Comparison of extraction techniques for measuring exchangeable cations in calcareous soils.
Cation exchange capacity (CEC) measurement in alkaline soils is complicated by the need to remove soluble cations and the requirement to buffer extraction solution at a pH to prevent soluble carbonates from dissolving (Tucker 1971, 1974; Ross 1995). Method 15C1 from the Australian Soil and Land Survey Handbook Australian Laboratory Handbook of Soil and Water Chemical Methods" (Rayment and Higginson 1992), colloquially known as the Tucker method, is often quoted as the preferred technique for alkaline soils (Rengasamy and Churchman 1999). Due to the high labour requirements of this method, very few routine testing laboratories have adopted it without significant modification. The labour-intensive steps identified in Method 15C1 requiring multiple leach robes and multiple solvent pre-washes reflect the earliest protocols published on the measurement of exchangeable cations in calcareous soils (Tucker 1954). End-over-end slow speed mixing machines and single solvent pre-washes replaced leach columns and multiple solvent pre-washes on the basis of improved precision and superior leaching of soluble salts (Tucker 1971, 1974, 1985; FAO 1990).
Sodic soils contain high concentrations of exchangeable sodium. An exchangeable sodium percentage (ESP) value of 5% (McIntyre 1979) has been defined as the threshold between sodic and non-sodic soil for Australian conditions. When using solvents containing water and ethanol as a prewash to remove soluble salts, sodic soils disperse, which could result in a loss of clay particles prior to extracting the exchangeable cations. Polyvinyl alcohol (PVA) has long been used to condition soils to reduce dispersion. It is common laboratory practice to add PVA to ethanolic (but not ethylene glycol) pre-wash solutions to reduce soil dispersion and the loss of clay particles, and improve clarity of extracts (Tucker 1985; B. Shelly, pers. comm.).
The origin of CEC lies in the negative charges carried by soil particles, usually clay, organic matter, and sesquioxides. These charges fall into 2 categories, either permanent or variable (e.g. pH-dependent). The variable charge component is altered by both concentration and valence of the ions in the equilibrium solution. In addition, specific adsorption of both cations and aniominan have marked effects on the manner and conditions in which CEC is determined (Sumner and Miller 1996). Consequently, CEC is operationally defined by its pH, concentration, nature, and valence of the ions, and buffer capacity of the extracting salt solution. The values for CEC are empirically determined by the methods used to remove the soluble salts and extract the exchangeable cations. Varying these conditions will yield different results.
Complex, labour-intensive procedures such as those described by Amrhein and Suarez (1990) for determining sodium and calcium in calcareous and gypsiferous soils preclude widespread adoption in routine testing laboratories. Simplicity and reduced number of analytical steps for calculating the effective CEC (ECEC), which is determined by summing the extracted exchangeable bases, has become the industry standard procedure rather than measuring CEC (NSW Agriculture and Fisheries 1989; Amacher et al. 1990; Incitec Fertilizers 1995).
Comparison of extraction methods involves 2 considerations: agreement and conversion. Study of agreement between methods begins with the assumption that exchangeable cation methodology cannot be regarded as measuring a true quantity. All methods will invariably include soluble, exchangeable, and extractable cations to different degrees. Therefore, interest lies in finding whether a method provides results that are directly comparable to another, rather than providing an absolute value. If methods are found to differ, they may do so in a consistent and predictable manner. In that case, interest may lie in converting the results from one method to results that would have been obtained via a different method.
The objective of this paper is to compare the performance (agreement and convertibility) of 9 separate ECEC extractions on calcareous soils. The solvents used to remove soluble salts and means of extractions were the variables. Additionally, use of unbuffered salt extraction was included in the comparisons. The time involved to conduct the extraction along with correlations between methods formed part of the assessment.
Materials and methods
The 30 soil samples compared in this assessment were from a farmer-managed experiment at 'Woodgmin' near Boggabri (150[degrees]09'E, 30[degrees]76'S) in north-western New South Wales (NSW), Australia. The soils were classified as fine, thermic, montmorillonitic, Typic Haplusterts (Soil Survey Staff 1996). The soils had both high salinity and sodicity in the subsoil. Electrical conductivity (ECe) and exchangeable sodium percentage (ESP) in the 0.6-1.8 m depth were 8.9 dS/m and 24%, respectively. This soil type is viewed as characteristic for the land under cotton cultivation in southern Queensland and north-western NSW. The high salinity of these calcareous soils made them desirable as test sites for these exchangeable cation extraction comparisons.
