Content uploaded by Rebecca E Drenovsky
Author content
All content in this area was uploaded by Rebecca E Drenovsky
Content may be subject to copyright.
Soil Water Content and Organic Carbon Availability Are Major
Determinants of Soil Microbial Community Composition
R.E. Drenovsky, D. Vo, K.J. Graham and K.M. Scow
Department of Land, Air and Water Resources, University of California, Davis, One Shields Avenue, Davis, CA 95616-8627, USA
Received: 23 September 2003 / Accepted: 26 November 2003 / Online publication: 23 September 2004
Abstract
Exploration of environmental factors governing soil mi-
crobial community composition is long overdue and now
possible with improved methods for characterizing mi-
crobial communities. Previously, we observed that rice
soil microbial communities were distinctly different from
tomato soil microbial communities, despite management
and seasonal variations within soil type. Potential con-
tributing factors included types and amounts of organic
inputs, organic carbon content, and timing and amounts
of water inputs. Of these, both soil water content and
organic carbon availability were highly correlated with
observed differences in composition. We examined how
organic carbon amendment (compost, vetch, or no
amendment) and water additions (from air dry to
flooded) affect microbial community composition. Using
canonical correspondence analysis of phospholipid fatty
acid data, we determined flooded, carbon-amended (+C)
microcosm samples were distinctly different from other
+C samples and unamended ()C) samples. Although
flooding without organic carbon addition influenced
composition some, organic carbon addition was neces-
sary to substantially alter community composition. Or-
ganic carbon availability had the same general effects on
microbial communities regardless of whether it was
compost or vetch in origin. In addition, flooded samples,
regardless of organic carbon inputs, had significantly
lower ratios of fungal to bacterial biomarkers, whereas
under drier conditions and increased organic carbon
availability the microbial communities had higher pro-
portions of fungal biomass. When comparing field and
microcosm soil, flooded +C microcosm samples were
most similar to field-collected rice soil, whereas all other
treatments were more similar to field-collected tomato
soil. Overall, manipulating water and carbon content
selected for microbial communities similar to those ob-
served when the same factors were manipulated at the
field scale.
Introduction
Our knowledge of soil microbial communities is rapidly
expanding with the explosion of new methods available
for characterizing organisms in nature [13, 14, 39]. Soil is
one of the most diverse habitats known for microor-
ganisms [38], and new taxa are discovered virtually every
time a new soil community is described [e.g., 10, 27].
Identifying the biotic and abiotic factors that deter-
mine community composition is an important subdis-
cipline of ecology. Although there is debate among
ecologists over how discrete plant community boundaries
are, there is general agreement that plant community
types can be recognized according to their composition
(e.g., coastal sage scrub, heathland, and deciduous for-
est). In addition, the distribution of these communities
across landscapes can be predicted based on environ-
mental variables, such as climate, soil type, and altitude
[3, 42], and biotic interactions such as dispersal and
competition [20, 31, 37]. Understanding the major de-
terminants of soil microbial communities, on the other
hand, has yet to be achieved. Plant community structure
is hypothesized to be a major determinant of microbial
community composition; however, this is not always
supported by data [e.g., 7]. A number of edaphic, cli-
matic, and environmental factors also have been hy-
pothesized and demonstrated to influence microbial
communities [2, 9, 19, 21, 24, 32, among others]. Dif-
ferences in taxonomic and functional diversity between
microbial communities [33] can feed back into changes
in soil and ecosystem processes [28, 40]. Insights into
microbial community composition and the factors that
determine them may improve our understanding of bi-
ogeochemical processes [15], food web dynamics [16,
Correspondence to: R.E. Drenovsky; E-mail: redrenovsky@ucdavis.edu
424 DOI: 10.1007/s00248-003-1063-2 dVolume 48, 424–430 (2004) dSpringer Science+Business Media, Inc. 2004
29], biodegradation processes [12], and overall soil
quality [9, 41].
Among the multiple factors potentially influencing
microbial community composition, available organic
carbon and soil water content are particularly important
[33]. Organic carbon availability limits microbial com-
munities in most soils [1], and additions of labile organic
material rapidly alter microbial communities by selecting
for populations that are most competitive in terms of
growth rates and ability to absorb nutrients. Soil water
content influences communities both directly and indi-
rectly through impacts on oxygen concentrations and
nutrient availability. Flooding reduces soil oxygen levels
and selects for facultative and obligate anaerobic micro-
organisms, whereas soil desiccation lowers microbial ac-
tivity, in general, and selects for fungi and spore formers
[33].
