Assessment of potentially reactive pools of aluminum in Andisols using a five-step sequential extraction procedure

Abstract

A five-step sequential extraction procedure was developed for the partitioning of soil aluminum (Al) into five fractions: exchangeable, weakly organic bound, strongly organic bound, inorganic non-crystalline and crystalline. The results obtained by the sequential extraction method for Al were compared with those estimated from single extractions using identical operating conditions applied in each individual sequential fraction. The Al content in the extracts was measured by inductively coupled plasma optical emission spectrometry. The results indicated that the first four steps [potassium chloride (KCl), copper chloride (CuCl₂), sodium pyrophosphate (Na₄P₂O₇), acid ammonium oxalate ((NH₄)₂C₂O₄)] in the sequential procedure could be as effective as single extraction methods at estimating exchangeable Al, weakly organic bound Al, strongly organic bound Al, and non-crystalline Al. However, the crystalline Al content by sequential procedure was not in agreement with single extraction procedures. Further, the sequential extractions resulted in more consistent estimates of the aluminum/silicon (Al/Si) molar ratio and allophane and crystalline Al contents than single extractions method. Results of X-ray diffraction on the soil samples confirm the presence of appreciable amounts of at least three types of crystalline minerals, including gibbsite, vermiculite or halloysite. Our result implies that the sequential method can be more reliable in estimating the various Al forms in Andisols.

Keywords:

• Al fractionation
• Andisol
• single extraction procedure
• sequential extraction procedure
• Allophane
• crystalline aluminum.

Introduction

Aluminum (Al) is one of the major elements in the earth's crust, comprising about 7% of its mass. When soils become acidic as a result of natural processes or human activities, Al is solubilized into the toxic trivalent cation, Al³⁺. Micromolar concentrations of Al³⁺ can inhibit root growth within minutes or hours of many agriculturally important plant species (Kochian 1995), resulting in poor growth and productivity. The toxic effect of certain Al species on different plants has been well documented (Lukaszewski and Blevins 1996; Takabatake and Shimmen 1997). However, some plant species, for example tea plants (Camellia sinensis L.) which naturally grow in soils containing elevated levels of Al, have the ability to accumulate unusually high concentrations of Al in mature leaves without any impact on their growth and development (Konishi et al. 1985; Matsumoto et al. 1976). Although the Al accumulation potential is dependent on the ability of plants to absorb and translocate the metal from root to shoot, Al bioavailability in rhizosphere soil is considered to be a critical factor. Since a large proportion of Al in soils is bound to organic and inorganic soil constituents, it is generally unavailable for root uptake.

A number of chemical extraction methods have been developed to identify various aluminum fractions in the soils and characterize their bioavailability. For instance, single extraction methods, including extractions with potassium chloride (KCl), copper chloride (CuCl₂), sodium pyrophosphate (Na₄P₂O₇), acid ammonium oxalate ((NH₄)₂C₂O₄), and cold sodium hydroxide (NaOH), are widely used in the quantification of Al fractions in acid soils. Aluminum extracted by KCl is considered to represent salt-exchangeable Al, while Al extracted with CuCl₂ is used to estimate the potentially reactive non-exchangeable Al (Al_Cu) (Juo and Kamprath 1979). Likewise, extraction with Na–pyrophosphate at pH 10 (Al_p) has been proposed to estimate the Al in organic complexes. Thus the difference between Al_Cu and Al_p is considered to represent the Al that forms stable complexes with organic matter (Urrutia et al. 1988). Acid ammonium oxalate is commonly used to dissolve total non-crystalline Al (Al_o), including short-range order Al hydroxides and oxihydroxides, and Al bounded to organic matter (Higashi and Ikeda 1974; Kodama and Schnitzer 1971; McKeague and Day 1966; McKeague et al. 1971; Theng et al. 1982; Wada 1977). Similarly Al extracted with 0.5 M NaOH (Al_n) is used to estimate the total free Al pool in allophane, imogolite and gibbsite (Darke and Walbridge 1994; Eduardo Garcia-Rodeja et al. 2004; Wada 1980). Although the single extraction methods are efficient for the estimation of particular fractions of Al, various uncertainties are also associated with the use of these extractants. For instance, although CuCl₂ removes Al from organic matter via ligand exchange reactions, it can extract some inorganic Al from interlayer clay minerals (Hargrove and Thomas 1981; Juo and Kamprath 1979). Likewise Na₄P₂O₇ may extract some interlayer Al and relatively labile forms of surface-precipitated Al (Page and Kimpe 1989; Paterson et al. 1993; Soon 1993). Similarly, Kaiser and Zech (1996) have shown that pyrophosphate extracted amorphous Al hydroxides as well as crystalline gibbsite. It has been established that acid oxalate also extracted a portion of the hydroxy-Al interlayer material from 2:1 layer silicates (Farmer et al. 1988; Iyengar et al. 1981; Paterson et al. 1993). Eduardo Garcia-Rodeja et al. (2004) also reported that the amount of Al extracted by acid oxalate was significantly greater than that by cold NaOH.

