 

.I- I“ . .09“ «I; I
WLIWMMLMUIN. .
m.nm%uh

1’". “(arena
3... C .

. I k (.3... D..-

. Hunt]:

. ”paw? Wan...

a . .

3.. msgsé.-. .
“ﬁat-Lt!“ 1|.f .

I. it‘d-v- "1. '5... :3!va

1.1L.

l
.6!
n

.I. I 1!.
. .L. 2.5:;
my»? .2...

.11.}?
asiuxxv

ill
AI :1:
I... ‘6viv: t )3”
.l l . II...
P: 3'1!

 

5
. LIBRARY
2009) Michigan State
niversity

This is to certify that the
dissertation entitled

WHOLE-PLANT RESOURCE ECONOMIES AND ASSOCIATED
MORPHOLOGICAL AND PHYSIOLOGICAL TRAITS: TOWARDS A
MECHANISTIC UNDERSTANDING OF PLANT RESPONSES TO
RESOURCE VARIATION

presented by

Justin Michael Kunkle

has been accepted towards fulfillment
of the requirements for the

 

 

 

Doctoral degree in Forestry
/ ’1‘—
EW naurek/
L ”4 - (’29?
’- ate

MSU is an afﬁnnative-action, equal-opportunity employer

1-.-I-n--A-h-I-l-l-I-l_.-.-l-.-

O.-----a-n-c-o-a-I-n-n-n-l-.L-g_-_l_-

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

 

DATE DUE DATE DUE DATE DUE

h®$éwh

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProi/Acc&Pres/ClRC/DateDue indd

 

WHOLE-PLANT RESOURCE ECONOMIES AND ASSOCIATED
MORPHOLOGICAL AND PHYSIOLOGICAL TRAITS: TOWARDS A
MECHANISTIC UNDERSTANDING OF PLANT RESPONSES TO RESOURCE
VARIATION
By

Justin Michael Kunkle

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Forestry

2008

ABSTRACT
WHOLE-PLANT RESOURCE ECONOMIES AND ASSOCIATED
MORPHOLOGICAL AND PHYIOLOGICAL TRAITS: TOWARDS A
MECHANISTIC UNDERSTANDING OF PLANT RESPONSES TO RESOURCE
GRADIENTS
By
Justin Michael Kunkle

Differences in plant resource economies (i.e., resource-use efﬁciency, resource
access and storage capacity) and related plant traits may underlie species speciﬁc
variation in growth and survival across resource gradients. In my dissertation, I combine
potted plant and ﬁeld studies to explore functional traits as the basis of these mechanisms
and how they relate to variation in whole-plant performance over resource and
disturbance gradients.

First, with respect to soil N availability, I investigated how ﬁne root dimensions
and N concentrations vary during senescence for N fertilized and unfertilized seedlings of
species that differ in soil N affinity (Populus tremuloides Michx, Acer rubrum L., A cer
saccharum Marsh. and Betula alleghaniensis Britton). Senescence-related decreases in
root mass per length and root length indicated substantial root mass loss among species.
Mass-based root N increased from live to dead roots 10-3 5% among species, whereas N
decreased in dead roots when values were corrected by changes in mass (-12 to -28%).
My data along with re-analyzed values from the literature suggest that N resorption may

occur in ﬁne roots, which would lead to increased whole-plant N use-efﬁciency.

Second, I quantiﬁed interrelations of whole-plant total non-structural
carbohydrates (TNpr), relative growth rates (RGR) and associated functional traits for

seedlings of 36 temperate and boreal species grown in a common low-light environment.

Across species, plant traits related to surface area for above- and below-ground resource

capture were strongly related to RGR, whereas proportional allocation to root mass was
the strongest predictor of TNpr, Although RGR and TNpr were negatively
correlated, when RGR was normalized for plant mass effects, RGR was weakly, but

positively related to TNpr, Furthermore. independent of plant mass, carbon

conservation traits were positively related to RGR and TNpr, In contrast to previous
research, my ﬁndings suggest that in low light environments, independent of mass
effects, traits that increase growth also increase TNpr.

