been reported by several authors (Barrow, 1984; Toor et al., 1997; Sims et al., 1998; Leclerc et al., 2001; Ige et al., 2005).

Path and Multiple Regression Analyses for the 115 Soils using PRI as the Dependent Variable
Of all the soil properties evaluated, seven variables, namely CO₃^2–, sand, Ca_M3, Mg_M3, Al_Ox, Mn_M3, and Fe_Ox were retained in the stepwise backward regression model as being statistically significant to the prediction of PRI with a coefficient of determination (R²) of 0.88 and a residual (U) of 0.35. A coefficient of determination (R²) of 0.88 showed that the model explained most (88%) of the variation in the P retention capacity of the soils. Test of collinearity among independent variables confirmed the presence of significant dependence among variables (data not presented).

Exchangeable Ca and Mg were significantly related with PRI and the regression models using each of these two parameters were able to explain 52 and 46%, respectively, of the variation in the P retention capacity of the soils when used as independent variables for PRI prediction. However, these two variables were not included in the regression model as they did not make significant contribution to the P retention capacity of the soils. Conversely, Ca_M3 contributed significantly to the model. This may due to the fact that Ca_M3 and Mg_M3 were able to account for the role of Ca_EX and Mg_EX as well as other extractable Ca and Mg that were not on the exchange sites (e.g., easily dissolved Ca and Mg from soil minerals).

The partitioning of the correlation coefficient between PRI and soil properties into direct and indirect effects is shown in Table 4. The direct effect of CO₃^2– on PRI (D = –0.13) was negative and not significant. The significant positive correlation between CO₃^2– and PRI observed in the correlation analysis (Table 3) was due to the significant indirect effect of Ca_M3 (D = 0.39; Table 4).

View this table: [in this window] [in a new window]   Table 4. Direct and indirect effects of soil properties on P Retention Index obtained from the single point adsorption for the 115 soils in comparison to the simple correlation coefficient.

Although all the variables analyzed were significantly related to PRI, only Ca_M3, Mg_M3, and Al_Ox had significant direct effects on the P retention capacity of the soils with Ca_M3 having the highest direct effect (D = 0.62). The significant direct effect of Ca_M3 and Mg_M3 supported the selection of these two variables as the best two-variable predictive model for P retention of Manitoba soils (Ige et al., 2005). Kleinman and Sharpley (2002) also reported a significant relationship between soil P retention capacity and Ca_M3 for a group of alkaline soils. However, contrary to the conclusion of Ige et al. (2005) that Al_Ox was not important in the prediction of P retention capacity for neutral to calcareous soils, path analysis showed that Al_Ox was important and needed to be included in the prediction equations. Even though its correlation with PRI was low (r = 0.39), the significance of this correlation was due mainly to its direct influence on PRI. Thus, in predicting the P retention capacity of these neutral to calcareous soils, Al_Ox content should still be taken into consideration. This might justify the conclusion of some authors (Samadi and Gilkes, 1998; Bertrand et al., 2003) who reported the significant effect of Al_Ox on P retention capacity of calcareous soil. Zhang et al. (2005) also reported significant direct effect of Al_Ox on P sorption capacity and a nonsignificant direct effect of Ca_M3 on P sorption contrary to our observation. This is probably because Zhang et al. (2005) worked mainly with acid soils having pH < 7 and soils with lower Ca_M3 than those used in our study. Ammonium oxalate extractable Fe (Fe_Ox) and Mehlich-3 Mn did not have significant direct effect on PRI. The low, significant correlation observed between these soil properties and PRI was due to the nonsignificant indirect effects of other soil properties.

Ige et al. (2005) reported that increasing the number of variables in the regression equation for estimating the P retention capacity improved the goodness of fit of the regression equation, as evaluated by the values of the coefficient of the determination. Path analysis, however, showed that increasing the number of variables by including variables that were collinearly related to variables already in the predictive model may reduce the accuracy of the regression equation and add unnecessary effort and expense for determining PRI.

