ARTICLE IN PRESS

Soil Biology & Biochemistry 39 (2007) 684–690

www.elsevier.com/locate/soilbio

Short communication

Bacterial diversity of terra preta and pristine forest soil from the
Western Amazon
Jong-Shik Kima, Gerd Sparovekb, Regina M. Longoc, Wanderley Jose De Meloc,
David Crowleya,
a
Department of Environmental Sciences, University of California, Riverside, CA, USA
b
Department of Soil Science, ESALQ, University of Sao Paulo, Piracicaba CP 9, CEP 13418-900, Brazil
c
Department of Technology, Universidade Estadual Paulista, Jaboticabal, SP, CEP 14884-900, Brazil
Received 25 May 2006; received in revised form 1 August 2006; accepted 11 August 2006
Available online 18 September 2006

Abstract

The survey presented here describes the bacterial diversity and community structures of a pristine forest soil and an anthropogenic
terra preta from the Western Amazon forest using molecular methods to identify the predominant phylogenetic groups. Bacterial
community similarities and species diversity in the two soils were compared using oligonucleotide fingerprint grouping of 16S rRNA gene
sequences for 1500 clones (OFRG) and by DNA sequencing. The results showed that both soils had similar bacterial community
compositions over a range of phylogenetic distances, among which Acidobacteria were predominant, but that terra preta supported
approximately 25% greater species richness. The survey provides the first detailed analysis of the composition and structure of bacterial
communities from terra preta anthrosols using noncultured-based molecular methods.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Bacterial diversity; Forest soils; Microbial ecology; Terra preta

1. Introduction
Terra preta anthrosols in the Amazon basin are nutrient
rich soils that contain thick, dark colored, surface horizons
with a high organic matter content (Lima et al., 2002) and
are noted for their exceptional fertility and ability to
accumulate stable organic carbon. Prior studies suggest
these soils were formed by pre-Columbian, Amerindians,
who practiced a ‘‘slash and char’’ agriculture (Mann,
2002). The ability of terra preta to accumulate stable
organic matter has attracted attention as a possible means
for increasing fertility and carbon storage in tropical soils
(Sombroek, 1966; Lehmann et al., 2003), and has provoked
questions regarding the role of microorganisms in the
formation and plant growth promotion characteristics of
these soils (Thies and Suzuki, 2003). The objective of this
study was to carry out a survey of the bacterial diversity in
Corresponding author. Tel.: +1 951 827 3785; fax: +1 951 827 3993.
E-mail address: crowley@ucr.edu (D. Crowley).
terra preta and pristine forest soil as the basis for studies on
the ecology of these forest soils.
To characterize bacterial diversity in the terra preta and
pristine forest soil, samples were collected from two
adjacent locations in the Jamari National Forest, of
Rondonia, Brazil (latitude: 81 450 0 S, longitude: 631 270 0
W). The soil from the pristine forest was a clay loam ultisol
having a 1 cm A horizon under a layer of moist duff. The B
horizon consisted of a highly oxidized ultisol with an acid
pH of 3.1 and had negligible organic matter. The terra
preta samples were collected at a site two hundred meters
distant where the soil contained a deep, organic matter
rich, sandy loam in the A horizon that was approximately
1 m thick before transitioning into the underlying clay
ultisol. Intact soil cores measuring 10 cm in depth and 8 cm
in width were collected from the surface horizon below the
organic litter layer using sterile, stainless steel soil coring
sleeves. The cores were shipped immediately to the
University of California, Riverside, USA, where they were
frozen at 80 1C. At the time of analysis, three cores from

