Vertical distribution of chloroform-extractable green fraction in several Andosol and Cambisol profiles

Abstract

The vertical distribution of the content of the chloroform-extractable green fraction (CEGF) in the soil profiles of four Japanese Andosols, one Japanese Cambisol and one Nepalese Cambisol was examined using a colorimetric method with an ultraviolet and visible spectrophotometer. In addition, the relationship between the CEGF content and the soil chemical properties was examined by statistical analysis. CEGF was detected in all the soil samples, and the content ranged from 0.02 to 1.16 g kg⁻¹. The CEGF content in the surface A horizon of the Andosol tended to be higher than in the Cambisol. The vertical distribution of the CEGF content in three soil profiles (two Andosol profiles and one Nepalese Cambisol) revealed maximum values in the subsurface horizon. In the Japanese Cambisol, the content of the CEGF was highest in the surface horizon. In the remaining two Andosol profiles containing a buried humus horizon, the content of the CEGF was highest in the buried humus horizon and the distribution was considerably different from that of organic carbon. Statistical analysis showed that the content of the CEGF displayed a significant positive correlation (P < 0.001, n = 36) with the contents of dithionite-citrate-bicarbonate-extractable Al (Ald), and with oxalate-oxalic acid-extractable Fe and Al (Feo and Alo, respectively). Furthermore, the content of the CEGF showed a significant positive correlation with the content of Alo (r = 0.944, P < 0.01, n = 6) in the soil with the highest content of CEGF in the profile. It was, therefore, suggested that CEGF is actively produced in humus-rich soils like Andosols, and that Ald, Feo and, in particular, Alo were associated with the vertical distribution of CEGF in the soil profiles.

Key words:

• chloroform-extractable green fraction
• Pg
• sclerotia
• soil green pigment
• vertical distribution

INTRODUCTION

In a previous paper, a colorimetric method developed for the estimation of the content of the chloroform-extractable green fraction (CEGF) in soils was described (Kobayashi et al. 2005). The CEGF, which exhibits strong absorption maxima at 454, 440 and 418 nm in chloroform, is considered to be of microbial origin and one of the components of, or a closely related substance to, the green fraction of humic acid (HA) called Pg (Kumada 1987; Watanabe et al. 1996) as previously reported (Kobayashi et al. 2005). It is important to note that in a previous paper (Kobayashi et al. 2005), CEGF was found to occur in various soils and that the method for estimating its content could be applied to various soils. Pg is distributed worldwide in various soils (Kumada 1987; Zancada et al. 2003) and numerical studies on the distribution, origin (Kumada and Hurst 1967; Sato 1976b; Valmaseda et al. 1989) and chemical properties (Kumada and Sato 1980; Watanabe et al. 1996) of Pg have been conducted in view of the importance of Pg, as one of the components of HA (Kumada 1987), and of its role in the mobilization and transport of iron through its chelating ability (Nakabayashi and Wada 1991). Furthermore, it has been reported that a Pg-like pigment occurs in the soil fulvic acid fraction (Schnitzer and Skinner 1968), humin fraction (Tsutsuki and Kuwatsuka 1989) and in dark-colored soil fungi (Sato 1976b; Valmaseda et al. 1989). On the basis of these studies, it is assumed that both Pg and the Pg-like pigment are distributed worldwide in various soils. However, the quantitative distribution of these soil green pigments is poorly documented because of the lack of appropriate methods for determining the content, except for Pg, and because the method for estimating Pg content in soil cannot be applied to humus-rich soils (Sato 1974). It was, therefore, concluded that the factors described above could not be well established without quantitative determination of Pg and Pg-like pigments in various soils.

Although CEGF is only a closely related substance to Pg, it is, at present, the only fraction that enables the estimation of the content of green pigments in various soils. Therefore, studies on the quantitative distribution of CEGF in various soils should contribute to studies on soil green pigments including Pg. The purpose of the present study was to investigate the vertical distribution of the CEGF content in four Andosol profiles and two Cambisol profiles using the method described by Kobayashi et al. (2005). In addition, the relationship between the CEGF content and several soil chemical properties was examined.