Path and Multiple Regression Analysis for the 26 Soil Subset using Langmuir Adsorption Maximum as the Dependent Variable
In a parallel analysis conducted using the traditional Langmuir adsorption maximum (S_max), Al_M3, Ca_M3, Mg_M3, and Fe_Ox were the four variables that were statistically significant (p = 0.1) in predicting the P retention of the soil, using stepwise backward regression, with a R² value of 0.86 and U value of 0.37. Path analysis showed that of these four variables, Ca_M3, Mg_M3, and Al_M3 were the only variables that had significant direct effects on the P retention capacity of the soils (Table 5). The direct effect of Ca_M3, Mg_M3, and Al_M3 may be due to the acid nature of Mehlich-3 extracting agent, which is able to dissolve other reactive forms of Ca, Mg, and Al that are important to soil P retention in addition to Ca, Mg, and Al at the exchange sites. Even though simple correlation analysis indicated that the correlation between S_max and Al_M3 was not significant, the direct effect of Al_M3 was significant. This further shows that simple correlation analysis is not enough in describing the relationship between P retention and soil properties.

View this table: [in this window] [in a new window]   Table 5. Direct and indirect effects of soil properties on Langmuir adsorption maximum obtained from the multipoint adsorption isotherm for a subset of 26 soils.

	CONCLUSIONS

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

 
Not all soil variables that exhibited significant correlation with soil P retention ability had a significant direct effect on P retention. Thus, in formulating equations for estimating P retention capacity of soils, simple regression analysis alone may not be adequate in determining which soil variables are to be included in the equation. However, path analysis was able to distinguish between direct and indirect effects of soil properties on P retention, making it possible to determine soil variables that directly influence P retention. To estimate the PRI of neutral to calcareous soils of Manitoba the three variables that should be taken into consideration are Ca_M3 and Mg_M3 and Al_OX. However, for the classical Langmuir adsorption maximum (S_max), Ca_M3, Mg_M3, and Al_M3 should be used. The advantage of these three variables is that they all can be determined from a single extraction procedure. Although soil texture was significantly related to the P retention capacity; the direct effect was not significant.

	ACKNOWLEDGMENTS

 
The authors acknowledged the Manitoba Livestock Manure Management Inc. (MLMMI), the Sustainable Development Innovation Fund (SDIF) and the Manitoba Rural Adaptation Council (MRAC) for making funds available for this study.

	NOTES

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

 
Abbreviations: CEC, cation exchange capacity; subscript EX, exchangeable; subscript M3, Mehlich-3 extractable; NTA, sodium nitrilotriacetate; subscript OX, oxalate extractable; PRI, P retention index.

Received for publication September 28, 2005.

	REFERENCES

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

 