0038-0717/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.soilbio.2006.08.010
 ARTICLE IN PRESS
J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690
each site were combined into one composite sample
and sieved to pass a 1.0 mm screen. The composite
samples were analyzed for their chemical and physical
characteristics and metal contents (Supplemental
Table 1, online supplementary material) and were subjec-
ted to molecular analyses to identify the predomi-
nant bacteria and to describe the bacterial community
structures at different phylogenetic distances based on
comparisons of 16S rRNA gene sequences recovered from
the samples.
The analysis of bacterial diversity employed a two step
procedure in which the first step was to sort 16S rRNA
clone libraries with a method referred to as oligonucleotide
fingerprinting of ribosomal genes (OFRG) (Valinsky et al.,
2002). This was followed by sequencing of random clones
from the taxonomic clusters generated by OFRG to
identify the predominant bacteria. The gene sequences
were also used to construct phylogenetic trees based on
actual sequence data. To produce the 16S rRNA gene clone
libraries, near full-length (1465 bp) small-subunit rRNA
gene sequences were PCR amplified from soil DNA. The
purified DNA was ligated into pGEM-T (Promega) and
transformed into competent Escherichia coli JM109 (Pro-
mega). Libraries of the rRNA genes were represented by
768 clones for each sample, which were arrayed on
replicate nylon membranes for hybridization with 33P
DNA end-labeled oligonucleotide probes as described by
Valinsky et al., (2002). Signal intensities with background
correction were obtained using ImaGene 4.0 software
(Biodiscovery) and were transformed for every probe
into three values: 0, 1, and N, where 0 and 1 indicate
negative and positive hybridization events, respectively,
and N indicates an uncertain assignment. This process
creates a hybridization fingerprint for each clone, which is
a vector of values resulting from hybridizations with all
probes. Estimates of diversity using 16S rRNA gene
sequences were examined based at 95%, and 90%
similarities using the PHYLIP-formatted distance matrices
generated by the computer program PAUP 4.0 (Sinauer
Associates, Inc.).
To identify the taxonomic groups separated by OFRG,
nucleotide sequences were determined for 76 and 49 rRNA
gene clones that were randomly selected from within each
of the clusters generated by OFRG. The 16S rRNA gene
sequences were submitted to the BLAST server (Basic
Local Alignment Search Tool), National Center for
Biotechnology Information (NCBI) to determine the
closest matching sequences in the GenBank and to infer
phylogenetic affiliations. Novel sequences were deposited
at GenBank and were assigned accession numbers
AY326512–AY326636.
Phylogenetic trees based on the sequenced clones were
constructed using the 16S rRNA gene sequence data.
Nucleotide sequences were aligned using Clustal X, after
which an evolutionary distance matrix was generated
using the program MEGA3 (Kumar et al., 2004) using
the N-J method. Bootstrap analyses of the neighbor-
685
joining data were conducted based on 1000 samplings to
assess the stability of the phylogenetic relationships.
Rarefaction curves and estimates of species diversity at
different phylogenetic distances were determined using
the computer program DOTUR (Schloss et al., 2004).
Comparisons of the species coverage and overall commu-
nity similarities at different phylogenetic distances were
determined using the program LIBSHUFF (Singleton
et al., 2001).
2. Results and discussion
The Amazon contains diverse soils of which only a few
have been characterized with respect to their microbiology.
To date there has been only one prior survey of Amazon
soils that used molecular methods to analyze soil microbial
diversity in which 100 clones were analyzed (Borneman and
Triplett, 1997). The survey data presented here thus
provide a more comprehensive survey and the first analysis
of terra preta using molecular methods. Inspection of the
taxonomic trees generated by OFRG showed that the two
soils differed with respect to the number of operational
taxonomic units (OTUs), and the number of branches that
represent different taxonomic groups, with terra preta
having overall greater diversity (Fig. 1). As shown using
rarefaction, there was significant separation of OTU
richness (95% similarity level) as sample size increased
above 50 for each sample set (Fig. 2). Enumeration of
the unique 16S rRNA gene sequences yielded 396
OTUs from terra preta as compared to 291 OTUs in the
forest soil (Table 1). Calculation of Shannon index values
using DOTUR yielded significantly different (P40:05)
values for terra preta and forest soil (5.2 and 4.37,
respectively). Other commonly used measures of diversity
including Simpsons Index of Diversity and Chao I, which
employ independent mathematical approaches to measur-
ing diversity revealed the same phenomenon. Altogether,
these analyses suggest that bacterial species richness was
approximately 25% greater for terra preta than that in the
forest soil.
Using the program LIBSHUFF to compare the simila-
rities of the communities from each soil at different
phylogenetic distances, it was determined that overall
coverage of sequence diversity was high, with values of
60% and 68% for the forest soil and terra preta samples
sets, respectively (Fig. 3). Comparison of the forest soil
sample set against terra preta revealed that all of the
OTU diversity in the forest soil was represented in terra
preta and that the forest sample was not signifi-
cantly different from the terra preta (P ¼ 0:64; data not
shown). In contrast, comparison of the terra preta versus
the forest soil showed that terra preta was significantly
different (P ¼ 0:05), in that it contained additional
sequences that did not occur in the forest soil. The greatest
difference in 16S rRNA gene sequences between terra
preta and the forest soil occurred for OTUs having up to
5% evolutionary distance (95% 16S rRNA similarity),
 ARTICLE IN PRESS
686 J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690

