a] J".

k...
..
:.
3.5%.
, _ .
. awn}
.- flit-.3 -- 1 .n— .
-1... r.........§......u. . nu... wink... .
k .1... :55... t... 3 z.
7 . éflnolili ~.W7.42t£
xv . . .

,. m... ..
$9.... «mm,

37..

..rl,.f £3)
. . .v
.3.
.. .1. 5
‘9..."
3.9%
.. a: . .
u
1.14 .
7.. I...

a z.
.7... :1... .

0:3! .1.
”5.4.2.... u

. ..
. 1... . ._ \2
.35.. filer .rv
. I 133-
a".
31.: 3... . .
‘2... z 1.... 5;...
3.2.3.... 1...... .
s x
.r...
..
=1
.: .
.5... .
r... . .
.35....méu.‘ ..
12.2. n..- :1...
VIII». H.311... l.n3_.-....a..3
. n . . ..
77-7 ivwx.$xs,.
Soliflln ,
.533...
1...
SJ.

:

.10 .u ..
x."

11......
I... (3:...
1L... .

. m“:

I. 35.2.:
L 1.3.:
.2.

13)....
a. .3, 9n
:3... 3....

‘.-\IO.VI.
u.1ls?

I. , ugt'v“
E59213",

IQN: 3.

 

 

llllllllllllllllll”(Illllllllllllllllllllllllllllllllllllll

193 01020 6104

This is to certify that the

dissertation entitled

DOMINANCE, DIVERSITY, AND RESOURCE COMPETITION

IN OLD-FIELD PLANT COMMUNITIES

presented by

Lisa Ellen Huberty

has been accepted towards fulfillment
of the requirements for '

Ph.D. degreein Botany and Plant Pathology

 

 

Wm...) :1. 414...;

Major professor

Date /0 NOU‘l’mbLg, 97":

MS U Lt an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY
Michigan State
Unlverslty

 

 

 

PLACE N RETURN BOXto monthl- Mum your record.
TO AVOID FINES Man on or Mon duo din.

DATE DUE D%T§WE DATE DUE

\ .

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

molommmwowlm

 

. DOMINANCE, DIVERSITY, AND RESOURCE COMPETITION
IN OLD-FIELD PLANT COMMUNITIES

By
Lisa Ellen Huberty

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

W. K. Kellogg Biological Station,
and
Department of Botany and Plant Pathology

1994

ABSTRACT

DOMINANCE, DIVERSITY, AND RESOURCE COMPETITION
IN OLD-FIELD PLANT COMMUNITIES

By
Lisa Ellen Huberty

I conducted species removal and resource addition experiments in a
first- and second-year old—field in southwestern Michigan to test two
mechanistic hypotheses about the relationship between diversity and
productivity: 1) plant diversity declines with increasing fertility due to
increasing competition intensity from the dominant species, and 2) the time
of resource addition within a season influences the magnitude of competition
intensity, and consequently the effect of a dominant species on community
structure.

In the first-year annual old-field the effect of the dominant species,
Winn; album, on subordinate community biomass, evenness, diversity,
and nitrogen acquisition depended on the time of nitrogen addition. When
nitrogen was added early in the growing season (early July) both the
dominant species and the subordinate community increased in biomass and
total nitrogen content relative to the unfertilized control. However, when
nitrogen was added mid- or late-season, the dominant species pre-empted the
added nitrogen and exhibited a disproportionate increase in biomass relative to
the subordinate community. The subordinate plant community had increased
tissue nitrogen concentrations but did not increase in biomass. The degree of
canopy development appears to determine the ability of dominant and
subordinate species to acquire added nutrients and allocate them to increased
growth rather than storage, and thus, influences the magnitude of the effect

of the dominant species on diversity and community structure.

 

 

 

 

In t

of species
--n'i

but not fit
addition r
this effect
of life-his:
nitrogen ;
higher in
resources
Fir
evaluate ‘

exDerirne

In the second-year old-field community the composition and abundance
of species in the subordinate community varied seasonally, while Cm
canadgnsis was the clear biomass dominant. Biomass evenness and diversity
but not richness declined with increasing Q, mademis biomass. Nitrogen
addition reduced evenness and diversity but not richness, but the magnitude of
this effect did not vary with the time of nitrogen addition. Rather the biomass
of life-history/ growth form groups and individual species varied with time of
nitrogen addition. Light penetration was lower and nitrogen availability was
higher in nitrogen addition plots early and mid-season, but by mid-August,
resources declined to the same low level in control and nitrogen addition plots.

Finally, I present a general graphical and statistical approach to
evaluate the absolute and relative response of a species biomass hierarchy to

experimental treatments or environmental perturbations.

I did this for me.

iv

ACKNOWLEDGMENTS

I am deeply grateful to my adviSor, Kay Gross, for her advice,
encouragement, patience, humor and unwavering support, and for helping
me develop a taste for Swiss cheese. I appreciated the discussions with her at
all the stages of my research and her critical review of numerous drafts of the
chapters. My other committee members critically reviewed an earlier draft of
my dissertation and contributed substantially to my professional development
at KBS. Special thanks to Mike Klug for introducing me to the world of
microbial ecology, to Susan Kalisz for introducing me to field methods in
evolution and natural selection and SAS, and to G. Phil Robertson for
stimulating my thinking about relating community ecology to ecosystem
processes. Each of the KBS faculty have provided me excellent training in
their ecology and evolution courses and seminars.

My fellow graduate students at KBS have been a great source of
stimulation and support. I am especially grateful to Bryan Foster for critical
discussions about my results and moral support at critical times, and to Paco
Moore for rescuing Chapter 4 from a black hole. I thank Walter P. Carson,
Kevin Kosola, and Bryan Poster for critically reviewing earlier drafts of the
chapters.

1 gratefully acknowledge the field and laboratory assistance of Olivia
Damon, Bryan Foster, Derek Fox, Proserpina Gomez, Kerri Gorentz, Sandy
Halstead, Heather Hobbie, Leslie Jagger, Michelle Keagle, Janice Krumm,

Carolyn Miller, Mindy Watts, and Pam Woodruff. I am grateful to Carolyn
v

Hammarskjold for her assistance obtaining references from the library, and to
John Gorentz and Stephan Ozminski in computer services for their technical
assistance. Bryan Foster and Lolita Krievs created the map in the introduction.
I thank Barb Eagle, Peggy Schultz, Jim Bever, Ann Miner, Rob Aiken,
and Marg Lansing for walking with me along the difficult and easy parts of
my path. The KBS quilting crew helped me balance the left and right sides of
my brain and kept me laughing. I am deeply grateful to Dan Dettweiler for his
gentleness, patience, support and encouragement during all the phases of my
work, for helping me balance my interests, and for cooking delicious,
nutritious meals.
This work was supported in part by several NSF grants: BIR 91-2006 for
the Center for Microbial Ecology at Michigan State University, DEB 92-1 1771
for Long-Term Ecological Research in Agricultural Ecology at Michigan State
University, and DIR 91-13598 for the Research Training Group at Kellogg

Biological Station.

vi

TABLE OF CONTENTS

LIST OF FIGURES .......................................................................................................... .x
CHAPTER 1 INTRODUCTION ........................................................................................ 1
Thesis organization ......................................................................................... 3
Description of field sites ................................................................................. 4
Description of plant communities ................................................................ 6
CHAPTER 2 DOMINANCE, DIVERSITY, AND RESOURCE COMPETITION IN AN
ANNUAL OLD-FIELD PLANT COMMUNITY ................................................................. 8
INTRODUCTION .................................................................................................. 9
METHODS ............................................................................................................ 12
Study site ............................................................................................... 12
Experimental design ........................................................................... 13
Measures of biomass, light penetration and nitrogen
acquisition ............................................................................................ 14
Indices of diversity and competition intensity ............................. 15
Statistical analysis ............................................................................... 17
RESULTS ............................................................................................................. 18
Biomass responses ............................................................................... 18
Effects on species richness, evenness, and diversity ................... 26
Competition intensity ......................................................................... 28
Light availability ................................................................................. 3O
Nitrogen acquisition by the dominant species .............................. 36
Nitrogen acquisition by the subordinate plant community ....... 38
Summary of results ............................................................................. 40
DISCUSSION ........................................................................................................ 44
Community biomass responses and nitrogen acquisition ........... 44
Diversity and community structure ................................................. 48
Magnitude of competition intensity and mechanisms of
plant competition ................................................................................ S 1
CHAPTER 3 Effect of time of nitrogen addition on community structure
and resource availability in an old-field plant community ................................ 57
INTRODUCTION ................................................................................................. .58
METHODS ........................................................................................................... .60

vii

Study site ............................................................................................... 60

Experimental design ........................................................................... 60
Soil sampling and nitrogen analysis ............................................... 61
Light measurements ........................................................................... 63
Statistical analyses .............................................................................. 63
RESULTS ............................................................................................................ .64
Seasonal patterns of community biomass, diversity, and
structure ................................................................................................ 6 4
Effect of time of nitrogen addition on total community
biomass, species richness, diversity and evenness ...................... 69
Responses of the dominant species and subordinate
community ............................................................................................ 69
Responses of life history/ growth form groups and
individual species to the time of nitrogen addition ...................... 77
Effects of time of nitrogen addition on seasonal patterns of
resource availability ........................................................................... 80
DISCUSSION ....................................................................................................... 86
Seasonal development of community structure ............................ 88
Effects of time of nitrogen addition on community
structure ................................................................................................ 89
Seasonal patterns of resource availability ..................................... 92
CHAPTER 4 RESPONSE OF A SPECIES BIOMASS HIERARCHY TO
EXPERIMENTAL TREATMENTS: A GRAPHICAL AND STATISTICAL METHOD ......... 96
INTRODUCTION. ................................................................................................. 9 7
METHODS ............................................................................................................ 100
Testing rank correlation of species abundances .......................... 100
Graphical and statistical analysis ..................................................... 100
Field experiments ................................................................................ 106
RESULTS ............................................................................................................. 107
Annual old-field: Relationship between species frequency
and abundance ..................................................................................... 107
Annual old-field: Correlations of species rank abundance ........ 1 10
Annual old-field: Absolute biomass responses .............................. 1 13
Annual old-field: Relative biomass responses .............................. 1 17
Second-year old-field: Absolute biomass responses ..................... 121
Second-year old-field: Relative biomass responses ..................... 126
DISCUSSION ........................................................................................................ 126
Consequences of spatial variability of species
distributions ......................................................................................... 126
Responses of the species biomass hierarchy to nitrogen
additions ................................................................................................ 132
Absolute and relative responses ....................................................... 133
Analogies to other approaches ......................................................... 133
CHAPTER 5 SUMMARY AND CONCLUSIONS .............................................................. 135
LIST OF REFERENCES ..................................................................................................... 140

viii

LIST OF TABLES

Table 1. Effects of species removal and time of nitrogen addition on total
community biomass, subordinate community biomass, ChenQdemm album
biomass, species richness, biomass evenness (J), and species diversity (H')
(ANOVAs). Error df for each ANOVA = 9. Treatment means are shown in Figure

2 and 5. ........................................................................................................................... 20

Table 2. Pairwise comparisons of the effect of time of nitrogen addition

on total community biomass, subordinate community biomass,

W album biomass, species richness, biomass evenness (J),

and species diversity (H') using Dunnett's t-test (alpha = 0.05, df = 9,

critical value of Dunnett's t = 2.812). ........................................................................ 21

Table 3. Five—letter Bayer codes (WSSA 1984) for weed species and their
corresponding scientific and common names. ...................................................... 25

Table 4. Effect of time of nitrogen addition on the absolute and relative
competition intensity indices (1 standard error, see Methods for
calculation of absolute and relative competition intensity). .............................. 29

Table 5. Results of analysis of variance for effects of species removal and time

of nitrogen addition on light penetration, tissue nitrogen concentration and

total nitrogen content of the dominant species, W 31mm, and the
subordinate community. Error df for each ANOVA = 9. Treatment means are
shown in Figures 8, 9, and 10. ................................................................................... 34

Table 6. Pairwise comparisons of the effect of time of nitrogen addition

on light penetration, tissue nitrogen concentration and total nitrogen

content using Dunnett's t-test (alpha = 0.05, df = 9, critical value of

Dunnett's t = 2.853) ..................................................................................................... .35

Table 7. Frequency of species in phenology plots (maximum = 4 plots)
and final harvest of treatment plots (August 16, maximum = 24 plots ). ........... 67

Table 8. Results of analysis of variance for time of nitrogen addition on
community biomass, species richness, biomass evenness, and biomass

diversity (numerator df = 5, error (if = 18). Dunnett’s t-test was used to

test for significant differences between the control and each time of

nitrogen addition (* = significant at alpha = 0.05, df = 18). ................................. 71

Table 9. Seasonal patterns in average soil nitrate and ammonium
concentrations (ug N/ g soil) in control (no nitrogen addition) and
nitrogen addition treatments in a second-year Michigan old-field. ................. 81

Table 10. Number of replicates out of four in which species were present and
their mean biomass (g/mz) in parentheses. ........................................................ 108

Table 1 1. Number of species occurring in both the control and nitrogen

addition treatments that were included in the correlation and

regression analyses, and number of species that were excluded because

they did not appear in both the control and the nitrogen addition

treatment. ..................................................................................................................... .1 1 1

Table 12. Spearman and Kendall tau-b correlation coefficients for rank
of species mean biomass in control and nitrogen addition plots. ...................... 1 12

Table 13. Regression statistics for the relationship between mean biomass of
species in control plots (no nitrogen added) and the change in mean biomass
between treatment (nitrogen added at different times of the growing season)

and control plots (absolute response, see Methods). Only species that occurred
in both control and treatment are included in the regression analyses. Values

in boldface indicate that the slope or y-intercept differs significantly from

zero. The corresponding plots appear in Figs. 4, 5, 6. .......................................... 1 15

Table 14. Regression statistics for the relationship between mean biomass of
species in control plots (no nitrogen added) and the relative change in mean
biomass between treatment (nitrogen added at different times of the growing
season) and control (relative response, see Methods). Only species that

occurred in both control and treatment are included in the regression

analyses. Values in boldface indicate that the slope or y—intercept differs
significantly from zero. The corresponding plots appear in Figs. 7, 8, and 9.
.......................................................................................................................................... 120

Table 15. Regression statistics for the relationship between mean biomass of
species in control plots (no nitrogen added) and the change in mean biomass
between treatment (nitrogen added at different times of the growing season)

and control plots (Absolute response, see Methods) in the second-year old field.
Only species that occurred in both control and treatment are included in the
regression analyses. Values in boldface indicate that the slope or y-intercept
differs significantly from zero. The corresponding plots appear in Figures 10

and 1 1. ............................................................................................................................ 1 2 5

Table 16. Regression statistics for the relationship between mean biomass of
species in control plots (no nitrogen added) and the relative change in mean
biomass between treatment (nitrogen added at different times of the growing
season) and control (Relative response, see Methods) in the second-year old-

field. Only species that occurred in both control and treatment are included in
the regression analyses. Values in boldface indicate that the slope or y-

intercept differs significantly from zero. The corresponding plots appear in
Figures 12 and 13. ........................................................................................................ 129

LIST OF FIGURES

Figure 1. Locations of the annual old-field site (A) and the second-year
old-field site (B) at W. K. Kellogg Biological Station, Kalamazoo County,

Michigan. .............................................................................................................

Figure 2. Effects of the removal of Cbangpbgllum album and the time of
nitrogen addition on a) total community biomass, b) subordinate

community biomass, and C) Wu; album biomass (error bar- = 1

standard error). ..................................................................................................

Figure 3. Species biomass responses to nitrogen addition: a) early
nitrogen addition (July 9), b) mid-season nitrogen addition (July 24),
and c) late season nitrogen addition (August 7) when the dominant

species, Cbenobbdlum album, was present. The corresponding species
names for the five-letter Bayer codes on the x-axis are listed in Table 3.

Figure 4. Species biomass responses to nitrogen addition: a) early
nitrogen addition (July 9), b) mid-season nitrogen addition (July 24),
and c) late season nitrogen addition (August 7) when the dominant
species, Chenbpbdlum album, was removed. The corresponding species
names for the five- letter Bayer codes on the x—axis are listed in Table 3.

Figure 5. Diversity and community structure responses to the removal
of Chgnbpbgllum album and the time of nitrogen addition: a) species
richness, b) biomass evenness, and c) biomass diversity (error bar = 1

standard error). ..................................................................................................

Figure 6. Relationship between the biomass of the dominant species,
album, and a) absolute competition intensity and b)

relative competition intensity. ........................................................................

Figure 7. Relationship between total plant community biomass and a)

absolute competition intensity and b) relative competition intensity.

Figure 8. Effect of the removal of the dominant species and time of
nitrogen addition on percent light penetration to the soil surface

measured August 1. .............................................................................................

Figure 9. Tissue nitrogen concentration of C, album (a) and total
nitrogen content of C, album (b) in response to removal of C, flbum and
time of nitrogen addition. Note that Q. mm in removal treatment were

seedlings that emerged after the removal treatment in early July. .........

........ .S

......... 19

....... 23

....... 24

........ .27

......... 31

........ 32

........ .33

........ 37

Figure 10. Tissue nitrogen concentration (a) and total nitrogen content
(b) of the subordinate plant community in response to removal of the
dominant species and time of nitrogen addition. .................................................. 39

Figure 1 1. Relationship between tissue nitrogen concentration and
species biomass in control plots and mid-season (July 24) nitrogen
addition plots. ............................................................................................................... 4 1

Figure 12. Relationship between (a) tissue nitrogen concentration of

species in control and mid-season (July 24) nitrogen addition plots, and

(b) total nitrogen content of species in control and mid-season (July 24)

nitrogen addition plots. .............................................................................................. 4 2

Figure 13. Seasonal pattern of above-ground biomass production of
different life history groups in a second-year Michigan old-field
(arrows indicate time of nitrogen addition). ......................................................... 65

Figure 14. Seasonal patterns of a) species richness, b) biomass
evenness, and c)biomass diversity in a second-year Michigan old-field. ....... 66

Figure 15. Above-ground plant community biomass by life history
groups at harvest in control (no nitrogen added) and 5 times of nitrogen
addition treatments in a second-year Michigan old-field. .................................. 70

Figure 16. Species richness (a), species biomass evenness (b), and

species biomass diversity (c) at harvest in control plots (no nitrogen

added) and 5 times of nitrogen addition treatments in a second-year

Michigan old-field (error bar = 1 standard error). ............................................... 72

Figure 17. Above-ground biomass of the dominant species, Ema

ganadensls, and the subordinate community in control plots (no

nitrogen added) and 5 times of nitrogen addition treatments in a second-

year Michigan old-field (error bar = 1 standard error). ..................................... 73

Figure 18. Relationship between total community biomass per plot and
a) species richness, b) species biomass evenness, and c) species
biomass diversity. ........................................................................................................ 7 5

Figure 19. Relationship between biomass of the dominant species,

Cbnxza canadensis, per plot and a) species richness, b) species biomass
evenness, and c) species biomass diversity. .......................................................... 76

Figure 20. Above-ground biomass of seven common species in control
plots and nitrogen addition treatments in a second-year Michigan old-
field plant community (error bar = 1 standard error). ........................................ 79

Figure 21. Seasonal patterns of mineral nitrogen levels (NO3-N + NH4-

N) in control plots and 5 times of nitrogen addition treatments in a

second-year Michigan old-field. Arrows indicate time of nitrogen

addition. Note the log scale on the y-axis. .............................................................. 83

xii

Figure 22. Seasonal patterns of potential nitrogen mineralization rate
in control plots and 5 times of nitrogen addition treatments in a second-
year Michigan old-field. Arrows indicate time of nitrogen addition. .............. 84

Figure 23. Seasonal patterns of soil moisture in control plots and 5 times
of nitrogen addition treatments in a second-year Michigan old-field.
Arrows indicate time of nitrogen addition. ............................................................ 85

Figure 24. Seasonal patterns of light penetration to the soil surface in
control plots and 5 times of nitrogen addition treatments in a second-
year Michigan old-field. Arrows indicate time of nitrogen addition. .............. 87

Figure 25. Absolute biomass response to an experimental treatment.

Each point represents a single species. The dashed line represents no

change in species abundance in response to the treatment. Points

above the dashed line represent species that increased in response to

the treatment ................................................................................................................ 101

Figure 26. Potential responses of species to an experimental treatment.

The dashed line represents the null hypothesis and the solid line

represents the fitted regression line. In (a), all species increase in

biomass in response to the treatment and the magnitude of their

response is independent of their biomass in the control. In (b), the

biomass of species of all abundances increased and species of larger

biomass in the control have larger magnitudes of response in the

treatment. In (c), species of low abundance in the control decrease in

biomass and species of high abundance in the control increase in

biomass in response to the treatment. ..................................................................... 103

Figure 27. Potential relative responses of species to an experimental
treatment. The dashed line represents the null hypothesis and the solid
line represents the fitted regression line. In (a), all species increase in
biomass in response to the treatment and the magnitude of their
relative response is larger for species of lower abundance in the

control than for species of higher abundance. In (b), the biomass of
species of all abundances increases and species of all abundances in the
control have equivalent or almost equivalent relative biomass
responses. In (c), species of low abundance in the control have a large
negative relative biomass response and species of high abundance in
the control have a large positive relative biomass response to the
treatment. ...................................................................................................................... 105

Figure 28. Absolute species biomass responses to a) early, b) mid-, and
c) late season nitrogen addition with C, flbum present in the community
(dashed line = null hypothesis .................................................................................. 1 14

Figure 29. Absolute species biomass responses to a) early, b) mid—, and

c) late season nitrogen addition with C, album present in the community

but excluded from the regression analysis (dashed line = null

hypothesis ..................................................................................................................... 1 l 6

xiii

Figure 30. Absolute species biomass responses to a) early, b) mid-, and
c) late season nitrogen addition with C, album removed from the
community (dashed line = null hypothesis ............................................................ 1 18

Figure 31. Relationship between the biomass of species in the control

plots and the relative change in mean biomass of species with a) early,

b) mid-, and c) late season nitrogen addition and C, album present in

the community (dashed line = null hypothesis ..................................................... 1 19

Figure 32. Relative species biomass response to a) early, b) mid-, and c)
late season nitrogen addition with C, album excluded from the
regression analysis (dashed line = null hypothesis ............................................. 122

Figure 33. Relationship between the biomass of species in the control

plots and the relative change in mean biomass of species with a) early,

b) mid-, and c) late season nitrogen addition and L album removed

from the community (dashed line = null hypothesis ........................................... 123

Figure 34. Absolute species biomass responses to 5 times of nitrogen
addition in a second-year old-field in southwestern Michigan (dashed
line = null hypothesis ................................................................................................. 124

Figure 35. Absolute species biomass responses to 5 times of nitrogen
addition in a second-year Michigan old-field. The dominant species,
Cbnyza ganadgnsls, is excluded from the regression analysis (dashed

line = null hypothesis ................................................................................................. 127

Figure 36. Relationship between the biomass of species in the control
plots and the relative change in mean biomass of species at 5 times of
nitrogen addition in a second-year Michigan old-field. (dashed line =
null hypothesis ............................................................................................................ 128

Figure 37. Relationship between the biomass of species in the control

plots and the relative change in mean biomass of species at 5 times of

nitrogen addition in a second-year Michigan old-field. The dominant

species, Cbnyza canadenm, is excluded from the regression analysis

(dashed line = null hypothesis .................................................................................. 130

xiv

CHAPTER 1
INTRODUCTION

INTRODUCTION

A commonly observed pattern in nature is that at some spatial scales
species diversity increases and then declines with increasing productivity
(reviews by Tilman 1982, Tilman and Pacala 1993, Rosenzweig and Abramsky
1993). This pattern has been well-documented in a variety of terrestrial plant
communities (reviews by Tilman 1982, T ilman and Pacala 1993, DiTommaso and
Aarssen 1989), but the mechanism(s) underlying this pattern is often debated
(Tilman 1993).

Numerous mechanisms have been proposed to explain the decline in
plant diversity at high levels of fertility. Some authors have proposed that
competition intensity increases with increasing productivity, leading to the
loss of species (Grime 1973, 1979). Specifically, competition intensity for
resources from one or a few dominant species may underlie the pattern of
decreasing diversity with increasing fertility (Grime 1987). Newman ( 1973)
Goldberg and Miller (1990) suggested that as productivity increases light
penetration declines and shade-intolerant or slow-growing species, and
seedlings experience higher mortality rates. Tilman ( 1982) argued that spatial
heterogeneity in resources is higher in moderately productive habitats. Each
of these hypotheses invokes plant competition for resources as an important
mechanism leading to the decline in plant diversity with increasing
productivity. This dissertation focuses on the mechanism of reduction of
species diversity through increased competition intensity for resources from

one or a few dominant species as fertility increases (Grime 1987).

3
There is a good deal of controversy in the literature about the

mechanisms of plant competition for resources in productive habitats. One

view, associated most frequently with Tilman (1982, 1988) but supported by
others (Newman 1973, 1983, Grubb 1985, Taylor et a1. 1990), argues that as soil
fertility increases and above-ground plant biomass increases, light becomes
relatively more limiting than soil nutrients. This leads to a trade-off in above—
and below—ground resource competition and no change in the magnitude of
competition intensity with increasing fertility. The opposing view, most
commonly associated with Grime (1973, 1979, 1987) but also articulated by
others (Huston 1979, Southwood 1988, Keddy 1989a, Bertness 1991, Campbell et
al. 1991b) argues that both above- and below-ground competition intensity

increase with increasing soil fertility.

