PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

OCH-"0 9 “ta-7
" V ‘1'“ xi Qua-

 

SEP 2 2 20.7

 

 

EBUIZUU
5

no
'17

A

 

I‘D

A r
7 T

Qp‘fQ 3 '0 7

 

 

 

 

 

 

 

 

 

 

 

6’01 cJCIRC/DateDuepBS—p. 15

 

LINKING PLANT COMMUNITIES TO SOIL MICROBIAL COMMUNITIES
AND PROCESSES IN OLD-FIELDS

By

Laura C. Broughton

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Zoology
Ecology, Evolutionary Biology, and Behavior Program

2001

ABSTRACT

LINKING PLANT COMMUNITIES TO SOIL MICROBIAL COMMUNITIES
AND PROCESSES IN OLD-FIELDS

By

Laura C. Broughton

The resources that support soil microbial communities are primarily derived from
plants, so the soil microbial community should respond to changes in plant diversity
or productivity, particularly if changes in the plant community affects the quality or
quantity of available resources. I investigated the role of soil and plants on the
structure and function of the soil microbial community by conducting observational

and experimental studies and two manipulative greenhouse experiments.

I examined the relationship between plant diversity and productivity and soil
microbial community structure and function along a topographic gradient in a
successional old—ﬁeld in Michigan. Variation in plant productivity was confounded
by changes in plant community diversity and edaphic characteristics, so I could not
determine which of these variables caused the observed changes in the soil microbial

community.

To further investigate the relationship between the soil microbial community and
plant species diversity, I sampled soils from a set of experimental grassland plant
communities established as part of the BIODEPTH experiment at Silwood Park,
England. Plant species diversity, functional group diversity, and species composition

varied across treatments. I found that plant diversity signiﬁcantly affected soil

microbial community structure. However, N-mineralization rates and microbial

respiration responded to variation in plant community composition, but not diversity.

In a greenhouse experiment I examined how variation in soil fertility inﬂuenced the
soil microbial community. I found that soil origin had strong effects on the structure
and function of the soil microbial community. Higher fertility soils had higher
organic nitrogen pools and microbial activities and more eukaryotes in the microbial
community. In addition, the presence of Andropogon gerardi also affected the
structure and function of the soil microbial community. However, the magnitude of

the plant effect on soil microbial respiration was inversely related to soil fertility.

In a second greenhouse experiment I further explored the plant species effect on the
soil microbial community. I found strong effects of both plant species identity and
soil origin on the structure and function of the soil microbial community. In
particular, the presence of a legume (T rifolium pratense) increased soil nitrogen
cycling processes. Plant species identity had a small effect on soil microbial

community structure, but it was dwarfed by the soil origin effect.

Results from these studies indicate that several aspects of the plant community,
including diversity, composition and individual plant species identities, can strongly
inﬂuence the structure and function of the soil microbial community. However, other
environmental factors that affect soil quality can have strong and persistant effects on

the soil microbial community.

ACKNOWLEDGMENTS

I would like to thank my advisor, Kay Gross, for her unlimited support, patience, and
guidance over the last six years. She has been particularly understanding this past year
as I have faced health problems that interfered with my work. I will be forever grateful
to Kay. I would also like to thank the other members of my committee, Steve
Hamilton, Mike Klug, and Gary Mittelbach for their guidance and help in shaping the

path of my graduate career.

This work could not have been completed without the help of the members of Kay
Gross’ lab. I would particularly like to thank Carol Baker, Karen Tindall, and Jaime
Fiordalis for their technical assistance and help with ﬁeld work; and Bryan Foster,
Andrea Corbett, Melissa McCormack, Heather Reynolds, Wendy Goodﬁiend, Greg
Houseman, Rich Smith, and Sarah Emery for their intellectual feedback and

companionship.

I am grateful to Michel Cavigelli, Mike Kaufman, Christine Easley, and Terry Marsh
for help with laboratory techniques. The statistical work could not have been done
without timely help from Carl Ramm, Chris Blackwood, and Hal Collins. Additionally,
the BIODEPTH portion of my work could not have been completed without the help of
my collaborator, Andy Hector, and John Lawton, Asher Minns, and Ellen Bazeley-

White. Special thanks to participants in the summer 1996 Advanced Field Ecology and

iv

Evolution course at the W. K. Kellogg Biological Station who helped with the initial

soil and plant sampling for the ﬁeld survey at Lux Arbor.

Living and working at the Kellogg Biological Station can be a lonely endeavor. For
their fellowship, dinner groups, game nights, and great parties, I am grateful to my
fellow graduates students, particularly Eric Thobaben, Tara Darcy, and Natalie Dubois.
I enjoyed singing with Eric in the Bach Festival Chorus and talking about the trivial but
fun aspects of our lives. Stefanie Whitmire has been my constant companion for the
past four years and I could not have made it through without her. I am eternally

grateful to Stefanie.

For their encouragement, love, and support throughout my life, I would like to thank
my parents, Sharon and Gary Broughton. I could never have made it this far without
them. My brother, Keith Broughton, has also shown constant support for my chosen

path.

The research was funded by NSF grants to the Center for Microbial Ecology at
Michigan State University, a Research Training Grant at the W. K. Kellogg Biological
Station, and a Dissertation Completion Grant from the College of Natural Science at
Michigan State University. I obtained additional monies from the Michigan State
University Department of Zoology, the Michigan State University Ecology,

Evolutionary Biology, and Behavior Program, and Sigma Xi.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. ix
LIST OF FIGURES ............................................................................. xii
INTRODUCTION .............................................................................. 1
THESIS OVERVIEW ................................................................. 2
FIELD SITES ........................................................................... 6
REFERENCES ......................................................................... 9

CHAPTER 2: PATTERNS OF DIVERSITY IN PLANT AND SOIL
MICROBIAL COMMUNITIES ALONG A PRODUCTIVITY GRADIENT

IN A MICHIGAN OLD FIELD ............................................................... 10
INTRODUCTION ..................................................................... 10
METHODS .............................................................................. 12

Site Description ................................................................ 12
Sampling Design and Characterization of the Gradient .................. 12
Characterization of the Soil Microbial Community ....................... 14
Statistical Analyses ............................................................ 16
RESULTS ............................................................................... l7
Characterization of the Gradient ............................................. 17

Plant and Soil Microbial Community Relationships with
Productivity ............................................................ 1 9
Compositional Shifts in the Plant and Soil Microbial Communities... 21
DISCUSSION .......................................................................... 26
REFERENCES ......................................................................... 34

CHAPTER 3: LINKING PLANT COMMUNITY DIVERSITY TO SOIL
MICROBIAL COMMUNITIES: AN EXPERIMENTAL EVALUATION FROM

THE BIODEPTH EXPERIMENT ............................................................ 37
INTRODUCTION ..................................................................... 37
MATERIALS AND METHODS .................................................... 39

Site .............................................................................. 39
Soil Sampling and Analysis .................................................. 40
Statistical Analysis ............................................................ 46
RESULTS ............................................................................... 48
Legacies of Site Preparation ................................................. 48
Effects of Species Richness and Functional Group Richness
Treatments on the Soil Microbial Community ................... 48
Effects of Functional Groups or Individual Species on the Soil
Microbial Community ............................................... 53

vi

Effects of Plant Community Composition (Mixture) on the Soil

Microbial Community ............................................... 54
Changes in Soil Microbial Community Composition .................... 54
DISCUSSION .......................................................................... 58
Plant Diversity Effects on the Soil Microbial Community ............. 59

Plant Community Composition Effects on the Soil Microbial
Community ......................................................... 61
Legacy Effects on the Soil Microbial Community ....................... 63
REFERENCES ......................................................................... 66

CHAPTER 4: PLANT-MEDIATED EFFECTS OF SOIL ORIGIN ON THE

COMPOSITION AND FUNCTION OF SOIL MICROBIAL COMMUNITIES ...... 69
INTRODUCTION ..................................................................... 69
METHODS .............................................................................. 71
Site Descriptions ............................................................... 71
Experimental Design .......................................................... 72
Data Collection ................................................................ 74
Statistical Analyses ............................................................ 77
RESULTS ............................................................................... 79
Plant Communities ............................................................ 79
Effects of Soil Origin on Soil and Microbial Processes .................. 80
Effects of Andropogon gerardi on Soil and Microbial Processes ...... 83
Plant and Soil Effects on Soil and Microbial Processes .................. 85
Plant and Soil Effects on Soil Microbial Community Structure. . . . . 87
DISCUSSION ........................................................................... 95
Site Fertility Effects on Soil Microbial Processes ........................ 96
Historical Plant Diversity Effects on Soil Microbial Processes ......... 97
The “Plant” Effect on Soil Microbial Processes .......................... 100
Plant and Soil Effects on Soil Microbial Community Structure ........ 102
REFERENCES ......................................................................... 104

CHAPTER 5: AN EXPERIMENTAL EVALUATION OF THE EFFECTS OF
DIFFERENT PLANT SPECIES ON THE STRUCTURE AND FUNCTION OF

SOIL MICROBIAL COMMUNITIES ....................................................... 108
INTRODUCTION ..................................................................... 108
METHODS .............................................................................. 110

Site and Species Descriptions ............................................... 110
Experimental Design .......................................................... 1 12
Data Collection ................................................................ 1 15
Statistical Analyses ............................................................ 1 16
RESULTS ............................................................................... 118
Plant Communities ............................................................ l 18
Time Effects on Soil and Microbial Processes ............................ 119
Plant Species Responses to Soil Types ..................................... 119

vii

Effects of Soil Origin on Soil and Microbial Processes .................. 125

Plant Effects on Soil and Microbial Processes ............................ 128
Interactive Effects of Plant Species and Soil Origin on Soil and
Microbial Processes .................................................. 129

Plant and Soil Effects on Soil Microbial Community Structure. . . 130
A Comparison of the Microbial Communities within and outside the

Root Exclosures ..................................................... 135
DISCUSSION ........................................................................... 138

The Effects of Soil Origin on Soil and Microbial Processes and Soil
Microbial Community Structure .................................... 139

Plant Effects on Soil and Microbial Processes and Soil Microbial

Community Structure ................................................ 140

Implications of the Similarity of Soil Microbial Communities in the
Root Exclosures and the Rhizosphere Soil ........................ 145
REFERENCES ......................................................................... 147
CHAPTER 6: SUMMARY .................................................................... 151

viii

LIST OF TABLES

Table 1.1. Plant productivity and diversity of abandoned old ﬁelds from which soil was
collected for the greenhouse experiments. Values for peak aboveground plant
biomass (an estimate of primary productivity), species richness, and mean
percent organic matter are expressed as mean 1" standard deviations (n = 6).
Values that are signiﬁcantly different for a given variable based on Fisher’s
LSD test are indicated by different letters .............................................. 7

Table 2.1. Plant species composition in the clusters from the K—means cluster analysis
shown in Figure 2.3B. Plant species are listed from most common (by biomass)
to least common. Only species that have a total biomass > 1 g for the cluster are
listed. Nomenclature follows Gleason and Cronquist (1991) ..................... 25

Table 2.2. Fatty acids used in K—means cluster analysis of FAME proﬁles .............. 27

Table 3.1. (A) Plant species pool used in establishing the experimental communities
and (B) the plant communities sampled for the soil microbial community. Plots
sampled only in year 3 (1998) of the experiment are in italics. The control and
Rumex plots were not maintained through year 4 and could not be re-sampled.
Plots sampled in both years are indicated in bold. Other plots were sampled
only in year 4 (1999) of the experiment. Communities are grouped by number
of plant functional groups (FG) present ............................................... 41

Table 3.2. Analysis of variance model with sequential sums of squares used to evaluate
the effects of the plant community manipulations on soil and microbial
community parameters .................................................................. 47

Table 3.3. Signiﬁcance and direction (for diversity) of treatment effects on soil and
microbial parameters for both years as detected by Analysis of Variance using
Type I sums of squares (model in Table 3.2) for (a) species diversity, functional
group diversity, and composition and (b) presence/absence of legumes or forbs.
ANCOVA results using above-ground plant biomass as the covariate are shown
in parentheses if the effect changed in signiﬁcance. NS=not signiﬁcant. . . . . ...50

Table 3.4. Fatty acids used in principal components analysis of PLF A proﬁles ........ 51

Table 4.1. Plant productivity and diversity of abandoned ﬁelds from which soil was
collected for the greenhouse experiment. Values for Annual Net Primary
Productivity, species richness, and mean percent organic matter are expressed as
mean i standard deviations. Values that are not signiﬁcantly different for a
given variable based on Fisher’s LSD test have the same letter ................... 80

Table 4.2. Effect of soil origin and presence of A. gerardi on (A) plant and (B) soil and
microbial variables as detected by Analysis of Variance. NS = not signiﬁcant, p

ix

> 0.05. The time factor investigates the results of harvesting half the
experiment at 12 weeks, the other half at 16 weeks ................................. 81

Table 4.3. Effect of soil origin and presence of A. gerardi on the soil microbial
community as detected by Distance-Based Redundancy Analysis (db-RDA) of
CLPP proﬁles. (A) Signiﬁcance values for the permutation tests on the
environmental factors of the RDA. (B) Variance explained by species data and
species-environment correlations for the RDA ....................................... 88

Table 4.4. Fatty acids used in principal components analysis of PLFA proﬁles. I
describe fatty acids using standard nomenclature where the total number of
carbon atoms appears before the colon and the total number of C-C bonds
appears after it. Cyclo-propane analogs are indicated by "cyclo," and the
location of the epoxy bond is indicated by a "c" followed by two numbers. If
the cis or trans conﬁguration is unknown, the word "at" is used. The number
following "cis," "trans," or "at" indicates the location of the double bond in
relation to the carboxyl end of the molecule. Fatty acids with the same retention
time are grouped as "sum in feature" and given a unique number designation..91

Table 4.5. Effect of soil origin, time of harvest, and presence of A. gerardi on the soil
microbial community as detected by Distance-Based Redundancy Analysis of
PLF A proﬁles. (A) Signiﬁcance values for the permutation tests on the
environmental factors of the RDA. (B) Variance explained by species data and
species-environment correlations for the RDA ....................................... 92

Table 5.1. Plant species by functional group and aboveground biomass at the two ﬁeld
sites from which soil was collected. Functional groups are coded by Grass (G),
Forb (F), Legume (L), and Woody (W). Plant species used in this greenhouse
experiment are in bold ................................................................. 111

Table 5.2. Treatments used in the Randomized Complete Block Design to test for the
effects of plant species on the soil microbial community. Each treatment was
replicated three times within three Time blocks on two different soils. Species
planted were G = Bromus inermis, F = Solidago canadensis, L = T rifolium
pratense. 0 = absent, + = present, C = control (no plants) ........................ 114

Table 5.3. Summary of ANOVA results analyzing effect of soil origin and (A)
individual plant treatments, (B) individual plant treatments without no-plant
controls, or (C) factorial plant treatments on plant and soil and microbial
variables. Species codes in 3C are: BROIN, Bromus inermis; SOOCA,
Solidago canadensis; TRFPR, T rifolium pratense. NS = not signiﬁcant, p >
0.05. ANCOVA results, using total plant biomass as the covariate, are shown in
parentheses if the effect changed in signiﬁcance ................................... 122

Table 5.4. Root to shoot ratios for each plant species in each soil and treatment. Values
are mean 3: standard error. Signiﬁcance values from the ANOVA are listed in
Table 5.3C ............................................................................... 126

Table 5.5. Fatty acids used in principal components analysis of PLFA proﬁles. I
describe fatty acids using standard nomenclature where the total number of
carbon atoms appears before the colon and the total number of C-C bonds
appears after it. Cyclo-propane analogs are indicated by "cyclo," and the
location of the epoxy bond is indicated by a "c" followed by two numbers. If
the cis or trans conﬁguration is unknown, the word "at" is used. The number
following "cis," "trans," or "at" indicates the location of the double bond in
relation to the carboxyl end of the molecule. Fatty acids with the same retention
time are grouped as "sum in feature" and given a unique number
designation ............................................................................... 13 1

Table 5.6. Effect of soil origin, time, replicate, and presence of Bromus inermz's
(BROIN), Solidago canadensz's (SOOCA), and T rifolium pratense (TRFPR) on
the soil microbial community as detected by Distance-Based Redundancy
Analysis (db-RDA) of PLFA proﬁles. (A) Signiﬁcance values for the
permutation tests on the environmental factors of the RDA. (B) Variance
explained by species data and species-environment correlations for the
RDA ...................................................................................... 134

Table 5.7. Effect of exclosure, soil origin, time, and presence of Bromus inermis
(BROIN), Solidago canadensis (SOOCA), and T rifolium pratense (TRFPR) on
the soil microbial community as detected by Distance-Based Redundancy
Analysis (db-RDA) of PLFA proﬁles. (A) Signiﬁcance values for the
permutation tests on the environmental factors of the RDA. (B) Variance

explained by species data and species-environment correlations for the
RDA ...................................................................................... 137

xi

Figure

LIST OF FIGURES

1.1. Map of the W. K. Kellogg Biological Station with codes illustrating the
location of the six ﬁeld sites (after Foster 1996). Table 1.1 describes these
sites .......................................................................................... 8

Figure 2.1. (A) Factor loadings plot for PC 1 and 2 from the PCA of the June 1996

sampling for light at ground level, soil moisture, soil inorganic N, and peak
plant biomass; PC 1 and 2 account for 64.8% and 17.1% of the variation,
respectively. Correlations between factor 1 and light (r = -0.77), moisture (r =
0.73), N (r = 0.80), and biomass (r = 0.91) were signiﬁcant at p < 0.001;
correlations between factor 2 and light (r = 0.56) and moisture (r = 0.56) were
signiﬁcant at p < 0.005. (B) The relationship between PC 1 and distance down
slope for 34 of the 35 sampling points, y = -1.69 + 0.058x; Lines depict
regression and 95% conﬁdence intervals .............................................. 18

Figure 2.2. The relationship between productivity index and several variates describing

Figure

the plant and microbial communities at this site. (A) Plant species richness, y =
10.18 - 0.71x; (B) The number of carbon sources metabolized by the soil
microbes after 48 hours; (C) Biolog average well color development after 48
hours, y = 0.89 + 0.14x; (D)ug COz-C/ g dry soil/ hour evolved from SIR
control after 40 hours, y = 0.54x — 1.55; (E) ug COZ-C/ g dry soil/ hour evolved
from the SIR glucose addition treatment after 40 hours, y = 0.33x — 2.80. Lines
depict regression and 95% conﬁdence intervals, 11 = 34 ............................ 20

2.3. Results from K-means cluster analysis evaluating the changes in
production and composition of the plant and soil communities in relation to
transect position at this site. (A) Contour plot of above-ground plant biomass,
values are g/ m2; (B) Plant species composition: cluster 1 has no dominant
species, cluster 2 is dominated by Rubus sp., cluster 3 by Solidago canadensis,
cluster 4 by S. canadensis and Poa pratense, cluster 5 by Agropyron repens,
cluster 6 by Poa pratense, and cluster 7 by Poklfvgonum amphibium var.
emersum (see Table 1 for species lists). (C) BiologT carbon source utilization
proﬁles separated into two clusters based on AWCD; and (D) FAME proﬁles:
cluster 2 had smaller proportions of 18:1 cis 9, 16:0, and summed in feature 9
(18:2 cis 9, 12 and 18:0 anteiso, Table 2) than cluster 1 ........................... 23

Figure 3.1. Plant community composition effects on soil microbial biomass in (A) year

3 and (B) year 4 of the Silwood Park BIODEPTH study. Values are Mean :1: 1
SE. Plant community identiﬁcation codes are from Table 3.1B. No-plant
control plots are represented by a blank circle, while non-manipulated reference
communities are represented by stars. Monocultures are indicated by circles

and mixtures by triangles. Plots that have Hypochaeris radicata have hatch-
marks ...................................................................................... 49

xii

Figure

Figure

3.2. Relationship between number of culturable soil bacteria and (A) plant
species richness (R2=0.294, p<0.01), (B) number of plant functional groups
(R2=0.288, p<0.01), and (C) aboveground plant biomass in year 3 (R2=0.573,
p<0.01). No-plant control plots are represented by blank circles, while non-
manipulated reference communities are represented by stars. Monocultures are
indicated by circles and mixtures by triangles. Plots that have Lotus
corniculatus have hatch-marks slanting up to the right; plots that have
Hypochaeris radicata have hatch-marks slanting down to the right. The two
plots with both species are mottled .................................................... 52

3.3. Plant diversity effects on the metabolic activity of the soil microbial
community as measured by Community Level Physiological Proﬁles (CLPP) in
year 3 (n = 28). Principal component axis 1 accounted for 46.1% of the total
variation. Principal component axis 3 accounted for 7.6% of the overall
variation and was driven by the ability to metabolize alpha-D-lactose.
Signiﬁcance values for diversity effects are listed in Table 3.3A. Number of
functional groups is represented by symbol shape. No-plant control plots are
represented by circles, and non-manipulated control communities labeled with
“R”. Number of plant species (0, l, 4, 8, or 11) are labeled ........................ 55

Figure 3.4. Plant community composition effects on the structure of the soil microbial

community, measured with Phospholipid Fatty Acid proﬁles of 1998 soil
samples. Principal component axis 4 accounted for 4.5% of the overall variation
and reﬂected the amount of 15:0 anteiso in the PLFA proﬁles. Plant community
identiﬁcation codes are from Table 3.1B. No-plant control plots are represented
by a blank circle. The reference plots were not included in this analysis.
Monocultures are indicated by circles and mixtures by triangles .................. 57

Figure 4.1. Total plant biomass of A. gerardi produced at 12 and 16 weeks in relation

to variation among sites in 1998 ﬁeld above-ground plant biomass. Soils are
coded as in Table 4.1. Values are mean :1: standard error, n = 8. Signiﬁcance
values from the ANOVA are listed in Table 4.2A ................................... 82

Figure 4.2. Variation among sites in (A) total inorganic nitrogen, (B) N-mineralization

Figure

rate, (C) nitriﬁcation rate, (D) microbial respiration, (E) number of culturable
bacteria, and (F) microbial biomass. Sites are arranged in order of increasing
productivity, as listed in Table 4.1. For each variable, samples are distinguished
between treatments with Andropogon gerardi (solid) and no-plant controls
(open symbol). Values are mean i standard error, n = 8 for all except number
of culturable bacteria and microbial respiration which have n = 4. Signiﬁcance
values from the ANOVA are listed in Table 4.28 ................................... 84

4.3. The magnitude of the plant effect on soil and microbial processes across
sites in relation to the plant biomass gradient for (A) total inorganic nitrogen,
(B) N-mineralization rate, (C) Nitriﬁcation rate, (D) microbial respiration, (E)
number of culturable bacteria, and (F) microbial biomass. Values are mean i-

xiii

Figure

Figure

Figure

Figure

Figure

Figure

standard error, n = 8 except number of culturable bacteria and microbial
respiration where have n = 4 ............................................................ 86

4.4. Soil origin and plant effects on the structure of the soil microbial
community as measured by CLPP. CLPP patterns are distinguished between soil
microbial communities from the A. gerardi (solid symbols) and no plant
treatments (open symbols). Soil microbial communities in soils from different
sites are indicated by symbols: circles, BA; triangles, FK; upside-down
triangles, LL; diamonds, MK; stars, PL; pentagons, UL. Signiﬁcance values
from the RDA are listed in Table 4.4 ...................... ' ............................ 89

4.5. Distance-based Redundancy Analysis of PLFA proﬁles, investigating the
effects of soil origin, plant, and time effects on the structure of the soil
microbial community. The PLFA proﬁles are distinguished between soil
microbial communities from the A. gerardi (solid symbols) and no plant
treatments (open symbols). Soil microbial communities in soils from different
sites are indicated by symbols: circles, BA; triangles, FK; upside-down
triangles, LL; diamonds, MK; stars, PL; pentagons, UL. Labels for phospholipid
fatty acids are listed in Table 4.3. Signiﬁcance values from the RDA are listed
in Table 4.5 ............................................................................... 93

5.1. Total plant biomass of Bromus inermis, Solidago canadensis, and T rifolium
pratense, coded by A) soil origin and plant treatment, treatment codes are in
Table 5.2; and plant species by plant treatment for soil from B) Field K (PK)
and C) Upper Louden (UL). For (A), values are mean :1: standard error; for (B)
and (C), values are means. Signiﬁcance values from the ANOVA are listed in
Table 5.3A ............................................................................... 120

5.2. Root biomass of Bromus inermis, Solidago canadensis, and T rifolium
pratense, coded by (A) soil origin and plant treatment, treatment codes are in
Table 5.2; and plant species by plant treatment for soil from (B) Field K (PK)
and (C) Upper Louden (UL). For (A), values are mean :1: standard error; for (B)
and (C), values are means. Signiﬁcance values from the ANOVA are listed in
Table 5.3A ............................................................................... 121

5.3.Variation among plant treatments in (A) total inorganic nitrogen, (B) N-
mineralization rates, (C) nitriﬁcation rate, (D) microbial respiration, and (E)
microbial biomass. For each variable, samples are distinguished between the
two sites: FK (ﬁlled circles) and UL (hollow circles). Values are mean :1:
standard error. Signiﬁcance values from the ANOVA are listed in Table
5.3A ....................................................................................... 127

5.4. Distance-based Redundancy Analysis of Phospholipid Fatty Acid proﬁles,
investigating the effects of soil origin, plant, and time effects on the structure of
the soil microbial community. (A) Canonical principal component plot for the
full RDA. The PLFA proﬁles are distinguished between soil microbial

xiv

Figure

Figure

communities from FK soil (solid symbols) and UL soil (open symbols). Soil
microbial communities in soils with different plant communities are indicated
by symbols: circles. controls; triangles, monocultures; four or ﬁve-sided
polygons, mixtures. Labels for phospholipids are listed in Table 5.4.
Signiﬁcance values from the RDA are listed in Table 5.5A ...................... 132

5.5. (A-C) Partial canonical principal component plots derived ﬁom the
redundancy analysis of PLFA proﬁles. (A) Bromus inermis axis (1 of 1) with
soil origin, replicate, Solidago and T rifolium treatments, and time partialled out.
(B) Solidago canadensis axis (1 of 1) with soil origin, replicate, Bromus and
T rifolium treatments, and time partialled out. (C) T rifolium pratense axis (1 of
1) with soil origin, replicate, Bromus and Solidago treatments, and time
partialled out. Labels for phospholipids are listed in Table 5.4. Signiﬁcance
values from the RDA are listed in Table 5.5A ...................................... 133

5.6. Distance—based Redundancy Analysis of Phospholipid Fatty Acid proﬁles,
investigating the effects of exclosure, soil origin, plant, and time effects on the
structure of the soil microbial community. Canonical principal component plot
for the full RDA. The PLFA proﬁles are distinguished between soil microbial
communities from FK soil (solid symbols) and UL soil (open symbols) ....... 135

XV

 

 

CD

IOC

Pia

CHAPTER 1

INTRODUCTION

The role of species diversity and composition in ecosystems is increasingly under
scrutiny due to concerns about the potential impacts of the current rapid decline in the
Earth’s biodiversity. Species diversity may have important effects on key ecosystem
functions like nutrient cycling, water quality, and productivity Rosenzweig 1995,
Tihnan 1996). Recent investigations into the relationship between diversity and
function have focused mainly on how changes in primary producers and consumers
affect ecosystem processes (Schlapfer and Schmid 1999, Rosenzweig 1995). The
interaction between aboveground and belowground (soil) communities in mediating
these processes has been less studied (Schlapfer and Schmid 1999, Ohtonen et al.
1997). To understand controls on diversity and the role of diversity in ecosystem
function it is important to understand the relationships among organisms in the
ecosystem. While considerable attention has been paid to factors that affect the
composition and function of communities of macroorganisms, very little is known
about the factors that affect the structure of soil microbial communities (Ohtonen et al.

1997, Tiedje 1995).

Resources available to soil microorganisms are primarily derived from plants. Most of
the carbon and nitrogen entering the soil matrix results from litterfall, root exudates, or
root death (Paul and Clark 1996). As a result, the composition and productivity of the

plant community inﬂuences the soil microbial community. Similarly, the productivity

OI

mic

cor
C la
by

Sci

relt‘

qua
con

9C0.

The

or diversity of the plant community may be affected by processes mediated by soil
microorganisms (e.g. N-mineralization rates). By performing key steps in the cycling
of nutrients (carbon, nitrogen, and phosphorus, among others), the soil microbial
community plays an essential role in the functioning of terrestrial ecosystems (Paul and
Clark 1996). In most temperate grassland and forest systems, plant growth is limited
by nitrogen (or a combination of nitrogen and phosphorus, Shaver and Chapin 1980,
Schmidt et al. 1997, Jonasson et al. 1999). The soil microbial community controls the
release of inorganic nitrogen to plants; however, the soil microbial community is most
often limited by carbon (Zak et al. 1994). Therefore, the rate at which limiting
nutrients are made available to plants is likely to be inﬂuenced by the amount and
quality of carbon available to soil microorganisms. Consequently, changes in the plant
community likely will change the soil microbial community and potentially affect

ecosystem function.

Thesis Overview

I am interested in the inﬂuence of the plant community on the structure of soil
microbial communities and the processes they mediate. The challenge is to distinguish
between the direct effects of plants (through changes in soil carbon inputs) and indirect
effects (due to soil characteristics) on the structure of the soil microbial community. In
this dissertation, I explore the relationship between the plant and soil microbial
communities through a combination of observational studies and manipulative

experiments at two different scales: the plant community scale and the individual plant

mi

ant

C1

C01

pro
pro
bep

H01

COm

1mg

species scale. I addressed the following questions through my research: (1) How does
plant productivity affect the activity and structure of the soil microbial community? (2)
How does plant community diversity affect the structure and function of the soil
microbial community? (3) How does plant community composition affect the structure
and function of the soil microbial community? (4) How do soil factors inﬂuence the
structure and function of the soil microbial community and can plants mediate soil

effects?

Chapter 2 examines the ﬁrst two questions on the relationships between plant
community productivity and diversity and the soil microbial community. I compared
patterns of diversity in the plant and soil microbial communities along a productivity
gradient in an old ﬁeld at the Lux Arbor Reserve at the W. K. Kellogg Biological
Station in southwestern Michigan. The sampled gradient had a high diversity — low
productivity plant community on the ridge top that graded into a low diversity —high
productivity plant community down-slope. There was a strong positive relationship
between above ground plant biomass and soil microbial respiration at the site.
However, this association was confounded by changes in edaphic characteristics
(moisture and nitrogen) and with the composition of the plant community that also
varied along the gradient. Distinguishing plant from soil effects on the soil microbial
community is a necessary ﬁrst step in determining factors that structure the
composition and affect the function of soil microbial communities. In the following
chapters I describe the results of ﬁeld and greenhouse experiments designed to

investigate the independent effects of plants and soils on the soil microbial community.

