Genesis and mineralogy of soils formed on uplifted coral reef in West Timor, Indonesia

Geoderma 154 (2010) 544–553

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r.com/ locate /geoderma

Genesis and mineralogy of soils formed on uplifted coral reef inWest Timor, Indonesia

Welhelmus Mella, Ahmet R. Mermut ⁎University of Saskatchewan, Department of Soil Science, Saskatoon SK., Canada S7A 5A8

⁎ Corresponding author.E-mail address: [email protected] (A.R. Mermut).

0016-7061/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.geoderma.2009.10.021

a b s t r a c t

Article history:Received 15 August 2008Received in revised form 10 October 2009Accepted 31 October 2009Available online 2 December 2009

Keywords:Coral reefsWest TimorAlfisolsMollisolsSoil mineralogyLimestone residuum

Many scientists believe that soils forming on, or in association with, limestone are derived from consolidatedlimestone rocks or the limestone residuum. The objective of this study was to determine the genesis ofreddish Alfisols and black Mollisols formed on uplifted coral reefs in West Timor, Indonesia. A total of 24Alfisols and 23 Mollisols were studied and sampled. Rock samples were also collected from the surface, fromwithin the profile, and from the bedrock of representative profiles. Dissolution of the coral limestonesyielded 1.7% insoluble residue. Oxides of Si, K, Ca, and Mg were present at higher concentrations in theinsoluble rock residues than in the bulk soils and clay fractions. Concentrations of Al and Fe oxides werehigher in the soils than in the residues. Considering these, possibility exists that exogenic materials(atmospheric dusts or volcanic ash) may have also contributed to the formation of these soils studied.Alfisols were dominated by kaolinite followed, in a decreasing order, by vermiculite, gibbsite, and quartz,whereas Mollisols were predominantly smectite and mica followed by kaolinite and quartz. The mineralogyof the insoluble residues was similar to that of the clay of Mollisols. The presence of kaolinite in residues andin the clay of both soils suggests that kaolinite may have been derived from the residue. Therefore, weconcluded that weathering of the underlying limestone is the main source contributed to the development ofboth the Alfisols and the Mollisols. Based on the principles of mineral transformation in the soil, particularlyhigher kaolinite content, we conclude that Alfisols occurring in more stable surfaces and are stronglyweathered than Mollisols.

ll rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Coral reefs consist mostly of sedimentary limestone with highcalcium carbonate content, often N95% by weight, with minimalinsoluble residue consisting mainly of silicate minerals (Pettijohn,1957; Macleod, 1980). The paucity of silicate minerals, the silicatemineralogy, and the sharp contact between the soil and limestone,raise questions as to whether limestone is the sole parent material ofassociated soils (Ruhe et al., 1961; Ives et al., 1976; Olson et al., 1980;Muhs et al., 1990; Nihlén and Lund, 1995; Durn et al., 1999; Muhs,2001). In most cases, eolian, fluvial, alluvial, or colluvial depositioncontributed to the formation of the soils, as well as limestoneweathering (Merino and Banerjee, 2008). Many workers believe thatsoils formed on, or in association with, limestone are derived fromconsolidated limestone rocks or the limestone residuum (Stace, 1956;Barshad et al., 1956; Glazovskaya and Parfenova, 1974; Duchaufour,1982; Nagatsuka et al., 1983; Esteban and Klappa, 1983; Rabenhorstand Wilding, 1986; Moresi and Mongelli, 1988; Boero and Schwert-mann, 1989; and Yassoglou, et al., 1997).

The objective of this study was to explain the genetic pathways ofreddish-colored Alfisols and black Mollisols that are formed on orassociated with coral reef limestone in West Timor, Indonesia.

2. Materials and methods

2.1. Study area and sampling protocol

This studywas undertaken in the Soe region, in south-centralWestTimor, Indonesia (Fig. 1). The study area was located on coral reeflimestone uplifted in the Late Quaternary (Van Bemmelen, 1949;Jouannic et al., 1988) at 9°55′S and 124°39′E. The present surface is800–900 m above sea level, with a mean annual temperature of 21 °Cand a mean annual rainfall of 1800 mm. The rainy season normallystarts in October and ends in May, with a peak that occurs inDecember and January (Badan Meteorologi dan Geofisika Kupang,2000).

Soils formed on the limestone in West Timor are classified asAlfisols and Mollisols, which according to Duchaufour (1982) can beconsidered to be Terra Rossa or Red Fersiallitic and Calcimagnesiansoils, respectively. Alfisols are red and occur on level to gently sloping(b5%) landscapes; whereas the Mollisols are black and are situatedon somewhat steeper (5–10%) slopes. Alfisols are usually under

Fig. 1. Location of the study area, West Timor, Indonesia. The inset is the Indonesian archipelago to north and Australia in the southeast corner (modified from Barlow et al., 1990).

545W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

eucalyptus savanna stands (Eucalystus alba, Lantana camara, andgrasses). In the last decade, the savanna has been invaded by Siamweed (Cromolaena odorata), which replaces lantana shrubs. On theother hand, the Mollisols are generally covered by mixed-deciduousforest stands that consist of leguminous and non-leguminous treesand shrubs.

Twenty four profiles from Alfisols (12 profiles each from cultivatedand uncultivated lands) and 23 profiles from Mollisols (13 fromcultivated area and 10 from native lands) were excavated anddescribed in the field. Similar numbers of profiles from each soilorder were selected for a realistic comparison. The Alfisols wereexcavated down to 120 cm, usually the depth to bedrock. Thesampling depth for Mollisols depends on the depth at which bedrockwas encountered, ranging from 23 cm to 70 cm. Limestone rocks werealso collected from the surface, throughout the profiles, and from thebedrock of representative Alfisols (4 sites) and Mollisols (6 sites).

All rock samples were crushed into small fragments, washedseveral times with distilled water to remove any soil, and then soakedfor 30 min in 0.05 M HCl to remove any vestiges of soil. The rockpieces were then washed once again with distilled water, oven driedat 550 °C for 24 h and pulverized to pass a 1-mm sieve. They werethen dissolved in 1 M NaOAc (sodium acetate) buffered at pH 4 withacetic acid. The insoluble residue was collected, dried and weighed tocalculate the amount of insoluble residue, and saved for chemical andmineralogical analyses. Rock powder also was dissolved in 1 N HCl tocalculate the Mg/Ca ratio of the limestones. The amounts of Ca andMgin the solution were determined by atomic absorption spectroscopy(AAS).