Thesis organization

This thesis addresses several questions regarding the consequences of
variation in the seasonal timing of nitrogen pulses for diversity, community
structure, and resource competition in old-field plant communities. It is
divided into five chapters. The first chapter provides an orientation to the
ecological questions that motivated this work, and describes the plant
communities and study sites where the research was conducted. Chapter 2
describes the effects of removing a dominant species and varying the seasonal
timing of nitrogen addition on diversity, community structure, indices of
competition intensity, and nitrogen uptake by the dominant species and the
subordinate community in a one-year old-field. In Chapter 3, the effects of
varying the seasonal timing of nitrogen addition on community structure and
resource availability in a second-year old-field community are examined. The

seasonal patterns of diversity and community structure in the second-year

4
old-field community are described. The results demonstrate a link between the

time of nitrogen addition and seasonal patterns of mineral nitrogen
availability, nitrogen mineralization, and light depletion. Chapter 4 describes
a graphical and statistical method for examining shifts in species biomass
hierarchies in response to an experimental treatment. The analysis
demonstrates the differences between absolute and relative responses to
nitrogen additions. Chapter 5 summarizes the conclusions of the preceding
chapters and discusses the implications of this research for further study of
the relationship between plant diversity and productivity and mechanisms of

plant competition.

Description of field sites

The first experiment was conducted in south-western Michigan at the
Long Term Ecological Research (LTER) site in agricultural ecology (labeled A
in Figure 1) at the W. K. Kellogg Biological Station (KBS, 85° 22' W longitude,
42° 24' N latitude). For more than 20 years prior to the establishment of this
LTER site in 1989, the land had been under continuous corn (Zga mm)
production, except for strip-cropping of wheat and corn for two years in the
1970's, and production of alfalfa on the western side of the site for 4 years in
the 1980's. The soil is primarily Kalamazoo silt loam (Typic Hapludalf, sandy to
silty clay loam), and is well-drained and of moderate fertility (Whiteside et al.
1959). One of the eight treatments in this LTER is an old-field succession
treatment abandoned from agricultural cultivation in spring 1989. At the
north end of each of the six replicate 0.9 hectare plots in the succession
treatment is a 20 x 30 m area plowed and disced each year in late April or early

May to maintain an annual plant community.

 

 

 

 

 

 

 

 

"I
I
i

 

 

Figure 1. Locations of the annual old-field site (A) and the second-year old-
field site (B) at W. K. Kellogg Biological Station, Kalamazoo County, Michigan.

 

6
The second experiment was conducted in a second-year abandoned old-

field (labeled B in Figure 1) at Michigan State University’s W. K. Kellogg
Biological Station in southwestern Michigan (85° 23’ W longitude, 42° 24’ N
latitude). The old-field is one in a chronosequence of 1-hectare fields planted
to soybeans and abandoned in sequential years since 1989. The field in this
study was planted to soybean (Glycine. max (L.) Merr.) in May 1991 with no
fertilizer application, and was abandoned in the fall after the soybean harvest.

The soil is of the Kalamazoo series.

Description of plant communities

Plant succession following abandonment from agriculture in
southwestern Michigan is characterized by an initial dominance by annual
grasses and dicots in the first year and a rapid succession to mixture of
biennials, perennials, winter annuals, and summer annuals in the second year
(Gross and Werner 1982). Perennial species, such as Sglldagb species, Aster
species, and Daucus carbta often dominate by the third year after
abandonment. Wind dispersed woody species can colonize in the first few
years post-abandonment.

To address the effects of a dominant species on community structure,
biomass, and nitrogen acquisition, the first experiment (described in Chapter
2) was conducted in an annual old-field plant community located in the
successional treatment of the Long Term Ecological Research (LTER) site in
agricultural ecology at W.K. Kellogg Biological Station (KBS). This annual
plant community is maintained by annual plowing in late April or early May.
The species composition is typical of agricultural fields in southwestern

Michigan in the first year of abandonment. The dominant dicots included

ChemodmmalbmnAmaranthnsretmflexusAmbmmmemisnfeha and

7
Abunlbn mgbpbrastl. The dominant monocots included maria faben, 5,

111mm. 8. 21211113. Ranicmn dichmgmiflnmm. and E. capillate. Winter
annuals and perennials were less abundant than the summer annuals. The
winter annuals included Bubanea vulgarls, Qxalls 51211913, and Stellaria media.
Perennial species, such as Apbcynum gannlbmum, mm m m
spp., and Rumex spp., are patchily distributed on the 50 hectare scale, but
locally abundant on the one-hectare scale.

To address the effects of the time of nitrogen addition on community
structure and resource availability, the second experiment (described in
Chapter 3) was conducted in an second-year old-field. The experiment covered
an area of 1 1 x 8 m. The vegetation was dominated by summer annuals
(Abmilonthmhtasii. Amatamhustetmflexus, andChenQnodiumalbiun) in
1992 and by Cbnyza ganadgnsis (a winter annual or biennial) with a diverse
mixture of winter annuals, biennials, perennials, and summer annuals during

the summer of 1993 when the study was conducted.

CHAPTER 2

DOMINANCE, DIVERSITY, AND RESOURCE COMPETITION
IN AN ANNUAL OLD-FIELD PLANT COMMUNITY

9
INTRODUCTION

A common pattern in herbaceous plant communities is that species
diversity declines with naturally or experimentally increased fertility (Reed
1977, Bakelaar and Odum 1978, Silvertown 1980, Inouye et al. 1987, Tilman 1987,
Carson and Barrett 1988, Gurevitch and Unnasch 1989, Goldberg and Miller
1990; reviews by Tilman 1982, 1988, DiTommaso and Aarssen 1989). Numerous
mechanisms involving variation in local extinction and colonization
probabilities have been proposed to explain these patterns (see references in
Tilman 1993). The decrease in plant species diversity when soil nutrients are
abundant is often correlated with the dominance of one or a few species, and
the mortality or low recruitment of rare (= smaller stature) species (Willis
1963, Tilman 1987, Goldberg and Miller 1990). Plant community biomass often
increases in response to increasing fertility, but individuals or species may
have differential abilities to capture, store, assimilate, and allocate added
nutrients to growth (Pate 1983, Campbell et al. 1991a, Grime 1994) resulting in
a biomass or size-based hierarchy of competitive ability and consequent
decreases in species diversity. This chapter focuses on an experimental test of
the mechanism(s) of reduction of species diversity through increased
competition intensity for resources from one or a few dominant species as
fertility increases (Grime 1987).

There are two alternative views about the nature of plant competition in
productive habitats. One view, associated most frequently with Tilman (1982,
1988) but supported by others (Newman 1973, 1983, Grubb 1985, Taylor et al.
1990), argues that as soil fertility increases and above-ground plant biomass
increases, light becomes relatively more limiting than soil nutrients. This
leads to a trade-off in above- and below-ground resource competition, but no

change in the magnitude of competition intensity, with increasing fertility.

10
The opposing view, most commonly associated with Grime (1973, 1979, 1987) but

also articulated by others (Huston 1979, Southwood 1988, Keddy 1989a, Bertness
1991, Campbell et al. 1991b) argues that both above- and below-ground
competition intensity increase with increasing soil fertility. Recently, Grime
(1994) has argued that there is a covariance between the resource demands of
plant roots and shoots. In order to acquire biomass for exploring the above-
ground environment for light, the shoot requires water and nutrients
provided by the roots. In order for the roots to explore the soil for water and
nutrients they require carbon fixed by the leaves. This interdependence of
photosynthesis and nutrient capture underlies Grime’s hypothesis that
competition intensity for both light and nutrients increases with soil fertility.
Increasing competition intensity from the dominant plant species in
fertile environments may alter two community attributes that determine
species diversity as measured by the Shannon diversity index (H'): species
richness and evenness. First, by having disproportionate access to resources
above and/ or below ground, the dominant species may acquire more added
resources and allocate them to growth. The positive feedbacks between
asymmetry in size, resource capture, and growth, called resource preemption
(Grace 1990) or asymmetric competition (Weiner 1986, 1990, Wilson 1988), lead
to low evenness of biomass distribution across species within a growing
season, and thus, lower diversity as measured by H'. Second, the dominant
species may reduce species richness by reducing the germination, survival,
growth, and reproduction of less common species. This is one set of
mechanisms thought to generate the descending portion of the hump-shaped
response of species diversity to resource richness (Grime 1979, Huston 1979,

Tilman 1982).

1 1
A common experimental method to determine the effect of a dominant

competitor or predator on species diversity and community structure in plant
and animal communities is to remove the dominant species (Connell 1961,
Paine 1966, Dayton 1975, Lubchenco 1980, Armesto and Pickett 1986). Several
studies in terrestrial plant communities have demonstrated that removing the
dominant species can increase the diversity of the subordinate plant
community (Abul-Fatih and Bazzaz 1979, Armesto and Pickett 1985, Gurevitch
and Unnasch 1989, Keddy 1989b). However, such results are not universal;
other studies have shown that the removal of the dominant species leads to a
decrease or no change in the species diversity of the subordinate community
(Pinder 1975, Allen and Forman 1976, Hils and Vankat 1982). To test the
hypothesis that increasing competition intensity from the dominant species
results in decreasing diversity with increasing soil fertility, the dominant
species must be removed over a range of soil fertility. If the hypothesis is
correct, then the removal of the dominant species should increase species
richness, evenness, and/ or diversity, and the magnitude of this effect should
increase at higher fertility levels. I am aware of only one study where the
response of the plant community to removal of a dominant plant species has
been examined at two levels of soil fertility (Gurevitch and Unnasch 1989).
They found that removal of the dominant perennial grass species, Dagglls
glbmerata, from a 24-year-old field increased richness and diversity, and the
effect on richness was greater in fertilized plots than in control plots.

Plant species vary in phenology (Swieringa and Wilson 1972, Mellinger
and McNaughton 1975, Reed 1977), and these differences in growth phenology
can affect resource demands and resource depletion patterns. Consequently, a
species’ response to a resource pulse may depend on the time of resource

addition relative to their development status (Mengel and Kirkby 1978, Watson

12
1984, Benner and Bazzaz 1985, 1988) and/or the degree to which a

biomass/height hierarchy has developed among individuals or species
(Benner and Bazzaz 1987). Reed ( 1977) showed that in an annual old-field a
single large pulse of nitrogen early in the season reduced diversity more than
the same total amount of nitrogen applied in smaller amounts over the
growing season. However, I am aware of no studies in natural plant
communities that have varied the time of a single large pulse of nitrogen
during the growing season to address whether the stage of development of the
biomass/ height hierarchy influences the response of dominant and
subordinate species to increased soil fertility.

I conducted a species-removal experiment and varied the time of
nitrogen addition to address the following questions about competition
intensity, community structure, and resource competition in an annual old-
field plant community: (1) What is the effect of the dominant plant species on
the biomass and the species richness, evenness, and diversity of the
subordinate plant community? (2) How does the impact of the dominant
species change when the time of a nutrient pulse is varied within a growing
season? (3) What is the effect of the dominant species and the time of
nitrogen addition on light penetration to the soil surface and nitrogen

acquisition by the subordinate plant community?

METHODS
Study site
The experiment was conducted in south—western Michigan at the Long
Term Ecological Research (LTER) site in agricultural ecology at the W. K.
Kellogg Biological Station (KBS, 85° 22' W longitude, 42° 24' N latitude) of

Michigan State University. For more than 20 years prior to the establishment

13
of this LTER site in 1989, the land had been under continuous corn (Zea mays)

production, except for strip-cropping of wheat and corn for two years in the
1970's, and production of alfalfa on the western side of the site for 4 years in
the 1980's. The soil, primarily Kalamazoo silt loam (Typic Hapludalf, sandy to
silty clay loam), is well-drained and of moderate fertility (Whiteside et al.
1959). The mean annual temperature is 9°C (30 year mean) and annual
precipitation is 860 mm/ y, with about half falling as snow in the winter
months. One of the eight treatments in this LTER is an old-field succession
treatment abandoned from agricultural cultivation in spring 1989. At the
north end of each of the six replicate 0.9 hectare plots in the succession
treatment is a 20 m x 30 m area plowed and disced each year in late April or
early May to maintain an annual plant community. The most common and
abundant species in this community include Chanbpbdlum album (common
lambsquarters), Amaramhus [bubflgxus (redroot pigweed), Abunlbn
W (velvet leaf), Ambrbsia anemlsilfglla (common ragweed), Setarla
[abezi (giant foxtail), S. 11.1mm (yellow foxtail), 5. mm: (green foxtail),
Eamcum dlghmbmlflgmm (fall panicum), and P_. capillane (witchgrass). The
field experiment described in this chapter was conducted in this annually

disturbed area.

Experimental design

The experimental design included 4 blocks of the LTER succession
treatment X 2 species removal levels (no removal and removal of the dominant
species, Changbbdmm album) X 4 times of nitrogen addition (none and 0, 2, and
4 weeks after removing C album). Four blocks where Q album had been
abundant during the previous three years were chosen for the experiment.

The treatments were arranged in a complete randomized block design using 1

14
m2 plots (2.0 X 0.5 m). A 1 m border with the plant community left intact

separated the plots on each side. Chanbpbgllum album was removed between
July 6 and 8, 1992 (about 5 weeks after the soil had been plowed and disced) by
pulling or clipping seedlings. By this time most of the germination of C. album
had occurred, and the plants were a maximum of 10 cm tall. The biomass of Q
album seedlings removed in July accounted for about 22 percent of the final
biomass of the unfertilized plant community. The removal treatment was not
repeated during the growing season. Thus, a small amount of Q album (1 to 3
percent of the total community biomass) was present in the removal
treatments at the final harvest. Nitrogen was applied in a single pulse at 32 g
N/m2 as ammonium nitrate (320 kg N/ha, typical of fertilizer levels for corn
production in this region) 0, 2 and 4 weeks after removing Q album. The time
of nitrogen addition treatments are henceforth referred to as early- (July 9),
mid- (July 24) and late-season (August 7). No nitrogen was applied to the

control plots.

Measures of biomass, light penetration and nitrogen acquisition

The ability of the plants to respond to this nutrient pulse was measured
as the biomass and nitrogen acquired above-ground at harvest on August 21,
two weeks after the final nutrient addition. The plants were harvested by
clipping the stems at the soil surface, separated by species, dried at 60°C, and
weighed to 1 0.01 g.

To determine plant tissue nitrogen content, the dried plant material
(including all above-ground tissue) was ground by species on a rough gauge
screen using a Wiley mill and a fine gauge screen using a Cyclotec grinding
mill (Tecator Model 1093). The finely ground samples were dried at 48°C for

12-24 hours before weighing duplicate samples of about 1 mg for nitrogen

15
analysis. The duplicate samples were analyzed for total nitrogen content using

a Carlo-Erba nitrogen/carbon analyzer (Model 1500 Series 2). Tissue nitrogen
concentration is defined as the percent nitrogen content of the above-ground
tissue. Total nitrogen content per species was calculated by multiplying the
tissue nitrogen concentration of each species by its above-ground biomass.
The total nitrogen content of the subordinate community is the sum per plot of
the total nitrogen content of each species. Subordinate community nitrogen
concentration was calculated by dividing the total nitrogen content of the
subordinate community by the biomass of the subordinate community in each
plot.

light measurements were taken on August 1 between 1 1:00 and 13:00
using a 40-cm long integrating light probe (Sunfieck PAR ceptometer,
Decagon Devices Model SF-80, Pullman, WA) to determine the effect of the
dominant species and the time of nitrogen addition on light penetration to the
soil surface. Photosynthetically active radiation (PAR) was measured above
the plant canopy and 1 cm above the soil surface with the light probe oriented
west to east along the center of the Z-m length of the plot. Light penetration is
expressed as the percent of PAR above the plant canopy penetrating to 1 cm
above the soil surface. One light reading per plot was taken and light
penetration is expressed as the mean percent light penetration in four
replicate plots. Because the light measurements were taken before the August
7 nitrogen addition, only the control, early- and mid-season nitrogen

treatments are included in this analysis.

Indlces of diversity and competition intensity
Species richness was calculated as the number of species per plot. The

Shannon diversity index was calculated as H' = - 2 pi (log pi), where pi is the

16
proportion of biomass per plot accounted for by species i. The Pielou index of

biomass evenness was calculated as J = H' / Hmax. where Hmax = log of species
richness. Species richness, evenness and diversity were calculated per plot,

and are expressed as the mean of four replicates. Evenness and diversity were
calculated based on biomass rather than numbers of individuals because
biomass has been shown to be an important predictor of competitive ability in
plants (Gaudet and Keddy 1988 and references therein; review by Goldberg and
Barton 1992).

Two measures of competition intensity were calculated for the control
and each time of nitrogen addition. Absolute competition intensity, a measure
of the reduction of subordinate community biomass by the dominant species,
was calculated as ACI = (Br - BC), where Br = the biomass of the entire
subordinate community when Q album was removed and BC = the biomass of
this assemblage of species when C. album was present in a plot. The index of
absolute competition intensity may increase in response to nitrogen addition
because the performance of the subordinate community increases with
increasing fertility in the absence of the dominant species (increase Br) apart
from the decline in subordinate community performance in the presence of
the dominant species (decreasing BC) (Reader et al. 1994). Thus, an index of
relative competition intensity was calculated as RCI = (Br - BC) / By. This gave
a measure of the reduction of the subordinate community biomass by the
dominant species relative to its biomass unimpeded by the dominant species at
a given level of nutrient addition. These indices are modifications of the more
typical form of the equations in which the impact of competitors (the entire
surrounding community) is measured on a single target individual (Wilson
and Keddy 1986, Reader and Best 1989, Wilson and Shay 1990, Wilson and Keddy
1991, Wilson and Tilman 1991a, 1993, Reader et al. 1994). However, because the

17
objective was to test the hypothesis that competition intensity from the

dominant species increases with increasing fertility, the formula for
quantifying competition intensity was modified to assess the response by the

subordinate community.

Statistical analysis

Analysis of variance was used to test for the main effects of removal of
C. album and time of nitrogen addition on the plant biomass variables, species
richness, evenness, diversity, plant tissue nitrogen concentration, and total
nitrogen content. The three way interaction of block x removal x nitrogen
time was used as the error term to test for the significance of the main effects
and their interactions. The values of the residuals were independent of the
means for the biomass and diversity variables. Thus, the analyses of variance
for these variables were performed on the untransformed data. Tissue
nitrogen concentrations and percent light penetration values were arcsin
square root transformed prior to the analysis of variance. The ANOVA for total
nitrogen content was performed on the untransformed data. A significant
main effect of time of nitrogen addition in the ANOVA was further decomposed
using the Dunnett's t-test to make pairwise comparisons between control and
nitrogen addition at the three times of nitrogen addition. The Dunnett's t-test
allowed a determination of which times of nitrogen addition had significant
effects on community biomass, richness, evenness, diversity, and plant
nitrogen content. Analysis of variance was used to test the effect of the time
of nitrogen addition on absolute and relative competition intensity. Linear
regression was used to test the relationship between total community biomass
and Q album biomass and the indices of competition intensity. The SAS

procedure GLM (version 6) was used for all statistical analyses (SAS 1989).

18

RESULTS
Biomass responses

Removal of the dominant species, C album, significantly reduced the
biomass of the whole community (Figure 2a, Table 1); the subordinate
community did not increase in biomass enough to compensate for the removal
of C. album within this single growing season(Figure 2a, b). Community
biomass increased significantly in response to nitrogen addition both in
treatments where Q album was present and removed. However, the effect of
nitrogen addition on community biomass declined when the addition came
later in the growing season; the positive effect of nitrogen addition on
community biomass was significant by Dunnett's t-test only for the early- and
mid-season nitrogen addition treatments (Table 2).

The biomass of the subordinate plant community increased
significantly with nitrogen addition and the magnitude of the response
depended on the time of the nitrogen addition and whether or not Q album was
removed (Figure 2b, Table 1). When C. album was present, the subordinate
community biomass increased in response to the early-season addition of
nitrogen, but not to the mid- or late-season addition of nitrogen. When Q
album was removed, the subordinate community biomass increased with both
the early- and mid-season addition of nitrogen. However, the magnitude of the
biomass increase in the subordinate community declined when the nitrogen
addition came later in the season (Figure 2b). These results suggest that a
strong dominance hierarchy develops over the season in this annual plant
community and affects the ability of subordinate members of the community

to respond to a nutrient pulse.

l9

 

(a) E] No removal
1000 -
I C. album removed
800 "‘
600 "

 

Total community biomass
(s/mAZ)

 

 

 

 

 

 

 

None July 9 July 24 Aug 7
Time of nitrogen addition

600 - (b) D No removal
I C. album removed
500 "

400-
3001

(s/mAZ)

200 ‘

Subordinate
community biomass

100-

 

 

 

 

     

None July 9 July 24 Aug 7
Time of nitrogen addition

(C) CI No removal

700 - I C. album removed

600-1 l l

500 - .
400- l
300 '-
200 '-
100 -

0 — . . .
None July 9 July 24 Aug 7
Time of nitrogen addition

 

C. album biomass (g/mAZ)

 

 

 

 

 

 

 

 

 

Figure 2. Effects of the removal of Chmbbbdlum album and the time of
nitrogen addition on a) total community biomass, b) subordinate community
biomass, and c) Chenbbouium album biomass (error bar = 1 standard error).

20

8888:: u 858 #

H00.0 V m .13. mH00 V m .3. um0.0 V m a. H0._H.0 V hm.

 

 

 

55.828

mo; .33. Edm no; * Ram .33. whom i. 3.0 06m .3. Sam 880% «.8808

Ed .3. v0.3 mud n. mm.m .3. deN .3. wad 0.3 i. 05‘ $28.43» mmanE

mm; NWN 3.0 mm; .P ova” .33. Rama v.3 .. 3mm «886M.— 86on

i. 3.0 .3... 3.3 v.3 .23. NW: mmm8oE 833d

«.8865 583888

:4 Cd oo.~ 3. 3.3 a N2: 3. Cd mAm 3. 32¢ 388833

@3805

Mai" .13. wNK~ cm; .33. 00.5..V sashm€MN .13. m¢.mN Mdm .33. V0.MN kuadgou Each

.3 m a a a a NH 682 a 882 289$
@3683: mumu838 ”1.88:: m188a8 ~188=c mun—£083:

3688 «83 8082!.“ 2

~83 x xuoE x xuoE x 368mm .6 o8F 8388. x068

 

 

8083.28 Ho 8.88

.m 88 N wan—mm 8 850% oh." 838 “5888... .m u <>oz< 58 88 .6 8.2m Am<>oz<v
a: .0838 863m 88 .5 80885 $8805 .mmogut 80on £8803 g Sauna £8803 585888

@852023. £8805 583888 :39 so 8092:.“ cowoba no .083 88 3688 882% mo 38% A 2an

21

Table 2. Pairwise comparisons of the effect of time of nitrogen addition on
total community biomass, subordinate community biomass, Wm
Mm biomass, species richness, biomass evenness (J), and species diversity
(H') using Dunnett's t-test (alpha = 0.05, df =- 9, critical value of Dunnett's t =

2.812).

 

of nitrcgen addition

 

Variable MSD 1* MSE 1 none v. none v. none v.
early- mid-season late-season
season

Total community 82.08 3408.6 320.19 *** 185.28 *** 62.72

biomass

Subordinate community 103.49 5419.2 205.52 *** 53.42 9.27

biomass

Q album biomass 176.18 7852.5 222.35 *** 256.40 *** 104.54

Species richness 2.00 2.03 -0.63 0.75 -0.25

Biomass evenness 0.1 1 0.006 -0.07 -0.12 *** -0.03

Species diversity 0.22 0.03 -0.19 -0.25 *** -0.08

 

t MSD - minimum significant difference

1- MSE = mean standard error

22
The dominant species, C. album, increased significantly in biomass in

response to the early and mid-season nitrogen addition (Figure 2c, Table 2).
Cbenbbbdlum album also increased in biomass in response to the late-season
nitrogen addition, but this effect was not significant. In the mid- and late
season nitrogen addition treatments the increase in total community biomass
was accounted for entirely by the increase in C. album biomass; the
subordinate community did not increase in biomass at either of these times
when C. album was present (Figure 2b).

The responses of individual species in the subordinate community to the
time of nitrogen addition depended on whether Q album was present in the
community (Figure 3a-c) or removed (Figure 4a-c) from the community.
When 9. album was present, the biomass of several of the subordinate species
increased in response to the early-season nitrogen addition (Figure 3a). The
final biomass of Ambrosia memisiifnlia 5mm fallen. Abutilen themhtam.
S, lumsmm, and Digitana sangmnalls doubled relative to the unfertilized
control when nitrogen was added early in the season (July 9). However, when
nitrogen was added mid-season (July 24), few of the subordinate species
increased in biomass. In fact, several of the species that increased in biomass
in the early season nitrogen addition treatment declined in biomass when
nitrogen was added mid-season (Figure 3b). Likewise, few species responded
strongly to the late-season nitrogen addition when Q album was present
(Figure 3c).

In contrast, when Q album was removed from the community, the
biomass of several species in the subordinate community increased in
response to both mid- and late-season nitrogen addition (Figure 4b, c).
However, no single subordinate species acquired the biomass dominance of C.

album. There was some variation in which subordinate species responded most

23

 

 

 

 

 

 

 

 

 

 

(a) 259481112
{<3 UNoNaddition
g .NaddedJuly9
:33
E
‘3.
i
m HIE-$7755.;rfiuéuw
駧§§§§§§§§ iézéé
(b)
g EINoNaddition
(g I N added July 24
E
E
'5.
(C) 7S259363
E? DnoNaddition
g 50 I N added August 7
g x
.5. ‘
g 25
l
3 0] I'll.
§§§E§E§§§§§§§§§§55§§§§

Figure 3. Species biomass responses to nitrogen addition: a) early nitrogen
addition (July 9), b) mid-season nitrogen addition (July 24), and c) late season
nitrogen addition (August 7) when the dominant species, Cbengbbdlum album,
was present. The corresponding species names for the five-letter Bayer codes
on the x—axis are listed in Table 3.

24

    

(a) 75 83 86
S DNoNaddition
(g .NaddedJuly9
V 50
g I
.5
g 25
50mm”..-
~ m o >.