 

Ioz
strut
ﬁmi
Silu
exp:
were
sml

conu

with .
pknn
anesi
4i. 1 (
Plant.
pmdm

dﬁl‘cm

To address questions of how plant community diversity and composition inﬂuence the
structure and function of the soil microbial community, I sampled soil communities
from a series of experimental grassland plots from the BIODEPTH experiment at
Silwood Park, England. This research is described in Chapter 3. In the BIODEPTH
experiment, the number of plant species and the number of plant functional groups
were varied to create plant communities with different plant diversities on the same
soil. I found that plant community biomass, composition, and diversity all affected the

composition and several functional traits of the soil microbial community.

Many ﬁeld studies have detected differences among soil microbial communities
sampled from sites with contrasting plant communities (Zak et al. 1994, Grayston and
Campbell 1996, Chapter 2), but in these studies plant and soil effects are confounded.
Soil in different sites has been shaped by a variety of factors besides differences in
plant community composition, such as parent material, disturbance and management
regimes. Therefore, differences among soil microbial communities sampled from sites
with different plant communities cannot be attributed solely to the differences in the
plant communities because the soil characteristics and histories also differ. To
investigate the direct effects of soil origin on the soil microbial community (question
4), I conducted a greenhouse experiment that compared soils from six different local
plant communities. The six sites differed in fertility, soil organic matter, and plant
productivity, and these factors had detectable, correlated effects on soil processes. To

determine if plants could mediate these differences, a common plant species,

 

.‘I’lti
COII

som

860

soil

com
com
and
and
thrc
dclc
{Um

soil

This
11111:

disc

Andropogon gerardi, was grown in each soil. Interestingly, when soil microbial
communities were grown in the presence of this species, the effect of soil origin on

some functions (6. g. microbial respiration and N-mineralization rates) was diminished.

Because soil microorganisms are dependent on carbon, and most available carbon in
soil comes from plants, the identity of the plant species supplying carbon to the
microorganisms may inﬂuence the structure and function of the soil microbial
community. In Chapter 5, I describe the results of a greenhouse experiment in which I
compared the effects of three plant species, grown in two distinct soils, on the structure
and function of the soil microbial community. Soils from two of the old ﬁelds (high
and low fertility) used in the previous experiment were planted with all combinations of
three plant species common to local old ﬁelds. The experiment allowed me to
determine that (1) different plant species can have unique effects on the structure and
function of soil microbial communities and (2) the effects of different plant species on

soil microbial community structure and function are non-additive.

This dissertation suggests that both the origin of the soil and the presence of a plant
inﬂuence the structure and functioning of the soil microbial community. In Chapter 6 I

discuss the overall conclusions from this collection of ﬁeld and greenhouse studies.

 

 

Fie

 

 

abai
“as
con-

Elba

Field Sites

The greenhouse experiments presented in Chapters 4 and 5 of this dissertation were
conducted with soils from six successional old—ﬁelds at the W. K. Kellogg Biological
Station of Michigan State University in southwestern Michigan (Kalamazoo County;
42° 24’ N, 85° 24’ W). The sites varied in fertility, species richness, and dominant
plant type, but all were located on Kalamazoo sandy loam soil (Table 1.1, Figure 1.1,
Burbank et al. 1992). The six sites also differed in past land use and time since
abandonment ranging in age from 20 to 50 years. McKay (MK) ﬁeld was abandoned
from agriculture in 1973; a section was plowed once in 1981 and then re-abandoned
(Burbank et al. 1992). Both the Upper (UL) and Lower (LL) Louden ﬁelds were
abandoned from agriculture in 1951 (Burbank et al. 1992). The Bailey (Ba) ﬁeld site
was farmed until ten years prior to this sampling (K.L. Gross, personal
communication). The Pond Lab Orchard (PL) and Field K (FK) sites had been

abandoned for at least twenty years (Foster 1996).

Soils from all six sites were used in the ﬁrst greenhouse experiment (Chapter 4; Table
1.1, Figure 1.1). Soils from two of the six sites (a high and a low fertility site) were
used for the second greenhouse experiment (Chapter 5; sites FK & UL; Table 1.1,

Figure 1.1).

 

Ll

PL

FK

Table 1.1. Plant productivity and diversity of abandoned old ﬁelds from which soil was
collected for the greenhouse experiments. Values for peak aboveground plant biomass
(an estimate of primary productivity), species richness, and mean percent organic
matter are expressed as mean i standard deviations (n = 6). Values that are signiﬁcantly
different for a given variable based on Fisher’s LSD test are indicated by different

 

 

 

 

letters.
Site Dominant Peak Plant Biomass Species diversity Soil Organic
Plant Form (standing + litter, g/mz) (#/m2) Matter (%)
MK Grass 188 i 16 a 2.2 i- 0.4 a 2.40 i 0.43 a
UL Forb 320i23b 15.8:1.0e 3.17:0.23b
Ba Forb 424i52c ll.3il.0d 3.03:0.17b
LL Grass 432 i 27 c 8.5 i 0.4 c 3.84 i: 0.32 0
PL Grass 480 i 48 c 5.7 i 0.8 b 3.63 i 0.28 c
FK Grass 592 i 22 d 1.3 i 0.2 a 3.84 i 0.20 c
7

.mozm omoﬁ montage 2 033. A33 Sumo.“ 8me 35 Eva xv. 2: mo
c2802 05 wcumbmszm mecca HEB cosﬁm Hammo—omm wwozox .M .3 05 «o 932 ._._ onE

‘

 

 

25-532 MP...—

 

 

.._.._J

 

 

 

 

 

 

Fc

Cu

lor

0hr
Pau
Ros.

Sch]

Schn

Shari

Iicdh

TUHHM

Zak. J.

 

REFERENCES

Burbank, D. H., K. S. Pregitzer, and K. L. Gross. 1992. Vegetation of the W. K.
Kellogg Biological Station. Michigan State University Agricultural Experiment
Station Research Report Number 510.

Foster, B. L. 1996. Plant competition and diversity in relation to productivity in old-
ﬁeld plant communities. Ph.D. Dissertation. Michigan State University, East
Lansing, MI, USA.

Grayston, S. J. and C. D. Campbell. 1996. Functional biodiversity of microbial
communities in the rhizospheres of hybrid larch (LarL'x eurolepis) and Sitka
spruce (Picea sitchensis). Tree Physiology 16: 1031-8.

Jonasson, S., A. Michelsen, I.K. Schmidt, and E.V. Nielsen. 1999. Responses in
microbes and plants to changed temperature, nutrient, and light regimes in the
arctic. Ecology 80: 1828-1843.

Ohtonen, R., S. Aikio, H. Wire. 1997. Ecological theories in soil biology. Soil
Biology and Biochemistry 29: 1613-1619

Paul, EA. and FE. Clark. 1996. Soil Microbiology and Biochemistry. 2nd ed. San
Diego, CA: Academic Press, Inc.

Rosenzweig, ML. 1995. Species Diversity in Space and Time. Cambridge University
Press, Cambridge.

Schlapfer, F. and B. Schmid. 1999. Ecosystem effects of biodiversity: a classiﬁcation
of hypotheses and exploration of empirical results. Ecological Applications 9:
893-912.

Schmidt, I.K., A. Michelsen, and S. Jonasson. 1997. Effects of labile soil carbon on
nutrient partitioning between an arctic graminoid and microbes. Oecologia 112:
557-565.

Shaver, GR. and F .S. Chapin, III. 1980. Response to fertilization by various plant
growth forms in an Alaskan tundra: nutrient accumulation and growth. Ecology
61: 662-675.

Tiedje, J. M. 1995. Approaches to the comprehensive evaluation of prokaryote
diversity of a habitat. In Microbial Diversity and Ecosystem Function, eds. D.
Allisopp, R.R. Colwell, and BL. Hawksworth. Wallingford, UK: CAB
International.

Tilman, D. 1996. Biodiversity: Population versus ecosystem stability. Ecology 77:
350-363.

Zak, J. C., M. R. Willig, D. L. Moorhead, and H. G. Wildman. 1994. Functional
diversity of microbial communities: A quantitative approach. Soil Biology and
Biochemistry 26: 1101-1108.

 

 

 

Th:
20er

prm

lntr

Com 1‘.
inﬂu,
MlCrr;
strum
(Obit)
Comm

 

 

CHAPTER 2

PATTERNS OF DIVERSITY 1N PLANT AND SOIL MICROBIAL COMMUNITIES
ALONG A PRODUCTIVITY GRADIENT IN A MICHIGAN OLD-FIELD

The following chapter was published as the article: Broughton, LC. and KL. Gross.

2000. Patterns of diversity in plant and soil microbial communities along a

productivity gradient in a Michigan old-ﬁeld. Oecologia 125: 420-427.

Introduction

A central question in ecology is why there are so many different organisms on the earth
(Hutchinson 1959). Much of the work focusing on macroorganisms has emphasized
the role of factors such as productivity, disturbance, energy, predation, resources,
stochasticity, and colonization in determining the diversity of plant and animal
communities (Rosenzweig 1995). Considerably less is known about what factors
inﬂuence the abundance and diversity of microorganisms (Tiedje 1995).
Microorganisms have rarely been incorporated into studies of mechanisms that may
structure diversity-productivity relationships for plants and other macroorganisms
(Ohtonen et al. 1997, Schléipfer and Schmid 1999). Although plant and soil
communities are functionally linked, few studies have examined how patterns of

diversity in plant and soil microbial communities co-vary.

10

 

SO;

nit:

5011

C017.

COI‘.~

inx'e

 

 

 

an i
diii
(6g
Olilt

Dri'

mic

199

CO]
Sui

Ihi

Soil microbial communities are often limited by carbon (D. R. Zak et al. 1994) or
nitrogen (Zak et a1. 1990). Because the extant plant community is usually the main
source for both of these resources, the composition and diversity of the soil microbial
community may be closely associated with the plant community. It is difﬁcult to assess
composition and diversity of soil microbial communities. As a result, most
investigations use techniques that assay different aspects of a subset of the microbial
community. Two tools that are commonly used by ecologists to characterize the soil
microbial cormnunity are the Biolog assay and fatty acid methyl ester (FAME) proﬁles.
Biolog assays sole-carbon-source utilization by the microbial community and provides
an index of functional diversity. Biolog proﬁles have been successfully used to
differentiate soil microbial communities associated with different plant communities
(e.g. J. C. Zak et al. 1994; Goodfriend 1998), especially when used in concert with
other techniques, like fatty acid methyl ester (FAME) proﬁles (e.g. Buyer and
Drinkwater 1997). FAME proﬁles reﬂect the phenotypic composition of the soil
microbial community (Tunlid and White 1992) and can be used to distinguish among
microbial communities with different compositions (Haack et al. 1995; Cavigelli et al.

1995).

I investigated the relationship between the structure and activity of the soil microbial
community and its relationship to the plant community within an ecologically variable
site. I hypothesized that the structure of the soil microbial community would vary at
this site in relation to: (1) soil characteristics, (2) plant productivity, and (3) plant

diversity. I investigated the relationship between the soil microbial community and

11

 

 

(A

these three variates along a topographic productivity gradient in a mid-successional old-

ﬁeld in southwestern Michigan.

Methods

Site Description

The study site was in a mid-successional abandoned ﬁeld at the Lux Arbor Reserve of
the W. K. Kellogg Biological Station (KBS) in southwestern Michigan. This site had
been abandoned approximately 25 years from agricultural production and during that
period had not been grazed, burned, or otherwise managed. Successional ﬁelds in this
area typically attain a stable species composition of herbaceous perennial 5 to 25 years
after abandonment (Huberty et al. 1998). There has been no apparent change in the
plant community at this site over the past ten years (K.L. Gross, personal
communication). The study site was located along a gentle slope, approximately 15°
from the top to the bottom of the hill, along which there were apparent changes in plant
species composition and productivity. The soil at the site is Kalamazoo sandy loam soil

and does not vary across the study area.

Sampling Design and Characterization of the Gradient

1 established ﬁve parallel transects, 7.5 m apart, perpendicular to the slope of the hill

and sampled soil and vegetation in seven 0.25 m2 plots placed at 10 m intervals along

12

.Ab‘

 

 

chp
dr\ 7‘.

pan

 

seas
and

bior

Ur

each transect (n = 7 per transect). To characterize the gradient, I measured light at

ground level, soil moisture, soil inorganic nitrogen, and aboveground plant biomass.

Aboveground plant biomass and species composition were sampled in June 1996 by
clipping the plants at ground level (0 cm above the soil surface), sorting by species,
drying at 60°C for 48 hours, and weighing. To better estimate peak plant biomass,
particularly at the more mesic end of the gradient, which was dominated by wann-
season grasses, the same plots were re-clipped in July 1996. Samples were processed
and treated as before. Peak plant biomass was calculated as June biomass + July

biomass (both living and standing dead).

Light availability at ground level was determined prior to clipping in June and July.
Measurements were made at midday (1100-1400 hours EDT) using a Sunﬂeck PAR
Ceptometer (Decagon Devices, Inc.). I measured photosynthetically active radiation
(PAR) in full sun 1 m above the plots and took four measurements of PAR at ground
level within each plot (cardinal directions). I averaged these four data points to obtain

an estimate of the percentage of ﬁlll sunlight penetrating to ground level.

Soils were also sampled in June and July. For the soil analyses, I aggregated ﬁve 2.5
cm diameter by 10 cm deep soil cores taken from each 0.25 m2 plot in an X-shaped

pattern. Samples were placed in sealed plastic bags and kept on ice for up to 6 hours
until they could be returned to the laboratory. There, they were passed through a 2-mm

sieve and sub-sampled for gravimetric soil moisture and nitrogen content within 24

13

llOl

 

Silll‘

ﬁll.

 

 

hours of sampling. Samples were kept at 4°C until processed. Gravimetric soil
moisture was determined by weight loss after drying 10 to 15 g soil at 105°C for 24

hours. For the nitrogen assays, I extracted 20 g fresh soil in 100 ml 1 M KCl. These

samples were shaken for 1 minute, settled for 24 hours at room temperature, and
ﬁltered through a 1 pm Gelman glass ﬁlter. The N03' and NH4+ concentrations of the

extracts were determined using an Alpkem Auto-Analyzer. The remaining soil was
used to characterize the soil microbial community. Soil for FAME analyses was kept at

-20°C until the fatty acids were extracted.

Characterization of the Soil Microbial Community

I characterized the soil microbial community from samples taken in July using a
modiﬁed substrate-induced respiration (SIR) method, carbon-source utilization
(Biolog) and FAME proﬁles. SIR assesses the microbial biomass of the soil microbial
community and is a good indicator of microbial respiration (Hassink 1993). For SIR

microbial biomass, soil slurries were shaken with and without glucose in Erlenmeyer
ﬂasks sealed with paraﬁlm, and the headspace C02 was measured. For the control, I
combined 25 g soil and 25 ml water in 125 ml Erlenmeyer ﬂasks, and for the glucose-
addition I substituted 25 ml 30 mg ml'I glucose for the water. Both sets were shaken
for 2 hours at 22°C. After 2 hours, I transferred 5 ml of the headspace gas to a serum

vial and measured the initial C02 on an ADC series EGA infrared COz gas analyzer

l4

(I1

sha

For

buf

Chi

Thn
in 0
prcc
knox
take]
oﬁcr
Slant
Smur
Hand
aVera
reﬁdii

Slmlh

(The Analytical Development Co. Ltd., Hoddesdon, Herts., UK). The ﬂasks were

shaken for another 38 hours and headspace C02 again measured.

For the Biolog assay, 1 g of fresh, sieved soil was shaken with 99 m1 of 1% phosphate
buffer solution for 20 min and 150 pl of the solution was transferred into each well of a
GN Biolog microtiter plate (95 Carbon sources; Biolog, Inc., Hayward, CA 94545).
Three replicate plates were inoculated for each plot. The plates were incubated at 25°C
in the dark and optical densities were measured after 24 and 48 hours using an Emax
precision microplate reader (Molecular Devices Corp., Menlo Park, CA). It is well
known that inoculation densities from a standard amount of soil can vary for samples
taken from different environments (Konopka et al. 1998). Optical density measures are
often standardized to account for differences in inoculation densities; however, the
standardizations have been criticized for not accurately reﬂecting growth across
samples with different compositions (Konopka et al. 1998). Therefore, instead of
standardizing optical densities, I chose to take advantage of differences and used
average well color development (AWCD) from Biolog (corrected within plate for water
reading) as an index of microbial respiration. Because proﬁles at 24 h and 48 h were

similar only the results from the 48 h time point are presented here.

To obtain fatty acids for FAME analysis, I ﬁrst extracted the lipids from whole soil
samples for 2 h using a mixture of dichloromethane (DCM):methanol:phosphate buffer
(1:2:0.8 v/v/v), following a modiﬁed Bligh-Dyer procedure (Bligh and Dyer 1959). I

then saponiﬁed the samples using 1 ml NaOl-l (15% w/v) in methanol (50% v/v) at

15

 

Me

31011

nUml
Cafbc

$700;

Stan’s

T0 01

Comp(

100°C for 30 min and methylated the sample with 2 ml 6N HCl in methanol at 80°C for
10 min. I extracted the fatty acid methyl esters into 1.25 ml (1:1 v/v) methyl-tert—butyl
ether-hexane for 10 min and washed the extract with 3 ml 1.2% NaOH. FAME
analyses were carried out using a HP 5890 series H gas chromatograph (Hewlett
Packard Co., Palo Alto, CA) equipped with a 7673 autosampler and ﬂame ionization
detector (Microbial ID Inc., Newark, DL). Peaks were identiﬁed by comparison with
an external standard. I performed all analyses on the fatty acid proportions of the total

peak area to correct for differences in overall peak area.

I describe fatty acids using standard nomenclature where the total number of carbon
atoms appears before the colon and the total number of CC double bonds appears aﬁer
it. Cyclo-propane analogs are indicated by "cyclo," and the location of the epoxy bond
is indicated by a "c" followed by two numbers. If the cis or trans conﬁguration is
unknown, the word "at" is used. The number following "cis", "trans" or "at" indicates
the location of the double bond in relation to the carboxyl end of the molecule. A
number before "OH" indicates the location of the hydroxyl group in relation to the

carboxyl end of the molecule. Those fatty acids with the same retention time are

grouped as "sum in feature" and given a unique number designation.

Statistical Analyses

To obtain an index of productivity along the gradient I performed a principal

components analysis (PCA) on those variates expected to be closely related to

16

‘ﬁ

 

 

1110.1
100

1110

Res

(ll 4

productivity: light at ground level, gravimetric soil moisture, soil inorganic N, and peak
plant biomass. To examine the productivity-diversity patterns, I regressed plant species
diversity, number of carbon sources metabolized (Biolog), AWCD (from Biolog), and
SIR microbial biomass against this index of productivity. Changes in plant community
composition along the gradient were evaluated with indirect gradient correspondence
analysis on species-speciﬁc aboveground plant biomass. To visually compare the plant
and soil community patterns, I performed K-means cluster analysis on plant species-
speciﬁc biomass data, carbon source utilization proﬁles (Biolog), and fatty acid methyl
ester proﬁles (FAME). Multidimensional scaling (MDS) was used to predict the
number of clusters expected for the Biolog and FAME proﬁle data. As there were
more parameters than samples for the Biolog data, I randomly split the parameters into
two subgroups that were run through all analyses independently. The results of these

two independent analyses were consistent, so the data from only one is presented.

Results

Characterization of the Gradient

Light at ground level, gravimetric soil moisture, soil inorganic N, and peak plant

biomass all co-varied along the topographic gradient. Light availability at ground level

(%PAR) decreased from 85% at the crest of the hill to 3% at the base of the hill.

Gravimetric soil moisture increased from 15% to 36%, soil inorganic N increased from

17

20>..ch 00:20:08 33 30 30230ch
830: 0.33 mxwmod + $2- .I. a .323 3:933. mm 2.: .20 Vm 0.: 32m :30: 032.2: :00 _ Um
300250: 9:30:28 051 Amv .390 v Q :0 300:3me 203 Gmd n c 03208 :5 Smd u 5 2w: 000
N 00:02 300250: 30:20:00 Sood v a :0 300:2:me 0.53 Cod n c 30803 :00 Aowd H 5 Z Ammo n
c 23208 ﬁnd- u b Em: 0:0 _ .582 300250: 30:29:00 0203833: 30:22, 0.: :0 $2 .2 000
{exec 8.: 330000 N 0:0 2 DA $3803 E20 203 :30 .2 030302 20m 6.3208 :8 205. 030%
a £0: 50 05388 82 as; 20.8 <00 a: 800 N 05 _ 00 50 ca .0532 502 2V .2 2&5

 

 

 

05 8% .560 8:320 220.48 2 00
co cm ow on on 3 o o._ m6 o6 We- o. T
0 _ _ T _ 0 P _ _ _ _ _
. 1 v- I r o T
.8 o v a
one u w. . f m- M
m
m I. I m.ol d
M. D
"r. mars—em “cw—mm 7»
IA \1/
w - -2 u
m %
vim Z 2530.: /\
D as
U 1 mi \ r We
2332 :8
i < I o._

 

 

 

 

 

 

18

[a J
i)

W

IP;
of r
bot:
inor‘
has:
accu
nm’k"
grou:

addiz

 

3.29 to 18.58 pg N/ g dry soil, and above-ground plant biomass increased from 89.2 to

309.6 g m'2 along the slope.

I performed a PCA to obtain an index of productivity that incorporated these measures
of resource availability plus above-ground plant biomass. One sample point at the
bottom of the hill was excluded from the analysis because of an abnormally high
inorganic soil N value (IO-fold higher than the median). The ﬁrst principal component
based on resource levels measured in June and peak plant biomass (June + July)
accounted for 64.8% of the variation in the data set (X = 2.594). Soil moisture,
nitrogen, and peak plant biomass were positively correlated with PCI, whereas light at
ground level was negatively correlated with PCI (Figure 2.1A). PC2 accounted for an
additional 17.1% of the variation, but showed no pattern in relation to the gradient.
Therefore, I used PCl as an index of productivity in the remaining analyses (r2: 0.50,
Figure 2.1B). A PCA performed on the same variates from the July sampling was
indistinguishable from the PCA on the June data, so I will present and use only the June

results here.

Plant and Soil Microbial Community Relationships with Productivity

Plant species richness declined with increasing productivity (PCl) at this site, but
productivity accounted for little of the variation in diversity (r2 = 0.17, Figure 2.2A).

This relationship was clearly driven by two low diversity points at the high end of the

19

6m u : .2988 0082.800 $3 :8 8888.0: 880: 8:5 .9: 1. fund u a .080:
ow 8:: 80880: 8:68 080:8 Em 0:: 80: 8208 80: :80 b: w \U-~OU w: Am: Hm: I 5%: u x .080: ov 8:: .8800
Em 80: 8208 80: E00 b: m CHOU was: as: .o + mad u a .080: we 8:: 8088—08: 880 :03 0m08>0 w0_0_m CV 580:
w: 8:: 8:088 :00 0:: 0:: 8288208 80:80 8:80 .8 8:80: 0: H 5: ”025: .. 3.2 u a .0880: 880% 82: 2: 0:0
0:: 8 8:8:8800 3:088 88 82: 0:: 35:08: 82:3 .0880 :8 08:8 53:038.: :0038: 8:30:20: 0: H .N.N 08m;

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

08:5 £38285 08:5 b38885
vmm_o_-m-m-v. va:o_-m-m-v..
_ . - . r. F . + FL . . . _ . _ . D
1 :5.on .0 D 160v: .0m..
1 $.oﬂmm ...N \n} —No"~1~— LN“)
«no «a
1v 3 D
.0 ..o km
:W .33
1» m. hm.
2 m 0m.
N.— .... 1 .mA
m —W l 0 Q km.-
iww m . . _ J _ _ . d . .Ux
08:5 838208: 08::— b_>:0:_005 08:5 538.805
VMN_0_-~-M-Y em~_0_-m-m-v amm_o_-~-m-v
Frrb...... ._.._..p. F..:..LL_1N
1 .808 80 1 . 5N N 1 8.0!:
~81”: 1 0m m 1 . ton”: hm
h.° 8 .0. o. w.- TC m.-
m 4 o o o CVW m I” n
0: m. L .. . . 00m 1. m.
v . . .. m. m 2M
M coco .0 ”no bwmu 0.. m
h. 3 n . .. W m 88
a 4 O O 0 job ﬁOb EM“
0 3
00 1 m 00 m Em...
q 4 a 4‘4 - - d u u 4 d 1 _4 — d - m

20

produ
char

(0.25

854 5
hi ghes
nucrol
Idaho:
”ﬁgun

nncrol

COnSUI

Cchn

bECaug

COIN!)

To Vi
Sﬁpar
PAM
mom
(th

productivity gradient (Figure 2.2A). However, averaging over transects there was a
clear decline in species richness across the gradient: an average of 13 species per plot
(0.25 m2) were found at the top of the hill, where productivity was lowest, while as few
as 4 species per plot were found at the bottom of the hill, where productivity was
highest (Figure 2.2A). Although the number of carbon sources metabolized by the
microbial community varied from 23 to 75 across this gradient, there was no

relationship between the number of carbon sources metabolized and productivity

(Figure 2.2B). However, AWCD at 48 hours, which could be indicative of either
microbial respiration or biomass, increased from 0.4 to 1.4 as productivity increased (r2
= 0.52, Figure 2.2C). Similarly, SIR basal activity rate after 40 hours also increased

along the gradient (r2 = 0.21, Figure 2.2D), as did the rate at which glucose was

consumed between 2 and 40 hours (r2 = 0.44, Figure 2.2B). I used the SIR control and

experimental treatments separately in this analysis as measures of microbial respiration

because the SIR time course was insufﬁcient to determine microbial biomass.

Compositional Shifts in the Plant and Soil Microbial Communities

To visually compare patterns in the plant and microbial communities, I performed
separate K-means cluster analyses on the plant species biomass data and the Biolog and
FAME proﬁles of the soil microbial communities. A plot of above-ground plant
biomass across the study site clearly shows the topographic-productivity gradient
(Figure 2.3A) and allows visual comparisons to plant diversity and microbial

community measures (Figure 2.3B-D). There were compositional changes in the plant

21

PW“..—

Figure 2.3. Results from K-means cluster analysis evaluating the changes in production
and composition of the plant and soil communities in relation to transect position at this
site. (A) Contour plot of above-ground plant biomass, values are g/ m2; (B) Plant species
composition: cluster 1 has no dominant species, cluster 2 is dominated by Rubus sp.,
cluster 3 by Solidago canadensis, cluster 4 by S. canadensis and Poa pratense, cluster 5
by Agropyron repens, cluster 6 by Poa pratense, and cluster 7 by Polygonum amphibium
var. emersum (see Table 2.1 for species lists). (C) BiologTM carbon source utilization
proﬁles separated into two clusters based on AWCD; and (D) FAME profiles: cluster 2

had smaller proportions of 18:1 cis 9, 16:0, and summed in feature 9 (18:2 cis 9, 12 and

18:0 anteiso, Table 2.2) than cluster 1.

22

Distance Down Slnpc (m)

 

Distance Down Slope (m)

Distance Down Slope (m)

 

 

 

 

 

7.5

Transect (m)

 

10‘

20‘

30-4 0

40 ‘

60“ "

 

 

High AWCD

:9 s '." '1' I—

 

 

I l I l

7.5 15.0 22.5 30.0
Transect (m)

15.0 22.5 30.0

{V 1') h—

 

10-

Distance Down Slope (m)

 

   
 

 

 

 

 

 

 

 

60d 0 I Q r—
I I I r I
0.0 7.5 15.0 22.5 30.0
Transect (m)
D
L l 1 l 4
0— a a Q 9 9 I-
104
E
o 20" ’
c.
.9
m
E 304
o
D
<1)
8 40-~
£3
2
Q
50— a 1 o .. .—
60— I {32 A -'-'- ‘: I—
T I I I I
0.0 7.5 15.0 22.5 30.0
Transect (m)
Figure 2.3.

23

comn‘
indlca
l
comp?
analys
produt
comm
n0 €le

perenn

Agropi

 

 

Plots ;

domin;

MDS <
relied 1
with n
numbC
AWCI
the [(1
lower
There

the Sr:

community along the productivity gradient. MDS on the species biomass data
indicated three to seven valid clusters: I chose to illustrate seven in order to more
completely portray the variation in plant composition (Figure 2.3B). The cluster
analysis revealed an inverse relationship between plant community diversity and
productivity. More speciﬁcally, the cluster analysis showed a shift in the plant
community from the top of the slope, where there was a mixed community of forbs and
no clearly dominant species (Table 2.1, cluster 1), to a mid-slope region dominated by
perennial herbs (Table 2.1, clusters 3-4) to a low diversity community dominated by
Agropyron repens near the bottom of the slope (Table 2.1, clusters 5-7, Figure 2.33).
Plots at the bottom of the hill with highest soil moisture and productivity were

dominated by Polygonum amphibium var. emersum (water smartweed) (Figure 2.33).

MDS of the Biolog proﬁles indicated two strong clusters; however, cluster formation
relied solely on AWCD and not number or type of carbon sources. This is consistent
with the soil microbial community - productivity relationship (Figure 2.2B,C). The
number of carbon sources was not related to the productivity index (Figure 2.2B), while
AWCD was signiﬁcantly related to the productivity index (Figure 2.2C). Similarly, in
the K—means cluster analysis, sites located at the top of the slope were characterized by .
lower AWCD, while the sites at the base of the hill had higher AWCD (Figure 2.3C).
There was no difference in the number or types of carbon sources metabolized across

the gradient (data not shown).

24

 

Table
shom
comir
Nome

 

Clustc
-1311)”

1(13)

"‘1 (6)

Table 2.1. Plant species composition in the clusters from the K-means cluster analysis
shown in Figure 2.38. Plant species are listed from most common (by biomass) to least
common. Only species that have a total biomass > 1 g for the cluster are listed.
Nomenclature follows Gleason and Cronquist (1991).