In previous field studies of California agricultural
soils, we found soil water content and organic carbon
amendments were strongly related to the microbial
community composition of lowland rice soils, as indi-
cated by phospholipid fatty acid (PLFA) profiles [5].
Using direct gradient analysis, soil water content was
found to correlate most strongly with the observed dif-
ferences in community composition, as indicated by
substantial differences between communities in flooded
versus unsaturated soils. Following 2 years of treatment,
organic carbon amendment was also significantly corre-
lated with community differences, regardless of how
plant residues were incorporated into the soil. In addi-
tion, we observed that rice soil microbial communities
were distinctly different from those in unflooded tomato
soils, despite variations in organic carbon and water in-
puts within each set of soil samples [6]. Specifically,
flooded rice soils had a greater relative abundance of
branched fatty acids and a lower relative abundance of
monounsaturated fatty acids (indicators of high substrate
availability), fungal biomarkers (in general, obligate aer-
obic organisms), and actinomycete biomarkers (typically
associated with drier soils).
These differences between rice and tomato microbial
communities led us to ask whether we could induce
similar differences within one soil type by simply ma-
nipulating soil water content and organic carbon availa-
bility under controlled, laboratory conditions. Using
microcosm experiments, we could eliminate other vari-
ables, especially biotic factors such as crop type, root
mass turnover, and root carbon exudation. This study
addresses three questions: (1) Do soil microbial com-
munities change substantially when removed from their
corresponding plant communities? (2) What is the rela-
tive importance of soil water content and organic carbon
availability in shaping soil microbial community com-
position? (3) Can microbial community composition be
predicted based upon changes in environmental varia-
bles, such as soil water content and organic carbon
availability? We addressed the first question by compar-
ing microbial community composition, using phospho-
lipid fatty acid (PLFA) analysis, in tomato soils following
collection and then after laboratory incubation for 20
days, with the expectation that removal of plant inputs
would substantially change soil microbial communities.
For the second question, we measured changes in tomato
soil microbial communities incubated at four different
moisture contents (from air dry to flooded), with or
without organic carbon amendment. We hypothesized
that flooding and organic carbon inputs would influence
microbial community composition, with flooding se-
lecting for communities with lower fungal:bacterial ratios
and organic carbon amendment increasing microbial
biomass and the prevalence of monounsaturated fatty
acids. For the third question, we compared the micro-
cosm data to field data from rice and tomato soils to
address whether soil water content and organic carbon
availability could explain the differences in microbial
community composition we observed in field studies. We
hypothesized that flooding a commonly unsaturated soil
would cause its microbial communities to take on the
traits of an agricultural soil that is saturated during part
of the year.
Methods
Experimental Conditions. The top 15 cm of Yolo silt
loam was collected during spring 1998 from the SAFS
agricultural plots at the University of California, Davis
(plots described in [11, 30, 34]). These fields had been
managed for 8 years using conventional management
practices and planted in tomatoes at the time of soil
sampling. All soils were passed through a 4-mm pore size
sieve and air dried. Triplicate 75 g (dry weight) soil mi-
crocosms were assigned randomly to a factorial combi-
nation of four water levels (air dry, half field capacity,
field capacity, and flooded) and three carbon additions
(compost, vetch, or no amendment). To each organic
carbon–amended microcosm, 0.15 g of compost or vetch
was added. Soil water content was maintained gravi-
metrically throughout the experiment. Changes in mi-
crobial community composition were assessed using
PLFA analysis. PLFAs are components of cell membranes
that are rapidly degraded following cell death [23] and so
are representative of living soil microorganisms. Micro-
cosm soil was subsampled for PLFA analyses prior to
moisture and organic carbon addition and then at 2 or 6,
7 or 11, and 16 or 20 days following water and organic
carbon additions.