• Bache, B.W., and E.G. Williams. 1971. A phosphate sorption index for soil. J. Soil Sci. 22:289–301.
• Barrow, N.J. 1978. The description of phosphate adsorption curves. J. Soil Sci. 29:447–462.[CrossRef]
• Barrow, N.J. 1984. Modelling the effects of pH on phosphate sorption by soils. J. Soil Sci. 35:283–297.
• Basta, N.T., D.J. Pantone, and M.A. Tabatabai. 1993. Path analysis of heavy metal adsorption by soil. Agron. J. 85:1054–1057.[Abstract/Free Full Text]
• Belsley, D.A., E. Kuh, and R.E. Welsch (ed.). 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. John Wiley & Sons Inc., New York.
• Bertrand, I., R.E. Holloway, R.D. Armstrong, and M.J. McLaughlin. 2003. Chemical characteristics of phosphorus in alkaline soils from southern Australia. Austr. J. Soil Res. 41:61–76.[CrossRef]
• Borling, K., E. Otabbong, and E. Barberis. 2001. Phosphorus sorption in relation to soil properties in some cultivated Swedish soils. Nutr. Cycl. Agroecosyst. 59:39–46.
• Burt, R., M.D. Mays, E.C. Benham, and M.A. Wilson. 2002. Phosphorus characterization and correlation with properties of selected benchmark soils of the United States. Commun. Soil Sci. Plant Anal. 33:117–141.[CrossRef]
• Chien, S.H., and W.R. Clayton. 1980. Application of Elovich equation to the kinetics of phosphate release and sorption in soils. Soil Sci. Soc. Am. J. 44:265–268.[ISI]
• Daly, K., D. Jeffrey, and H. Tunney. 2001. The effect of soil type on phosphorus sorption capacity and desorption dynamics in Irish grassland soils. Soil Use Manage. 17:12–20.
• Freese, D., S.E.A.T.M. Van der Zee, and W.H. Van Riemsdijk. 1992. Comparison of different models for phosphate sorption as a function of the iron and aluminum oxides of soils. J. Soil Sci. 43:729–738.[CrossRef]
• Ige, D.V., J.A. Adepetu, O.A. Obi, and M.T. Adetunji. 2003. Phosphate sorption potential as a predictor of plant available phosphate in soils of southwestern Nigeria. African Soils 33:13–27.
• Ige, D.V., O.O. Akinremi, D.N. Flaten, B. Ajiboye, and M.A. Kashem. 2005. Phosphorus sorption capacity of alkaline Manitoba soils and its relationship to soil properties. Can. J. Soil Sci. 85:417–426.
• Indiati, R., and A.N. Sharpley. 1997. Changes in some soil phosphorus availability parameters as induced by phosphorus addition and soil sorption properties. Commun. Soil Sci. Plant Anal. 28:1565–1578.
• Johnston, M.A., N. Miles, and G.R. Thibaud. 1991. Quantities of P fertilizer required to raise the soil test value. South Afr. J. Plant Soil 8:17–21.
• Kinniburgh, D.G. 1986. General purpose adsorption isotherms. Environ. Sci. Technol. 20:895–904.[CrossRef]
• Kleinman, P.J.A., and A.N. Sharpley. 2002. Estimating soil phosphorus sorption saturation from Mehlich-3 data. Commun. Soil Sci. Plant Anal. 33:1825–1839.[CrossRef]
• Krishnasamy, R., and K.K. Mathan. 2001. Path coefficient analysis of zinc and boron adsorption in soils. Commun. Soil Sci. Plant Anal. 32:465–475.[CrossRef]
• Kutner, M.H., C.J. Nachtsheim, and J. Neter. 2004. Applied linear regression models. 4th ed. McGraw-Hill, Irwin, New York.
• Legendre, P., and L. Legendre. 1998. Numerical ecology. 2nd English ed. Elsevier Science B.V. Amsterdam, The Netherlands.
• Leclerc, M.L., M.C. Nolin, D. Cluis, and R.R. Simard. 2001. Grouping soils of the Montreal Lowlands (Quebec) according to fertility and P sorption and desorption characteristics. Can. J. Soil Sci. 81:71–83.
• Li, C.C. 1975. Path analysis—A premier. Boxwood Press, Pacific Grove, CA.
• Lookman, R., K. Jansen, R. Merckx, and K. Vlassak. 1996. Relationship between soil properties and phosphate saturation parameters. A transect study in Northern Belgium. Geoderma 69:265–274.
• Maguire, R.O., R.H. Foy, J.S. Bailey, and J.T. Sims. 2001. Estimation of the phosphorus sorption capacity of acidic soils in Ireland. Eur. J. Soil Sci. 52:479–487.[CrossRef]
• Samadi, A., and R.J. Gilkes. 1998. Forms of phosphorus in virgin and fertilized calcareous soils of Western Australia. Austr. J. Soil Res. 36:585–601.[CrossRef]
• SAS institute. 2000. SAS user's guide. Version 8.2 ed. SAS Inst., Cary, NC.
• Sims, J.T., R.R. Simard, and B.C. Joem. 1998. Phosphate loss in agricultural drainage: Historical perspective and current research. J. Environ. Qual. 27:277–293.[ISI]
• Toor, G.S., G.S. Bahl, and A.C. Vig. 1997. Pattern of P availability in different soils as assessed by the adsorption equations. J. Ind. Soc. Soil Sci. 45:719–723.
• Walker, J.H. 2004. Teaching regression with simulation. p. 2096–2102. In R.H. Ingalls et al. (ed.) Proceedings of the 2004 Winter Simulation Conference. 5–8 Dec. 2004. Washington, DC. Omnipress, Washington, DC.
• Williams, W.A., M.B. Jones, and M.W. Demment. 1990. A concise table for path analysis statistics. Agron. J. 82:1022–1024.[Abstract/Free Full Text]
• Wright, S. 1921. Correlation and causation. J. Agric. Res. 20:557–585.
• Yuan, T.L., and D.E. Lucas. 1982. Retention of phosphorus by sandy soils as evaluated by adsorption isotherms. Proc. Soil Crop Sci. Soc. Fla. 11:197–201.
• Zhang, H., J.L. Schroder, J.K. Fuhrman, N.T. Basta, D.E. Storm, and M.E. Payton. 2005. Path and multiple regression analyses of phosphorus sorption capacity. Soil Sci. Soc. Am. J. 69:96–106..

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Ige, D. V. Articles by Flaten, D. N. Search for Related Content PubMed Articles by Ige, D. V. Articles by Flaten, D. N. Agricola Articles by Ige, D. V. Articles by Flaten, D. N. Related Collections Sorption/Exchange Nutrient Cycling Soil Chemistry