MATERIALS AND METHODS

Soil samples

The soil profiles used in the present study were as follows: three Japanese Silic Andosols, namely Sugadaira-G (SGG), Sugadaira-F (SGF) and Soni (SN), one Japanese Umbric Andosol, Gyosei (GS), and one Japanese Dystric Cambisol, namely Kuroiwa (KI), where G and F stand for grassland and forest, respectively. In addition, one Nepalese Dystric Cambisol profile, Jiri-F (JR), was used. Although the JR profile contained hydroxyanthraquinone pigments at the same level as the Japanese Cambisol (Suzuki et al. 1999b), the soil environment was very different from that of the Japanese soil. Detailed information about the sites and soil properties was presented in Suzuki et al. (1999a). Soil samples were taken from each horizon of the soil profile. The sample was air-dried and passed through a 2-mm mesh sieve. A portion of the sample was ground to pass through a 0.5-mm mesh sieve.

Soil chemical analyses

Exchangeable Al (Ex. Al) of the soil samples was extracted with 1 mol L⁻¹ KCl (Bertsch and Bloom 1996) and the concentration was subsequently determined using Shimadzu ICPS-1000III (Shimadzu Co., Kyoto, Japan). The remaining chemical data for the soil samples were from Suzuki et al. (1999b).

Classification of HAs and estimation of Pg content in soils

The HAs of the soil samples were prepared according to the method described by Oba (1964). The HAs were dissolved in 0.1 mol L⁻¹ NaOH and used for the classification of HAs and for the estimation of the Pg content in soil. The log (A ₄₀₀/A ₆₀₀) (= Δlog K, A _λ: absorbance at λ nm in alkaline solution) value of the HA solutions was determined using the method of Oba (1964). The volume of 0.02 mol L⁻¹ KMnO₄ consumed by the HA solutions (30 mL), except for the HAs from the JR profiles, was determined using the method of Oba (1964). Subsequently, the volume was transformed into the organic carbon concentration using a conversion factor (0.0157) described by Ikeya and Watanabe (2003). For the JR profiles, the organic carbon concentration of the HA solution was determined using a dissolved C analyzer (Shimadzu TOC-V_CPH), according to the method of Ikeya and Watanabe (2003). The classification of HAs was carried out using the A ₆₀₀/C-log (A ₄₀₀/A ₆₀₀) diagram (A ₆₀₀/C; the absorbance at 600 nm mg⁻¹ C mL⁻¹ of HA in 0.1 mol L⁻¹ NaOH) described by Ikeya and Watanabe (2003). Based on their log (A ₄₀₀/A ₆₀₀) and A ₆₀₀/C values, HAs were classified into 4 types, Type A, Type B, Type P and Type Rp, and their degree of humification increased in the order of Type A, Type B and Type Rp (Kumada 1987). Type P HA is a group with absorption maxima near 615, 570 and 450 nm in alkaline solution because of the presence of Pg (Kumada 1987). Subdivision of HAs was carried out using the method of Kumada et al. (1967). In this method, Type P HA is subdivided into Type P_±, P₊, P₊₊ and P₊₊₊, depending on the intensities of the absorption maxima near 615, 570 and 450 nm because of the presence of Pg. In cases where the same absorption maxima were found in the HA except for Type P, they were also subdivided into Type P HA (e.g. A_± and Rp₊₊).

The Pg content in the HA solutions prepared was estimated according to the method developed by Sato (1974). This method is a geometrical method based on the absorption curve of Type P HA in alkaline solution.