 

Cluster

(#plots)
1 (13)

2 (2)

3 (9)

4 (6)

5 (2)

6(1)

7 (2)

Plant species

Centaurea maculosa, Rubus occidentalis, Hieracium sp., Achillea
millifolium, Rubus allegheniensis, Rumex acetosella, Poa compressa,
Panicum sp., A gropyron repens, Solidago canadensis, Potentilla recta,
Solidago graminifolia, Poa pratense, Aster pilosus, Plantago lanceolata,
Phleum pratense. Cerastium vulgatum, Daucus carota, T rifolium pratense.
Lespedeza capitata, Dactylis glomerata

Rubus occidentalis, Poa pratense. Solidago canadensis, Phleum pratense,
Hieracium sp., Rumex acetosella, Agropyron repens, Polygonum
amphibium var. emersum, Panicum sp.

Solidago canadensis, Rubus occidentalis, Poa pratense, Achillea
millifolium, Monardaﬁstulosa, Phleum pratense, Poa compressa, Rumex
acetosella, Agropyron repens, Daucus carota, Potentilla recto, Solidago
graminifolia, Cornus racemosa, Apocynum cannabinum, Hieracium sp.,
Lespedeza capitata, Rubus allegheniensis, Centaurea maculosa, T rifolium
pratense. Cerastium vulgatum, T araxacum oﬂicinale, Rumex crispus,
Hypericum perforatum

Poa pratense, Solidago canadensis, Agropyron repens, Achillea millifolium,
Monardaﬁstulosa, Phleum pratense. Aster strigosa, Potentilla recta,

T araxacum oﬂicinale, Galium aparine, Daucus carota, Rumex crispus,
Solidago graminifolia, Rumex acetosella

Agropyron repens, Solidago canadensis, Monardaﬁstulosa, Polygonum
amphibium var. emersum, Rubus occidentalis, Solidago graminofolia,
Galium aparine, Poa pratense, Achillea millifolium

Poa pratense, Agropyron repens, Solidago canadensis, Achillea millifolium,
Polygonum amphibium var. emersum

Polygonum amphibium var. emersum, Agropyron repens, Polygonum
persicaria, Solidago canadensis, Rumex acetosella, Poa pratense, Poa
compressa

 

25

 

Altho
no dl.‘
used :
analys
18:1 C
than 3
did n1“

diversi

 

Discus

I had h
be rel;
Becaus
index

Comnu
Comml
[he di\
Plant I

Although there was variation in the FAME proﬁles from these samples, MDS showed
no distinct clusters in this data set along the productivity-diversity gradient. Fatty acids
used in MDS and cluster analyses are listed in Table 2.2. When I forced the cluster
analysis to create two clusters, soils from cluster 1 had larger proportions of fatty acids
18:1 cis 9, 16:0, and summed in feature 9 (18:2 cis 9, 12 and 18:0 anteiso, Table 2.2)
than soils from the cluster 2 (Figure 2.3D). However, the cluster-based FAME proﬁles
did not show any pattern concordant with peak plant biomass (Figure 2.3A), plant

diversity (Figure 2.3B), or Biolog AWCD patterns (Figure 2.3C).

Discussion

I had hypothesized that the structure of the soil microbial community at this site would
be related to plant community diversity, plant productivity, or soil characteristics.
Because these three factors covaried at this site (Figure 2.1), I combined them into an
index of productivity, but still could not detect any relation to the soil microbial
community structure. Neither Biolog nor FAME assays of the soil microbial
community were strongly related to variation in productivity. There were changes in
the diversity and composition of the plant community associated with soil fertility and
plant biomass; however, these differences in plant community composition had no

detectable effect on the composition of the soil microbial community.

I did ﬁnd evidence, however, that suggested the respiration (or biomass) of the soil

microbial community varied in relation to plant productivity, paralleling the edaphic

26

 

Tabl

F311}
T210
11:0
C911
14:0
15:0
15:0
15:1
15:0
16:0
16:1
16:1
16:0
150 1
amei
1?:0
sumr

Table 2.2. Fatty acids used in K-means cluster analysis of FAME proﬁles.

 

 

Fatty Acid

12:0 17:0 anteiso

11:0 iso 30H 17:1 cis 10

C9 dicarboxylic acid 17:0 cyclo

14:0 18:3 cis 6,12, 14
15:0iso 18:1cis9

15:0 anteiso 18:0

15:1 cis7 19:0 cycloC11-12
15:0 18:0 20H

16:0 iso 20:4 cis

16:1 cis 9 20:0

16:1 cis 11 22:0

16:0 23:0

iso 17:1 G 22:0 20H
anteiso 17:1 at 9 24:0

17:0 iso 23:0 20H

summed feature 9:
18:2 cis 9, 12; 18:0 anteiso

summed feature 10:
18:1 cis 11; 18:1 trans 9; 18:1trans6

27

 

 

C0111 1‘

relate
can b
“V0

differ
AWC
comn
is 815.
and p
micm

resmr

Most
3C1'035
PrOdu
COmn.
Comm
Chang
along

111% El

gradient. I have two lines of evidence that support the idea that soil microbial
community respiration (or biomass) increases with plant productivity at this site. First,
I detected an increase in AWCD of the Biolog plates in relation to productivity (Figure
2.2C). Although Biolog AWCD is not a direct measure of respiration, it is strongly
related to inoculum density (Garland and Mills 1991, Haack et al. 1995) and, as such,
can be interpreted as an indicator of total number of bacteria (biomass). Conversely,
two wells with the same inoculation density may differ in AWCD because of
differences in microbial respiration (Konopka et al. 1998). In either case, the higher
AWCD in the sites at the base of the hill indicates a more productive microbial
community and this corresponds to areas along the gradient where the plant community
is also the most productive. This is consistent with the higher amounts of N, moisture,
and plant biomass at the base of the hill, which should make more C available to the
microorganisms. Secondly, the modiﬁed SIR analysis indicates higher rates of

respiration at the base of the hill where productivity was highest (Figure 2.2D,E).

Most studies that have reported changes in soil microbial community composition
across community types have sampled sites that differed in plant species composition,
productivity, and soil type. From these studies, it is unclear whether the plant
community or the underlying edaphic factors are inﬂuencing the soil microbial
community structure. For example, J. C. Zak et al. (1994) used Biolog to investigate
changes in functional diversity of the soil microbial community from grasslands located
along an elevational and moisture gradient in New Mexico. They found differences in

the Biolog proﬁles of the soil microbial community from six distinct plant communities

28

 

 

dong
sues.
[ﬂant
Goodl
cmnm
LautC<
edaph

Bkﬂcu

SCVCTL
acnls]
dlSIlnE
1n the
(180551

mICrok

along this gradient. However, because soil characteristics also varied among these
sites, it is not clear whether the differences in Biolog proﬁles were due to changes in
plant community composition, edaphic factors, or some other variable. Similarly,
Goodfn'end (1998) found that Biolog distinguished among the soil microbial
communities of eight sites representing a variety of wetlands in the southwestern
United States. However, it was not clear whether plant community composition or
edaphic characteristics were more important in inﬂuencing the grouping of those

Biolog proﬁles into habitat types.

Several authors have argued that phospholipid fatty acids (PLFA’s), a subset of fatty
acids present in the phospholipid membrane, may provide a more sensitive indicator to
distinguish among microbial communities. Phospholipid fatty acids break down easily
in the soil and are thus thought to represent the active soil microbial community
(Bossio and Scow 1998). Zelles et al. (1992) used PLFA proﬁles to compare soil
microbial community patterns in grassland and agricultural ﬁelds under different
management regimes and found that proﬁles differed among the different ﬁelds, but
they did not distinguish between plant community and edaphic effects. Bossio et al.
(1998) concluded that soil type has stronger effects on the soil microbial community
structure than plant community type. They found that the addition of a cover crop (an
increase in plant community diversity over time) was less inﬂuential in changing PLF A
proﬁles than soil type. The differences in edaphic characteristics at this site, although

substantial, were not as striking as differences between soil types would be.

29

 

There are several possible reasons why I did not detect changes in the soil microbial
community composition along this gradient: (1) there is no connection between the
structure of the soil microbial community and the soil characteristics, plant diversity, or
plant productivity; (2) the soil microbial community structure is very stable and
affected mainly by factors like long-term plant community composition or historical C
inputs to the soil; (3) the spatial scale or time of year I sampled was inappropriate for
detecting differences in the soil microbial community; or (4) the techniques I used to
assay the soil microbial community were not speciﬁc enough to detect what differences

were there.

The ﬁrst two reasons seem unlikely because there should be a linkage between the
microbial (consumer) community and the resources (plant carbon) that they utilize
(Paul and Clark 1996). Much of the carbon available to soil microorganisms is being
provided to the soil microorganisms each year by the extant plant community, and
although this is a successional community, the plant community composition at this site
has remained stable for the past decade (K. L. Gross, personal communication). Even
if soil microbial community structure is not affected by plant community composition,
increasing plant diversity or productivity should provide additional resources to the
extant soil microbial community and thus inﬂuence soil microbial community
composition. Additionally, past agricultural use at this site likely would have depleted
soil C (Drinkwater et al. 1998, Robertson et al. 1993), and therefore made the current
community inputs of C important in determining the structure and activity of the soil

microbial community.

30

 

 

 

It is
NEW:
temp
chos
plant
reve;
mid-1
samp
carbc

rhizo

It is p
“ﬁght
Westr
T111205
and C
01 rhi
Hon-e
5133113
more

Produ,

d1 1TB 1' c

It is difﬁcult to know if sampling at a different time of year or spatial scale would have
revealed associations between the plant and microbial communities at this site. Both
temporal and spatial scales are important in the observation of ecological phenomena. I
chose to sample in mid-summer on the assumption that at this time of year both the
plant and soil microbial communities would be most active. Other researchers have
revealed associations between plant and microbial communities using soil sampled in
mid-summer. Bossio et al. (1998) detected differences in PLFA patterns of soil
sampled in July from different agricultural treatments in California. Similarly, using
carbon source utilization patterns, Westover et al. (1997) differentiated among

rhizosphere soils sampled in August from several grass species in Washington.

It is possible that if this sampling had been done at a smaller, more ﬁne-grained scale I
might have detected associations between microorganisms and speciﬁc plant species.
Westover et al. (1997) detected differences among soil microbial communities of
rhizosphere soils of several grass species in both the ﬁeld and greenhouse. Grayston
and Campbell (1996) used Biolog to differentiate between the microbial communities
of rhizosphere soils from two tree species, Larix eurolepis and Picea sitchensis.
However, others have found associations between plant and microbial communities at
spatial scales similar to the scale used in this study. Plant community composition is
more likely to affect soil microbial community composition than plant diversity or
productivity. A recent experimental study by Wardle et al. (1999) did detect

differences in PLFA composition of the soil microbial community that were

31

 

The
hnnta
there
to dist
used 1
1995;
lechni

Closet:

signiﬁcantly related to the plant removal treatments. This suggests that 3-4 years of
abandonment is sufﬁcient to detect changes in the soil microbial community.
Broughton et al. (2001) saw a similar relationship between plant community
composition and PLFA patterns of the soil microbial community at the Silwood Park

BIODEPTH site after three years (results presented in Chapter 3).

The inadequacy of tools to assess microbial diversity has been a long-standing
limitation to this understanding of soil microbial communities (Tiedje 1995). While
there are clearly limitations to the ability of ﬁmctional tools such as Biolog and PLF A
to distinguish among microbial communities, as noted above, a number of studies have
used these tools to successfully differentiate among communities (Zelles et al. 1992,
1995; J. C. Zak et al. 1994; Goodfriend 1998). The development of molecular
techniques more sensitive to shifts in composition may reveal natural shifts from one
closely related microorganism to another along a gradient, just as there are shifts among
closely related plant species along gradients. Additionally, molecular techniques may
allow us to better address the roles of dominance and plasticity in structuring soil

microbial communities.

The correlative nature of this study does not allow us to determine what factors may
underlie the observed variation in the soil microbial community at this site. Despite our
expectation that there should be a close association between the plant community and
the soil microbial community, at this spatial scale (within a site), using these tools, I

were not able to detect any association between plant community composition and soil

32

 

micr
and
that

patte

microbial community composition. The similarity between patterns of plant biomass
and soil microbial respiration is intriguing, however, and suggests that the resources
that limit each of these communities co-vary. In contrast, the differences between
patterns of plant diversity and soil microbial community structure suggest that different

mechanisms are responsible for structuring diversity in these associated communities.

33

 

Brou
Buye
Cavh

h
\

Dnnk

Gaﬂa

GOOd]

Bra-VS

Haack

REFERENCES

Bligh E. G., W. J. Dyer. 1959. A rapid method of total lipid extraction and puriﬁcation.
Canadian Journal of Biochemistry and Physiology 37: 911-917

Bossio, D. A. and K. M. Scow. 1998. Impacts of carbon and ﬂooding on soil
microbial communities: Phospholipid fatty acid proﬁles and substrate utilization
patterns. Microbial Ecology 35: 265-278.

Bossio, D. A., K. M. Scow, N. Gunapala, and K. J. Graham. 1998. Determinants of
soil microbial communities: effects of agricultural management, season, and soil
type on phospholipid fatty acid proﬁles. Microbial Ecology 36, 1-12.

Broughton, L.C., K.L. Gross, and A. Hector. 2001. Linking plant community diversity
to soil microbial communities: an experimental evaluation from the BIODEPTH
experiment. Journal of Ecology (submitted).

Buyer, J. S. and L. E. Drinkwater. 1997. Comparison of substrate utilization assay and
fatty acid analysis of soil microbial communities. Journal of Microbiological
Methods 30: 3-11.

Cavigelli M. A., G. P. Robertson, and M. J. Klug. 1995. Fatty acid methyl ester
(FAME) proﬁles as measures of soil microbial community structure. Plant and
Soil 170: 99-113

Drinkwater L. E., P. Wagoner, and M. Sarrantonio. 1998. Legume-based cropping
systems have reduced carbon and nitrogen losses. Nature 396: 262-265

Garland J .L. and A. L. Mills. 1991. Classiﬁcation and characterization of heterotrophic
microbial communities on the basis of patterns of community-level sole-carbon-
source utilization. Applied and Environmental Microbiology 57: 2351-2359

Gleason H. A. and A. Cronquist. 1991. Manual of Vascular Plants of Northeastern
United States and Adjacent Canada. The New York Botanical Garden, New
York.

Goodﬁiend W.L. 1998. Microbial community patterns of potential substrate utilization:
a comparison of salt marsh, sand dune, and seawater-irri gated agronomic
systems. Soil Biology and Biochemistry 30: 1169-1176

Grayston SJ. and C. D. Campbell. 1996. Functional biodiversity of microbial
communities in the rhizospheres of hybrid larch (Larix eurolepis) and Sitka
spruce (Picea sitchensis). Tree Physiology 16: 1031-8

Haack, S.K., H. Garchow, M.J. Klug, and L. J. Fomey. 1995. Analysis of factors
affecting the accuracy, reproducibility, and interpretation of microbial
community carbon source utilization patterns. Applied and Environmental
Microbiology 61: 1458-1468

Hassink J. 1993. Relationship between the amount and the activity of the microbial
biomass in Dutch grassland soils: Comparison of the fumigation-incubation
method and the substrate-induced respiration method. Soil Biology and
Biochemistry 25: 533-538

34

Huberty L. E., K. L. Gross, and C. J. Miller. 1998 Effects of nitrogen addition on
successional dynamics and species diversity in Michigan old-ﬁelds. Journal of
Ecology 86: 794-803

Hutchinson GE. 1959. Homage to Santa Rosalia or Why are there so many different
kinds of animals? American Naturalist 93: 145-159

Konopka A., L. Oliver, and R. F. Turco, Jr. 1998. The use of carbon substrate
utilization patterns in environmental and ecological microbiology. Microbial
Ecology 35: 103-115

Ohtonen R., S. Aikio, and H. Wire. 1997. Ecological theories in soil biology. Soil
Biology and Biochemistry 29: 1613-1619

Paul, EA. and FE. Clark. 1996. Soil Microbiology and Biochemistry. 2nd ed. San
Diego, CA: Academic Press, Inc.

Robertson G. P., J. R. Crum, and B. G. Ellis. 1993. The spatial variability of soil
resources following long-term disturbance. Oecologia 96: 451-456

Rosenzweig, ML. 1995. Species Diversity in Space and Time. Cambridge University
Press, Cambridge.

Schléipfer, F. and B. Schmid. 1999. Ecosystem effects of biodiversity: a classiﬁcation

of hypotheses and exploration of empirical results. Ecological Applications 9:
893-912.

Tiedje, J. M. 1995. Approaches to the comprehensive evaluation of prokaryote
diversity of a habitat, pp. 73-87, In Microbial Diversity and Ecosystem Function
(eds D. Allisopp, R.R. Colwell, and D.L. Hawksworth). CAB International,
Wallingford, UK.

Tunlid, A., and DC. White. 1992. Biochemical analysis of biomass, community
structure, nutritional status, and metabolic activity of microbial communities in
soil, pp. 229-262, In Soil Biochemistry, vol. 7 (eds G. Stotsky and J .M. Bollag).

Wardle, D.A., K.I. Bonner, G.M. Barker, G.W. Yeates, K.S. Nicholson, R.D. Bardgett,
R.N. Watson, and A. Ghani. 1999. Plant removals in perennial grassland:
vegetation dynamics, decomposers, soil biodiversity, and ecosystem properties.
Ecological Monographs 69: 535-568.

Westover, K. M., A. C. Kennedy, and S. E. Kelley. 1997. Patterns of rhizosphere
microbial community structure associated with co-occurring plant species.
Journal of Ecology 85: 863-873.

Zak D. R., D. R. Gri gal, S. Gleeson, and D. Tilman. 1990. Carbon and nitrogen cycling
during old-ﬁeld succession: constraints on plant and microbial biomass.
Biogeochemistry 11: 111-129

Zak D. R., D. Tilman, R. R. Parmenter, C. W. Rice, F. M. Fisher, J. Vose, D.
Milchunas, and C. W. Martin. 1994. Plant production and soil microorganisms

in late-successional ecosystems: a continental-scale study. Ecology 75: 2333-
2347

35

Zak, J. C., M. R. Willig, D. L. Moorhead, and H. G. Wildman. 1994. Functional
diversity of microbial communities: A quantitative approach. Soil Biology and
Biochemistry 26: 1101-1108.

Zelles, L., Q.Y. Bai, T. Beck, and F. Beese. 1992. Signature fatty acids in
phospholipid fatty acids and lipopolysaccharides as indicators of microbial

biomass and community structure in agricultural soils. Soil Biology and
Biochemistry 24: 317-323.

36

CHAPTER 3

LINKING PLANT COMMUNITY DIVERSITY TO SOIL MICROBIAL
COMMUNITIES: AN EXPERIMENTAL EVALUATION FROM THE BIODEPTH
EXPERIMENT
These results have been submitted to Journal of Ecology in an article: Broughton, L.C.,
K.L. Gross, and A. Hector. 2001. Linking plant community diversity to soil microbial

communities: an experimental evaluation from the BIODEPTH experiment. Journal of

Ecology (submitted).

Introduction

Most studies to date investigating the relationship between species diversity and
ecosystem function have focused on how changes in primary producers and consumers
affect ecosystem processes (Schlapfer and Schmid 1999, Rosenzweig 1995). The
interaction between above-ground and below-ground (soil) communities in mediating
these processes has been less studied (Schlapfer and Schmid 1999, Ohtonen et al.
1997). While considerable attention has been paid to factors that affect the composition
and function of communities of macroorganisms, very little is known about the factors
that affect the structure of soil microbial communities (Ohtonen et al. 1997, Tiedje

1995).

Soil microorganisms play an essential role in the functioning of terrestrial ecosystems

because the soil microbial community provides key steps in the cycling of nutrients

37

 

 

 

31;

prc

pen
C011.
0111.
may
or d
llllCr.
C0mr

€COS}

T0 dc
C0mm

$thm

(carbon, nitrogen, and phosphorus, among others) through the ecosystem (Paul and
Clark 1996). In most temperate grassland and forest systems, plant growth is limited
by nitrogen (or a combination of nitrogen and phosphorus, Shaver and Chapin 1980,
Schmidt et al. 1997, Jonasson et al. 1999). The soil microbial community controls the
release of inorganic nitrogen to plants; however, the soil microbial community is most
often limited by carbon. Therefore, the rate at which limiting nutrients are made
available to plants is likely to be inﬂuenced by both the amount and quality of carbon

provided by plants and available to soil microorganisms.

Because plant species differ in carbon content and quality, plant species identity has the
potential to affect nutrient process rates through litter quality effects, which
consequently affect the soil microbial community (Paul and Clark 1996, Wardle and
Giller 1996). As a result, the composition and productivity of the plant community
may inﬂuence the soil microbial community. Similarly, differences in the productivity
or diversity of the plant community may be affected by processes mediated by soil
microorganisms (e.g. N-mineralization rates). Consequently, changes in the plant
community and the resulting change in the soil microbial community potentially affect

ecosystem function.

To determine the effect of plant community diversity on soil microbial community
diversity and processes, I studied the soil microbial community in experimental plant
communities at the BIODEPTH site in Silwood Park, UK, where plant community

structure and diversity were experimentally manipulated. I asked the following

38

questions: (1) How does plant community diversity affect the structure of the soil
microbial community? (2) How does plant community composition affect the structure
of the soil microbial community? (3) How does plant productivity affect the
relationship between plant community diversity and the soil microbial community? and
(4) Do speciﬁc functional groups or plant species have detectable effects on the soil

microbial community?

Materials and Methods

Site

The study was conducted at the Imperial College site at Silwood Park, Ascot, UK
(National Grid Reference 51°22’N, 00°37’W) and was part of the BIODEPTH
experimental network of sites (BIODiversity and Ecosystem Processes in Terrestrial
Herbaceous systems, Hector et al. 1999). The site was previously used for horse-
grazing and has sandy-loam soil with an average pH of 5.26. In Fall 1995, the ﬁeld
was fenced, herbicided (Round Up, Dow Elanco), and tilled (Hector et al. 2000). The
soil was fumigated in April 1996 with methyl bromide (Check Fumigation Ltd.,
Reading, UK) to remove the soil seed bank. Fumigation should also have killed much
of the soil microbial community. In May 1996, two replicate blocks each with 33 plant
assemblages, plus no-plant controls, were established in 2 x 2 m plots. The
assemblages consisted of different combinations of plant species that varied in species

richness and ﬁrnctional group richness (Hector et al. 1999). The species sown were all

39

 

herbaceous perennials representative of grassland species common to this part of
England. Maximum species richness was 11 species per 2 m x 2 m quadrat, the
average number of species in the area. Plant species were classiﬁed into one of three
functional groups: legumes, non-leguminous forbs, and grasses. To address concerns
about individual species effects on species richness curves (see Huston 1997), species
thought to have strong effects on productivity were included in all mixtures. To
minimize plant biomass effects, all mixtures included at least one grass species. This
limits our ability to detect the effects of individual plant species and to evaluate the
effects of grasses. Plant assemblages were maintained by hand-weeding for all 4 years
of the experiment. Several undisturbed reference plots were also established adjacent to
the manipulated plots. Peak plant biomass was clipped 5 cm above-ground level both

years to provide an estimate of annual net primary productivity (Hector et al. 1999).

Soil Sampling and Analysis

To determine the relationship between plant diversity and the soil microbial
community, I selected a subset of plots, encompassing the full range of species and
functional group diversity, to sample for soil and microbial characteristics. In October
of year 3 of the experiment (1998) I sampled soils from 28 plots (2 replicates of 14
different plant compositions, Table 3.1). The following year (September 1999, year 4),
I sampled soils from 36 plots (2 replicates of 18 different plant compositions, Table
3.1). These included 12 of the 14 plots sampled in year 3, plus an additional 6 plots to

expand the coverage of the species richness gradient. 1 were unable to re-sample the

40

 

 

Tab
and
$8.111I
lo:
ih bl
the -
pres
(a)
£58"
:1 gl‘r’
..~1 / 0f:
prartl
Ant/i
odor.
Arr/r
clam
C 1 Hr
Cl‘lSlr.
Dar! 1
F8511!
H()/( I.
Liam“.
P/l/t’li’
Trr’sor
ﬂares.

 

Table 3.1. (A) Plant species pool used in establishing the experimental communities
and (B) the plant communities sampled for the soil microbial community. Plots
sampled only in year 3 (1998) of the experiment are in italics. The control and Rumex
plots were not maintained through year 4 and could not be re-sampled. Plots sampled
in both years are indicated in bold. Other plots were sampled only in year 4 (1999) of
the experiment. Communities are grouped by number of plant functional groups (F G)

 

 

present.

(a)

Grasses Abbrev. Legumes Abbrev. Forbs Abbrev.

Agrostis capillaris AgC Lotus LC Achillea AM

Alopecurus AP corniculatus millefolium

pratensis Medicago ML Cerastium CF

Anthoxanthum AO lupulina fontanum

odoratum T rifoli um TR Hypochaeris HR

Arrhenatherum AE repens radicata

elatius T rifoli um TP Plantago PL

Cynosurus CC pratense lanceolata

cristatus Vicia hirsuta VH Potentilla PE

Dactylis glomerata DG V icia saliva VS erecta

F estuca rubra FR Vicia VT Rumex acetosa RA

Holcus lanatus HL tetrasperma Stellaria SG

Luzulla campestris LC graminea

Phleum pratense PhP T araxacum TO

T risetum TF oﬂicinale

ﬂavescens Veronica VC
chamaedrys

 

41

 

Table 3.1 (cont’d).

 

 

 

 

(b) E E E
Trt 0 g Trt 1 FG g Trt 2 FG g Trt 3 FG
Code PG 5 Code i Code 3 Code
Co co {1 AB 1 15 AE, LC lRer reference
n- E 2 FR E 1 plots
tro 5 3 HR 3 16 AE, FR, LC, :
ls g 4 RA g TR g 11 AE, TR,
i 5 LC 1 5 RA, HR
56 TR 3 18 AgC, AE, FR, 5
5 5 HL, AM, HR, 5 12 FR,AE,
:13 AgC, FR, ; RA, PL ; LC, TR,
g HL, AE g ; HL, RA,
1 a 17 AgC, AE, FR, 5 PL, AM
3 14 AgC,AE, 3 HL, TR, LC, 3
5 FR, HL, 2 TP, vs 5 19 HL, AgC,
§ AP, A0, 2 3 LC, TR, FR,
g CC, TF g 8 AgC, AE, ; PL, RA, HR
5 i FR.HL,RA, i
5 7 AgC,AE, 3 PL, AM, so, 310 AgC,AE,
; FR,HL, ; VC, PE, TO 5 FR,HL,
E AP,AO, 3 : LC,TR.
g CC, TF, g 9 AgC, AE, 5 AM, CF,
5 DG, PhP, 1 FR, HL, TR, 5 HR, PL,
3 LZ 3 LC, TP, vs, 5 RA
; ; ML, VH, VT 3
42

control (no plant) and Rumex monoculture plots because the plots were not maintained
in the fourth year. In year 4, I focused on a subset of the soil and microbial variables.
In both years the sampled plots included undisturbed reference plots and combinations

of 4, 8, and 11 species varying from 1 to 3 functional groups.

I sampled soil to a depth of 10 cm, then sieved the sample through a 3.35 mm sieve,
and stored it in sealed plastic bags at 4°C until analyzed. All analyses were done at the
WK. Kellogg Biological Station of Michigan State University within 3 days of

sampling.

I determined gravimetric soil moisture for each sample by drying 10 g soil at 105°C for
48 hours (Nelson and Sommers 1982). A subsample of the dried soil was ashed at
500°C for 4 hours to determine organic matter content (Nelson and Sommers 1982).
Soil pH was determined using a Corning pH meter 420 after mixing 5 g of air-dried soil
in 50 ml millipure H20. For nitrogen analyses, I extracted 20 g of fresh soil in 100 ml
1M KCl. The samples were shaken for 1 min and allowed to settle for 24 h at room
temperature. The supernatant mixture was ﬁltered through a l-um Gelman glass-ﬁber
ﬁlter and NO3' and NH4+ concentrations were measured using Alpkem auto-analyzer.
To determine potential N-mineralization and nitriﬁcation rates, a companion 20 g
sample was incubated for 21 days at 25°C and 15% humidity and then extracted using

the same methods as above.

43

 

1 determined microbial biomass using the chloroform fumigation incubation method
(Paul et al. 1999b). Two 25 g soil samples were pre-incubated for 5 days then one
sample was fumi gated with chloroform for 24 hours to kill the microorganisms. After a

vacuum was created and the chloroform evaporated, 0.5 g of original soil was added to

both samples. I measured initial headspace C02 and accumulated C02 after 10 days on

an ADC series EGA infrared C02 gas analyzer (The Analytical Development Co. Ltd.,
Hoddesdon, Herts., UK). I calculated microbial biomass as [ 1.73 * (10 day

accumulated COz-C — initial COz-C for the fumigated samples) — 0.56 * (10 day

accumulated COz-C — initial COz-C for the control samples)] (Paul et al. 1999b). To
determine microbial respiration I used a separate set of 10 g soil samples that were pre-

incubated 5 days in a 160 ml glass qorpak bottle. I measured initial headspace CO; and

accumulated C02 after 1 and 5 days and calculated the rate of COz-C respired per day.

Community-level physiological proﬁles (CLPP) were determined using both Biolog
and Ecolog plates (Biolog, Inc., Hayward, Calif, USA). For both assays, 1 g of fresh,
sieved soil was shaken with 99 ml 1% phosphate buffer solution for 20 min. 150 1.11 of
the mixture was transferred to each well of the microtiter plate (GN Biolog, 95 Carbon
sources + 1 non-Carbon control; or Ecolog, 3 replicates of 31 Carbon sources + 1 non-
Carbon control). The plates were incubated in the dark at 25 °C and optical densities
were measured at 14 h intervals from 0 h to 64 h using an Emax precision microplate

reader (Molecular Devices Corp., Menlo Park, Calif, USA). Because the 5 incubation

44

 

tin:

165'.

CW:

 

 

(VarI
polar
phos

met

pl

0\

times gave consistent results and the Biolog and Ecolog plates were similar in their

results, I present here data only from the 64 h Ecolog measurements.