PLFA Analyses. At each sampling date, 8 g (dry
weight) of microcosm soil was extracted following a
R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES 425
modified Bligh and Dyer method [4]. Polar lipids (in-
cluding phospholipids) were separated from neutral li-
pids and glycolipids using solid-phase extraction columns
(Supelco, Bellefonte, PA). Following mild alkaline met-
hanolysis of the polar lipid fraction, the resulting fatty
acid methyl esters (FAMEs) were extracted with two
aliquots of hexane. The hexane was evaporated under N
2
gas, and the FAMEs were redissolved in hexane con-
taining the internal standard 19:0. Samples were analyzed
using capillary gas chromatography, and peaks were
identified using bacterial FAME standards and MIDI
peak identification software (Microbial ID, Newark, DE).
Peak identification was verified by comparing mass
spectrometry EI spectra to spectra from standards. Mo-
lecular weights were confirmed with chemical ionization
spectra, using a Varian 3400 gas chromatograph inter-
faced with a Finnigan ITD 806 mass spectrometer.
Fatty acid nomenclature denotes the number of
carbons: number of double bonds, followed by double
bond location(s) from the methyl (x) end of the mole-
cule. For example, 16:1x5 indicates a fatty acid with 16
carbons and a double bond at carbon 5. Cis and trans
geometry are indicated by the suffixes c and t. The pre-
fixes a and i indicate anteiso and iso branching; 10Me
specifies a methyl group on the 10th carbon from the
carboxyl end of the molecule; OH indicates a hydroxyl
group; and cy indicates cyclopropane fatty acids. In
Fig. 1A, some fatty acids are indicated by ‘‘sum’’ followed
by a number. These summed features indicate two or
more fatty acids having the same retention time that
therefore cannot be resolved into individual fatty acids.
Statistical Analyses. Both correspondence analy-
sis (CA) and canonical correspondence analysis (CCA)
were used to analyze the microcosm data. CA also was
used to compare data from microcosm soils to the SAFS
tomato and commercial rice field soils. CA is an indirect
gradient analysis method; consequently, no explanatory
(environmental) variables are included in the analysis.
This method maximizes the correlation between fatty
acids and samples. Fatty acid scores and sample scores are
obtained simultaneously, allowing relationships between
treatments and fatty acid patterns to be inferred from
plots of the data. CCA can detect relationships between
environmental variables (in our experiment, soil water
content and organic carbon availability) and fatty acid
and sample patterns [17, 35]. Since fatty acid, sample,
and environmental variable scores are obtained simulta-
neously, relationships between samples, treatments, and
fatty acid patterns can be determined from biplots of the
data. On the biplots, soil water content and organic
carbon availability (labeled as ‘‘water’’ and ‘‘carbon’’ in
the ordination diagrams) are plotted as centroids (dis-
crete points plotted in the ordination diagram), indicat-
ing the quadrants most closely associated with these
variables. All multivariate statistical analyses were con-
ducted using CANOCO for Windows [36].
Univariate statistical methods were used to test
treatment differences in fatty acid loadings at the end of
the experiment. The ratio of fungal to bacterial biomass
(18:2x6,9c/(i15:0 + a15:0 + 15:0 + i16:0 + 16:1x5c +
i17:0 + a17:0 + 17:0cy + 18:1x7c + 19:0cy) [5] was an-
alyzed using analysis of variance (ANOVA). Total mi-
crobial biomass (approximated by total nanomoles of
PLFA) was analyzed by ANOVA after data were weighted
Figure 1. Results from the CCA of the microcosm PLFA and
environmental variable data. (A) Ordination biplot of the fatty
acids and environmental variable scores. Three circles were added
to this biplot following statistical analysis to aid in identifying the
plotted fatty acids. The circle farthest to the left includes the fatty
acids i15:0, 16:0, 16:1x5c, 16:1x7t, and i17:1x5. The middle circle
includes the fatty acids 16:1x11c, i17:0, 17:0cy, 18:0, and sum 7.
The circle farthest to the right includes the fatty acids 10Me 16:0,
10Me 17:0, 17:1x9c, and sum 9. (B) Ordination biplot of the
sample and the environmental variable scores. Each sample point is
the average of three treatment replicates. Black squares indicate +C
samples, and gray circles indicate )C samples. Following statistical
analysis, circles were added to the biplots to indicate treatment
groupings, but these circles do not indicate confidence ellipsoids.
In both plots (A and B) the environmental variables are plotted as
discrete points.
426 R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES
by water treatment due to nonhomogeneous variance.