All reagents were of analytical grade and high purity. Extraction and prewash solutions were prepared with ammonium chloride (AnalaR BDH Prod 10073D), ammonia solution 28% (AnalaR BDH Prod 104405Q), barium chloride (Univar Ajax Finechem 408486), and ethanol 100% undenatured (Chem. Supply 615798, ethylene glycol (1,2-ethanediol, AnalaR BDH Prod 103244N), which was deionised with amberlite ion exchange resin-150 (Merck 1.15965.0500), glycerol 99%+ (Aldrich cat 13,487-2), hydrochloric acid (Univar Ajax Finechem B/No. A3F034), methanol (Omnisolv Ajax Chemicals 408486), nitric acid (Aristar BDH Prod 450043), and polyvinyl alcohol (MW approx. 22000, BDH Prod 305735B).
The cation calibration standards were prepared by diluting 10000 [micro]g/mL standard solutions prepared by Perkin Elmer Australia in 2.5% nitric acid. A caesium solution (5000 mg/L) prepared from caesium chloride (Univar Asia Pacific Speciality Chemicals B/No. F2J070) was added to the analytical stream via a mixing module constructed from a t-connector with mixing coils provided by Lachat Instruments flow injection analyser. Caesium flow rate was 1.4 mL/min. Caesium saturates the plasma with ions to control the easily ionisable effects observed when high concentrations of cations are analysed in an axial orientation (Wu and Hieftje 1994; Ryan 1997).
Extractions were conducted using a Heidolf Reax 2 tumbler or a NSW Agriculture designed tumbler consisting of 3 separate rollers. A 50-place leaching rack was prepared using a commercially available plastic storage box (80 by 45 by 14.5 cm). Holes were drilled in the lid to accommodate the 15-mL centrifuge tubes (Sarstedt 62554.001) and holes were drilled in the bottom to fit the sample bottles (Cospack Pty Ltd, 125 mL natural barrel bottle).
All measurements were made using matrix-matched standard solutions, using a Spectro (Kleve, Germany) [Ciros.sup.CCD] inductively coupled plasma-atomic emission spectrometer (ICP-AES) in an axial orientation. Three sets of instrument operating conditions were employed during this investigation. Each set of conditions was optimised according to the solvent strength and dilution required for each analysis. The instrumental parameters are summarised in Table 1.
The samples were dried at 40[degrees]C and ground to pass 2-mm sieves prior to testing. In order to facilitate comparisons, each method has been arbitrarily assigned a method letter as an identifier. Method A was based on Rayment and Higginson's (1992) procedure 15 C1 without variation. The leaching column consisted of 15-mL centrifuge tubes with a 1-mm hole drilled in the base. A small amount of acid-washed cotton wool was placed over the hole and covered with 2-mm acid-washed sand (Unilab, Asia Pacific Speciality Chemicals B/No. F2L167). Each column was conditioned with 10 mL of 0.1 M nitric acid and 10 mL of Type-1 water before adding any soil. Soil (2.5 g) was mixed with approximately an equal amount of acid-washed sand and placed within the column. The soluble salts of the sample were removed with 12.5 mL 60% w/v aqueous ethanol (665 mL of a solution containing 96% absolute ethanol, 2% water, and 2% methanol) brought to 1 L with Type-1 water. This was followed by a second wash with 12.5 mL of 20% w/v glycerol. The samples were then leached with 4 additions of 12.5 mL alcoholic (60% w/w as above) 1 M ammonium chloride solution buffered to pH 8.5 into a plastic bottle containing 40 mL 0.5 M HC1. The volume was then brought to 100 mL with 10 m L 1 M alcoholic ammonium chloride. The extract solution was 36% w/w ethanol and at a 40 times dilution factor (2.5 g:100 mL). These extracts were diluted 10 times and analysed using ICP-AES conditions shown for method number I (Table 1). This extraction took 4.5 days to complete with 15 h hands-on time to finish. Many of the samples were extremely slow draining, reflecting the high clay content.