Third, I examined the relationship of plant traits to tolerance (i.e., survival) of

water deﬁcits using a conceptual framework that classiﬁed traits into water-use efﬁciency

(WUE) and water access (Waccess) categories. Seedlings of eight tree species differing in

soil resource afﬁnity were transplanted across glacial landforms with differences in water

holding capacity. I found that both the ability maintain positive photosynthetic rates (i.e.,

Waccess) and high photosynthesis per unit water loss (i.e.,WUE) during drought enhanced

seedling survival. Across species. increased Waccess was realized via deeper rooting
which was positively related to seed and seedling Size. Interspeciﬁc variation in WUE
was positively related to area-based leaf N (leaf Nam). Thus, differential expression of

these traits may partly underlie interspecifc differences in growth and survival responses,
which likely contribute to the observed species distribution patterns across glacial

landforms in northwestern Michigan.

To my parents, brother and Miss Heidi F rei for their enduring support and encouragement

ACKNOWLEDGEMENTS

First of all, I would like to thank my parents for the incredible guidance and
advice that they have provided me throughout the years. Both of my parents worked
extremely hard and sacriﬁced many things to see me through my college years at
Franklin and Marshall and I will always be grateful. In addition, my parents and brother
have always encouraged and supported me in every aspect of my life, and without that I
would never be where I am today. My family possesses an amazing work ethic and this
is something that has been passed down through the generations and this mindset has
enabled me to persevere through many of life’s challenges, including my PhD program. I
also would like to thank the rest of the Kunkle-Kratzer family for all of the letters of
support and care packages that helped me stay motivated during the last few weeks
leading up to my defense.

I cannot thank Miss Heidi Frei enough for her love and support, especially during
the last few months leading up to my dissertation defense. Heidi is truly one of the most
selﬂess people I know and she sacriﬁced so much to help me achieve my goals. She also
introduced me to long distance running and its many rewards. In addition, she Showed
me that there is so much more to life than scientiﬁc research and I look forward to the
future as we build a life together.

My advisor, Mike Walters has been a source of inspiration and knowledge that
has helped me tremendously in my development as a scientist. His work ethic and
passion in the ﬁeld are truly contagious. Over the years we have spent a lot of time

together in the laboratory and ﬁeld and I hope that he enjoyed this time as much as I did.

Besides providing me with substantial ﬁnancial support and guidance, he also allowed
me the freedom to pursue many of my research interests.

My committee members, Mike Walters, David Rothstein, Rich Kobe and Bert
Cregg provided guidance and scientiﬁc expertise that greatly improved my dissertation
project. David was kind enough to adopt me as an honorary member of the Rothstein lab
and provided me with full access to lab supplies and the C/N elemental analyzer. He also
took the time to talk to me about a post-doctoral position that I was offered on the Big
Island of Hawaii. I am especially grateful that all of my committee members were all
willing to read my dissertation with less than a week before my defense date.

I would also like to thank my undergraduate mentor, Dr. Timothy Sipe for
introducing me to the many aspects of ecological research in forested ecosystems. He
provided me with a solid foundation in how to design and carry out research projects,
which prepared me well for all of the challenges of a PhD project. Before working with
Tim at the Harvard Forest, I didn’t know how I could turn my interest in nature into a
career. His dedication to students and teaching is remarkable and Franklin and Marshall
College is very fortunate to have him as a faculty member.

And ﬁnally, I wanted to thank all of the organizations that provided generous
ﬁnancial support for my dissertation research. Michigan State University helped initiate
my work with a Plant Science Fellowship, and continued their support with a Dissertation
Completion Fellowship and supplemental funding through the Department of Forestry.
The National Science Foundation supported me as a graduate research assistant through a

research grant awarded to Mike Walters and Rich Kobe. NSF also supported me through

vi

the GK-12 fellowship program. The Hanes Fund of the Michigan Botanical Club

provided funding that helped me to expand the scope of my dissertation research.

vii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES ........................................................................................................ xvii
INTRODUCTION .............................................................................................................. 1
CHAPTER 1