The objective of this work was to compare, using some Japanese soil samples, Al fractions estimated from a sequential extraction procedure with those obtained by single extraction methods employing similar operating conditions to the corresponding sequential individual steps. The advantages and drawbacks of the two procedures were considered and discussed in the light of available literature.

Materials and methods

The soils

The soil samples used in this study were collected from 10 different regions of Japan. The samples were taken from the surface layer (0–20 cm).

General characterization of soil samples

The soils were air-dried, ground, and passed through a 1-mm screen to remove rocks, roots, and other large particles. Soil pH was measured in water (pH-H₂O) and 1 M KCl (pH-KCl). Fluoride reactivity was measured as the pH in 1 M sodium fluoride (NaF) after 2 min of equilibration (Fieldes and Perrot 1966). Total aluminum oxide (Al₂O₃) and silicon dioxide (SiO₂) contents of the soils were measured by X-ray fluorescence analysis (Primini, Rigaku, Japan). Total organic carbon content was determined using a vario MAX CN elemental analyzer (Elementar Analysen systeme GmbH, Hanau, Germany). Phosphate adsorption was determined by the standard method according to (Nair et al. 1984). Clay minerals of the soils were determined by X-ray diffraction (Wada 1966).

Aluminum extraction procedures

Single extraction procedures

Extractions consisted in shaking soil samples with solutions given below (Table 1), thus producing extracts containing one or several specified forms of Al. Following commonly used interpretations, the Al pools were classified as follows:

Alk	=	exchangeable Al.
Alp-AlCu	=	Al bounded to organic matter, in order of decreasing stability.
Alo-Alp	=	Al in allophane and imogolite.
Aln-Alo	=	crystalline Al.

Table 1. Experimental conditions of the single extraction procedure

Extracting reagent	Soil : solution	Shaken time	Liquid separation	References
1 M KCl (Alk)	5 g : 30 mL	Shaken for 30 min (extraction was repeated four times)	Centrifugation at 3000 rpm for 15 min and filtration with advantec paper #6	Lin and Coleman 1960
0.5 M CuCl2 (pH 2.8) (AlCu)	5 g : 50 mL	Shaken for 5 min, equilibrated for 12 h and shaken for 30 min	Filtration with advantec paper #6	Juo and Kamprath 1979
0.1 M Na–pyrophosphate at pH 10 (Alp)	1 g : 100 mL	Shaken for 16 h	Centrifugation at 2500 rpm for 15 min and filtration with advantec paper #6	Bascomb 1968
0.2 M ammonium oxalate at pH 3 (Alo)	1 g : 50 mL	Shaken for 4 h in dark	Centrifugation at 2500 rpm for 15 min and filtration with advantec paper #6	Blackmore 1981
0.5 M NaOH (Aln)	1 g : 100 mL	Shaken for 16 h	Centrifugation at 2500 rpm for 15 min and filtration with advantec #6 paper	Hashimoto and Jackson 1960

Sequential extraction procedure

The reagents and operating conditions for the sequential extraction procedure are presented in Fig. 1. The procedure was based on the single extraction methods but the shaking time and soil/extractant ratio in the first two steps were modified. The extraction was carried out in 50 mL polyethylene tubes, which were also used for centrifugation to minimize the possible loss of solid in the centrifuge-washing steps. After each extraction step the supernatant liquid was separated from the solid phase by centrifugation at 2500 rpm for 15 min. It was then decanted into polyethylene vessels and stored at 4°C before analysis.