(b) 75 136
G EINoNaddition
Te I N added July 24
3o 50
§ Q
E I
g 25"

§E§§§§E§§§§§§§§§§§EE%§
(C) 75
A EINoNaddltion
28‘ I N added August 7
B, 50
g I
395 25
8
i,|l|lI....__

~ an o
gaééigééiééééiéééiééié

Figure 4. Species biomass responses to nitrogen addition: a) early nitrogen
addition (July 9), b) mid-season nitrogen addition (July 24), and c) late season
nitrogen addition (August 7) when the dominant species, Chenbpbdlum album,

was removed. The corresponding species names for the five-letter Bayer codes
on the x-axis are listed in Table 3.

25

Table 3. Five-letter Bayer codes (WSSA 1984) for weed species and their

corresponding scientific and common names.

Bayer code Scientific name
Annual dicots

ABUTH Amman W
AMARE Awnings retroilem
AMBEL Ambrosia anemisiifelia
CHBAL Chmdinm album
POLPB Eelxeemim Imam
SOLNI Selanum nigmm
Annual monocots

DIGSA Dieltaria saneninam
ECHCG Echinnclna mks-2am
PANCA Panicnm miller:

PAN DI Eanicnm dichotomiflnmm
SETFA Setaria faberi

SETLU Setaria linemen:

SEWI Setaria midis

Winter annuals and biennials

BARVU Barbarea mm:
OXAST Qxalis stride

STEME Stellaria media
Perennial dicots

APCCA Amnm cannibinnm
ASCSY Asclsnias Mara
LACCA Lactma canadensis
LACSE Lactnca serriela

MELAL Silene alha

RUMAA Rumex media
RUMCR Bumex 911512115

RUMOB Rumex Qmuaifnlius
TAROF Iamacmn officinfle
Perennial monocots

CYPF-S 9mm mlentm
Perennial nitrogen-fixers

TRFRE Irifelium Lem
TRFHY Irifelinm hybridnm
WOODY

Common name

velvet leaf

redroot pigweed
common ragweed
common lambsquarters
ladysthumb

common nightshade

large crabgrass
barnyard grass
witchgrass

fall panicum
giant foxtail
yellow foxtail
green foxtail

yellow rocket
yellow woodsorrel
common chickweed

hemp dogbane
common milkweed
tall lettuce

prickly lettuce
campion

red sorrel

curly dock
broadleaf dock
dandelion

yellow nutsedge

white clover
hybrid clover

unidentified woody
biomass

26
strongly at each time of nitrogen addition (Figure 4a-c). For example,

memisfifblla, and S, fabgri responded strongly to the early-season nitrogen
addition and less to the mid- and late-season nitrogen addition treatments.
Digitarla Wand Eanlcum dlgbbtbmlflbmm responded little to the
early-season nitrogen addition, but doubled in biomass when nitrogen was
added mid— or late-season. In an absolute sense, the strongest responses to
nitrogen addition occurred among subordinate species with the highest
biomass in control plots. None of the species of low biomass in the control
plots responded to nitrogen addition as strongly as did the abundant

subordinate species (Figure 4a-c).

Effects on species richness, evenness, and diversity

The removal of the dominant species, C. fibum, resulted in a marginally
significant increase in species richness (Figure 5a, Table 1). There was no
significant effect of time of nitrogen addition on species richness (Table 1).

The removal of C, album in control and nitrogen addition plots resulted
in a significant increase in biomass evenness across species as measured by
the Pielou evenness index (Figure Sb, Table 1). The time of nitrogen addition
also significantly affected evenness and this was most pronounced when Q.
album was present in the community (Figure 5b). Biomass evenness in
treatments where C. album was present decreased most relative to the control
when nitrogen was added mid-season and less when nitrogen was added early
or late in the season (Table 2). Even when Q, album was removed, biomass
evenness declined slightly when nitrogen was added early and mid-season,

suggesting that a new dominance hierarchy was being established among the

27

 

 

 

12 q (a) El Noremoval
a I C. album removed
10 -
33
a s -
fi
1: 6 -
a 4-
8
n. 2 -
m
0 .
None July 9 July 24 August 7
Time of nitrogen addition
A . (b) E] No removal
:3 I C. album removed
a 0.6 -
W
o
d l
8 o 4 1
> .
u I
0 2
E .
2 .
an
0.0 -
None July 9 July 24 August 7
Time of nitrogen addition
A (C) E] No removal
E5 ' I C. album removed
1.50 -
>.
H
E
E 1.00 - ‘
'u ‘ .
2
u 0.50 -
a .
2
m 0.00 -

 

None July 9 July 24 August 7
Time of nitrogen addition

Figure 5. Diversity and community structure responses to the removal of
W and the time of nitrogen addition: a) species richness, b)
biomass evenness, and c) biomass diversity (error bar— = 1 standard error).

28
remaining species. The magnitude of the effect of the dominant species on

biomass evenness was greatest with the mid-season addition of nitrogen.

The effect of removing Q album and the time of nitrogen addition on
the Shannon diversity index paralleled the effects on evenness (Figure 5c,
Tables 1 and 2). Removal of Q album significantly increased H' in the control
and nitrogen addition plots. This result is interesting because, it shows the
strong effect of Q album on the structure of the subordinate community.
When Q album was removed and no nitrogen was added, the biomass of the
subordinate community did not change (Figure 2b), but the distribution of
biomass among species (evenness and H') became more equitable (Figure Sb,
c). This suggests that the removal of Q album resulted in a redistribution of
resources among the remaining species, resulting in a change in the biomass
hierarchy, but no increase in the total biomass of the community. The time of
nitrogen addition significantly affected the diversity index (Table 1). Early,
mid-, and late season nitrogen addition reduced H' relative to the unfertilized
controls (Figure 2c); however this effect was significant only in the mid—
season nitrogen addition treatment (Table 2). The decline in H' with addition
of nitrogen was most pronounced when Q album was present. The magnitude

of the effect of Q album on H' was greatest when nitrogen was added mid-

season (Figure Sc).

Competition intensity

Absolute and relative competition intensity from the dominant species,
Q album, increased when nitrogen was added (Table 4). However, there was
considerable variation among replicates in both indices of competition
intensity and thus the standard errors were large. Consequently, the effect of

the time of nitrogen addition on absolute competition intensity (the absolute

29

Table 4. Effect of time of nitrogen addition on the absolute and relative
competition intensity indices (:1: standard error, see Methods for calculation of
absolute and relative competition intensity).

 

 

Time of nitrogen Absolute Relative

addition competition intensity competition intensity
None 2.6 :I: 36 0.01 :I: 0.19

July 9 94.4 a: 70 0.21 :I: 0.18

July 24 177.8 :I: 39 0.57 :I: 0.14

August 7 56.4 x 28 0.27 :I: 0.16

 

30
reduction of subordinate community biomass by Q album) was only

marginally significant (53,12 = 2.53, p = 0.11, r2 = 0.39). The effect of time of
nitrogen addition on relative competition intensity was not statistically
significant (F3,12 = 1.80, p = 0.20, r2 = 0.31). Both indices of competition
intensity were greatest when nitrogen was added mid-season, consistent with
the result that biomass evenness and diversity in the presence of the dominant
species declined most when nitrogen was added mid-season.

Both the absolute and relative indices of competition intensity increased
significantly as the biomass of the dominant species, Q album, increased
(Figure 6a, b, absolute index, F1,14 = 4.25, p = 0.06, r2 = 0.23; relative index, F1,14
= 7.1 1, p = 0.02, r2 = 0.34). However, there was not a significant relationship
between the total biomass of the plant community and either the absolute or
relative index of competition intensity (Figure 7a, b, absolute index, F1,14 =
0.48, p = 0.50, r2 = 0.03; relative index, 51,14 = 0.40, p = 0.54, r2 = 0.03). This may

be due to the single anomalous point in the lower right corner of the plots.

Light availability

light penetration to the soil surface on August 1 was significantly
higher in plots where the dominant species, Q album, was removed than in
plots where it was not removed (Figure 8, Table 5). Nitrogen addition reduced
light penetration in both the removal and no removal treatments (Figure 8,
Table 6). light penetration on August 1 was lower in the early season
nitrogen addition treatment than in the mid-season nitrogen addition
treatment, suggesting that delaying nitrogen addition within a season affected
light depletion by the plant community. The relative magnitude of the effect

of the dominant species on light availability increased with nitrogen addition

(Figure 8). Wm album reduced light penetration by 38 percent in

31

(a)

300 -

 

Absolute competition intensity

 

 

rA2 = 0.23
p = 0.06
O
-100 - . . . . . .—.—-—-I
0 200 400 600 800

Biomass of dominant species,
C. album (g/mAZ)

(b)

1.0-

rA2 = 0.34
p = 0.02

 

Relative competition intensity

 

 

‘005 ' I ' I ' l ' l
0 200 400 600 800
Biomass of dominant species,

C. album (g/mAZ)

Figure 6. Relationship between the biomass of the dominant species,
album, and a) absolute competition intensity and b) relative

competition intensity.

32

 

 

 

 

(a)

>. 300 -
t:

g I

I
E 200 - '
I

a I
.9.
a

:3 100 - I

a I

o

u 0 r42 = 0.03
a ‘ p = 0.50
E II.

8
.o I
< -100 . , . . . . . , I .
200 400 600 800 1000 1200
Total plant community biomass (g/m2)
(b)

>‘ 1.0 "

x I

so

:1

‘3

I I

on I I

g 0.5 1

~.°: - -

”J ' I

a - '-

E

u 0.0 d rAZ = 0.03
E I p = 0.54
1'5 I I I

'5';

'0-5 ' I ' l ' I ' I ' l
200 400 600 800 1000 1200

Total plant community biomass (g/m2)

Figure 7. Relationship between total plant community biomass and a) absolute
competition intensity and b) relative competition intensity.

33

0'6 - I No removal

D C. album removed
0.5 ‘
l

0.3 -'

 

 

 

0.2 4

0.1-

Percent light penetration to soil surface

 

 

 

   

 

0.0 -
None July 9 July 24

Time of nitrogen addition

Figure 8. Effect of the removal of the dominant species and time of nitrogen
addition on percent light penetration to the soil surface measured August 1.

34

88808:: 1 8:: u

Hood v m .53. ”~06 v m .5 uno.0 v m e ”cad v AZ.

 

 

 

 

 

 

 

3.0 3.4 a omé .3. mm: is Hvdm I 3.: $6 383888
388893
HOTH «:1 H5.” .4. HV.“ «:4. cm.“ .8411 ofieHHN «:1 59.” ”$.O NW.©H a g5
.88QO 82:88
8888
88.38 :38.
RA 3 was £4 .2; Eu: «.3. 91% a 26 Red 383888
888883
.P wm.~ 8.0 $6 .3. meme .2; 2.? 3d 8.0 ems: 883 888a
.88on 8888
8088888
88.58 83F
AN
1 083 .88 .8 8:5
4 wwd 4 No.0 3.0 :5 EN: 2:. 2.2 H is. omwm mad 808888 Ewfl
m m m m m m Na 89: 288$
@1888: mu888 mua888 m188=8 8188:: MICUEsc
13088 088 80388 2
083 x moo—m x 0.008 x 8308mm 8 08¢. 388mm 808
8088.88 8 8.88
.3 88 d .w

8.88m 8 8508 0.8 888 H888; .mn <>Oz< 88 88 8 8.2m $883888 «888093 m5 88 .8438
Bag 880% 888820 05 8 8088 88.58 :33 88 8088888 88.58 088 808888
88. no 80828 88.58 do 083 88 3888 80QO 8 388. .88 8883 8 maimed .8 3888 .m 8an

35

Table 6. Pairwise comparisons of the effect of time of nitrogen addition on
light penetration, tissue nitrogen concentration and total nitrogen content
using Dunnett's t-test (alpha = 0.05, df = 9, critical value of Dunnett's t = 2.853).

 

Pairwise comparison of time
of nitrogen addition

 

Variable MSD T MSE t none v. none v. none v.
early- mid- late-
season season season

Light penetration .078 .00298 *** ***

Tissue N concentration

Dominant species,

Chenonodium album 8.489E-6 *** *** ***

Subordinate community .002 2.84E-6 *** ***

Total N content

Dominant species,

Chenemdium album 1.92 *** *** ***

Subordinate community 1.45 *** *** ***

 

T MSD - minimum significant difference

1 MSE = mean standard error

36
the plots with no nitrogen addition, 87 percent in the early season nitrogen

addition treatment, and 51 percent in the mid-season nitrogen addition
treatment.
Nitrogen acquisition by the dominant species

There were significant effects of species removal and time of nitrogen
addition on the tissue nitrogen concentration of the dominant species, (1,
album, at harvest (Table 5). Because the C. album removal was done only once
(in early July), any individuals that germinated later in the season remained
in the removal treatments, but were shorter than the rest of the subordinate
plant community. This provided an opportunity to examine how the ability to
acquire nitrogen differs when a species is a subordinate in the ground-layer
(and light-limited) or a dominant in the canopy. The tissue nitrogen
concentration of Q album was significantly higher in the removal plots,
where it was a ground layer subordinate, than in the control plots, where it
was the dominant species in the canopy (Figure 9a). Furthermore, the tissue
nitrogen concentration of C, album increased as the nitrogen addition came
later in the growing season when it was both the dominant and a subordinate
member of the community (Table 6, Figure 9a).

The effects of species removal and time of nitrogen addition on the total
nitrogen content of C, album were significant (Table 5). The total nitrogen
content of (3. album was much higher in the control plots where it was the
canopy dominant than in removal plots where it was the ground layer
subordinate (Figure 9b). This result indicates that as a subordinate, C. album
acquired some of the added nitrogen (i.e., had higher nitrogen concentration
than the unfertilized control), but did not allocate it to growth, and suggests
that C, album may have been light-limited when it was a subordinate member

of the community. When Q. abum was the canopy dominant, its total nitrogen

 

37

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g 0.04 . (a) Cl C. album dominant
8 I C. album subordinate
u'
'H

° 0.03 -

Cl
3

H

a

b

g 0.02 u

c:

o

u

d

a

O 0.01 'i

b
8

0

:3
g 0.00 --
I“ None July 9 July 24 Aug 7

Time of nitrogen addition

" 12 -

28 (b) D C. album dominant
\ I I C. album subordinate
an

V 10 -

E I

a 3 -

d

“a

a 6 i

C!

3

8

u 4 d

c:

u

g 2 .1

.13

a

g 0 , 1 ____.____F_
[— None July 9 July 24 Aug 7

Time of nitrogen addition

Figure 9. Tissue nitrogen concentration of 5; album (a) and total nitrogen
content of Q, album (b) in response to removal of Q. album and time of nitrogen
addition. Note that 5;, album in removal treatment were seedlings that emerged
after the removal treatment in early July.

38
content was 4 to 5 times higher and its biomass was 2 times higher in plots

where nitrogen was added than in control plots, indicating that it had allocated
nitrogen uptake to growth and was not light limited. When it was the canopy
dominant, the increase in total nitrogen content of Q album relative to the

unfertilized control was maintained at each time of nitrogen addition (Table 6).

Nitrogen acquisition by the subordinate plant community

There was a significant effect of removing Q album on the tissue
nitrogen concentration of the subordinate community in control plots and
nitrogen addition treatments (Table 5, Figure 10a). The subordinate
community had higher tissue nitrogen concentrations when Q album was
removed than when it was present (Figure 10a). The total nitrogen content of
the subordinate community was also significantly higher when Q album was
removed (Table 5, Figure 10b).

The effect of the time of nitrogen addition on the tissue nitrogen
concentration of the subordinate community was significant (Table 5). The
tissue nitrogen concentration of the subordinate community was higher in
the mid- and late season nitrogen addition treatments than in the control plots
(Table 6, Figure 10a). The subordinate community acquired some of the mid-
and late season nitrogen addition (i.e. increased its tissue nitrogen
concentration relative to the unfertilized control). But in the mid- and late
season nitrogen treatments, it did not allocate the additional nitrogen to
growth when Q album was present (Figure 2b). However, the subordinate
community increased tissue nitrogen concentration, total nitrogen content,
and growth in the mid- and late season nitrogen addition treatments when Q

album was removed. As had been observed in Q album, the tissue nitrogen

39

(a) D No removal
I C. album removed

 

S”
O
1

E"
O
1

f"
O
1

 

Tissue nitrogen concentration
of subordinate community

 

 

 

 

 

 

 

 

0.0 -—
None July 9 July 24 Aug 7
Time of nitrogen addition
10 '-
(b) El No removal

3 I C. album removed

<
“a a 8 ‘
.. B
8
3:3 6 .
3 a
5 a
3’ o
:1 U 4 1
“a
E .8
as. 2 -

.o

:1

a

0 A;

 

 

 

 

 

 

None July 9 July 24 Aug 7
Time of nitrogen addition
Figure 10. Tissue nitrogen concentration (a) and total nitrogen content (b) of

the subordinate plant community in response to removal of the dominant
species and time of nitrogen addition.

40
concentration of the subordinate community increased as the nitrogen

addition came later in the season (Figure 10a).

Tissue nitrogen concentrations of individual species declined with
species biomass in both the control and mid-season nitrogen addition plots
(Figure 1 1). In both the control and nitrogen addition treatment, tissue
nitrogen concentrations leveled off above a species biomass of 100 g/m2,
suggesting that the species of large biomass were able to convert added
nitrogen to biomass. This suggestion is supported by two results. First, the
higher elevation of the curve for the nitrogen addition treatment indicates
that most species across the range of biomass values increased nitrogen
acquisition with the mid-season nitrogen addition. Second, most of the species
increased nitrogen concentrations when nitrogen was added mid-season
(Figure 12a). That is, most of the species are above the dotted line where tissue
nitrogen concentrations of individual species are equivalent in the control
and nitrogen addition treatment (Figure 12a). However not all species in the
community were able to allocate the acquired nitrogen to growth and increase
their total nitrogen content. The canopy-dominant species increased their
total nitrogen content disproportionately to the sub-canopy and ground-layer

species (Figure 12b).

Summary of results

The effect of the dominant species on subordinate community biomass,
community structure, and nitrogen acquisition varied with the time of the
nitrogen pulse. The effects of Q album removal and the time of nitrogen
addition on evenness and diversity (Figure 5a - c) were consistent with their
effects on the biomass (Figure 2a - c) and nitrogen uptake patterns (Figures 9

and 10) of the dominant species and the subordinate community. When

41

l y = 2.2457 + -0.43078*LOG(x) RAz - 0.611 c No nitrogen applied
5.01 y = 2.6858 + -0.26772*LOG(x) RAZ =- 0.190 I Nitrogen applied July 24

 

 

 

 

 

Tissue nitrogen concentration

 

 

I ' I ' I
400 600 800

Species biomass (g/mAZ)

Figure 1 1. Relationship between tissue nitrogen concentration and species
biomass in control plots and mid-season (July 24) nitrogen addition plots.

42

 

 

0.06 " I
g 35’ (a) ,I o CANOPY SPECIES
=5 3 o ,I’ o GROUND LAYER SPECIES
H I
3 g 0.05 _ [I A SUBCANOPY SPECIES
0 .4 I
U :1 I
d 'u I
8 V I’

«I I
a a 0.04“ ’l \
I) o I
M u I
E g A 00 . .
8 8 0.03 . E Q I tissue nitrogen concentrations
A l .
0’ I equivalent In control and nitrogen

8 g 0 g [I 0 addition plots
:3 g 0.02 . M“ A
3 '8 I’
'8 E I’
a a 0.01 . r . I 8 .
‘0 '~ 0.01 0.02 0.03 0.04 0.05 0.06

Species tissue nitrogen
concentration in control plots

 

 

,3 12 ' (b) . ,4’ o CANOPY SPECIES
5 a ’I’ o GROUND LAYER SPECIES
a. a 10 - [I A SUBCANOPY SPECIES
0 O I,
a a 0 x
8 g 8 - O I, \
u I
d ”
g I: I total nitrogen content
0 6 -I I
g go ’I equivalent in control and
a .8 ”I nitrogen addition plots
_. c: -
:3 a 4
S o ,’
s , 0
§ 3 2 ' A z’
a“)
m a O A I I I I I I

O 2 4 6 8 10 12

Species total nitrogen content
in control plots

Figure 12. Relationship between (a) tissue nitrogen concentration of species
in control and mid-season (July 24) nitrogen addition plots, and (b) total
nitrogen content of species in control and mid-season (July 24) nitrogen
addition plots.

43
nitrogen was added early in the season, both Q album and the members of the

subordinate community increased in biomass and nitrogen content relative to
the unfertilized control. Likewise, when nitrogen was added early in the
season, the removal of Q album had a small effect on evenness and diversity.
When nitrogen was added mid-season, however, the Q album response was
disproportionately greater than that of the subordinate community:
Wm album increased tissue nitrogen concentration, total nitrogen
content, and biomass relative to the subordinate community. The tissue
nitrogen concentration of the subordinate community increased relative to

the unfertilized control, but the added nitrogen was not allocated to growth
when the dominant species was present. When the dominant was removed, the
subordinate community was able to respond to a mid-season nitrogen addition
by increasing tissue nitrogen concentrations, total nitrogen content and
biomass. Likewise, the competition intensity indices and the magnitude of the
effect of Q album on evenness and diversity were greatest when nitrogen was
added mid-season. When nitrogen was added late in the season, the dominant
species increased biomass, tissue nitrogen concentration, and total nitrogen
content. The subordinate community increased tissue nitrogen concentrations
significantly with the late-season nitrogen addition, but did not allocate the
acquired nitrogen to growth (i.e. biomass and total nitrogen content did not
increase). When Q album was removed, the subordinate community increased
tissue nitrogen concentrations, biomass, and total nitrogen content in
response to the late-season nitrogen addition. With the late-season nitrogen
addition, the competition intensity indices and the magnitude of the effect of
the dominant species were smaller than in the mid-season nitrogen addition

treatments.

44
DISCUSSION

Community biomass responses and nitrogen acquisition

To my knowledge, this is the first field study that explicitly tests the
impact of within-season variation in the time of a soil resource pulse on the
competitive interactions and consequences for diversity in a terrestrial plant
community. These results Show that the time of a nitrogen addition determines
the magnitude of the competitive effect of the dominant species on
subordinate community structure, biomass, and resource acquisition. When
the dominant species was present and nitrogen was added early in the growing
season, both the dominant species and the subordinate community could
acquire the additional nitrogen and convert it to a significant increase in
biomass. However, when nitrogen was added later in the growing season (late
July or early August) only the dominant species, Q album, could increase in
biomass; the subordinate community did not increase in biomass. When the
dominant species was removed, the subordinate community increased biomass
in both the early- and mid-season nitrogen addition treatments. These results
are consistent with the hypothesis that, with increased fertility, the
magnitude of the competitive effect of the dominant species on the
subordinate community increases (Grime 1987).

The difference in the response of the dominant species and subordinate
community to nitrogen addition may have been influenced by the degree to
which a biomass/ height hierarchy had developed among species when the
nitrogen pulse was added. Early in the growing season, the plant community
had not developed a distinct canopy, sub-canopy, and ground layer. Thus, the
biomass dominant (Q album) and subordinate species were likely of more
equal size than later in the season, and thus more equivalent competitors

(Miller and Werner 1987). By mid-season (late JuIY), there may have been a

45
more developed biomass/ height hierarchy in the community, allowing the

dominant species to have disproportionate access to resources and
consequently a larger competitive effect on the subordinate community. A
more detailed examination of the physiognomy and phenology of the growth
of species in this community would be required to confirm this assumption.
Miller (1987) has shown that early advantages in emergence or growth may
become magnified through the growing season in these annual old-field
communities. Several Studies of the response of old-field plant communities to
increased nutrient availability have shown that diversity and evenness

peaked early in the growing season (May or June) and declined as the season
progressed (Bakelaar and Odum 1978, Wakefield and Barrett 1979, Hyder and
Barrett 1986, Carson and Barrett 1988, Carson and Pickett 1990), supporting the
suggestion by Miller and Werner (1987) that competitive interactions may be
weak early in the season and then increase later in the season when plants
begin to overlap spatially. These patterns may be characteristic of fertile sites
where light competition will be important due to overlap of species canopies
and disproportionate growth of the dominant species and where asymmetric
competitive interactions would preclude a more even representation of species
during the whole growing season.

One consequence of the seasonal development of a competitive
hierarchy in this plant community is that the resources limiting canopy and
sub-canopy species changed over the growing season. The patterns of
nitrogen acquisition suggest that early in the season the whole plant
community was nitrogen-limited. The analysis of tissue nitrogen
demonstrated that the dominant and subordinate species responded similarly to
a nitrogen pulse early in the season by increasing growth, tissue nitrogen

concentrations, and total nitrogen content. Later in the season, however, as

46
the canopy closed, the subordinate community was limited by light and the

dominant species was only limited by nitrogen. When nitrogen was added mid-
season, the dominant species acquired the added nitrogen and allocated it to
growth. The subordinate community was able to acquire some of the mid-
season nitrogen addition, but, because of light limitation, it could not allocate it
to growth. Consequently, tissue nitrogen concentration of the subordinate
community was significantly higher in the mid-season nitrogen addition than
in the unfertilized control. However, when the dominant species, Q album,
was removed the subordinate community could respond to the mid— and late
season nitrogen addition by increased nitrogen uptake and growth. The data
on light penetration to the soil surface in these plots in early August (between
the mid- and late season nitrogen addition) support the conclusion that growth
of Q album controlled light levels to the subordinate community. On August 1
light penetration was significantly lower in the early-season nitrogen

addition plots than in the mid-season nitrogen addition plots when Q album
was present.