MEASUREMENT BASES FOR FINE ROOT N CONCENTRATIONS:
IMPLICATIONS FOR SENESCENCE-RELATED N LOSS IN TEMPERATE TREE

SEEDLIN GS ....................................................................................................................... 6
Abstract ................................................................................................................... 6
Introduction ............................................................................................................. 7
Materials and Methods ............................................................................................ 9
Results ................................................................................................................... 15
Discussion ............................................................................................................. 18

CHAPTER 2

PLANT TRAIT CORRELATES OF WHOLE-PLANT CARBOHYDRATE STORAGE
AND RELATIVE GROWTH RATE IN TEMPERATE AND BOREAL TREE

SEEDLINGS: ARE THERE TRADEOFF S? ................................................................... 32
Abstract ................................................................................................................. 32
Introduction ........................................................................................................... 34
Materials and Methods .......................................................................................... 38
Results ................................................................................................................... 49
Discussion ............................................................................................................. 54

CHAPTER 3

ASSOCIATION OF MORPHOLOGICAL AND PHYSIOLOGICAL TRAITS WITH

NORTHERN TEMPERATE TREE SPECIES LANDFORM AFFINITY ...................... 74
Abstract ................................................................................................................. 74
Introduction ........................................................................................................... 76
Materials and Methods .......................................................................................... 81
Results ................................................................................................................... 95
Discussion ........................................................................................................... 104

APPENDIX ..................................................................................................................... 152

REFERENCES ............................................................................................................... 205

viii

LIST OF TABLES

Table 1.1. Results from ANOVA and summary of ﬁne root nitrogen (N)
concentrations ................................................................................................................... 24

Table 1.2. Results from ANOVA and summary of changes in ﬁne root N (AN) during
senescence ......................................................................................................................... 25

Table 1.3. Root dimensional changes, corrected change in ﬁne root N (AN) during
senescence and root Ca concentrations ............................................................................. 26

Table 1.4. Live and dead ﬁne root N concentrations, mass-based changes in ﬁne root N

(ANmass) and changes in ﬁne root N corrected for mass loss (ANconected) for data
gathered from published studies (adapted from Gordon and Jackson, 2000) ................... 27

Table 2.1. Summary of seedling characteristics for angiosperm and gymnosperm
species ............................................................................................................................... 62

Table 2.2. Correlation matrices for germinant mass, ﬁnal mass, relative growth rate
(RGR), residuals of RGR vs. germinant mass, whole-plant total non-structural

carbohydrates (TNpr) and residuals of TNpr vs. ﬁnal mass. The top number is the

correlation coefficient for angiosperrns only and the bottom number represents the
coefficient for all species. * P < 0.05, ** P < 0.01, *** P < 0.0001 ................................. 66

Table 2.3. Correlation statistics for interrelationships between plant functional traits and
germinant mass, ﬁnal mass, relative growth rate (RGR), residuals of RGR vs. germinant
mass, whole-plant total non-structural carbohydrates (TNpr) and residuals of TNpr
vs. ﬁnal mass. The top number is the correlation coefﬁcient for angiosperms only and
the bottom number represents the coefﬁcient for all species. * P < 0.05, ** P < 0.01, ***
P < 0.0001 ......................................................................................................................... 67

Table 2.4. Multiple regression models of relative growth rate, relative growth rate with

initial mass as a covariate, whole-plant total non-structural carbohydrates, and root mass
ratio. Models were developed by ﬁrst including the strongest bivariate predictor (Table

2.1), then adding the variable with the second strongest bivariate predictor, and its

interaction. Additional variables were left in the model if adjusted R2 values and Pratt

indices indicated that their inclusion explained additional variance in the predicted
term ................................................................................................................................... 68

Table 3.1. Mean species basal area across glacial landforms in Manistee National Forest,
near Cadillac, MI (condensed from Host and Pregitzer 1992). Outwash has the lowest
water holding capacity and rich moraines the highest .................................................... 112

Table 3.2. Mean, standard deviation, ranges and Pearson’s correlation for the different
indices of light availability used in this study ................................................................. 113

Table 3.3. Results of a standard least squares linear model for main effects of site (n = 6)
on gravimetric soil moisture (%) across different sampling dates and averaged across the
growing season ................................................................................................................ 114