2.2. Physical and chemical analyses

Soil pH was measured with a portable pH meter on 1:1 soil/watermixture according to McLean (1982). Two core samples were takenfrom each horizon in the field by driving (horizontally or vertically) aring sampler of 5.5 cm in diameter and 4 cm in length. Core sampleswere oven dried at 105–110 °C andweighed. Bulk density values were

calculated using the dry soil mass and the known core volume (Blakeand Hartge, 1982).

Particle size analysis on b2 mm soil followed the pipette method(Sheldrick and Wang, 1993) after removing organic matter usingcommercial bleach (sodium hypochlorite, NaOCl) as outlined byAnderson (1963). Iron oxides were not removed for this analysis (Geeand Bauder, 1986), exchangeable cations were extracted using 0.5 NBaCl2–0.05 N triethanolamine adjusted to pH 8 and cation exchangecapacity (CEC) by 1 N NH4OAc as extracting solutions according toHendershot and Lalonde (1993). The leachate was collected todetermine the concentrations of major cations (Ca2+, Mg2+, K+ andNa+) using AAS. The soil was subsequently re-suspended, filtered, andwashed with 200 mL NH4OAc to replace the adsorbed Ba2+. The Ba2+

in the NH4OAc extract wasmeasuredwith AAS to represent the CEC ofthe soil samples.

Iron (Fe), aluminum (Al), and manganese (Mn) were extractedwith dithionite-citrate, acid ammonium oxalate, and sodium pyro-phosphate extraction methods as outlined by Ross and Wang (1993).All filtrates were collected separately and analyzed using AAS tomeasure dithionite (d), oxalate (o) and pyrophosphate (p) extract-able Fe, Al, and Mn.

Clay-sized particles were obtained from 10 selected soil profiles(four Alfisols and six Mollisols). The clay fractionation was performedafter removing the organic matter only, without the removal of ironoxides and carbonates. Since H2O2 can also oxidize calcium carbonate,soil organicmatter was removed by using commercial bleach (sodiumhypochlorite, NaOCl) as outlined by Anderson (1963). Clay fraction-ation was based on the procedure outlined by Jackson (1969). Specificsurface area of the clay-size particles was determined using the EGMEmethod as outlined by Carter et al. (1986).

Total elemental analysis was performed on whole soil samples,clay fractions, and insoluble limestone residue. Closed vessel hydro-fluoric acid digestion was adapted for total elemental analysis (Limand Jackson, 1982). The resulting solution was analyzed using AAS todetermine total Si, Al, Fe, Ca, Na, K, Mg, Ti, and Mn. The concentrationof P was determined using a Technicon Autoanalyzer II. All elements

Table 1Insoluble residue content of limestones from different locations in Alfisols andMollisolsfrom West Timor, Indonesia. Expressed in percent of oven dry weight of insolubleresidue.

Soil Surface rock In-profile rock Bedrock

Insoluble residue (%)

AlfisolsMean 1.4 a† (n=4)‡ 1.3 a (n=4) 1.5 a (n=4)SD 0.4 0.6 0.6

MollisolsMean 1.5 a (n=6) 2.0 a (n=6) 1.4 a (n=6)SD 1.0 0.8 0.7

†Mean values followed by the same letters within columns are not significantly different atα=0.05.‡n is the number of samples.

546 W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

were expressed as oxide. Note that total Ti, Mn, and P contents werenot determined for clays.

2.3. Soil mineralogical and morphological analyses

X-ray diffraction (XRD) was used to characterize the insolublelimestone residue and clay-size particles. The following treatmentswere used to prepare subsamples for XRD analyses: Mg saturation atroom temperature (Mg), Mg-glycerol saturation (Mg-Gly), K satura-tion at room temperature (K-RT), and K saturation and heating to550 °C (K-550). The samples, as slurry, were mounted on glass slidesevenly as described by Jackson (1969) and analyzed with a RigakuRotaflex X-ray diffractometer with Fe Kα radiation and a graphitecrystal monochromator.

Fourier-transform infrared (FTIR) spectrometry was carried out oninsoluble limestone residues and clay-sized particles from the soils.The spectra were generated using a Bio-Rad FTS-40 spectrophotom-eter (Digilab Division, Cambridge, MA). Prior to exposure, sampleswere prepared using the alkali halide pressed-pellet method (Whiteand Roth, 1986). Spectra were collected at 256-scan number, 4 cm−1

wave number resolutions, and at the mid-range infrared region(4000–400 cm−1 or 2.5 to 25 μm).

Scanning electron microscopy (SEM) studies were carried out oncoral reef samples. Pieces of limestone were mounted on aluminumstubs with double-sided tape, gold-coated, and examined with aPhilips 505 SEM and by JSM-840A Scanning Microscope (JEOL). Thepresence of microorganisms in the limestone was first noted duringthe dissolution experiment of coral limestones. A fragment of thesame rock was sterilized (outside part of the rock) with 1.05% NaOCl(commercial bleach) for 10 min, ground and cultured on 0.3%Trypticase-Soy Agar (TSA). The growing organism was then dried

Table 2Elemental analysis (oxide forms) of insoluble residue of coral reef limestone collected fromWest Timor, Indonesia.

Soil Location† SiO2 Al2O3 Fe2O3 K2O

%

Alfisols Bed rock (4) 44.6 22.8 8.2 3.0In-profile (4) 44.5 23.4 7.2 3.0Surface (4) 45.1 23.0 7.0 3.6Mean 44.7 a 23.1 a 7.5 a 3.2 aSD 3.8 1.7 1.3 0.6

Mollisols Bed rock (6) 51.3 20.6 7.8 2.7In-profile (6) 48.5 20.1 7.4 2.8Surface (6) 40.3 18.9 5.2 2.7Mean 46.7 a 19.9 b 6.8 a 2.7 bSD 7.0 2.2 1.7 0.5

Mean values in the same column followed by different letters are significantly different at† In brackets following locations are the sample sizes. ‡molar ratio=(% oxide/molecular w

and mounted with a double-sided tape on an aluminum stub andexposed to SEM.

2.4. Statistical analyses

All the statistical analyses were performed using SPSS program(ver. 11.5 for Windows) to perform analysis of variance (ANOVA) ofsingle treatments and their interactions, as well as to compare meanvalues of different treatments. Due to the imbalances in soil depth, theANOVA was performed by using general linear model procedure.

3. Results and discussions

3.1. Nature of coral reef limestone

Due to the sub-aerial exposure, the coral reef limestone dissolvedto form soil and karst facies (Esteban and Klappa, 1983; Trudgill,1985). It is important to note that the soil lies directly on top of thekarst with a sharp contact, which is in agreement with theobservations of Olson et al. (1980). This karstic feature plays animportant role in the subsequent weathering that produces theoverlying soil cover (Sweeting, 1972; Trudgill, 1985). At a microscopiclevel, the limestone consists of macro- and micro-pores that mayfacilitate the movement of water and microorganisms (endolithicorganisms) into the limestone and promote weathering (Golubić andSchneider, 1979; Ehrlich, 1990).