Post hoc Tukey’s tests were used to determine differences
between treatment means. All univariate data were ana-
lyzed with SAS [26].
Results
Microcosm PLFA Profiles. We compared microbial
community composition by PLFA analysis in soils either
amended or not amended with compost at the four water
levels. With CA analysis 49.0% and 26.5% of the varia-
tion in the PLFA data could be explained by the first and
second axes, respectively. The sample and fatty acid re-
lationships were very similar to those detected with CCA
(see below); therefore, only CCA plots are presented.
CCA was used to relate the environmental variables
(organic carbon availability and soil water content) with
microbial community composition (Fig. 1A,B). Since
CCA is a constrained analysis and there were only two
discrete environmental variables, 100% of the fatty acid-
environment variance is explained by the first two axes
(58.1% and 41.9%, respectively). If only the percent
variation explained by the fatty acids is considered, the
first axis describes 25.7% of the variation, and the second
axis describes 18.6% of the variation. Contrary to ex-
pectations, there was not a strong effect of duration of
incubation on PLFA profiles in most treatments, indi-
cating composition changed rapidly following water and/
or organic carbon addition and then was stable over the
remaining incubation period. PLFA composition either
changed little (all )C and £field capacity samples) or if
it did change, it happened rapidly (all flooded samples,
especially +C samples). Both soil water content and or-
ganic carbon availability were significant explanatory
environmental variables (P= 0.005 for both variables), as
determined by the Monte Carlo permutation test. Or-
ganic carbon–amended samples plotted lower on the
second axis (closer to the carbon centroid), and flooded
samples (both +C and )C) plotted higher on the second
axis (closer to the water centroid). Flooding had the
strongest effect on microbial community composition
when organic carbon also was added to the samples.
Without organic carbon amendments, microbial com-
munity composition in flooded samples shifted only
slightly from that in )C, unflooded soils, whereas +C,
flooded samples plotted as a distinct group much further
from their +C, unflooded counterparts. With increased
soil water content, samples were enriched in most
straight-chain, saturated fatty acids but were less enriched
in the fungal biomarker (18:2x6,9c), which plotted in the
quadrant opposite the water centroid. Also, most bran-
ched fatty acids were more prevalent in +C samples.
Thus, flooded, +C treatments were enriched in the sat-
urated fatty acids and reduced in the fungal biomarker,
18:2x6,9c. In contrast, all other treatments were enriched
more in monounsaturated fatty acids, the fungal bio-
marker 18:2x6,9c, and 10-methylated fatty acids.
The fungal:bacterial ratio was affected by a significant
carbon by water interaction (P= 0.0001; Fig. 2A).
Overall, flooded treatments (with or without organic
carbon addition) had lower fungal:bacterial ratios than
drier soils, with the flooded, +C treatment having the
lowest proportion of fungi relative to bacteria. In the )C
treatments the fungal:bacterial ratio decreased linearly
with increasing soil water content. In contrast, in +C
samples that were not flooded (air dry, half field capacity,
and flooded) the fungal:bacterial ratio was fairly similar,
with the strongest suppression occurring in the flooded
treatment. These changes in the fungal:bacterial ratio
were driven by strong decreases in fungal biomarker
biomass, with increased soil water content strongly sup-
pressing fungi but having little effect on bacterial bio-
marker biomass (data not shown). Although there was a
significant interaction of organic carbon and water ad-
Figure 2. Fungal to bacterial ratio (A) and total nanomoles of
PLFA (nmol g
)1
DW soil) (B) in each microcosm treatment
(n= 3, bars are means ± SE). In (A), letters indicate a significant
difference between treatment means following a post hoc Tukey’s
test (a= 0.05). Although there was a significant carbon*water
interaction for total nanomoles of PLFA (P= 0.01), means were
not significantly different following a post hoc Tukey’s test.
R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES 427
dition on total nanomoles of PLFA (P= 0.01), an esti-
mate of microbial biomass, there were no significant
differences between means based on a post hoc Tukey’s
test (Fig. 2B). However, there was a trend for organic
carbon addition to increase microbial biomass, especially
in treatments with greater water availability.