Terra Preta Preserved Forest

Nitrospira
Proteobacteria
Chloroflexi
Acidobacterium Bacillus Proteobacteria

Acidobacterium

Acidobacterium

alpha-Pro

Actinobacteria

CFB
gamma-Pro

beta-Pro
Verrucomicrobia Acidobacterium

Actinobacteria

alpha-Pro

Verrucomicrobia
Actinobacteria
Delta-Pro
Chloroflexi
gamma-Pro
beta-Pro
Planctomycetes
Proteobacteria Acidobacterium

Verrucomicrobia Actinobacteria
0.05

Fig. 1. Taxonomic cluster analysis of 16S rRNA gene sequences from terra preta and adjacent pristine forest soil based on oligonucleotide fingerprinting
of 16S rRNA gene sequences. The 16S rRNA clone libraries were generated for single samples consisting of three composited replicate soil cores (litter
layer removed) taken from the top 10 cm of the soil profile at each location.

corresponding to species and genera. The coverage curves
converged at 95% evolutionary distance and then sepa-
rated again at distances from 96% to 90%, indicating that
there were also significant differences between the commu-
nities at deeper phylogenetic levels. Complete coverage of
the two communities was obtained at greater than 90%
similarity, suggesting that the large data set that was
analyzed here, with ca. 700 clones per sample, represented
all of the major bacterial phyla that were present in the
soils.
The differences in community composition revealed by
the LIBSHUFF analysis were confirmed by the phyloge-
netic analyses based on DNA sequencing. Identities of the
major bacterial groups by direct sequencing of clones
representing different clusters revealed 14 phylogenetic
groups in terra preta, as compared to 9 from the forest soil
(Fig. 4). Among the major groups represented, Acidobac-
terium was predominant, comprising approximately 30%
of the bacteria in terra preta and 50% of the forest soil.
The phylogenetic trees included two possible new clades
of Acidobacterium that were distinct from the pre-
viously described subgroups for this phylum. Other
bacterial groups that were common to both two soils
included the Proteobacteria, Actinobacteria, Planctomy-
cetes, and Verrucomicrobia. Representatives of the a, b,
and g Proteobacteria were similarly abundant in terra
preta; whereas mostly a Proteobacteria were found in the
forest soil.
Acidobacterium sp are common in many forest soils
around the world, comprising from 12% of the non-
cultured species surveyed in Austrian forest soils under
pine to 35% in spruce-fir-beech soils in Europe (Hackl
et al., 2004). At least four subgroups of these bacteria have
been recognized previously using specific 16S rRNA gene
primers in a survey of 43 soils and sediments representing a
range of environments (Barns et al., 1999). This earlier
survey suggested that the A and G subgroups are
ubiquitous, whereas the Y and O subgroups are largely
absent from acid soils. The A subgroup is thought to be the
most phylogenetically diverse, with deep branches in the
phylogenetic tree that may represent distinct lineages. Our
results are largely in agreement with these suppositions.
Here, most of the cloned bacterial sequences were in the
A and G subgroups. Terra preta contained sequences from
the Y subgroup, which purportedly favors non-acid soils.
Only the A subgroup was found in the acid soil from the
pristine forest. In the phylogenetic tree generated by DNA
sequencing, the BLAST analysis revealed three possible
 ARTICLE IN PRESS
J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690 687

350

Terra Preta
300
Forest Soil

250

Number of OTUs observed

200

150

100

50

0
40 80 120 160 200 240 280 320 360 400 440 480 520 560 600 640 680
Number of sequences sampled

Fig. 2. Rarefaction analysis of 16S rRNA gene sequences from terra preta and forest soil. Gene sequences were differentiated by oligonucleotide
fingerprint grouping at 0.05 dissimilarity level. Vertical bars indicate 95% confidence intervals.