3.2. Insoluble residue

The coral contains 1.3 to 2.0% insoluble residue (Table 1). Previousstudies found that the insoluble residue concentration in limestones isb5% (Pettijohn, 1957) and in pure calcite minerals b0.2% (Deer et al.,1967). Thus, the underlying coral reef in the study area is a high-gradelimestone with minor amounts of silicate minerals.

The limestone rocks had similar contents of insoluble residueregardless of the source. On average, the residue content was 1.7%.This means that about 29 and 70 m of limestone would need to bedissolved by weathering to form the 50-cm and 120-cm deep soilprofiles of the Mollisols and Alfisols, respectively, assuming noaddition of silicate minerals from other sources and no significantremoval of residue left behind by the dissolution of calcium carbonaterock.

Jouannic et al. (1988) suggested that, based on the age of fossilsfound in the Central Graben area of West Timor (around the studysite), the uplift of coral reefs began 700,000 years before present.Assuming a limestone dissolution rate of 10 to 72 mm per 1000 years,as suggested for karst of tropical regions by Sweeting (1972), thethickness of limestone that may have been dissolved during the last700,000 years should be between 7 and 49 m. Although this estimate

surface, in-profile, and bedrock of Alfisols (Terra Rossa) and Mollisols (Rendzina) from

TiO2 MgO CaO MnO Na2O SiO2 ‡/Al2O3

1.9 2.2 5.2 2.5 0.3 3.452.1 2.4 2.9 1.9 0.2 3.351.8 2.1 2.4 1.5 0.4 3.451.9 a 2.3 a 3.1 a 2.0 a 0.3 a 3.410.3 0.3 3.9 1.8 0.11.5 2.8 2.9 0.4 0.3 4.391.3 2.4 3.2 0.2 0.2 4.261.3 1.9 1.9 0.4 0.2 3.761.4 b 2.4 a 2.7 a 0.3 b 0.3 a 4.120.2 0.5 1.3 0.3 0.1

α=0.05.eight).

Fig. 2. Scanning electron micrographs of coral reef limestones. (A) needle-shape calcite; (B) calcite. Bar is 10 μm.

547W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

seemswithin the range required to produce the soils, it is possible thatthe soils may have received materials from surrounding areas(allochtonous) especially in the case of the Alfisols (Terra Rossa) asalso suggested by Yassoglou et al. (1997), or from aerial deposition ofvolcanic and continental dusts (Ruhe et al., 1961; Ives et al., 1976;Macleod; 1980; Muhs et al., 1990; Nihlén and Lund, 1995; Durn et al.,1999; and Muhs, 2001).

3.3. Elemental analysis

The insoluble residue of limestone extracted from Alfisols containsmore Al2O3, K2O, TiO2, and MnO than limestones under Mollisols(Table 2). The percentages of other oxides (SiO2, Fe2O3, MgO, CaO, andNa2O), however, were not significantly different. This may indicatethat the limestones from the representative regions of the two soilorders have slightly different silicate impurities. Compared to thechemical composition of limestones mentioned by Pettijohn (1957),the SiO2 content of the Timor limestone is lower but the other majorelements are higher.

3.4. Mineralogy and morphology of coral reef limestone and residues

Acid dissolution of the limestones from all locations in both Alfisolsand Mollisols reveals that the average Mg/Ca ratio in the rocks is 0.01(data not presented), indicating that neither the kind of soil nor thelocation influence the Mg/Ca ratios of the limestone. The low Mg/Caratio also indicates that the coral reef limestones in the study area are

Fig. 3. Scanning electron micrograph of void-filling needle-shaped minerals (aragoniteneedles?) in coral rock samples from the study area (Alfisol region). Bar is 10 μm.

dominated by CaCO3 (calcite and/or aragonite) (Doner and Lynn,1989). This is consistent with the small amount of acid insolubleresidue. The presence of calcite is revealed by X-ray diffractograms(XRD) and the shape observed with the SEM (Figs. 2 and 3).

Although there are needle-shaped minerals forming in the voids(Fig. 3), aragonite is not necessarily present in the rocks. Aragonite isunstable in sub-aerial conditions and is rapidly transformed intocalcite (Pettijohn, 1957). However, aragonite can be formed assubmarine precipitates and preserved in micro voids of upliftedPleistocene coral reefs (Pingitore, 1971). Despite uncertainty as towhether these needle-shaped minerals are aragonite, their presencein the voids reduces the internal pore space in the limestones and alsofunctions as a cementing agent in the diagenetic process.

The XRD of limestone residue samples from the bedrock ofrepresentative Alfisols and Mollisols shows that they contain smectite(d=14.0 Å for Mg and expanding to 17–18 Å for Mg-glycerol treatedclays), mica (d=10, 5, and 3.3 Å), kaolinite (d=7.0 Å), and somequartz (d=3.3, 4.2 and 4.3 Å) (data not presented) (Berry, 1974;Whitting and Allardice, 1986). This means that the limestone rocksfrom different locations for both soil orders have relatively uniformclay mineralogy. Tarzi and Paeth (1975) reported similar results in astudy of limestones from Lebanon.

Infrared spectra of the acid insoluble residues from limestone rocksof the Alfisol and Mollisol regions showed the same clay mineralogy,regardless of locations, consistent with the findings of the XRDstudies. Smectite (represented by montmorillonite at 3621, 1630, and1029 cm−1), kaolin minerals (represented by kaolinite at 913, 796,

Fig. 4. Scanning electron micrograph of fungal hyphae penetrating soft limestonecollected from Mollisol regions. Bar length is 0.1 μm.

Fig. 5. Scanning electron micrographs of fungal hyphae and the resulting tunnels (arrows) in a soft limestone collected from the Mollisol region. Bar length is 0.1 mm.

548 W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

754, 696, 533, 473 and 431 cm−1), and possibly brucite (3696 cm−1),quartz (798 cm−1), and mica (477 cm−1) were identified (data notpresented) (Kodama, 1985). A wave number of 3621 cm−1 may beshared between montmorillonite and kaolinite and “fire clay” (kaolinclay for making bricks) (Van der Marel and Beutelspacher, 1976).

Durn et al. (1999) found neither kaolinite nor smectite in theinsoluble residue of Upper Jurassic and Cretaceous limestones; themostprevailing clay minerals were illites, mica, and quartz. In dry areas,Barshad et al. (1956) found montmorillonite, kaolinite, and quartz inUpper Cenomenian limestones. Montmorillonite and mica were alsoreported in the insoluble residue of Middle Ordovician limestones (Rayand Gault, 1961).