Method B was the only testing conducted outside NSW Agriculture's Environmental Laboratory. The Department of Primary Industries, State Chemistry Laboratory of Victoria, located at Werribee did this testing. This procedure is a validated in-house method reported to be equivalent to Rayment and Higginson's (1992) method 15C1 (B. Shelley, pers. comm., 2003). In this method soluble salts were removed from 4 g of soil with 25 mL of a Sarina ethanol (the same composition as that identified as the first soluble salt prewash of Method A) containing 0.05% w/v polyvinyl alcohol. This mixture was shaken on a reciprocating shaker for 5 min at 150-200 cycles/min. The samples are then centrifuged and the washing solution discarded. The samples were then sequentially extracted with 3 additions of 25 mL and one addition of 15 mL extracting solution. The composition of this extraction solution is approximately 60% v/v denatured ethanol and 1 M ammonium chloride solution buffered to pH 8.5. Each extraction was by reciprocating shaker for 5 min at 150-200 cycles/min. These extracts were combined with 10 mL 2 M HN[O.sub.3] to bring the final volume to 100 mL and then analysed on ICP-AES following appropriate dilution.
The extraction used for Method C was conducted exactly as Method A except that the ethanolic prewash contained 0.05% PVA.
Method D consisted of a single extraction using the solvents identified in Rayment and Higginson's (1992) method 15C1. This extraction was conducted on 2 g of soil in a 50-mL centrifuge tube. The first prewash utilised 25 mL of 60% (v/v) ethanol mixed with the soil for 30 min in an end-over-end tumbler at 6 r.p.m. followed by centrifuging and decanting. The final solution used to remove soluble salts consisted was 25 mL of 20% (v/v) glycerol under the same conditions as the first rinse. Exchangeable cations were shaken with 60% w/w ethanolic 1 M ammonium chloride solution buffered at pH 8.5. This extraction was conducted by adding 40 mL of the extraction solution to the wetted soil and extracting for 16 h at 6 r.p.m.. The extract was diluted 10 times to yield a solution that was 6% ethanol and a dilution factor of 100 before analysis by ICP Method 1 (Table 1). Total extraction time was 1 day with 5 h hands-on.
Method E was carried out exactly as was done for Method D except that the ethanolic prewash was made to contain 0.05% PVA.
Method F was Gillman and Sumpter's (1986) 0.1 M N[H.sub.4]Cl:0.1 M Ba[Cl.sub.2] extraction of exchangeable bases. This method consisted of 4 g of air-dried soil being pre-washed 2 times with 40 mL of 60% w/v ethanol shaken for 30 min. These prewashes were followed by a single 30-min rinse using 40 mL 20% aqueous glycerol. The exchangeable cations were extracted with 40 mL of 0.1 M N[H.sub.4]Cl:0.1 M Ba[Cl.sub.2] extracted for 30 min. Extraction time was 1 day with 6.5 h hands-on from the analyst. These extracts were diluted 5 times and the exchangeable cations analysed by ICP-AES Method 2.
Methods G, H, and l were procedurally based on Tucker (1974), except for variations in the prewash methodology. Method G employed two 30-min end-over-end tumbler washes to remove the soluble salts. These included one with 20 mL of 70% v/v aqueous ethanol followed by another with 20 mL of 17% v/v aqueous glycerol. Method H used the same wash solutions as Method G, except that the ethanol wash contained 0.05% PVA. The prewash solution for Method I consisted of 10% v/v deionised ethylene glycol in 90% v/v ethanol. These methods used 2 g of soil and 4 separate extractions for 30 min using 20 mL alcoholic (60% w/w) 1 M ammonium chloride buffered at pH 8.5 extractions. A final extraction consisted of 30 min with 15 mL alcoholic (60% w/w) 0.05 M ammonium chloride buffer at pH 8.5. These extracts were combined and brought to 100 mL with 4 M HCl. Extraction time for each of these methods was 2 days and required 7.5 h of operator time. This solution was diluted 4 times and analysed by ICP-AES Method 3, requiring greater power to handle the higher concentration of ethanol (15%). Table 2 contains a brief summary of the differences between the extraction methods.