Estimation of CEGF content in soils

The CEGF of each soil sample (0.5–2 g) was extracted three times with HCl-dimethylsulfoxide (DMSO) solvent (DMSO in 0.2 mol L⁻¹ HCl) at a soil/solvent ratio of 1/20 (w/v) for 24 h in the first extraction, 1/20 (w/v) for 6 h in the second extraction and 1/10 (w/v) for 30 s in the third extraction. In the present study, the ratio of extractant was twofold that indicated in our previous paper (Kobayashi et al. 2005) because it was considered that CEGF could not be completely extracted from the

Figure 1  Visible spectra of chloroform-extractable fraction in chloroform from (a) Andosol (Sugadaira-G [SGG]) and (b) Cambisol (Kuroiwa [KI]) profiles.

soil with the ratio used in the previous paper. After alkalinization of the combined extracts, the supernatant was acidified and then extracted with chloroform. A detailed procedure for the extraction of CEGF was described in Kobayashi et al. (2005). The component in the chloroform solution was designated as chloroform-extractable fraction (CEF). The content of CEGF in the chloroform solution of CEF from each soil sample was estimated using the colorimetric method described in Kobayashi et al. (2005).

Ultraviolet and visible spectral analysis

The ultraviolet and visible (UV–Vis) spectra of HAs in 0.1 mol L⁻¹ NaOH and of CEFs in chloroform were measured using a JASCO V-530 spectrophotometer (Jasco Co., Tokyo, Japan).

Statistical analysis

All statistical analyses were carried out using SPSS for Windows (ver. 11.5 J, SPSS Inc., Tokyo, Japan). Correlations were obtained using the Pearson correlation coefficient. P < 0.05 was considered to be significant.

RESULTS

Classification of HAs and Pg content in soils

Table 1 shows the CEGF and Pg contents, soil chemical properties and the properties of HAs in the soil samples. In all Andosol profiles (SGG, SGF, SN and GS), most of the HAs were classified into Type A. In the Cambisol profiles, HAs from the KI profile were classified into Type P or Type Rp, and those from the JR profile were classified into Type P.

Although Pg was detected in all samples (Table 1), the method for estimating the content of Pg in the soil could only be applied for the Type P HA soil (Sato 1974). It was, therefore, impossible to estimate the content of Pg in soils except for the JR profiles and the remaining soils with Type P HA (Table 1). In the JR profile, the Pg content, which was high in the A1 and AB horizons (0.17 and 0.16 g kg⁻¹, respectively), markedly decreased, thereafter, with depth (Table 1).

UV–Vis spectral analysis

The visible spectra of the chloroform solution of CEF from the Andosol (SGG) and Cambisol (KI) profiles are shown in Fig. 1. Characteristic absorption maxima near 583, 565, 542, 454, 440 and 418 nm were detected in all horizons of the two profiles. These absorption maxima were also detected in all horizons of the remaining four profiles (SGF, SN, GS and JR), and they were almost identical to those of CEGF reported in Kobayashi et al. (2005). The relative absorption intensity at 454, 440 and 418 nm, however, differed among the soil horizons (Fig. 1). Therefore, the ratios of E _440CEGF to E _454CEGF and of E _418CEGF to E _454CEGF (E _λCEGF : absorbance at λ nm due to CEGF) from the SGG and KI profiles were calculated using the method described in Kobayashi et al. (2005), and the ratios are presented in Table 2. The ratios were different among the soil horizons from the SGG and the KI profiles and this difference was also observed in the remaining soil profiles.

Distribution of CEGF in soil profiles

The CEGF was detected in all soil samples and the content ranged from 0.02 to 1.16 (g kg⁻¹), as shown in Table 1.

Table 1 Chloroform-extractable green fraction (CEGF) and Pg contents, soil chemical properties and properties of humic acids