Figure 1. Schematic diagram of the sequential extraction procedure employed. Abbreviations: CuCl₂, copper chloride; KCl, potassium chloride; Na₄P₂O₇, sodium pyrophosphate; NaOH, sodium hydroxide; NH₄-Oxalate, ammonium oxalate. s-Al_k, exchangeable aluminum (Al) sequentially extracted; s-Al_Cu, Al weakly bounded to organic matter sequentially extracted; s-Alp, Al strongly bounded to organic matter sequentially extracted; s-Al_o, Al in allophane or immoglite sequentially extracted; s-Aln, crystalline Al sequentially extracted. fr., fraction.

Aluminum in the extracts and silicon (Si) extracted with (NH₄)₂C₂O₄ were measured in duplicates using inductively coupled plasma spectroscopy (SPS3100, SII Nano Technology Inc., Japan).

Soil classification

The soils studied were classified into “allophanic” (A) and “non-allophanic” (N) by using the following criteria:

Non-allophanic	=	Alp/Alo ≥ 0.5%; Alp/Alo < 0.5%, Sio% < 0.5% and clay% ≥ 8%.
Allophanic	=	Alp/Alo < 0.5%, and Sio% ≥ 0.5%; Alp/Alo < 0.5%, Sio% < 0.5% and clay% < 8%, as defined by Masahiko Saigusa et al. (1993).

Allophane calculation

Allophane contents were calculated as follows (Dahlgren 1994):

The factor f depends on the Si/Al molar ratio and distinguishes Si-rich allophane (Al/Si, 1:1) and Al-rich allophane or imogolite (Al/Si, 2:1). Factor f can be determined as follows:

For an Al/Si ratio of 1:1, the factor is 5, for one of 2:1 the factor is 7.

Results and discussion

General characteristics of the soils

Table 2 shows some of the general parameters analyzed in the soils. The pH in water ranged between 4.4 and 6.9. Among the soils studied, A1 and A2 showed the highest aluminum oxide (Al₂O₃) content (33%). Similarly silicon dioxide (SiO₂) content in soil N2 was higher than those observed in other soils. The highest carbon content (12.8%) was recorded in the N3 soil.

Table 2. Location, pH, carbon content, phosphate retention and some general properties of the studied soils

Soil	Location	pH-H2O	pH-KCl	pH-NaF	Al2O3 (%)	SiO2 (%)	C (%)	Pads (mg kg^−1)	Clay (%)
A1	Tsukuba Ibaraki	6.6	4.6	11.1	33	55	4.41	196	–
A2	Katsuma Tochigi	5.9	5.4	11.2	33	57	1.68	239	–
A3	Nisine Iwate	6.6	6.0	9.5	25	57	4.87	152	–
A4	Kuroiso Nasu	6.9	5.7	10.8	26	61	4.81	159	–
A5	Takizawa Iwate	5.8	5.0	10.9	22	44	7.57	211	–
A6	Kawatabi Miyagi	5.4	4.8	11.4	22	68	0.68	154	7.2%
N1	Kawatabi Miyagi	6.0	4.5	11.4	20	67	10.42	183	–
N2	Kawatabi Miyagi	4.6	4.1	11.4	19	71	10.52	147	–
N3	Tonomi Hyogo	4.4	3.9	11.1	21	69	12.75	187	–
N4	Towono Iwate	6.1	5.4	10.2	22	53	6.57	114	–
Soil pH was measured in water (pH-H2O), 1 M potassium chloride (pH-KCl), 1 M sodium fluoride (NaF). Aluminum oxide (Al2O3), silicon dioxide (SiO2), carbon (C), phosphorus adsorption (Pads), “—” non-analyzed.

Table 3. Aluminum (Al) concentration (in g kg⁻¹) of each fraction extracted by single extraction methods, Al_p/Al_o ratio and classification of the soils studied