The results from exchangeable cation testing will inevitably vary when different extraction techniques arc employed. Since the procedures for measuring exchangeable cations are intended to measure the same entity, there may be a functional relationship established between the results that would enable conversion. This relationship may be expressed so that measurements from one method (x) may be converted to those by another method (y) using an equation of form: y = a + bx.
Estimation of a and b would seem to be a simple application of linear regression. However, in this case the regression parameters would be poor estimates of the true values. The reason is that we do not directly observe x and y, we observe proxy values v and w where:
v = x + [e.sub.1] and w = y + [e.sub.2]
The measurement errors [e.sub.1] and [e.sub.2] have zero means and constant variances [[sigma].sup.2.sub.1] and [[sigma].sup.2.sub.2], respectively. Thus there is a structural relationship between v and w of form:
w = a + bv + [e.sub.2] - b[e.sub.1]
that is ignored by simple linear regression of w on v. It can he shown that the main consequence of this is that the simple regression estimate of b will under estimate the true value. An estimate of the parameters under the structural relationship is available if the error variances ([[sigma].sup.2.sub.1] and ([[sigma].sup.2.sub.2]) are known (Kendall and Stuart 1979). Unfortunately, this is rarely the case.
However, as described by Kendall and Stuart (1979) and applied by Lewis et al. (1991), mathematical limits for the estimate of the slope parameter b are given by:
regression of w on v < |b| < inverse of regression of v on w
Lewis et al. (1991) suggested that if the limiting regressions were 'close' (i.e. correlation between v and w is 'high') then the line bisecting this structural region can be taken as a reasonable estimate of the true structural relationship.
This leads to estimation of conversion factors between any pair of methods (v and w for consistency with previous notation) as follows:
(1) Obtain least squares regression estimates of [a.sub.1], [b.sub.1], [a.sub.2], and [b.sub.2] as follows:
v = [a.sub.1] + [b.sub.1]w
w = [a.sub.2] + [b.sub.2]v (giving v = -[a.sub.2] / [b.sub.2] + 1 / [b.sub.2] w)
(2) Obtain the estimate of the structural relation:
v = A + Bw
A = [a.sub.1] - [a.sub.2] / [b.sub.2] / 2
B = [b.sub.1] + 1 / [b.sub.2] / 2
At this stage, average absolute prediction error is proposed as a measure of the 'goodness of fit' for the structural relationship. Figure 1 displays an example of this process using methods A and B for extraction of sodium. The tables containing the conversion equations and absolute error estimates are available as an accessory publication.
[FIGURE 1 OMITTED]
Results and discussion
Exchangeable sodium agreement between the 9 methods is shown in Fig. 2; method B extracted more sodium than other methods, as shown with the dotted line (1:1 line of agreement) being consistently higher than the solid line (structural relationship). Only 2 pairs of methods indicated some degree of agreement with the dotted line and the solid line nearly overlapping. Methods A and D had some similarity as did methods F and G. In methods A and D the only difference in procedure was the method of extraction. Exchangeable cations in Method A were extracted in leach tubes, whereas Method D used a tumbler. Method F was the Gilman and Sumpter unbuffered salt extraction with prewash and Method G was a conventional calcareous soil extract buffered at pH 8.5.
[FIGURE 2 OMITTED]
The ethanolic extractions with PVA (Methods C, E, H) had higher sodium values than the companion method without PVA (Methods A, D, G). Method I, which used ethylene glycol as the solution for washing soluble salts, had higher sodium values than most ethanol washes other than those determined by method B. This suggested that ethanol removed more exchangeable sodium than ethylene glycol, as reported by Tucker (1971, 1974).
Although the correlation between methods shown in Table 3 indicates a strong association between the methods (range = 0.94-0.99), it must be noted that the soil samples include a cluster with relatively low sodium values. These low values are exerting a strong influence on the correlation estimates. It could be reasonably conjectured that correlations would be somewhat lower if the samples were more uniformly distributed across the range of sodium levels. The conversion error was elevated for Methods B, D, and E compared with the other extractions tested.
Figure 3 provides a graphical depiction of the agreement between the different extraction techniques for exchangeable potassium. Once again, the State Chemistry Laboratory Method (B) had the highest exchangeable potassium. Methods B, G, H, and I showed the greatest level of agreement. The common factor among these similar methods was that each had multiple sequential extractions rather than a single leach or single extraction. These similar methods covered each prewash variable tested, ethanol by itself, ethanol with PVA, and ethylene glycol, indicating that the selection of tested solutions used to remove soluble salts had little effect on the measured concentration of potassium.