Soil sample (Horizon)	Depth (cm)	CEGF (g kg^−1)	Pg (g kg^−1)	O-C^† (g kg^−1)	pH^† (H2O)	Ex. Al (cmol kg^−1)	Fed^† (g kg^−1)	Ald^† (g kg^−1)	Feo^† (g kg^−1)	Alo^† (g kg^−1)	Humic acid
											A 600/C	log (A 400/A 600)	HA type^‡
Silic Andosols
Sugadaira-G (SGG)
A1	(0–10)	1.00	–	149	6.16	0.03	26.3	39.2	12.7	62.5	4.81	0.58	B±
A2	(10–21)	1.16	–	169	5.76	0.15	14.9	31.6	14.0	46.9	6.24	0.54	A±
A3	(21–35)	1.10	–	129	5.73	0.34	34.8	62.5	8.9	56.4	6.92	0.54	A±
A4	(35–62)	0.90	–	128	5.72	0.28	36.9	60.6	18.7	95.4	8.26	0.53	A±
Bw1	(62–80)	0.26	–	25	5.83	0.02	21.3	27.9	11.4	73.7	5.04	0.57	A±
Bw2	(80–100)	0.18	–	17	5.85	0.01	22.8	27.6	11.7	77.2	5.54	0.61	A±
Sugadaira-F (SGF)
A1	(0–7)	0.78	–	158	5.11	1.01	21.1	35.7	10.6	44.5	4.77	0.58	B±
A2	(7–17)	1.04	–	132	5.28	0.73	12.8	30.9	14.3	58.8	5.44	0.55	A±
A3	(17–31)	0.90	–	74	5.78	0.15	24.9	41.8	9.0	63.0	6.09	0.56	A±
A4	(31–47)	0.53	–	40	5.89	0.09	31.7	33.0	7.9	57.5	6.34	0.55	A±
Bw1	(47–67)	0.11	–	10	5.68	0.11	37.3	22.2	9.2	30.9	5.47	0.58	A±
Bw2	(67–86)	0.06	–	7	5.80	0.08	32.0	18.3	8.4	25.6	5.12	0.60	A±
Soni (SN)
A1	(0–20)	0.55	–	142	5.19	0.90	27.2	33.7	12.2	28.6	9.64	0.57	A±
A2	(20–25)	0.67	–	92	5.11	0.82	35.9	39.8	14.7	32.7	4.95	0.55	B±
2A3	(25–35)	0.80	–	69	5.20	0.27	31.1	44.2	16.9	43.6	10.23	0.52	A±
3A4	(35–58)	0.87	–	58	5.10	0.26	22.5	41.9	9.8	41.8	6.35	0.52	A±
3AB	(58–67)	0.46	–	28	5.00	0.31	18.6	29.1	9.1	35.7	5.58	0.53	A±
3Bw	(67–85)	0.13	–	14	5.11	0.54	23.1	21.2	6.1	18.0	5.13	0.57	A±
3BC	(85–110)	0.10	–	9	5.12	0.74	19.5	14.0	6.0	10.9	4.09	0.58	B±
Umbric Andosol
Gyosei (GS)
A1	(0–10)	0.51	–	177	4.18	1.89	23.8	14.8	7.3	11.4	5.81	0.53	A±
A2	(10–16)	0.45	–	124	4.58	2.10	32.0	18.0	8.7	13.6	6.29	0.53	A±
A3	(16–35)	0.61	–	65	4.90	1.32	28.4	23.3	17.5	18.5	5.58	0.52	A+
2A4	(35–41)	0.71	–	56	4.90	0.82	23.0	24.2	19.4	24.8	5.99	0.52	A+
2AB	(41–47)	0.30	–	26	4.81	1.13	43.6	28.6	11.7	19.1	6.39	0.54	A±
2Bw	(47–76)	0.05	Tr.	9	5.03	1.04	33.2	15.5	10.9	10.3	3.19	0.58	P+
Dystric Cambisols
Japanese soil Kuroiwa (KI)
A1	(0–3)	0.38	–	98	4.53	1.93	16.7	11.5	6.9	7.4	1.15	0.74	Rp±
A2	(3–9)	0.27	–	33	4.90	1.65	18.0	12.8	5.8	8.4	1.56	0.73	Rp+
AB	(9–19)	0.21	0.06	20	4.85	1.19	20.4	13.1	5.8	9.3	2.10	0.65	P++
Bw1	(19–27)	0.10	–	13	4.82	1.26	19.2	11.8	6.2	9.5	1.58	0.71	Rp++
Bw2	(27–37)	0.05	Tr.	8	4.98	1.36	16.5	10.1	4.7	7.8	1.85	0.64	P±
BC	(37–65)	0.02	Tr.	4	5.00	1.36	11.7	6.2	3.2	5.5	1.67	0.69	P±
Nepalese soil Jiri-F (JR)
A1	(0–4)	0.38	0.17	57	5.90	0.86	18.4	11.6	12.6	11.8	2.72	0.60	P±
AB	(4–12)	0.52	0.16	43	5.65	0.84	22.7	14.2	11.8	12.5	2.98	0.56	P±
Bw1	(12–22)	0.32	0.04	27	5.70	0.77	21.9	12.7	11.7	11.9	2.52	0.56	P±
Bw2	(22–41)	0.24	0.04	15	5.70	0.71	22.5	10.7	9.4	9.6	2.26	0.55	P±
BC	(41–60)	0.05	0.01	11	5.89	0.70	23.0	8.6	7.5	6.8	2.52	0.60	P±
O-C, content of organic carbon in soil samples; Fed and Ald, amounts of dithionite-citrate-bicarbonate-extractable Fe and Al in the soil samples, respectively; Feo and Alo, amounts of oxalate-oxalic acid-extractable Fe and Al in the soil samples, respectively; Tr., trace amount (< 0.01 g kg^−1); –, not determined. ^†Data were cited from the report of Suzuki et al. (1999b).
^‡Subdivision of humic acids (HAs) was used according to the method described by Kumada et al. (1967).