To determine whether organic carbon type influ-
enced fatty acid composition under different water re-
gimes, we compared vetch-amended, compost-amended,
and unamended samples incubated under the four water
levels. When these samples were analyzed with CA (data
not shown), vetch-amended and compost-amended soils
had very similar microbial communities which diverged
strongly from those of the unamended soils. Thus, re-
gardless of organic carbon type, flooded samples were
similar in fatty acid composition (less enriched in
18:2x6,9c, higher in saturated fatty acids) and were most
dissimilar from other samples.
Tomato, Rice, and Microcosm PLFA Profiles. Our
third objective was to determine whether imposing or-
ganic carbon and water additions on tomato soil could
shift its microbial community to reflect that of rice soils
(i.e., that responses of the microbial community to water
and organic carbon addition are common across soil
types). To facilitate this comparison, we combined our
microcosm data with data previously collected in tomato
and rice fields [6] and conducted a CA (Fig. 3A,B). To-
gether, the first two axes describe more than 75% of the
variation in fatty acid composition, with 65.9% and 9.6%
of the variation described by the first and second axes,
respectively. As when the microcosm data were analyzed
separately, the flooded, +C microcosm treatments
grouped apart from all other microcosm samples. These
samples were most similar to the rice-field soil PLFA
profiles. In contrast, the remaining microcosm samples
were more similar to the tomato-field soils than to the
rice-field soils. Rice-field soils and flooded, +C micro-
cosm samples were more strongly enriched in saturated
fatty acids and reduced in the fungal biomarker fatty acid
(18:2x6,9c). In contrast, the remainder of the microcosm
samples and the tomato-field soils were more strongly
enriched in the fungal biomarker fatty acid and mono-
unsaturated fatty acids.
Discussion
Our microcosm results supported that soil water content
and organic carbon availability (but not carbon type) are
major determinants of microbial community composi-
tion, at least in Yolo soil. The effect of soil water content
on microbial community composition was most pro-
nounced in the flooded treatments, as might be expected
since the prevailing electron acceptors strongly influence
microbial community composition [33]. Although
flooding changed microbial community composition in
all treatments along the same trajectory, the response was
most pronounced in the organic carbon–amended sam-
ples. Without organic carbon inputs, the microbial
community most likely lacked sufficient organic carbon
and energy sources to grow and ‘‘replace’’ itself with a
new community. Another potential impact of carbon
additions may have been development of anoxic condi-
tions in carbon-amended microcosms, in contrast to
unamended microcosms that more likely remain aerobic
in the absence of a strong oxygen demand.
Removing soils from the field, and away from
growing plant inputs, had little effect on microbial
community composition within the incubation period, as
evidenced by minimal changes in composition during
Figure 3. Results from the CA of microcosm and field PLFA data.
(A) Ordination plot of fatty acid scores. As in Fig. 1A two circles
were added to the plot following analysis to aid in identification of
the plotted fatty acids. The circle at the origin contains the fatty
acids 16:1 2OH, 10Me 16:0, i17:0, and 10Me 17:0. The circle in the
top left quadrant includes the fatty acids a15:0, i15:0, and 16:0. (B)
Ordination plot of sample scores from the same analysis. Black
circles indicate the tomato-field data, gray circles indicate the rice-
field data, and white squares indicate the microcosm data. As in
Fig. 1B, circles were added to the plot following statistical analysis
to indicate treatment groupings, but these circles do not indicate
confidence ellipsoids.
428 R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES
incubation of unflooded treatments. These treatments
also were relatively similar in fatty acid composition to
tomato soils analyzed immediately after collection from
the field. Thus, disturbance of the soil during sample
preparation, including sieving, and incubation under
laboratory conditions did not substantially alter the mi-
crobial communities. This preservation of the microbial
community present in the field suggests that many mi-
croorganisms may be protected within small soil aggre-
gates during sample processing and do not require
constant plant inputs to maintain their presence. There is
the possibility that more substantial changes in microbial
community composition could have occurred with in-
cubation times longer than the 20 days investigated in
our study.