Table 3.4. Linear relationship of leaf-level photosynthesis (Aarea) with photosynthetic
photon ﬂux density (PPFD) at very low soil water availability (See methods and
Appendix, Table A4. for more details about soil moisture categories). Multiple linear

regression models of Aarea as a function of PPF D in combination with plant traits (leaf

nitrogen, whole-plant mass, root area, root mass ratio, speciﬁc root area,
root depth) ....................................................................................................................... 1 15

Table 3.5. Linear relationship of leaf-level photosynthesis (Ama) with photosynthetic
photon ﬂux density (PPF D) at low soil water availability (See methods and Appendix,
Table A4. for more details about soil moisture categories). Multiple linear regression
models of Aam as a function of PPF D in combination with plant traits (leaf nitrogen,
whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth) ................. 1 17

Table 3.6. Linear relationship of leaf-level photosynthesis (Am-ea) with photosynthetic
photon ﬂux density (PPFD) at moderate soil water availability (See methods and

Appendix, Table A4. for more details about soil moisture categories). Multiple linear
regression models of Aarea as a function of PPF D in combination with plant traits (leaf

nitrogen, whole-plant mass, root area, root mass ratio, speciﬁc root area,
root depth) ....................................................................................................................... 1 19

Table 3.7. Linear relationship of leaf-level photosynthesis (Ama) with photosynthetic
photon ﬂux density (PPF D) at high soil water availability (See methods and Appendix,
Table A4 for more details about soil moisture categories). Multiple linear regression
models of Aarea as a function of PPF D in combination with plant traits (leaf nitrogen,
whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth) ................. 121

Table 3.8. Multiple linear regression models of leaf-level photosynthesis (Amea) during

very low soil moisture at OW 1 with PPFD as a covariate, root area, root depth or both
root depth and root area .................................................................................................. 123

Table 3.9. Multiple linear regression models of leaf-level photosynthesis (Aarea umol m-

2 s") as a function of photosynthetic photon ﬂux density (PPFD, umol m'2 s"), root
depth (cm) and their interaction within the three well-drained sites (OW 1, IC 1, MOR 1)
during very low soil moisture conditions ........................................................................ 124

Table 3.10. Summary of linear regression analyses for data in Figure 3.15. In all cases,
the dependent variable is residuals of survival (SURVmsid) vs. loglo canopy openness
nonlinear functions from Figure 3.14. Deﬁnitions for abbreviations and units for
independent variables are as follows: leaf nitrogen (N, pg cm'z), whole-plant mass (g).
root surface area (cmz), RMR (root mass ratio, g g'l), SRA (speciﬁc root area, cm2 g'l).
root depth (cm), Aarea (leaf-level photosynthesis during the peak of the drought, pmol
lC302 m.2 s"), WUE (water use efﬁciency during the peak of the drought, mmol C02 mol-
H20) ............................................................................................................................. 125

Table 3.11. Simple and multiple linear regression models of residSURv for OW l with

whole-plant mass, root area, root depth or a combination of these seedling characteristics
......................................................................................................................................... 126

Table 3.12. Leaf-level light compensation points (LCP) and respiration rates (R1) of
study species from shaded understory or greenhouse conditions. Data compiled from
unpublished studies and from the literature. Species values with multiple superscripts
represent averages from cited studies ............................................................................. 127

Appendix

Table A.l. Species summary of leaf-level C02 gas-exchange vs. PPF D (umol m.2 5")

linear regression equations and associated tests of signiﬁcance and R2 values. The
resulting species-level regression equations were used to: (1) estimate photosynthesis at
30 umol rn'2 s'l (Ama); (2) quantum yield (i.e., slope = QY); and (3) light compensation
point (i.e., PPF D level at which photosynthesis = 0, LCP) ............................................ 153

Table A2. Means, standard deviations and ranges of measurement times (0.0—24.00 h
local time) across sites and measurement periods during the 2002 growing season ...... 155

