Efficiency of gypsum application to acid Andosols estimated using aluminum release rates and plant root growth

Abstract

A great deal of information on the efficiency of gypsum or phosphogypsum to ameliorate acidity in highly weathered soils is available, but only limited information is available on the efficiency in acid Andosols, which possess large amounts of active aluminum (Al). We examined the effectiveness of gypsum application to non-allophanic Andosols (one humus-rich A horizon and two B horizons poor in humus) using extractable soil Al analyses (batch and continuous extraction methods) and a cultivation test using burdock (Arctium lappa). With gypsum amendment, pH(H₂O) values of the soil decreased from 4.5–4.7 to 4.2–4.4, whereas the treatment made almost no difference to the values of pH(KCl). Total active Al (acid oxalate-extractable Al) was hardly affected by gypsum for all samples. Potassium chloride-extractable Al definitely decreased with the addition of gypsum in all soils; however, the decrease was small (0.1–1.4 cmol_c kg⁻¹) and the values still exceeded “the threshold of 2 cmol_c kg⁻¹” for inducing Al toxicity in sensitive plants (4.4–8.6 cmol_c Al kg⁻¹). The change in Al solubility with gypsum application represented by Al release rates from soils using continuous extraction methods with a dilute acetate buffer solution (10⁻³ mol L⁻¹, pH 3.5) differed greatly among the soil samples: The release rate of one of the B horizon samples decreased by 71%, certainly showing the insolubilization of Al compounds, whereas the release rates of the A horizon sample showed almost no change. These changes in Al solubility were well correlated with the plant root growth. Root growth was improved with gypsum in the B horizon sample, whereas improvement was not observed in the A horizon soil. The decrease in the rate of Al release of another B horizon soil with gypsum treatment was smaller (by 20–34%), possibly because of lower pH values after gypsum application (pH[H₂O] of 4.2–4.3). In the B horizon soil, root growth improved only slightly. Thus, the effectiveness of gypsum application to acid Andosols appeared to be largely influenced by soil humus contents and slight differences in soil pH values, and corresponded to a decrease in Al release rates using the continuous extraction method.

Key words:

• aluminum release rates from soils
• aluminum toxicity
• Andosols
• exchangeable aluminum
• gypsum

INTRODUCTION

Andosols, the representative soils formed from volcanic ejecta, generally possess excellent physical properties such as low bulk density, high permeability and high water holding capacity (Nanzyo et al. 1993a). In contrast to these good physical properties, Andosols show high phosphate sorption capacity, which is the main chemical problem in their agricultural use. Another chemical problem of the soils is soil acidity. Acid injury to plant roots, mainly aluminum (Al) toxicity, usually occurs in Andosols poor in allophanic materials (non-allophanic Andosols) (Nanzyo et al. 1993b). Because exchange acidity (y₁, 1 mol L⁻¹ KCl exchangeable acidity) is a good indicator for Al toxicity, toxic Al is believed to be mainly the exchangeable Al on the permanent charge of 2:1 aluminosilicates (Saigusa et al. 1980). However, in non-allophanic Andosols dominated by Al-humus complexes, the exchangeable Al is largely controlled by organically complexed Al (Takahashi et al. 1995, 2003, 2006). This indicates that these soils contain a large number of precursors of toxic Al; nontoxic Al can easily be converted to toxic Al by decreasing the pH or increasing the ionic strength. Non-crystalline or poorly-crystalline minerals, such as allophane and imogolite, show a high buffer capacity to Al dissolution in their dominance (allophanic Andosols). In fact, allophanic Andosols seldom exhibit Al toxicity. In cultivated allophanic soils, however, lowering soil pH by the heavy application of fertilizers resulted in dissolution of Al from the allophanic materials (Matsuyama et al. 2005).