For the PLFA analysis, I extracted lipids from 6 g whole soil samples for 2 h using a
mixture of dichloromethane (DCM): methanol: phosphate buffer (1:2:0.8 v/v/v),
following a modiﬁed Bligh-Dyer procedure (Bligh and Dyer 1959). Phase separation
was achieved by adding DCM and saturated sodium bromide solution (1:4 v/v). I
isolated the phospholipid fatty acids from the dried lipid extracts by solid phase
extraction. The lipid material was added to a polar column consisting of 100 mg silica
(Varian Bond Elut LRC Columns, Product # 1211-3010). Lipids of low or intermediate
polarity were eluted with chloroform and acetone and discarded. Subsequently,
phospholipid fatty acids were eluted with 1.5 m1 methanol for preparation of fatty acid
methyl esters. I saponiﬁed the samples using 1 m1 NaOH (15% w/v) in methanol (50%
v/v) at 100°C for 30 min and methylated the samples with 2 m1 6M HCl in methanol at
80°C for 10 min. I extracted the fatty acid methyl esters into 1.25 ml (1:1 v/v) methyl-
tert-butyl etherhexane for 10 min and washed the extract with 3 ml 1.2% NaOH.
Phospholipid amounts were measured using a HP 5890 series 11 gas chromatograph
(Hewlett Packard Co., Palo Alto, Calif, USA) equipped with a 7673 autosampler and
ﬂame ionization detector (Microbial ID Inc., Newark, Del., USA). Peaks were
identiﬁed by comparison with an external standard. I performed all analyses on the
phospholipid fatty acid proportions of the total peak area to correct for differences in

overall peak area.

45

 

 

ator
C it
indi.
unlo
the 1

acid:

 

numl

Star

13;.

in

I describe fatty acids using standard nomenclature where the total number of carbon
atoms appears before the colon and the total number of C-C bonds appears after it.
Cyclo-propane analogs are indicated by "cyclo," and the location of the epoxy bond is
indicated by a "c" followed by two numbers. If the cis or trans conﬁguration is

unknown, the word "a " is used. The number following "cis, trans," or "a " indicates
the location of the double bond in relation to the carboxyl end of the molecule. Fatty
acids with the same retention time are grouped as "sum in feature" and given a unique
number designation. For analysis, I included only those phospholipid fatty acids that
were present in all samples and reported their abundance as the proportion of the total

phospholipid fatty acid amount in each sample. Of the 25 lipids detected, 16

phospholipid fatty acids met this criterion in both years.

Statistical Analysis

Using an analysis of variance (ANOVA) model and sequential (Type 1) sums of
squares, I tested the effect of number of species (richness), number of functional
groups, block, and plant community composition (MIXTURE) on the following
response variables: pH, soil moisture, soil organic matter, total N, N-mineralization
rate, nitriﬁcation rate, number of culturable bacteria, microbial respiration, microbial
biomass, CLPP, and PLFA proﬁles (Table 2). The effect of number of species was
tested separately from the effect of number of functional groups; therefore, both tests
were non-conservative. Hector et al. (1999) found that aboveground plant biomass

increased with increased numbers of plant species at this site (and most of the other

46

 

1211.
the
par.

Son
11.:
810
Oh i
1111'
Bio
[111
lune
legit

F

P111): I
Plan‘
8101‘
Dit't‘
Mitt
ago

 

 

 

Table 3.2. Analysis of variance model with sequential sums of squares used to evaluate
the effects of the plant community manipulations on soil and microbial community
parameters.

 

 

Source of variation Mean square Variance ratio
Main GLM model (ANOVA):

BIOCK M83 M83/ MSBtM
Diversity MSD MSD / MSM
Mixture MSM MSM / MSBI-M
Block*Mixture MSWM --

[Where Diversity is 1) species diversity, 2)
functional group diversity, 3) presence/absence of
legumes, or 4) presence/absence of forbs]

Plant Biomass as a covariate (ANCOVA):

Plant Biomass (covar) MSC MSC / MSM
BIOCK M83 M83/ MSBt-M
Diversity MSD MSD / MSM
Mixture MSM MSM / MSBtM
Block*Mixture MSB-M --

 

BIODEPTH sites, species number, p<0.001; functional group number, p<0.01).
Consequently, I used aboveground plant biomass as a covariate in these analyses to

investigate the inﬂuence of plant diversity independent of plant biomass effects (Table

3.2).

I used Principal Components Analysis (PCA) to determine if there was an underlying
structure to the soil microbial community as detected by CLPP and PLFA. I used these
principal component axes as response variables in the ANOVA and ANCOVA to test
for main treatment effects of plant species richness, functional group richness, and

community composition on soil microbial community metabolic activity and structure.

47

Results

Legacies of Site Preparation

The effect on the soil microbial community of soil fumigation with methyl bromide
during site preparation is clearly illustrated by the soil microbial biomass data. In both
years, the undisturbed (non-fumigated) reference plots had the highest microbial
biomass, as much as four times higher than the manipulated plots (Figure 3.1). Even
after 4 years (1999 sampling), the biomass of the soil microbial communities had not

recovered from the disturbance effect of fumigation (Figure 3.1B).

Effects of Species Richness and Functional Group Richness Treatments on the Soil

Microbial Community

The initial analyses of these data excluded aboveground plant biomass as a covariate
and I found little effect of species richness or functional group richness on soil or
microbial community parameters (Table 3.3A). There were a few exceptions. In year

3, the numbers of colony forming units (CFU’s) after 48 hours increased at higher

species richness (Figure 3.2A, r2=0.29, p< 0.01). Similarly, number of CFU’s after 48

hours increased at higher functional group richness (Figure 3.2B, r2=0.29, p< 0.01).

However, neither of these relationships was signiﬁcant when aboveground plant
biomass was used as a covariate in the year 3 analysis (Table 3.3A). The number of

culturable bacteria was signiﬁcantly correlated with aboveground plant biomass in year

48

9.35-53: 3:: 33.35 atmcﬁombﬂ 3:: 35 $07. .8353: 3 3:ng :5 33:6 3 @8365 2:
85:28:02 28m 3 cap—omega: 8: 33:36:50 00:20.3: 333358-:0: 223 .286 x52: : E “68:88:“: 2: 32:
65:8 22902 .m:.m 038. Eat 2: 3:3 5:85:52 €55.88 Ema .mm _ H :32 2: 82:5 .353 Imam—DOE
fem vooB=m of no v SPA Am: 2.: m :8» A5 E 3:62: 3595:: :8 :0 ﬂoat» 52895: $53888 Ema ._.m “:sz

arm—.5923? $3.2m m N o m m N 5mm

 

+ﬁ45511ﬁ71___

: :

a m
A. m 4
.W m - m

LF.

 

 

 

 

 

ﬁﬂﬁﬁﬂﬁl

m

1&1

.L

I

J

Zoom-
.oow-
.o
.oow
CON
oom
.oov
com
com

 

([Ios Kip 8 /;) 8n) ssequIq [etqorotw

NF :owm m N. o m e m Ntomoo

 

ﬁﬁﬁﬁﬁr

 

4-

 

 

 

0

OO
O
\—

OF.

.IN

0) no N 8 L0 V
([105 Kip 3 /3 Bit) sseurorq lerqomr

49

Table 3.3. Signiﬁcance and direction (for diversity) of treatment effects on soil and
microbial parameters for both years as detected by Analysis of Variance using Type I
sums of squares (model in Table 3.2) for (a) species diversity, functional group
diversity, and composition and (b) presence/absence of legumes or forbs. ANCOVA
results using above-ground plant biomass as the covariate are shown in parentheses if
the effect changed in signiﬁcance. NS=not signiﬁcant

 

 

(a) Plant Species Diversity F G Diversity Composition
Variable: Biomass

Year 3

pH <0.05 NS NS < 0.05 (< 0.05)
Soil moisture <0.10 + < 0.01 (< 0.05) NS NS

Soil organic matter NS NS NS NS
Total N NS NS NS NS
N-mineralization NS NS NS NS

rate

Nitriﬁcation rate NS NS NS < 0.05 (NS)
Culturable bacteria + <0.05 + < 0.05 (NS) + < 0.05 (NS) NS
Microbial NS NS NS < 0.01 (< 0.01)
respiration

Microbial biomass NS NS NS NS
CLPP PC] + <0.01 + < 0.05 (NS) + < 0.05 (< 0.10) NS
CLPP PC3 NS NS + < 0.10 (<0.05) NS
PLFA PCl - <0.05 NS NS NS
PLFA PC2 NS - <0.05 (<0.10) NS NS

PLF A PC4 NS NS NS <0.05
Year 4

Soil moisture NS NS NS NS
Total N + <0.10 - <0.10 (NS) NS NS
N-mineralization NS NS NS NS

rate

Nitriﬁcation rate NS NS NS NS
Microbial NS NS NS NS
respiration

Microbial biomass NS NS NS NS
PLFA NS NS NS NS

 

50

Table 3.3 (cont’d).

 

 

(b) Variable: Plant Biomass +/- Legumes +/- Forbs
Year 3

pH NS NS NS

Soil moisture + <0.05 NS + <0.10 (<0.10)
Soil organic matter NS NS NS

Total N NS NS NS
N-mineralization rate NS NS NS
Nitriﬁcation rate NS NS NS
Culturable bacteria + <0.01 NS + < 0.05 (NS)
Microbial respiration NS NS + < 0.05 (<0.05)
Microbial biomass NS NS + < 0.05 (<0.05)
CLPP PC4 NS + < 0.05 (0.10) + < 0.10 (< 0.10)
CLPP PCS NS NS - < 0.01 (<0.01)
PLF A NS NS NS

Year 4

Soil moisture NS NS NS

Total N NS + <0.05 (<0.05) NS
N-mineralization rate NS + <0.05 (<0.10) NS
Nitriﬁcation rate NS NS NS
Microbial respiration NS NS + <0.10 (<0.10)
Microbial biomass NS NS NS

PLFA NS NS NS

 

Table 3.4. Fatty acids used in principal components analysis of PLF A proﬁles.

 

Phospholipid Fatty Acids

 

C9 Dicarboxylic acid
14:0

15:0 iso

15:0 anteiso

16:0 iso

16:1 cis 9

16:1 cis 11

16:0

17:0 iso

17:0 anteiso

17:0 cyclo

18:2 cis 12

18:1 cis 9

18:0

19:0 cyclo c11-12
summed feature 8: 18:1 trans 9

 

51

.339: 2: 36on 50: 5:3 32: 03: of. .Ewm: on: 9 :30: wEEEm 3.38.53: 02:: 88:8:
atmgoomxt 02:: :9: 32a ”Em: on: o: a: @352: 358.53: 96; 2:339:58 32:: 3:: 8...: 32m .mo_w::_:: B 3:238 28
83:6 :3 38:65 0:: 85:30:32 9:3 3 325850: 0:: mo_:_::EEo: 858.3: “BEE—€278: 223 520:: :23: E 820850:
0:: 82: .0550 233-02 .Codvq .mnmdnmmv m :3» E 3:805 :52: 33833:: ADV cc: .Codv: .wwmdnemv 820% 3:285: ES:
:0 :onEz: Am: fodva .vomdummv 33:2: 36on :52: A<V :5: 3:208 :8 2:83:20 :0 :opEzc 503:0: 9:32:33: .N.m Saw:—

ANEEV 39:05 E2: :::O:wo>o:<
cog 8o— oow coo co: com o

anew 3:285: mo :38: Z

 

 

860% Ema? 82:32

 

 

m N _ o N: w e o
_ ~ ~ — i1 — Mul O _ q _ _lO _ q _ :10
a o W 4 o N o
w 4 00— w .1 4 O m 0
O l
e e. e )m m 18:)m m 18:)
4 co 3 4 . J 3
4 o 13 ma 4 ® mm 0 a 4 o 13
q 1:8 n m u n .1... n
H m. . 108x w e . roamx
4 0 u 4 I o l.
‘ 9mm. ‘ 0. m. a Or.)
18: (m a... m (
a w room w room
mm 4 ova
u 1:9. u.
m 4:8 < 18:.

 

 

1131; re saguoloojo iaqum N

52

3 (F 1;
numk
With }
I
numb

when

plots

E ffcct

Becau
0f prc
pararm
moistu
year 3.

diSIingl

 

0f ’egUﬂ

and N- j

The Sm;

Very few

3 (Figure 3.2C, r2: 0.57, p< 0.01). PLF A PC2 was also signiﬁcantly related to species

number (Table 3.3A). Soil moisture and CLPP PCl were both positively correlated
with plant species number, while CLPP PCI and PC3 were positively correlated with
number of ﬁmctional groups (Table 3.3A). These relationships were still signiﬁcant
when plant biomass was included as a co-variate; however, when the no-plant control

plots were excluded from the analyses, the relationships were not signiﬁcant.

Effects of Functional Groups or Individual Species on the Soil Microbial Community

Because grass species were present in all mixtures, I were only able to test for the effect
of presence/absence of forbs and legumes on the soil and microbial community
parameters (Table 3.2). The presence of forbs had a signiﬁcant positive effect on soil
moisture, number of culturable bacteria, microbial respiration, microbial biomass in
year 3, and microbial respiration in year 4 (Table 3.38). CLPP PC4 and PCS also
distinguished among plots with and without forbs in year 3 (Table 3.3B). The presence
of legumes inﬂuenced CLPP PC4 in year 3 and corresponded with increased Total N

and N-mineralization rates in year 4.

The small number of plots sampled only allowed us to evaluate the direct effects of a
few species on the soil microbial community; most species were either in all or only a
very few mixtures. I were able to detect the effects of two species, Lotus corniculatus
(legume) and Hypochaeris radicata (forb), on soil and microbial parameters. Soil from

L. corniculatus monocultures had a signiﬁcantly lower pH than all other plots. The L.

53

corniculatus monocultures also were distinguished from the other plots by having
higher values for CLPP on the third PC (Figure 3.3), indicating the soil microorganisms
in these plots were better able to metabolize a-D-lactose. Hypochaeris radicata
monocultures had the highest soil microbial biomass of manipulated plots in both years
(Figure 3.1A,B), suggesting a strong plant species effect. However, I saw no evidence
that the presence of this species in mixtures increased the overall soil microbial biomass

for the mixture.

Effects of Plant Community Composition (Mixture) on the Soil Microbial

Community

I were able to detect plant community composition effects on the soil microbial
community even when aboveground plant biomass was a covariate. Year 3 PLFA PC4
was signiﬁcantly related to plant species composition. Soil microbial respiration was
signiﬁcantly related to plant community composition in year 3 (Figure 3.1A, Table
3.3A, p< 0.05), but not in year 4 (Figure 3.1B). Soil microbial biomass in these plots

was consistent across years (r2=0.39, p <0.01), although not related to diversity or

composition of the plant community.

Changes in Soil Microbial Community Composition

There was considerable variation in CLPP among plots, and much of this could be
accounted for by the plant diversity or composition treatments. PCl explained 46.1%

of the variation and was correlated with increased species diversity and increased

54

CLPP PC3 (7.6%)

— 1 at.
1 11 .1

Oil. R0 211‘} 1. %§;ﬁgﬁl}ﬁ4

l 1

 

_Number of PG ' 0

 

 

. 0
-3: ii 1
. 34 1 1 I

 

-4 '
-4 -3 -2 -1 O l 2
CLPP PCl (46.1%)

Figure 3.3. Plant diversity effects on the metabolic activity of the soil
microbial community as measured by Community Level Physiological
Proﬁles (CLPP) in year 3 (n = 28). Principal component axis 1 accounted
for 46.1% of the total variation. Principal component axis 3 accounted for
7.6% of the overall variation and was driven by the ability to metabolize
alpha-D-lactose. Signiﬁcance values for diversity effects are listed in
Table 3.3A. Number of functional groups is represented by symbol
shape. No-plant control plots are represented by circles, and non-
manipulated control communities labeled with “R”. Number of plant
species (0, l, 4, 8, or 11) are labeled.

55

nmnh
150“
\anar
560 w
p<0l’
was $1
anabs
tﬂant
posuil
(Tabh:

negau'

 

The V3
m'0 hr
indicat
Comnn
1411 1
more 1
indicat
PC3ac
(lite-r31“.
Vallalir

abere

numbers of functional groups (Table 3.3A), while PC2 accounted for an additional
15.0% of the variation, but was not signiﬁcantly correlated with any explanatory
variables. However, PC3 (which accounted for 7.6% of the variation in the CLPP data
set) was signiﬁcantly related to number of functional groups (Figure 3.3, Table 3.3,
p<0.05). PC3 was driven by the ability to metabolize a-D-lactose. This relationship
was still signiﬁcant when aboveground plant biomass was included as a covariate in the
analysis, indicating that the functional group diversity effect was independent of any
plant biomass effect. CLPP PC4 accounted for 6.8% of the variation and was
positively correlated with both the presence of legumes and the presence of forbs
(Table 3B, p<0.05, p<0.10), while PCS accounted for 5.4% of the variation and was

negatively correlated with the presence of forbs (Table 3B, p<0.01).

The variation in PLFA proﬁles of the soil microbial community showed structure at
two levels. PC 1 (45.3% of the variation) was signiﬁcantly related to Block (p<0.01),
indicating location in the ﬁeld was important in structuring the soil microbial
community at this site. Soils from the ﬁrst block contained more C9 dicarboxylic acid,
14:0, 15:0 iso, 16:0 iso, 17:0 iso, and 17:0 anteiso; while the second block contained
more 16:1 cis 11, 18:2 cis 12, 18:1 cis 9, and summed in feature 8: 18:1 trans 9,
indicating a higher proportion of eukaryotes, most likely fungi (Cavigelli et al. 1995).
PC2 accounted for 21.5% of the variation and was signiﬁcantly correlated with species
diversity (Table 3.3A, p<0.05). PC3 accounted for an additional 16.8% of the total
variation, but was not signiﬁcantly correlated with plant diversity, plant productivity, or

aboveground biomass. A much smaller amount of the variation in PLFA proﬁles (PC4,

56

.0

o

00
l

 

 

0.00 - E . E r E

-0.08 _ —~ -_

PLFA PC4 (4.5%)
f

 

 

 

 

-O.16 “‘

JJJJJIJJJJJJJ

 

C°\m°> use 3°°°tx°00

Figure 3.4. Plant community composition effects on the structure of the soil
microbial community, measured with Phospholipid Fatty Acid proﬁles of 1998
soil samples. Principal component axis 4 accounted for 4.5% of the overall
variation and reﬂected the amount of 15:0 anteiso in the PLFA proﬁles. Plant
community identiﬁcation codes are from Table 3.1B. No-plant control plots are
represented by a blank circle. The reference plots were not included in this
analysis. Monocultures are indicated by circles and mixtures by triangles.

57

...1

r..- -.

3.99“
p <0,

micrt

 

 

Discu

1 EXP:
sigm I"
comm
rather
range

effects

 

3.9%) was signiﬁcantly related to plant community composition (Figure 3.4, Table 3.2,
p <0.05). PC4 reﬂected the amount of 15:0 anteiso in the PLFA proﬁles. As with

microbial biomass and CLPP, this relationship was independent of plant biomass.

Discussion

I expected to ﬁnd that the composition and diversity of the plant community would
signiﬁcantly affect soil microbial diversity and productivity. I found that plant
community diversity and composition affected soil microbial community structure
rather than processes (Table 3.3A,B). Although I expanded the number of plots and
range of treatments sampled in year 4, I detected fewer diversity and composition
effects than in year 3 and the signiﬁcant variables were inconsistent across years (Table
3.3A,B). In most cases, the initially detected effects of plant species number or number
of functional groups on the soil microbial community were reduced or became non-
signiﬁcant when above-ground plant biomass was included as a covariate in these
analyses. Plant diversity effects above and beyond those effects on the microbial
community that could be accounted for by productivity were detected for CLPP and
PLFA proﬁles in year 3. However, CLPP proﬁle differences seem to be driven
primarily by the no plant control plots. Only year 3 did I detect a shift in the soil

microbial community (PLF A PC2) in response to plant species diversity (Table 3.3A).

58

Plant Diversity Effects on the Soil Microbial Community

A number of studies have suggested that there should be a relationship between plant
diversity and soil microbial diversity (Ohtonen et al. 1997, Wardle and Giller 1996,
Schlapfer and Schmid 1999). In most ﬁeld studies of plant community effects on soil
microbial communities several explanatory variables (e.g. plant diversity, plant
productivity, and plant community composition) are confounded. For example, J.C.
Zak et al. (1994) detected differences in the structure of the soil microbial community
(using Biolog) along an elevational and moisture gradient in the Chihuahuan Desert at
the Jomada Long-Tenn Ecological Research site. Similarly, Goodfriend (1998) used
CLPP patterns to distinguish among the communities at eight sites representing a
variety of wetland communities. Broughton and Gross (2000) examined characteristics
of the soil microbial community composition along a natural topographic, productivity
and diversity gradient at a site in southwestern Michigan and found a correlation
between the productivity of the plant and soil microbial communities, but no
relationship between plant diversity and soil microbial community composition (results
in Chapter 2). However, in all of these studies, the plant communities sampled were
from different sites in which there were likely concomitant changes in soil
characteristics, so the inﬂuence of plant community composition differences could not

be assessed independently of differences in edaphic characteristics.

In contrast, Wardle et al. (1999) showed plant effects on the soil microbial community,

which were not confounded by either soil or management effects. They removed

59

subsets of the plant community (that varied in number and functional group
composition) from a New Zealand perennial grassland. PLFA patterns distinguished
among soils from the plant removal treatments, suggesting plant community

composition effects on soil microbial community structure (Wardle et al. 1999).

The BIODEPTH experiment provides a unique opportunity to examine both species
diversity (through number and functional group) and composition effects on the soil
microbial community. Because the Silwood Park site preparation included fumigation
with methyl bromide after tillage to destroy the seed bank, the soil microbial
community was “standardized” before the initiation of treatments. This allowed us to
control for the effects of soil factors and focus solely on the manipulated plant diversity

and composition treatments as explanatory factors for the soil microbial community.

A recent paper by Stephan et al. (2000) from the Swiss BIODEPTH site reported a
relationship between plant species richness and functional diversity of the culturable
soil microbial community as measured by CLPP. They found that increased plant
species richness and plant functional diversity increased the overall catabolic activity
diversity in CLPP. However, Stephan et al. (2000) did not include a measure of plant
biomass as a covariate in these analyses, so it is not clear the extent to which the plant
diversity effect on the culturable soil bacteria is due to a correlated plant productivity
effect on the microorganisms. Hector et al. (1999) reported a strong relationship
between plant diversity and plant biomass at the Swiss site. At higher plant diversities,

aboveground plant biomass is greater, likely making labile carbon available to the soil

6O

microorganisms. Unlike the Silwood Park site, soils at the Swiss site were not
fumigated prior to the establishment of the diversity treatments (Hector et al. 1999,
Spehn et al. 2000a). These two sites also differed in the range of species diversity used
in the experiment: the Silwood site had a maximum species richness of l 1, whereas the
Swiss site had a maximum of 32 (Hector et al. 1999, Spehn et al. 2000b). These
differences in range of species diversity examined at the two sites is reﬂective of the
natural diversity at these sites. This may also affect the ability to detect plant species

diversity effects on the soil microbial community at these two sites.

CLPP catabolic activity has been found to be strongly related to innoculum density
(Garland and Mills 1991, Haack et al. 1994) and, thus, can be an approximate indicator
of overall bacterial number. I did not detect a signiﬁcant relationship between CLPP
overall catabolic activity and plant diversity at the Silwood site. However, the increase
in culturable bacteria with both increased plant species diversity (Figure 3.2A) and
increased plant functional group diversity (Figure 3.2B) reﬂects the underlying
relationship between plant diversity and plant biomass. The number of culturable
bacteria in the soil is clearly correlated with the overall aboveground plant biomass of

the plot (Figure 3.2C).

Plant Community Composition Effects on the Soil Microbial Community

In year 3 (but not year 4), I found plant community composition effects on some soil

and microbial parameters (Table 3.3A). Some previous studies have shown plant

61

COII‘.

 

soil
eff:
Tn,-

Cor

community composition to be more important than species number or number of
ﬁmctional groups in inﬂuencing ecosystem processes. Hooper and Vitousek (1998)
concluded that plant community composition better explained variation in nutrient
cycling processes in a Californian serpentine grassland than number of functional
groups. In a comprehensive study using plant removals, Wardle et al. (1999) found
signiﬁcant effects of plant community composition on several different trophic levels,
including the soil microbial community, and ecosystem properties. These results
suggest that individual plant species may inﬂuence communities and processes

independent of any diversity or productivity effects.

I also found that plant community composition signiﬁcantly affected some soil and
microbial parameters at the Silwood site, but these effects varied across years. The
presence of legumes was positively correlated with CLPP PC4 (year 3), Total N (year
4), and N-mineralization rates (year 4; Table 3.3B). As legumes are symbiotic with
nitrogen-ﬁxing bacteria in their roots, it is not surprising to ﬁnd an effect of legumes on
soil microbial processes. However, I were not able to determine whether the legume
effects were due to a particular legume species, because both L. corniculatus and
Trifolium repens were present in all mixtures containing legumes. However, L.
corniculatus monocultures differed from the other monocultures and mixtures in both
pH and CLPP proﬁles. At the Swiss BIODEPTH site, Stephan et al. (2000) found that
legumes had positive effect on overall CLPP catabolic activity, and Spehn et al.
(2000a) reported a positive effect of legumes on microbial biomass. In contrast to this

study, Stephan et al. (2000) were able to detect the effect of the presence of a speciﬁc

62

met.

 

SM:
duh
bye

I990

 

Lego

SUQ.
reﬂ

suc

Mr

ﬁn:

legume, T. repens, on the soil microbial community. The presence of T. repens was
positively correlated with CLPP catabolic activity and number of carbon sources
metabolized at the Swiss site (Stephan et al. 2000). Although both the Silwood and
Swiss sites had L. corniculatus and T. repens grown in monocultures and mixtures;
differences between these sites in the speciﬁc effects of these species suggest a species

by environment interaction as seen for aboveground biomass (Table 3.3, Hector et a1.

1999).

Legacy Effects on the Soil Microbial Community

N-mineralization rates did not differ signiﬁcantly among mixtures; however, the
reference plots consistently had lower N-mineralization rates than the treatment plots.
N-mineralization rates typically are higher in earlier successional sites and decline over
time (Schlesinger 1997). These differences in N-mineralization rates may reﬂect the
successional status of the treatment and reference plots. This temporal change may
reﬂect the immobilization of nitrogen by the soil microbial community in later

successional plots and a more mature soil microbial community (Schlesinger 1997).

Microbial biomass measurements from the reference communities at Silwood indicate
that the treatment plots still had not recovered from fumigation aﬂer 4 years. Because
microbial respiration and microbial biomass may be correlated with aboveground plant
biomass (Broughton and Gross 2000), and plant diversity is correlated with

aboveground plant biomass, I would expect to see higher microbial activities and

63

biomasses at higher plant diversity levels. Spehn et al. (2000a) observed a positive
relationship between soil microbial biomass and plant species diversity at the Swiss site
where the soil was not fumigated. The fact that I did not see a relationship between
plant diversity and soil microbial biomass at the Silwood site suggests that the soil
microbial community is still recovering from the severe disturbance of methyl bromide
application. Alternatively, it may be that different relationships between plant diversity
and soil microbial biomass emerge under different local conditions. Studies at other
BIODEPTH sites may help to resolve this issue. Disturbance effects on the soil
microbial community may persist for decades and make it difﬁcult to detect current
plant species or diversity effects on the soil microbial community. Buckley and
Schmidt (2001) found that there was little difference between the soil microbial
communities of a continuously tilled agricultural site and a companion successional site
(abandoned for 12 years) in southwestern Michigan. A nearby reference ﬁeld (never-
tilled) had a distinct soil microbial community (detected using rRNA) from either the

tilled or successional ﬁelds (Buckley and Schmidt 2001).

The results from this study provide some evidence that there is an overall plant
diversity effect on the soil microbial community. However, both the productivity and
the composition of the plant community can have effects on the biomass and structure
of the soil microbial community. In addition, disturbance effects on the soil microbial
community may persist for quite some time. Longer-term studies that can separate soil

microbial community function from structure are necessary to better investigate the role

64

that plants play in how the soil microbial community mediates the ﬂow of carbon and

nitrogen through the ecosystem.

65

REFERENCES

Bligh, E.G. and W. J. Dyer. 1959. A rapid method of total lipid extraction and
puriﬁcation. Canadian Journal of Biochemistry and Physiology 37 : 911-917.

Broughton, LC. and K. L. Gross. 2000. Patterns of diversity in plant and soil microbial
communities along a productivity gradient in a Michigan old-ﬁeld. Oecologia
125: 420-427.

Buckley, DH. and T. Schmidt. 2001. The structure of microbial communities in soil
and the lasting impacts of cultivation. Microbial Ecology 42: 11-21.

Cavigelli, M.A., G. P. Robertson, and M. J. Klug. 1995. Fatty acid methyl ester
(FAME) proﬁles as measures of soil microbial community structure. Plant and
Soil 170: 99-113.

Garland, J.L. and A. L. Mills. 1991. Classiﬁcation and characterization of heterotrophic
microbial communities on the basis of patterns of community-level sole-carbon-
source utilization. Applied and Environmental Microbiology 57: 2351-2359.

Goodfriend, W.L. 1998. Microbial community patterns of potential substrate
utilization: a comparison of salt marsh, sand dune, and seawater-irrigated
agronomic systems. Soil Biology and Biochemistry 30: 1169-1176.

Haack, S.K., H. Garchow, D.A. Odelson, L.J. Fomey, and M.J. Klug. 1994. Accuracy,
reproducibility, and interpretation of fatty acid methyl ester proﬁles of model

bacterial communities. Applied and Environmental Microbiology 60: 2483-
2493.

Hector, A., A. J. Beale, A. Minns, S. J. Otway, and J. H. Lawton. 2000. Consequences
of the reduction of plant diversity for litter decomposition: effects through litter
quality and microenvironment. OIKOS 90: 357-371.

Hector, A., B. Schmid, C. Beierkuhnlein, M. C. Caldeira, M. Diemer, P. G.
Dimitrakopoulos, J. A. Finn, H. Freitas, P. S. Giller, J. Good, R. Harris, P.
Hogberg, K. Huss-Danell, J. Joshi, A. Jumpponen, C. Komer, P. W. Leadley,
M. Loreau, A. Minns, C. P. H. Mulder, G. O’Donovan, S. J. Otway, J. S.
Pereira, A. Prinz, D. J. Read, M. Scherer-Lorenzen, E-D. Schulze, A-S. D.
Siamantziouras, E. M. Spehn, A. C. Terry, A. Y. Troumbis, F. I. Woodward, S.
Yachi, and J. H. Lawton. 1999. Plant diversity and productivity experiments in
European grasslands. Science 286: 1123-1127.