Table 2 Values of E440CEGF/E454CEGF and E418CEGF/E454CEGF ratios calculated from the visible spectra of the chloroform-extractable fraction (CEF) from the Sugadaira-G (SGG) and Kuroiwa (KI) profiles

Soil sample (Horizon)	E 440CEGF/E 454CEGF	E 418CEGF/E 454CEGF
SGG
A1	0.85	0.56
A2	0.86	0.54
A3	1.17	0.74
A4	1.29	0.82
Bw1	1.62	1.04
Bw2	2.74	1.69
KI
A1	0.76	0.56
A2	0.80	0.48
AB	0.89	0.56
Bw1	1.06	0.65
Bw2	1.22	0.75
BC	1.10	0.69
CEGF^†	1.24	0.90
^†Data from Kobayashi et al. (2005). CEGF, chloroform-extractable green fraction.

The highest content of CEGF was recorded in the A2 horizon of SGG. The average content of CEGF in the surface A horizon of the four Andosol profiles (SGG, SGF, SN and GS) was 1.00, 0.78, 0.57 and 0.54 g kg⁻¹, respectively. In the two Cambisol profiles (KI and JR), the content was 0.30 and 0.38 g kg⁻¹, respectively.

Figure 2 shows the vertical distribution of the contents of CEGF and organic carbon (O-C) in the soil profiles. The content of O-C tended to decrease with depth in all soil profiles except for the A2 horizon of the SGG profile.

In the SGG, SGF and JR profiles (Fig. 2a,b,f, respectively), the content of CEGF was highest in the subsurface horizon (1.16, 1.04 and 0.52 g kg⁻¹, respectively) and it decreased with depth in the same way as the O-C content. The distribution of the CEGF and Pg contents in the JR soil profile differed from each other (Fig. 2f, Table 1, respectively). In the KI profile, the distribution of the content of CEGF showed a similar pattern to that of O-C (Fig. 2e). In contrast, the distribution of the CEGF content in the SN and GS profiles, which contained a buried humus horizon, tended to increase up to the buried A horizon (the 3A4 horizon of SN and the 2A4 horizon of GS) and then decreased with depth. In these two profiles, the distribution differed considerably from that of O-C (Fig. 2c,d, respectively).