[FIGURE 3 OMITTED]
The regression statistic calculated for potassium in Table 4 covered a larger range than the sodium correlations (range = 0.63-0.96). Method A had the overall lowest correlation for potassium compared with all other methods. Method A also recorded the highest levels of absolute error when converting between methods.
Exchangeable magnesium agreement is shown in Fig. 4. There was a high level of agreement between most extraction methods for magnesium. Significant similarities exist between Methods A & H, B & I, B & C, B & D, C & D, F & I, G & H. Except for F & I, the methods with the highest level of agreement shared similar prewash solutions.
[FIGURE 4 OMITTED]
The ethanolic extractions with PVA (Methods C, E, H) did not indicate any affect on the magnesium values compared with the companion method without PVA (Methods A, D, G).
The correlation coefficients for magnesium between each pair of methods are displayed in Table 5, ranging from 0.89 to 0.98. Overall, there was a high degree of association between the methods and the absolute error as a function of concentration was low.
Exchangeable calcium agreement between the 9 methods is shown in Fig. 5. From this figure, it can be seen that Methods A and C extracted more calcium and that these methods had a functional relationship with each other. Methods A and C showed some similarity with calcium extraction results observed with Methods F and G. These methods were similar in that they all had similar aqueous ethanol and aqueous glycerol soluble salt prewash. The extraction solvent for Method F was different from the other methods in that it was an unbuffered dilute salt solution rather than ethanolic 1 M ammonium chloride buffered at pH 8.5.
[FIGURE 5 OMITTED]
The ethanolic extractions with PVA (Methods C, E, H) showed very similar calcium concentrations compared with the companion method without PVA (Methods A, D, G). In fact, the solid lines that indicate structural relationship and dotted lines that show the 1:1 line of agreement virtually overlap when comparing Methods A to C, D to E, and G to H. This provides strong evidence that PVA in the prewash does not exert any influence on the extraction of exchangeable calcium.
Correlation coefficients calculated for calcium between each pair of methods are displayed in Table 6. The range (0.57-0.96) of correlations for calcium was much greater than for the other exchangeable cations, with 13 out of the 36 comparisons being below 0.9. This indicates statistically less agreement with exchangeable calcium comparisons than with the other exchangeable cations extracted. Method F recorded lower correlation values for calcium than the other extractions. Method B was most similar to I, as was Method G to H.
The absolute conversion error was elevated in most comparisons between Methods F, the unbuffered salt extraction, compared with the other calcium extractions tested. Methods B and H had the lowest absolute conversion errors.
The ECEC (summation of the exchangeable cations) agreement between the different extraction techniques may be seen in Fig. 6. Methods A, C, G, H and Methods B, E and I yielded similar concentrations. Since Methods A, C, G, and H had the same prewash and extraction solvent, these similarities are expected. It is noted that Method A and C were leach methods and Methods C and H had 0.05% PVA in the ethanol prewash. This provides evidence that similar results may be obtained for ECEC concentration independent of extraction mode (leach or tumbler) and PVA status.
[FIGURE 6 OMITTED]
It is surprising that Method B, F, and 1 yielded similar ECEC results. Each of these extraction methods used different prewash solvent(s), and the extraction solvent in Method F consisted of a low molarity unbuffered salt solution, while the other two techniques utilised highly buffered, high molarity aqueous ethanolic salt solution for extraction of cations.
The ECEC correlations reported in Table 7 showed that many of the methods yielded overall results that were similar, with correlation coefficients ranging from 0.87 to 0.99. As was the case with calcium, Method F had the lowest ECEC correlation values compared with the other methods.
Overall there was less variation; the observed relationship (solid line) was generally closer to the 1:1 line compared with the regression analysis for the individual exchangeable cations. This would be expected because the differences in each cation tend to cancel out when summed. As an example, Method A had the second lowest exchangeable sodium and the highest exchangeable calcium, whereas Method B had the highest exchangeable sodium and the lowest exchangeable calcium.
Method F generally had higher conversion errors when compared to the other extraction techniques.