Table A3. Results of a standard least squares linear model for main effects and
interactions of measurement periods (n = 4), site (n = 6) and species (n = 8) on
measurement times of leaf-level gas-exchange .............................................................. 156

Table A4 Summary of soil moisture categories for gas-exchange analyses. Categories
were based on variation in volumetric soil moisture which was measured concurrently
with gas—exchange measurements during the 2002 growing season across seedling
transplant plots ................................................................................................................ 157

Table A5. Results of a standard least squares mixed linear model for main effects and

interactions of canopy openness (%), site (n = 6) and sampling date (n = 6) on
gravimetric soil moisture (%) across seedling transplant plots ...................................... 158

xi

Table A6. Results of a standard least squares mixed model for the main effects and

interactions of loglo (whole-plant mass) (g), site (n = 6) and species (n = 8) on logjo (root
mass) (g). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002 ................................................................. 159

Table A7 Results of a standard least squares mixed model for the main effects and
interactions of loglo (root mass) (g), site (n = 6) and species (21 = 8) on logjo (root area)

(cmz). The model is based on data are from all individual seedlings that were harvested
from transplant plots in June 2002 .................................................................................. 159

Table A8. Results of a standard least squares mixed model for the main effects and

interactions of loglo (whole-plant mass) (g), site (n = 6) and species (n = 8) on loglo (root
depth) (cm). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002 ................................................................. 160

Table A.9. Results of a standard least squares mixed model for the main effects and
interactions of loglo (whole-plant mass) (g), site (n = 6) and species (22 = 8) on loglo (leaf

area) (cmz). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002 ................................................................. 160

Table A.lO. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all species combined
within OW l (well-drained outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 161

Table A] 1. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all species combined
within 0W2 (sub-irrigated outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 162

Table A.l2. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all species combined
within IC 1 (well-drained ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 163

Table A.l3. Standardized major axis regression slopes (0t), elevations ([3), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all species combined
within IC 2 (sub-irrigated ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 164

xii

Table A.l4. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of Signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all Species combined
within MOR l (well-drained moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 165

Table A.15. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root mass (g) for individual species and all species combined
within MOR 2 (sub-irrigated moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 166

Table A.l6. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root

mass (g) and root area (cmz) for individual species and all species combined within OW
1 (well-drained outwash site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 167

Table A.l7. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root

mass (g) and root area (cmz) for individual species and all species combined within OW
2 (sub-irrigated outwash site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 168

Table A.l8. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root

mass (g) and root area (cmz) for individual species and all species combined within IC 1
(well-drained ice contact site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 169

Table A.l9. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root

mass (g) and root area (cmz) for individual species and all species combined within IC 2

(sub-irrigated ice contact site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 170

Table A.20. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root
mass (g) and root area (cmz) for individual species and all species combined within MOR
1 (well-drained moraine site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 171

Table A2]. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between root

xiii

mass (g) and root area (cmz) for individual species and all species combined within MOR
2 (sub-irrigated moraine site). Analyses are based on data are from all individual
seedlings that were harvested from transplant plots in June 2002 .................................. 172

Table A.22. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within OW 1 (well-drained outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 173

Table A.23. Standardized major axis regression slopes (ct), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within OW 2 (sub-irrigated outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 174

Table A.24. Standardized major axis regression slopes (a), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within IC 1 (well-drained ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 175

Table A.25. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within IC 2 (sub-irrigated ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 176

Table A.26. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within MOR 1 (well-drained moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 177

Table A.27. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and root depth (cm) for individual species and all species combined
within MOR 2 (sub-irrigated moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 178

Table A.28. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and leaf area (cmz) for individual species and all species combined
within OW 1 (well-drained outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 179

xiv

Table A.29. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and leaf area (cm ) for individual species and all species combined
within OW 2 (sub-irrigated outwash site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ............. 180

Table A.30. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between

whole-plant mass (g) and leaf area (cmz) for individual species and all species combined
within IC 1 (well-drained ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 181

Table A3]. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole—plant mass (g) and leaf area (cmz) for individual species and all species combined
within IC 2 (sub-irrigated ice contact site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 182