3.5. The role of microorganisms in limestone weathering

The presence of fungal hyphae in the limestone was first notedduring the dissolution experiment of coral limestones. Evidence, inthe form of bore holes and tunnels, of the role of microorganisms inthe weathering of limestones can be observed through the scanningelectron micrographs (Figs. 4 and 5).

We interpret that there are twomodes of action ofmicroorganismsin the weathering process. First, the microorganisms actively boretheir way into the limestone; through the dissolution of CaCO3 via theproduction of CO2 and then carbonic acids, and organic acids releasedby organisms (Fig. 4). This process results in the formation of tunnels

Fig.6. Scanning electron micrograph of cultured fungus derived from inside of a softlimestone sampled from the Mollisol region. Bar length is 20 μm.

with specific patterns, which are similar to what was pointed out byGolubić and Schneider (1979). These holes and tunnels are round andsmooth, and are almost the same in diameter as the organism.

The second mode is by passively penetrating the interior of thelimestone throughpre-existingfissures and cavities thatmay have beengenerated by submarine abrasion (Golubić and Schneider, 1979),karstification processes (Esteban and Klappa, 1983; Trudgill, 1985), orby the presence of fossils. This mode is likely for soft limestone as inthe case of the study area. Judging from the size of the hyphae (25 to30 μmin diameter) in these examples, themicroorganism is interpretedto be a fungus.When themicroorganismwas cultured in a Petri dish, thepresence of septa confirmed its identification as a fungus (Fig. 6). Notethat the diameter of the hyphae is about 30 μm, which is about thesame size as shown in Fig. 4.

3.6. Soils

The clay content of Alfisols increases with depth; whereas inMollisols, it tends to remain relatively constant throughout theupper part of the profile, and then decreases sharply below the Bhorizon. The clay content of Mollisols is somewhat higher than inAlfisols and it increases from the surface soil to subsoil. However, theincrease in clay content in the Mollisols is not as striking as in theAlfisols. A Bt (argillic horizon) occurs at about the 15-cm depth in

Fig. 7. Interaction effect of depth category and soil order on clay content. A=Ap or Ahhorizons; Bt=argillichorizon;Ck=calcic horizons. Soil depth represents themidpoints ofdepth intervals of each horizon. Averages of 12 cultivated and 12uncultivatedAlfisols, and13 cultivated and 10 uncultivated Mollisols.

Table 3Effect of the soil type, management practice, and soil depth on bulk density in the soils studied.

Cultivated Alfisols Cultivated Mollisols Uncultivated Alfisols Uncultivated Mollisols

Depth Bulk density Depth Bulk density Depth Bulk density Depth Bulk density

(cm) (mg m−3) (cm) (mg m−3) (cm) (mg m−3) (cm) (mg m−3)

0–14 1.50 (12) 0–8 1.17 (13) 0–8 1.38 (12) 0–10 1.17 (10)14–26 1.62 (12) 8–28 1.33 (12) 8–20 1.50 (12) 10–22 1.37 (10)26–120 1.75 (12) 28–40 1.48 (9) 20–40 1.62 (12) 22–42 1.44 (10)

40–56 1.55 (4) 40–120 1.79 (10) 42–55 1.55 (6)56–65 1.80 (2) 55–65 1.50 (1)

Numbers in brackets following the bulk density values are the sample sizes.

549W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

Alfisols, down to the bottom of the profile in cultivated area, andbelow 35-cm depth under native conditions. According to SoilSurvey Staff (1999) if a pedon has more than 40% clay in the surfacehorizon, that the subsurface horizon must have at least an 8%absolute increase in clay content within a vertical distance of 30 cmto qualify as an argillic.

The clay contents are shown in Fig. 7 and bulk density in Table 3. Eventhough the textural analysis of the Alfisols and Mollisols indicates a claytexture throughout the profiles, the ANOVA showed that the clay contentvalues are significantly influenced by the interaction of soil order andmanagement practice aswell as by the interaction of soil order and depth.The bulk density values are affected only by single factors of soil order,management practice, and depth.

Total extractable iron (Fed) is greater in the Alfisols than theMollisols; whereas Feo and Fep do not differ consistently between thetwo soil orders (Table 4). In the Alfisols, Fed–FeoNFeo–Fep, whichaccording to Ross and Wang (1993) indicates that these soils contain

Table 4Extractable Fe, Al, and Mn of representative profiles of Alfisols and Mollisols.

Horizon Depth Claycontent

Dithionite

Fe Al Mn

%

AlfisolsMnelalete 1 (cultivated)

Ap 0–13 62 6.12 0.59 0.42AB 13–25 73 6.17 0.62 0.36Bt 25–120 91 6.47 0.73 0.29

Nenobila 2 (uncultivated)Ah 0–6 87 5.52 0.50 0.19AB 6–22 88 5.89 0.60 0.20Bt 22–41 90 6.35 0.64 0.29Bt 41–10 91 6.68 0.62 0.12

MollisolsNefobaun 1 (cultivated)

Ap 0–10 73 0.80 0.28 0.16B 10–47 90 0.90 0.29 0.18B 47–60 72 0.69 0.18 0.08

Ajaomaknao 2 (uncultivated)Ah 0–11 88 0.91 0.22 0.18B 11–24 73 1.15 0.25 0.19B 24–39 85 1.22 0.24 0.20B 39–63 79 1.09 0.24 0.22

Oelnunuh 1 (cultivated)Ap 0–8 83 2.15 0.40 0.40B 8–20 86 2.10 0.42 0.37B 20–31 88 2.00 0.46 0.29B 31–57 83 2.12 0.46 0.23Ck 57–67 77 0.78 0.20 0.09

Buat 1 (uncultivated)Ah 0–12 81 1.30 0.28 0.22AB 12–29 96 2.03 0.45 0.19B 29–44 85 1.60 0.46 0.26B 44–59 81 1.90 0.46 0.24B2 59–70 87 1.91 0.54 0.13Ck 70–80 43 0.56 0.16 0.04

primarily crystalline Fe-oxyhydroxides. The Mollisols, with Fed–FeobFeo–Fep, contain primarily poorly crystalline Fe-oxyhydroxides.Higher Fep contents in the Mollisols, especially in the topsoil,presumably reflects the higher organic matter content of the Mollisolsrelative to the Alfisols, in that pyrophosphate preferentially extractsorganically complexed Fe and Al.