The amount of operator time to conduct extractions is one of the important considerations when selecting analytical methods. The time taken for extracting within the NSW Agriculture laboratory was recorded to the nearest half hour during this assessment. Method B was analysed by the SCL of Victoria as a routeing analytical test, so the amount of operator time remains uncertain. The numbers of washes and extraction steps for Method B were similar to Method I, so it is assumed these have similar extraction time requirements.
The methods are shown in order of most to least time for extraction: A & C (15 h) > B, G, H, & 1 (7.5 h) > F (6.5 h) > D & E (5 h). Clearly, the time required for Methods A and C using the leach columns with clayey, calcareous soils would preclude consideration for use as routine laboratory testing if the other results are seen to be equivalent.
As expected, differences were observed with the specific exchangeable cation dependent on the method used for extraction. The ethanol-soluble salt rinse methods recorded the lowest concentration of exchangeable sodium. Since it is well documented (Tucker 1971, 1974; Amacher et al. 1990) that the extraction solutions evaluated in this assessment are capable of removing soluble sodium, it appears that the ethanol rinse procedures removed a greater fraction of the exchangeable sodium. Addition of PVA to control dispersion reduced the loss of sodium, but not to the extent observed for Method B employed by the State Chemistry Laboratory and the Method I, which used ethylene glycol as the prewash solvent.
This study found that there was good agreement with most individual exchangeable cations and ECEC when assessing the leach method, A and C with tumbler extractions that used the same prewash and extraction solutions G and H. This finding demonstrates that the mode of extraction should not be a central consideration when selecting methods, except when evaluating the time requirements. The leach method required at least twice the time of all other extraction techniques.
The results from one extraction technique may easily be converted to another with a high degree of certainty. This means that testing results determined by one method may be easily converted for purposes of comparison to another method.
Use of unbuffered salt extraction as described by Gillman and Sumpter (1986) and Amacher et al. (1990) was included in these comparisons. The unbuffered salt extraction results showed good agreement with many methods, including the Tucker (1974) and State Chemistry laboratory methods. The benefits of using this extraction include ease of measurement on ICP-AES and accurate exchangeable calcium removal without dissolving carbonates. This procedure has good international acceptance (Amacher et al. 1990, Rayment and Higginson 1992; Sumner and Miller 1996) for weathered soil and it appears to be generally satisfactory for calcareous soils. This study found two areas of concern that require additional evaluation prior to recommending full acceptance of this extraction technique over others. It generally showed higher absolute conversion errors than the other methods and extracted less potassium than the sequential extraction techniques (Methods B, G, H, and 1).
Method I (Tucker 1974) had many positive considerations. This extraction had high exchangeable sodium and potassium values, showed overall good agreement with the Gillman and Sumpter method and the State Chemistry Laboratory procedure, was relatively simple and quick to extract, and has good international acceptance for measuring exchangeable cations in alkaline soils (FAO 1990; Ross 1995; Podwojewski and Petard 1996; Ellington et al. 1997; Rieu et al. 1998). The major drawback from this procedure is from the higher ethanol content in dilutions requiring greater expertise in ICP-AES analysis compared with lower ethanol dilutions.
On the basis of the comparisons examined in this paper, the Tucker (1974) or the Gillman and Sumpter extraction (1986) procedures perform favourably compared with the other extraction procedures when assessing exchangeable cations in calcareous soils.