Lime has generally been applied to acid soils to counter Al toxicity. However, liming of surface soils is not effective for the amelioration of subsoil acidity because of its low solubility. Gypsum or phosphogypsum, which have higher rates of solubility, are believed to ameliorate subsoil acidity with surface application. Considerable information on their efficiency to ameliorate acidity in highly weathered soils is available (Farina and Channon 1988; Pavan et al. 1982; Shainberg et al. 1989; Sumner et al. 1986; Sumner and Carter 1988; Sumner 1993). In contrast, there is very little information on the efficiency of gypsum or phosphogypsum in Andosols or Andisols and the tendency of the efficiency among soils has not been clear. Mora et al. (1999) showed that gypsum treatment of acid Andisols rich in organic matter increased soil pH slightly, reduced exchangeable Al and increased the yield of ryegrass in a similar way to calcitic and dolomitic liming. The effect of gypsum is also related to the supply of sulfur (Mora et al. 1999). Saigusa and co-workers (Saigusa et al. 1996; Saigusa and Toma 1997; Toma and Saigusa, 1997) reported that phosphogypsum application to an Andosol with low humus content was effective for plant root elongation and reduction of exchange acidity, whereas the application to an Andosol with high humus content was barely effective. Inoue et al. (2001) also reported that the application of gypsum to an acid Andosol with high humus content scarcely reduced the exchangeable Al of the soil. Thus, consistency has not been shown in the effect of gypsum or phosphogypsum to Andosols or Andisols. Although Saigusa et al. (1996), Saigusa and Toma (1997) and Toma and Saigusa (1997) found an effect of phosphogypsum on one Andosol with low humus contents, as previously noted, the universality of the effectiveness on such Andosols has not yet been confirmed.

Variation in the efficiency of gypsum or phosphogypsum to acid Andosols may be related to the complexity of Al solubility of the soil. Acid Andosols contain large amounts of active Al in various forms, such as easily exchangeable Al, organically complexed Al and allophanic materials. The exchangeable Al in Andosols, which is used as an indicator of toxic Al, was largely influenced by organically complexed Al and by allophanic materials as well (Dahlgren et al. 2004). This indicates that exchangeable Al in Andosols is not the Al form directly reflecting the Al solubility or toxicity of soils, although many previous studies evaluated the efficiency of gypsum using exchangeable Al.

In the present study, we investigated the efficiency of gypsum application to non-allophanic Andosols (one A horizon sample and two B horizon samples) using soil analyses and plant culture tests in a closed experimental system in the laboratory. The effect on extractable Al from soil samples was evaluated not only by using batch extractions but also by using the Al release rate calculated from continuously extracted Al concentrations from soils using acidic buffer solution (pH 3.5, 10⁻³ mol L⁻¹ sodium acetate) (Dahlgren and Walker 1993; Takahashi et al. 1995). It was suggested that the method of Al release rates from soils would be a good indicator for the effectiveness of gypsum application to acid Andosols (Toma 1996).

MATERIALS AND METHODS

Soil samples and their characterization

One A horizon sample and two B horizon samples of non-allophanic Andosols were used in this study (Table 1). As a comparison, we used a Kakogawa Yellow soil sample.

Table 1 Sampling points, classification and clay mineralogy of the soil samples

Sample name	Horizon	Sampling point	Classification (US Soil Taxonomy)	Crystalline clay minerals in clay fraction	References
Kawatabi	A	Sawai, Naruko Town, Miyagi Prefecture	Pachic Melanudand	HIV >> Chl, It	Ito and Saigusa (1996)
Fujisawa	B	Fujisawa, Kitakami City, Iwate Prefecture	Andic Dystrudept	HIV >> Chl, It	Shoji et al. (1985)
Rokuhara	B	Kanegasaki, Kitakami City, Iwate Prefecture	Andic Dystrudept	HIV >> Chl, It	Shoji et al. (1985)
Kakogawa	B	Kakogawa City, Hyougo Prefecture	Typic Hapludult	Kt >> 2:1 min	Nanzyo et al. (1995)
Chl, chlorite; HIV, hydroxy-interlayered vermiculites; It, illite; Kt, kaolinite; 2:1 min, 2:1 type minerals.

This pedon was classified as Udults in Soil Taxonomy and the clay fraction was dominated by kaolinite (Table 1).

Air-dried, fine-earth fractions (<2 mm) were used for the laboratory analyses that follow. The soil pH was potentiometrically measured in water and 1 mol L⁻¹ KCl (soil:solution = 1:2.5) following a 30 min and 1 h equilibrium period, respectively. The amount of organic carbon was determined using dry combustion. Exchangeable bases were determined using extraction with 1 mol L⁻¹ ammonium acetate (pH 7). Selective dissolution techniques were used to determine the various pools of Al, Fe and Si. They are: (1) 1 mol L⁻¹ KCl extractable Al (exchangeable Al, Al_K) (Blakemore et al. 1981), (2) 0.5 mol L⁻¹ CuCl₂ extractable Al (Al_Cu) (Hargrove and Thomas 1981), (3) sodium pyrophosphate (pH 10) extractable Al and Fe (Al_p, Fe_p) (McKeague 1967), (4) acid ammonium oxalate (pH 3) extractable Al, Fe and Si (Al_o, Fe_o, Si_o) (McKeague 1976), (5) dithionite–citrate–bicarbonate extractable Fe (Fe_d) (Mehra and Jackson 1960). The contents of Al and Fe in all the extracts were measured using atomic absorption spectrophotometry and the Si content was colorimetrically determined. All soil analyses were carried out in duplicate.