Extractable Al values are greater in the Alfisols than in theMollisols. However, the crystallinity of Al in both soil orders cannot beconcluded because most of the Ald values are lower than Alo,especially in the Mollisols (Table 4). This is understandable, becausealthough Ald may be derived from Al oxyhydroxide (Cho andMermut,1992), most of the Ald originates from substituted Al in crystalline andpoorly crystalline Fe oxides (Darke and Walbridge, 1994; Dahlgren,1994). Darke andWalbridge (1994) stated that this procedure cannotbe used to predict the crystallinity of Al oxides, due to the smallamount of Al that can be removed by the dithionite extractionprocedure.

Oxalate Pyrophosphate Fed–Feo Feo–Fep

Fe Al Mn Fe Al Mn

1.78 0.57 0.34 0.06 0.08 0.02 4.34 1.721.65 0.63 0.26 0.07 0.10 0.01 4.52 1.581.76 0.76 0.18 0.03 0.13 0.00 4.71 1.73

0.84 0.46 0.16 0.04 0.07 0.03 4.68 0.800.90 0.51 0.16 0.07 0.10 0.02 4.99 0.831.06 0.61 0.22 0.03 0.14 0.01 5.29 1.031.14 0.65 0.08 0.01 0.14 0.00 5.54 1.13

0.69 0.50 0.12 0.03 0.06 0.02 0.11 0.660.76 0.62 0.15 0.04 0.08 0.01 0.14 0.720.39 0.39 0.05 0.02 0.05 0.00 0.30 0.37

0.73 0.41 0.16 0.10 0.08 0.08 0.18 0.630.85 0.49 0.18 0.07 0.07 0.05 0.30 0.780.79 0.49 0.19 0.04 0.06 0.02 0.43 0.750.77 0.41 0.19 0.02 0.03 0.00 0.32 0.75

1.61 0.51 0.33 0.05 0.05 0.02 0.54 1.561.72 0.46 0.29 0.14 0.16 0.01 0.38 1.581.46 0.47 0.24 0.13 0.08 0.01 0.54 1.331.14 0.45 0.18 0.10 0.07 0.00 0.98 1.330.31 0.20 0.05 0.01 0.21 0.00 0.47 0.30

1.27 0.29 0.19 0.21 0.07 0.12 0.03 1.061.40 0.58 0.17 0.14 0.10 0.03 0.63 1.261.38 0.61 0.24 0.13 0.11 0.02 0.22 1.251.41 0.52 0.20 0.17 0.10 0.05 0.49 1.241.28 0.67 0.10 0.08 0.09 0.01 0.63 1.200.22 0.19 0.02 0.00 0.03 0.00 0.34 0.22

550 W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

Almost all fractions of the extractable Fe, Al, and Mn vary betweensoil orders (Table 4); the only exception being Feo, Mnd, and Mnp.Dithionite and pyrophosphate extractable Fe contents (Fed and Fep,respectively) are also influenced by the soil depth. It is also shownthat among the extractable Mn fractions, only Mnp exhibitedstatistically relevant trends.

As can be seen in Fig. 8, exchangeable K andNa in both cultivated andnon-cultivated Mollisols are similar. In Alfisols, however, exchangeableCa, Mg, K, and Na are higher in cultivated soils. Regardless of manage-ment practices, Mollisols have a higher CEC, exchangeable Ca and K, andlower exchangeableNa than theAlfisols,whichmight be due to the highcontent of organic matter of the Mollisols. It is well established thatincrease of organic matter will increase the CEC and change the balanceof exchangeable cations in exchange sites. In theMollisols, the presence

Fig.8. Average values of CEC, base saturation, and exchangeable Ca2+, Mg2+, K+,

of clastic limestone (as small as sand-sized particles) in the profile likelycontributes to the predominance of Ca2+ on the exchange sites.

In terms of ANOVA (data is not given), CEC and exchangeablecations, are influenced by the interaction of soil order and manage-ment practice and, except for exchangeable Na, by soil depth.Therefore, the effect of management practice on these properties isexplained by considering the soil orders.

3.7. Elemental analysis of bulk soil and clay size particles of soil

There are differences in total elemental compositions between thetwo soil orders, especially Al2O3, Fe2O3, MgO, CaO, TiO2, and MnOcontents (Table 5). The contents of SiO2, K2O, Na2O, and P2O5, however,

and Na+ in the profiles of cultivated and uncultivated Alfisols and Mollisols.

Table 5Elemental composition of the Alfisol and Mollisol pedons studied. Presented as 95%confidence interval.

Oxides Alfisols Mollisols

SiO2 (%) 42.6±3.3 a 45.8±3.5 aAl2O3 (%) 23.6±1.0 a 14.8±1.8 bFe2O3 (%) 13.1±0.4 a 9.3±1.1 bK2O (%) 2.0±1.9 a 1.2±0.2 aMgO (%) 0.5±0.2 a 6.2±3.3 bCaO (%) 0.3±0.1 a 1.0±0.1 bNa2O (%) 0.1±0.01 a 0.4±0.2 aTiO2 (%) 3.1±0.1 a 1.9±0.2 bMnO (%) 1.6±0.5 a 0.9±0.1 bP2O5 (%) 0.2±0.1 a 0.2±0.04 aAl2O3+Fe2O3 (%) 36.7 24.1SiO2/Al2O3* 3.18 5.45

Numbers in the same row followed by different letters are significantly different atα=0.05.*Molar ratio=% oxide/molecular weight.

551W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

appear to be similar. Statistical analyses also showed that, except forMgO, elemental composition did not change significantly with depth.

Alfisols contain almost 9% more Al and 4% more Fe oxides than theMollisols. Titanium and Mn oxides are also higher in Alfisols than inMollisols. Since Al2O3 and Fe2O3 contents are higher in the Alfisolsthan the Mollisols, and assuming both formed from the same parentmaterial, it can be concluded that the Alfisols are in a more advancedstage of soil development than the Mollisols (Birkeland, 1999). This issupported by the fact that the SiO2/Al2O3 ratio of the Alfisols is lower(3.18) than that of the Mollisols (5.45). This is likely the higherdissolution of Si in Alfisols in comparison toMollisols, as they occur onmore stable surfaces. These results indicate that Alfisols and Mollisolsare different not only in appearance, but also in composition.

On average, the clay contains less SiO2, K2O, MgO, and Na2O buthigher Al2O3 and Fe2O3 in the Alfisols than theMollisols (Table 6). Thisbrings about higher values of Al2O3+Fe2O3 and lower molar ratio ofSiO2/Al2O3 in the Alfisols than in the Mollisols, supporting the ideathat Alfisols are more developed than Mollisols (Bech et al., 1997).