Relationship between CEGF content and soil chemical properties

The CEGF content exhibited a significant positive correlation with the O-C content (P < 0.001, n = 36) in all soil samples. However, the distribution of the CEGF content in the SN and GS profiles, which contained a buried humus horizon, differed considerably from that of O-C. Correlation analyses between the contents of CEGF and O-C in the soil profiles was, therefore, carried out separately in soil profiles that contained a buried humus horizon and in those without one. As a result, the content of CEGF did not show any significant relationship with that of O-C (n = 13) in the former two profiles (SN and GS). In contrast, there was a significant relationship between the content of CEGF and O-C (r = 0.922, P < 0.001, n = 23) in the latter four soil profiles (SGG, SGF, KI and JR). Figure 3 shows the relationship between the content of CEGF and dithionite-citrate-bicarbonate-extractable Fe and Al (Fed and Ald, respectively), and CEGF and oxalate-oxalic acid-extractable Fe and Al (Feo and Alo, respectively) in all soil samples. A significant positive correlation (P < 0.001, n = 36) was found between the content of CEGF and Ald, Feo and Alo (Fig. 3b,c,d, respectively) in all soil samples. As the content of O-C showed a significant correlation with Ald (P < 0.01), Feo (P < 0.05) and Alo (P < 0.05) in all soil samples, a partial correlation analysis between the CEGF content and Ald, Feo and Alo was carried out to exclude the influence of O-C. As a result, the CEGF content showed a significant correlation with Ald (r = 0.693, P < 0.001), Feo (r = 0.504, P < 0.01) and Alo (r = 0.535, P < 0.01). In contrast, there was no significant relationship between the content of CEGF and Fed (Fig. 3a), or with the pH (H₂O) value or the concentration of Ex. Al in all soil samples (n = 36). In addition, we determined whether a relationship between the CEGF content and the chemical properties of the soil with the highest content of CEGF in the profile could be found. As a consequence, the content of CEGF showed a significant correlation with Alo at P < 0.01 (r = 0.944, n = 6: the content of Alo did not show a significant relationship with O-C), as shown in Fig. 3d, but not with the remaining parameters.

DISCUSSION

Distribution of CEGF in soil profiles

The CEGF content in the surface A horizon of the Andosol tended to be higher than the content in the Cambisol. In general, Andosol is a humus-rich soil with a higher organic carbon content than other soils. Therefore, CEGF, which is assumed to be of microbial origin (Kobayashi et al. 2005), may be actively produced in a humus-rich soil like Andosol. Sato (1976a) pointed out that humus-rich soils might promote the activity of Pg-producing fungi. Although CEGF is only a closely related substance to Pg, this may support our concept.

Based on the results of the distribution of CEGF content, it appears that the vertical distribution of CEGF

Figure 2  Vertical distribution of chloroform-extractable green fraction (CEGF) (•) and organic carbon (◊) contents in the soil profiles. (a) Sugadaira-G (SGG), (b) Sugadaira-F (SGF), (c) Soni (SN), (d) Gyosei (GS), (e) Kuroiwa (KI) and (f) Jiri-F (JR).

content differed considerably depending on whether the soil profile contained a buried humus horizon (SN and GS) or did not contain one (SGG, SGF, KI and JR). In the latter profiles, the soil horizon with the highest content of CEGF did not necessarily have the highest content of O-C, although CEGF content showed a significant relationship with O-C content. A clear reason for these results cannot be determined at present. However, sclerotia of Cenococcum graniforme (currently C. geophilum), which is a resting structure of ectomycorrhizal fungi and is considered to be one of the sources of CEGF (Kobayashi et al. 2005), were distributed in various soils and their contents differed among the soil horizons (Dahlberg et al. 1997; Kumada and Hurst 1967; Watanabe et al. 2002). Watanabe et al. (2002) reported that in a certain soil profile, the density of sclerotia in the buried humus horizon was higher than that in the surface A horizon in a Japanese volcanic ash soil. Although the distribution of sclerotia in the soil profiles used in the present study were not examined, the distribution of the CEGF content may be associated with the distribution of CEGF producers such as sclerotia, rather than with the O-C content in the soil profiles. Statistical analysis showed that the content of CEGF displayed a significant positive correlation with Ald, Feo and, in particular, with Alo. Nakabayashi et al. (1982) and Nakabayashi and Wada (1991), who investigated the vertical distribution of Pg content and the factors controlling Pg content in one brown forest soil profile, assumed that part of the Pg was combined with Fe and then transferred to the underlying