The flooded samples had increased levels of straight-
chain fatty acids (indicative of bacterial biomass) and
decreased levels of the fatty acid 18:2x6,9c (often con-
sidered a fungal biomarker), similar to what was observed
in flooded rice soils in the field [5, 25]. What was sur-
prising, however, was that at moisture contents £field
capacity, the existing community changed little, possibly
reflecting the communities’apparent adaptation to low-
moisture regimes. California agricultural fields are sub-
jected to extreme wet/dry cycles throughout the growing
season, and research suggests surface microorganisms can
adapt readily to these changes [18]. In addition, respi-
ration data suggest growth occurred in both the half-
field-capacity and field-capacity samples but not in the
air-dry samples (D. Vo, unpublished data). Together, the
lack of change in fatty acid composition and the respi-
ration data imply that the original members of the soil
community grew in similar relative proportions
throughout the study in the 50% and 100% field capacity
samples. In the air-dry samples, these data indicate that
the original populations lived but did not actively grow
during the incubation period.
Carbon addition to soil has altered specific microbial
fatty acids in several studies. In rice soils organic carbon
addition increased monounsaturated fatty acids (pur-
ported indicators of high substrate availability) relative to
other fatty acids [5]. Manure additions also enriched
monounsaturated fatty acids in a Tennessee no-till agri-
cultural soil [22]. In our study, however, there was no
relationship between organic carbon inputs and mo-
nounsaturates. What we did find, however, was that +C
samples had decreased levels of 10-methylated fatty acids
(often considered as biomarkers for actinomycetes) and
cyclopropyl fatty acids. Similar reductions in 10-meth-
ylated fatty acids also were measured in sucrose-amended
subarctic heath soils [29] and the manure-amended ag-
ricultural soil referred to above [22]. Although low car-
bon availability was associated with high relative
proportions of branched fatty acids in rice soils [5], we
observed the opposite trend in our microcosm soils.
Microbial community responses to organic carbon
amendment were not influenced by carbon type. The
same general effects on microbial communities were as-
sociated with both compost, which consisted largely of
poultry manure and straw, and vetch, a leguminous cover
crop. The C:N ratio of the compost and vetch were
comparable (18.7, compost; 15.0, vetch) [11], and thus,
nutritionally, the two organic carbon sources may not
have been very different for microbial populations. It is
possible that more divergent organic carbon sources may
lead to greater differences in community composition.
However, our findings were consistent with a study of
farming systems in Pennsylvania in which management
history (conventional versus organic) was more impor-
tant than type of crop residue (vetch versus corn and rye)
in influencing microbial community composition, as
measured by total soil fatty acid methyl esters [8]. The
C:N ratios of the two inputs in this study differed sub-
stantially, yet there was still only a minor effect on com-
munity composition, leading the authors to conclude that
the overriding factor in determining community com-
position was site history rather than amendment origin.
In conclusion, we found that simple manipulation of
organic carbon inputs and soil water content caused
microbial communities of a typically unsaturated soil to
adopt some of the characteristics of a periodically flooded
agricultural soil, in a manner predicted from field ob-
servations. These changes occurred rapidly without the
influence of other biotic factors and were not related to
carbon type. When considering hypotheses linking mi-
crobial functions to microbial community composition,
one cannot ignore the strong effects of abiotic factors in
shaping microbial community composition.
Acknowledgments
The authors thank the U.S. Environmental Protection
Agency’s Center for Ecological Health Research at U.C.
Davis, the NIEHS Superfund Basic Research Program
(2P42 ESO4699), and a grant from the Kearney Foun-
dation of Soil Science. The comments of K.M. Batten,
K.A. Hicks, and three anonymous reviewers significantly
improved the manuscript.