Table 1. Operating conditions for ICP AES determinations Argon Auxillary Nebuliser coolant argon argon Method no. Power Pump speed flow flow flow (W) (flow) (L/min) (L/min) (L/min) 1 1400 2 (2.1 mL/min) 13 1.0 0.8 2 1400 2 (2.1 mL/min) 13 0.9 0.9 3 1600 2 (2.1 mL/min) 12 0.7 1.0 Table 2. Method extraction summary L, Leach; T, tumbler or shaker; E, ethanol prewash; E-G, ethanol gylcerol prewash; 1, ethylene glycol, (1,2 ethanediol) prewash; P, polyvinyl alcohol in prewash; S, sequential extractions: X, single extraction Method ID Summary A Rayment and Higginson Method 15C1, L/E-G/S B State Chemistry Laboratory of Victoria, T/E/P/S C Modified Method 15C1, L/E-G/P/S D Modified Method 15C1, T/E-G/X E Modified Method 15C1, T/E-G/P/X F Gillman and Sumpter's compulsive exchange (unbuffered), T/E-G/S G Tucker 1974, T/E-G/S H Tucker 1974, T/E-G/S/P 1 Tucker 1974. T/1/S Table 3. Matrix of sodium correlations between each pair of methods A B C D E F G H B 0.97 C 0.98 0.99 D 0.96 0.98 0.98 E 0.96 0.98 0.98 0.97 F 0.97 0.99 0.99 0.98 0.97 G 0.97 0.95 0.97 0.94 0.95 0.95 H 0.98 0.98 0.99 0.97 0.97 0.98 0.99 I 0.97 1.00 0.99 0.98 0.97 0.99 0.95 0.97 Table 4. Matrix of potassium correlations between each pair of methods A B C D E F G H B 0.83 C 0.79 0.95 D 0.63 0.84 0.89 E 0.74 0.88 0.88 0.86 F 0.75 0.88 0.83 0.82 0.79 G 0.80 0.94 0.93 0.90 0.89 0.88 H 0.82 0.93 0.93 0.87 0.89 0.92 0.99 I 0.80 0.94 0.96 0.89 0.88 0.87 0.96 0.95 Table 5. Matrix of magnesium correlations between each pair of methods A B C D E F G H B 0.94 C 0.95 0.97 D 0.95 0.96 0.95 E 0.92 0.95 0.95 0.96 F 0.93 0.91 0.89 0.94 0.92 G 0.97 0.96 0.96 0.98 0.94 0.95 H 0.97 0.98 0.96 0.97 0.97 0.93 0.98 I 0.95 0.97 0.96 0.98 0.96 0.91 0.97 0.98 Table 6. Matrix of calcium correlations between each pair of methods A B C D E F G H B 0.90 C 0.92 0.93 D 0.87 0.92 0.90 E 0.87 0.91 0.91 0.87 F 0.70 0.64 0.63 0.70 0.57 G 0.95 0.92 0.95 0.94 0.87 0.74 H 0.96 0.92 0.92 0.93 0.88 0.76 0.97 I 0.91 0.96 0.96 0.93 0.91 0.70 0.95 0.94 Table 7. Matrix of ECEC correlations between each pair of methods A B C D E F G H B 0.95 C 0.95 0.97 D 0.92 0.96 0.95 E 0.92 0.96 0.95 0.95 F 0.88 0.87 0.84 0.90 0.85 G 0.97 0.98 0.97 0.97 0.94 0.91 H 0.98 0.98 0.96 0.97 0.96 0.91 0.99 I 0.95 0.98 0.97 0.98 0.96 0.88 0.99 0.98
We thank Dr N. R. Hulugalle from Australian Cotton Cooperative Research Centre and NSW Agriculture, Australian Cotton Research Institute, for raising the issue of comparative testing of calcareous soils and providing the soils used in these comparisons, and Bruce Shelly from the State Chemistry Laboratory (Vie.) for providing the SCL procedure used to extract calcareous soils and providing additional soil for testing. NSW Agriculture's Diagnostic and Analytical Services are recognised for supplying the resources to undertake this testing at its Wollongbar Environmental Laboratories. We are grateful for the technical assistance of C. Hunt, P. Gupta, S. Judkins, G. Conner, and J. Rust. We are also indebted to Dr P. Slavich for his helpful comments during manuscript preparation.
Amacher MC, Henderson RE, Breithaupt MD, Seale CL, LaBauve JM (1990) Unbuffered and buffered salt methods for exchangeable cations and effective cation-exchange capacity. Soil Science Society of America Journal 54, 1036-1042.
Amrhein C, Suarez DL (1990) Procedure for determining sodium-calcium selectivity in calcareous and gypsiferous soils. Soil Science Society of America Journal 54, 999-1007.
Ellington A, Badawy NS, Ganning GW (1997) Testing gypsum requirements for dryland cropping on a red-brown earth. Australian Journal of Soil Research 35, 591-607.
FAO (1990) Management of gypsiferous soils. FAO Soils Bulletin NO. 62. FAO, Rome.
Gillman GP, Sumpter EA (1986) Modification of the compulsive exchange method for measuring exchange characteristics of soils. Australian Journal of Agricultural Research 24, 61-66.