Soil	Alk	AlCu	Alp	Alo	Aln	AlCu–Alk	Alp–AlCu	Alo–Alp	Aln–Alo	Alp/Alo	Classification
A1	0.013	1.8	6.9	52.9	41.1	1.7	5.1	46.0	−11.8	0.1	Allophanic
A2	0.009	4.2	4.5	64.8	49.9	4.2	0.3	60.3	−14.9	0.1	Allophanic
A3	0.015	0.4	3.2	14.5	11.0	0.4	2.8	11.3	−3.6	0.2	Allophanic
A4	0.004	1.0	5.0	25.0	14.6	1.0	4.0	20.0	−10.4	0.2	Allophanic
A5	0.006	1.8	9.9	35.9	21.2	1.8	8.1	26.0	−14.7	0.3	Allophanic
A6	0.053	1.7	3.9	17.2	15.6	1.7	2.2	13.3	−1.6	0.2	Allophanic
N1	0.168	3.6	14.9	18.2	14.7	3.5	11.2	3.3	−3.5	0.8	Non-allophanic
N2	0.455	4.1	16.9	18.2	15.1	3.6	12.8	1.3	−3.0	0.9	Non-allophanic
N3	0.554	4.0	18.0	18.1	21.4	3.5	13.9	0.1	3.3	1.0	Non-allophanic
N4	0.004	1.6	6.6	12.8	7.5	1.5	5.1	6.1	−5.3	0.5	Non-allophanic
Abbreviations: Alk, exchangeable Al; AlCu, potentially reactive non-exchangeable Al; AlP, Al in organic complexes. Alo, total non-crystalline Al; Aln, total free Al; AlCu–Alk, Al weakly bounded to organic matter; Alp–AlCu, Al strongly bounded to organic matter; Alo–Alp, Al in allophane and imogolite; Aln-Alo, crystalline Al.

Aluminum fractionation by single extraction methods

Extractable Al contents in soils applying single extraction procedure are shown in Table 3. Extractable Al by three single extraction methods from KCl to pyrophosphate is in the order: Al_k < Al_Cu < Al_p. This was in good agreement with the results from Eduardo Garcia-Rodeja et al. (2004) and Fransco Matus et al. (2008). Extactable Al by pyrophosphate varied between 3.2 g kg⁻¹ and 18.0 g kg⁻¹, and was the highest extractable Al form after Al_n and Al_o; extractable Al in ammonium oxalate ranged from 12.8 g kg⁻¹ to 64.8 g kg⁻¹ and exhibited the greatest mean value; extractable Al in NaOH ranged from 7.5 g kg⁻¹ to 49.9 g kg⁻¹; excluding soil N3, Al_n values of the other soils were less than Al_o values, causing values of [Al_n–Al_o] negative. It indicates that there are no crystalline minerals in these soils.

The aim was to separate samples according to properties that are mainly related to the crystalline minerals, non-crystalline alumino-silicates, and Al–organic matter complexes (Masahiko Saigusa et al. 1993). According to Shoji et al. (1993) and Dahlgren et al. (2004), in Japan, allophanic andisols are typically younger than 10,000 years and have a high content of allophanic constituents, which contrasts with non-allophanic Andisols that are typically older than 10,000 years and have a low content of allophanic constituents. All the allophanic soils (A1–A6) had little pyrophosphate-extractable Al (<1%) but high oxalate-extractable Al (> 1%), causing Al_p/Al_o values <0.5%. Soil A2, is Akadama soil and developed from the loamy layer of the Kanto District. Although the soils A1 and A2 had the highest Al₂O₃ content (33%), they showed the lowest Al_p/Al_o ratio, indicating that these two soils are richer in allophane or imoglite than the other allophanic soils studied.

The non-allophanic soil, N1, was collected from the cultivated field, soil N2 from a no-till soil, soil N3 from a forest, and soil N4 from a vegetable plot. Carbon contents in non-allophanic soil, N1, N2 and N3 (>10%) were much higher than in allophanic soils. Since these soils are rich in pyrophosphate-extractable Al, they showed higher Al_p/Al_o ratio than allophanic soils. Further the results revealed that the non-allophanic soils are rich in Al–organic matter complex.

Aluminum fractionation by sequential extraction methods

Extractable Al contents in soils applying sequential extraction procedure are shown in Table 4. It is evident that the five-step sequential procedure (Sum-s-Al) extracted more Al (13.9–71.4 g kg⁻¹) than the single oxalate method (Al_o, 12.8–64.8 g kg⁻¹). Further, the fourth step (s-Al_o) of sequential procedure extracted greatest amount of Al in allophanic soils. It indicates that Al in allophanic soils mainly exist as non-crystalline Al. However in the case of non-allophanic soils, the highest amounts of Al found in the third step (s-Al_p) indicating the greatest proportion of Al in non-allophanic soils was associated with organic bound fraction. In contrast to estimation using the single extraction method (Al_n–Al_o), the positive s-Al_n values obtained with the sequential extraction, indicated the presence of crystalline-bound Al in both allophanic and non-allophanic soils.