References
1. Alden, L, Demoling, F, Ba
˚a
˚th, E (2001) Rapid method of deter-
mining factors limiting bacterial growth in soil. Appl Environ
Microbiol 67: 1830–1838
2. Ba
˚a
˚th, E, Anderson, TH (2003) Comparison of soil fungal/bacterial
ratios in a pH gradient using physiological and PLFA-based
techniques. Soil Biol Biochem 35: 955–963
3. Bedford, BL, Walbridge, MR, Aldous, A (1999) Patterns in nutrient
availability and plant diversity of temperate North American
wetlands. Ecology 80: 2151–2169
4. Bligh, EG, Dyer, WM (1959) A rapid method of lipid extraction
and purification. Can J Biochem Phys 39: 911–917
R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES 429
5. Bossio, DA, Scow, KM (1998) Impacts of carbon and flooding on
soil microbial communities: phospholipids fatty acid profiles and
substrate utilization patterns. Microb Ecol 35: 265–278
6. Bossio, DA, Scow, KM, Gunapala, N, Graham, KJ (1998) Deter-
minants of soil microbial communities: effects of agricultural
management, season, and soil type on phospholipids fatty acid
profiles. Microb Ecol 36: 1–12
7. Broughton, LC, Gross, KL (2000) Patterns of diversity in plant and
soil microbial communities along a productivity gradient in a
Michigan old-field. Oecologia 125: 420–427
8. Buyer, JS, Drinkwater, LE (1997) Comparison of substrate utili-
zation assay and fatty acid analysis of soil microbial communities. J
Microbiol Methods 30: 3–11
9. Calderon, FJ, Jackson, LE, Scow, KM, Rolston, DE (2001) Short-
term dynamics of nitrogen, microbial activity, and phospholipid
fatty acids after tillage. Soil Sci Soc Am J 65: 118–126
10. Dunbar, J, Barns, SM, Ticknor, LO, Kuske, CR (2002) Empirical
and theoretical bacterial diversity in four Arizona soils. Appl En-
viron Microbiol 68: 3035–3045
11. Gunapala, N, Scow, KM (1998) Dynamics of soil microbial bio-
mass and activity in conventional and organic farming systems.
Soil Biol Biochem 30: 805–816
12. Hanson, J, Macalady, JL, Harris, D, Scow, KM (1999) Linking
toluene degradation with specific microbial populations in soil.
Appl Environ Microbiol 65: 5403–5408
13. Hedrick, DB, Peacock, A, Stephen, JR, Macnaughton, SJ, Brugg-
emann, J, White, DC (2000) Measuring soil microbial community
diversity using polar lipid fatty acid and denaturing gradient gel
electrophoresis data. J Microb Methods 41: 235–248
14. Hill, GT, Mitkowski, NA, Aldrich-Wolfe, L, Emele, LR, Jurkonie,
DD, Ficke, A, Maldonado-Ramirez, S, Lynch, ST, Nelson, EB
(2000) Methods for assessing the composition and diversity of soil
microbial communities. Appl Soil Ecol 15: 25–36
15. Keith-Roach, MJ, Bryan, ND, Bardgett, RD, Livens, FR (2002)
Seasonal changes in the microbial community of a salt marsh,
measured by phospholipid fatty acid analysis. Biogeochemistry 60:
77–96
16. Laakso, J, Seta
¨la
¨, H, Paloja
¨rvi, A (2000) Influence of decomposer
food web structure and nitrogen availability on plant growth. Plant
Soil 225: 153–165
17. Lindstrom, ES (2000) Bacterioplankton community composition
in five lakes differing in trophic status and humic content. Microb
Ecol 40: 104–113
18. Lundquist, EJ, Scow, KM, Jackson, LE, Uesugi, SL, Johnson, CR
(1999) Rapid response of soil microbial communities from con-
ventional, low input, and organic farming systems to a wet/dry
cycle. Soil Biol Biochem 31: 1661–1675
19. Macalady, JL, Mack, EE, Nelson, DC, Scow, KM (2000) Sediment
microbial community structure and mercury methylation in
mercury-polluted Clear Lake, California. Appl Env Microbiol 66:
1479–1488
20. Nilsson, C, Andersson, E, Merritt, DM, Johansson, ME (2002)
Differences in riparian flora between riverbanks and river lake-
shores explained by dispersal traits. Ecology 83: 2878–2887
21. Pankhurst, CE, Yu, S, Hawke, BG, Harch, BD (2001) Capacity of
fatty acid profiles and substrate utilization patterns to describe
differences in soil microbial communities associated with increased
salinity or alkalinity at three locations in South Australia. Biol
Fertil Soils 33: 204–217
22. Peacock, AD, Mullen, MD, Ringelberg, DB, Tyler, DD, Hedrick,
DB, Gale, PM, White, DC (2001) Soil microbial community re-
sponses to dairy manure or ammonium nitrate applications. Soil
Biol Biochem 33: 1011–1019
23. Pinkart, HC, Ringelberg, DB, Piceno, YM, Macnaughton, SJ,
White, DC (2002) Biochemical approaches to biomass measure-
ments and community structure analysis. In: Hurst, CJ, Crawford,
RL, Knudsen, GR, McInerney, MJ, Stetzenbach, LD (Eds.) Manual
of Environmental Microbiology, 2nd ed., ASM Press, Washington,
DC, pp 101–113
24. Powell, SM, Bowman, JP, Snape, I, Stark, JS (2003) Microbial
community variation in pristine and polluted nearshore Antarctic
sediments. FEMS Microbiol Ecol 45: 135–145
25. Reichardt, W, Mascarin
˜a, G, Padre, B, Doll, J (1997) Microbial
communities of continuously cropped, irrigated rice fields. Appl
Environ Microb 63: 233–238
26. SAS Institute (2001) SAS/STAT User’s Guide. Versions 8. SAS
Institute, Cary, North Carolina
27. Schadt, CW, Martin, AP, Lipson, DA, Schmidt, SK (2003) Seasonal
dynamics of previously unknown fungal lineages in tundra soils.