Incitec Fertilizers (1995) 'Soil interpretation manual." (Incitec: Morningside, Qld)
Kendall MG, Stuart A (1979) 'The advanced theory of statistics. Vol. 2. Inference and relationship.' 4th edn (Griffin: London)
Lewis PA, Jones PW, Polak JW, Tillotson HT (1991) 'The problem of conversion in method comparison studies. Applied Statistics 40, 105-112.
McIntyre DS (1979) Exchangeable sodium, subplacticity and hydraulic conductivity of some Australian soils. Australian Journal of Soil Research 17, 115-120.
NSW Agriculture and Fisheries (1989) 'Biological and Chemical Research Institute soil testing methods and interpretation.' (NSW Agriculture & Fisheries: Rydalmere, NSW)
Podwojewski P, Petard J (1996) Compariasons entre differentes methods d'extraction des sels solubles et des caions echangeables darts un vertisol a gypse de Nouvelle Caledonie. Etude et Gestion des Sols 3, 193-206.
Rayment GE, Higginson FR (1992) 'Australian laboratory handbook of soil and water methods.' (Inkata Press: Melbourne)
Rengasamy P, Churchman GJ (1999) Cation exchange capacity, exchangeable cations and sodicity. In 'Soil analysis--an interpretation manual'. (Eds KI Peverill, LA Sparrow, DJ Reuter). pp. 147-157. (CSIRO Publishing: Collingwood, Vic.)
Rieu M, Vaz R, Cabrera F, Moreno F (1998) Modelling the concentration or dilution of saline soil-water systems. European Journal of Soil Science 49, 53-67. doi:10.1046/J.1365-2389. 1998.00137.X
Ross DS (1995) Recommended methods for determining soil cation exchange capacity. In 'Recommended soil testing procedures for the northeastern United States'. 2nd edn (Eds JT Sims, A Wolf) pp. 62-69. (University of Delaware Press: Newark, DE)
Ryan AJ (1997) Direct analysis of milk powder on the Liberty Series Il ICP-AES with the axially viewed plasma. Varian Corporation. ICP Instruments At Work, ICP-21, Melbourne.
Soil Survey Staff (1996) 'Keys to Soil Taxonomy.' 7th edn (Natural Resources Conversation Service of the United States Department of Agriculture: Washington, DC)
Sumner ME, Miller WP (1996) Cation exchange capacity and exchange coefficients. In 'Methods of soil analysis. Part 3. Chemical methods'. Number 5 in the Soil Society of America Book Series. pp. 1201-1229. (Soil Science Society of America, Inc: Madison, WI)
Tucker BM (1954) The determination of exchangeable calcium and magnesium in carbonate soils. Australian Journal of Agricultural Research 5, 706-715.
Tucker BM (1971) Basic exchangeable cations in soils. CSIRO Australia, Division of Soils, Technical Paper No. 8.
Tucker BM (1974) Laboratory procedures for cation exchange measurements on soils. CSIRO Australia, Division of Soils, Technical Paper No. 23.
Tucker BM (1985) Laboratory procedures for soluble salts and exchangeable cations in soils. CSIRO Australia, Division of Soils, Technical Paper No. 8.
Wu M, Hieftje GM (1994) The effect of easily ionized elements on analyte emission efficiency in inductively coupled plasma spectrometry. Spectrochimica Acta 49, 149-161. doi: 10.1016/0584-8547(94)80014-6
C. G. Pierce (A) and S. Morris (B)
(A) Corresponding author; Department of Environment and Conservation (NSW), Analytical and Environmental Chemistry Section, PO Box 29, Lidcombe, NSW 2825, Australia; email: email@example.com
(B) Wollongbar Agricultural Institute, Bruxner Highway, Wollongbar, NSW 2477, Australia.
Manuscript received 15 September 2003, accepted 26 February 2004
|Printer friendly Cite/link Email Feedback|
|Author:||Pierce, C.G.; Morris, S.|
|Publication:||Australian Journal of Soil Research|
|Date:||May 1, 2004|
|Previous Article:||Field calibration of a capacitance soil water probe in heterogeneous fields.|
|Next Article:||A field study of phosphorus mobilisation from commercial fertilisers.|