Table A.32. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between

whole-plant mass (g) and leaf area (cmz) for individual species and all species combined

within MOR 1 (well-drained moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 183

Table A.33. Standardized major axis regression slopes (or), elevations (B), conﬁdence
intervals and associated tests of signiﬁcance for log-log linear relationships between
whole-plant mass (g) and leaf area (cmz) for individual species and all species combined
within MOR 2 (sub-irrigated moraine site). Analyses are based on data are from all
individual seedlings that were harvested from transplant plots in June 2002 ................. 184

Table A.34. Results of a standard least squares mixed model for the main effects and
interactions of loglo (photosynthetic photon ﬂux density, PPFD) (umol m'2 s"), loglo

(leaf Nam) (pg cm'z) and site (n = 6) on water-use efﬁciency (WUE) under very low soil
moisture (see methods for soil moisture classiﬁcation scheme) ..................................... 185

Table A.35. Results of a standard least squares mixed linear model for main effects and
interactions of canopy openness (%) and site (n = 6) on gravimetric soil moisture (%)
across seedling transplant plots. Models were evaluated for ﬁve different sampling dates
(16 May, 25 June, 9 July, 25 July, 20 August, 8 September) and averaged across the 2002
growing season ................................................................................................................ 186

XV

Table A.36. Results of a standard least squares mixed linear model for main effects and
interactions of canopy openness (%) and site (21 = 6) on in situ nitrogen mineralization
rates averaged across the 2002 growing season .............................................................. 187

Table A.37. Results of a standard least squares mixed model for the main effects and

interactions of logjo (photosynthetic photon ﬂux density, PPFD) (pmol m.2 s"), logjo
(root depth) (cm) and Site (n = 6) on leaf-leaf photosynthesis under very low soil moisture
(see methods for soil moisture classiﬁcation scheme). The model is based on plot-level
averages of PPFD and root depth for each respective species ........................................ 188

Table A.38. Summary of linear regression analyses for data in Appendix, Figure A.lO.
In all cases, the dependent variable is deviation in species survival from plot average
survival (i.e., calculated as: species average plot survival - overall average plot survival

for all species) vs. loglo canopy openness from Figure 3.14. Deﬁnitions for
abbreviations and units for independent variables are as follows: leaf nitrogen content (N,

pg cm'2 ), whole-plant mass (g), root surface area (cmz), RMR (root mass ratio, g g"),
SRA (speciﬁc root area, cm2 g'l), root depth (cm), Ama (leaf-level photosynthesis during
the peak of the drought, umol COz m'2 s"), WUE (water use efficiency during the peak
of the drought, mmol C02 mol'I H20) ........................................................................... 189

xvi

LIST OF FIGURES

Figure 1.1. Proportion of ﬁne root length across root diameter classes for both live and
dead root categories. Root length data from individual seedlings were combined for all
species and fertilizer treatments ........................................................................................ 29

Figure 1.2. Representative examples of live and dead ﬁne root segments for A.
saccharum. Roots were procured from individual seedlings in late-September 2003 at the
Tree Research Center, Michigan State University. Images were acquired with a ﬂatbed
scanner (Epson Expression 1680, Nagano, Japan) at a resolution of 400 DPI and later
edited for inherent root color differences and shadowing with Adobe Photoshop 7.0
(Adobe Systems Inc., San Jose, California). All roots in this image are < 2.0 mm in
diameter ............................................................................................................................. 30

Figure 1.3. Live (LRN) versus dead (DRN) ﬁne root Nlength for four tree species. Data
represent individual seedlings. In a mixed linear least squared model partial P values for
LRN, species and their interaction as predictors of DRN were P = 0.0007, 0.0023, and
0.0614, respectively. Summary statistics for signiﬁcant (P < 0.05) regressions are:

Overall relationship, DRN = 0.332 + 0.765(LRN), R2 = 0.57, P < 0.0001, n = 55; P.
tremuloides, DRN = -0.025 + 0.924(LRN), R2 = 0.50, P = 0.0218, n = 10; A. rubrum,
DRN = -0.162 + 0.993(LRN), R2 = 0.51, P = 0.0029, n = 15 .......................................... 31