3.8. Clay mineralogy

The X-ray diffraction patterns of the clay (data not presented)show a distinct difference between Alfisols and Mollisols. In theAlfisols, Mg-saturated clays showed a sharp peak at 14.5 Å — thatshifts to 11.3 Å when K-saturated and heated to 550 °C. This partialcollapse upon heating is diagnostic for hydroxy-interlayered vermic-ulite. Peaks at 7.2 and 3.5 Å that disappear when samples are heatedto 550 °C are attributable to kaolinite; the peak at 4.8 Å is indicative ofgibbsite; and the peaks at 4.4 and 3.3 Å are indicative of quartz (Berry,1974; Whitting and Allardice, 1986). The composition of the Alfisolclay mineralogy does not change with soil depth.

Table 6Confidence interval at 95% of elemental analysis of clay size particles of selected soilsprofiles.

Oxides Alfisols Mollisols

SiO2 (%) 37.7±2.2 a 45.9±1.0 bAl2O3 (%) 33.5±1.0 a 25.1±1.1 bFe2O3 (%) 11.1±2.1 a 9.6±0.8 bK2O (%) 0.2±0.1 a 1.1±0.1 bMgO (%) 0.6±0.1 a 1.6±0.2 bCaO (%) 0.3±0.2 a 0.6±0.5 aNa2O (%) 0.5±0.1 a 1.3±0.2 bAl2O3+Fe2O3 (%) 44.6 34.7SiO2/Al2O3⁎ 1.98 3.22

Numbers in the same row followed by different letters are significantly different atα=0.05.*Molar ratio=% oxide/molecular weight.

The clay mineralogy of Mollisols is more variable among the soilprofiles and, in general, the patterns are different from those ofAlfisols. Sharp peaks at 14.0 Å in Mg saturated clays shifted to 18 Åwith Mg-glycerol solvation, indicating the presence of smectite. Thepeaks at 10, 5, and 3.3 Å (present for all the treatments) indicate themica. Peaks at 7.0 and 3.5 Å that disappear when heated to 550 °Cconfirm the presence of kaolinite. Calcite is present in Ck horizon ofthe Mollisols, indicated by the sharp peak at 3.0 Å.

The XRD patterns for other profiles, without the Ck, were similarall throughout the profile. We conclude that theMollisols aremineral-ogically different from the Alfisols. Indeed, the Mollisols containmainly a mixture of smectite, mica, and kaolinite, whereas the Alfisolsare predominantly kaolinite, vermiculite, and gibbsite with a minoramount of quartz.

3.9. Comparison of the properties of insoluble rock residue, soil and claysize particles

Oxides of Si, K, Mg, and Ca are higher in the insoluble residue ofrocks than in the clay of both soil orders (Fig. 9). Conversely, there arelesser amounts of Al and Fe oxides in the rock residues than in the clayand bulk soil. The compositional differences between the mainlysilicateminerals in the rock residues and the soils are not large and areconsistent with changes that can be expected with weathering. Forexample, the relatively insoluble Al2O3 and Fe2O3 are enriched in theclays, while more mobile elements such as Si, K, Mg, Ca, and Na arerelatively depleted in the clays.

The acid insoluble residues of limestones collected in both Alfisoland Mollisol regions are predominantly smectite, mica (muscoviteand illite), and kaolinite. Thus, the clay mineralogy of the insolubleresidue from both regions is more similar to the claymineralogy of theMollisols than that of the Alfisols. However, since there is kaolinite inboth residues and in the soil clay particles, it can be inferred that thesoil kaolinite has been inherited from the residue. Zie et al. (1982)reported the presence of kaolinitic clay in limestone rocks from China.

Upon weathering, the limestone dissolves and the silicate‘impurities’ are concentrated into the environment. This “freshly”released residue forms initial soil cover. Mollisols represent an early tomoderate stage of soil development and possess mainly the samekinds of minerals as those found in the insoluble residue. With timeand under suitable soil environments the minerals undergo transfor-mations. Mica loses its interlayer K (depotassication) and some of itslayer charge to form expansive 2:1 minerals such as smectite andvermiculite (Fanning et al., 1989). Under the conditions of low soilorganic matter, moderate acidity (pH∼5), and moderately activeleaching, vermiculite and smectite will, in turn, transform intohydroxy-interlayered minerals (Douglas, 1989; Borchardt, 1989;Barnhisel and Bertsch, 1989). As the landscape evolves and drainageimproves, vermiculite and smectite lose Si (desilication) and gain H+

Fig. 9. Elemental composition of insoluble rock residue and clay samples from Alfisolsand Mollisols.

552 W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

to form kaolinite and, if they contain Fe, iron oxides will alsoprecipitate (Borchardt, 1989). Under appropriate conditions, thekaolinite is in equilibriumwith gibbsite through the loss (desilication)or the gain (silication) of Si.

3.10. Soil classification

Soil physical, chemical, and mineralogical analyses indicate thatthe red soils are best classified as Alfisols and the black soils asMollisols. The red soils have argillic horizons and base saturationsN35%, diagnostic of Alfisols (Soil Survey Staff, 1999). Here the Bthorizon is in a direct contact with the underlying coral rocks. The colorof the entire argillic horizon has a hue of 10R (redder than 2.5YR),moist value of less than 3, and the dry value is the same as the wetvalue. This means that the soil is a Typic Rhodudalf (Soil Survey Staff,1999). The black soils have a mollic epipedon and the base saturationof all horizons above the lithic contact is N50%. Here the mollicepipedon is b50 cm thick. We propose to classify the Mollisols asHaprendolls (Soil Survey Staff, 1999).

4. Conclusions

The insoluble rock residues in the limestones found in the Alfisol andMollisol regions have the same physical, chemical, and mineralogicalproperties. Acid dissolution of the limestone yielded ≤1.7% insolubleresidue, which consists of montmorillonite, kaolinite, and mica. By as-suming zero transportation of soil materials into or out of the area, itwould require the weathering of about 29 m and 70 m of limestones toform 50-cm and 120-cm profiles of Mollisols and Alfisols, respectively.Microorganisms, especially fungi, may have played major roles in theweathering of the limestones through biochemical boring and dissolutionprocesses.

Alfisols and Mollisols are physically, chemically, and mineralogi-cally different. Alfisols have higher bulk density, pH, and total Fe, Al, Ti,and Mn throughout the soil profile than Mollisols. On the other hand,Mollisols possess higher CEC and exchangeable base cations, and totalCa and Mg than Alfisols. Oxides of Si, K, and P are similar in both soils.