The patterns of nitrogen acquisition in this study are consistent with
the results of Carson and Pickett (1990) showing a shift in resource limitation
from nitrogen early in the growing season to water and then light later in the
growing season in a SQlldagQ-dominated old-field. These results clarify the
interpretation by Goldberg and Miller (1990) that addition of a resource
limiting early in the season (nitrogen) will reduce diversity more than
addition of a resource limiting later in the season (water). They suggested that
early season nitrogen addition would lead to earlier canopy closure and
greater mortality due to light limitation. Indeed, early nitrogen addition did
lead to earlier canopy closure in this study. However, the early season

nitrogen addition was acquired by both dominant and subordinate species and

47
allocated to biomass. Though early nitrogen addition may lead to earlier

canopy closure, more species in the community would be able to increase in
biomass if the growth phenologies among Species were relatively
synchronous and the nitrogen addition came before a dominance hierarchy
was established.

The importance of position in the canopy for resource acquisition was
further demonstrated by the difference in nitrogen acquisition by a single
Species, Q album, when it was a canopy dominant and a ground layer
subordinate member of the community. When Q alblun was the subordinate
member of the community, it increased tissue nitrogen concentrations
significantly in response to nitrogen addition but did not allocate the acquired
nitrogen to above-ground biomass, suggesting that it was light-limited. When
it was the canopy-dominant species, it increased nitrogen concentration in
response to nitrogen addition and it allocated the nitrogen acquired to growth.
Hence, when Q album was a canopy-dominant, it had higher total nitrogen
content in nitrogen addition plots than control plots.

The decline in the ability of the plant community to increase biomass in
response to the late-season nitrogen pulse may have been due to harvesting
the plants before they could allocate nitrogen to growth. The tissue nitrogen
concentration of both the dominant species and the subordinate community at
harvest increased as the nitrogen addition came later in the season. The
plants clearly were able to take up the late-season nitrogen addition, but could
not convert it to biomass, probably because they were harvested before they
could do so. The dominant species and subordinate species appeared to be
physiologically capable of taking up added nitrogen at each of these three

stages in their phenology.

48
Diversity and community structure

The effect of the dominant species on subordinate species biomass and
nitrogen acquisition had important consequences for diversity and
community structure. These results are consistent with the predictions of the
"resource richness hypothesis" (Grime 1979, Huston 1979, Tilman 1982) that
the negative effect of the dominant species, Q album, on the subordinate
species performance should be correlated with a reduction in the diversity
indices. These results are also partially consistent with the prediction that the
impact of the dominant species on the diversity indices should be greater in
nitrogen addition treatments than in the control plots.

Remoyal effects. - In this study, the removal of the dominant species
resulted in a small, but significant, increase in species richness. There were
one to two more species in the community when Q album was removed.
However, the removal of the dominant species had a stronger impact on
biomass evenness than on species richness. Thus, the Shannon diversity
index increased significantly when Q album was removed. It appears that,
within a single growing season, the removal of the dominant species had a
stronger effect on the distribution of biomass across species than on the net
recruitment and mortality of species. Chenbpbdlum album may have been
removed too late in the season and a canopy established too rapidly for new
species to germinate and grow. In addition, it may take several generations
for responses in recruitment and mortality of Species to manifest. Within a
single season, however, rearrangement of the biomass hierarchy after
removal of the dominant species may create the conditions for differential
recruitment and mortality in following years. The effects of the removal of Q
album are consistent with other studies that have demonstrated an increase in

richness, evenness, or diversity after removal of the dominant species (Abul-

49
Fatih and Bazzaz 1979, Armesto and Pickett 1985, Keddy 1989b, Gurevitch and

Unnasch 1989). The effects of the removal of Q album contrast with studies
where the removal of the dominant species resulted in either no change or a
decline in species diversity (Pinder 1975, Allen and Forman 1976, Hils and
Vankat 1982). The studies in which Species diversity did not respond to the
removal of a dominant species were conducted in older fields or in younger
fields dominated by biennials, where the dominant species accounted for a
small proportion of total community biomass. In contrast, species diversity
increased upon removal of the dominant species in young fields and in a 7-
year field, where the dominant species accounted for more than 80 percent of
the total community biomass. Similarly, in my study, Q album accounted for 75
to 90 percent of community biomass.

Ilmg of W additlgn. - In this study, the response of diversity and
community structure to the time of nitrogen addition differed depending on
whether the dominant species was present in the community. When Q albmn
was present, the time of nitrogen addition had no significant impact on species
richness. The range of standing crop created by nitrogen addition to an
already fertile Site may not have been sufficient to produce a response in
species richness. Moore and Keddy ( 1989) have demonstrated that richness
and standing crop were unrelated within community types. However, over a
wide range of standing crops that included different types of communities,
there was a curvi-linear relationship between richness and standing crop. In
addition, the effects of nitrogen addition on species richness may not manifest
in a single season.

However, the nitrogen addition significantly reduced biomass evenness,
and this effect was strongest when nitrogen was added mid-season.

Consequently, the Shannon diversity index also was lower when nitrogen was

50
added, and it was lowest when nitrogen was added mid-season. Thus, like the

removal treatment, the time of nitrogen addition appeared to have a stronger
effect on the distribution of biomass across species than on the net
recruitment or mortality of Species. The strongest decline in evenness and
diversity occurred with the mid-season nitrogen addition where Q album
biomass doubled and the subordinate community biomass did not change.
These results are consistent with the hypothesis that the disproportionate
increase in the growth of the dominant species when fertility increases leads
to a decline in community diversity (Grime 1979, 1987) and with the results of
the experiment by Gurevitch and Unnasch (1989) where the effect of the
removal of the dominant species, Damdls glbmerata, on richness and evenness
was stronger in fertilized plots than in control plots. However, this study
shows that the seasonal time of the nitrogen pulse was crucial in determining
the magnitude of the effect of the dominant Species at higher fertility levels.
When the dominant species was removed, the time of nitrogen addition
had no significant impact on species richness. The effect of time of nitrogen
addition on evenness in the subordinate community was similar to, but weaker
than, the effect of nitrogen addition when the dominant was present. Thus,
there appeared to be more equitable access to the nitrogen pulses applied at all
times during the season when the dominant was removed. Consequently,
when the dominant was removed there was no significant change in the
Shannon diversity index with different times of nitrogen addition. This
pattern is consistent with the result that, when the dominant was removed, the
subordinate community doubled in biomass in response to both early- and mid-

season nitrogen addition.

Thus, in this experiment, dominance by W album led to a

decline in evenness and diversity with increasing fertility. When Q album

S 1
was removed, no single species consistently replaced it as the dominant

species. Several C3 dicots (Abunlgm Wand Ambmsla W),
a C4 dicot (Amarantbus mmflgxus), and several C4 grasses (Panigum
dichntomiflonim Setariafabm'i. mmmm andcxnenis
esgllentus) all responded positively to the early nitrogen addition.
CbgnMum album has several traits that may contribute to its
disproportionate growth in response to nitrogen addition. First, Q album is an
early-emerging, fast growing, C3 dicot (Chapin 1980, Miller 1987, Grime 1988)
potentially achieving an early size advantage. Second, Q album is capable of
consuming nitrogen in excess of the immediate needs for growth. My results
showed that the total nitrogen content of Q album increased 4 to 5-fold while
its biomass increased by about 2-fold in response to nitrogen addition. Third,
Cbmbpbdlum album has a tall, erect growth form with the potential to branch
profusely above ground, thereby depleting light to shorter plants. However, I
do not suggest that Q album is the only species that could function as the
biomass dominant and reduce the subordinate community performance and
diversity in this type of community. In a nearby annual old-field community,
where Ambrbsla anemiafifolla, was the numerical and biomass dominant and
Q album was a sub-dominant species, water addition reduced the diversity of

the community (Goldberg and Miller 1990).

Magnitude of competition intensity and mechanisms of plant
competition

The results of this experiment are consistent with the hypothesis that
increasing competition intensity along a gradient of increasing nutrient
availability and plant biomass leads to a decrease in species diversity (Grime

1973, 1979, Huston 1979, Southwood 1988, Keddy 1989a, Campbell et al. 1991b).

52
However, in this annual old-field community, the seasonal time of nitrogen

addition and presumably the Stage of development of the species biomass
hierarchy influenced the magnitude of the increase in competition intensity
caused by fertilization. The indices of competition intensity were greatest
when nitrogen was added at mid-season, and smaller when nitrogen was added
early or late in the growing season. The intensity of competition from the
dominant species was greatest at the time of nitrogen addition when the
dominant species most severely reduced evenness, diversity, and subordinate
community biomass. These results are consistent with the few studies that
report an increase in competition intensity with increasing productivity
(Wilson and Keddy 1986, Reader and Best 1989). However, these results contrast
with the larger number of recent studies that demonstrate no relationship
between fertility or productivity and competition intensity (Gurevitch et al.
1990, Wilson and Shay 1990, Wilson 1991, Wilson and Keddy 1991, Wilson and
Tilman 1991a, 1993, and Campbell and Grime 1992). There are at least two
explanations for these differences in results. These explanations clarify some
of the conflict in the literature about the magnitude of competition intensity
and mechanisms of plant competition along productivity gradients.

First, most of the studies on competition intensity (Wilson and Shay
1990, Wilson 1991, Wilson and Keddy 1991, Wilson and Tilman 1991a, 1993,
Reader and Best 1989, and Reader et al. 1994) measure either absolute or
relative competition intensity as the effect of all neighbors on the
performance of a Single target. However, implicit in the hypothesis for the
mechanism underlying the negative relationship between diversity and
fertility is the idea that asymmetric competition from one or several dominant
species (Allen and Forman 1976, Grime 1987, Miller and Werner 1987, Carson

and Pickett 1990) leads to reduced recruitment, growth, and reproduction of

S3
subordinate species. To develop an index and experimental design that is more

consistent with the logic of this hypothesis, the removal of the dominant
species at different fertility levels may be a more appropriate experimental
design. The design in this study measures the effect of a single dominant
species on the performance of the subordinate community rather than the
effect of the entire community on a single individual. Using a single target
individual may exaggerate the competition intensity relative to the effect of a
dominant species on the subordinate community. The whole plant community
may have a stronger competitive effect (sensu Goldberg and Werner 1983,
Goldberg 1990) on a single target individual than a dominant species has on
the subordinate community. Alternatively, the ability of the subordinate
community to respond (sensu Goldberg and Werner 1983, Goldberg 1990) to a
dominant species may be stronger than the response ability of a single target
individual to the whole plant community. If either of these alternatives is
true, then the magnitude of competition intensity in this and other studies
may not be comparable. The approach of removing the dominant species may
be appropriate under conditions where asymmetric interactions are likely to
occur (i.e. in plant communities where light is limiting). Miller and Werner
(1987) have demonstrated that competitive interactions in a similar annual
old—field plant community located 1 km from this LTER site were primarily
asymmetric. The result observed in my study that competition intensity was
significantly correlated with the biomass of Q album supports the conclusion
that asymmetric competition from the dominant species is important in
determining community structure and diversity in this annual old-field
community.

Second, the range of fertility and biomass levels in this and other

studies may differ substantially. The early-successional old-fields in

S4
southwestern Michigan where these experiments were conducted may

represent a gradient from moderate to high fertility, whereas the Cedar Creek
site in Minnesota (Wilson and Tilman 1991a, 1993) and the North American
prairie site in Saskatchewan (Wilson and Shay 1990) may represent a gradient
from lower fertility to moderate fertility. In addition, the rate of nitrogen
addition in this experiment was twice that applied by Wilson and Tilman
(1991a, 1993) and four times that applied by Wilson and Shay (1990) and these
studies were conducted in perennial grass-dominated communities. However,
the above-ground biomass gradient in my study (200 - 1 100 g/m2) was
comparable to that in Wilson and Tilman ( 1991a) (400 - 1200 g/mz) and Wilson
and Tilman (1993) (10 - 1000 g/m2 ), but not with that in Wilson and Shay
(1990) (100 - 450 g/mz).

Wilson and Tilman (1991a, 1993) demonstrated that below-ground
competition intensity decreased and above-ground competition increased
along a nutrient gradient. As a result of this negative correlation between
above- and below-ground competition intensity, total competition intensity
remained constant as nutrient availability increased. In this study, the
fertility gradient may represent a gradient of primarily increasing above-
ground competition. The subordinate plant community in the presence of the
dominant species was able to acquire added nitrogen, but when the addition
came mid- or late season, the subordinate community did not allocate the
additional nitrogen to growth. This suggests that nitrogen addition increased
competition intensity for light. The trade-offs in above- and below-ground
competition intensity that Wilson and Tilman (1991a, 1993) observed may not
have occurred in my study where competition intensity increased when

nitrogen was applied. Competition intensity may increase with fertility only

SS
in fertile Sites where there can be strong and increasing competition for

light.

Finally, the nitrogen acquisition results observed in this study and the
patterns of nitrogen acquisition by plants of different life-histories described
by Pate (1983) provide a possible explanation for why resource demand and
acquisition by roots and shoots, and thus, above- and below-ground
competition intensity, in an annual plant community may not be as tightly
linked as suggested by Grime (1994). Pate (1983) describes several patterns of
nitrogen acquisition in annuals from numerous plant families in Australia.
Several of the families that he describes, including Chenopodiaceae,
Amaranthaceae, Polygonaceae, occur in this Michigan old—field plant
community. According to Pate ( 1983) some annuals have shoot-dominated
nitrogen assimilation. They transport nitrate via the xylem and store nitrate
reserves in the shoot later reducing it to amino compounds in the leaves. Pate
(1983) suggests that root demand for nitrogen can be satisfied by phloem
transport of reduced nitrogen from the Shoot to the roots. This mechanism of
nitrogen uptake and assimilation would result in a decoupling of the demands
of the root and shoot within a season. Early season nitrogen uptake can be
stored as nitrate and converted to biomass later in the season when soil
nitrogen has been severely depleted. However, competition with other plants
will alter the potential allocation of nitrogen uptake to growth. If light
availability declines significantly as it did in this study, subordinate plants
will store the nitrogen taken up early in the season or in response to a pulse
and not allocate it to growth. The time scale (within and between seasons)
over which plants integrate their depletion of resources and their response to

resource depletion (sensu Goldberg 1990) may complicate predictions about

56
whether above- and/or below-ground resource competition predominates in

productive plant communities.

CHAPTER 3

EFFECT OF TIME OF NITROGEN ADDITION ON COMMUNITY STRUCTURE
AND RESOURCE AVAILABILITY IN AN OLD-FIELD PLANT COMMUNITY

57

58
INTRODUCTION

Increasing attention is being paid to the patterns of exploitation of
Spatial and temporal resource heterogeneity by plants (Campbell and Grime
1989, Naeem and Colwell 1991, Caldwell and Pearcy 1994) and the role of spatial
and temporal heterogeneity in maintaining species coexistence (review by
Tilman and Pacala 1993). Several studies have demonstrated the importance of
temporal variation in resources for plant performance. Yearly fluctuations in
rainfall favored different Species in different years in a coastal dune
grassland (van der Maarel 1981). Chapin and Shaver (1985) found that the
relative abundances of tundra species depended on annual variation in
environmental conditions (temperature, nutrients, and light). Carson and
Pickett (1990) demonstrated that three subcanopy species in a New Jersey old-
field were limited by different resources at different times of the season and
that the severity of resource limitation varied from year to year. They
suggested that resource heterogeneity could maintain species coexistence.
Relatively little attention has been paid to the consequences of within-season
temporal variation in resource availability for community structure in
terrestrial plant communities. This study examines the effects of varying the
within-season time of nitrogen addition on diversity, community structure,
and resource availability of an early successional (second-year) old-field.

Community level responses to within-season temporal variation in
resource availability might be expected because plant species vary in the time
of emergence, growth, flowering, fruiting, and senescence within a season
(Swieringa and Wilson 1972, Mellinger and McNaughton 1975, Reed 1977).
Differences in growth phenology can affect resource demands and resource
depletion patterns of different species. Their response to resources may

depend on the time of resource addition relative to their development status

59
(Mengel and Kirkby 1978, Watson 1984, Benner and Bazzaz 1985, 1988), and

relative to the development of the biomass hierarchy among individuals or
species (Benner and Bazzaz 1987, Chapter 2). I am aware of only two studies in
natural plant communities that have explored community level consequences
of within-season temporal variation in resource availability. In an annual
old-field plant community, Reed (1977) showed that a single large pulse of
nitrogen early in the season reduced diversity more than the same total
amount of nitrogen applied in smaller presses over the growing season.
However, the magnitude and time of nutrient pulses were confounded in that
experiment. In a species removal experiment (Chapter 2), I varied only the
time of nitrogen addition to an annual plant community while keeping the
magnitude of nitrogen addition constant. The dominant species reduced
biomass evenness and diversity most with the mid-season (early July)
nitrogen addition, suggesting that the degree of development of the biomass
hierarchy influenced the response of dominant and subordinate species to
nitrogen addition. In this study, the seasonal time of nitrogen addition was
varied in a second-year old field, because it has a greater number of growth
forms, life history types (annuals, biennials, and perennials), and growth
phenologies than a first-year annual old-field plant community. The pattern
of development of a biomass hierarchy and consequently the response to time
of nitrogen addition may differ from that in an annual old-field.

The objectives of this study were: (1) to document the growth
phenologies of individual species in a second-year old-field plant community
and the seasonal development of plant community structure, (2) to explore the
effects of varying the within-season time of nitrogen addition on diversity,
community structure and biomass, (3) to document the effect of the plant

community on seasonal patterns of resource availability, and (4) to link

60
seasonal patterns of resource availability with patterns of community

structure.

METHODS
Study site

The study was conducted from May to August of 1993 in a second-year
abandoned old-field at Michigan State University’s W. K. Kellogg Biological
Station in southwestern Michigan (42° 24’ N latitude, 85° 23’ W longitude). The
old-field is one in a chronosequence of l-hectare fields planted to soybeans
and abandoned in sequential years since 1989. The field in this study was
planted to soybean (Glycine max (L.) Merr.) in May 1991 with no fertilizer
application, and was abandoned in the fall after the soybean harvest. The
vegetation was dominated by summer annuals (Abufilbn membrasu,
Amaramhusretmflexus. and Chenomdiumflbum) in 1992 and by Coma
canaglensls (horseweed) with a diverse mixture of winter annuals, biennials,
perennials, and summer annuals during the summer of 1993 when this study

was conducted. The soil is of the Kalamazoo series (fine-loamy, mixed, mesic).

Experimental design

The experimental design included 6 treatments (no nitrogen addition
and nitrogen added at five different times during the growing season)
replicated four times (24 treatment plots). The treatments were arranged in a
completely randomized design using 1-m2 plots (0.5 x 2.0 m) separated by a l m
buffer on each side. Nitrogen was applied in a single pulse at 32 g N/m2 as dry
ammonium nitrate pellets (320 kg N/ acre, typical of fertilizer levels for corn
production in this region) on either May 11, May 26, June 17, July 8, or July 26.
The times of nitrogen addition were chosen to include periods where different

plant life history groups would be actively growing.

61
West of each treatment plot was a paired 0.5 x 2.0 m plot which I refer to

as a phenology plot. All of the vegetation was sampled from four replicates of
the phenology plots on the day that nitrogen was added to the paired treatment
plots. The vegetation samples from the phenology plots were used to estimate
above-ground community biomass, species diversity, and community structure
at the time when the nitrogen was added to the treatment plots and to
document the seasonal development of community structure in this field. All
of the above-ground plant biomass was harvested from the control and
nitrogen addition treatment plots on August 16 by clipping stems 1 cm above
the soil surface. Plant samples from the phenology plots and the final harvest
of the treatment plots were separated by species, dried at 60°C, and weighed to t
0.01 g.

Species richness was calculated as the number of species per plot. The
Shannon diversity index was calculated as H' = - 2 pi (log pi), where pi is the
proportion of biomass accounted for by species i. The Pielou index of biomass
evenness was calculated as J = H' / Hmax. where Hmax = log of species richness.
Species richness, evenness and diversity were calculated per plot, and are
expressed as the mean of four replicates. Evenness and diversity were
calculated based on biomass rather than numbers of individuals because
biomass has been shown to be a good predictor of competitive ability in plants
(Gaudet and Keddy 1988 and references therein; review by Goldberg and
Barton 1992).

Soil sampling and nitrogen analysis

Soil was sampled from the experimental plots on June 2, June 16, July 6,
and July 22 for analysis of mineral nitrogen availability (NO3-N + NH4—N) and
for laboratory incubations to determine potential nitrogen mineralization

rate. Soil sampling dates were chosen to follow the dates when nitrogen was

62
added to the treatment plots. Soil sampling dates were constrained by seasonal

rainfall patterns which influenced whether the soil was moist enough to be
able to sample the appropriate depth (10 cm) or dry enough to pass easily
through the soil sieve. Three 2.5-cm diameter, 10-cm deep soil cores were
collected from random locations in each experimental plot. The three cores
were composited and passed through a 2-mm sieve. A S-g sample of soil was
extracted in 50 m1 of 2M KCL for analysis of the initial mineral nitrogen
content. The extracts were shaken for 1 minute and left at room temperature
for 24-hours before filtering through a l-micron Gelman (Ann Arbor, MI,
USA) glass filter. Filtered sample extracts were stored in 7-ml plastic vials at
4°C until analyzed for nitrate and ammonium concentrations. About 25 g of
the remaining soil was placed in a loosely capped 50 ml Qorpak bottle. These
samples were incubated at 25°C and 80% relative humidity for 28 days and then
subsampled and extracted as for the initial soil samples. Gravimetric soil
moisture content was determined for each composite sample by drying soil
subsamples at 105°C for 48 hours. Soil moisture as a measure of resource
availability was calculated as a percent of dry soil mass. Soil moisture for
correcting the nitrogen content of soil extracts was calculated as a percent of
wet soil mass.

The soil extracts were analyzed colorimetrically for nitrate and
ammonium using an Alpkem Segmented Flow Analyzer (Model 3550). The N03-
N and NH4—N concentrations of the extracts were corrected for variation in
initial soil moisture and are expressed as ug NO3-N or NH4-N per gram dry soil.
Available soil nitrogen is expressed as the sum of initial NO3-N and NH4-N
concentrations. Potential mineralization rate (ug mineral-N/ g soil/ day) was

calculated as the difference between the mineral-N (NO3-N + NH4-N)

63
concentrations of initial and post-incubation soil subsamples divided by 28

days.

Light measurements

The effect of the plant community on light availability at the soil
surface was measured on May 13, May 20, June 16, June 23, July 15, and August
13 using a 40-cm long integrating light probe (Sunfleck PAR ceptometer,
Decagon Devices Model SF-80, Pullman, WA). The dates of light measurements
were constrained to cloudless days between 1 1:00 and 13:00 and to follow the
dates of nitrogen addition. Photosynthetically active radiation (PAR) was
measured above the plant canopy and 1 cm above the soil surface with the
light probe oriented north-south along the center of the 2-m length of the
plot. Light penetration is expressed as the percent of PAR above the plant
canopy penetrating to 1 cm above the soil surface. One light reading was
taken per plot and treatment levels of light penetration are expressed as the

mean percent light penetration over four replicate plots.

Statistical analyses

Analysis of variance (ANOVA) of the effect of time of nitrogen addition
on each variate was performed using the GLM procedure in SAS version 6
(1989). The error term was the replicate x time of N addition interaction term.
Analysis of variance was performed on log-transformed plant biomass data.
Percent light penetration values were arcsin square root transformed before
analysis of variance. The GLM procedure was also used to perform correlations

and regressions.

64
RESULTS
Seasonal patterns of community biomass, diversity, and structure
Total community biomass in the unfertilized phenology plots peaked in
early May, then declined and increased again to 337 g/m2 by mid-August
(Figure 13). Mean species richness increased from 15 to 20 species early in the
season, and then declined slightly throughout the growing season to 18 species
by mid-August (Figure 14a). There was a turnover of Species composition as
well as a change in Species richness over the growing season (Figure 14a and
Table 7). The decline in richness and biomass in May corresponds to the
decline of winter annuals in the community and an increase in the biomass
and species richness of biennials, perennials, and summer annuals as the
growing season progressed (Figure 13 and 14a). Annual and perennial
monocots were less common than dicots throughout the season. The Species
biomass evenness index increased early in the season from 0.38 to 0.67, and
then declined throughout the season to 0.49 (Figure 14b). The seasonal pattern
of Species biomass diversity reflected the richness and evenness patterns,
peaking in May at 2.0 and then declining to 1.4 by mid-August (Figure 14c).
The peak in total community biomass in May corresponded to the early
peak in richness, evenness and diversity. Early in the season there was a
positive correlation between community biomass and the three diversity
measures. As the season progressed total community biomass increased and
the diversity measures declined, creating a negative association between
community biomass and diversity. This pattern suggests that a biomass
hierarchy developed during the growing season and that increasing biomass
of the dominant species resulted in a decline in the three diversity measures.
The dominant species, sza canadensis, increased from 3 percent of

community biomass in May to 51 percent of community biomass in mid-August.

65

400 _ Perennial nitrogen fixers
Perennial monocots
Perennial dicots
Biennials

Annual monocots
Annual dicots
Winter annuals

300 -

ISIEDII

 

200 1

Plant biomass (g/mAZ)

 

100 -

 

130 150 170 190 210 230

A A
May 11 May 26 June 17 July 8 July 26

Julian date

Figure 13. Seasonal pattern of above-ground biomass production of different
life history groups in a second-year Michigan old—field (arrows indicate time
of nitrogen addition).

Species richness
H ._. N
o m o
l l l

U]
L

 

 

 

Perennial nitrogen fixers
Perennial monocots
Perennial dicots
Biennials

Annual monocots
Annual dicots

Winter annuals

 

I I l
147 168 182
Julian date

.0
0‘
I

p
.h
n

 

I
202

 

 

Species biomass evenness
O
U!
l

p
ua

l l l l
120 140 160 180 200
Julian date

I l
220 240

 

 

>~ c

< > it.“ i

a I ‘ .