The rates of Al release from the <0.5 mm soil fraction were determined using a stirred flow cell as described by Dahlgren and Walker (1993) and Takahashi et al. (1995). A reaction solution consisting of 10⁻³ mol L⁻¹ sodium acetate adjusted to pH 3.5 was used for the Al release experiment. The reacted solution was continuously injected into an autoanalyzer using the pyrocatechol violet method for quantification of monomeric Al (Rogeberg and Henriksen 1985). The measurement was carried out in triplicate.

Gypsum treatment

Field moist, undried soil samples were used for the gypsum treatment. The following rates of gypsum (CaSO₄·2H₂O) were mixed well with the soil samples: 4.3 g kg⁻¹ soil samples (Gypsum 1) and 8.6 g kg⁻¹ soil samples (Gypsum 2). The rate of gypsum in the Gypsum 2 plot was based on the Ca amount of the lime requirement of Kawatabi A horizon soil. The mixture of soil samples and gypsum was incubated at field water capacity (60% of maximum water holding capacity) at room temperature for 30 days.

Subsamples of the mixture were air-dried and passed through a 2-mm or 0.5-mm sieve. The same soil characterization analyses used for the original samples were carried out for the gypsum-treated samples except for total carbon content and phosphate retention. The remaining subsamples were used for plant culture.

Plant culture

A plant culture was carried out in the gypsum-treated and original soils. The soil samples were added to 200 mL pots (50 g dry soil basis for Kawatabi soil and 100 g for the other soils). The water content was adjusted and kept at 70% of the maximum water holding capacity. Ten seedlings of burdock (Arctium lappa cv. Kantan) per one pot were sown and cultured at room temperature (23–25°C). After a 2-week culture, the plants were harvested and the length of the roots and shoots was measured. Withering individuals were excluded from the measurement.

Statistical analyses

We used JMP 4.0 (SAS Institute Inc.) for the statistical analyses. A two-way anova (gypsum treatments × soils) was carried out for Al_o, Al_Cu–K and Al_K values. Mean comparisons were carried out using Tukey's tests for the total released Al from the soil samples at 60 min calculated from the experiment of Al release rates, and for the results of the plant culture experiment.

RESULTS

Soil characterization

Kawatabi A, Fujisawa B and Rokuhara B soils showed characteristics for Andosols, such as high acid oxalate-soluble Al and Fe (Al_o and Fe_o) concentrations and high phosphate retention (Table 2). Among these soils, the total carbon concentration of Kawatabi A soil was very high, whereas the concentration in Fujisawa B and Rokuhara B soils was low. All the soil samples contained small amounts of allophanic materials (Si_o 0.7–2.8 g kg⁻¹). The dominant crystalline minerals in the clay fractions of all the soil samples were chloritized 2:1 minerals (Table 1).

In contrast to the Andosols, Kakogawa Yellow soil contained small amounts of Al_o and Fe_o (Table 2).

Table 2 Selected properties of the soil samples

	T-C (g kg^−1)	pH (H2O)	pH (KCl)	Exchangeable cations (cmolc kg^−1)					AlCu (g kg^−1)	Alp (g kg^−1)	Alo (g kg^−1)	Feo (g kg^−1)	Fed (g kg^−1)	Sio (g kg^−1)	P-retn (%)
				Ca	Mg	K	Na	Al
Kawatabi A	111	4.5	4.0	0.59	0.23	0.37	0.14	4.88	8.34	14.7	18.0	10.0	15.3	2.53	93.5
Fujisawa B	7.7	5.1	3.8	0.11	0.92	0.19	0.28	7.20	3.01	4.7	10.3	13.9	33.6	2.80	87.0
Rokuhara B	6.6	4.7	3.9	0.25	0.19	1.01	0.09	8.99	2.99	4.8	7.5	8.8	34.2	0.67	81.5
Kakogawa B	2.8	4.6	3.6	0.16	0.43	0.13	0.13	5.45	0.92	1.0	2.0	1.5	16.7	0.14	41.0

The large difference between the concentrations in dithonite–citrate–extractable Fe (Fe_d) and Fe_o (Table 2) indicates that almost all of the free iron is crystalline. The clay fraction was dominated by kaolinite (Table 1). These properties of Kakogawa Yellow soil are similar to those of Ultisols or Oxisols in which the efficiency of gypsum treatment is well known (Farina and Channon 1988; Pavan et al. 1982; Sumner et al. 1986).