Crystalline iron oxides are most prominent in Alfisols, whereas inMollisols poorly crystalline iron oxides aremore abundant. In general, theconcentrations of all iron oxides are higher in Alfisols than in Mollisols.More organically bound iron (Fep) occurs in Mollisols than in Alfisols.Since the Alfisols contain vermiculite, kaolinite, and gibbsite, and theMollisols contain montmorillonite, mica, and kaolinite minerals, weconclude that the Alfisols are at a more developed stage than Mollisols.This is also supported by analysis of metal oxides especially Si, Fe, and Aloxides.

We conclude that Alfisols and Mollisols have developed mainlyfrom the weathering of limestone. However, dust storms or volcanicash have likely contributed to the soil formation as well.

References

Anderson, J.U., 1963. An improved pretreatment for mineralogical analysis of samplescontaining organic matter. Clays Clay Min. 10, 380–388.

Badan Meteorologi dan Geofisika Kupang. 2000. Data iklim Provinsi Nusa TenggaraTimur (Climatic data of the Province of East Nusa Tenggara).Unpublished data.

Barlow, C., Gondowarsito, R., Birowo, A.T., Jayasuriya, S., 1990. InternationalDevelopment Issue. : Australian Gov. Publ. Service, vol. 13. Australian InternationalDevelopment Assistance Bureau, Canberra.

Barnhisel, R.I., Bertsch, P.M., 1989. Chlorites and hydroxy-interlayered vermiculite andsmectite. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in soil environments: Soil Sci.Soc. Am. Book Series no. 1. Soil Sci. Soc. Am, Madison, WI, pp. 729–828.

Barshad, I., Havely, E., Gold, H.A., Hagin, J., 1956. Clay minerals in some limestone soilsfrom Israel. Soil Sci. 81, 423–437.

Bech, J., Rustullet, J., Garrigó, J., Tobías, F.J., Martínes, R., 1997. The iron content of somered Mediterranean soils from northeast Spain and its pedogenic significance. In:Mermut, A.R., Yaalon, D.H., Kapur, S. (Eds.), RedMediterranean soils: Catena SpecialIssue no. 28, pp. 211–229.

Berry, L.G., 1974. Selected powder diffraction data for minerals. Joint Committee onPowder Diffraction Standards, Swarthmore, PA.

Birkeland, P.W., 1999. Soils and geomorphology. Oxford University Press, New York.Blake, G.R., Hartge, K.H., 1982. Bulk density. In: Page, A.L (Ed.), Methods of soil analysis.

Part 2. Chemical and microbiological properties: Agronomy no. 9. Am. Soc.Agronmy, Soil Sci. Soc. Am., Madison, WI, pp. 363–375.

Boero, V., Schwertmann, U., 1989. Iron oxide mineralogy of Terra Rossa and its geneticimplications. Geoderma 44, 319–327.

Borchardt, G., 1989. Smectite. In: Dixon, J.B.,Weed, S.B. (Eds.), Minerals in soil environments:Soil Sci. Soc. Am. Book Series no. 1. Soil Sci. Soc. Am, Madison, WI, pp. 675–727.

Carter, D.L., Mortland, M.M., Kemper, W.D., 1986. Specific surface. In: Klute, A. (Ed.),Methods of soil analysis Part 1. Agronomy no. 9: Agron. Soc. Am, Soil Sci. Soc. Am.,Madison, WI, pp. 413–423.

Cho, H.D., Mermut, A.R., 1992. Pedogenesis of two forest soils (Kandiustults) derivedfrom metamorphosed granite in Korea. Soil Sci. Soc. Am. J. 56, 517–525.

Dahlgren, R.A., 1994. Quantification of allophane and imogolite. In: Luxmoore, R.J. (Ed.),Quantitiativemethods in soil mineralogy. Soil Sci. Soc. Am,Madison,WI, pp. 430–451.

Darke, A.K., Walbridge, M.R., 1994. Estimating non-crystalline and crystallinealuminum and iron by selective dissolution in a riparian forest soil. Commun.Soil Sci. Plant Anal. 25, 2089–2101.

Deer, W.A., Howie, R.A., Zussman, J., 1967. An introduction to the rock-formingminerals. Longmans, Green and Co., Ltd, London.

Doner, H.E., Lynn, W.C., 1989. Carbonate, halide, sulfate, and sulfideminerals. In: Dixon,J.B., Weed, S.B. (Eds.), Minerals in soil environments: Soil Sci. Soc. Am. Book Seriesno. 1. Soil Sci. Soc. Am, Madison, WI, pp. 279–330.

Douglas, L.A., 1989. Vermiculites. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in soilenvironments: Soil Sci. Soc. Am. Book Series no. 1. Soil Sci. Soc. Am, Madison, WI,pp. 635–674.

Duchaufour, P., 1982. Pedology. Pedogenesis andclassification.GeorgeAllen&Unwin, Boston.Durn, G., Ottner, F., Slovenec, D., 1999. Mineralogical and geochemical indicators of the

polygenetic nature of Terra Rossa in Istria, Croatia. Geoderma 91, 125–150.Ehrlich, H.L., 1990. Geomicrobiology. Marcel Dekker, Inc, New York.Esteban, M., Klappa, C.F., 1983. Subaerial exposure environment. In: Scholle, P.A.,

Bebout, D.G., Moore, C.H. (Eds.), Carbonate depositional environments: TheAmerican Association of Petroleum Geologists Memoir, vol. 33, pp. 1–54.

Fanning, D.S., Keramidas, V.Z., El-Desoky, M.A., 1989. Micas. In: Dixon, J.B., Weed, S.B.(Eds.), Minerals in soil environments: Soil Sci. Soc. Am. Book Series no. 1. Soil Sci.Soc. Am, Madison, WI.

Gee, G.W., Bauder, J.W., 1986. Particle size analysis. In: Page, A.L. (Ed.), Methods of soilanalysis. Part 1. Physical and mineralogical methods: Agronomy no. 9. Am. Soc.Agronmy, Soil Sci. Soc. Am., Madison, WI, pp. 383–411.

Glazovskaya, M.A., Parfenova, E.I., 1974. Biochemical factors in the formation of TerraRossa in the Southern Crimea. Geoderma 12, 57–82.

Golubić, S., Schneider, J., 1979. Carbonate dissolution. In: Trudinger, P.A., Swaine, D.J.(Eds.), Biogeochemical cycling of mineral-forming elements. : Studies in Environ-mental Science, vol. 3. Elsevier Sci. Publ. Co, Amsterdam, pp. 107–129.

Hendershot,W.H., Lalonde,H., 1993. Ion exchangeandexchangeable cations. In:Carter,M.R.(Ed.), Soil sampling and methods of analysis. Lewis Publ, Boca Raton, pp. 167–176.