Figure 2.]. Box plots of tissue-level (Lf = leaves, St = stem, Rt = roots) non-structural
carbohydrate concentrations of angiosperms (n = 28) and gymnosperm (n = 8) species

(a). Lower and upper ends of the boxes represent the 25t and 75t percentile, lower and

upper whiskers represent the 10th and 90th percentile and the horizontal lines within the
boxes represent the median. Tissue-level TNC concentration means that do not share a
common letter are signiﬁcantly different (P < 0.05, Tukey-Kramer HSD). Total non-
structural carbohydrate (TNC) partitioning for angiosperms and gymnosperms (b).
Signiﬁcant t-test statistics (* P < 0.0001) indicate differences between plant orders
(angiosperms, gymnosperms). In TNC partitioning to speciﬁc organs (leaves, stems,
roots). Results of ANOVA for TNC as a function of order, organ (leaves, stems, roots)
and their interactions are indicated as: *** P < 0.0001 .................................................... 69

Figure 2.2. Relationships between relative growth rate and germinant mass (a), total non-
structural carbohydrates and ﬁnal mass (b), total non-structural carbohydrates and
relative growth rate (0), and total non-structural carbohydrates and the residuals of the
regression of relative growth rate vs. germinant mass (d). The inset on (d) is for the
residuals of the residuals of the regression of relative growth rate vs. germinant mass for
angiosperms only and has the same axis scales as the larger ﬁgure panel. In the larger
panels, soild lines are regressioin ﬁts of all data, and hatched lines are for angiosperrms,

and in (a) and (c) for Quercus spp. Correlation statistics for these relationships are in
Table 2.2 ........................................................................................................................... 70

xvii

Figure 2.3. Relative growth rate vs. leaf area ratio (a), and residuals of the regression of
relative growth rate vs. germinant mass vs. leaf partitioning ratio (b), and vs. leaf light
compensation point (e). Corresponding correlation statistics are in Table 2.3. See Figure
2.2 legends for other details .............................................................................................. 71

Figure 2.4. Relationships of total non-structural carbohydrates with root mass ratio (a),
leaf partitioning ratio (b), and leaf-level light compensation point for photosynthesis (c).
Corresponding correlation statistics are in Table 2.3. See Figure 2.2 legends for other
details ................................................................................................................................ 72

Figure 2.5. Seedling survival vs. the residuals of the regression of relative growth rate vs.
germinant mass for all data (larger panel) and of seedling survival vs. the residuals of the
regression of relative growth rate vs. germinant mass for angiosperms for the inset panel.
Fits are for all data (larger panel) and angiosperms only (inset). Pearson correlations are r
= 0.52, P = 0.001 for all data and r = 0.55, P = 0.002 for angiosperms ........................... 73

Figure 3.1. Map of study area with arrow pointing to Lake, Wexford and Manistee
counties, in the northern lower peninsula of Michigan ................................................... 128

Figure 3.2. Mean gravimetric soil water availability at 0-20 cm depth for study sites
located on different glacial landforms (OW = outwash, IC = ice contact, MOR =
moraine) and daily precipitation from May 1 to Septmeber 15, 2002 (Wellston-Tippy
Dam Weather Station). Sites followed by a 1 are well-drained, whereas sites followed by
a 2 are sub-irrigated ........................................................................................................ 129

Figure 3.3. Vertical proﬁles (0-20 cm, 20-40 cm, 40-100 cm) of mean gravimetric soil
water availability on different glacial landforms (OW = outwash, IC = ice contact, MOR
= moraine) on July 25, 2002 (i.e., peak of the drought). Sites followed by a 1 are well-
drained, whereas sites followed by a 2 are sub-irrigated. Results of ANOVA for soil
water with site, depth and their interaction as factors ..................................................... 130

Figure 3.4. Means (i 1 SD) of in situ nitrogen mineralization rates across landforrn sites
for different measurement dates and averaged across the growing season. For pairwise
comparisons of sites, means followed by a different letter are signiﬁcantly different at P
< 0.05 according to Tukey HSD ..................................................................................... 131