'U 1.5. II ‘~‘~ ________ i

an "

g l

a 10- f

.2

.0

a 0.5—

.E.’

U

U

a. 0.0 l l I l I I
120 140 160 180 200 220 240

Julian date

Figure 14. Seasonal patterns of a) species richness, b) biomass evenness, and
c)biomass diversity in a second-year Michigan old—field.

Table 7. Frequency of species in phenology plots (maximum = 4 plots) and

final harvest of treatment plots (August 16, maximum = 24 plots ).

Scientific name

Winterannuals
Anthemiscmula
Amhislsmsismaliana
mm
mm
lamspp-
Lamasemnla
isnidiumcammm
Malxanezlsfla
Stellariamedia
WW

Summer annual grasses
Banimmcanillane
Ranicumdichomim'flamm
EQaannua
Setariaiumcens

S . 'Il’

Summer annual dicots
Abutilonmmhtaati
Amaranmusretmflexus
E l . ME 1'
C] l' I]

Q l' .

m. l
W .
W2 1 WI .
mummspp-

Biennials
Cirsiumaltissimum
Cirsiummlzam
maximum
Emacanadensis
Retentiilamnmica
mm

Perennial grasses
Amman teams
12m 219mm
lunms tennis

Bayer
code

ANT CO
ARBTH
BARVU
CAPBP
LAC??
LACSE
LEPCA
LEPDB
LEPVI
MALNE
STEME
VEROF
VERPG

PANCA
PANDI
POAAN
SETLU
SETVI

ABUTH

AMBEL
CHEAL
OXAST
POLAV
POLPE
POLPY
POLSPP

CIRAL
CIRVU
BRIAN
ERICA
PTLNO
VESTH

AGRRE
DACGL
IUNTE

May May
12 27
4 4
2 4
l 3
1
2
2
1
4 4
4 4
4 4
2
4 4
4 4
4 4
1
2
4 3
1
4
3
2
2

June
17

2

3

(A)

wA-b-AN-b

July

1

H

A

H-k-b-hs-h

July
21

A-h-b-hN-A

HN-h-h

August
16

24

22
24

Table 7 (cont’d).

Perennial grasses
(cont’d)
Boa amnesia

Perennial dicots
Achillea millefglimn
ASE; spp-
Cerastium mlaamm

Unidentified composite
Daucus canola
Hieracium spp-
Madigaga mama
Medicaga satin

811m: alha

Planing lancealata
We: mum
W Lecta

Rumex manna

Rumex 91211151321111:
Solarium nismm
Santiago canadansis
Iaraxacum officinala
Spammia biflata
malaria nadaliata
2191a spp-

unidentified woody spp

Perennial nitrogen
fixers

Infalium hxbxidum
Intolium mamas:
Intolium census

POACO

ACHMI
ASTPI
CERVU
com?
DAUCA
11112.7?
MBDLU
MEDSA
MELAL
PLALA
PTLAG
PTLRC
RUMCR
RUMOB
SOLNI
SOOCA
TAROF
TJDBI
TJDPE
won
WOODY

TRFHY
TRFPR
TRFRE

68

H4:

N t—Iv—iv—IA

HN

r—ANHA

NA

OJNA

24

69
Effect of time of nitrogen addition on total community biomass,

species richness, diversity and evenness

Total community biomass in the nitrogen addition treatments was twice
that of the control for all five times of nitrogen addition (Figure 15).
Community biomass in nitrogen addition plots was significantly greater than
the control plots for all times of nitrogen addition except the last addition on
July 26 (Table 8). Total community biomass increased most relative to the
control when nitrogen was added mid-season on June 17, and the magnitude of
biomass response to nitrogen declined when nitrogen was added later on July 8
and July 26. Total community biomass accumulation in the phenology plots
was beginning to level off by mid-July (Figure 13); lower productivity may
thus explain the weaker response of community biomass to nitrogen added
later in the season.

There was no significant effect of the time of nitrogen addition on
species richness at harvest (F5,13 = 0.38, p = 0.85; Figure 16a). Species biomass
evenness was lower in the nitrogen addition treatments than in the control,
but the effect of time of nitrogen addition was not statistically significant
(F5,13 = 1.35, p = 0.29; Figure 16b). There was little variation in species biomass
evenness with the time of nitrogen addition. The change in the diversity
index reflected the change in the evenness index (Figure 16c). Species
biomass diversity declined with each time of nitrogen addition relative to the
control, but the effect of time of nitrogen addition was not significant (F5,13 =

1.61, p = 0.21; Figure 16c).

Responses of the dominant species and subordinate community
The clear biomass dominant in this community was Coma W, a
tall, erect, leafy-stemmed biennial (Figure 17). The biomass of Coma

7O

Perennial nitrogen fixers
Perennial monocots
Perennial dicots
Biennials

Annual monocots
Annual dicots

Winter annuals

    
   

800-

/ /////§ /{/,/, ”I /

III§DII

600~

400-

Plant biomass (x g/mAZ)

200

 

None May 11 May 26 June 17 July8 July 26

Time of nitrogen addition

Figure 15. Above-ground plant community biomass by life history groups at
harvest in control (no nitrogen added) and 5 times of nitrogen addition
treatments in a second-year Michigan old-field.

71

Table 8. Results of analysis of variance for time of nitrogen addition on
community biomass, species richness, biomass evenness, and biomass diversity
(numerator df = 5, error df - 18). Dunnett’s t-test was used to test for
significant differences between the control and each time of nitrogen addition
(* =- significant at alpha = 0.05, df 18).

 

ANDMA Qunnettlnztest
Variable 1.2 F p None v None v None v None v None v
May 11 May 26 Jlt;ne July 8 July 26
Community variables
Total community biomass 0.63 6.06 0.002 * * * *
Species richness 0.10 0.38 0.85
Biomass evenness 0.27 1.35 0.29
Biomass diversity 0.31 1.61 0.2 1
Dominant species
(99mm canadansis)
biomass 0.44 2.82 0.05 * * *
Subordinate community
biomass 0.22 1.00 0.45
Life history groups
Winter annuals 0.59 5.17 0.004 * *
Summer annual dicots 0.26 1.26 0.32
Summer annual monocots 0.40 1.90 0.16
Biennials 0.41 2.49 0.07 *
Perennial dicots 0.04 0.15 0.98
Perennial monocots 0.20 0.17 0.91
Perennial nitrogen fixers 0.80 13.95 0.0001 * * *
Common species
Mama 101122.215 0.65 6.79 0.001 * *
Ambrosia W 0.26 1.24 0.33
W album 0.13 0.53 0.74
Qxalis grim 0.39 2.3 1 0.09
Rotmtilla amantea 0.16 0.66 0.66
Santiago canadansis 0.10 0.40 0.84
mm mm 0.89 28.13 0.0001 * * *

 

72

Species richness
8
l

 

 

0 _ None May 11 May 26June 17July 8 July 26
0.61 (b)
0.5 -
0.4 -
0.3 1
0.24

0.1-J

Species biomass evenness

 

0.0 ‘

 

None May 11 May 26June 17July 8 July 26

(C)

1.50 -

  

1.00 -

0.50 ‘

 

Species biomass diversity

0.00 -

None May 11 May 26June 17 July 8 July 26
Time of nitrogen addition

Figure 16. Species richness (a), species biomass evenness (b), and species

biomass diversity (c) at harvest in control plots (no nitrogen added) and 5
times of nitrogen addition treatments in a second-year Michigan old-field

(error bar = 1 standard error).

73

I Conyza canadensis

700 A E] Subordinate community

600 -

500 -

Plant biomass (x g/mAZ)

 

 

None May 11 May 26 June 17 July 8 July 26

Time of nitrogen addition

Figure 17. Above-ground biomass of the dominant species, (mum canadensis,
and the subordinate community in control plots (no nitrogen added) and 5
times of nitrogen addition treatments in a second-year Michigan old-field
(error bar = 1 standard error).

74
canadensis doubled in response to each time of nitrogen addition (Figure 17).

The effect of nitrogen addition on C, magmas biomass was significant on the
first (May 1 1), third (June 17), and fourth (July 8) nitrogen addition (Table 8).
The magnitude of biomass response to the second (May 26) addition was as
large as the response to other times of nitrogen addition, but there was too
much variation in biomass with the second nitrogen addition to detect a
difference. The subordinate community had a much smaller biomass response
to the nitrogen additions than the dominant species (Figure 17). The biomass
of the subordinate community was slightly higher in nitrogen addition plots
than in the control plots, but the effect was not significant for any time of
nitrogen addition (Table 8).

The reduction in evenness and diversity with nitrogen addition may
have been a result primarily of the disproportionate increase in the biomass
of the dominant species, Conyza canadensis. Species richness was not
significantly related to total community biomass (Figure 18a) or to the biomass
of Q canadensis (Figure 19a). However, community structure was closely
related to total community biomass and the biomass of Q, canadensis. Both
evenness and diversity declined with increasing total community biomass
(Figure 18b and c) and increasing biomass of C. canadensis (Figure 19b, c).
The r2 values for the regression of evenness and diversity on the biomass of Q
canadensis (0.67 and 0.72, respectively) were higher than the r2 values for the
regression of evenness and diversity on total community biomass (0.47 and
0.50, respectively), suggesting that the biomass of the dominant species, C.
canadensis, explained variation in evenness and diversity better than did total
community biomass. The negative relationship between the biomass of Q
canadensis and the measures of evenness and diversity suggest that

increasing fertility, accompanied by increased biomass of the dominant

75

 

 

 

 

 

  

 

  

 

(a)
I
I I
a 20 d I I I I
a I I
0 I I
a 15 a I ' I '
’c' I u I r2 = 0.05
" I I
an F = 1.12
a 10 " p = 0.30
.2
U
a s .
U)
0 I I T I l
200 400 600 800 1000 1200
§ 0.6 a
d
d
3
o 0.5 a
3 r2 = 0.47
a F = 19.18
2 0.4 .. p = 0.0002
.0
33 0.3 -
U
0
O.
V)
0-2 I I I I 1
200 400 600 800 1000 1200
E? 1.8- (c)
n
*6 1.6 a " n
3.
U 1.4 -
a r2 = 0.50
a 1.2 - 1= .. 22.08
.9 u p = 0.0001
n 1.0 -
8
'6 0.8“ . i . I
a
(I: 0.6 I I r T fi
200 400 600 800 1000 1200

Total community biomass (g/mAZ)

Figure 18. Relationship between total community biomass per plot and a)
species richness, b) species biomass evenness, and c) species biomass

diversity.

76

 

 

 

 

 

 

 

 

a
A A
3 20 I A A a ‘
9 A A
B ‘ i ‘ i
u 15 - A A as a
”£1 ‘ A ‘ A a A
r2 - 0.06
3:; 10 - F = 1.45
g p = 0.24
a.
m 5 .
O T I r T I
0 200 400 600 800 1000
3
o
d
C!
u
>
u
g r2 = 0.67
E F - 44.75
.9. p - 0.0001
.0
a
u
'6
o
3' 0.2 . r . . .
0 200 400 600 800 1000
>, 1.8 ‘
H
E
u 1.6 -
.2
'9 1.4 -
3
“ 1.2 - r2 - .72
g F - 55.68
3 1_0 a p = .0001
Q
.2 0.8 -
u
u
3- 0.6 . 1 . . .
0 200 400 600 800 1000

Biomass of Conyza canadensis (g/mAZ)

Figure 19. Relationship between biomass of the dominant species, Conyza
canadensis, per plot and a) species richness, b) species biomass evenness, and
c) species biomass diversity.

77
species, leads to decreased diversity in this plant community, but that there

was not sufficient power to detect the effect of nitrogen statistically in the

ANOVA in Table 8.

Responses of life history/ growth form groups and individual
species to the time of nitrogen addition

Although the effects of time of nitrogen addition on richness, evenness,
and diversity were not significant, the response of several life history groups
and individual species to the time of nitrogen addition varied significantly.
The effect of time of nitrogen addition on the biomass of life history groups
was significant for winter annuals, biennials, and perennial nitrogen fixers
(Table 8). The biomass of winter annuals at harvest was significantly greater
in the first and second nitrogen addition treatments than in the control
(Figure 15, Table 8), but their biomass in later nitrogen addition treatments
was not significantly greater than the control. The biomass of biennials
increased significantly over the control only with the third nitrogen addition.
The biomass of perennial nitrogen fixers decreased significantly relative to
the control in the first, second, and third nitrogen addition treatments, but not
with the fourth and fifth nitrogen addition. These differences among life
history/ growth form groups may have offset each other at the final harvest,
resulting in no significant effect of time of nitrogen addition on the indices of
species diversity and community structure. Because the diversity and
evenness indices are based on the proportional biomass of individual species
and because the growth phenology of species within life/ history growth form
groups varied (Table 7), I expected to find that individual species within life
history/ growth form groups differed in their response to the time of nitrogen

addition.

78
Seven species in addition to the dominant species, Conyza canadensis,

were chosen to analyze individual species responses to the time of nitrogen
addition. These species were chosen because they were abundant in all
replicates of all treatments (11 = 24) and because they represent the most
frequent and abundant life history/ growth form groups in this community.

The effect of the time of nitrogen addition on the biomass of W
mllgarls, a winter annual, was significant (Table 8). The biomass of B, means
increased significantly in response to the first and second nitrogen addition,
but not to the later nitrogen additions (Figure 20). The biomass of Ambmsla
anemlsllfella, a tall erect summer annual, increased in response to all times of
nitrogen addition (Figure 20), though the biomass response was not
statistically significant (Table 8). Ambrosia was able to reach the canopy in
this community and was the same height as C, eanadensls during the growing
season. There was no significant effect of time of nitrogen addition on the
biomass of another erect summer annual, Cbmepedlum album (Table 8). In
this plant community, C, alme was common in the subordinate community but
did not make it to the canopy. It accounted for only 1 to 4 percent of
community biomass at harvest. The effect of time of nitrogen addition on the
biomass of Qxalls 51121913. a low-growing summer annual, was marginally
significant (Table 8). Qxalls m tended to increase in biomass with the later
nitrogen additions, but did not respond to the first two early nitrogen additions
(Figure 20).

The effect of time of nitrogen addition on the biomass of two perennial
species, Retentilla argentea and Selldage canadensis, was not significant (Table
8). However, these species provide examples of the response of two perennials
with different growth forms and phenologies in this community. Retentllla

argemea, has a short spreading growth form, and was an abundant member of

79

Barbarea vulgar-is Poten tilIa argen tea

 

   

   

None May ll May 26 June 17 July 8 July 26 None May llMay 26June 17 July 8 July 26
Ambrosia artemisiifoiia g; Solidago canadensis
<
E
\
no
v
n
In
(I
E
O
.1
.D
H
I
d
E
None May 11 May 26 June 17 July 8 July 26 None May llMay 26June 17 July 8 July 26
A Chenopodium album so Tr If 01' um r epens
N
<
50
5
no 40
V
a 30
d
a 20
.9.
,9 10
a o
% None May 11 May 26 June 17 July 8 July 26

 

None May 11 May 26 June 17 July8 July 26 Time Of nitrogen addition
Oxalis stricta

1.50

.9 .0 2"
8 8 8

None May llMay 26June 17 July 8 July 26
Time of nitrogen addition

Figure 20. Above-ground biomass of seven common species in control plots
and nitrogen addition treatments in a second—year Michigan old-field plant
community (error bar a 1 standard error).

80
the subordinate community throughout the growing season (Table 7). It

responded more to the second and third nitrogen additions and less to the very
early and very late additions (Figure 20). Selidage canadensis forms a rosette
and then develops to an erect growth form. It never attained canopy status in
this growing season, though it grows to a height of at least 1 m and becomes
the dominant species later in later successional old-fields in this region. It
tended to be suppressed by nitrogen additions early in the season and did not
increase in response to any times of nitrogen addition (Figure 20) as did
Betentllla angemea. The response of Infemlm repens, a low-growing,
nitrogen-fixing species, was striking (Table 8). Its biomass at harvest was
significantly reduced by the first three nitrogen additions and not by the
fourth and fifth nitrogen additions (Figure 20). This variation in the
responses of life history/ growth forms and several individual species suggests
that the position in the canopy and time of active growth relative to the
development of the biomass hierarchy may influence access to resources and
plant growth in this plant community. The absence of a response in the
measures of community structure (evenness and diversity) obscured these

individual species responses.

Effects of time of nitrogen addition on seasonal patterns of
resource availability

In unfertilized plots, soil mineral nitrogen levels (N03-N + NH4—N)
declined during the growing season from 3.3 ug/ g soil on June 2 to 2.5 ug/ g
soil on July 22 (Table 9). Soil ammonium concentrations in unfertilized plots
were 2 to 4 times higher than soil nitrate concentrations throughout the
growing season (Table 9). In this community experimental nitrogen addition

at different times of the growing season provided a transient increase in

81

Table 9. Seasonal patterns in average soil nitrate and ammonium
concentrations (ug N/ g soil) in control (no nitrogen addition) and nitrogen
addition treatments in a second-year Michigan old-field.

w
Time of soil June 6 June 16 July 6 July 22
sample

 

NO3-N NH4-N NO3-N NH4-N NO3-N NH4-N NO3-N NH4-N

Time of

N

addition

No N 1.2 2.1 0.4 1.4 0.4 1.7 0.7 1.8
May 11 44.1 43.5 12.3 10.8 1.2 2.3 1.3 1.8
May 26 50.4 44.4 16.5 17.0 0.2 1.7 0.8 1.8
June 17 2.7 2.0 0.8 1.5 4.4 7.1 0.5 3.1
July 8 1.8 2.3 0.8 1.7 0.9 2.2 32.8 34.8

July 26 1.2 2.1 0.3 1.9 0.2 1.4 0.9 1.5

82

available soil nitrogen. Addition of 32 g/m2 of ammonium nitrate to the
experimental plots increased mineral nitrogen availability by an order of
magnitude (Figure 21, May 1 l and 26 additions). By the end of the season,
however, nitrogen levels had declined to similar levels in the control and

early season nitrogen addition treatments, suggesting that plants could deplete
soil nitrogen to similar levels in unfertilized control plots and fertilized
treatment plots or that the losses of nitrogen to plant uptake, leaching,
denitrification

and microbial immobilization were equivalent in both control and nitrogen
addition treatments.

The patterns of mineral nitrogen levels in Figure 21 suggest that the
plant community was depleting soil nitrogen faster in the middle of the
growing season (June) than early in the growing season (May). When plots
were fertilized early in the season (May 1 1) total mineral nitrogen in these
plots was 87.6 ug N/ g soil within 23 days after the ammonium nitrate addition.
In contrast, later in the season (mid-June addition) the plant community drove
soil mineral nitrogen to 7.1 ug N/ g soil within 19 days after adding ammonium
nitrate. This assumes that losses of nitrogen to leaching, denitrification, and
microbial immobilization were minimal relative to plant uptake.

Potential nitrogen mineralization rates declined between June 2 and
July 22 regardless of the time of nitrogen addition (Figure 22). Thus, both
nitrogen availability and the rate of nitrogen supply declined during the
growing season. The nitrogen mineralization rate of plots fertilized on July 8
was greater than the control on July 22, suggesting that the microbial
community may have been nitrogen-limited in mid-July.

Soil moisture in the control and treatment plots increased from June 2 to

June 16 and then decreased by July 22 (Figure 23). By the July 6 and 22

83

 

—-I— No N added

100- 3‘ ---O-- May 11
' "m."- May 26
""0"" June 17
----- km July 8
................... July 26

 

 

 

10‘

ug mineral-N / g dry soil

 

 

1

 

I I I I I I I I
130 140 150 160 170 180 190 200 210

A A A A A
May 11 May 26 June 17 July 8 July 26

Julian date

Figure 21. Seasonal patterns of mineral nitrogen levels (NO3-N + NH4-N) in
control plots and 5 times of nitrogen addition treatments in a second-year

Michigan old-field. Arrows indicate time of nitrogen addition. Note the log
scale on the y—axis.

84

 

 

 

 

 

 

 

No N added
M

0.6 . ay 11
3 May 26
2 June 17
g July 8
“A 111 26
‘5 >. J y
N :1
=5 '9 0.4 -
.. a
g 8
a S
8 I
85
E an 0.2 ~
Cl 3 \
a '-
H .
I: . \
8 Ag
8 "I

0.0 I I I I I I I—"‘l

130 140 150 160 170 180 190 200 210
A A A A A
May 11 May 26 June 17 July 8 July 26

Julian date

Figure 22. Seasonal patterns of potential nitrogen mineralization rate in
control plots and 5 times of nitrogen addition treatments in a second-year
Michigan old-field. Arrows indicate time of nitrogen addition.

85

 

 

 

 

 

 

 

—l— NoNadded
0.14- "'13-- May 11
-.-.-...-. May 26
----°---
0.12 _ June17
u """ 4"" July 8
E "iv-"m July 26
:3 0.10 -
o
E
:1 0.08 -
o
a
E 0.06 -
u
U
In
0
9* 0.044
0.02 -
0.00 r I I I I I I I
130 140 150 160 170 180 190 200 210
A A A A A
May 11 May 26 June 17 JUIY 8 July 22

Julian date

Figure 23. Seasonal patterns of soil moisture in control plots and 5 times of
nitrogen addition treatments in a second-year Michigan old-field. Arrows
indicate time of nitrogen addition.

86
sampling date, soil moisture was lower in plots fertilized early in the season

on May 1 1, May 26 and June 17 than in control plots and the treatments
fertilized later in the season. This suggests that soil moisture was depleted
more in plots that had higher plant biomass.

Percent light penetration to the soil surface declined from a grand
mean of 80% in mid-May to 19% in mid-August (Figure 24). The most rapid
light depletion in control plots occurred between mid-June and mid-July when
above-ground plant productivity was greatest (Figure 13). The time of
nitrogen addition influenced the rate, magnitude, and timing of light
depletion. Light was depleted earlier, more rapidly, and to a lower level in
plots with early nitrogen additions (May and June) than in the control plots
(Figure 24). The earlier in the season that nitrogen was applied the earlier in
the season light penetration to the soil surface was depleted. The rank of
percent light penetration to the soil surface on June 23 corresponded to the
time of nitrogen addition up to that time; light was depleted to the lowest level
when nitrogen was added May 1 1. It is important to note that although there
were effects of the time of nitrogen addition on light penetration at mid-
season (June 16 and 23), light levels were depleted to similar levels in all
treatments by mid-July and remained low until the last measurement in mid-

August.

DISCUSSION
The results of the phenology study showed that: 1) a species biomass
hierarchy developed, 2) species diversity and evenness declined , and 3)
species composition changed during the growing season in this second-year
old-field. These patterns create the potential for variation in response of

community structure to the time of nitrogen addition. The nitrogen addition

Percent light penetration

 

87

 

 

 

 

 

 

1'0 " Time of N application
. + none
0.9 ‘ \O‘. p‘ "D" May 11
i 3, ‘~- ‘~~. - ---O--- May 26
0.8-J . ‘ ----o---- June 17
I” ‘s‘. s.‘ .Y """ A "" JUly 8
' 3...; ' "***“ July 26
0.7 - O "t: .......
I: K's. ........ A
‘:. ..- '1 "A
0.6 ‘ "M" o- ‘5‘ ‘ I ‘0‘
I \ .\ ‘0
\ ‘, I. ‘0
III \\\ \.\. " a“
0.5 - , , ‘- '.
I X ‘2 l ‘.
D, \\ .‘ " X.
0.4- ‘\ '3, '1“.
\ ‘- '. '
\ ‘, I s.
\\ ‘I I 1 \“n. .
0.3 q u -‘ I. "a... ’."
\ ' 'u " ,.-"
\ ‘30. r" -43
‘ ‘- “\ o°;.— ”
0.2 - ‘I ‘I \‘s l a ’o ’ ‘-
\ 1 .:~°-... \
‘\‘- ......... ‘ ékf‘". -. '~-. .,.v"‘
0.1 d \‘b.— ’l’ \“ ”‘,o" .......
I s .v' "
b/’ ‘O" A
0-0 I I r I I I I I I I
130 140 150 160 170 180 190 200 210 220 230
A A A A A
May 11 May 26 June 17 July 8 July 26

Julian date

Figure 24. Seasonal patterns of light penetration to the soil surface in control
plots and 5 times of nitrogen addition treatments in a second-year Michigan
old-field. Arrows indicate time of nitrogen addition.

88
study showed that: 1) species richness, evenness, and diversity at the final

harvest did not vary significantly with the time of nitrogen addition, 2)
anyza canadensis, a tall biennial, dominated community biomass and was
associated with the reduction in biomass evenness and diversity, 3) variation
in the response of individual life history/ growth form groups and individual
species to the time of nitrogen addition did occur and obscured changes in the
diversity indices, and 4) although there are transient nitrogen addition
treatment effects on mineral nitrogen levels, WW3,
and light levels, these resource levels decline to the same low levels in control

plots and nitrogen addition plots by the end of the season.

Seasonal development of community structure

From mid- to late May species richness, evenness, and diversity
increased with increasing total community biomass. After this early season
peak, species richness, evenness, and diversity declined as total above-ground
community biomass increased until the final harvest in mid-August. The
observed seasonal decline in species biomass evenness demonstrates that a
biomass hierarchy developed in this plant community during the growing
season. Biomass was more equitably distributed across species early in the
season than late in the season. Similar patterns of decline in species evenness
over the growing season have been observed in other old-field studies
(Bakelaar and Odum 1978, Wakefield and Barrett 1979, Hyder and Barrett 1986,
Carson and Barrett 1988). Because individual plant biomass is an important
plant trait determining competitive ability, (Gaudet and Keddy 1988, review by
Goldberg and Barton 1992), the seasonal development of the biomass hierarchy
among species may influence the strength of competitive interactions and the
ability of different members of the community to respond to a nutrient pulse.