Hooper, D. U. and P. M. Vitousek. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

Huston, MA. 1997. Hidden treatments in ecological experiments: re-evaluating the
ecosystem function of biodiversity. Oecologia 110:449-460.

J onasson, S., A. Michesen, I. K. Schmidt, and E. V. Nielsen. 1999. Responses in
microbes and plants to changed temperature, nutrient, and light regimes in the
arctic. Ecology 80: 1828-1843.

66

Nelson, D.W. and LE. Sommers. 1982. Total carbon, organic carbon, and organic
matter, pp. 539-580, In_ Methods of Soil Analysis Part 2 - Chemical and
Microbiological Properties. ASA Special Publ. No. 9. ASA, SSSA, Madison,
WI.

Ohtonen, R., S. Aikio, H. Véire. 1997. Ecological theories in soil biology. Soil
Biology and Biochemistry 29: 1613-1619

Paul, EA. and FE. Clark. 1996. Soil Microbiology and Biochemistry. 2"d ed. San
Diego, CA: Academic Press, Inc.

Paul, E.A., D. Harris, H. P. Collins, U. Schulthess, and G. P. Robertson. 1999a.
Evolution of C02 and soil carbon dynamics in biologically manage, row-crop
agroecosystems. Applied Soil Ecology 11: 53-65.

Paul, E.A., D. Harris, M. J. Klug, and R. W. Ruess. 1999b. The determination of
microbial biomass, pp. 291-317, In Standard Soil Methods for Long-Term
Ecological Research (eds G.P. Robertson, D.C. Coleman, C.S. Bledsoe, & P.
Sollins) Oxford University Press, New York.

Rosenzweig, ML. 1995. Species Diversity in Space and Time. Cambridge University
Press, Cambridge.

Schlapfer, F. and B. Schmid. 1999. Ecosystem effects of biodiversity: a classiﬁcation

of hypotheses and exploration of empirical results. Ecological Applications 9:
893-912.

Schmidt, I.K., A. Michelsen, and S. Jonasson. 1997. Effects of labile soil carbon on
nutrient partitioning between an arctic graminoid and microbes. Oecologia 112:
557-565.

Shaver, GR. and F. S. Chapin, III. 1980. Response to fertilization by various plant
growth forms in an Alaskan tundra: nutrient accumulation and growth. Ecology
61: 662-675.

Schlesinger, W.H. 1997. Biogeochemistry An Analysis of Global Change. Academic
Press. San Diego, CA.

Spehn, E.M., J. Joshi, B. Schmid, J. Alphei, and C. Komer. 2000a. Plant diversity
effects on soil heterotrophic activity in experimental grassland ecosystems.
Plant and Soil 224: 217-230.

Spehn, E.M., J. Joshi, B. Schmid , M. Diemer, and C. Komer. 2000b. Aboveground
resource use increases with plant species richness in experimental grassland
ecosystems. Functional Ecology 14: 326-337.

Stephan, A., A. H. Meyer, and B. Schmid. 2000. Plant diversity affects culturable soil
bacteria in experimental grassland communities. Journal of Ecology 88: 988—
998.

Tiedje, J. M. 1995. Approaches to the comprehensive evaluation of prokaryote
diversity of a habitat, pp. 73-87, In Microbial Diversity and Ecosystem Function
(eds D. Allisopp, R.R. Colwell, and D.L. Hawksworth). CAB International,
Wallingford, UK.

67

Wardle, DA. and K. E. Giller. 1996. The quest for a contemporary ecological
dimension to soil biology. Soil Biology and Biochemistry 28: 1549-1554.

Wardle, D.A., K.I. Bonner, G.M. Barker, G.W. Yeates, K.S. Nicholson, R.D. Bardgett,
R.N. Watson, and A. Ghani. 1999. Plant removals in perennial grassland:

vegetation dynamics, decomposers, soil biodiversity, and ecosystem properties.
Ecological Monographs 69: 535-568.

Zak, J. C., M. R. Willig, D. L. Moorhead, and H. G. Wildman. 1994. Functional
diversity of microbial communities: A quantitative approach. Soil Biology and
Biochemistry 26: 1101-1108.

68

CHAPTER 4

PLANT-MEDIATED EFFECTS OF SOIL ORIGIN ON THE COMPOSITION AND
FUNCTION OF SOIL MICROBIAL COMMUNITIES

Introduction

While considerable attention has been paid to factors that affect the composition and
function of communities of macroorganisms, very little is known about the factors that
affect the structure and function of soil microbial communities (Ohtonen et al. 1997,
Tiedje 1995). To understand how changes in the structure of the soil microbial
community affect ecosystem functions, I must ﬁrst investigate what factors inﬂuence
soil microbial community structure and function. Plant community composition can be
stable for long time periods, but can also vary depending on factors such as disturbance
history and successional status. In contrast, soil characteristics change much more
slowly on average than the plant community (e.g. soil quality). Consequently, soil
characteristics may have a more consistent effect on the soil microbial community than
plants. While the soil has a large reserve of relatively recalcitrant carbon that is less
available to microorganisms, much of the labile carbon available to the soil microbial
community is derived from recent plant production (Paul and Clark 1996). Because the
plant community is dynamic and the main source of carbon for the soil microbial
community, the current plant community should have a big effect On structure and
function. To date, many studies have investigated the role of edaphic factors in
structuring the soil microbial community, while fewer have addressed the effects of

plants (Metting 1993, Schlapper and Schmid 1999, Hooper et al. 2000). Additionally,

69

very few have sought to distinguish between the effects of soil and the effects of plants

on the soil microbial community.

Several studies that have reported plant community effects on soil microbial
communities confound direct plant-mediated effects with soil effects. For example,
J .C. Zak et al. (1994) detected differences in the structure of the soil microbial
community (using Biolog) from plant communities that occur along an elevational and
moisture gradient in the Chihuahuan Desert at the Jomada Long-Term Ecological
Research site. Similarly, Goodﬁ'iend (1998) used Biolog to distinguish among the soil
bacterial communities at 8 sites representing a variety of wetlands across a salinity
gradient. However, in both these studies, the plant communities sampled were from
different sites in which there were likely concomitant changes in soil characteristics, so
the inﬂuence of plant community composition differences could not be assessed
independently of differences in edaphic characteristics. Other studies in single sites
have found little change in microbial communities in soils sampled from different plant
communities (e.g. Buckley and Schmidt 2001). After 10 years of plant community
divergence resulting from vastly different agricultural management, Buckley and
Schmidt (2001) could detect no differences among rRNA patterns of the soil microbial
communities among treatments at the Kellogg Biological Station’s Long Term
Ecological Research site, suggesting that the soil microbial community structure was
still dominated by the inﬂuence of the past land use and soil quality or, perhaps, that the
various plant forms contributed similarly to the sustenance of the microbial assemblage

despite our perception that they might differ in this respect. The plant communities

70

sampled ranged from successional ﬁelds to poplar plantations to conventional com. All
had been under similar management (conventional corn) prior to treatment

implementation (Robertson et al. 1993).

In this study, I am interested in distinguishing the relative importance of variability in
soil characteristics and the current plant community in controlling soil microbial
community structure and composition (see Figure 1.1). Speciﬁcally, I used a
manipulative greenhouse experiment to investigate whether (1) soils from different
plant communities that differ in fertility vary in the composition and function of the soil

microbial community and (2) plants can mediate these effects.

Methods

Site Descriptions

I selected six successional old-ﬁelds at the W. K. Kellogg Biological Station of
Michigan State University in southwestern Michigan to compare the soil microbial
communities of sites with different plant communities. The sites varied in fertility,
species richness, and dominant plant type, but all were located on Kalamazoo sandy
loam soil. I determined plant species composition at each site in six 0.5 m x 2 m plots
in August 1998. To estimate above-ground net primary productivity (ANPP), I clipped
aboveground biomass at ground level from a 0.5 m x 0.5 m plot located within the plots

used to assess species diversity and composition.

71

The six sites also differed in past land use and ranged in time since abandonment from
20 to 50 years. McKay ﬁeld was abandoned from agriculture in 1973; a section was
plowed once in 1981 and then re-abandoned (Burbank et al. 1992). Both the Upper and
Lower Louden ﬁelds were abandoned from agriculture in 1951 (Burbank et al. 1992).
The Bailey ﬁeld site was farmed until ten years prior to this sampling (K.L. Gross,
personal communication). The Pond Lab Orchard and Field K sites had been

abandoned for at least twenty years (Foster 1996).

Experimental Design

To determine if site differences in ANPP and species composition had detectable
effects on the soil microbial community I incubated soils from each site in the
greenhouse and evaluated the soil microbial communities 12 to 16 weeks later. To
determine if plants could mediate these differences, I sowed half the pots with
Andropogon gerardi, a C4 grass native to Michigan prairies. I collected approximately
10 kg of soil in June 1998 from the top 15 cm of each ﬁeld in the same area from which
species diversity and plant biomass were sampled. The soil was sieved to 4 mm and
thoroughly mixed. Soil was stored at room temperature until the experiment was

established in the greenhouse (less than 2 weeks).

I used a randomized complete block design for the greenhouse experiment to test for

the effects of soil origin and plant effects on the soil microbial community: 6 soils x 2

72

treatments, with 8 replicates of each. The two treatments were control (no-plant) and
plant (Andropogon gerardi). I chose A. gerardi because it is a native C4 grass that can
grow in all of these ﬁelds, although it was not present in our soil collection sites and is
rare in these communities because it is out-competed by naturalized C3 grasses (Foster
1996). This allowed us to measure the effects of a relatively novel plant on the soil
microbial communities present in each site. Seeds of A. gerardi were collected from
local ﬁelds in autumn 1997, and stored at room temperature in the laboratory until used

for these experiments the following summer.

Soils were placed in 5 cm diameter x 20 cm deep pots and kept well-watered with de-
ionized water to avoid adding nutrients or contaminants. Temperature in the
greenhouse ranged from 25 to 40 °C; light availability was controlled through a 12 h

light/ 12 h dark cycle.

Treatments were randomly assigned within replicates. Andropogon gerardi was added
as 2 week old seedlings; all seedlings were germinated in a sterile sand medium in 3
grth chamber and were less than 2 cm in height when transplanted. I estimated
initial biomass by drying a representative subset of the seedlings at the time of
transplantation. The experiment ran for a total of 16 weeks. I harvested the experiment
in two segments because of the number of samples and the time required to process
each sample: 4 replicates were harvested at week 12, and the remaining 4 replicates

were harvested at week 16. Thus, time was an additional factor in the ANOVA.

73

Data Collection

Differences among sites and treatments in soil fertility were assessed by (1) grth of
Andropogon gerardi, (2) inorganic nitrogen pools and N—mineralization rate, and (3)
soil organic matter. Shoot and root biomass of Andropogon gerardi were harvested
separately and dried at 60°C for 48 h. I separated root biomass from the soil during
sieving; roots were rinsed thoroughly in de-ionized water before drying. I used the
change in total plant biomass to estimate the relative growth rate (RGR) as [In (total
plant biomass) — In (initial plant biomass)] / number of days between harvest and

planting.

I sieved the soil through a 2 mm sieve, and stored it in sealed plastic bags at 4°C until
analysis. All analyses were done within 3 days of sampling, except PLFA. Soil for

PLFA analyses was kept at —80°C until the fatty acids were extracted.

I determined gravimetric soil moisture for each sample by drying 10 g soil at 105°C for
48 hours (Nelson and Sommers 1982). A subsample of the dried soil was ashed at
500°C for 4 hours to determine organic matter content (Nelson and Sommers 1982).
For nitrogen analyses, I extracted 20 g of fresh soil in 100 m1 1M KCl. The samples
were shaken for l min and allowed to settle for 24 h at room temperature. The
supernatant mixture was ﬁltered through a l-um Gelman glass ﬁlter and N03 and
NH4+ concentrations were measured using Alpkem auto-analyzer. To determine

potential N-mineralization and nitriﬁcation rates, a companion 20 g sample was

74

incubated for 21 days at 25°C and 15% humidity and then extracted using the same
methods as above. The remaining soil was used to characterize the soil microbial

community.

I assessed differences in soil microbial community production among the sites and
treatments by (1) microbial biomass C, (2) microbial respiration, and (3) plate counts
(number of colony-fonning units). I determined microbial biomass using the
chloroform fumigation incubation method (Paul et al. 1999). Two 25 g soil samples
were pre-incubated for 5 days then one sample was fumigated with chloroform for 24
hours to kill the microorganisms. After a vacuum was created and the chloroform
evaporated, 0.5 g of original soil was added to both samples. I measured initial
headspace C02 and accumulated C02 after 10 days on an ADC series EGA infrared
CO; gas analyzer (The Analytical Development Co. Ltd., Hoddesdon, Herts., UK). I
calculated microbial biomass as [ 1.73 * (10 day accumulated COz-C — initial COz-C
for the fumigated samples) - 0.56 "‘ (10 day accumulated COz-C — initial COz-C for the
control samples)] (Paul et al. 1999). To determine microbial respiration I used a
separate set of 10g soil samples that were pre-incubated 5 days in a 160 ml glass qorpak

bottle. I measured initial headspace C02 and accumulated CO; aﬁer 1 and 5 days.

I determined the number of colony-forming units by mixing 5 g of fresh, sieved soil
into 1% phosphate buffer to reach a ﬁnal dilution of 10'6 g soil/ ml. I plated this
solution on minimal media (R2A agar plates) and incubated the plates at 25 °C and then

counted the number of colony-forming units after 24 h and 48 h.

75

I assessed soil microbial community structure differences among sites and treatments
by (1) Community-level physiological proﬁles (CLPP) and (2) phospholipid fatty acid
(PLFA) proﬁles. Community-level physiological proﬁles (CLPP) were determined
using Biolog GN plates (Biolog, Inc., Hayward, Calif, USA) and reﬂect the range and
amount of carbon sources or resources that can be metabolized by the community
(Konopka et al. 1998). For the assay, 1 g of fresh, sieved soil was shaken with 99 ml
1% phosphate buffer solution for 20 min. 150 pl of the mixture was transferred to each
well of the microtiter plate (GN Biolog, 95 Carbon sources + l non-Carbon control).
The plates were incubated in the dark at 25 °C and optical densities were measured at
24 h intervals from 0 h to 96 h using an Emax precision microplate reader (Molecular
Devices Corp., Menlo Park, Calif, USA). Because the 5 incubation times gave
consistent results, I present here data only from the 96 h Biolog measurements. Optical

densities (intensity of resource use) were used in the RDA analysis.

For the PLFA analysis, I extracted lipids from 6 g whole soil samples for 2 h using a
mixture of dichloromethane (DCM): methanol: phosphate buffer (1:2:0.8 v/v/v),
following a modiﬁed Bligh-Dyer procedure (Bligh and Dyer 1959). Phase separation
was achieved by adding DCM and saturated sodium bromide solution (1:4 v/v). I
isolated the phospholipid fatty acids from the dried lipid extracts by solid phase
extraction. The lipid material was added to a polar column consisting of 100 mg silica
(Varian Bond Elut LRC Columns, Product # 1211-3010). Lipids of low or intermediate

polarity were eluted with chloroform and acetone and discarded. Subsequently,

76

phospholipid fatty acids were eluted with 1.5 ml methanol for preparation of fatty acid
methyl esters. I saponiﬁed the samples using 1 ml NaOH (15% w/v) in methanol (50%
v/v) at 100 °C for 30 min and methylated the sample with 2 ml 6M HCl in methanol at
80 °C for 10 min. I extracted the fatty acid methyl esters into 1.25 ml (1:1 v/v) methyl-
tert—butyl etherhexane for 10 min and washed the extract with 3 ml 1.2% NaOH.
Phospholipid amounts were measured using a HP 5890 series II gas chromatograph
(Hewlett Packard Co., Palo Alto, Calif, USA) equipped with a 7673 autosampler and
ﬂame ionization detector (Microbial ID Inc., Newark, Del., USA). Peaks were
identiﬁed by comparison with an external standard. For analysis, I included only those
phospholipid fatty acids that were present in greater than 50% of samples and reported
their abundance as the square root of the proportion of the total phospholipid fatty acid
amount in each sample (Hellinger transformation). Of the 70 lipids detected, 30

phospholipid fatty acids met this criterion (Table 4.3).

Statistical Analyses

I used a randomized complete block design (ANOVA) model to test the effects of soil
origin, presence/absence of plant, and time on the following response variables: soil
moisture, soil organic matter, plant biomass, relative growth rate, total N, N-
mineralization rate, nitriﬁcation rate, number of culturable bacteria, microbial

respiration, and microbial biomass.

77

I used a modiﬁed redundancy analysis (RDA, Legendre & Anderson 1999) to
determine the relationship between the environmental factors (soil origin,
presence/absence of plant, time of harvest) and the two measures of soil microbial
community structure, CLPP and PLFA. This is a relatively new, powerful technique
for multivariate analysis. RDA is a multiple regression technique that reduces the
number of variables necessary to explain the variation in a data set by creating
composite variables. In addition, RDA compares a second matrix that describes the
environment in which the original variables were measured. This new technique also
uses permutations to allow for statistical tests of how these composite variables vary
with the explanatory variables to determine the strengths of the signiﬁcance of any

environmental correlations with measures of the soil microbial community.

Because the CLPP and PLFA data matrices have many zeros, I transformed the CLPP
and PLFA data using a Hellinger transformation (a square root transformation of
relative abundance, Legendre & Gallagher, in press). The RDA procedure involved:
(1) the creation of a matrix of dummy variables corresponding to the randomized
complete block design (modeled from the experimental design: soil origin,
presence/absence of Andropogon gerardi, time at harvest), (2) redundancy analysis of
the relationship between the principal coordinates (matrix of optical density or
phospholipid fatty acid data) and the environmental variables (matrix of dummy
variables in (1)), and (3) implementation of a Monte Carlo permutation test to estimate
the statistical relationship between the two matrices (Legendre & Anderson 1999).

This analysis allows us to test which factors from the experimental design (soil origin,

78

presence/absence of Andropogon gerardi, time at harvest) are signiﬁcantly related to
the variation in the CLPP and PLF A patterns. The modiﬁed RDA is a better statistical

technique than regular ordination techniques because it allows for signiﬁcance testing.

Results

Plant Communities

The six ﬁeld sites varied in ANPP, species diversity, and soil organic matter (Table
4.1). McKay Field (MK) had the lowest ANPP and a low species diversity and was
dominated by Agropyron repens, a C4 perennial grass. Both Bailey (Ba) and Upper
Louden (UL) had moderate ANPP and high species diversities and were dominated by
diverse forb communities. In contrast, Lower Louden (LL) and the Pond Lab Orchard
(PL) ﬁeld had moderate ANPP and species diversities and were dominated by
graminoids. Lower Louden was dominated by Bromus inermz's, a C3 perennial grass,
although perennial forbs such as Solidago canadensis, Daucus carota, T araxacum
oﬂicionale, and Hieracium sp. contributed signiﬁcant biomass to total ANPP. The
Pond Lab Orchard site was dominated by several C3 species: Bromus inermis,
Agropyron repens, and Poa pratensis. Field K (FK) had the highest ANPP and lowest

species diversity and was dominated by Bromus inermis (Table 4.1)

79

Table 4.1. Plant productivity and diversity of abandoned ﬁelds from which soil was
collected for the greenhouse experiment. Values for Annual Net Primary Productivity,
species richness, and mean percent organic matter are expressed as mean i standard
deviations. Values that are not signiﬁcantly different for a given variable based on
Fisher’s LSD test have the same letter.

 

 

Site Dominant Peak Plant Biomass Species diversity Soil Organic
Plant Form (standing + litter, g/mz) (#/m2) Matter (%)
MK Grass 188 i 16 a 2.2 i 0.4 a 2.40 i 0.43 a
UL Forb 3203223 b 15.8i1.0e 3.17:0.23b
Ba Forb 424i52c 11.3i1.0d 3.03:0.17b
LL Grass 432 i 27 c 8.5 i 0.4 c 3.84 :t 0.32 0
PL Grass 480 i 48 c 5.7 i 0.8 b 3.63 i 0.28 c
FK Grass 592 i 22 d 1.3 i 0.2 a 3.84 i 0.20 c

 

Effects of Soil Origin on Soil and Microbial Processes

The differences among sites in plant community productivity were reﬂected in the
growth of Andropogon gerardi in the greenhouse (p < 0.001, F = 40.3, Table 4.2A,
Figure 4.1A). A. gerardi grown in soils from more productive sites had greater total
biomass (Table 4.2A, Figure 4.1). The positive relationship between A. gerardi
production and 1998 ﬁeld above-ground plant biomass (Figure 4.1) suggests that the
ranking of sites based on plant productivity also reﬂected differences in fertility.

Interestingly, time had no effect on any of the measured plant growth or soil or

80

Table 4.2. Effect of soil origin and presence of A. gerardi on (A) plant and (B) soil and
microbial variables as detected by Analysis of Variance. NS = not signiﬁcant, p > 0.05.
The time factor investigates the results of harvesting half the experiment at 12 weeks,

the other half at 16 weeks.

 

 

 

 

 

(A)

Plant Variable BLOCK SOIL TIME SOIL*TIME

Total Biomass (g) NS < 0.001 NS NS

Root Biomass (g) 0.045 < 0.001 NS NS

Shoot Biomass (g) NS < 0.001 NS NS

RGR (g/day) NS < 0.001 < 0.001 NS

Plant Height (cm) NS < 0.001 NS NS

(B)

Soil or Microbial Variable BLOCK SOIL PLANT TIME SOIL*PLANT
Percent Organic Matter < 0.01 < 0.001 NS -- NS
Soil Moisture NS < 0.01 < 0.05 NS < 0.05
Total Inorganic Nitrogen NS < 0.001 < 0.001 NS < 0.001
N-mineralization Rate NS NS < 0.001 NS < 0.05
Nitriﬁcation Rate NS NS < 0.001 NS < 0.01
Microbial respiration NS < 0.001 < 0.001 -- < 0.05
Culturable Bacteria (CFU’s) NS < 0.001 NS -- < 0.05
Microbial Biomass < 0.01 NS NS NS NS

81

 

 

 

10 Han7est ' T '
A 9 12 wks FK
:9 8 C l6wks T
3 LL PL i3
e 6- r i
"U
8
Q4 1
:5 4" MK UL ; “
is 3 i Ba
8
be 2 -— -
V’

l l l l

 

 

 

0
l 00 200 300 400 500 600
Field Above-Ground Biomass (g/mz)

Figure 4.1. Total plant biomass of Andmpogon
gerardi produced at 12 and 16 weeks in relation to
variation among sites in 1998 ﬁeld above-ground
plant biomass. Soils are coded as in Table 4.1.
Values are mean i standard error, n = 8.
Signiﬁcance values from the ANOVA are listed in
Table 4.2A.

82

microbial variables measured, indicating that the 4 weeks difference in harvesting
replicates had no discemable effect on the results (Table 4.2A&B). Consequently I

combined data from the two sampling intervals for the subsequent analyses.

Soil origin signiﬁcantly affected total inorganic nitrogen (p < 0.001, F = 14.9, Table
4.2B, Figure 4.2A), microbial respiration (p < 0.001, F = 17.8, Table 4.2B, Figure
4.2D), and the number of colony-forming units (p < 0.001, F = 12.2, Table 4.2B, Figure
4.2E), but did not inﬂuence N-mineralization rate (Table 4.23, Figure 4.28),
nitriﬁcation rate (Table 4.2B, Figure 4.2C), or microbial biomass (Table 4.2B, Figure
4.2F). In general, sites with higher fertility soils had higher soil microbial respiration,

and higher nitrogen pools in the absence of plants.

Effects of Andropogon gerardi on Soil and Microbial Processes

The presence of Andropogon gerardi signiﬁcantly affected several soil and microbial
characteristics and processes (Table 4.28). The presence of Andropogon gerardi
decreased soil moisture (p < 0.05, F = 4.5, Table 4.2B), total inorganic nitrogen (p <
0.001, F = 643.6, Table 4.2B, Figure 4.2A), N-mineralization (p < 0.001, F = 14.9,
Table 4.2B, Figure 4.2B) and nitriﬁcation rates (p < 0.001, F = 18.6, Table 4.2B, Figure
4.2C), and increased soil microbial respiration (p < 0.001, F = 322.7, Table 4.2B,
Figure 4.2D). The most dramatic effect was on total inorganic nitrogen; the presence of

Andropogon gerardi reduced nitrogen to similar low levels in all soils (Figure 4.2A).

83

.mmé 03:: :_
08m: 0:: <>OZ< 0:: 80.: m03:> 0080585 .v n : 0>:: 2033 8:888: 33088: :8: $883 03838: :0 8080:
808:0 =: :0: w n 2 .:0:8 080:8: H ::08 0:: m0:_:> .2008»: :003 30:88 8:30: 0:: 8:08 888% zomomoﬁﬁx
5:5 3:08:08 :0038: 005,383.06 0:: 838:0. 63:8? 8:0 :0... .3: 03:... 8 00%: m: 53:02:08 818088
.«0 80:0 8 888:: 0:: 8:5 8:805 83088 0: 0:: 888:: 0383.30 :0 808:: m: 803888: 83088
R: .0::: 83:05:08 ADV .08: :0_::N__:8:_8-Z Amv .:0w0::: 08:80:: .80: 03 8 00:0. w:08: :0_::_::> .Né 0:33:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0:5 0:5 03m
v7.— 15 s: :m ADM—2 vi .5 x: :m ADM—2 v.“— 15 s: :m JD 22
— q n 4 - _ o u A a u q a o N 8 1ﬁ — _ ﬂ 4 o )
.n
) . a m .. w
.n 2.: l.
ﬁr Au 18N “W M T w «W. LON) m. T e - 1% 7m M.J.
F 0m . DJ: 0 6 3 «J m.
. + . .2: z. m. .H. n y. e - .w m...
. mm - .0va r + -Epu
8 8 r. o e o m m
r 180 D. m. nwr m. s .m
0 . ,u w ( m o m. m
C m m # .00 m. r o w + -2 N. w
.l 160” ”h s ..V m- u
. ( W M
_ . : L : GOA: : : _ _ _ _ Ow : h : : r : ON
25 : 25 m 35 :
vi 4: .3 :m 43% . vi .5 ‘3 :m .5 v.2 .2 vi .5 ‘5 :m :5 v=2
: _ i : : 4 mo mm + l 4 : q : vo N f O f O O O
\l N
w w. W: - La
1 o o o 430% w I o o .0: Wm e we;
.0 Q0 0 . p D.W. _l IOV
m. W. 6 v :w ..M a :W
J S u m: u. l C L
. : ++12mm - hon: H 8m.
mm mm . + c::
A: be .(A\ m T A: .9800 0.. m
( 8:3 0
p _ _ : _ h o._ : r : _ _ _ o._ . _ : : 50850:. oo—

 

 

 

 

 

 

 

 

 

 

U m <

84

uaSon! N orueﬁioul [moi

Similarly, the presence of A. gerardi decreased N-mineralization and nitriﬁcation rates
to similar low levels in all soils (Figure 4.ZB&C). In contrast, soil microbial respiration
increased in the presence of A. gerardi, but the magnitude of this effect decreased with
fertility (Figure 4.2D). Although the number of culturable bacteria varied across sites
(Table 4.2B), and there was a signiﬁcant plant x site interaction, there was no consistent
effect of A. gerardi on this variable across sites. The presence of A. gerardi also did not
have a consistent effect on either the number of culturable bacterial colonies (Table 2B,

Figure 2B) or microbial biomass (Table 2B, Figure 2F).

Plant and Soil Effects on Soil and Microbial Processes

For some variates, the soil x plant treatment interaction (i.e. the magnitude of the A.
gerardi effect) varied across the sites and appeared to be related to soil fertility. To
evaluate this relationship, I estimated the magnitude of the relative “plant effect” on
these variables by calculating the relative difference in the variable in the plant versus

no-plant treatments ((plant — control)/ control).

The magnitude of the effects of A. gerardi on soil and microbial variables is illustrated
in Figure 4.3. For total nitrogen, the magnitude of the effect of A. gerardi varied with
site fertility and was inversely related to plant biomass (Figure 4.3A, sites are ranked by
fertility as per Figure 4.1, R2 = 0.18, p < 0.01). The effect of A. gerardi on soil

microbial respiration also varied with site fertility and was inversely related to plant

85

.w u : 0>:: 0:033 :03::38: 33808 3:: :30:0:3 038330 :0 838:: 308:0 w
u 2 .880 8:88: H :88 0:: 8:_:> 8:803 33808 E: 3:: £3083 038830 :0 838:: Am: 603838:
33828 5: .0::: 83:03:52 6: .0::: :03:N:::0:_8-Z Amv .:0w0::_: 0_::w:0:_ 3:0: A<v :0: 803:3 8:803
8:3 0:: 0: 53:3: :_ 8:: 80:0: 88000:: 33828 3:: =0: :0 80::0 8:3 0:: :0 0:88:68 03:, .3: 0::wi

 

 

 

 

 

 

 

 

 

 

 

8%
8%
vi 1i 44 :m 15 v32 . vi xi 15 :m 15 v3): cT
T A _ : _ : moi m d A A J : _
M
M. .l 1m.OI
r ..o.o w... +
m.
W 1 rod
1 -2 m.
m. r 1 md
:
I -o: 3:. 1 1o.—
0
m
: : : : r : m._ $ b b : : : m.—
0:_m : 33 m
vi xi :4 :m 43 v3): . - vi 4: j :m 15 v32 .
: A : . _ ﬂ mm m : _ 4 _ J 4 m _n
I -0.m- m... 83?:
M. found . u
1 -m T a r o 0 w to _
4 p .
I to T N
o o o I.
o n 1 -md-
r ,: 10:. M
m
.l loo W. T. lo.o
r .00 H
W
_ : : : : : O._ 9 _ _ _ : F :

 

 

 

 

 

 

9198 UOlIBZHBJQUlLU-N Pall/819193 qgv 19 saguologgo JaqumN pangpelau

2::

 

 

 

 

 

'

 

Sod v i
mmd H mm

:

_

vi 1i 1: :m 43 v32 . 8
1 4 q _ : _ CO H.
530v: m.
+ Sou”: . m.
r m o W
M
3
, + + m
+ m
m.
I 2 w
.m.
m
: : Ti: : : ON m.
U
0:3

vi xi 1: :m a: v.2 .