The soil pH(H₂O) of the four soil samples ranged from 4.5 to 5.1, showing extreme acidities. The amounts of exchangeable Al (Al_K) were very high (range: 4.8–9 cmol_c kg⁻¹) and exceeded the threshold of 2 cmol_c kg⁻¹ for inducing Al toxicity in sensitive plant roots (Saigusa et al. 1980; Soil Survey Staff 1999).

Effect of gypsum amendment on the chemical properties and Al solubility of soils

The pH(H₂O) values decreased markedly in most soils with gypsum treatment, by as much as 0.7 units (Table 3). This is more attributable to the salt effect of gypsum leading to an increase in H⁺ and Al ions than to the opposite reaction of ligand exchange of OH⁻ and SO²⁻ ₄ (Garrido et al. 2003; Shainberg et al. 1989). The change in the pH(H₂O) values of Kawatabi A soil was small, reflecting the stronger buffer reaction of this soil. The pH(KCl) values were not affected or were only slightly lower with gypsum treatment.

The concentration of Al_o, which corresponds to total active Al, was scarcely changed by the application of gypsum in all soil samples (Fig. 1). The Al_p concentration, oraganically complexed Al, was also barely affected (data not shown). The difference between copper chloride-extractable Al (Al_Cu) and exchangeable Al (Al_K) (Al_Cu–K) is believed to be organically complexed Al for soils with high humus contents (Aitken 1992; Conyers 1990; Juo and Kamprath 1979). The Al_Cu–K is also attributable to Al-polymeric species and metastable forms of Al hydroxides and hydrous oxcides in soils with low humus contents

Table 3 Soil pH values before and after gypsum treatment

	Control	Gypsum 1	Gypsum 2
pH(H2O)
Kawatabi A	4.5	4.3	4.3
Fujisawa B	5.1	4.5	4.4
Rokuhara B	4.7	4.3	4.2
Kakogawa B	4.6	4.0	4.1
pH(KCl)
Kawatabi A	4.0	3.9	3.9
Fujisawa B	3.8	3.7	3.7
Rokuhara B	3.9	3.7	3.8
Kakogawa B	3.6	3.6	3.6
The difference in the measured pH values between the two repetitions was less than 0.03.

Figure 1  Effect of gypsum treatments on acid oxalate-extractable Al (Alo). The results of a two-way anova showed no significant difference in the Alo values among gypsum treatments (P > 0.05), but there was a significant difference in Alo values among soils (P < 0.01).

Figure 2  Effect of gypsum treatments on the value of the difference between CuCl2-extractable Al and KCl-extractable Al (AlCu–K). The results of a two-way anova showed no significant difference in the AlCu–K values among gypsum treatments (P > 0.05), but there was a significant difference in AlCu–K values among soils (P < 0.01).

(Juo and Kamprath 1979). The Al_Cu–K values slightly increased with gypsum in Kawatabi A soil (Fig. 2). This is likely to have occurred because organically complexed Al is more easily extracted with CuCl₂ by gypsum treatment (Garrido et al. 2003; Vizcayno et al. 2001). However, the values did not increase and even slightly decreased in the other two Andosol and Kakogawa soil samples (Fig. 2), indicating that the formation of the Al polymers with gypsum treatment was not significant. The amounts of exchangeable Al (Al_K) definitely decreased with gypsum in all soils (Fig. 3). However, the decrease was small (0.1–1.4 cmol_c kg⁻¹) and still exceeded the threshold of 2 cmol_c kg⁻¹ for inducing Al toxicity in sensitive plants.

Thus, all the batch methods of Al extraction were not significantly affected by the application of gypsum. On the contrary, Al release rates using continuous extraction of soils with a dilute acetate buffer solution (10⁻³ mol L⁻¹,

Figure 3  Effect of gypsum treatments on KCl-extractable Al (AlK). The results of a two-way anova showed significant differences in the AlK values both among gypsum treatments (P < 0.01) and among soils (P < 0.01).

pH 3.5) significantly changed with gypsum treatment in some soil samples. As seen in Fig. 4, after gypsum treatment, the Al release rate from Fujisawa soil dropped considerably. The decrease rate as compared with the Control plot at 20 min was 53% for the Gypsum 1 plot and 71% for the Gypsum 2 plot. The decrease in Al release rates was also observed in Rokuhara soil; the decrease rate at 20 min was 20% for Gypsum 1 and 34% for Gypsum 2 (Fig. 5). Figure 6 shows the comparison in the decrease of Al_K and total released Al calculated from the Al release rates during the 60 min experiment. In Fujisawa soil, the decrease in the total released Al far exceeded the decrease in Al_K. This trend was also observed in the Rokuhara soil, but the decrease in Al release was less than that in Fujisawa soil. This is probably because of the lower pH(H₂O) values of the Rokuhara soil after the gypsum treatments.