Miller and Werner (1987) suggested that competitive interactions may be less

89
intense earlier in the season when plants are small and do not overlap

spatially. If the degree of development of a biomass hierarchy determines
access to resources or success in converting resources to biomass, then the
subordinate species may have better access to added nutrients earlier in the
season than later in the season when dominance (in height and/or biomass)
by a few species is well established.

Species composition also changed seasonally in this plant community.
Winter annuals dominated early in the season and biennials, perennials, and
summer annuals later in the season. Within these life history groups, there
was also variation in the growth phenology of species. The results of the
phenology study describe a community with the potential of species to respond
to a nutrient pulse depending on the stage of their growth phenology and the

development of the biomass hierarchy of competitors.

Effects of time of nitrogen addition on community structure

In this study applying nitrogen at different times of the growing season
caused an increase in soil resources at different stages of development of the
plant biomass hierarchy. Although nitrogen addition reduced biomass
evenness and diversity in this community, the time of the nitrogen addition
did not influence species richness, biomass evenness, or biomass diversity.
This suggests that the degree of development of the biomass hierarchy as
documented in the phenology study did not influence the overall response of
community structure to the time of nitrogen addition. However, community
level measures of diversity and structure, such as the Shannon diversity index
and the Pielou evenness index do not account for changes in species
composition or changes in rank abundance of species in a community (Heske

et. a1 1993).

90
Although the time of nitrogen addition did not influence diversity and

community structure it did influence the final biomass of life history/ growth
form groups and individual species. The response of life history/ growth form
groups and individual species can be related to their growth phenology and
their position in the developing biomass hierarchy. Winter annuals, and
Barbarea yulgaris in particular, increased biomass significantly with the

early season nitrogen additions. These positive responses corresponded to the
time of active growth of winter annuals in this community and with a
relatively poorly developed biomass hierarchy early in the season (i.e. high
evenness). Light availability was high in early May and the summer annuals
and biennials had just begun to increase in biomass. The response of winter
annuals to the early nitrogen additions was probably larger than indicated by
the final harvest because most of the winter annuals appeared to increase
biomass in response to early nitrogen additions but they did not persist to the
final harvest.

The perennial species were subordinate members of the community for
most of the growing season in this second-year old field. As a group,
perennial dicots did not respond significantly to the time of nitrogen addition.
The responses of Petentllla angemea, a perennial dicot, to the time of nitrogen
addition were similar to those of Barbarea maria, a winter annual.

29121111113 argentea tended to increase in biomass relative to the controls in
response to early season nitrogen additions but not to later nitrogen additions.
Rotennlla memea appeared in the phenology samples early in the season
before a canopy of Cenyza canadensis and summer annuals developed. Thus, P_,
argemea may have been able to respond to nitrogen additions before

becoming light limited by the canopy later in the season.

 

91
In contrast, another perennial, Selldage canadensis, did not increase in

biomass relative to the control in response to nitrogen addition and tended to
be suppressed by early season nitrogen additions. Though Sellglagg eanaglemm
has a tall erect growth form as an adult and often dominates old-field
communities later in succession (Bazzaz 1990), as a juvenile in early
successional fields it has a rosette growth form and a relatively slow growth
rate (Bazzaz 1990). It did not appear frequently in the phenology samples until
mid-June. Thus, due to its growth phenology and subordinate position in the
biomass hierarchy at this stage of succession Selidage canadensis could not
turn added nitrogen into biomass within this growing season. However,
because it can store nitrogen acquired this season in structural roots, it does
have a growth advantage early the following growing season (Bazzaz 1990).
Among the subordinate community, Irifellum nepens, a low-growing
nitrogen—fixing perennial, showed the strongest biomass response to the time
of nitrogen addition. The final biomass of I, repens was significantly lower in
the first three nitrogen addition treatments than in the control plots.
However, I, rem biomass was equivalent in the two later nitrogen addition
treatments and in the control plots. The final biomass of I. repens increased as
nitrogen additions came later in the season. Iridium repens grew early in
the season before the summer annuals, biennials, and perennials emerged,
and it has an extended period of growth following flowering early in the
season. Early nitrogen additions may have increased the growth and
competitive effect of Llama madensls and some of the taller subcanopy
species on I. repens. Nitrogen-fixing species tend to be shade intolerant.
Thus, delaying nitrogen additions may have delayed the effect of the dominant
species on I. mbens by delaying the depletion of light near the soil surface.

This correlation between light availability, growth of the dominant and,

92
growth of I, renew; would need to be tested by experimentally removing

canopy species. However, the response of I, repens in this study is consistent
with the dominance of nitrogen fixers in low nutrient high light conditions
and a decline in their abundance in high nutrient low light conditions
(Tilman 1982).

In this study it appears that the variation in growth phenology and
position in the biomass hierarchy led to different responses among species -
and life history/ growth form groups to the time of nitrogen addition. The
responses of individual species to the time of nitrogen addition appear to have
offset each other, resulting in no effect of the time of nitrogen addition on the

indices of diversity and community structure. This result contrasts with a

 

lr

study in a newly abandoned old-field plant community dominated by summer
annuals where the seasonal time of nitrogen addition influenced the biomass
evenness and diversity of the community at harvest (Chapter 2). Germination
and growth in the first-year old-field dominated by annuals was relatively
synchronous compared to the second-year old-field in this study. There was
no winter annual component in the annual old-field community because they
were removed by plowing and discing in early May. Biennials and perennials
were also rare in the summer annual community. Thus, there was little
opportunity for early growing and late growing species to offset each other in
their responses to the time of nitrogen addition. The members of the
competitive hierarchy in the annual old-field were established early and

maintained throughout the growing season.

Seasonal patterns of resource availability
To my knowledge this is the first study of the effects of the time of

nitrogen addition on above- and below-ground resource availability measured

93
at several times during the growing season in a natural plant community.

Most studies have measured resource availability late in the season near the
time of harvest of the plant biomass and have been able to show significant
effects of nutrient additions or disturbance on light or soil nitrogen levels
(e.g. Tilman 1987, Goldberg and Miller 1990, Wilson and Tilman 1991a, 1991b).
However, the seasonal variation in both light, mineral nitrogen availability,
and nitrogen supply rate in this study demonstrates that single measurements
of resource availability within a season do not adequately describe resource
availability and may obscure the effects of experimental treatments on
resource availability. Wilson and Tilman (1993) also found seasonal variation
in light penetration and mineral nitrogen availability in a nitrogen-poor sand
plain in Minnesota.

Mid-season levels of light penetration to the soil surface declined when
nitrogen was added to this plant community. By late July, however, light
penetration to the soil surface was similar in fertilized and control plots. Thus,
if light measurements had been taken only at the end of the growing season,
the conclusion would have been that nitrogen addition had no effect on light
availability at the soil surface. The patterns of light availability in this study
contrast with those of Tilman (1984, 1987) and Wilson and Tilman (1993), where
the unfertilized plant community did not deplete light availability as low as did
the fertilized community by September and October. In the Minnesota sand
plain, light penetration to the soil surface declined when nitrogen was added
but the difference between fertilized and unfertilized plots was maintained
from mid-July through October. The vegetation on the Minnesota sand plain
has lower biomass and height than the vegetation in the old-field in this study.
Thus, the unfertilized sand plain vegetation may not deplete light availability

to the levels observed in the unfertilized plots in this study. light availability

94
at the soil surface in unfertilized plots in the Tilman (1984) study was 87% of

light availability above the canopy in July. In the Wilson and Tilman study
(1993) light penetration to the soil surface was 35-45% of incident radiation.
In contrast, mean light penetration in August in unfertilized plots in this
Michigan old-field site was much lower (19%).

This study also demonstrated the direct link between the time of
nitrogen addition and the timing of light depletion. Delaying nitrogen f
addition delayed the depletion of light availability at the soil surface. Early :
season nitrogen addition lead to an earlier, more rapid, and a greater

magnitude of light depletion than later season nitrogen additions. It appears

 

that if nitrogen is added when the plant community growth rate is highest,
then light is depleted more rapidly.

Mineral nitrogen concentrations (= resource availability) and potential
nitrogen mineralization rates (= resource supply rate) declined during the
growing season in both fertilized and unfertilized plots. While the nitrogen
additions increased mineral nitrogen availability immediately following the
addition, nitrogen levels were depleted to the same low level in fertilized and
control plots by July. Assuming that losses to leaching, denitrification, and
microbial immobilization were minimal, these results indicate that the plant
community could take up the added nitrogen rapidly and that increased soil
fertility by adding ammonium nitrate was maintained only over brief periods.
Thus, these different times of nitrogen addition functioned as pulses of
resource availability in this plant community.

Several studies have shown that mineral nitrogen availability and
nitrogen mineralization vary seasonally in grasslands (Davy and Taylor 1974,
1975, Gupta and Rorison 1975, Taylor et al. 1982), sand dunes, and old-fields

(Robertson and Vitousek 1981). But few studies in natural plant communities

95
have documented mineral nitrogen levels (Wilson and Tilman 1993) or

nitrogen mineralization rates (Pastor et al. 1987) over the growing season in
fertilized and control plots. Wilson and T ilman (1993) also found that, by
September and October, nitrogen was depleted to about 1 ug/ g soil in both
fertilized and control plots in a Minnesota sand plain. This demonstrates the
potential for soil nitrogen levels to be depleted to the same low levels in
habitats of differing productivity and the potential for nitrogen to become
more limiting to plant productivity as the growing season progresses. Pastor
et al. (1987) also found that nitrogen mineralization varied with month of the
year, but that nitrogen additions did not affect nitrogen mineralization rates

except in the oldest field (>100 yr).

CHAPTER 4

RESPONSE OF A SPECIES BIOMASS HIERARCHY TO EXPERIMENTAL TREATMENTS:
A GRAPHICAL AND STATISTICAL METHOD

96

-I... _= -" _-. --' JI- Flu-'1. —

 

97
INTRODUCTION

There is a growing interest in evaluating the influence of competitive
interactions among species at the community level (Goldberg 1994). However,
there is a lack of experimental designs and analytical approaches to examine
the response of many individual species to an experimental treatment. There
are numerous indices available to ecologists for evaluating the impact of
experimental treatments or environmental perturbations on species diversity '
and community structure (Magurran 1988). For example, species richness can
be used to evaluate whether the number of species present in a community

responds to a treatment. Diversity indices, such as the Shannon (H’) and

 

Brillouin (HB) indices, incorporate the number of species and their relative
abundance. The Pielou evenness index quantifies how evenly abundance (in
biomass or numbers) is distributed across species. To measure convergence
and divergence in community composition, similarity indices, such as the
Sorenson or Jaquard indices, are commonly used. Log abundance curves can
be used to detect whether the proportional abundance of dominant or
subordinate species changes with treatment. Rank statistics, such as Kendall’s
coefficient of rank correlation, can test whether the order of species
abundance remains constant in response to the treatment. However, the
identity of individual species is lost in all of these analyses of species diversity
and community structure. An experimental treatment may result in no
change in the species richness, the Shannon index or the Pielou evenness
index, while changes do occur in individual species rank abundance (Samson
et. a1 1992, Heske et. al 1993). None of these community level indices allow us to
examine individual species abundance responses in a multi-species context.
Nor do they permit a determination of whether the responses of species to a

treatment are dependent on some characteristic of the species.

98
In this chapter, I describe a graphical and statistical method to address

two null hypotheses about the response of a species abundance hierarchy to
an experimental treatment or an environmental perturbation. The first null
hypothesis is that there is no change in the hierarchy of species abundance
in response to an experimental treatment. The second null hypothesis is that
the magnitude and direction of the response of species to a treatment is
independent of their abundance in the control (or pre-perturbation
treatment).

The question of whether the absolute or relative response is more
appropriate for measuring species response to competition has raised a major
controversy in the literature in recent years (Campbell and Grime 1992, Grace
1993, Wilson and Tilman 1993, Goldberg 1994, Reader et. a1 1994). This problem
extends to measuring the response of species to any experimental treatment or
environmental perturbation. The absolute and relative responses of species to
a treatment may differ and lead to conflicting results and conclusions. For
example, species A may increase from 0.5 in the control to 1.0 in the treatment,
and species B may increase from 5.0 to 5.5. The absolute responses of the two
species are equivalent (0.5), but the response of species A relative to the
control value ( 100%) is larger than that of species B ( 10%). It is not yet clear
theoretically when to use absolute or relative changes in abundance. We are
only beginning to accumulate empirical patterns of absolute and relative
responses of species to competition in the literature. Therefore, in this
chapter, I calculate both the absolute and relative measures of species
responses to an experimental treatment.

Specifically, I examine the response of species in old-field communities
to nitrogen additions. The importance of biomass as a predictor of competitive

ability and the potential for asymmetric and transitive competitive

 

99
interactions suggest that the species responses to nitrogen additions should be

dependent upon the biomass of the species in the control. The biomass of a
species is an important predictor of competitive ability in plants (Miller and
Werner 1987, Gaudet and Keddy 1988 and references therein, review by
Goldberg and Barton 1992). On a population level, Wiener and Thomas (1986)
concluded that asymmetric competition is common in plant monocultures and
leads to size hierarchies in annual plant populations (Weiner and Solbrig 1984
and Weiner 1985). On a community level, Miller and Werner (1987) showed
that, in an annual old-field plant community, competitive interactions were
primarily asymmetric. Keddy and Shipley (1989) reviewed eight published
experiments and found that pairwise interactions were asymmetric and that
competitive interactions were transitive (for competing species A, B, and C, if
the effect of A on B > the effect of B on A, and the effect of B on C > the effect of
C on B, then the effect of A on C > the effect of C on A).

The graphical and statistical method described in this chapter addresses
three specific questions about species responses to nitrogen addition in old-
field plant communities. First, does the hierarchy of species abundances in
the control plots remain constant when nitrogen is added at different times of
the growing season? Second, does the magnitude and direction of species
responses to the nitrogen addition depend on the abundance of the species in
the control (i.e. do more abundant species have a larger response to the
nitrogen addition than do rare or less common species)? Third, are the
patterns of absolute and relative responses to nitrogen addition similar among
species of different abundance? In this chapter, I first present a graphical
and statistical approach to testing the second null hypothesis. Then I test the

efficacy of this method using data from a species removal and nitrogen

 

100
addition experiment in a newly abandoned old-field (Chapter 2) and a nitrogen

addition experiment in a second-year old-field (Chapter 3).

METHODS

Testing rank correlation of species abundances

I used Kendall’s coefficient of rank correlation in this analysis to test
the first hypothesis of whether the rank order of species is the same in the
control and treatment plots. The mean species biomass values across four
replicates within a treatment were ranked in descending order. The CORR
procedure in SAS version 6 (SAS 1989) was used to test for correlations between
the rank order of species abundance in control plots and nitrogen addition
treatments. Only species that were present in both the control and nitrogen

addition treatment within a comparison were used in the correlation analysis.

Graphical and statistical analysis

To test the second null hypothesis that the magnitude and direction of
species biomass responses to the nitrogen addition are independent of the
abundance of the species in the control plots, the absolute biomass response to
the nitrogen addition treatment was plotted against the mean species biomass
in the control (see example in Figure 25). The absolute biomass response is AB
= Bt - BC, where Bt is the mean biomass of the species in the treatment, and BC
is the mean biomass of the species in the control. Only the species that are
present in both the control and treatment are included in the regression.
Thus, this method is most appropriate for situations where species composition
in the control and treatment are similar and the communities are relatively
species rich so that there are a sufficient number of data points for a

regression.

101

 

 

 

Figure 25. Absolute biomass response to an experimental treatment. Each
point represents a single species. The dashed line represents no change in
species abundance in response to the treatment. Points above the dashed line
represent species that increased in response to the treatment; points below the
line represent species that decreased in response to the treatment. The solid
line is the fitted regression line through the data.

 

102
The null hypothesis that the species biomass responses to the treatment

are independent of their biomass in the control plots is the line of slope = 0,
indicated by a dashed line in Figure 25. The y-intercept may be greater than
or equal to zero. Points above the dashed line in Figure 25 represent species
that increase in biomass in response to the treatment. Points below the dashed
line represent species that decrease in biomass in response to the treatment.
The slope of the fitted regression line, a solid line in Figure 25, indicates
whether the magnitude and direction of species biomass responses to the
treatment are dependent on their biomass in the control plots. The criterion
for rejecting the null hypothesis is provided by the standard regression
analysis testing whether the slope and y-intercept are different from zero.
Several examples of potential absolute biomass responses are illustrated
in Figure 26. The dashed line represents the null hypothesis. In Figure 26a,
the slope of the regression, represented by a solid line, equals zero and the
intercept is significantly greater than zero. The interpretation of this
response is that species of all abundances in the control plots increased in
biomass by the same absolute amount in response to the treatment and this
response was independent of the biomass of the species in the control. In
Figure 26b, the y-intercept = 0, and the slope of the regression is significantly
greater than zero. The interpretation of this response is that the absolute
response of species to the treatment was dependent on their abundance in the
control plots. Species of all abundances in the control plots increased in
biomass in response to the treatment, but the absolute responses of the more
abundant or dominant species were larger than that of the least abundant
species. In Figure 26c, the slope of the regression is significantly greater
than zero, and the y-intercept is significantly less than zero. In this case, the

absolute responses of the species are dependent on their biomass in the

 

103

 

 

 

 

 

 

(a)
U slope = O
a? ... T T y-intercept > 0
an
0 ————————
B
C
(b)
U
m L..-
at: "’ A Slope > O
0 ________ . y-lntercept = 0
B
(c) C
U
a? $10pe > O
m H y-intercept < 0

 

 

 

C

Figure 26. Potential responses of species to an experimental treatment. The
dashed line represents the null hypothesis and the solid line represents the
fitted regression line. In (a), all species increase in biomass in response to the
treatment and the magnitude of their response is independent of their biomass
in the control. In (b), the biomass of species of all abundances increased and
species of larger biomass in the control have larger magnitudes of response in
the treatment. In (c), species of low abundance in the control decrease in
biomass and species of high abundance in the control increase in biomass in
response to the treatment.

104
control. However, the more abundant or dominant species increase in biomass

and the least abundant species decrease in biomass in response to the
treatment.

The r2 value of the fitted regression line indicates the degree of random
variation of species responses to the treatment. A small r2, associated with a
large amount of scatter, indicates that the magnitude and/or direction of the

response of different species to the treatment varied randomly with respect to

 

r
their abundance in the control plots. Various response patterns could produce H
a low r2 value. For example, all of the species responses to the treatment could .1
be positive, but the magnitudes of response could vary independently of the ‘
biomass of species in the control plots. Or, the responses could be both positive L

and negative, but vary randomly with respect to abundance in the control
plots.

To test whether the relative biomass response of species to the treatment
is dependent on the biomass of species in the control, the relative biomass
response, %AB, can be calculated by dividing the absolute biomass response by
the biomass of the species in the control, (Bt — BC) / BC. Figure 27 shows
several potential relative biomass responses that are analogous to the
responses in Figure 26. In Figure 27a the subordinate species have a larger
relative biomass response than the dominant species. Although species of all
abundances in the control increase by the same absolute amount (Figure 26),
the absolute response is a larger proportion of subordinate species biomass.
The slope is less than zero and the y-intercept is greater than zero. In Figure
27b, the relative biomass response to the treatment is equal for species of all
abundances in the control. The slope = 0 and the y-intercept is positive.
Whether the slope =- 0 in this case depends on the magnitude of the slope of the

response in Figure 27b. In Figure 27c the relative response is large and

105

 

 

 

 

(a)
U
e slope < 0
m A y-intercept > 0
d
O ________
B
(b) C
0
no slope = 0
E T 1. y—intercept > 0
<l
O _________
B
(C) C
U
co slope > 0
E y-intercept < 0
<1 0

 

 

 

Figure 27. Potential relative responses of species to an experimental
treatment. The dashed line represents the null hypothesis and the solid line
represents the fitted regression line. In (a), all species increase in biomass in
response to the treatment and the magnitude of their relative response is
larger for species of lower abundance in the control than for species of
higher abundance. In (b), the biomass of species of all abundances increases
and species of all abundances in the control have equivalent or almost
equivalent relative biomass responses. In (c), species of low abundance in the
control have a large negative relative biomass response and species of high
abundance in the control have a large positive relative biomass response to
the treatment.

106
negative for species of low abundance in the control plots and large and

positive for species of high abundance. The slope of the response is positive

and the y-intercept is negative.

Field experiments
The data used to test the efficacy of this graphical and statistical method

are from two field experiments conducted in a newly-abandoned old field and a

 

second-year old-field at Michigan State University’s W. K. Kellogg Biological '-
Station in southwestern Michigan (Chapter 2 and 3, respectively). The soils,

land-use history of the sites, and the details of the field experimental designs

are described in Chapters 2 and 3. In summary, the experimental design for

the study in a newly abandoned field at the Long Term Ecological Research I.

(LTER) site in agricultural ecology included 4 blocks X 2 species removal levels
(no removal, removal of the dominant species, Chenepedlum album) X 4 times
of nitrogen addition (none and 0, 2, and 4 weeks after removing Q album). The
treatments were arranged in a complete randomized block design using 1 m2
plots (2 m X 0.5 m). Cbenepedlum album was removed between July 6 and 8,
1992 by pulling or clipping seedlings. Nitrogen was applied in a single pulse
at 32 g N/m2 ammonium nitrate (320 kg/ha, typical of fertilizer levels for corn
in this region) 0, 2, and 4 weeks after removing Q album. This experimental
design produced 6 pairs of comparisons of species biomass: control v. early
(July 9), mid- (July 24) and late (August 7) season nitrogen addition with Q
album present, and control v. early, mid-, and late season nitrogen addition
with Q album removed experimentally. The response of the plant community
to a nutrient pulse was measured as above-ground biomass at harvest on
August 21 two weeks after the final nitrogen addition. The plants were
harvested by clipping the stems at the soil surface. Plants were separated by

species, dried at 60°C, and weighed to = 0.01 g.

107
The experimental design in the second-year old-field included 6

treatments (no nitrogen addition and nitrogen added at five different times
during the growing season) replicated four times (24 treatment plots). The
treatments were arranged in a completely randomized design using l-m2 plots
(0.5 x 2.0 m) separated by a 1 m buffer on each side. Nitrogen was applied in a
single pulse at 32 g N/m2 as dry ammonium nitrate pellets on either May 1 1,
May 26, June 17, July 8, or July 26, 1993. All of the above-ground plant biomass :-
was harvested from the control and nitrogen addition treatment plots on
August 16 by clipping stems 1 cm above the soil surface. Plant samples from

the phenology plots and the final harvest of the treatment plots were

 

separated by species, dried at 60°C, and weighed to 1 0.01 g. L
The correlation and regression analyses were conducted using the CORR

and GLM procedures, respectively, in SAS version 6 (SAS 1989). Only data

points for species that were present in both the control and treatment

comparison were used in the analyses.

RESULTS

Annual old-field: Relationship between species frequency and
abundance

Only species that were present in both the control and the nitrogen
addition treatment within a comparison were used in the correlation and
regression analyses. In the annual old-field, the species that were excluded
because they appeared in only one of the pair in a comparison tended to be
species of low mean biomass and low frequency (Table 10). There was a
significant positive correlation between the mean biomass of species and their
frequency across replicates (Pearson r = 0.35, p = 0.0001). The replicates over

the large spatial area of this experiment occur over a SO-hectare area of the

108

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HQNHH. 8N H. HHN H. 30 H. 2.3 m Re N 2.3 N 2.3 H. 5%

HwNHH HHH N HmN. N 8.8 H 83H 2.: «Huang

HNo. N HHHN. H HMS H SHN HNHN gas”

so: Em H HmN. H 8.: “Huang

3.8 H SNHHN 02 N SNN H 8.5 H 8.5 H HN.NH.HN CNH N 8333

HHHHHHH odeN :3 H HH.H.HH HNHHN .22: 8.3 H gunman

HmNHH $.va H; H gag

HH.H..HH HHHH 8N. H and;

83 H HNmH N :5 N «:33

:3 H «gang

HNOH H H83 H Woo; H «gag

33 H 83 H HH.H..HH :3 H $3 H .5: 3683833

88933

HH..mH.HH. 2.2? $.31. 2.: m H3; HN.mHHH. HNéH. 3.9 H. Egg

3: H Hm: N No; H No; H Hm: H gag

83 H 8N: so; H H3 H :NHN 83 N gang

3H,: 8.: H HNHH H Hag

HmN H an;

2%. H Hem H HméN 83H 38%

H: H HHHLH am. H 3N: HH.N. H SH: 22%

38:. 2% H. 8.3 8.2 H. HNdHHN 8.: H. SHNHN Hue 4 Egg

NEH HHN H 3%: ANN. H HNSH 2.0. H 338%

3.3. a.» H. $.va 8.: m H1884 3m; H. HNHHHHi. swig H. Huang

3.: NH. H AN: 33 H HH.o.HH 6H. H $3 H 353%

oN H HHH H Managua

E: 30.: SNH N HE; Nm. H So. H 83 N gang

mNm: 8.2: 6.3 H HNHNH H mNN H 3m H 2.6 H $.98 H «Hagan;

HHNNHH. 682:. WNHH H. 3.03 H. w.H.H. H. on H. m.mm H. 8.9.1. 338835585

6.34 $61.. Nd» H. 3.5 H. HHNN H. HEN H. Ham H. Hm.H.NHH. 93855333

N a: 3. 2+ H862 2 62 N a: 3. 2.. N a 2+ 6 H: 2+ 683262 806 m
639:“: 8% d H535 3% d

.3355de

E ANSRV $385 538 :05 U5 3085 who; 86on 533 E .38 .Ho 50 88338 Mo $2832 .9 2an

109

 

 

 

 

 

 

 

«3803
3.3 H .383 35chqu

3m; a 3”; a QC; H Ame; fl av; H gag
:3 H 383%
83 N 313838338

8.503 .OH 2an

1 1 0
LTER site at KBS. There is considerable variation in species composition in

these annual old-fields (Gross and Huberty, in prep). Thus, several rare
species with patchy distributions at the SO-hectare scale did not appear
consistently in the experimental plots. Between 6 and 12 species with mean
biomass ranging from 0.003 to 8.3 g/m2 were excluded from the six
comparisons leaving 12 to 18 species in common for each comparison (Table

1 1). The values used in the correlations and regressions are the mean biomass

 

T
I
of the species across four replicates. Thus, a species is represented once in 2.
each of the correlation and regression analyses. 3
Annual old-field: Correlations of species rank abundance L

Within a single growing season, the overall biomass hierarchy of
species was maintained when nitrogen was added to the annual plant
community at different times during the growing season. The Kendall tau-b
correlations between the rank abundance of species in the control and
nitrogen addition treatments were significant when the dominant species,
Chengpgdium album, was present and when it was removed experimentally
(Table 12). Therefore, the first null hypothesis that the rank of species
abundances does not change with nitrogen additions can be accepted.