Science 301: 1359–1361
28. Schimel, J (1995) Ecosystem consequences of microbial diversity
and community structure. In: Chapin, FS, Ko
¨rner, C. (Eds.) Arctic
and Alpine Diversity: Patterns, Causes, and Ecosystem Conse-
quences, Springer-Verlag, Berlin, pp 239–254
29. Schmidt, IK, Ruess, L, Ba
˚a
˚th, E, Michelson, A, Ekelund, F, Jo-
nasson, S (2000) Long-term manipulation of the microbes and
microfauna of two subartic heaths by addition of fungicide, bac-
tericide, carbon, and fertilizer. Soil Biol Biochem 32: 707–720
30. Scow, KM, Somasco, O, Gunapala, N, Lau, S, Venette, R, Ferris, H,
Miller, R, Shennan, C (1994) Changes in soil fertility and biology
during the transition from conventional to low-input and organic
cropping systems. Cal Agr 48: 20–26
31. Smith, MD, Knapp, AK (2001) Size of the local species pool
determines invasibility of a C
4
-dominated grassland. Oikos 92:
55–61
32. Steenwerth, KL, Jackson, LE, Calderon, FJ, Stromberg, MR, Scow,
KM (2002) Soil microbial community composition and land use
history in cultivated and grassland ecosystems of coastal California.
Soil Biol Biochem 34: 1599–1611
33. Sylvia, DM, Fuhrmann, JJ, Hartel, PG, Zuberer, DA (1999) Prin-
ciples and Applications of Soil Microbiology. Prentice Hall, Upper
Saddle River, NJ
34. Temple, SR, Friedman, DB, Somasco, O, Ferris, H, Scow, KM,
Klonsy, K (1994) An interdisciplinary, experiment station-based
participatory comparison of alternative crop management systems
for California’s Sacramento Valley. Am J Alternative Agr 9: 64–71
35. ter Braak, CJF (1986) Canonical correspondence analysis: a new
eigenvector technique for multivariate direct gradient analysis.
Ecology 67: 1167–1179
36. ter Braak, CJF, S
ˇmilauer, P (1998) CANOCO Reference Manual
and Users Guide to Canoco for Windows: Software for Canonical
Community Ordination. Version 4. Microcomputer Power, Ithaca,
NY
37. Theodose, TA, Bowman, WD (1997) The influence of interspecific
competition on the distribution of an Alpine graminoid: evidence
for the importance of plant competition in an extreme environ-
ment. Oikos 79: 101–114
38. Torsvik, V, Goksøyr, J, Daae, FL (1990) High diversity in DNA of
soil bacteria. Appl Environ Microbiol 56: 782–787
39. Torsvik, V, Øvrea
˚s, L (2002) Microbial diversity and function
in soil: from genes to ecosystems. Curr Opin Microbiol 5: 240–
245
40. Waldrop, MP, Balser, TC, Firestone, MK (2000) Linking microbial
composition to function in a tropical soil. Soil Biol Biochem 32:
1837–1846
41. Yao, H, He, Z, Wilson, MJ, Campbell, CD (2000) Microbial bio-
mass and community structure in a sequence of soils with in-
creasing fertility and changing land use. Microb Ecol 40: 223–237
42. Young, TP, Peacock, MM (1992) Giant Senecios and alpine veg-
etation of Mount Kenya. J Ecol 80: 141–148
430 R.E. DRENOVSKY ET AL.: WATER AND CARBON INFLUENCE MICROBIAL COMMUNITIES