Table 1
1.00
Measures of bacterial species diversity based on OFRG analysis of 0.0030
microbial communities from terra preta and Amazon forest soil 0.80 0.0025
Coverage, C

(Cx-Cxy)2
Index Terra preta Forest soil 0.0020
0.60
OTU richness 0.0015
Unique OFRG groups / clones 396/742 291/768 0.40
0.0010
0.05 evolutionary distance 287 196
0.20
0.10 evolutionary distance 210 142 0.0005
Shannon index 5.2 4.37
0.00
Simpsons index 0.015 0.34
0.0 0.1 0.2 0.3 0.4 0.5
Estimated richness Chao1 (0.05 413 277
evolutionary distance) Evolutionary distance, D

new subgroups of Acidobacterium. Two new clades of
Acidobacterium were found in the terra preta, one that was
most closely related to the G subgroup, and a second well-
defined cluster of five clones that were most closely related
to the Y-subgroup. The new clade of Acidobacterium in the
forest soil was comprised by four clones that occurred
within the A-subgroup and were most similar to previous
accessions designated as Holophaga.
Based on prior knowledge, soil pH is likely to be one of
the most important selection factors affecting the microbial
species composition in different soils (Fierer and Jackson,
2006). Here, both soils supported abundant plant
growth and carried similar above ground vegetation
consisting of a diverse community of Amazon forest
tree species and understory plants. The relationship
between organic matter inputs from different overstory
trees, rhizosphere effects, and the bacterial community
Fig. 3. LIBSHUFF analysis of 16S rRNA gene sequences from terra preta
(closed circles) compared to adjacent forest soil (open circles) showing
differences in composition of the bacterial communities (y-axis) over a
range of evolutionary distances (D) (x-axis). Solid lines indicate the value
of (CX-CXY)2 for samples at each value of D. Broken lines indicate the
P ¼ 0:05 value of (CX-CXY)2 for the randomized samples.
composition of forest soils is still not understood, but likely
contributes to differences in diversity and community
composition along the forest floor. There was also
increased earthworm activity and soil aggregation in terra
preta that may provide a wide range of niches and thereby
contribute to increased bacterial species diversity. Addi-
tional surveys and comparisons at different locations will
be needed to characterize other locations with terra preta.
It may then be possible to unravel the ecology of these
bacterial communities and study the role of specific
bacterial groups that contribute to the many interesting
properties of these soils.
 ARTICLE IN PRESS
688 J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690

91 869-1 Uncultured Acidobacteria (95%)

99 22-1 Uncultured Acidobacteria ( 95%)
91 26-1 Uncultured Acidobacteria (94%)
G
73 99-1 Uncultured Acidobacteria (96%)
210-1 Uncultured Acidobacteria (92%)
99 79 1192-1 Uncultured Acidobacteria (95%)
997-1 Uncultured Acidobacteria (92%) New
99 483-1 Uncultured Acidobacteria (87%)
99 894-1 Uncultured Acidobacteria (93%)
291-1 Uncultured Acidobacteria (95%)
1086-1 Uncultured Acidobacteria (92%) Acidobacterium
18-1 Uncultured Acidobacteria (92%)
83 991-1 Uncultured Acidobacteria (97%) A
99 1271-1 Uncultured Acidobacteria (94%)
97 1272-1 Uncultured Acidobacteria (94%)
9999 1005-1 Uncultured Acidobacteria (93%)
99 1091-1 Uncultured Acidobacteria (97%) Y
70 395-1 Uncultured Acidobacteria (93%)
1093-1 Uncultured Acidobacteria (91%)
99 1267-1 Uncultured Acidobacteria (92%)
99 1363-1 Uncultured Acidobacteria (93%) New
99 7-1 Uncultured Acidobacteria (93%)
87 501-1 Uncultured Acidobacteria (91%)
81 31-1 Uncultured Acidobacteria (93%)
40-1 Streptomyces kasugaensis (94%)
99 37-1 Streptomyces kasugaensis (94%)
34-1 Streptomyces kasugaensis (94%)
99 42-1 Streptomyces kasugaensis (94%) Actinobacteria
41-1 Streptomyces kasugaensis (94%)
75 911-1 Uncultured Rubrobacteridae (93%)
15-1 Uncultured Rubrobacteridae (95%)
99 43-1 Paenibacillus macerans (95%)
99 1078-1 Bacillus cereus (98%)
99 1077-1 Bacillus cereus (99%) Bacillus
993-1 Bacillus bataviensis (96%)
99 307-1 Bacillus bataviensis (97%)
98 1180-1 Uncultured Chloroflexi (87%)
1067-1 Uncultured Chloroflexi (92%) Chloroflexi
99 12-1 Uncultured Chloroflexi (94%)
98 99 4-1 Parachlamydia acanthamoebae (93%) Chlamidiae
3-1 Parachlamydia acanthamoebae (94%)
99 27-1 Uncultured Verrucomicrobia (95%)
28-1 Uncultured Verrucomicrobia (95%) Verrucomicrobia
30-1 Uncultured Verrucomicrobia (95%)
1164-1 Nitrospira moscoviensis (96%)
1163-1 Nitrospira moscoviensis (95%) Nitrospira
99 6-1 Uncultured Bacteroidetes (92%)
5-1 Uncultured Bacteroidetes (92%) Bacteroidetes
99140-1 candidate division TM7 (89%) TM7
227-1 Afipia genosp (93%)
99 128-1 Afipia genosp (93%)
99
98 594-1 Uncultured alpha proteobacterium (97%) Alpha Proteobacteria
99 711-1 Bradyrhizobium japonicum (98%)
1100-1 Bradyrhizobium japonicum (98%)
597-1 Mesorhizobium sp. (90%)
54 776-1 Uncultured gamma proteobacterium (92%)
47-1 Nitrosococcus halophilus (90%) Gamma Proteobacteria
82
99 1083-1 Uncultured beta proteobacterium (95%)
99 966-1 Uncultured beta proteobacterium (89%)
1251-1 Burkholderia sp. (98%)
99 99 592-1 Uncultured beta proteobacterium (94%) Beta Proteobacteria
78 591-1 Uncultured beta proteobacterium (94%)
142-1 Burkholderia sp. (98%)
141-1 Burkholderia sp. (98%)
2-1 Unknown
46-1 Uncultured Planctomycete (93%) Planctomycetes
99 426-1 Anaeromyxobacter dehalogenans (90%)
293-1 Uncultured delta proteobacterium (92%)
89 685-1 Geobacter sulfurreducens (87%)
99 1074-1 Geobacter sulfurreducens (87%)
24-1 Syntrophus gentianae (88%) Delta Proteobacteria
900-1 Uncultured delta proteobacterium (92%)
221-1 Uncultured delta proteobacterium (95%)
94 772-1 Uncultured delta proteobacterium (95%)
99 1154-1 Uncultured delta proteobacterium (96%)
1001-1 Uncultured delta proteobacterium (96%)
98