_ _ 4i 4 : _ OWG O.
+ a

r mmod-

oood-

mhwd-

 

Omwd-

 

uoﬁongN [910 .L pangnelou

86

biomass (Figure 4.3D, R2 = 0.26, p < 0.01). However, the signiﬁcant site x plant
interactions observed for N-mineralization rate (Table 4.28, p < 0.05, F = 2.7),-
nitriﬁcation rate (Table 4.28, p < 0.01, F = 3.6), and number of culturable bacteria
(Table 4.28, p < 0.05, F = 3.0) were not related to site fertility or Andropogon gerardi
production (Figure 4.3B, C, & E). Similarly, variation in microbial biomass was not

related to the magnitude of the plant response (Table 4.2B, Figure 4.3F).

Plant and Soil Effects on Soil Microbial Community Structure

CLPP of the soil microbial community varied among the soils from the six sites and
also responded to the presence of A. gerardi (Table 4.3, Figure 4.4). The RDA shows
that the presence of A. gerardi (Trt) was signiﬁcantly related to the ordination of the
CLPP proﬁles, as was soil origin (Table 4.3A). Figure 4.4 shows the separation of
samples coded by treatment. Axis I of the RDA accounted for 20.4% of the variance in
optical density data, 35.9% of the variance in the optical density-environment
relationship and had an optical density -environment correlation of 0.926 (Table 4.3B).
Axis 2 accounted for 12.0% of the variance in optical density data, 22.9% of the
variance in the optical density -environment relationship and had an Optical density - .
environment correlation of 0.946. Axis 3 accounted for 7.1% of the variance in optical
density data, 12.5% of the variance in the optical density -environment relationship and
had an optical density -environment correlation of 0.921. Axis 4 accounted for 5.7% of
the variance in optical density data, 10.0% of the variance in the optical density -

environment relationship and had an optical density -environment correlation of 0.897.

87

Table 4.3. Effect of soil origin and presence of A. gerardi on the soil microbial
community as detected by Distance-Based Redundancy Analysis (db-RDA) of CLPP
proﬁles. (A) Signiﬁcance values for the permutation tests on the environmental factors

 

 

 

of the RDA. (B) Variance explained by species data and species-environment
correlations for the RDA.

(A)

Factor Lambda F-stat p-value % variance

Plant 0.08 5.593 0.0010 7.6

Soil 0.36 5.340 0.0010 36.4

Plant*Soil 0.57 4.290 0.0010 56.7

(B)

 

Axis Eigenvalue Species-Environment Cumulative % Cumulative%

 

Correlation Variance of Variance of Species-
Species Data Environment
1 0.204 0.926 20.4 35.9
2 0.120 0.946 32.3 57.0
3 0.071 0.921 39.4 69.5
4 0.057 0.897 45.1 79.5

 

88

 

its

Canonical PC2, 12.0%
Or 0

 

 

 

 

0

Canonical PC 1 , 20.4%

1

Androgogon
Soil 0 +
MK <> 0
UL O 0
Ba 0 0
LL "'7 V
PL it ir
F K :L A

Figure 4.4. Soil origin and plant effects on the structure of the soil
microbial community as measured by CLPP. CLPP patterns are
distinguished between soil microbial communities from the A. gerardi
(solid symbols) and no plant treatments (open symbols). Soil microbial
communities in soils from different sites are indicated by symbols:
circles, BA; triangles, FK; upside-down triangles, LL; diamonds, MK;
stars, PL; pentagons, UL. Signiﬁcance values from the RDA are listed in

Table 4.4.

89

There were also some signiﬁcant interactions between the presence of A. gerardi and
soil origin (Table 4.3A), indicating that the presence of a plant did not have uniform
effects on the CLPP proﬁles across all soils. In general, CLPP patterns in the presence
of A. gerardi loaded lower on canonical PC] and canonical PC2; however, CLPP
patterns from UL soil were markedly different from all other soils regardless of the

presence of A. gerardi (Figure 4.4).

The RDA for the PLF A proﬁles of the soil microbial community also revealed variation
among the six sites and detected an effect of the presence of A. gerardi (Table 4.5,
Figure 4.5). Axis 1 accounted for 20.8% of the variance in phospholipid fatty acid data,
47.2% of the variance in the phospholipid fatty acid-environment relationship and had a
phospholipid fatty acid-environment correlation of 0.879 (Table 4.5B). Axis 2
accounted for 9.2% of the variance in phospholipid fatty acid data, 21.2% of the
variance in the phospholipid fatty acid-environment relationship and had a
phospholipid fatty acid-environment correlation of 0.890 (Table 4.5B). Axis 3
accounted for 5.1% of the variance in phospholipid fatty acid data, 11.6% of the
variance in the phospholipid fatty acid—environment relationship and had a
phospholipid fatty acid -environment correlation of 0.790 (Table 4.5B). Axis 4
accounted for 3.0% of the variance in phospholipid fatty acid data, 6.7% of the variance
in the phospholipid fatty acid-environment relationship and had a phospholipid fatty
acid-environment correlation of 0.630 (Table 4.5B). The presence of A. gerardi (Trt)
was signiﬁcantly related to the ordination of the PLFA proﬁles, as were all levels of

soil origin (MK, BA, LL, PL, and FK), and the time at harvest (Table 4.5A).

90

Table 4.4. Fatty acids used in principal components analysis of PLFA proﬁles. I
describe fatty acids using standard nomenclature where the total number of carbon
atoms appears before the colon and the total number of C-C bonds appears after it.
Cyclo-propane analogs are indicated by "cyclo," and the location of the epoxy bond is
indicated by a "c" followed by two numbers. If the cis or trans conﬁguration is
unknown, the word "at" is used. The number following "cis," "trans," or "at" indicates
the location of the double bond in relation to the carboxyl end of the molecule. Fatty
acids with the same retention time are grouped as "sum in feature" and given a unique
number designation.

 

Phospholipid Fatty Acids
C9 Dicarboxylic acid
14:0 iso

14:0

15:0 iso

15:0 anteiso

15:0

16:0 iso

16:1 cis 7

16:1 cis 9

16:1 cis 11

16:0

iso 17:1 G

17:0 iso

17:0 anteiso

17:0 cyclo

16:1 20H

18:1 9 trans alcohol
18:2 cis 12

18:1 cis 9

18:1 cis 13

18:0

19:1 at 11 alcohol
19:0 cyclo c11-12
19:0 cyclo 11-12 20H
22:0

22:0 20H

24:0

Coprostane
Unknown 25.339
Summed feature 8: 18:1 trans 9

 

 

91

Table 4.5. Effect of soil origin, time of harvest, and presence of A. gerardi on the soil
microbial community as detected by Distance-Based Redundancy Analysis of PLFA
proﬁles. (A) Signiﬁcance values for the permutation tests on the environmental factors of
the RDA. (B) Variance explained by species data and species-environment correlations
for the RDA.

 

 

 

(A)
Factor Lambda F -stat p-value % variance
Plant 0.06 9.189 0.0010 6.4
Time 0.01 1.859 0.0540 1.3
Soil 0.31 9.037 0.0010 31.3
Plant*Time 0.09 4.351 0.0010 9.0
Soil*Plant 0.42 5.607 0.0010 42.1
Soil*Time 0.36 4.799 0.0010 36.4
All 0.51 3.285 0.0010 51.2

(B)

 

Axis Eigenvalue Species-Environment Cum%Variance Cum%Variance

 

Correlation of Species Data of Species-
Environment
0.208 0.879 20.8 47.2
0.092 0.890 30.0 68.0
0.051 0.790 35.1 79.6
0.030 0.630 38.1 86.3

 

92

 

Canonical PC2, 9.2%

I
;=‘¢
l
D
l

 

1’ o
0
v 0 0’ 90 <>
0 ,3 v K K ’ ; ’
. ﬂ iii}; if? 4; O

 

 

-2 -1 0 1
Canonical PCl, 20.8%

2

 

Androgogon
S_0i|f 0 +
MK 0 6
UL O 0
Ba 0 0
LL 7 V
PL it? *
FK it A

Figure 4.5. Distance-based Redundancy Analysis of PLFA proﬁles,
investigating the effects of soil origin, plant, and time effects on the
structure of the soil microbial community. The PLFA proﬁles are
distinguished between soil microbial communities from the A. gerardi
(solid symbols) and no plant treatments (open symbols). Soil microbial
communities in soils from different sites are indicated by symbols:
circles, BA; triangles, FK; upside-down triangles, LL; diamonds, MK;
stars, PL; pentagons, UL. Labels for phospholipid fatty acids are listed
in Table 4.3. Signiﬁcance values from the RDA are listed in Table 4.5.

93

Figure 4.5 shows the separation of samples by environmental factors and fatty acids.
There was a signiﬁcant interaction between the presence of A. gerardi and time at harvest
(Table 4.5A, Plant*Time). Soils with plants tended to cluster higher on canonical PC1
(Figure 4.5). This was related to higher amounts of high carbon chain phospholipid fatty
acids (22:0, 24:0, 22:0 20H, and unknown 25.339), indicating more eukaryotes were

present in soils with plants.

Additionally, there was also a signiﬁcant interaction between the presence of A. gerardi
and soil origin (Table 4.5A: Soil*Plant), indicating that the presence of a plant did not
have uniform effects on the PLFA proﬁles across all soils. The higher fertility soils
(Ba, LL, PL, and FK) had similar PLFA patterns and responded the same way to the
presence of A. gerardi (an increase in canonical PC1, Figure 4.5); however, the PLFA
patterns of the soils from the two low productivity sites did not change in response to
the presence of A. gerardi and were different from the patterns of the high fertility sites
(Figure 4.5). The UL site was different from all other sites in the amounts of some
monounsaturated fatty acids on canonical PC2 (UL site had higher amounts of 18:1 cis
13 and lower amounts of summed in feature 8). Soils from the MK site showed a
smaller elevation along canonical PC2 and increased numbers of eukaryotes (higher

canonical PC 1 ).

94

Discussion

Growing A. gerardi in these soils provided an independent assay of the potential
productivity of each of these sites (A. gerardi production) and a direct test of the
“plant” effect on soil microbial community structure and processes. This study
suggests the origin of the soil and the presence of a plant both inﬂuence the structure
and functioning of the soil microbial community. Most previous studies have been
unable to distinguish between the effects of plants and the effects of soil origin on the
structure of the soil microbial community. For example, Zelles et al. (1992)
distinguished among the soil communities of grassland and agricultural ﬁelds using
PLF A; however, both the soils and the plant communities differed among sites. Zelles
et al. (1992) were able to distinguish among management regimes, but it was not
possible to determine the relative effects of the soil versus the plants on these
differences. In addition, those studies which have attempted to distinguish between soil
and plant effects on the soil microbial community often did not measure the soil effect
independent of any plant inﬂuence. Grayston and Campbell (1996) used CLPP patterns
to differentiate between the soil microbial communities from the rhizospheres of hybrid
larch (Larix eurolepis) and Sitka spruce (Picea sitchensis) trees in woodland and forest
sites. However, the study does not estimate the magnitude of the plant effect on the
CLPP patterns because there was no independent measure of the CLPP patterns of the
soil microbial community in the absence of plants in these sites (Grayston and

Campbell 1996).

95

Site Fertility Effects on Soil Microbial Processes

In this study, soil origin had a signiﬁcant inﬂuence on soil properties controlled by the
soil microbial community and on the structure of the soil microbial community itself.
This suggests that historical factors of the soil can have persistent effects on the soil
microbial community, while the extant plant community is a major source of labile
carbon and can inﬂuence the structure of the soil microbial community and,
consequently, ecosystem functioning through the soil microbial community. I have
presented site fertility as the driving factor explaining the relationships between soil
origin and the soil and microbial properties that I measured in this study. However,
other factors besides site fertility differed among these sites (Table 4.1). There were
some differences in percent soil organic matter across sites, and species diversity varied
dramatically among sites. Bossio et al. (1998) have shown that enrichment of organic
matter through agricultural management produces recognizable differences in the PLFA
patterns from the soil microbial communities from various management regimes
(organic, low-input, and conventional farming). The results from this study do not
change if I rank the sites by soil organic matter rather than ﬁeld above-ground biomass
(data not shown). Both soil organic matter and above-ground biomass are surrogates
for site fertility. Bossio et al. (1998) suggested that higher soil organic matter should
lead to greater soil microbial biomass, but this study does not support this assertion.
Increased organic matter inputs generally occur in agricultural systems that are being

managed organically; perhaps natural gradients in soil organic matter should not be

96

expected to reﬂect the same pattern because the systems have already had time to reach

an equilibrium in soil organic matter turnover.

Historical Plant Diversity E ﬂeas on Soil Microbial Processes

Another major difference among these ﬁeld sites was the current plant species
composition and diversity. There appears to be a unimodal relationship between above-
ground plant biomass and species diversity across these six sites with low diversity,
grass-dominated communities at both the lowest (MK) and highest (FK) fertility sites
(Table 4.1). However, the soil and microbial properties of ﬁeld MK were consistently
more similar to the other low fertility site (UL) rather than FK, the other low species
diversity site (Figure 4.2). This suggest that it is fertility more than diversity or

composition that inﬂuences the soil microbial community.

To better understand the impact of the global decline in species diversity due to human
activities, many researchers have been investigating the relationship between species
diversity and ecosystem function. The ‘rivet hypothesis’ proposes that each species
contributes something unique to ecosystem function, and so ecosystem function
declines as biodiversity declines (Ehrlich and Ehrlich 1981, Lawton 1984). A
contrasting hypothesis suggests that species are redundant and that ecosystem function
only declines when functional groups are missing from an ecosystem (Walker 1992,
Lawton 1994). Finally, Lawton (1994) proposed that ecosystem ﬁmction changes when

a species is lost, but the direction and amount of that change are not predictable.

97

Observational studies and manipulative experiments investigating these theories have
provided mixed results. My work in Chapter 2 (Broughton and Gross 2000) showed a
signiﬁcant effect of plant species diversity on the respiration or biomass of the soil
microbial community, but these results were confounded with plant productivity and
edaphic changes. My work at the Silwood, England BIODEPTH site (Chapter 3) found
a positive relationship between plant diversity and two measures of microbial
community structure (CLPP and PLFA). The Swiss BIODEPTH experiment has
shown positive relationships between plant diversity and plant biomass (Spehn et al.
2000a, Spehn et al. 2000b), soil microbial respiration and functional diversity (Stephan
et al. 2000), microbial biomass (Spehn et al. 2000a), and earthworm population density
(Spehn et al. 2000a). In contrast, Wardle has consistently shown no relationship
between diversity and ecosystem function in a series of plant removal experiments in
New Zealand perennial grasslands (Wardle et al. 1999, Wardle et al. 2000, Wardle and
Nicholson 1996). In a plant removal study in a North American grassland, Symstad et
al. (1998) showed a positive relationship between plant species diversity and
productivity, but no relationship between plant diversity and other ecosystem functions.
Mikola and Setala (1998) found unpredictable ecosystem functioning responses to
changes in species diversity when studying the phenomenon in a simple (three trophic

level) decomposer food web from the soil of a pine forest in Finland.

In this experiment, I found no relationship between species diversity and soil microbial

processes, regardless of the presence of Andropogon gerardi. Soils from communities

98

with high plant species diversity did not have higher or more efﬁcient ecosystem
processes or higher soil microbial activities or biomass, as suggested by some recent
theory (Loreau 2000). These results are consistent with Symstad et al.’s (1998) plant
species removal study in which they detected the effect of declining species richness on
productivity, but not nitrogen retention, suggesting that the relationship between
species diversity and ecosystem processes is not necessarily consistent or predictable.
In this study, the one exception was microbial biomass. The pots without A. gerardi
showed a positive relationship between the microbial biomass and the plant species
diversity of the communities from the soils were taken. This relationship disappeared
when A. gerardi was grown in the soil, however, suggesting that the present plant

community can have a large effect on the present soil microbial community.

Several studies have shown plant composition (rather than plant diversity) effects on
ecosystem functioning. The researchers reason that the quality of the carbon available
to the microbial community is important and so the identity of the plant species
providing that carbon should inﬂuence how ecosystem functions change (Paul and
Clark 1996). Wardle et al. (1999) saw plant composition effects on PLFA patterns in
the soils from a plant removal experiment in New Zealand grasslands, while Symstad et
a1. (1998) found plant composition effects on productivity and nitrogen retention.
Hector et al. (2000) detected a relationship between the species composition of litter
and the decomposition rate at the Silwood Park, England BIODEPTH site. Hooper and
Vitousek (1998) investigated the relationship between plant composition and nutrient

cycling in experimental plots on serpentine soil in California. They determined that

99

plant community composition accounted for much more of the variation in nutrient
cycling processes than did plant functional group diversity alone (Hooper and Vitousek
1998). Knowing the identity of the plant species involved allowed for a much better
explanation of changes in inorganic N pools, soil moisture, microbial biomass, and
microbial immobilization as a result of the experimental manipulations (Hooper and
Vitousek 1998). I found some evidence to support this view in the relationships
between plant composition and microbial respiration and soil PLFA patterns at the
Silwood Park, England BIODEPTH site (Chapter 3). Historical plant composition
effects may be contributing some of the variation in this study in the effects of soil
origin on soil and microbial processes, but I are unable to test this assertion due to the

design of the study.

The ‘Plant’ Effect on Soil Microbial Processes

Not only was soil origin important in structuring the soil microbial community in this
experiment, the presence of a plant also had important effects on the structure and
function of the soil microbial community. The presence of A. gerardi drove down
nitrogen pools, but increased microbial respiration without affecting microbial biomass,
suggesting an increase in the turnover rate of the soil microbial community and faster

nutrient cycling.

Other studies have shown plant effects on soil and microbial processes. Bachmann and

Kinzel (1992) detected differences in the amounts of amino acids and sugars and the

100

rates of C02 evolution and some enzyme activities in the rhizosphere soils of six
different plant species grown in four different soils. As in our study, Bachmann and
Kinzel (1992) were able to detect strong plant effects regardless of soil origin and
strong plant-soil interactions, although the magnitude and direction of change for
enzyme activities and resource amounts were not consistent for different plant species.
However, Bachmann and Kinzel (1992) did not measure nitrogen transformation rates,
nitrogen pools, or microbial respiration or biomass. Groffrnan et al. (1996) investigated
the relative roles of plant versus soil effects on the soil microbial community by
measuring microbial biomass and activity and nitrogen transformation rates of soil
taken from a range of old—ﬁeld sites. The two experiments used various combinations
of 10 plant species in monoculture and 4 soils that had been established for 4 years.
Groffman et al. (1996) concluded that the main driver for microbial biomass and
activity was soil type rather than plant species, although they suggested that plant

effects might become more important aﬁer a longer period of time.

In this study, the magnitude of the plant effect on microbial respiration and total
inorganic Nitrogen pools was larger in higher fertility soils than low fertility soils. One
possible explanation for this result is that in lower fertility soils, there is a greater
possibility that the resources supporting the soil microbial community are coming from
the extant plant community. Consequently, the presence of a plant constantly providing
resources to the soil microbial community could allow for a more active microbial

community. This is in contrast to the possibility of increasing the soil microbial

101

biomass using the inﬂux of new resources, which Bossio et al. (1998) suggest is often

the result of increasing organic matter inputs.

Plant and Soil Effects on Soil Microbial Community Structure

Both soil origin and the presence of Andropogon gerardi had signiﬁcant effects on
potential metabolic diversity (measured with CLPP patterns) and phenotypic diversity
(measured with PLFA proﬁles) of the soil microbial community in this study.
Communities from ﬁeld UL soils had strikingly different PLFA and CLPP patterns
from communities grown in other soils; additionally, the presence of A. gerardi had no
effects on CLPP and PLFA patterns for UL soils. The UL ﬁeld is a highly diverse, low
productivity site and the composition of the soil microbial community does not seem to

respond quickly to changes in the plant community.

The soil microbial communities detected in MK soil (another low fertility site) had
PLFA patterns similar to ﬁeld UL, but had similar CLPP patterns similar to the higher
productivity sites. This indicates that structurally different microbial communities
(from ﬁelds MK and UL) are capable of consuming the same resources. Other studies
have found that community function may not change when community structure does.
Buyer and Drinkwater (1997) detected differences in PLFA patterns between replicates
of manipulations of different management treatments involving different crop residues

but saw no differences in CLPP patterns. Similarly, IbekI and Kennedy (1998)

102

determined that PLFA proﬁles were more sensitive than CLPP patterns to differences

in the soil microbial community grown in two different soils under six plant treatments.

Like field UL, ﬁeld MK is a low productivity ﬁeld; however, MK also has low organic
matter is dominated by a grass, Agropyron repens. If the long chain phospholipid fatty
acids indicate mycorrhizae, the historical presence of A. repens plants in MK may
explain why microbial communities incubated in MK soils have a higher proportion of
eukaryotes than microbial communities grown in other soils, regardless of the presence
of Andropogon gerardi. Although, the long chain phospholipid fatty acids may only be

indicative of more plant material present in the soil.

A central goal in ecology is to determine the factors controlling the abundance and
distribution of species. Microorganisms in their natural habitats are only recently being
studied in ecology. This study suggests that both the history of the soil and the plants
presently growing in a community affect the structure and function of the soil microbial
community. Further studies should eludicate the relative importance of these two

factors in inﬂuencing the soil microbial community.

103

REFERENCES

Bachmann, G. and H. Kinzel. 1992. Physiological and ecological aspects of the
interactions between plant roots and rhizosphere soil. Soil Biology and
Biochemistry 24: 543-552.

Bligh E. G., W. J. Dyer. 1959. A rapid method of total lipid extraction and puriﬁcation.
Canadian Journal of Biochemistry and Physiology 37 : 911-917

Bossio, D. A. and K. M. Scow. 1998. Impacts of carbon and ﬂooding on soil
microbial communities: Phospholipid fatty acid proﬁles and substrate utilization
patterns. Microbial Ecology 35: 265-278.

Bossio, D. A., K. M. Scow, N. Gunapala, and K. J. Graham. 1998. Determinants of
soil microbial communities: effects of agricultural management, season, and soil
type on phospholipid fatty acid proﬁles. Microbial Ecology 36, 1-12.

Broughton, LC. and K. L. Gross. 2000. Patterns of diversity in plant and soil microbial
communities along a productivity gradient in a Michigan old-ﬁeld. Oecologia
125: 420-427.

Buckley, DH. and T. Schmidt. 2001. The structure of microbial communities in soil
and the lasting impacts of cultivation. Microbial Ecology 42: 11-21.

Burbank, D. H., K. S. Pregitzer, and K. L. Gross. 1992. Vegetation of the W. K.
Kellogg Biological Station. Michigan State University Agricultural Experiment
Station Research Report Number 510.

Buyer, J. S. and L. E. Drinkwater. 1997. Comparison of substrate utilization assay and
fatty acid analysis of soil microbial communities. Journal of Microbiological
Methods 30: 3-11.

Ehrlich, P. R. and A. H. Ehrlich. 1981. Extinction: The Causes and Consequences of
the Disappearances of Species. Random House, New York.

Foster, B. L. 1996. Plant competition and diversity in relation to productivity in old-
ﬁeld plant communities. Ph.D. Dissertation. Michigan State University, East
Lansing, MI, USA.

Goodfriend, W.L. 1998. Microbial community patterns of potential substrate
utilization: a comparison of salt marsh, sand dune, and seawater-irrigated
agronomic systems. Soil Biology and Biochemistry 30: 1169-1176.

Grayston, S. J. and C. D. Campbell. 1996. Functional biodiversity of microbial
communities in the rhizospheres of hybrid larch (Larix eurolepis) and Sitka
spruce (Picea sitchensis). Tree Physiology 16: 1031-8.

Groffrnan, P. M., P. Eagan, W. M. Sullivan, and J. L. Lemony. 1996. Grass species
and soil type effects on microbial biomass and activity. Plant and Soil 183: 61-
67.

104

Hamilton, E. W., HI, and D. A. Frank. 2001. Can plants stimulate soil microbes and
their own nutrient supply? Evidence from a grazing tolerant grass. Ecology 82:
2397-2402.

Hector, A., A. J. Beale, A. Minns, S. J. Otway, and J. H. Lawton. 2000. Consequences
of the reduction of plant diversity for litter decomposition: effects through litter
quality and microenvironment. OIKOS 90: 357-371.

Hooper, D. U., D. E. Bagel, V. K. Brown, L. Brossard, J. M. Daingerﬁeld, D. H. Wall,
D. A. Wardle, D. C. Coleman, K. E. Giller, P. Labelle, W. H. van deer Patten, P.
C. de Reiter, J. Ruse, W. L. Silver, J. M. Tiedje, and V. Walters. 2000.
Interactions between aboveground and belowground biodiversity in terrestrial
ecosystems: patterns, mechanisms, and feedbacks. Bioscience 50: 1049-1061.

Hooper, D. U. and P. M. Vitousek. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

Bikeway, A. M. and A. C. Kennedy. 1998. Phospholipid fatty acid proﬁles and carbon
utilization patterns for analysis of microbial community structure under ﬁeld
and greenhouse conditions. FEMS Microbiology Ecology 26: 151-163.

Konopka, A., L. Oliver, and RF. Turco, Jr. 1998. The use of carbon substrate
utilization patterns in environmental and ecological microbiology. Microbial
Ecology 35: 103-115.

Lawton, J. H. 1994. What do species do in ecosystems? OIKOS 39: 391-422.

Legendre, P. and M. J. Anderson. 1999. Distance-based redundancy analysis: testing
multispecies responses in multifactorial ecological experiments. Ecological
Monographs 69: 1 —24.

Legendre, P. and E. D. Gallagher. 2001. Ecologically meaningful transformations for
ordination of species data. Oecologia (in press).

Loreau, M. 2000. Biodiversity and ecosystem functioning: recent theoretical advances.
OIKOS 91: 3-17.

Metting, B. F., Jr. 1993. Structure and physiological ecology of soil microbial
communities, pp. 3-25, In Soil Microbial Ecology: Applications in Agricultural
and Environmental Management (ed. F. B. Metting, Jr.). Marcek Dekker, Inc.,
New York.

Mikola, J. and H. Setala. 1998. Relating species diversity to ecosystem functioning:
mechanistic backgrounds and experimental approach with a decomposer food
web. OIKOS 83: 180-194.

Nelson, D.W. and LE. Sommers. 1982. Total carbon, organic carbon, and organic
matter, pp. 539-580, In_ Methods of Soil Analysis Part 2 - Chemical and
Microbiological Properties. ASA Special Publ. No. 9. ASA, SSSA, Madison,
WI.

Ohtonen, R., S. Aikio, H. Vare. 1997. Ecological theories in soil biology. Soil
Biology and Biochemistry 29: 1613-1619

105

Paul, EA. and FE. Clark. 1996. Soil Microbiology and Biochemistry. 2nd ed. San
Diego, CA: Academic Press, Inc.

Paul, E.A., D. Harris, M. J. Klug, and R. W. Ruess. 1999. The determination of
microbial biomass, pp. 291-317, In Standard Soil Methods for Long-Term
Ecological Research (eds G.P. Robertson, D.C. Coleman, C.S. Bledsoe, & P.
Sollins) Oxford University Press, New York.

Robertson G. P., J. R. Crum, and B. G. Ellis. 1993. The spatial variability of soil
resources following long-term disturbance. Oecologia 96: 451-456

Schléipfer, F. and B. Schmid. 1999. Ecosystem effects of biodiversity: a classiﬁcation

of hypotheses and exploration of empirical results. Ecological Applications 9:
893-912.

Spehn, E.M., J. Joshi, B. Schmid, J. Alphei, and C. Korner. 2000a. Plant diversity
effects on soil heterotrophic activity in experimental grassland ecosystems.
Plant and Soil 224: 217-230.

Spehn, E.M., J. Joshi, B. Schmid , M. Diemer, and C. Komer. 2000b. Aboveground
resource use increases with plant species richness in experimental grassland
ecosystems. Functional Ecology 14: 326-337.

Stephan, A., A. H. Meyer, and B. Schmid. 2000. Plant diversity affects culturable soil
bacteria in experimental grassland communities. Journal of Ecology 88: 988-
998.

Symstad, A. J., D. Tilman, J. Wilson, and J. M. H. Knops. 1998. Species loss and
ecosystem functioning: effects of species identity and community composition.
OIKOS 81: 389-397.

Tiedje, J. M. 1995. Approaches to the comprehensive evaluation of prokaryote
diversity of a habitat, pp. 73—87, In Microbial Diversity and Ecosystem Function
(eds D. Allisopp, R.R. Colwell, and D.L. Hawksworth). CAB International,
Wallingford, UK.

Walker, B. H. 1992. Biodiversity and ecological redundancy. Conservation Biology
6: 18-23.

Wardle, D.A., K.I. Bonner, and GM. Barker. 2000. Stability of ecosystem properties

in response to above-ground functional group richness and composition.
OIKOS 89: 11-23.

Wardle, D.A., K.I. Bonner, G.M. Barker, G.W. Yeates, K.S. Nicholson, R.D. Bardgett,
RN. Watson, and A. Ghani. 1999. Plant removals in perennial grassland:

vegetation dynamics, decomposers, soil biodiversity, and ecosystem properties.
Ecological Monographs 69: 535-568.

Wardle, D. A. and K. S. Nicholson. 1996. Synergistic effects of grassland plant
species on soil microbial biomass and activity: implications for ecosystem-level
effects of enriched plant diversity. Functional Ecology 10: 410-416.

Zak, J. C., M. R. Willig, D. L. Moorhead, and H. G. Wildman. 1994. Functional
diversity of microbial communities: A quantitative approach. Soil Biology and
Biochemistry 26: 1101-1108.

106

Zelles, L., R. Rackwitz, Q. Y. Bai, T. Beck, and F. Beese. 1995. Discrimination of
microbial diversity by fatty acid proﬁles of phospholipid fatty acids and
lipopolysaccharides in differently cultivated soils. Plant and Soil 170: 115-22.

Zelles, L., Q.Y. Bai, T. Beck, and F. Beese. 1992. Signature fatty acids in
phospholipid fatty acids and lipopolysaccharides as indicators of microbial
biomass and community structure in agricultural soils. Soil Biology and

Biochemistry 24: 317-323.