Table 4. Aluminum (Al) concentration (in g kg⁻¹) of each fraction extracted by the five-step sequential procedure and the sum quantity of extractable Al (Sum-s-Al)

Soil	s-Alk	s-AlCu	s-Alp	s-Alo	s-Aln	Sum-s-Al
A1	0.049	1.7	3.5	36.3	17.6	59.1
A2	0.021	4.3	1.5	54.1	11.5	71.4
A3	0.015	0.4	2.5	10.4	11.1	24.4
A4	0.004	1.1	4.1	14.3	11.6	31.0
A5	0.006	1.8	9.5	20.4	8.3	40.1
A6	0.062	1.7	2.0	10.2	3.6	17.7
N1	0.143	2.8	10.6	3.2	2.8	19.6
N2	0.455	3.5	12.7	2.6	2.6	21.9
N3	0.554	3.5	11.2	3.1	7.2	25.5
N4	0.004	0.7	5.8	5.4	2.0	13.9
Abbreviations: s-Alk, exchangeable Al sequentially extracted; s-AlCu, Al weakly bounded to organic matter sequentially extracted; s-Alp, Al strongly bounded to organic matter sequentially extracted; s-Alo, Al in allophane or immoglite sequentially extracted; s-Aln, crystalline Al sequentially extracted.

Relationships amongst extractable aluminum between single extraction methods and sequential procedure

A simple correlation analysis was performed to assess relationships between the Al concentrations in the different Al fractions obtained by the single and sequential extraction procedures. The correlation coefficients for these relationships are listed in Table 5. It is evident that although the shaking time and soil/extractant ratio were modified in the first two steps, the amounts of Al released in the first four fractions extracted by the sequential procedure were correlated with the quantities extracted in the first four steps by the single extraction procedures. These results indicated that the first four steps in the sequential procedure could be as effective as the single extraction methods at estimating exchangeable Al, weakly organic bound Al, strongly organic bound Al, and non-crystalline Al. On the contrary, Al content extracted by Step 5, s-Al_n, in the sequential procedure was not correlated with the Al content in the Al_n–Al_o in the single procedure. This suggests the possibility that cold NaOH was not an efficient extractor in the crystalline Al form, especially without a pretreatment using (NH₄)₂C₂O₄ to remove amorphous Al and expose the crystalline Al form in the single extraction procedure.

Table 5. Relationships of extractable aluminum (Al) between single extraction methods and sequential method

Single methods	Alk	AlCu–Alk	Alp–AlCu	Alo–Alp	Aln–Alo
Sequential procedure	s-Alk	0.995***
	s-AlCu		0.938***
	s-Alp			0.934***
	s-Alo				0.985***
	s-Aln					0.318
P < 0.01, n = 10.
Abbreviations: Alk, exchangeable Al; AlCu-Alk, Al weakly bounded to organic matter; Alp-AlCu, Al strongly bounded to organic matter; Alo-Alp, Al in allophane and imogolite; Aln-Alo, crystalline Al.

Aluminum in allophane

According to Parfitt and Henmi (1982), a molar ratio (Al_o–Al_p)/Si_o close to 2 indicates the presence of allophane. The molar ratios of (Al_o–Al_p)/Si_o obtained from single extraction methods ranged from 1.6 to 5.1 in the allophanic soils and from 0.1 to 1.3 in non-allophanic soils (Table 6). However, with the sequential procedure, this ratio was close to 2 for all the allophanic soils tested, and ranged from 2.9 to 7.9 for non-allophanic soils (Table 6). These results clearly show evidence that the sequential procedure can distinguish between non-allophanic soils and allophanic soils based on the molar ratio (Al_o–Al_p)/Si_o. The molar ratio (Al_o–Al_p)/Si_o of soil A6 obtained from the single oxalate procedure had a very high value (5.1), because the single oxalate procedure could also extract Al from crystalline minerals (Farmer et al. 1988; Iyengar et al. 1981; Paterson et al. 1993) and probably also because of the lower content of carbon (Table 1) in this soil. This suggests that using single oxalate and pyrophosphate extractions may not be suitable for estimating the allophane contents in Andisols when the soil carbon content is low.