Ives, D.W., Furkert, R.J., Searle, P.L., Daly, B.K., Whitton, J.S., 1976. A soil sequencedeveloped over coral limestone on the uplifted coastal terraces of Sumba, East NusaTenggara. New Zealand. J. Sci. 19, 101–103.

Jackson, M.L., 1969. Soil chemical analysis-advance course. Published by the author (outof print). Department of Soils. University of Wisconsin, Madison, WI.

Jouannic, C., Hoang, C., Hantoro, W.S., Delionom, R.M., 1988. Uplift rate of coral reefterraces in the area of Kupang, West Timor: preliminary results. Palaeogeogr.Palaeoclimatol. Palaeoecol. 68, 259–272.

Kodama, H., 1985. Infrared spectra of minerals. Reference guide to identification andcharacterization of minerals for the study of soils. Research Branch, AgricultureCanada, Ottawa.

Lim, C.H., Jackson, M.L., 1982. Dissolution of total elemental analysis. In: Page, A.L. (Ed.),Methods of soil analysis. Part 2. Chemical and microbiological properties:Agronomy no. 9. Am. Soc. Agronmy, Soil Sci. Soc. Am., Madison, WI, pp. 1–12.

Macleod, D.A., 1980. The origin of the red Mediterranean soils in Epirus, Greece. J. SoilSci. 31, 125–136.

McLean, E.O., 1982. Soil pH and lime requirement. Methods of soil analysis. Part 2.Chemical and microbiological properties. : Agronomy no. 9. Am. Soc. Agronmy, SoilSci. Soc. Am., Madison, WI. A.L. Page, pp. 225–246.

Merino, E., Banerjee, A., 2008. Terra rossa genesis, implications for karst, and eoliandust; a geodynamic thread. J. Geol. 116, 62–75.

Moresi, M., Mongelli, G., 1988. The relation between the Terra Rossa and the carbonate-free residue of the underlying limestones and dolostones in Apulia, Italy. Clay Min.23, 439–446.

Muhs, D.R., Bush, C.A., Stewart, K.C., Rowland, T.R., Crittenden, R.C., 1990. Geochemicalevidence of Saharan dust parent material for soils developed on Quaternarylimestones of Caribbean and western Atlantic islands. Quat. Res. 33, 157–177.

Muhs, D.R., 2001. Evolution on Quaternary reef terraces of Barbados, West Indies. Quat.Res. 56, 66–78.

Nagatsuka, S., Kaneko, S., Ishihara, A., 1983. Soil genesis on the raised coral reef terracesof Ishigaki- and Okinawa-islands in the Ryuku Islands, Japan. Soil Sci. Plant Nutr. 29,343–354.

Nihlén, T., Lund, S.O., 1995. Influence of eolian dust on soil formation in the Aegian area.Z. Geomorphol. 39, 341–361.

Olson, C.G., Ruhe, R.V., Mausbach, M.J., 1980. The terra rossa limestone contactphenomena in karst, Southern Indiana. Soil Sci. Soc. Am. J. 44, 1075–1079.

Pettijohn, F.J., 1957. Sedimentary rocks. Harper & Row Publ, New York.Pingitore Jr., N.E., 1971. Submarine precipitation of void-filling needles in Pleistocene

coral. In: Bricker, O.P. (Ed.), Carbonate cements. : The Johns Hopkins UniversityStudies in Geology No. 19. The Johns Hopkins Press, Baltimore, pp. 68–71.

553W. Mella, A.R. Mermut / Geoderma 154 (2010) 544–553

Rabenhorst, M.C., Wilding, L.P., 1986. Pedogenesis on the Edwards Plateau, Texas:I. Nature and continuity of parent material. Soil Sci. Soc. Am. J. 50, 678–687.

Ray, S., Gault, H.R., 1961. Mineralogy of Jacksonburg (Middle Ordovician) formation inWestern Pennsylvania and western New Jersey. Bull. Am. Assoc. Petrol. Geol. 45,39–50.

Ross, G.J., Wang, C., 1993. Extractable Al, Fe, Mn, and Si. In: Carter, M.R. (Ed.), Soilsampling and methods of analysis. Lewis Publ, Boca Raton, pp. 239–246.

Ruhe, R.V., Dady, J.G., Gomez, R.S., 1961. Paleosol of Bermuda. Geol. Soc. Am. Bull. 72,1121–1142.

Sheldrick, B.H., Wang, C., 1993. Particle size distribution. In: Carter, M.R. (Ed.), Soilsampling and methods of analysis. Lewis Publ, Boca Raton, pp. 499–512.

Soil Survey Staff, 1999. Keys to soil taxonomy. United States Department of Agriculture,Natural Resources Conservation Service.

Stace, H.C.E., 1956. Chemical characteristics of Terra Rossa and Rendzinas of SouthAustralia. J. Soil Sci. 7, 280–293.

Sweeting, M.M., 1972. Karst landform. The Macmillan Press Ltd, London.Tarzi, J.G., Paeth, R.C., 1975. Genesis of a Mediterranean red and white Rendzina soil

from Lebanon. Soil Sci. 120, 272–277.

Trudgill, S., 1985. Limestones geomorphology. Longman, London.Van Bemmelen, R.W., 1949. The geology of Indonesia. : General geology of Indonesia

and adjacent archipelagoes, vol. 1A. Government Printing Office, The Hague.Van der Marel, H.W., Beutelspacher, H., 1976. Atlas of infrared spectroscopy of clay

minerals and their admixtures. Elsevier Scientific Publ. Co., Amsterdam.White, J.L., Roth, C.B., 1986. Infrared spectrometry. In: Page, A.L. (Ed.), Methods of soil

analysis. Part 1. Physical and mineralogical methods: Agronomy no. 9. Am. Soc.Agronomy, Soil Sci. Soc. Am., Madison, WI, pp. 291–330.

Whitting, I.D., Allardice, W.R., 1986. X-ray diffraction techniques. In: Page, A.L. (Ed.),Methods of soil analysis. Part 1. Physical andmineralogical methods: Agronomy no.9. Am. Soc. Agronomy, Soil Sci. Soc. Am., Madison, WI, pp. 331–362.

Yassoglou, N., C. Kosmas, and N. Moustakas. 1997. The red soils, their origin, properties,use, andmanagement in Greece. p. 28: 261–278. In A.R. Mermut, D.H. Yaalon, and S.Kapur (eds.) Red Mediterranean soils. Special issue, Catena 28.

Zie, Z, L. Daren, F.Molin, F. Boahua, and J. Taiquan. 1982. Kaolin deposits in China. p. 719–731.In H. Van Olphen and F. Veniale (eds.) Proc. Int. Clay Conf. 1981. Elsevier North HollandInc., New York.