In contrast, gypsum treatment did not affect the Al release rates in Kawatabi A soil, which is abundant in organic matter (Fig. 4). The Al release rates from Kakogawa soil were very slow and the effect of gypsum was not clear. Nevertheless, the application of gypsum resulted in the acceleration of the Al release rates in the first few minutes and the subsequent decrease in the rates (Fig. 5).

The decrease in the Al release rates indicates insolubilization of Al and the increase indicates solubilization, although the change in the specific Al fractions cannot be identified.

Figure 4  Aluminium release rates (upper) and cumulative amounts of Al released (lower) from Kawatabi A and Fujisawa B soils. The numbers in the lower figure of Fujisawa B show the amount of total released Al at 60 min. Data with different letters are significantly different at P < 0.05 using a Tukey's test. There was no significant difference in the amount of total released Al in Kawatabi A soil among the treatments.

Figure 5  Aluminium release rates (upper) and cumulative amounts of Al released (lower) from Rokuhara B and Kakogawa B soils. The numbers in the lower figures shows the amount of total released Al at 60 min. Data with different letters are significantly different at P < 0.05 using a Tukey's test.

Figure 6  Comparison of the decrease in AlK with gypsum treatment (Gypsum 1) with total released Al using acetate buffer solution.

Effect of gypsum application on plant growth

As shown in Fig. 7, the effects of gypsum application on the growth of plant roots and shoots differed greatly among the soils. In the Kawatabi soil, root growth was strongly restricted in the Control plot and the restriction of root growth was not improved with gypsum. In the Fujisawa soil, although Al injury was observed in roots in the Control, root growth recovered markedly in the Gypsum 1 and 2 plots and this influenced shoot growth. This improvement in root growth with the application of gypsum corresponds to the decrease in Al release rates from soils (Fig. 4). The decrease in Al release rates with gypsum in Rokuhara soil was remarkable compared with the rate in Fujisawa soil (Fig. 5). However, the effect of gypsum on plant root growth was not observed in Rokuhara soil (Fig. 7). Application of gypsum to Kakogawa soil reduced root growth (Fig. 7); this might correspond to the accelerated Al release from gypsum-amended soils during the first few minutes (Fig. 5) or to the decrease in soil pH with gypsum (Table 3).

DISCUSSION

A decrease in the Al solubility of soil with the application of gypsum is attributable to the formation of hydroxy Al polymers and Al hydroxysulfate minerals (Garrido et al. 2003; Saigusa and Toma 1997; Sposito 1989; Sumner 1993). The formation of Al compounds is expected to reflect some soil chemical data. In particular, the formation of hydroxy Al polymers would reduce Al_K and increase

Figure 7  Effect of the application of gypsum on shoot and root growth in burdock (n = 7–10). Error bars indicate standard deviation. Data with different letters are significantly different at P < 0.01 using a Tukey's test.

Al_Cu–K values (Garrido et al. 2003). In the present study, however, the increase in Al_Cu–K values with the application of gypsum was very small in Kawatabi and Kakogawa soils and a slight decrease in the values was observed in Fujisawa and Rokuhara soils (Fig. 2). The decrease of Al_K with gypsum was not as large as expected (0.1–1.4 cmol_c kg⁻¹, maximum 19% of decrease) (Fig. 3), and the amount of Al_K in all soils still exceeded the threshold of 2 cmol_c kg⁻¹ for inducing Al toxicity in sensitive plants. These results must be because of the low pH values for both pH (H₂O) and pH (KCl) of soils with gypsum amendment, indicating that the formation of hydroxy Al polymers was marginal.