Table 6. Comparison of molar ratio aluminum/silicon (Al/Si) of allophane and allophane content calculated by single extraction methods and sequential procedure

Soils	Single extraction methods			Sequential procedure
	Sio (g kg^−1)	(Alo–Alp)/Sio	Allophane (%)	s-Sio (g kg^−1)	(s-Alo)/(s-Sio)	s-Allophane (%)
A1	21.5	2.2	15.0	17.9	2.1	12.6
A2	32.9	1.9	23.0	33.9	1.7	23.7
A3	7.2	1.6	5.1	4.6	2.3	3.2
A4	8.5	2.4	6.0	5.5	2.7	3.9
A5	13.4	2.0	9.3	9.5	2.2	6.6
A6	2.7	5.1	1.9	4.4	2.4	3.1
N1	2.9	1.2	1.5	0.5	6.1	0.4
N2	2.8	0.5	1.4	0.4	6.4	0.3
N3	2.8	0.1	1.4	0.4	7.9	0.3
N4	4.8	1.3	2.4	2.0	2.9	1.4
Abbreviations: Sio, Si in allophane or immoglite; Alo-Alp, Al in allophane and imogolite; s-Sio, Si in allophane or immoglite sequentially extracted; s-Alo, Al in allophane or immoglite sequentially extracted; s-Allophane, sequentially extracted allophane.

Allophane contents in soils calculated by applying both the sequential and single extractions are shown in Table 6. The allophane contents obtained from single extraction methods ranged from 1.9% to 23% in allophanic soils, and from 1.4% to 2.4% in non-allophanic soils. The allophane content obtained from the sequential procedure were 3.1% to 23.7% for the allophanic soils, and 0.3% to 1.4% for non-allophanic soils. In general, allophane and imogolite are formed preferentially at pH >4.9, because at pH values <4.9 complexation of organic matter with Al leaves no available Al to react with silica to form allophane or imogolite (Mizota and Van Reeuwijk 1989; Shoji and Fujiwara 1984). In the present case, because the pH of N2 and N3 soils are lower than 4.9 (Table 2), there is negligible possibility of the presence of allophane. Interestingly our results also show the sequential extraction to be more effective than the single extraction in evidencing the low allophone content in soils N2 and N3. Hence, it is possible to conclude that the sequential procedure provides more discriminatory results than the single extraction methods.

Crystalline aluminum

Negative Al_n–Al_o values (crystalline Al contents) found in the studied soil samples using the single procedure indicated that more Al_o was extracted than was Al_n. Farmer et al. (1988) and Paterson et al. (1993) also reported that acid oxalate can extract a part of the hydroxy-Al from silicates interlayer. Similarly, Eduardo Garcia-Rodeja et al. (2004) also reported that (NH₄)₂C₂O₄ was more efficient than 0.5 M NaOH in extracting Al from organic carbon-rich horizons (> 10% carbon). In contrast to the single extraction procedure, the sequential procedure revealed the existence of crystalline clays or minerals (crystalline Al-ranged from 2.0 to 17.6 g kg⁻¹) in the soil samples. Thus in order to confirm the existence of Al in crystalline minerals, all the soil samples were subjected to X-ray diffraction. Scans of magnesium ion (Mg²⁺) saturated oriented specimens of the clay fraction are presented in Fig. 2. The results of X-ray diffraction on the soil samples confirm the presence of appreciable amounts of more than three types of crystalline minerals, including gibbsite, vermiculite or halloysite. Our result implies that the sequential method to be more reliable in estimating the crystalline Al contents in the soil.

Figure 2. X-Ray diffraction diffractograms (Cu radiation) of oriented clay (<2 µm) specimen, air dry, magnesium ion (Mg²⁺) saturated. The scans correspond to the 2θ scale. Ch., chlorite; Vt., vermiculite; Gb., gibbsite; Qz., quartz; Kt., kaolinite; Sp., serpentine; Cb., cristobolites.

Conclusion

The proposed five-step sequential procedure can be as effective as the single extraction methods at estimating exchangeable Al, weakly organic bound Al, strongly organic bound Al fractions. However, the single extraction methods showed a high risk of overestimating allophane contents of non-allophanic soils. Further, the sequential extractions were more discriminatory with respect to Al/Si molar ratio and allophane and crystalline Al contents than single extraction methods. Moreover, the sequential procedure seems to be more reliable in evaluating the crystalline minerals in the soil. Hence, the use of sequential extractions should allow one to evaluate the extractable Al in soil samples.

Acknowledgments

The authors are grateful to Professor Masami Nanzyo (Graduate School of Agricultrual Science, Tohoku University) for collecting soil samples from Kawatabi Miyagi and to Ms Noami for her assistance in sample preparation.