Although the rank abundance of species did not change with nitrogen
additions, the correlation analysis does not show whether species of different
abundances in the control plots responded positively or negatively to the
nitrogen additions or whether the magnitude of the response was dependent

on species abundance.

111

Table l 1. Number of species occurring in both the control and nitrogen

addition treatments that were included in the correlation and regression
analyses, and number of species that were excluded because they did not
appear in both the control and the nitrogen addition treatment.

 

 

 

 

 

Treatments compared No. of species present No. of species
in both treatments of excluded
comparison

Q album present

Control v. N added July 9 12 9

Control v. N added July 24 13 1 1

Control v. N added August 7 12 12

C. album removed

Control v. N added July 9 18 6

Control v. N added July 24 18 8

Control v. N added August 7 16 10

 

112

Table 12. Spearman and Kendall tau-b correlation coefficients for rank of

species mean biomass in control and nitrogen addition plots.

Variables N

 

 

 

Spearman p Kendall tau-b p
correlation correlation
coefficient coefficient
C. album present
Control v. N added July 9 12 0.85 0.0005 1.64 0.004
Control v. N added July 24 13 0.90 0.0001 0.79 0.0002
Control v. N added August 7 12 0.75 0.005 0.64 0.004
5;. album removed
Control v. N added July 9 18 0.73 0.0007 0.54 0.002
Control v. N added July 24 18 0.89 0.0001 0.74 0.0001
Control v. N added August 7 16 0.80 0.0002 0.58 0.002

 

 

1 1 3
Annual old-field: Absolute biomass responses

In an absolute sense, the positive response to the nitrogen additions was
larger for the dominant species than for the less abundant species in the
annual old-field (Figure 28). When the dominant species, Cbengbbdlum mm,
was present, the slopes of the regressions of species biomass in each of the
three nitrogen addition treatments against species biomass in the control plots
were significantly greater than zero (Table 13). W album,
represented by the isolated data point in the upper right of Figures 28a-c, was
abundant in the control plots and had a large absolute response to the
nitrogen addition. Consequently, it exerted a large influence on the
regression.

The dominant species, Q album, accounted for a large proportion of
community biomass in this study and had the greatest absolute increase in
biomass in response to the nitrogen additions. Consequently, the data is not
bivariate normally distributed when 5; album is included in the analyses.
When I; albmn was excluded from the regression analyses (Figure 29), the
results and interpretations of the regressions were different from those in
Figure 28. When nitrogen was added early in the season, the magnitude of the
positive response of species tended to increase with their abundance in the
control plots (Figure 29a). However, the slope of the relationship was not
significantly different from zero (Table 13). Thus, the response was
qualitatively similar to the results in Figure 28a. When nitrogen was added at
mid-season, the absolute change in biomass was negative for species of greater
than 20 g/m2 in the control and positive for species of less than 20 g/m2 in the
control (Figure 29b). The slope was significantly less than one and the y-
intercept was not significantly different from zero (Table 13). This result and

interpretation differs radically from that in Figure 28b and illustrates the

 

114

C. 311mm present
300 ' (a) N added July 9

 

 

 

 

 

 

U
m H
”100 I T I 1
0 100 200 300
30° ‘ (b) N added July 24
U
m
m H
‘100 | I I l
0 100 200 300
150‘ (c) N added August 7
U
m
m H

 

 

 

T I I
O 100 B 200 300

Figure 28. Absolute species biomass responses to a) early, b) mid-, and c) late
season nitrogen addition with Q album present in the community (dashed line
= null hypothesis; solid line = fitted regression line).

 

115

”(El ..u...._ -P I q‘... . ca...

HanucEwHw Ho: 385me8 .P
:80chme 8: u m: s

 

 

 

 

ma wmd m.m wmd o o wmd 3 n Hm=w=< 328 z .> 3.950
8: de o.m- Hood 3.“ Omd Hood 3 N 3% 338 Z .> 8.980
m: 3.0 ms vod vfio H.N.o vod 2 m 33 bound 2 .> 3ch0
3388 and? d
m: H.m.o wN N Hd H. mmd. £6 :6 H H n umsms< 328 z .> 65:00
8 NH..o N 86 and. 93 86 NH H.N .HHHHN 828 z .> H888
8: mod 9m 36 omd NNd mNd S m 33 noun“ 2 .> 65:00
flax—mam
conme8 80.: oops—88
:5 £535 and? d
8: Mad Wm- good 2.6 and 5006 NH 5 8&3. Boom z .> 8.980
vod 03. good mad cad Seed 3 N 3a 32:. z .> 3.950
*8: £6 H..H good wad nwd Hocod NH a 33 canon z .> 3.950
2835 g d

H9885

N8 58:35 258

a 4A .5“ a doom NC a 688 z E82388 H5803;

 

.o .m .H. .me S .839... 303 wEUHHoammte BE. .83 8on Anamowmdwfi 8.8wa 583::

4A .5 moo—m o5 85 88:2: 83209 5 8:85 88385... :23on3 m5 5 39:05 m8 5258.5 USN H9980
Son 5 63.3qu 85 86.4.9. boo .Amoofimz com @3098 82823 303 8ch 98 383.». mausouw
2t no 883 688% 8 328 53.25 6253.5 H8253 88803 888 S mango o5 USN Gonna cmwobE
05 moo—a 8.280 5 86on No $8805 S88 5253 952528 m5 .8“ 83338 :onmwNwom .3 «BE.

1 1 6
Q. 3.1mm present, but excluded from regression analysis

 

 

 

 

 

 

 

 

100 " (a) N added July 9
80 - °
60 -
U
a:
I 40 "
H
0
an o o
20 ' o
o«a----3-----------;---
'20 I I I I I
0 10 20 3O 40 50
U
m
| H
m
U
m
' H
at:
-30. o
'40 I I T I I
O 10 20 30 4O 50
B
C

Figure 29. Absolute species biomass responses to a) early, b) mid-, and c) late
season nitrogen addition with C. album present in the community but excluded
from the regression analysis (dashed line = null hypothesis; solid line = fitted
regression line).

 

1 1 7
influence that one species with a strong response can have on the regression

results and interpretation. When nitrogen was added late in the season, the
slope was only marginally different from zero and the y—intercept was zero
(Figure 29c and Table 13), suggesting that the negative response to nitrogen
additions was larger for the more abundant species.

When C, album was removed experimentally, the more abundant species
had a larger absolute increase in biomass in the early and mid-season
nitrogen addition treatments than did the less abundant species (Figure 30a,
b). The slopes of the regressions of absolute response against the biomass of
species in control plots were significantly greater than zero for the early and
mid-season nitrogen addition (Table 13). The slope was not significantly
different from zero when the dominant species was removed and nitrogen was
added late in the season (Figure 30c, Table 13), indicating that the response to
late season nitrogen additions was independent of the abundance of the

species in the control plots.

Annual old-field: Relative biomass responses

The relative biomass response to nitrogen addition was independent of
the abundance of the species in the control plots for all three comparisons
when Q, album was present in the plots and included in the regression
analyses (Figure 3 1). The slopes and the y—intercepts were not significantly
different from zero (Table 14). The relative response of the (L album, the only
species with mean biomass greater than 100 g/m2 in the control plots, was
positive in the early and mid-season nitrogen addition (Figure 3 1a, b).
However, the direction and magnitude of relative biomass responses among
species with mean biomass less than 100 g/m2 in the control plots were highly

variable. Consequently, when Q, album was present in the community and

 

 

118

C. album removed experimentally
80 (a) N added July 9

 

 

  
   
 

 

 

 

 

U
at:
' H
m
100 ‘ (b) N added July 24
80 - ‘
u 60 4
m
' ... 40 + ‘
an
20 .
0 ...A. ...............
A
'20 | I I l
O 20 4O 6O
6O -
(c) N added August 7
4o -
U ‘ ‘
”f' 20 -
H
In ‘ ‘
O 3 " - - " - - - 2 -----------
A
'20 | l l I
O 20 4O 60
B
c

Figure 30. Absolute species biomass responses to a) early, b) mid-, and c) late
season nitrogen addition with Q album removed from the community (dashed
line = null hypothesis; solid line = fitted regression line).

 

 

119

 

 

 

 

 

C. album present
3 ' I (a) N added July 9
I
u 2 " I
In
\
A u 1 -
u
"5' . -
m H
v o ---—. ------------------
I
-1 I I I fl
0 100 200 300
4 (b) N added July 24
u 3
m
2 2
U
can .
I l I I
H
5
o .-_ ------------------
'1 l l l
O 100 200 300
20 1 (c) N added August 7
U
an 10 q
\
A U
can
. a 0 ............. .3.-.
E
‘10 l I l I
0 100 B 200 300
c

Figure 31. Relationship between the biomass of species in the control plots
and the relative change in mean biomass of species with a) early, b) mid-,
and c) late season nitrogen addition and C, album present in the community
(dashed line = null hypothesis; solid line = fitted regression line).

 

120

 

835:me 8c zHHmnkumE H
88chme 88 n m: ..

 

 

 

 

m: de H.H de o mod de 0H m .6an2. panda z .> 3.980
Nod v.H omd 0 No.0 omd wH E .22. nouns z .> 3.980
m: NHd de $6 0 mod mmd wH m. 32. 889... z .> 3.980
H8888 94an .0
mod ~.m NHd .H 2.0. mmd NHd HH 5 ”6:93. 3pr z .> 3.980
86 ~.H mod mod. Hmd 006 NH VN baa cuppa z .> 3.980
m: and mod. 36 o 36 £6 H H m 32. pound z .> 3.980
39385
.833qu :8: 9,8988
82 .8888 833 d
8 m Hd w.~ H md o mod Hmd NH h 3&3. pound z .> 3.980
8 o~.o mmd cod o .8006 cad mH VN 33.. cocoa z .> 3.980
*8 ovd mmd cod o mod cod NH m 32. wanna z .> 3.980
8888 End? :0
988.89:

.80 8889: 8on

a 4A 80 a 8on N.— Q 388 z cemtndfiou 8258b.

 

 

.m 98 .w K .me 9 .28an 303 wfipcoamota 22. .83 88b huguSame
3th8 8889:-» 8 8on 85 85 8889 @3839 9 8:15 .8338 .868qu

85 5 H8939 8.8 88888.: 98 3.98u :82 9 H8888 85 38QO >30 .Hmuofimz

93. 6289.3 352mb 3.988 98 3038 9:288 85 8 883 898003 9" 328 58.95
88808.5 58253 310805 939 9 $923 «>923 m5 H98 628m 88.98 9: 303 35:8
5 $3QO 8 89:03 988 82588 8328928 85 80 8333a aonmmummm .vH 29d.

1 2 1
included in the regression analysis, the value of its relative response exerted

disproportionate influence on the regression.

When Q album was excluded from the regression analysis, the direction
and magnitude of the relative biomass response to the early season nitrogen
addition was independent of the species biomass in the control plots (Figure
32a, Table 14). When nitrogen was added mid-season, the slope was
significantly negative, and the y-intercept was significantly greater than
zero (Figure 32b and Table 14). The relative response of species of less than 20
g/m2 in the control plots was positive and the relative response of species of
more than 20 g/m2 was negative. The slope was negative and marginally not
significant when nitrogen was added late in the season (Figure 32c, Table 14).
The analysis excluding Q album in Figure 32 illustrates the influence of the
dominant species on the regressions in Figure 3 1.

When Q album was removed from the community experimentally, the
relative biomass responses to nitrogen addition were independent of the
biomass of the species in the control (Figure 33, Table 14). The direction and
magnitude of relative responses to each time of nitrogen addition were highly
variable. The positive relative responses tended to be larger and more

variable than the negative relative responses.

Second-year old-field: Absolute biomass responses

In the second-year old-field the absolute biomass response to each time
of nitrogen addition was positive and larger for species of greater biomass in
the control plots (Figure 34). The slopes were significantly greater than zero
and the y-intercepts were not significantly different from zero (Table 15). As
had been observed in the annual-field, the dominant species in the second-

year old-field, Conyza canadensis (a tall, densely leaved composite), accounted

 

122
Q album present, but excluded from regression analysis

- (a) N added July 9

 

 

 

 

 

3
o
o
u 2.. .
an
\
A
u 1-
In
, o o
H
a:
v
u
in
\
A
u
an
I
H
an
v
u
:I:
\ 10"
A
u
an
I
H
an O
V

 

 

Figure 32. Relative species biomass response to a) early, b) mid-, and c) late
season nitrogen addition with Q album excluded from the regression analysis
(dashed line = null hypothesis; solid line = fitted regression line).

 

123

5; album removed
200 . (a) N added July 9

 

 

 

 

 

u 150 'l‘:
:I:
2 100«
U
an
H
5
OTQJ-m-U-a- ------ a—--
'50 | l I I
O 20 40 60
8 (b) NaddedJuly 24
u 6
an
2 4
U
an n
. H 2 a D
5 n
0 E___Q_a__u ___________
'2 I I I I
20 4O 6O
101(c) Nadded August7
8d
U
an 6-
\
Au 44
an
I 2 a
a" "
o gum-na- ------ u---
-2 I I I I
o 20 B 40 60

Figure 33. Relationship between the biomass of species in the control plots
and the relative change in mean biomass of species with a) early, b) mid-,
and c) late season nitrogen addition and Q album removed from the

Community (dashed line = null hypothesis; solid line = fitted regression line).

 

_‘ _

400 -

(a) N added May 11

 

 

 

    

 

 

   
    
  

 

 

124

400 -

(d) N added July 8

 

 

 

   
    

 

 

U
:I:
' H
an
-100
-100 I ' T 1 I I I I I I
0 so 100 150 200 0 50 100 150 200
400- (b) Nadded May 26 200‘ (e) Nadded July 26 o
300- ' 150-
an 200- m U 100-
l I
m”1m- m“ w-
0 ' ---------------- 0 ----------------
'100 I y r I I '50 I I I I I
0 so 100 150 200 0 50 B 100 150 200
P
400' (c) Nadded June 17 o
300-
U
an 200 1
. H
an 100-
0 ________________
~100 I I I I I
0 50 B 100 150 200
C

 

Figure 34. Absolute species biomass responses to 5 times of nitrogen addition
in a second-year old-field in southwestern Michigan (dashed line = null
Ypothesis; solid line =- fitted regression line).

125

 

5:020:00 50: 00050.08 0.
0:005:00. 00: n m: ..

 

 

 

 

00 .0 NA- 5000.0 00.0 :0 800.0 NN 0m 33 00000 z .> 35:00
-.0 0.0. 5000.0 00.0 00.0 800.0 0m 0 33 00000 z .> 35:00
00.0 0.0. 0000.0 ~00 00.0 0000.0 3 S 0:2. 00000 z .> 35:00
3.0 ~.0 000.0 00.0 m m .0 000.0 0N 0N 02 00000 z .> 35:00
00.0 0.7 N000 3.0 00.0 N000 “N a H 02 00000 z .> 35:00
£0030 :2000000: 805
0003000 003000000 05300
000 0.0. 5000.0 3.0 No.0 500.0 mm 0N E3 00000 z .> 35:00
0. 50.0 0.0- 5000.0 hm.“ N00 5000.0 3 0 baa 00000 z .> 35:00
0N0 N0- 5000.0 05.“ 00.0 300.0 0N S 0:3 00000 z .> 35:00
~00 0.0. 5000.0 H0.“ 00.0 300.0 HN mm 02 00000 Z .> 35:00
000 0.0- 5000.0 004 00.0 5000.0 NN H H 00E 00000 Z .> 35:00
£03050 :00000000: 5
000205 330000.00 .0500
50000005

:00 50000055 0020

0. A 0 4A .30 0 0000 N: a #0008 2 :03000800 30850000.

 

A a 0:0 0H 00.3000 :0 000000 .503 050:0000500 00,—. .0000 :35
35:85:08 $00.00 50000055-» :0 000B 05 005 050805 0000030 S 00205 0003050 :00000000:
05 :0 000205 0.0 5:0:5005 0:0 35:00 500 5 00.9500 505 006000 030 .200 03 0000:0000
05 5 300502 000 00:003.: 03333 520 35:00 0:0 E00000 0530.50 05 00 0085 30.0000

50 00000 :000530 3080005 :002000 000805 :008 E 00:05 05 0:0 60000 :00053 05

303 35:00 5 0000000 00 000803 :008 :002000 0000:0000: 05 :00 8000050 30000000 .2 030,—.

126
for a disproportionate share of community biomass. None of the species except

Q canadensis had a mean biomass greater than 100 g/m2. As a result, the data
were not bivariate normally distributed and the dominant species may have
had disproportionate influence on the regression.

When Q canadensis was excluded from the regression analyses, the
absolute biomass responses were dependent on the biomass of the species in
the control for each time of nitrogen addition (Figure 35). The slopes were
significantly positive and the y-intercepts did not differ from zero (Table 15).
The values for the slope and r2 were lower when Q canadensis was excluded
from the regression analyses than when it was included (Table 15),

illustrating the influence of the large positive response of the dominant

 

‘l— ‘1 v_uu.-ri - -—._—

species on the regressions.

Second-year old-field: Relative biomass responses

The relative biomass response was independent of the biomass of the
species in the control for each time of nitrogen addition when Q canadensis
was included in the regression analysis (Figure 36, Table 16) and when it was
excluded (Figure 37, Table 16). The magnitudes of the relative responses were
much smaller and less variable for species of biomass greater than 10 g/m2 in
the control than for species less than 10 g/m2. The positive relative responses
of the species of very low biomass were larger and more variable than the

negative relative responses.

DISCUSSION
Consequences of spatial variability of species distributions
The variability of species frequency and abundance on a 50 hectare

scale in the annual old-field study created several challenges for the analysis

127

Conyza madam excluded from the analysis

 

 

 

 

   

 

 

 

 

 

 

 

80 ‘ (a) N added May 11 . 30 ‘ (d) N added July 8 ,
60 d
U U
an an
an H in H
80 ‘ (b) N added May 26 40 - (e) N added July 26 .
0 U 9' U
1:: H an H
O 20 B 40 60 80
C
100 ' (c) N added June 17 0
U
:1
m H

 

 

I I
O 20 B 40 60 80
C

Figure 35. Absolute species biomass responses to 5 times of nitrogen addition
in a second-year Michigan old- field. The dominant species, Gama canadensis,
is excluded from the regression analysis (dashed line= null hypothesis; solid
line = fitted regression line).

 

128

 

 

 

 

 

20 (a) N added May 11 30 (d) N added July 8
1
m U 5 U 20
2 10 E
0 U A u 10
I S a;
H
an O "’ .___._ _________ 2
v 0 0....-__0_ ___________ a 0
'S I I I I I '10 l I I l 1
O 50 100 150 200 O 50 100 150 200 Ir-=
30 (b) N added May 26 200 (e) N added July 26
u 2 U
“i 0 an 100
\ \
a: U 10 m u
. - 0 .-——0 ————————— o _-'
00
5 H 0 0'3“"- --------- .- e H
'10 I I I I 1 '100 I I I I I
0 50 100 150 200 0 503 10 150 200
c
2000 'l (c) N added June 17
1500 -
U
an
\ 1000 +
an 500 0
' H

 

5 Or0---o- _________ .,
-500

I I I

I j
0 50 B 10 150 200
C

Figure 36. Relationship between the biomass of species in the control plots
and the relative change in mean biomass of species at 5 times of nitrogen
addition in a second-year Michigan old-field. (dashed line = null hypothesis;
solid line = fitted regression line).

 

129

0:005:00.” 00: 50050005 5
0:009:03 00: n m: a

 

 

 

 

0N0 H.0 00.0 N0. 50 00.0 NN 0N 530 00000 Z .> 35:00
5 00.0 0.~ 00.0 00.0. 00.0 00.0 0.5 0 530 00000 z .> 35:00
0N 5H 0:50 00000 z .> 35:00
5 00.0 0.0 00 .0 00.0. m0.0 Om .0 0m 0N 502 00000 z .> 35:00
5 00.0 v.~ 00.0 3.0. ~00 00.0 HN H H 502 00000 z .> 35:00
530:0 :50000000 5000
00020000 “$5000.00 08300
mmd 0.0 05.0 00.0. 500.0 05.0 3 0m 53a 00000 z .> 35:00
5 00.0 v.~ H50 8.0. 000.0 :0 HN 0 53a 00000 Z .> 35:00
0~.0 H.00 5.0 5.0. 000.0 :0 0N 5H 0:3 00000 z .> 35:00
5 00.0 50.0 00.0 H00. 8.0 00.0 HN 0m 502 00000 z .> 35:00
5 00.0 N.~ 05.0 000.0. 000.0 05.0 NN HH 502 00000 z .> 35:00
530:0 :50000000 5
000505 00280050 «and
00000005

00.5 00000005 00000

0. A 0 -5 00.0 0 00on N0 0 H0005 2 :05000500 5050000.

 

.mH 0:0 NH 0000000 5 000000 550 050:00000000 000. .0000 5000

50:85:00 0000000 00000005.» 00 0050 05 005 0000005 0000030 5 000H0> 000305 :50000000
05 5 000505 000 3050005 0:0 35:00 500 5 00005000 005 0000000 530 030-03 0005

0:800 05 5 3000002 000 .0w:00m00 0200300 35:00 0:0 300000 053000 05 00 00:05 50000000
00 00000 :00055 505005 :002500 0.6.0505 :005 5 00:000 03030 05 0:0 500000 5005: 0:0
503 35:00 5 0000000 00 0.00505 :02: :003000 05300500 005 00.0 00500000 :5800000 .0H 0300.

130

Gama madman excluded from the analysis

 

 

 

 

 

 

 

20 (a) N added May 1 1 30 (d) N added July 8
u 15
m u 20
Z 10 f
a u A 0 10
I 5 an
0 H ' .n
V O _._.1‘__-_________: a“ 0 '.-.---+ -------
I I '10 I I I I I
60 30 O 20 4O 6O 80
30 (b) N added May 26 20° 1 (e) N added July 26
u U 150
0 20 2
\ A 0 100 d
A U 10 Fe
m; u 50 ..
H
I-I C 0 III
a O -c--""1 """"" ' V 0T900----0- ------- -o
-50
‘10 f I I I I I I I l j
0 20 4O 6O 80 O 20 B 40 6O 80
c
2000 1 (c) N added June 17
1500 .
U
0
- \ 1000 -
"‘ u
0 500 -
I
an H 0 f+oo----e -------- 0
'500 I I I j
0 20 B 40 60 80

Figure 37. Relationship between the biomass of species in the control plots
and the relative change in mean biomass of species at 5 times of nitrogen
addition in a second-year Michigan old-field. The dominant species, Cbnyza
canadansis, is excluded from the regression analysis (dashed line = null
hypothesis; solid line = fitted regression line).

 

 

1 3 1
and interpretation. Several species did not occur in both the control and

nitrogen addition treatments when mean species biomass was calculated.
Several species were locally abundant at the one hectare scale, but not
frequent at the SO hectare scale (Setafia m and Ambrbsla angmisiifblla) and
so did not appear in all 4 replicates of the treatments. Several species were of
low abundance at the one hectare scale, and low frequency across replicates,
particularly the early colonizing perennials, such as Amum mm
and Ellen: alba. In general, the frequency and abundance of species was
positively correlated (Pearson r = 0.35 and p = 0.0001).

One consequence of species not appearing in all replicates of all
treatments is that the sample size, and thus the power of the regression
analysis, is reduced. Between 6 and 12 species were excluded from the
regression analysis because they were not present in both the control plots
and nitrogen addition treatments within a comparison. One way to increase
the degrees of freedom is to plot species responses in each replicate separately.
However, this could create pseudo-replication because each species would be
represented by more than one data point in the regression analysis.

A second consequence of some species appearing in only a subset of
replicates is that the mean biomass response may be a poor estimate of the
performance of the species. The behavior of Selma [alga provides a good
example of a species that occurred infrequently on a 50 hectare scale, but was
locally abundant on a one hectare scale, and displayed a large negative
response to the nitrogen additions. Its response may have been the result of
sampling bias in either the control or treatment plots. Consequently, Selma
[aberi was an anomalous data point in Figures 29b and 29c, and may have

exerted disproportionate influence on the fitted regression lines.

 

132
Responses of the species biomass hierarchy to nitrogen additions

Each of the correlations between rank of species abundance in control
and rank of species abundance in the nitrogen addition treatment was
significant indicating that the general order of species abundance remained
constant within a growing season when nitrogen was added to the annual
community. Thus, we can accept the first null hypothesis that, in general,
there was no change in the species biomass hierarchy with any of the times of
nitrogen addition. Mitchley and Grubb (1986) also found a constancy of rank
order of plant species abundances across a site, between years, and between
sites in a chalk grassland.