0.05
(A)

Fig. 4. Phylogenetic trees for unique 16S rRNA gene sequences based on rRNA gene sequencing. (A) Terra preta soil (B) Forest soil.
 ARTICLE IN PRESS
J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690 689

98 340-2 Uncultured Acidobacteria (92%)

355-2 Uncultured Acidobacteria (91%)
100 339-2 Uncultured Acidobacteria (92%)
753-2 Uncultured Acidobacteria (92%)
100 169-2 Uncultured Acidobacteria (92%)
1209-2 Uncultured Acidobacteria (92%)
1496-2 Uncultured Acidobacteria (92%)
466-2 Uncultured Acidobacteria (93%) A
1151-2 Uncultured Acidobacteria (93%)
1215-2 Uncultured Acidobacteria (92%)
100 1216-2 Uncultured Acidobacteria (92%)
81-2 Uncultured Acidobacteria (94%) Acidobacteria
92-2 Uncultured Acidobacteria (94%)
100
100 1031-2 Uncultured Acidobacteria (95%)
1219-2 Uncultured Acidobacteria (88%)
244-2 Uncultured Acidobacteria (96%)
83 821-2 Uncultured Acidobacteria (94%)
760-2 Uncultured Acidobacteria (95%) New
99 100
99 958-2 Uncultured Acidobacteria (94%)
829-2 Uncultured Acidobacteria (93%)
648-2 Uncultured Acidobacteria (94%)
100
462-2 Uncultured Acidobacteria (92%) A
55-2 Uncultured Acidobacteria (96%)
100 157-2 Uncultured Acidobacteria (96%)
100
90-2 Frateuria aurantia (96%)
100 845-2 Uncultured gamma proteobacteria (96%) Gamma Proteobacteria
96 1052-2 Frateuria aurantia (96%)
99 1048-2 Burkholderia unamae (97%) Beta Proteobacteria
100 1131-2 Burkholderia tropica (98%)
1236-2 Uncultured gamma proteobacterium (94%)
100 52-2 Uncultured alpha proteobacterium (97%)
71 557-2 Bradyrhizobium genosp (99%)
99 98 722-2 Sphingomonas sp. (96%)
670-2 Afipia felis (89%)
100 178-2 Uncultured alpha proteobacterium (95%) Alpha Proteobacteria
100 288-2 Azospirillum amazonense (92%)
87 337-2 Acidosphaera rubrifacien (95%)
1503-2 Acidosphaera rubrifacien (91%)
100
100 1529-2 Acidosphaera rubrifacien (93%)
1202-2 Uncultured actinobacterium (92%)
100
1309-2 Acidothermus cellulolyticus (92%) Actinobacteria
100 1532-2 Frankia sp. (92%)
271-2 Unknown
100 1126-2 Uncultured planctomycete (90%)
345-2 Nostocoida limicola (96%) Planctomycetes
530-2 Parachlamydia acanthamoebae (93%) Chlamydiae
546-2 Uncultured Verrucomicrobia (96%)
100
241-2 Uncultured Verrucomicrobia (89%) Verrucomicrobia
100 941-2 Uncultured Verrucomicrobia (88%)