Figure 3  Relationship between the chloroform-extractable green fraction (CEGF) content and the contents of (a) Fed, (b) Ald, (c) Feo and (d) Alo. Parentheses indicate partial correlation coefficient excluding the influence of organic carbon (O-C). (–•–) all soil samples (n = 36); (–○–) soil samples with the highest content of CEGF in each soil profile (○, SGG [A2 horizon]; ▵, SGF [A2 horizon]; ◊, SN [3A4 horizon]; ▿, GS [2A4 horizon]; □, KI [A1 horizon]; , JR [AB horizon]). **P < 0.01; ***P < 0.001.

horizon by percolating water, and that the Pg–Fe complex was stabilized against bacterial decomposition. Tsutsuki and Kuwatsuka (1989) suggested that Pg was preserved in volcanic ash soils by the formation of a complex with ferric compounds. Although CEGF is only a closely related fraction to Pg, these minerals may be associated with stability and/or movement of CEGF in the soil profile. Further studies examining the relationship between the content of CEGF and the content of these minerals are necessary for understanding the stability and behavior of CEGF in soil.

In a previous paper (Kobayashi et al. 2005), CEGF was considered to consist of several fractions. The difference in the E _440CEGF/E _454CEGF and E _418CEGF/E _454CEGF ratios among the soil samples (Table 2), therefore, may reflect differences in the properties of CEGF, such as the composition of the fractions constituting CEGF. Furthermore, the difference in the ratios might reflect differences in the soil environment. These ratios, however, were not strictly correct because of the interference of a fraction that differed from CEGF (Fig. 1). It will be necessary to develop a separation and purification method for CEGF to investigate the distribution of CEGF content and its properties in detail.

The distribution of the CEGF content was different from that of Pg in the JR profiles. This indicates that CEGF is not identical to Pg. It will be necessary to examine the distribution of the CEGF and Pg contents in soils with Type P HA to clarify their precise relationship.

Occurrence of CEGF and Pg in buried humus horizons

The CEGF content was highest in the buried A horizons of the SN and GS profiles (Fig. 2c,d, respectively). Pg was detected in these buried humus horizons, although the content could not be estimated (Table 1). It is known that Pg, which is considered to be of microbial origin (Kumada and Hurst 1967), is distributed not only in various surface soils but also in buried humus horizon approximately 100,000 years bp (Watanabe and Kobayashi 2001). Watanabe and Kobayashi (2001) reported that the Pg absorption strength (Watanabe 1992; Pg absorption strength = (A ₆₁₅-A ₇₀₀)/(A ₆₀₀-A ₇₀₀), A _λ: absorbance at λ nm of HA in alkaline solution; Pg showed a characteristic absorption maximum at 615 nm in alkaline solution) might be used as an indicator reflecting the soil environment in the past. However, the use of the Pg absorption strength is not appropriate in the case of humus-rich soils because a brown fraction that differs from Pg strongly interferes with Pg absorption (Watanabe and Kobayashi 2001). In contrast, the method for estimating the content of CEGF in soil can be applied to various soils including humus-rich soils. Although more detailed studies examining the relationship between the contents of CEGF and Pg should be carried out, it might be possible to predict the soil environment in the past more precisely using the CEGF content and Pg absorption strength in combination.

Conclusion

In the present study, the vertical distribution of a soil green pigment in an Andosol profile was reported for the first time quantitatively by determination of the CEGF content. CEGF occurred not only in the A horizon, but also in the buried A and remaining horizons (B and BC horizons). The CEGF content in the A horizon of the Andosol was higher than that in the Cambisol. The content of CEGF was highest in the surface or subsurface horizon in the soil profiles that did not contain a buried humus horizon. In contrast, the maximum content of CEGF was observed in the buried humus horizon when the soil profiles contained this horizon. Some minerals, such as Ald, Feo and, in particular, Alo, are considered to be associated with the distribution of CEGF in the soil profile. In addition, it was revealed that the properties of CEGF differed among the soil samples.

Further studies examining the quantitative distribution of CEGF in soils should be carried out to emphasize the importance of the soil green pigments, including Pg.

ACKNOWLEDGMENT

The authors thank Dr Takeshi Suzuki, Faculty of Agriculture, Kobe University, for classification of humic acids.