The change in the Al solubility of soils with gypsum that was demonstrated by the Al release rates definitely differed among the soils (Figs 4,5). The Al release rates in Kawatabi soil did not decrease with the application of gypsum. This can be explained by the fact that almost all the active Al is in Al-humus complexes and this form of Al is hardly reduced and even increased by the application of gypsum as revealed by the Al_K and Al_Cu–K values. These results convince us that gypsum amendment to acid Andosol horizons rich in organic matter is scarcely effective (Inoue et al. 2001; Saigusa et al. 1996). In contrast, Al release rates from soil with low humus contents (Fujisawa and Rokuhara soils) decreased greatly with gypsum treatment. This may be explained by the fact that easily exchangeable Al absorbed on permanent negative charges of 2:1 minerals was exchanged and precipitated as hydroxy Al polymers, and the polymers absorbed irreversibly on the soil surface (Saigusa and Toma 1997). In the present study, however, the Al_Cu–K values did not increase with gypsum treatment and, thus, the formation of hydroxy Al polymers in these soil samples would be reduced. The decrease in the amounts of Al released by dilute acetate buffer largely exceeded the decrease in the amounts of Al_K in Fujisawa and Rokuhara soils (Fig. 6). The formation of Al hydroxysulfate minerals is another possibility for lowering the Al release rates (Shainberg et al. 1989; Sumner 1993). Although X-ray diffraction analysis did not show the presence of crystalline Al hydroxysulfate minerals in Gypsum 1 and 2 plots of Fujisawa and Rokuhara soils (data not shown), the formation of amorphous Al-sulfate compounds might exist. It has been very difficult to directly confirm the existence of these minerals in soils. Recently, Delfosse et al. (2005) found these minerals in the clay fractions of Andosols, simply by dispersing clays with Na-resins, with transmission electron microscopy coupled with energy-dispersive analysis.

Improvement in root growth with gypsum treatment was observed only in Fujisawa B soil, which contains small amounts of humus. An effect of gypsum was not observed in Rokuhara B soil with a low humus content and Kawatabi A with high humus content. In Kakogawa Yellow soil, gypsum application reduced root growth. The tendency of the gypsum effectiveness to root growth did not well harmonize with the decrease of exchangeable Al (Al_K)d with the addition of gypsum. (Figs 3,7), although Saigusa et al. (1996), Saigusa and Toma (1997) and Toma and Saigusa (1997) concluded that the improvement of plant root growth with gypsum or phosphogypsum was explained by a decrease in exchange acidity, which is mostly composed of exchangeable Al. The Al_K values in Fujisawa soil, in which root growth improved, decreased by up to 19% with gypsum (Gypsum 2 plot). However, the Al_K values of Gypsum 1 and 2 were still greater than 2 cmol_c kg⁻¹, the threshold for inducing Al toxicity. Although other soils also decreased Al_K values with gypsum, improvement of root growth was not observed in these soils.

A good correspondence between the effects of gypsum on root growth and Al release rate was observed in Kawatabi A and Fujisawa B soils (Figs 4,7). In Kawatabi soil, gypsum application did not affect either root growth or Al release rates. In contrast, Al release rates from Fujisawa soil decreased greatly with gypsum, and this reflected root growth. In Rokuhara soil, Al release rates also decreased in Gypsum 1 and 2 (Fig. 5), but root growth did not improve (Fig. 7). This may be because of the high level of Al release rates in Gypsum 1 and 2 plots, possibly because of the low pH values (pH[H₂O] of 4.2–4.3) after gypsum application. Although Al release was very slow from Kakogawa soil, gypsum application lead to an increase in Al release during the first few minutes (Fig. 5). This increase in Al release is related to further restriction of root growth (Fig. 7). As expected from the very low pH values of Kawatabi, Rokuhara and Kakogawa soils (Table 3), proton injury might have occurred in these soils (Foy 1984).

Acid Andosols contain a much greater amount of active Al compared with other acid soils. It was confirmed that, in the case of soils with high humus contents, organically complexed Al is dominant and this form of Al is hardly affected by gypsum application. It is unlikely to form enough hydroxy Al polymer even in B horizon soils poor in humus to consume high amounts of exchangeable and active Al when the soil pH does not increase with gypsum addition. In these cases, even H⁺ injury to plant roots may occur at pH levels below approximately 4.2 (Foy 1984). Therefore, materials that result in an increase in soil pH would be more effective for acid Andosols. Otherwise, materials, such as chelate compounds, that can make strong complexes with Al and remove them from soils into below subsoils would be expected.

Notes

Present addresses: Johnson and Johnson K. K., 5-2 Nishi-kanda 3-chome, Chiyada-ku, Tokyo 101-0065, Japan.

Shizuoka Agricultural Experiment Station, 678-1 Tomigaoka, Iwata, Shizuoka 438-0803, Japan.