0.05
(B)

Fig. 4. (Continued)

Acknowledgments
The authors gratefully acknowledge Lea Valinsky and
James Borneman for assistance with oligonucleotide
fingerprinting methods, and Steven Qi for technical
assistance.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found in the online version at doi:10.1016/j.soilbio.2006.08.
010
References
Barns, S.M., Takala, S.L., Kuske, C.R., 1999. Wide distribution and
diversity of members of the bacterial kingdom Acidobacterium in the
environment. Applied and Environmental Microbiology 65,
1731–1737.
Borneman, J., Triplett, E.W., 1997. Molecular microbial diversity in soils
from eastern Amazonia: evidence for unusual microorganisms and
microbial population shifts associated with deforestation. Applied and
Environmental Microbiology 63, 2647–2653.
Fierer, N., Jackson, R.B., 2006. The diversity and biogeography of soil
bacterial communities. Proceedings of the National Academy of
Sciences 103, 626–631.
Hackl, E., Zechmeister-Boltenstern, S., Bodrossy, L., Sessitsch, A., 2004.
Comparison of diversities and compositions of bacterial populations
inhabiting natural forest soils. Applied and Environmental Micro-
biology 70, 5057–5065.
Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: integrated software for
molecular evolutionary genetics analysis and sequence alignment. Brief
Bioinformation 5, 150–163.
Lehmann, J., Kern, D., Glaser, B., Woods, W., 2003. Amazonian Dark
Earths: Origin, Properties, Management. Kluwer Academic Publish-
ers, The Netherlands, 523pp.
Lima, H.N., Schaefer, C.E.R., Mello, J.W.V., Gilkes, R.J., Ker, J.C.,
2002. Pedogenesis and pre-Colombian land use of ‘‘Terra Preta
Anthrosols’’ (‘‘Indian black earth’’) of Western Amazonia. Geoderma
110, 1–17.
Mann, C.C., 2002. The real dirt on rain forest fertility. Science 297, 920–923.
Schloss, P.D., Larget, B.R., Handelsman, J., 2004. Integration of
microbial ecology and statistics: a test to compare gene libraries.
Applied and Environmental Microbiology 70, 5485–5492.
Singleton, D.R., Furlong, M.A., Rathbun, S.L., Whitman, W.B., 2001.
Quantitative comparisons of 16S rRNA gene sequence libraries from
environmental samples. Applied and Environmental Microbiology 67,
4374–4376.
Sombroek, W.G., 1966. Amazon Soils: A Reconnaissance of the Soils of
the Brazilian Amazon Region. Centre for Agricultural Publications
and Documentation, Wageningen, Netherlands, 300pp.
 ARTICLE IN PRESS
690 J.-S. Kim et al. / Soil Biology & Biochemistry 39 (2007) 684–690

Thies, J., Suzuki, K., 2003. Amazonian dark earths: biologi-
cal measurements. In: Lehmann, J., Kern, D.C., Glaser, B.,
Woods, W. (Eds.), Amazonian Dark Earths: Origin, Proper-
ties, Management. Kluwer Academic Publishers, The Netherlands,
pp. 232–287.
Valinsky, L., Vedova, G.D., Scupham, A., Alvey, S., Figueroa, A., Yin,
B., Hartin, R.J., Chrobak, M., Crowley, D.E., Jiang, T., Borneman, J.,
2002. Analysis of microbial community composition using oligonu-
cleotide fingerprinting of ribosomal RNA genes. Applied and
Environmental Microbiology 68, 3243–3250.