The second null hypothesis addressed whether the direction and
magnitude of the species biomass response to nitrogen addition depended on
the biomass of the species in the control plots. The slope of the relationship
between the absolute and relative responses of species biomass depended on
the time of nitrogen addition and whether Q album was present or removed
from the community. When nitrogen was added early in the season there was
at least a tendency for species of greater biomass in the control plots to have
larger absolute responses to the nitrogen addition whether Q album was
present or not (Figures 29a and 30a). The absolute and relative biomass
responses to the early nitrogen addition suggest that a competitive hierarchy
was not well established early in the season (Miller and Werner 1987). Species
of all abundances could increase biomass (in an absolute and relative sense) in
response to the early season nitrogen addition.

When nitrogen was added later in the season (late July and early
August), subordinate species of higher biomass in the control plots had a
larger negative absolute response to the nitrogen addition than species of

lower biomass. The slope of the relationship between absolute and relative

 

133
species response and species abundance in the control plots was significantly

negative. Presumably, this was a result of the stronger competitive effect
from Q album when nitrogen was added at mid—season. The dominant species,
Q album, doubled in biomass and reduced species diversity and evenness most
with the mid-season nitrogen addition (Chapter 1). When Q album was
removed experimentally, the absolute responses of the remaining species to
the mid-season nitrogen addition were generally positive and the magnitude
of the absolute response increased significantly with species biomass in the
control plots (Figure 30b). The relative responses were also generally positive
(Figure 33b), but the slope was not significantly different from zero.

In contrast to the annual community, the slopes for absolute responses
to nitrogen addition in the second-year old-field were always significantly
positive (Figure 35), and the slopes for the relative responses were never

significantly different from zero (Figure 37).

Absolute and relative responses

The magnitude of absolute responses tended to be larger among species
of high mean biomass in the control plots than species of low biomass. The
magnitude of relative responses tended to be higher among species of low
biomass in the control plots than species of high biomass. This is consistent
with the statement by Goldberg (1994) that the measure of relative change in
species abundance is more sensitive to changes in rare species than is the

measure of absolute change in abundance.

Analogies to other approaches
This graphical method is analogous to statistical and graphical
approaches in population biology that analyze genetic covariation between

performance of genotypes in different environments (Via and Lande 1985,

134
Shaw and Platenkamp 1993). These population methods compare the

performance of genotypes in different environments to determine whether
the direction and magnitude of response to environmental change is the same
among genotypes. In this analysis of species responses in a community, the
mean species biomass response is analogous to the genotype response in
population genetics analyses. The species response surfaces in this study are
the result of the ability of the species to respond to the nitrogen additions and
the changes in competitive interactions when nitrogen is added.

In community ecology, Pinder (1975) used a similar graphical method to

illustrate the response of net productivity of subordinate forb species as a

 

group to the removal of a dominant grass species, but he did not extend his I
graphical approach to evaluate species level responses. Silander and
Antonovics (1982) used perturbation response diagrams to examine the
responses of species abundances to the removal of a dominant species. The size
of the circles in the diagrams were proportional to the abundance of species in
the control and species removal treatments. However, it is difficult to discern
from the perturbation response diagrams in Silander and Antonovics (1982)
whether the species responses were related to their abundance in the control
plots.

This approach of looking at the relationship between species
performance in two environments is also analogous to the ecosystem level
approach used by Hendrickson et. a1 (1988 a, b) to quantify changes in the
water chemistry of lakes between 1974/ 75 and 1986 in the Thousand Lake
Survey in Norway. They plotted water chemistry parameters of the lakes in
1974/ 75 (pH, nitrate, sulfate, aluminum) against the parameters in 1986,

looking for responses to changes in industrial emissions.

CHAPTER 5

SUMMARY AND CONCLUSIONS

135

 

136

The species removal and nitrogen addition experiment in the annual
old-field demonstrated that competition intensity from the dominant species
varied with the time of nitrogen addition. The dominant species reduced
subordinate community biomass, evenness, and diversity most when nitrogen
was added mid-season (late July). When nitrogen was added early in the
growing season, both the dominant species and subordinate community
increased tissue nitrogen concentration, growth and total nitrogen content.
When nitrogen was added later in the season (late July or early August), the
dominant species increased tissue nitrogen concentration, growth, and total
nitrogen content. The subordinate community increased tissue nitrogen
concentration, but not growth and total nitrogen content. The degree of
canopy development appears to determine the ability of dominant and
subordinate species to acquire added nutrients and allocate them to increased
growth rather than storage, and thus, influences the magnitude of the effect
of the dominant species on diversity and community structure.

This pattern of response to nitrogen addition may be typical of habitats
where plant productivity is high, where a distinct canopy, sub-canopy, and
ground-layer can develop, and light limitation increases during the growing
season. The patterns of response to the time of nitrogen addition may differ
substantially in a habitat with lower fertility and productivity and shorter
vegetation or where a distinct canopy, sub-canopy, and ground layer do not
develop.

The extension of these results to the effects of natural patterns of

nutrient pulses must be done with care. The amount of the experimental

137
nitrogen pulse was very large, equivalent to the amount of nitrogen

mineralized during the entire growing season in these soils. The nutrient
pulses that occur in spring when microbial activity increases following snow
melt (“vernal dam”) and during the growing season following rainfall events
are much smaller in magnitude than the nitrogen pulses described in Chapter
2 and 3. We need more information about the consequences of natural
temporal variation in nutrient availability for competitive interactions and
community structure in terrestrial plant communities. However, I would
predict that in the annual community, the subordinate community would be
able to acquire and allocate nitrogen made available by microbial activity
early in the season better than nitrogen mineralized later in the season.

The decline in biomass evenness during the growing season in the
second-year old-field study demonstrated the development of a biomass
hierarchy. While there were seasonal changes in species composition and
abundance, there was a clear biomass dominance by Cgmyza canadensis. This
dominant species increased biomass disproportionately to the subordinate
community in response to each time nitrogen addition. Although biomass
evenness of the community declined with each time nitrogen addition, the
magnitude of the effect did not vary with time of nitrogen addition as it did in
the annual old-field. Rather, life-history/ growth form groups and individual
species in the subordinate community varied in their response to the time of
nitrogen addition. These individual responses appeared to compensate for
each other and thus, evenness and diversity did not vary with the time of
nitrogen addition in the second—year old-field. The variation in growth
phenologies of component species is greater in the second-year old-field with
winter annuals, biennials, and perennials than in the annual old field

consisting primarily of summer annuals.

 

138
The second-year old-field represents a transition between dominance

by annual species in the first year following abandonment from agriculture
and dominance by perennial species three to five years following
abandonment. Many of the ultimate dominant herbaceous species, such as
Sblldagg canadensis and Aster species colonize fields at this stage of succession.
Understanding the mechanisms of plant competition in this transitional
community is crucial to understanding how annual species are lost and
perennial species come to dominate as succession proceeds in this region. The
experiment in the second-year old-field demonstrated the potential for
variation in competitive interactions during the grong season. Plants that
establish or grow at different times of the growing season will likely be
confronted with different amounts and species of competitors and with
different availabilities of above- and below-ground resources. This may
influence their success in persisting for longer periods in these successional
communities.

The seasonal patterns of resource availability demonstrated that
nitrogen addition to the second-year old-field led to reduced light penetration
to the soil surface. At mid-season light penetration was lower and soil
nitrogen availability was higher in the nitrogen addition treatments than the
control. However, by August light penetration and nitrogen availability were
at the same low levels in the control and nitrogen addition plots. Few field
studies have documented seasonal patterns of resource availability in
terrestrial plant communities. The effects of competitors and experimental
treatments on resource availability are usually measured only at the end of the
experiment. In the second-year old-field, the effect of nitrogen additions on
resource availability was apparent early and mid-season but not at the end of

the season. Monitoring seasonal patterns of resource availability will be

 

139
important for understanding which resources limit plant productivity and for

untangling the mechanisms of plant competition.

The graphical and statistical method demonstrated that the slope of the
relationship between the absolute and relative responses of species to
nitrogen addition in the annual old-field and the biomass of the species in the
control plots depended on the time of nitrogen addition. The slope was not
significant when nitrogen was added early in the season in the annual old-
field, when the hierarchy among species was not established and species of all
abundances in the control plots increased biomass. However, the slope was
significant and negative when nitrogen was added mid- or late season. Species
of higher abundance in the control plots decreased in biomass and species of
low abundance increased in biomass. The negative slope was the result of
competition from the dominant species; the slope was positive when Q album
was removed from the community and nitrogen was added mid-season. In
contrast to the annual old-field, the slope of the relationship between absolute
biomass response to nitrogen addition and the biomass of species in the control
plots was always positive. This result suggests that the effect of Conyza
canadansls and Chengbbdium album on the biomass hierarchy differed. In
both studies, the absolute response was larger for species of high abundance
in control plots than for species of low abundance. The relative response was
larger for species of low abundance than for species of high abundance. This
pattern suggests that measures of absolute response are more sensitive
changes in rare species, and measures of relative response are more sensitive

to changes in abundant or dominant species.

LIST OF REFERENCES

 

LIST OF REFERENCES

Abul-Fatih, H. A. and F. A. Bazzaz. 1979. The biology of Ambrbsia mfida L. 1.
Influence of species removal on the organization of the plant
community. New Phytologist 83:8 1 3-8 1 6.

Allen, E. B. and R. T. T. Forman. 1976. Plant species removals and old-field
community structure and stability. Ecology 57:1233-1243.

Armesto, J. J. and S. T. A. Pickett. 1985. Experiments on disturbance in old-field
plant communities: impact on species richness and abundance. Ecology
66:230-240.

Armesto, J. J. and S. T. A. Pickett. 1986. Removal experiments to test mechanisms
of plant succession in old-fields. Vegetatio 66:85-93.

Bakelaar, R. G. and E. P. Odum. 1978. Community and population level responses
to fertilization in an old-field ecosystem. Ecology 59:660-665.

Bazzaz, F. A. 1990. Plant-plant interactions in successional environments. pp.
240-263 In Perspectives on Plant Competition (Eds. J. B. Grace and D.
Tilman) Academic Press, San Diego.

Benner, B. L. and F. A. Bazzaz. 1985. Response of the annual Abutllgn
111mm Medic. (Malvaceae) to timing of nutrient availability.
American Journal of Botany 72:320-323.

Benner, B. L. and F. A. Bazzaz. 1987. Effects of timing of nutrient addition on
competition within and between two annual plant species. Journal of
Ecology. 75:229-245.

Benner, B. L. and F. A. Bazzaz. 1988. Carbon and mineral element accumulation
and allocation in two annual plant species in response to timing
nutrient addition. Journal of Ecology 76:19-40.

Bertness, M. D. 1991. Zonation of Spaglna pagans and Spartan mamjflgna in a
New England salt marsh. Ecology 72:138-148.

Caldwell, M. M. and R. W. Pearcy. 1994. Exploitation of environmental

heterogeneity by plants: Ecophysiological processes above- and below-
ground. Academic Press, San Diego, USA.

140

 

141

Campbell, B. D. and Grime, J. P. 1989. A comparative study of plant
responsiveness to the duration of episodes of mineral nutrient
enrichment. New Phytologist 1 12:261-267.

Campbell, B. D. and J. P. Grime. 1992. An experimental test of plant strategy
theory. Ecology 73:15-29.

Campbell, B. D., J. P. Grime, and J. M. L. Mackey. 1991a. A trade-off between
scale and precision in resource foraging. Oecologia 87:532-538.

Campbell, B. D., J. P. Grime, and J. M. L. Mackey, and P. Jalili. 1991b. The quest
for a mechanistic understanding of resource competition in plant
communities: the role of experiments. Functional Ecology 5:241-253.

Carson, W. P. and G. W. Barrett. 1988. Succession in old-field plant communities:
effects of contrasting types of nutrient enrichment. Ecology 69:984-994.

Carson, W. P. and S. T. A. Pickett. 1990. Role of resources and disturbance in the
organization of an old-field plant community. Ecology 71:226-238.

Chapin, F. S., III. 1980. The mineral nutrition of wild plants. Annual Review of
Ecology and Systematics 1 1:233-260.

Chapin, F. S., III, and G. R. Shaver. 1985. Individualistic growth response of
tundra plant species to environmental manipulations in the field.
Ecology 66:564-576.

Connell, J. H. 1961. The influence of interspecific competition and other
factors on the distribution of the barnacle Chmamalus stellatus. Ecology
42:710-723.

Davy, A. J. and K. Taylor. 1974. Seasonal patterns of nitrogen availability in
contrasting soils in the Chiltern Hills. Journal of Ecology 62:793-807.

Davy, A. J. and K. Taylor. 1975. Seasonal changes in the inorganic nutrient
concentrations in Dgachambsla cagspltgsa (L) Beauv. in relation to its
tolerance of contrasting soils in the Chiltem Hills. Journal of Ecology
63:27-39.

Dayton, P. K. 1975. Experimental evaluation of ecological dominance in a rocky
intertidal algal community. Ecological Monographs 45:137-159.

DiTommaso, A. and L.W. Aarssen. 1989. Resource manipulations in natural
vegetation: a review. Vegetatio 84:9-29.

Gaudet, CL. and RA. Keddy. 1988. A comparative approach to predicting
competitive ability from plant traits. Nature 334:242-243.

Goldberg, D. E. 1990. Components of resource competition in plant communities.
Pages 27 - 49 in J. B. Grace and D. Tilman, editors. Perspectives on plant
competition. Academic Press, New York, New York, USA.

 

142

Goldberg, D. E. and A. M. Barton. 1992. Patterns and consequences of
interspecific competition in natural communities: A review of field
experiments with plants. American Naturalist 139:771-801.

Goldberg, D. E. and T. E. Miller. 1990. Effects of different resource additions on
species diversity in an annual plant community. Ecology 71:213-225.

Goldberg, D. E. and P. A. Werner. 1983. Equivalence of competitors in plant
communities, a null hypothesis and a field experimental approach.
American Journal of Botany 70:1098-1 104.

Goldberg, D. E. 1994. Influence of competition at the community level: An
experimental version of the null models approach. Ecology 75:1503-1506.

Grace, J. B. 1990. On the relationship between plant traits and competitive
ability. Pages 51-65 in J. Grace and D. Tilman, editors. Perspectives on
plant competition. Academic Press, New York, New York, USA.

Grace, J. B. 1993. The effects of habitat productivity on competition intensity.
Trends in Ecology and Evolution 8:229-230.

Grime, J. P. 1973. Competitive exclusion in herbaceous vegetation. Nature
242:344-347.

Grime, J. P. 1979. Plant Strategies and Vegetation Processes. John Wiley and
Sons, Chichester.

Grime, J. P. 1987. Dominant and subordinate components of plant communities:
Implications for succession, stability, and diversity. Pages. 413-428 in A.
J. Gray, M. J. Crawley, and P. J. Edwards, editors. Colonization, succession,
and stability. Blackwell, Oxford.

Grime, J. P. 1994. The role of plasticity in exploiting environmental
heterogeneity. Pages 1-19 in M. M. Caldwell and R. W. Pearcy, editors.
Exploitation of environmental heterogeneity by plants:
Ecophysiological processes above- and below ground. Academic Press,
New York, New York, USA.

Grubb, P. J. 1985. Plant populations and vegetation in relation to habitat
disturbance and competition: problems of generalization. Pages 595-621
in J. White, editor. The population structure of vegetation. Dr. W. Junk,
The Hague, The Netherlands.

Gupta, P. L. and l. H. Rorison. 1975. Seasonal differences in the availability of
nutrients down a podzolic profile. Journal of Ecology 63:521-534.

Gurevitch, J. and R. S. Unnasch. 1989. Experimental removal of a dominant
species at two levels of soil fertility. Canadian Journal of Botany 67:3470-
3477.

Gurevitch, J., P. Wilson, J. L. Stone, P. Teese, and R. J. Stoutenburgh. 1990.
Competition among old-field perennials at different levels of soil
fertility and available space. Journal of Ecology 78:727—744.

 

143

Henriksen, A. and D. F. Brakke. 1988. Increasing contribution of nitrogen to
the acidity of surface waters in Norway. Water, Air, and Soil Pollution
42:183-201.

Henriksen, A., L. Lien, F. S. Traaen, I. S. Sevaldrud and D. F. Brakke. 1988. Lake
acidification in Norway - present and predicted chemical status. Ambio
17:259-266.

Heske, E. J., J. H. Brown, and Q, Guo. 1993. Effects of kangaroo rat exclusion on
vegetation structure and plant species diversity in the Chihuahuan
Desert. Oecologia 95:520-524.

Hils, M. H. and J. L. Vankat. 1982. Species removals from a first-year old-field
plant community. Ecology 63:705-71 1.

Huston, M. A. 1979. A general hypothesis of species diversity. American
Naturalist 113:81-101.

Hyder, M. B. and G. W. Barrett. 1986. Effects of nutrient enrichment on the
producer trophic level of a six-year old-field community. Ohio Journal
of Science 86:10-14.

Inouye, R. S., N. J. Huntly, D. Tilman, J. R. Tester. 1987. Pocket gophers (ermys
Mus), vegetation, and soil nitrogen along a successional sere in
east central Minnesota. Oecologia 72:1 78-1 84.

Keddy, P. A. 1989a. Competition. Chapman and Hall, London, England.

Keddy, P. A. 1989b. Effects of competition from shrubs on herbaceous wetland
plants: a 4-year field experiment. Canadian Journal of Botany 67:708-
716.

Keddy, P. A. 1990. Competitive hierarchies and centrifugal organization in
plant communities. Pages 266-290 in J. B. Grace and D. Tilman, editors.
Perspectives on plant competition. Academic Press, New York, New
York, USA.

Keddy, P. A. and B. Shipley. 1989. Competitive hierarchies in herbaceous plant
communities. Oikos 54:234-241.

Lubchenco, J. 1980 Algal zonation in the New England rocky intertidal
community: an experimental analysis. Ecology 61:333-334.

Magurran, A. 1988. Ecological Diversity and Its Measurement. Princeton
University Press, Princeton, NJ.

Mellinger, M. V. and S. J. McNaughton. 1975. Structure and function of
successional vascular plant communities in Central New York.
Ecological Monographs 45:161-182.

Mengel, K. and E. A. Kirkby. 1978. Principles of plant nutrition. International
Potash Institute, Berne, Switzerland.

 

144

Miller, T. E. 1987. Effects of emergence time on survival and growth in an early
old-field plant community. Oecologia 72:272-278.

Miller, TE. and RA. Werner. 1987. Competitive effects and responses between
plant species in a first-year old-field community. Ecology 68:1201-12 10.

Mitchley, J. and P. J. Grubb. 1986. Control of relative abundance of perennials
in chalk grassland in Southern England I. Constancy of rank order and
results of pot- and field-experiments on the role of interference.
Journal of Ecology 74:1 139-1 166.

Moore, D. R. J., and P. A. Keddy. 1989. The relationship between species
richness and standing crop in wetlands: the importance of scale.
Vegetatio 79:99-106.

Naeem, S. and R. K. Colwell. 1991. Ecological consequences of heterogeneity of
consumable resources. pp. 224-255 In Ecological Heterogeneity (Eds. J.
Kolasa and S. T. A. Pickett) Springer-Verlag, New York.

 

Newman, E. I. 1973. Competition and diversity in herbaceous vegetation. Nature
244:310-31 1.

Newman, E. I. 1983. Interactions between plants. Pages 679-710 in O. L. Lange,
P. S. Nobel, C. B. Osmond, and H. Ziegler, editors. Physiological plant
ecology. 111. Responses to the chemical and biological environment.
Springer-Verlag, Berlin, Germany.

Paine, R. T. 1966. Food web complexity and species diversity. American
Naturalist 100:65-75.

Pastor, J ., M. A. Stillwell, and D. Tilman. Nitrogen mineralization and
nitrification in four Minnesota old fields. Oecologia 71:481-485.

Pate, J. S. 1983. Patterns of nitrogen metabolism in higher plants and their
ecological significance. Pages 225-255 in J. A. Lee, S. McNeill, I. H.
Rorison, editors. Nitrogen as an ecological factor. Blackwell, Oxford.

Pinder, J. E., III. 1975. Effects of species removal on an old-field plant
community. Ecology 56:747-751.

Reader, R. J. and B. J. Best. 1989. Variation in competition along an
environmental gradient: Hieraclum flmjlbundum in an abandoned
pasture. Journal of Ecology 77:673-684.

Reader, R. J. and 19 other authors. 1994. Plant competition in relation to
neighbor biomass: An intercontinental study with Pba matenals.
Ecology 75:1753-1760.

Reed, F. C. P., 111. 1977. Plant species number, biomass accumulation, and
productivity of a differentially fertilized Michigan old-field. Oecologia
30:43-53.

145

Robertson, G. P. and P. M. Vitousek. 1981. Nitrification potentials in primary
and secondary succession. Ecology 62:376-3 86.

Samson, D. A., T. E. Philippi and D. W. Davidson. 1992. Granivory and
competition as determinants of annual plant diversity in the
Chihuahuan desert. Oikos 65:61-80.

SAS. 1989. SAS/STAT user's guide, Version 6, Fourth edition. SAS Institute , Cary,
North Carolina, USA.

Shaw, R. G. and G. A. J. Platenkamp. 1993. Quantitative genetics of response to
competitors in Namgbmla menzlesil: A greenhouse study. Evolution
47:801-812.

Silander, J. A. and J. Antonovics. 1982. Analysis of interspecific interactions in
a coastal plant community - a perturbation approach. Nature 298:557-
560.

Silvertown, J. 1980. The dynamics of a grassland ecosystem: botanical
equilibrium in the park grass experiment. Journal of Applied Ecology
17:491-504.

Southwood, T. R. E. 1988. Tactics, strategies, and templets. Oikos 52:3-18.

Swieringa, S. and R. E. Wilson. 1972. Phenodynamic analyses of two first-year
old fields. American Journal of Botany 59:367-372.

Taylor, D. R., L. W. Aarssen, and C. Loehle. 1990. On the relationship between
r/K selection and environmental carrying capacity: a new templet for
life history strategies. Oikos 58:239-250.

Taylor, A. A., J. DeFelice, and D. C. Havill. 1982. Seasonal variation in nitrogen
availability and utilization in an acidic and calcareous soil. New
Phytologist 92:141-152.

Tilman, D. 1982. Resource competition and community structure. Princeton
University Press, Princeton, New Jersey.

Tilman, G. D. 1984. Plant dominance along and experimental nutrient gradient.
Ecology 65:1445-1453.

Tilman, D. 1987. Secondary succession and the pattern of plant dominance
along experimental nitrogen gradients. Ecological Monographs 57:189-
214.

Tilman, D. 1993. Species richness of experimental productivity gradients: How
important is colonization limitation? Ecology 74:2179-2191.

Tilman, D. 1988. Dynamics and structure of plant communities. Princeton
University Press, Princeton, New Jersey.

 

146

Tilman, D. and S. Pacala. 1993. The maintenance of species richness in plant
communities. pp. 13-25 In Species diversity in ecological communities:
historical and geographical perspectives (Eds. R. E. Ricklefs and D.
Schluter) University of Chicago Press, Chicago, USA.

van der Maarel, E. 1981. Fluctuations in a coastal dune grassland due to
fluctuations in rainfall: experimental evidence. Vegetatio 47:259-265.

Via, S. and R. Lande. 1985. Genotype-environment interaction and the
evolution of phenotypic plasticity. Evolution 39:505-522.

Wakefield, N. G. and G. W. Barrett. 1979. Effects of positive and negative
nitrogen perturbations on an old-field ecosystem. American Midland
Naturalist 101:159-169.

Watson, M. A. 1984. Developmental constraints: effect on population growth
and patterns of resource allocation in a clonal plant. American
Naturalist 123:41 1-426.

Weed Science Society of America. 1984. Composite list of weeds. Weed Science
32(Supplement 2).

Weiner, J. 1985. Size hierarchies in experimental populations of annual plants.
Ecology 66:743-752.

Weiner, J. 1986. How competition for light and nutrients affects size variability
in leera 1121:5219: populations. Ecology 67:142 5-142 7.

Weiner, J. 1990. Asymmetric competition in plant populations. Trends in
Ecology and Evolution 5:3 60-3 64.

Weiner, J. and O. T. Solbrig. 1984. The meaning and measurement of size
hierarchies in plant populations. Oecologia 61:3 34—336.

Weiner, J. and S. C. Thomas. 1986. Size variability and competition in plant
monocultures. Oikos 47:21 1-222.

Whiteside, E.P., l.F. Schneider, and R.L. Cook. 1959. Soils of Michigan. Michigan
State University Agricultural Experiment Station Special Bulletin 402,
East Lansing, Michigan, USA.

Willis, A. J. 1963. Braunton Burrows: the effects on the vegetation of the
addition of mineral nutrients to the dune soils. Journal of Ecology
5 1:3 5 3-3 74.

Wilson, J. B. 1988. The effect of initial advantage on the course of plant
competition. Oikos 51:19-24.

Wilson, 5. D. 1991. Variation in competition in eucalypt forests: the importance
of standardization in pattern analysis. Journal of Vegetation Science.
2:577—586.

147

Wilson, S. D. and P. A. Keddy. 1986. Measuring diffuse competition along an
environmental gradient: results from a shoreline community.
American Naturalist 127:862-869.

Wilson, 8. D. and P. A. Keddy. 1991. Competition, survivorship and growth in
macrophyte communities. Freshwater Biology 25:33 1-337.

Wilson, S. D. and J. M. Shay. 1990. Competition, fire, and nutrients in a mixed-
grass prairie. Ecology 71:1959-1967.

Wilson, 5. D. and D. Tilman. 1991a. Components of plant competition along a
productivity gradient. Ecology 72:1050-1065.

Wilson, S. D. and D. Tilman. 1991b. Interactive effects of fertilization and
disturbance on community structure and resource availability in an
old-field plant community. Oecologia 88:61-71.

Wilson, S. D. and D. Tilman. 1993. Plant competition and resource availability in
response to disturbance and fertilization. Ecology 74:599-61 1.

 

TE N V. LIBRRRIES

I

01LL06104

U

nrcuzcnu STA
llHlllllllWllWl"
31293