107

CHAPTER 5

AN EXPERIMENTAL EVALUATION OF THE EFFECTS OF DIFFERENT PLANT
SPECIES ON THE STRUCTURE AND FUNCTION OF SOIL MICROBIAL
COMMUNITIES

Introduction

Most of the carbon and nitrogen entering the soil results from litterfall, root exudates,
or root death of plants (Paul and Clark 1996). Because organic inputs from plants
species can differ in quantity, timing, and biochemistry, plant species identity has the
potential to affect microbial process rates through litter quality and root exudation
effects. As a result, the composition and productivity of the plant community may
inﬂuence the soil microbial community. Conversely, the productivity or diversity of
the plant community may be affected by processes mediated by soil microorganisms
(e. g. N-mineralization rates). Consequently, changes in the plant community and the
accompanying change in the soil microbial community potentially affect ecosystem

function.

To date, many studies have investigated the role of edaphic factors in structuring the
soil microbial community, while fewer have addressed the effect of the plant
community. Additionally, very few have sought to distinguish between the effects of
soil and the effects of plants on the soil microbial community. Most studies of plant
community effects on soil microbial communities confound direct plant effects with
soil effects. For example, J .C. Zak et al. (1994) detected differences in the structure of

the soil microbial community (using Biolog) along an elevational and moisture gradient

108

in the Chihuahuan Desert at the Jomada Long-Tenn Ecological Research site.
Similarly, Goodfriend (1998) used Biolog to distinguish among the microbial
communities at eight sites representing a variety of wetlands. However, in both of
these studies, the plant communities sampled were from different sites in which there
were likely concomitant changes in soil characteristics Consequently, the inﬂuence of
plant community composition differences could not be assessed independently of

differences in edaphic characteristics.

It remains unclear how the plant community affects the soil microbial community
relative to effects of soil environment on the microbial community. In an earlier
greenhouse experiment I used ﬁeld soils from six sites to distinguish the relative
importance of direct soil effects versus plant-mediated soil effects on the soil microbial
community in a controlled environment. Using a single species, Andropogon gerardi, I
found that microbial communities grown in the six soils differed in both structure and
function and that A. gerardi mediated those differences. To follow up on those results
and determine if plant species differed in their effects on the structure and function of
the soil microbial community, I conducted a greenhouse experiment in which soils from
two old ﬁelds that differed in fertility were planted with all combinations of three plant
species common to local old ﬁelds. I hypothesized that (1) plant species have unique
effects on the structure and ﬁinction of soil microbial communities, (2) the effects of
different plant species on soil microbial community structure and function are non-
additive, and (3) soil microbial communities close to the roots (rhizosphere) are

different from the communities in the bulk soil. To test whether soil microbial

109

communities in rhizosphere soil differed from those in bulk soil, 1 excluded roots from

a cylinder of soil in each pot.

Methods

Site and Species Descriptions

Soil was collected in October 1999 and October 2000 from two of the six successional
old-ﬁelds at the W. K. Kellogg Biological Station in southwestern Michigan sampled
for the experiment in Chapter 4. The sites differed in fertility, species diversity,
dominant plant type, and years since abandonment, but both were located on
Kalamazoo sandy loam soil (Table 5.1). To estimate plant species diversity, I
determined species composition in six 0.5 m x 2 m plots at each site in August 1998.
To estimate annual net primary productivity (ANPP), I clipped aboveground biomass at
ground level from a 0.5 m x 0.5 m plot located within the plots used to assess species

diversity.

The Upper Louden (UL) ﬁeld had moderate ANPP, high species diversity and was
dominated by diverse forb communities. In contrast, Field K (FK) had high ANPP, low
species diversity, and was dominated by Bromus inermis (Table 5.1). The UL ﬁeld had
been abandoned from agriculture for over ﬁfty years. Field K was once used as

pasture, but had been unmanaged for over twenty-ﬁve years.

110

Table 5.1 Plant species by functional group and aboveground biomass at the two ﬁeld
sites from which soil was collected. Functional groups are coded by Grass (G), Forb
(F), Legume (L), and Woody (W). Plant species used in this greenhouse experiment
are in bold.

 

 

Field Species Functional Biomass % of Total
group (g/mz) Biomass

F K Bromus inermis G 589.6 98.7
A gropyron repens G 8.0_, 1.3
Total 597.6

UL Poa compressa G 48.8 15.6
T rifolium pratense L 34.4 1 1.0
Andropogon virginicus G 30.6 9.8
Hieracium sp. F 30.5 9.8
Danthonia spicata G 26.3 8.4
Rudbeckia hirta F 20.6 6.6
Solidago nemoralis F 20.0 6.4
Solidago canadensis F 14.4 4.6
Aster sp. F 14.4 4.6
Antennaria plantaginifolia F 9.4 3.0
Rubus sp. W 9.3 3.0
Panicum sp. G 8.1 2.6
Achillea millifolium F 8.0 2.6
Centaurea maculosa F 7.0 2.3
Chrysanthemum Ieucanthemum F 5.5 1.8
Aster sp. F 4.4 1.4
Rumex acetosella F 3.8 1.2
Panicum sp. G 3.8 1.2
Solidago speciosa or juncea F 3.7 1.2
Other (5 species) 3F,2G 9.2 . 2.9
Total 312.3

 

To maximize the possible differences among the plant species, I selected three
perennial species that are representative of different functional groups in old-ﬁeld
communities: Solidago canadensis (forb), T rifolium pratense (legume), and Bromus

inermis (grass). Solidago canadensis is a native herbaceous perennial dicot that is

111

commonly dominant in higher productivity ﬁelds of southwestern Michigan (Foster
1996, Werner et al. 1980). Trifolium pratense is an herbaceous perennial legume
naturalized from Eurasia that occurs in ﬁelds across a broad range of productivities
(Scoggin 1978b). Bromus inermis is a C3 perennial grass dominant in high productivity

ﬁelds in southwestern Michigan and is naturalized from Eurasia (Scoggin 1978a).

Experimental Design

To distinguish direct plant effects versus plant-mediated soil effects on the soil
microbial community, I grew each species alone and in all combinations in soil from
the two sites described above in a greenhouse experiment. I also included no-plant

controls. There were root exclosures in each pot to separate bulk from rhizosphere soil.

Soil for these experiments was collected from both ﬁeld sites from the top 15 cm in the
same area from which species diversity and plant biomass were sampled. Because of
sample processing constraints, I conducted the experiment in three time blocks. Soil
collected in December 1999 was used for the ﬁrst two time blocks, while soil collected
in October 2000 was used for the third time block. The soil was sieved to 4 mm and
mixed to reduce variability. Soil was stored at room temperature until the experiment

was established in the greenhouse (less than 2 months).

I used a randomized complete block factorial design for the greenhouse experiment in

which I varied soil (2 sources - Soil) and plant species (3 species in monoculture and all

112

combinations, including no plants), with 3 replicates of each treatment (Replicate) at 3
different times (Time). The plant treatments used are described in Table 5.2. Soils
were placed in 5 cm diameter x 20 cm deep pots and kept well-watered with de-ionized
water to avoid adding nutrients or contaminants. Treatments were randomly assigned
to pots within replicate. To compare soil microbial communities between bulk and
rhizosphere soil, I included one root exclosure tube in each pot. Root exclosures were
sewn into 15 cm x 2 cm diameter cylinders from 20-micron mesh, and the seams were
sealed with silicone sealant. The exclosure tube was ﬁlled with soil and placed in the

center of the pot.

All plant species were added as 1 month old seedlings; all seedlings were germinated in
a sterile sand medium in a grth chamber and were less than 2 cm in height at
transplantation. One plant per species was added to each pot; for mixtures, seedlings
were planted at equal distances from each other with the root exclosure in the center. If
the transplant was unsuccessful, the seedling was replaced for up to four weeks into the

experiment.

Each segment of the experiment ran for a total of 18 weeks to ensure that the roots had

ﬁlled the pot. I ran the experiment in three time blocks because of the number of

samples and the time required to process each sample. Thus, time was an additional

113

Table 5.2. Treatments used in the Randomized Complete Block Design to test for the
effects of plant species on the soil microbial community. Each treatment was replicated
three times within three Time blocks on two different soils. Species planted were G =
Bromus inermis, F = Solidago canadensis, L = T rifolium pratense. 0 = absent, + =
present, C = control (no plants).

 

Plant Treatment Bromus inermis Solidago canadensis T rifolium pratense

 

C O O 0
G + 0 0
F O + O
L 0 0 +
FL 0 + +
GF + + 0
GL + O +
GFL + + +

 

114

factor in the ANOVA. The ﬁrst three replicates grew from December 1999 until May
2000, the second three replicates grew from February to June 2000, and the third set of
three replicates grew from November 2000 until March 2001. Temperature in the
greenhouse from 25 to 40 °C; light availability was controlled through a 12 h light/ 12 h

dark cycle.

Data Collection

Differences among treatments were assessed by (1) plant growth, (2) the inorganic
nitrogen pool, and (3) N-mineralization rate. At harvest, I determined the dry mass of
above and belowground plant biomass. I separated root biomass from the soil (and
other species’ roots) during sieving. To avoid confusing roots from different plant
species, root systems were kept intact. Roots were rinsed thoroughly in de-ionized
water before drying. The root and shoot biomass for each species were measured
separately. All plant material was dried at 60°C for 48 h. Shoot biomass was dried

separately from root biomass.

After the roots were removed, soil from the entire pot (rhizosphere inﬂuenced) was
passed through a 2 mm sieve and stored in sealed plastic bags at 4°C until analyzed.
Soil from the exclosures was sieved and stored separately. All analyses were done
within 3 days of sampling, except PLFA. Soil for PLFA analyses was kept at —80°C

until the fatty acids were extracted.

115

I determined gravimetric soil moisture, total inorganic nitrogen, and N-mineralization
and nitriﬁcation rates for each sample as described in Chapter 4. I assessed differences
in soil microbial community production among treatments by (1) microbial biomass C
and (2) microbial respiration. The method used to determine microbial respiration and
the chloroform fumigation incubation method used to determine microbial biomass are
described in Chapter 4 (see also Paul et al. 1999). I assessed soil microbial community
structure differences among treatments by PLFA proﬁles as described in Chapter 4.
For analysis, I included only those phospholipid fatty acids that were present in greater
than 50% of samples and reported their abundance as the square root of the proportion
of the total phospholipid fatty acid amount in each sample (Hellinger transformation).
Of the 70 lipids detected, 31 phospholipid fatty acids met this criterion for the entire
experiment (n = 144) and 34 phospholipid fatty acids met this criterion for the subset (n

= 58) analyzed with root exclosures (Table 5.4).

Statistical Analyses

I used a randomized complete block design (ANOVA) model to test the effects of soil
origin, presence/absence of each plant species, and time on the following response
variables: soil moisture, plant biomass, relative growth rate, total N, N-mineralization
rate, nitriﬁcation rate, microbial respiration, and microbial biomass (Table 5.3). The
plant treatments were analyzed in three ways. (1) I included all 8 plant treatments as
independent factors in the ANOVA (Table 5.3A) to assess the effect of each plant

treatment on the soil and microbial characteristics. (2) Because in Chapter 4 I found

116

large effects of the presence of a plant on many soil and microbial variables, to
determine the effects of the plants on soil and microbial characteristics, I ran the
ANOVA with only the 7 plant treatments, excluding the no-plant control (Table 5.3B).
(3) To determine the compositional effects of Bromus inermis, Solidago canadensis,
and T rifolium pratense on soil and microbial characteristics, I used a 2 x 2 x 2 factorial
design for the plant treatments in the ANOVA (Table 5.3C). As carbon should be
limiting to the microorganisms, the size of the microbial community should depend on
the input of carbon to the system. Because the plant material is the primary source of
new carbon to microorganisms, then the microbial biomass should be proportional to
the belowground plant biomass in any one spot. To determine if signiﬁcant treatment
effects on soil community structure or function were the indirect results of changes in
plant biomass, I also used root, shoot, and total plant biomass as co-variates in all three

analyses.

1 used a modiﬁed redundancy analysis (RDA, Legendre & Anderson 1999) to
determine the relationship between the environmental factors (soil origin,
presence/absence of each plant species, time of harvest) and the PLF A measure of soil
microbial community structure. I transformed the PLFA data using a Hellinger
transformation because this transformation does a good job of handling data matrices
with many zero values (Legendre & Gallagher, in press). The RDA procedure involved:
(1) the creation of a matrix of dummy variables corresponding to the randomized
complete block design (modeled from the experimental design: soil origin,

presence/absence of each plant, time at harvest), (2) redundancy analysis of the

117

relationship between the principal coordinates (matrix of optical density or
phospholipid fatty acid data) and the environmental variables (matrix of dummy
variables in (1)), and (3) implementation of a Monte Carlo permutation test to estimate
the statistical relationship between the two matrices (Legendre & Anderson 1999).
This analysis allows me to test which factors from the experimental design (soil origin,
presence/absence of each plant, time at harvest) are signiﬁcantly related to the variation
in the PLFA patterns. The modiﬁed RDA is a better statistical technique than regular

ordination techniques because it allows for signiﬁcance testing.

Results

Plant Communities

As presented in Chapter 4, the two ﬁeld sites differed signiﬁcantly in peak

aboveground plant biomass (live and litter) and species composition (Table 5.1). The

Upper Louden (UL) ﬁeld had 320 i 23 g/m2 aboveground plant biomass and 15.8 i 1.0
species per square meter, while Field K (FK) had 592 i 22 g/m2 aboveground plant

biomass and 1.3 i 0.2 species per square meter. Field K was dominated by Bromus
inennis with very little Agropyron repens (Table 5.1). Upper Louden ﬁeld supported a
complex forb-dominated community, which included Poa compressa, T rifolium
pratense, Andropogon virginicus, several Hieracium species, Danthonia spicata,

Rudbeckia hirta, several Solidago species, Antennaria plantaginifolia, Achillea

118

millefolium, Centauria maculosa, Chrysanthemum leucanthemum, several Panicum

species, and a species of Rubus (Table 5.1).

Time Effects on Soil and Microbial Processes

Time signiﬁcantly affected most soil and microbial variables (Table 5.3A). The ﬁrst
two time groups used soil collected in 1999, while the last time group used soil
collected in 2000. The last experimental group (time 3) had signiﬁcantly higher total
inorganic nitrogen pools (p < 0.001, F = 83.3), N-mineralization rates (p < 0.001, F =
43.5), and nitriﬁcation rates (p < 0.001, F = 39.4). Microbial respiration (p< 0.001, F =
8.0) and microbial biomass (p < 0.001, F = 17.8) also differed signiﬁcantly among time
blocks (Table 5.3A). The time effect seems to be the result of the two different soil
collection times. Because patterns were consistent among time blocks and there were
no signiﬁcant interactions between Time and other factors, I did not investigate the

Time effects in any further detail.

Plant Species Responses to Soil Types

Plants grown in soil from ﬁeld FK, the higher fertility site, attained higher biomass than
those grown in soil from ﬁeld UL (Figure 5.1A-C, Table 5.3B). There were also
signiﬁcant differences among the seven plant treatments in total plant biomass (Figure
5.1A, p < 0.001, F=5.2), root biomass (Figure 5.2A, p < 0.001, F = 11.2), and shoot

biomass (p< 0.05, F = 2.6). In FK soil, Bromus inermis had the highest biomass and

119

.<m.m 03::. 8 :08: 0:: <>OZ< 0:: 80:: 82:>
00::0::8w:m 8808 0:: 83:> AU: ::: Am: :0: ”:88 :::::::m a: ::08 0:: 82:> A3 :0: .33: :0::0»: 883 AU :8:
Aviv v: 20:: Am 80:: :0: :0: 80880:: 8:8 :3 80:00:: 8:8 ::: ”Gm 03: :. 8 0:: 80:00 80880:: 80880:: 8:8 :::
88:0 =0: 9:: :3 :0:00 0808:: 23:03:: :8: 88030800 0:038: 8220:: 880%: :0 8:803 8:3 :80 :1 ._.m 0:53:

 

AD 0
0808:: 88:03:: 9 vi 0
8.8080500 0:00:20: 0 x=Om

 

 

 

8.8202: 880:: I

80880:. 8:7:
3:040 :0 A: A : O

 

 

 

 

 

_ ﬂ _ o o _ — d A d _ a
E < l l a. 8 l
_ - m w u e -. e .H. m.
j .. N 8 N 8 r + 8
m. m. W m.
w m. + w
B B e
r r m «w m m w r + w
.m .0 .0.
U m <
_ L : _ : _ _ v v _ _ : : : : :
3:: 5:53 :25: 0:: v: 20::

80880; 8:i
3:640 :0 A: A : O

 

80888:. 8:5
4:030 :0 A: A : O

 

 

 

 

120

<0: 2%,: a :22: 2: <>0z< 2:: 82:

328 0083583 48:08 0:: 838 A0: :8: Am: :0: ”:80 ::::8::: H 8:08 0:: m08_:> A43 :0 : .33 80:80: 08:3 :0:
:8: CE: V: 20:: Am: 80:: :0: :0: 80880:: 8:8 :3 360:: 8:8 :8: ”N6 03: H 8: 0:: 8:00 80880:: 808::08 8:8
:8: 83:0 :0: £3 :3 :0:00 0808.8: 80:05:: :8: 88000800 0:00.00: 88:08.: 880$ :0 3:803 :00: .m.m 0:83:

0:80:08: 88:08:: 8
88000200 0:00.00: a
88:08 880$ I

80880; 8:7:
4:040 :0 A: A : 0

E:
<<<
r m<-

m<

 

 

 

 

_ _ _ h _ _ _

 

3:: 80:84 :0:—:3 D

O. ‘0. O.
— o o
(8) ssewoyg 1003

V!
I—

O.
N

wﬁDSH—NDPH «GM—Am

3:0A0:0x:: 1: : 0

 

0:: v: 22.8

<3. V? O.
'— O O
(8) ssmuoyg 100)]

V?
v-!

o.
N

 

ADO
vi.

£Om

 

 

 

80880; 8:5
3:040:0 A: H : 0

:_:__J.
86

 

 

 

 

 

 

0.0

md

o.:

W:

o.m

(8) ssmuogg 1003

121

Table 5.3. Summary of ANOVA results analyzing effect of soil origin and (A)
individual plant treatments, (B) individual plant treatments without no-plant controls, or
(C) factorial plant treatments on plant and soil and microbial variables. Species codes
in 3C are: BROIN, Bromus inermis; SOOCA, Solidago canadensis; TRFPR, T rifolium
pratense. NS = not signiﬁcant, p > 0.05. ANCOVA results, using total plant biomass
as the covariate, are shown in parentheses if the effect changed in signiﬁcance.

 

 

(A)

Variable SOIL TRT TIME SOIL*TRT

Total Biomass (g) < 0.001 < 0.001 < 0.05 < 0.001

Root Biomass (g) < 0.001 < 0.001 < 0.001 < 0.001

Shoot Biomass (g) < 0.001 < 0.001 NS < 0.001

Root to Shoot Ratio <0.01 < 0.001 NS < 0.05

Total Inorganic < 0.05 (NS) < 0.001 < 0.001 NS

Nitrogen

N-mineralization Rate < 0.01 NS < 0.001 NS
(< 0.05)

Nitriﬁcation Rate < 0.01 NS < 0.001 NS
(< 0.05)

Microbial respiration < 0.001 < 0.001 < 0.001 < 0.01
(< 0.01) (< 0.05)

Microbial Biomass NS NS < 0.001 NS

 

122

(B)

 

 

Variable SOIL TRT TIME SOIL*TRT

Total Biomass (g) < 0.001 < 0.001 < 0.05 NS

Root Biomass (g) < 0.001 < 0.00] < 0.00] < 0.05

Shoot Biomass (g) < 0.001 < 0.05 NS < 0.001

Root to Shoot Ratio <0.01 < 0.001 < 0.05 NS

(NS)

Total Inorganic Nitrogen < 0.05 NS < 0.001 NS

N-mineralization Rate < 0.01 NS < 0.001 NS
(< 0.05)

Nitriﬁcation Rate < 0.01 NS < 0.001 NS
(< 0.05)

Microbial respiration < 0.001 NS < 0.01 < 0.05
(< 0.01)

Microbial Biomass NS NS < 0.001 NS

 

123

 

(C)

 

 

Variable SOIL BROIN SOOCA TRIPR TIME
Total Biomass (g) < 0.001 < 0.001 < 0.05 < 0.001 < 0.05
Root Biomass (g) < 0.001 < 0.001 NS < 0.01 < 0.001
Shoot Biomass (g) < 0.001 < 0.01 < 0.001 < 0.001 NS
Root to Shoot Ratio < 0.001 < 0.001 < 0.05 < 0.05 < 0.05
Total Inorganic NS < 0.001 < 0.001 < 0.01 < 0.001
Nitrogen (< 0.05) (< 0.01) (< 0.05)
N-mineralization < 0.001 NS NS < 0.05 < 0.001
Rate (NS) (NS)
Nitriﬁcation Rate < 0.001 NS NS < 0.05 < 0.001
(NS) (NS)
Microbial respiration < 0.001 < 0.01 < 0.001 < 0.01 < 0.001
(< 0.05) (NS) (< 0.01) (NS) (< 0.05)
Microbial Biomass NS NS NS NS < 0.001

 

124

T rifolium pratense had the next highest biomass (Figure 5.1A-C). In UL soil, Bromus
inermis and Solidago canadensis had higher biomass than T rifolium pratense (Figure
5.1A-C). However, the diversity effect on productivity was non-additive: mixtures did
not have higher biomass than monocultures (Figure 5.1A). Bromus inermis and
T rifolium pratense dominated mixtures in FK soil (Figire 5.1B), while Bromus inermis

tended to dominate the mixtures in UL soil (Figure 5.1C).

Plants grown in the higher fertility soil (FK) had higher root biomass than those grown
in soil from ﬁeld UL (Figure 5.2A-C, p < 0.001, F = 21.9). Bromus inermis had the

highest root biomass in both PK and UL soils (Figure 5.2B,C, Table 5.3C).

Bromus inermis and T rifolium pratense grown in the lower fertility UL soil had higher
root to shoot ratios than plants grown in FK soil (Table 5.4). Bromus inermis, the
grass, had two times higher root to shoot ratios than both Solidago canadensis and
T rifolium pratense, regardless of soil origin (p < 0.001, F = 42.5). Plant species did not
appear to change their root to shoot ratios in response to the presence of competitors

(Table 5.4).

Effects of Soil Origin on Soil and Microbial Processes

Soils from the FK site had signiﬁcantly higher total inorganic nitrogen pools (p < 0.05,

F = 5.7, Figure 5.3A), N-mineralization rates (p < 0.01, F = 9.0, Figure 5.38),

nitriﬁcation rates, (p < 0.01, F = 8.5, Figure 5.3C), and microbial respiration (p < 0.001,

125

cc.—
2.—
id
4&0
No.0 H Sod
Smo H two
@od H co;
imp—U

8.. 558 .
83on m

cote Emucﬁm H

26H ocd

m_.onm.N

4+

No.0 H ovd

wm.oH3.N

4+

SEE 8m mo:_m>

mmd H «5.—

hmd H mud

m+

mod H end

Omd H ow.—

m+

8.5332 86on N

m_.ova._

:dH hwd

omd H mwd

deoﬁo

m_.onm._

nod H cod

3.0 H 05m

9:51-96:62

cod H nod

006 H 05.0

:deo;

2.5—:35:

3V 3.2295.
sea:

EV ﬂutmhomoo
Sagan

6v

2::on §Soxm

SE 5.53
.5an

3V 3.28on
naught

CV nwwtmnomoo
assay.

6V

.3:me 3 Beam

0:: v. 22m

.Um.m 038i E 3%: 2m <>OZ< 2: Eoﬁ mos—g oogocEﬂm
.EoEEu: 28 :8 some 5 $6on :83 some .59 8:8 80% 8 Hood .vw 2an

126

.<m.m 033. E 37.: 2a <>OZ<
2: Bot 82? 353$in :88 Emvcﬁm H SEE 2m 82a> Ame—”Eu 30:2: 15 98 @386 3:5 vi ”8%
oz: 2: 50253 uozﬂzwcumﬁ 2m 8353 diet? some com .3285 3582:. Am: 95 death—v.2 3580:: R: .28
:ocaomtzc ADV £88 :EENmEoEEi Amv .coon: oanSE .88 2V E 355235 :83 wcoEa c252; .m.m 2sz

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

H5860; Ema E258; EmE
iEO 40 LO :5 A m 0 U .50 40 m0 ‘5 i— n— O U
u — _ q A d 1 — O — d a N A m )
\.l 10 WHO
i . 18. cm W i W i 3 M.
D m. i . Av 1w :0 m
. O m + + 3 m..
i r 0 -SN z. m.. r 6 a / e
3 m... AV co .1
/ .l 10— p H
L. % Av S 8 .M 9
i a icon Pm. is L. m.
w L. m w i . + + is mm
as. m. x 5 o + + i: ../i m.
r .. (U. vi 0 I * w. u
m 58 .. . oi: M
_ _ p _ _ _ p r cam _ _ _ _ _ _ p F m—
EoEEoci Ema E0823; EmE 25:58:. EmE
.5040 LO 4... i— n. O U iEOi—O m0 i: i— m D U iEOi—O LO 4“— i. m 0 U
. _ d O ai 9 O _ 4 _ S — a O
Q% ) C H i: e 6 10— l
1 Aw * 1N aim 1. +% IN ml .N I . + ﬂ ION) m9
+ N W co W. I v V 10m“ H...
v 0 m ..v N u no u
1 i s/ U i 3 m r Low N m
co 0 D. mi I now /c.o
T P ic P m. i + S M m. cm. w
,M m m n... n a room m.
i in mun T iwmm i 13mm.
../.. m. D. 1 an H n
H i
P a .m e 1 3(aw
l lO—rw l lO—( m r. 1 9
i . 52 u
U r m <
. L _ _ ~ _ _ _ N~ _ _ _ _ _ _ _ N_ _ _ _ L _ L _ . :

 

 

 

 

 

 

 

 

 

 

 

 

127

F = 22.6, Figure 5.3D) than those from the UL sites. However, soil origin did not affect

soil microbial biomass (Table 5.3A, Figure 5.3B).

The differences in soil and microbial variables between the two sites were partly due to
differences in plant biomass produced on the two soils (Figure 5.1A-C, Table 5.3A,C).
When the analyses were re-run with total plant biomass as a covariate, total inorganic
Nitrogen was no longer affected by soil origin (Table 5.3A) and N-mineralization and
nitriﬁcation rates were not signiﬁcantly inﬂuenced by soil origin (Table 5.3C). In
contrast, when root biomass was the covariate, total inorganic Nitrogen (p < 0.05, F =
4.5) and N-mineralizaton (p < 0.01, F = 8.6) and nitriﬁcation rates (p < 0.01, F = 8.3)

were still signiﬁcantly affected by soil origin.

Plant Effects on Soil and Microbial Processes

The presence of any plant signiﬁcantly reduced total inorganic Nitrogen (Figure 5.3A,
p < 0.001, F = 13.0) and signiﬁcantly increased microbial respiration (Figure 5.3D, p <
0.001, F = 8.7) on both soils. However, the composition of the plant community had no
effect on either of these variables and there were no signiﬁcant plant treatment effects

when the no-plant controls were excluded from the analysis (Table 5.3B).

This experimental design allowed me to test for the effects of the presence of speciﬁc

plant species on these processes (Table 5.3C). The presence of T rifolium pratense was

signiﬁcantly correlated with higher N-mineralization rates (Figure 5.3B, Figure 5.4B,E,

128

 

p < 0.05, F = 5.1) and nitriﬁcation rates (Figure 5.3C, p < 0.05, F = 5.1) and this effect
was still detectable when root biomass was used as the covariate (p < 0.05, F = 4.5; p <
0.05, F = 4.5, respectively). However, increasing the number of plant species (or
functional groups, the same in this design) did not have additive effects on any of the
soil or microbial processes. Mixtures did not have detectably higher nitrogen pools or

process rates than monocultures (Figures 5.3A-E).

Interactive E ﬂeets of Plant Species and Soil Origin on Soil and Microbial Processes

In this study there were few signiﬁcant interactions between soil origin and plant
treatments that affected soil and microbial processes (Table 5.3A-B). Plant biomass
was signiﬁcantly affected by the interaction between soil origin and plant treatments
(Figure 5.1A, p < 0.001, F = 28.3). For example, Solidago canadensis had the lowest
total biomass in monoculture in FK soil, while T rifolium pratense had the lowest total
biomass in monoculture in UL soil (Figure 5.1A). Additionally, microbial respiration
was signiﬁcantly affected by the interaction between soil origin and plant treatment
(Table 5.3A, p < 0.01, F = 3.1), primarily because the control, Solidago canadensis
monoculture, and three species mixture (GFL) did not differ signiﬁcantly in microbial
respiration between the two soils, PK and UL (Figure 5.3D). Although there was no
signiﬁcant interaction between soil origin and plant treatment for N-mineralization and
nitriﬁcation rates (Table 5.3A,B), in mixtures in the FK soil both tended to decrease in

the presence of Solidago canadensis and increase in the presence of T rifolium pratense,

129

but were less responsive in the UL soil (Figure 5.3B-C). Total plant biomass and root

biomass had no effect on this relationship.

Plant and Soil Effects on Soil Microbial Community Structure

PLF A proﬁles of the soil microbial community varied between the soils from the two
sites and also responded to the presence of the three plant species (Table 5.6A). Soil
origin accounted for 12% of the variation in PLF A proﬁles (p < 0.001, F = 24.4), while
Time accounted for 10% (p < 0.001, F = 20.4) and Replicate an additional 5% of the
variation (p < 0.001, F = 10.6). The three plant species had small but signiﬁcant effects
PLF A proﬁles, each accounting for only 1% of the variation (Table 5.6A). Axis I from
the Redundancy Analysis accounted for 20.7% of the variance in phospholipid fatty
acid data, 41.1% of the variance in the phospholipid fatty acid -environment
relationship and had a phospholipid fatty acid -environment correlation of 0.853 (Table
5.58). Axis 2 accounted for 12.4% of the variance in phospholipid fatty acid data,
25.2% of the variance in the phospholipid fatty acid -environment relationship and had
a phospholipid fatty acid -environment correlation of 0.852 (Table 5.6B). Axis 3 and
Axis 4 accounted for 5.1% and 2.5% of the variance in phospholipid fatty acid data,

respectively (Table 5.68).

Figure 5.4 shows the separation of samples by environmental factors and fatty acids as
related to the axes from the Redundancy Analysis. PLFA proﬁles ﬁ‘om UL soils loaded

higher on PC2 than proﬁles from FK soils (Figure 5.4). UL soils had higher amounts

130

Table 5.5. Fatty acids used in principal components analysis of PLFA proﬁles. I
describe fatty acids using standard nomenclature where the total number of carbon
atoms appears before the colon and the total number of CC bonds appears after it.
Cyclo-propane analogs are indicated by "cyclo," and the location of the epoxy bond is
indicated by a "c" followed by two numbers. If the cis or trans conﬁguration is
unknown, the word "at" is used. The number following "cis," "trans," or "at" indicates
the location of the double bond in relation to the carboxyl end of the molecule. Fatty
acids with the same retention time are grouped as "sum in feature" and given a unique
number designation.

 

 

Phospholipid Fatty Acids Phospholipid Fatty Acids
(Full Experiment) (Exclosure subset)
C9 Dicarboxylic acid C9 Dicarboxylic acid
14:0 14:0

15:0 iso 15:0 iso

15:0 anteiso 15:0 anteiso

15:0 15:0

16:0 iso 16:0 iso

16:1 cis7 16:1cis7
16:1cis9 16:1cis9

16:1 cis 11 16:1 cis 11

16:0 16:0

iso 17:1 G iso 17:1 G

17:0 iso 17:0 iso

17:0 anteiso 17:0 anteiso

17:0 cyclo 17:0 cyclo

16:1 2OH 17:0

18:1 trans 9 alcohol 16:1 20H

18:2 cis 12 18:2 cis 12
18:1cis9 18:1cis9

18:1cis 13 18:1 cis 13

18:0 18:0

19:0 cyclo c11-12 19:1 trans 11

20:0 19:0 cyclo cl 1-12
19:0 cyclo 11-12 20H 20:4 cis 14

22:0 20:0

22:0 2OH 19:0 cyclo 11-12 20H
24:0 22:0

23:0 20H 22:0 20H
Coprostane 24:0

Unknown 25.339 23:0 20H
Cholesteryl-palmitate Coprostane

summed feature 8: 18:1 trans 9 Unknown 25.339
Cholesteryl-palmitate
summed feature 8: 18:1 trans 9 -
summed feature 12:

 

131

 

 

 

 

 

 

2
c,\"
‘1: 1— §0_il.
2 Plant FK UL
N" C o o
8 G A 2..
Te 0" F v
.9. L 4 <1
8 FL .1 O
5 CF I D
U -1" GL O O

GFL * H
_2 1 l 1
-2 -1 O 1 2

Canonical PC1, 20.7%

Figure 5.4. Distance-based Redundancy Analysis of Phospholipid Fatty Acid proﬁles,
investigating the effects of soil origin, plant, and time effects on the structure of the soil
microbial community. Canonical principal component plot for the full RDA. The PLFA
proﬁles are distinguished between soil microbial communities from FK soil (solid
symbols) and UL soil (open symbols). Soil microbial communities in soils with
different plant communities are indicated by symbols: circles. controls; triangles,
monocultures; four or ﬁve-sided polygons, mixtures. Labels for phospholipids are listed
in Table 5.4. Signiﬁcance values from the RDA are listed in Table 5.5A.

132

 

.<m.m 2an E Ba:
23 <QM on“ 8on was?» comma-«Ema 4mm 2an E @on2 2m mEazoanona com £083 .50 8:393 DE: 28
£58th owonzom. new 353% .8823“ .Ewto =8 53, 2 mo C mus 3:895. Snags-m CV .30 “SB—Etna be:
use .3533: 533x25 was 353m .8329: ...Ewto =8 53> 2 mo 3 £88 9.35.3386 omonmom. Amv .30 8:3th
be: can £3253: Eatobﬁ Ea omehaom 6:30:98 .Ewto =8 53> 2 90 C mos MESS 359$ $3 :8an
«mm-E mo marge 5:32:68 o5 Bow Bigot moo—Q Eoconﬁoo RENEE. 303058 :2th 6-5 .m.m oSwE

 

 

 

owtﬁoxm ESE-bus ﬂatmhoaoo omohzom $5on 353%.

+ o . + o . + o .

H H MOI H H MOI ‘ H “O:
r ivd- i + ind- Lio-
- $ -3- m - -3- m - -3-
i imd- m. i iod m. i imd-
1 -_.o- m i L6 m i L6-
1 iod M i 1N6 M i iod
- -3 .0 - -3 .0 - -3
i Lmd 1. r Lac .0 i imd
- -3 m .. -3 WW - -3
r + oil. 1 m 19o i <-v.o

_ _ md _ L 5o _ _ md

 

 

 

 

 

 

 

 

 

%Z'l ‘ I Dd 19311101193

133

Table 5.6. Effect of soil origin, time, replicate, and presence of Bromus inermis
(BROIN), Solidago canadensis (SOOCA), and T rifolium pratense (TRFPR) on the soil
microbial community as detected by Distance-Based Redundancy Analysis (db-RDA)
of PLFA proﬁles. (A) Signiﬁcance values for the permutation tests on the
environmental factors of the RDA. (B) Variance explained by species data and species-
environment correlations for the RDA.

 

 

 

 

(A)
Factor Lambda F—stat p-value % variance
Soil 0.12 24.4 0.001 12.2
BROIN 0.01 2.3 0.017 1.1
SOOCA 0.01 1.9 0.050 0.9
TRFPR 0.01 2.5 0.010 1.2
Time 0.10 20.4 0.001 10.3
Rep 0.05 10.6 0.001 5.3
Sum 0.50 2.7 0.001

(B)

 

Axis Eigenvalue Species-Environment Cum % Variance Cum % Variance of

 

Correlation of Species Data Species-Environment
1 0.207 0.853 20.7 41.1
2 0.124 0.852 33.1 66.3
3 0.051 0.761 38.2 76.5
4 0.025 0.589 40.6 81.4

 

134

of 18:1 cis 13, 16:1 cis 11, and 18:2 cis 12, indicating a higher fungal to bacterial ratio
and more gram-negative bacteria (Harwood and Russell 1984). Partial canonical
principal components derived from the redundancy analysis of PLFA proﬁles revealed

that soil microbial communities grown under Bromus inermis had higher amounts of

16:1 cis 9, 16:1 cis 11, 18:1 cis 9, and summed in feature 8 (Figure 5.5A).

Communities grown in the presence of Solidago canadensis had higher 18:1 cis 13,
16:1 cis 7, 16:1 cis 11, and summed in feature 8 (Figure 5.5B), while communities
grown in the presence of T rifolium pratense had higher 18:1 cis 13, summed in feature
8, 16:0 iso, iso 17:1 G, 16:1 cis 9, 17:0 cyclo, and 18:1 cis 9 (Figure 5.5C). This
indicates that the microbial communities under all three plants had a higher proportion
of gram-negative bacteria than microbial communities in the no-plant controls. In
addition, microbial communities grown in the presence of T rifolium pratense had a
higher proportion of gram-positive bacteria than soil microbial communities not
exposed to T rzfolium pratense. Overall there were few differences in PLFA proﬁles

among replicates (within time) or time.

A Comparison of the Microbial Communities within and outside the Root Exclosures

I used root exclosures to exclude rhizosphere effects and create “bulk” soil in these

pots. I compared the structure of the soil microbial community of the “bulk” soil to that

of the rhizosphere soil using PLFA proﬁles. The RDA of the PLFA proﬁles showed no

135

 

Canonical PC2, 10.1%

 

 

 

 

 

3 I I
<5
2 L— _
O
. Exclosure
' o +
1 — w .- 301! 0
Q UL O a.
““ PK 0 A
F» r?
0 H :2" ‘
L13: ;- _
Er
( (1
-1 l —
-2 -l 2

 

Canonical PC1, 28.9%

Figure 5.6. Distance-based Redundancy Analysis of Phospholipid Fatty Acid proﬁles,
investigating the effects of exclosure, soil origin, plant, and time effects on the
structure of the soil microbial community. Canonical principal component plot for the
full RDA. The PLFA proﬁles are distinguished between soil microbial communities
from FK soil (solid symbols) and UL soil (open symbols).

136

Table 5.7. Effect of exclosure, soil origin, time, and presence of Bromus inermis
(BROIN), Solidago canadensis (SOOCA), and T rifolium pratense (TRF PR) on the soil
microbial community as detected by Distance-Based Redundancy Analysis (db-RDA)
of PLF A proﬁles. (A) Signiﬁcance values for the permutation tests on the
environmental factors of the RDA. (B) Variance explained by species data and species-
environment correlations for the RDA.

 

 

 

(A)
Factor Lambda F-stat p-value % variance
Exclosure 0.02 1.4 0.110 1.8
Soil 0.25 20.0 0.001 24.8 _
BROIN 0.02 1.8 0.060 2.2 ‘
SOOCA 0.01 0.9 0.520 1.1 ‘i
TRFPR 0.03 2.6 0.002 3.3
Time 0.02 1.6 0.060 2.0 -
Sum 0.79 2.0 0.001

(B)

 

Axis Eigenvalue Species-Environment Cum%Variance Cum%Variance

 

Correlation of Species Data of Species-
Environment
1 0.289 0.976 28.9 36.8
2 0.101 0.920 39.0 49.6
3 0.085 0.898 47.5 60.4
4 0.066 0.907 54.1 68.8

 

137

effect of root exclosure on the structure of the soil microbial community. However,
like RDA of the PLFA proﬁles for the whole experiment (above), the RDA for the
PLFA proﬁles of the soil microbial communities within and outside the root exclosures
also revealed variation between sites and detected an effect of the presence of T rifolium
pratense on the soil microbial community (Table 5.7, Figure 5.6). Axis I accounted for
28.9% of the variance in phospholipid fatty acid data, while Axis 2 accounted for
10.1% of the variance in phospholipid fatty acid data (Table 5.68). Soil origin
accounted for 25% of the variation in PLFA proﬁles (Figure 5.6, p < 0.001, F = 20.0),
while Trifolium pratense accounted for 3% of the variation (p < 0.01, F = 2.6).
However, there was no effect of root exclosure on PLFA patterns (Figure 5.6),
indicating that resources that had inﬂuenced the microbial community in these pots
were able to travel through the soil and away from the roots. Differences detected

among plant species were not due to direct contact with the roots.

Discussion

1 expected to ﬁnd that plant species had unique effects on the structure and function of
soil microbial communities and that the effects of different plant species on soil
microbial community structure and function are non-additive. I found that plant species
had unique effects on soil microbial community structure (as indicated by PLFA) and
some soil processes like total inorganic Nitrogen pools, N-mineralization and
nitriﬁcation rates, and microbial respiration. However, I did not ﬁnd additive effects of

plant diversity, and soil effects were much stronger than individual plant effects on the

138

 

soil microbial community. This study suggests the origin of the soil and the presence
of a plant both independently inﬂuence the structure and functioning of the soil
microbial community. Most previous studies have been unable to distinguish between
the effects of plants and the effects of soil origin on the structure of the soil microbial

community.

As I saw in the previous chapter, soil process rates were higher and nitrogen pools were
lower in the higher fertility soil (FK). However, because of the longer time frame
during which these experiments were conducted (4 months in Chapter 4 versus 16
months in this study), Time did signiﬁcantly affect soil and microbial process rates.
The presence of a plant signiﬁcantly affected soil and microbial processes and soil
microbial community structure. In addition, the effects of different plant species on soil
microbial community structure and function were non-additive. Process rates from
mixtures could not be determined by summing process rates from monocultures of the
plant species included in the mixtures. Finally, contrary to our hypothesis, soil
microbial communities from the root exclosures did not differ structurally (as measured

by PLF A) from soil microbial communities in soil closely associated with plant roots.

The Effects of Soil Origin on Soil and Microbial Processes and Soil Microbial

Community Structure

In this study, soil origin had a signiﬁcant inﬂuence on soil properties controlled by the

soil microbial community and on the structure of the soil microbial community itself.

139

In fact, soil origin explained more variation in soil microbial community structure
(PLFA proﬁles) than any of the other explanatory variables. This suggests that the
history of the soil and plant community plays a major role in structuring the soil
microbial community and, consequently, inﬂuences ecosystem functioning through the
soil microbial community. Many studies have shown site or soil characteristics to be an
important inﬂuence on soil microbial processes and structure. Zelles et al. (1992)
differentiated among eight agricultural management treatments using PLFA. Zelles et
a1. (1995) distinguished among three different soils in farmland and grassland using
PLFA. Bossio et al. (1998) determined that soil type was the most important
environmental factor in structuring the soil microbial communities of sustainable
agriculture systems in California. Groffrnan et al. (1996) measured microbial biomass
and activity and nitrogen transformation rates of soil taken from a range of old-ﬁeld
sites. As with our study, Groffrnan et al. (1996) concluded that the main driver for

microbial biomass and activity was soil type.

Plant Effects on Soil and Microbial Processes and Soil Microbial Community

Structure

As I saw in Chapter 2 (Broughton and Gross 2000), Chapter 3 (Broughton and Gross
2001), and Chapter 4, in this study, plants can have signiﬁcant effects on soil and
microbial processes and soil microbial community structure. Plant effects are complex
and include effects that will be mediated through diversity, species composition, and

individual plant species effects. This study cannot distinguish among the effects of

140

 

species diversity, functional group diversity, and plant community composition as
aspects of the overall plant effect because each functional group is represented by only
one species. However, I can address whether species differ in their effects on soil and
microbial characteristics, whether plant species effects are detectable in mixtures, and
whether these effects are enhanced by diversity. Which component or components
(diversity, composition, individual plant species) are most important inﬂuencing the
soil microbial community can be addressed by other studies like the BIODEPTH

experiment in Chapter 3.

The results from this study do not support the hypothesis that increased plant species
diversity leads to increased soil and microbial processes because monocultures differed,
but mixtures did not. The mixtures did not have higher process rates than the
monocultures. Similarly, Wardle (Wardle et al. 1999, Wardle et al. 2000, Wardle and
Nicholson 1996) has consistently shown no relationship between diversity and
ecosystem ﬁinction in a series of plant removal experiments in New Zealand perennial
grasslands. Symstad et al. (1998) also found no relationship between plant diversity
and ecosystem functions (other than productivity) in a plant removal study in a North

American grassland.

However, other experiments investigating the relationship between plant diversity and
ecosystem function have shown variation in the relationship between plant diversity
and ecosystem function across habitats. In Chapter 2, Broughton and Gross (2000)

found that there was a signiﬁcant effect of plant species diversity on the respiration or

141

biomass of the soil microbial community. However, in that study, changes in diversity
were correlated with changes in edaphic variables. My work at the Silwood, England
BIODEPTH site (Chapter 3) also showed a signiﬁcant relationship between plant
diversity and two measures of microbial community structure (CLPP and PLFA).
Similarly, results from the Swiss BIODEPTH experiment have shown positive
relationships between plant diversity and plant biomass (Spehn et al. 2000a, Spehn et
al. 2000b), soil microbial respiration and functional diversity (Stephan et al. 2000),
microbial biomass (Spehn et al. 2000a), and earthworm population density (Spehn et al.

2000a)

In Chapter 4, I saw that soil effects alone can inﬂuence soil and microbial processes,
but these effects could be mediated by plants. In this study, the plant effect on soil and
microbial processes seemed to be limited only to the presence of a plant: the identity of
the plant mattered very little for soil processes in this study. The exception was
T rzfolium pratense, which did affect inorganic nitrogen pools and nitrogen process
rates. In addition, I did ﬁnd small differences among the plant treatments in their
effects on soil microbial community structure. At the Swiss BIODEPTH site, Stephan
et al. (2000) also detected a legume effect on soil processes. T rifolium repens
signiﬁcantly increased diversity and activity of catabolic proﬁles of the soil microbial
community (Stephan et al. 2000). Many studies have shown plant composition effects
on ecosystem functioning. Symstad et al. (1998) found plant composition effects on
productivity and nitrogen retention, and Wardle et al. (1999) observed plant

composition effects on PLFA patterns in the soils ﬁom a plant removal experiment in

142

New Zealand grasslands. Hector et al. (2000) detected a relationship between the
species composition of litter and the decomposition rate at the Silwood Park, England
BIODEPTH site. Hooper and Vitousek (1998) determined that plant community
composition accounted for much more of the variation in nutrient cycling processes on
serpentine soil in California than just plant functional group diversity. Broughton and
Gross found some evidence to support this view in the relationships between plant
composition and microbial respiration and soil PLFA patterns at the Silwood Park,

England BIODEPTH site (Chapter 3).

Plant community composition should inﬂuence the soil microbial community through
inputs of carbon into the soil. The quality and/or quantity of the carbon available to the
microorganisms should inﬂuence which microorganisms thrive in a particular
environment: therefore, the identity of the plant species providing that carbon should
inﬂuence how ecosystem functions change (Paul and Clark 1996). Plants provide
carbon to the soil in two ways: (1) litter and (2) root exudation. Hector et al. (2000)
found large effects of plant litter composition on litter chemistry and decomposition
rate at the Silwood Park, BIODEPTH site. However, this study did not last long
enough to test litter effects. All differences in soil and microbial processes and soil
microbial community structure among plant treatments must have been driven by
differences in root exudation and root turnover. A recent study by Hamilton and Frank
(2001) showed that herbivory could cause plants to stimulate rhizospheric microbial
communities to increase nitrogen cycling and make nitrogen more available to the

plants.

143

 

Several studies have detected differences among the rhizosphere soil microbial
communities from different plant species. Grayston and Campbell (1996) distinguished
between the rhizosphere of hybrid larch (Larix eurolepis) and Sitka spruce (Picea
sitchensis) trees using CLPP at a woodland site and two plantations. Garland (1996)
and Westover et al. (1997) also were able to differentiate the rhizosphere communities
of several herbaceous plant species both in the ﬁeld and greenhouse using Biolog.
Miethling et al. (2000) used CLPP and PLFA proﬁles to distinguish among soil
microbial communities from different ﬁelds planted with both alfalfa (Medicago
sativa) and rye (Secale cereale). Miethling et al. (2000) determined that plant species
identity was the most important factor in determining microbial community
characteristics in the rhizosphere. Bachmann and Kinzel (1992) showed that plant
species exude different amounts of organic metabolites, indicating that individual plants
might have quite different effects on the soil microbial community. These studies
suggest that plant species differences in root exudates may be important in
distinguishing the composition of the rhizosphere communities. The results from my
study also indicate that it is possible to distinguish among soil microbial communities
grown in the presence of different plant species, as I could distinguish among the plant

treatments with PLFA proﬁles.

Plant species effects on soil microbial processes are oﬁen detected in greenhouse

studies; however, these effects are much more difﬁcult to ﬁnd in the ﬁeld. Buckley and

Schmidt (2001) found that there was little difference between the soil microbial

144

 

 

communities of a continuously tilled agricultural site and a companion successional site
in southwestern Michigan that had different plant communities for 12 years. Plant
species effects on soil microbial processes may take a long time to manifest themselves
(Buckley and Schmidt 2001, Broughton and Gross 2000), despite the possibility that
plant species may need only to change the process rates of a small proportion of the soil

organic matter to have large effects on soil processes (Wedin and Pastor 1993).

Implications of the Similarity of Soil Microbial Communities in the Root Exclosures

and the Rhizosphere Soil

In this experiment, all of the soil in the pot was available to the plant roots, except the
soil in the exclosure in the center. Root production was high in the experiment. Most
of the plants in the experiment were root-bound by harvest time. The 20—micron mesh
of the exclosure allowed nutrients, water, microorganisms, and mycorrhizae to pass
through, but not plant roots. Consequently, there was no direct “contact” effect of
plants on the soil and the soil microbial communities inside the root exclosure. In
effect, the soil outside the exclosure was entirely rhizosphere soil, while the soil inside
the exclosure was the equivalent of bulk soil. However, I could not detect any
differences (in PLFA proﬁles) in the soil microbial communities in the two different
types of soil (rhizosphere and bulk). This implies that the soil microbial communities
in the rhizosphere and bulk soils were not signiﬁcantly different in structure. This
suggests that root exudates that can inﬂuence soil microbial community structure

moved freely from the soil near the plant roots to the soil within the root exclosure.

145

This effect implies that in ﬁeld settings I often do not ﬁnd plant effects in the bulk soil
because the resources (exudates) are used by rhizosphere microorganisms before the

resources have a chance to migrate away from the roots.

The soil microbial community plays a crucial role in the ﬂow of nutrients and energy
through the ecosystem. Changes in ecosystem function are intimately tied to the
composition and activity of the soil microbial community. This study has provided
evidence that soil is the most important factor inﬂuencing soil microbial communities,
but the extant plant community composition can also inﬂuence soil microbial
community structure. Better understanding the role microorganisms play in the overall
functioning of the ecosystem can only improve our ability to predict how these
ecosystems will change in the future under the inﬂuences of changing vegetation cover,
agricultural practices, invasive species, and other dynamic drivers of the plant-soil

ecosystem.

146

Wm‘

 

REFERENCES

Bachmann, G. and H. Kinzel. 1992. Physiological and ecological aspects of the
interactions between plant roots and rhizosphere soil. Soil Biology and
Biochemistry 24: 543-552.

Bossio, D. A. and K. M. Scow. 1998. Impacts of carbon and ﬂooding on soil
microbial communities: Phospholipid fatty acid proﬁles and substrate utilization
patterns. Microbial Ecology 35: 265-278.

Bossio, D. A., K. M. Scow, N. Gunapala, and K. J. Graham. 1998. Determinants of
soil microbial communities: effects of agricultural management, season, and soil
type on phospholipid fatty acid proﬁles. Microbial Ecology 36, 1-12.

Burbank, D. H., K. S. Pregitzer, and K. L. Gross. 1992. Vegetation of the W. K.
Kellogg Biological Station. Michigan State University Agricultural Experiment
Station Research Report Number 510.

Buyer, J. S. and L. E. Drinkwater. 1997. Comparison of substrate utilization assay and
fatty acid analysis of soil microbial communities. Journal of Microbiological
Methods 30: 3-11.

Foster, B. L. 1996. Plant competition and diversity in relation to productivity in old-
ﬁeld plant communities. Ph.D. Dissertation. Michigan State University, East
Lansing, MI, USA.

Goodfriend, W.L. 1998. Microbial community patterns of potential substrate
utilization: a comparison of salt marsh, sand dune, and seawater-irrigated
agronomic systems. Soil Biology and Biochemistry 30: 1169-1176.

Grayston, S. J. and C. D. Campbell. 1996. Functional biodiversity of microbial
communities in the rhizospheres of hybrid larch (Larix eurolepis) and Sitka
spruce (Picea sitchensis). Tree Physiology 16: 1031-8.

Groffman, P. M., P. Eagan, W. M. Sullivan, and J. L. Lemunyon. 1996. Grass species
and soil type effects on microbial biomass and activity. Plant and Soil 183: 61-
67.

Haack, S.K., H. Garchow, D.A. Odelson, L.J. Fomey, and M.J. Klug. 1994. Accuracy,
reproducibility, and interpretation of fatty acid methyl ester proﬁles of model
bacterial communities. Applied and Environmental Microbiology 60: 2483-
2493.

147

Hector, A., A. J. Beale, A. Minns, S. J. Otway, and J. H. Lawton. 2000. Consequences
of the reduction of plant diversity for litter decomposition: effects through litter
quality and microenvironment. OIKOS 90: 357-371.

Hooper, D. U. and P. M. Vitousek. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

Legendre, P. and M. J. Anderson. 1999. Distance-based redundancy analysis: testing
multispecies responses in multifactorial ecological experiments. Ecological
Monographs 69: l —24.

Legendre, P. and E. D. Gallagher. 2001. Ecologically meaningful transformations for
ordination of species data. Oecologia (in press).

Miethling, R., G. Wieland, H. Backhaus, and C. C. Tebbe. 2000. Variation of
microbial rhizosphere communities in response to crop species, soil origin, and
inoculation with Sinorhizobiom meliloti L33. Microbial Ecology 41: 43-56.

Nelson, D.W. and LE. Sommers. 1982. Total carbon, organic carbon, and organic
matter, pp. 539-580, In_ Methods of Soil Analysis Part 2 - Chemical and
Microbiological Properties. ASA Special Publ. No. 9. ASA, SSSA, Madison,
WI.

Ohtonen, R., S. Aikio, H. Wire. 1997. Ecological theories in soil biology. Soil
Biology and Biochemistry 29: 1613-1619

Paul, EA. and FE. Clark. 1996. Soil Microbiology and Biochemistry. 2nd ed. San
Diego, CA: Academic Press, Inc.

Paul, E.A., Harris, D., Klug, M.J., and Ruess, R.W. (1999) The determination of
microbial biomass, pp. 291-317, In Standard Soil Methods for Long-Term
Ecological Research (eds G.P. Robertson, D.C. Coleman, C.S. Bledsoe, & P.
Sollins) Oxford University Press, New York.

Scoggin, H. J. 1978a. The Flora of Canada. Pt 2: Pteridophyta, Gymnospermae,
Monocyotyledoneae. National Museum of Canada. Crown Copyrights, Ottawa,
Canada.

Scoggin, H. J. 1978b. The Flora of Canada. Pt 3: Dicotyledoneae (Sauruaceae to
Violaceae). National Museum of Canada. Crown Copyrights, Ottawa, Canada.

Spehn, E.M., Joshi, J., Schmid, B., Alphei, J., & Korner, C. (2000) Plant diversity
effects on soil heterotrophic activity in experimental grassland ecosystems.
Plant and Soil 224, 217-230.

148

 

 

Spehn, E.M., Joshi, J., Schmid, B., Diemer, M., & Komer, C. (2000) Aboveground
resource use increases with plant species richness in experimental grassland
ecosystems. Functional Ecology 14, 326-337.

Stephan, A., A. H. Meyer, and B. Schmid. 2000. Plant diversity affects culturable soil
bacteria in experimental grassland communities. Journal of Ecology 88: 988-
998.

Symstad, A. J., D. Tilman, J. Wilson, and J. M. H. Kn0ps. 1998. Species loss and
ecosystem functioning: effects of species identity and community composition.
OIKOS 81: 389-397.

Tunlid, A., and DC. White. 1992. Biochemical analysis of biomass, community
structure, nutritional status, and metabolic activity of microbial communities in
soil, pp. 229-262, In Soil Biochemistry, vol. 7 (eds G. Stotsky and J.M. Bollag).

Wardle, D.A., K.I. Bonner, and GM. Barker. 2000. Stability of ecosystem properties

in response to above-ground functional group richness and composition.
OIKOS 89: 11-23.

Wardle, D.A., K.I. Bonner, G.M. Barker, G.W. Yeates, K.S. Nicholson, R.D. Bardgett,
R.N. Watson, and A. Ghani. 1999. Plant removals in perennial grassland:

vegetation dynamics, decomposers, soil biodiversity, and ecosystem properties.
Ecological Monographs 69: 535-568.

Wardle, D. A. and K. S. Nicholson. 1996. Synergistic effects of grassland plant
species on soil microbial biomass and activity: implications for ecosystem-level
effects of enriched plant diversity. Functional Ecology 10: 410-416.

Wedin, D. A. and J. Pastor. 1993. Nitrogen mineralization dynamics in grass
monocultures. Oecologia 96: 186-192.

Werner, P. A., I. K. Bradbury, and R. S. Gross. 1980. The biology of Canadian weeds.
54. Solidago canadensis L. Canadian Journal of Plant Science 60: 1393-1409.

Westover, K. M., A. C. Kennedy, and S. E. Kelley. 1997. Patterns of rhizosphere
microbial community structure associated with co-occurring plant species.
Journal of Ecology 85: 863-873.

Zak, J. C., M. R. Willig, D. L. Moorhead, and H. G. Wildman. 1994. Functional
diversity of microbial communities: A quantitative approach. Soil Biol.
Biochem. 26: 1101-1108.

Zelles, L., R. Rackwitz, Q. Y. Bai, T. Beck, and F. Beese. 1995. Discrimination of

microbial diversity by fatty acid proﬁles of phospholipid fatty acids and
lipopolysaccharides in differently cultivated soils. Plant and Soil 170: 115-22.

149

 

Zelles, L., Q.Y. Bai, T. Beck, and F. Beese. 1992. Signature fatty acids in
phospholipid fatty acids and lipopolysaccharides as indicators of microbial

biomass and community structure in agricultural soils. Soil Biology and
Biochemistry 24: 317-323.

I. r

 

150

CHAPTER 6
SUMMARY

From the collection of ﬁeld surveys and manipulative experiments in the previous
chapters I can make several conclusions. This dissertation suggests the origin of the soil
and the presence of a plant both inﬂuence the structure and functioning of the soil
microbial community. However, plant effects were often not as strong as I had
expected and there were other factors inﬂuencing the structure and function of the soil

microbial community.

First, I found that legacy effects can last for years. The disturbance caused during site
preparation at the Silwood Park BIODEPTH site was still inﬂuencing the soil microbial

community four years later (Chapter 3).

Plant effects on the soil microbial community were often the result of plant biomass. In
Chapter 2, plant productivity had signiﬁcant effects on soil microbial respiration. In
Chapter 3 many plant effects on soil microbial community processes were associated

with larger amounts of plant matter rather than the diversity of the plant community.

As I saw in Chapter 2 (Broughton and Gross 2000), Chapter 3 (Broughton and Gross
2001), Chapter 4, and Chapter 5, plants can have signiﬁcant effects on soil and
microbial processes and soil microbial community structure. Plant effects are complex

and include effects that will be mediated through productivity (Chapters 2 and 3),

151

 

 

diversity (Chapter 3), species composition (Chapters 3 and 5), and individual plant

species effects (Chapters 4 and 5).

In Chapter 4, I saw that soil effects alone can inﬂuence soil and microbial processes,
but these effects could be mediated by plants. In Chapter 5, the plant effect on soil and
microbial processes seemed to be limited only to the presence of a plant: the identity of
the plant mattered very little soil processes in this study. The exception was T rifolium
pratense, which did affect inorganic nitrogen pools and nitrogen process rates. In
addition, I did ﬁnd small differences among the plant treatments in their effects on soil
microbial community structure. However, these unique effects of plant species did not
seem to be additive in mixtures. Finally, root exclosures had no effect on the soil
microbial community structure, indicating plant resources can migrate away from the
roots. The large number of ﬁndings in which plant effects are limited to rhizosphere
soils in the ﬁeld imply that these plant resources are used before they have the

opportunity to migrate away from the roots in the ﬁeld.

The soil microbial community is a complex part of the old-ﬁeld ecosystem that controls
many important ecological processes. A multitude of factors inﬂuence the structure
and ﬁrnction of this community. Understanding the roles each aspect of the community
plays in structuring the soil microbial community will allow better understanding of the

effects of global change on these important processes.

152