Figure 3.5. Species-level means (:1: SD) of leaf nitrogen content (N, pg cm'z) across ﬁeld
sites from the June 02 seedling harvest. Numbers within or above bars denote sample
sizes (number of plots) for each respective species. Sites are separated into well-drained
(outwash = OW 1, ice contact = IC 1, moraine = MOR 1) and sub-irrigated categories
(outwash = OW 2, ice contact = IC 2, moraine = MOR 2) (see methods for details).
Species are arranged in order of their drought tolerance. Species abbreviations are as
follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis ...................................................................................................... 132

xviii

Figure 3.6. Root depth (cm) expressed as species-level means (i SD) and as estimates at
a common whole-plant mass (see materials and methods) across ﬁeld sites from the June
02 seedling harvest. Numbers within or above bars denote sample sizes (i.e., number of
plots) for each respective species. (n.d.) indicates that the speciﬁed common mass was
beyond the range of individuals for a given species x site combination. Sites are arranged
top to bottom from the most xeric site to the most mesic site. Species are arranged in
order of their drought tolerance. Species abbreviations are as follows: Qv = Quercus
velulina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus serotina, Ar = Acer
rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba = Betula

alleghaniensis ................................................................................................................. 1 33

Figure 3.7. Relationships between root depth whole plant mass, root depth and published
values of seed mass. Relationships were examined within each of the study sites ....... 135

Figure 3.8. Leaf area ratio (cm2 g—l) expressed as species-level means (i SD) and as
estimates at a common whole-plant mass (see materials and methods) across ﬁeld sites
from the June 02 seedling harvest. Numbers within or above bars denote sample sizes
(i.e., number of plots) for each respective species. (n.d.) indicates that the speciﬁed
common mass was beyond the range of individuals for a given species >< site
combination. Sites are arranged top to bottom from the most xeric site to the most mesic
site. Species are arranged in order of their drought tolerance. Species abbreviations are
as follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis ...................................................................................................... 136

Figure 3.9. Leaf-level photosynthesis (Ama) as a function of photosynthetic photon ﬂux

density (PPFD, mmol m'2 $4) for sampling periods that varied in volumetric soil water
content (very low, low, moderate, high; see methods and Appendix, Table A4 for more
details). Linear regression summary statistics are provided within each respective graph
panel. See Tables 3.4-3.7 for parameter estimates ......................................................... 138

Figure 3.10. Multiple regression model of leaf-level photosynthesis (Ama, pmol m.2 s")
in relation to photosynthetic photon ﬂux density (PPFD, pmol m-2 s'l) and leaf Narea (pg

cm'z) across all sites during (1) very low, (2) low and (3) moderate soil moisture
conditions. Each datum represents a species >< plot mean. Regression models are as

follows: (1) very low, Am,l = ~2.978 + 0.949 (loglo PPF D) + 1.294 (Ioglo leaf Nam),
adjusted R2 = 0.31, n = 227, P < 0.0001; (2) low, Ama = -5.176 + 1.555 (loglo PPFD) +
2.115 (loglo leaf Nana) + 5.308 (logjo PPF D X logo leaf Nam), adjusted R2 = 0.49, n =
230, P < 0.0001; and (3) moderate, Aarea = -5.999 + 2.174 (loglo PPFD) + 1.833 (logjo

leafNam) + 7.307 (logjo PPFD x loglo leaf N,,,,), adjusted R2 = 0.69, n = 230,
P < 0.0001 ....................................................................................................................... 14o

xix

Figure 3.11. Multiple regression model of leaf-level photosynthesis (Aarea, pmol In2 S")

in relation to photosynthetic photon ﬂux density (PPFD, pmol m.2 S") and root depth
(cm) across all sites during (1) very low and (2) low soil moisture conditions. Each
datum represents a species >< plot mean. Regression models are as follows: (1) very low.
Am, = -1.612 + 0.31 (loglo PPFD) + 1.028 (logjo root depth), adjusted R2 = 0.33, n =
227, P < 0.0001; (2) low, Ama = -2.380 + 1.482 (loglo PPF D) + 0.911 (logjo root depth)
+ (loglo PPFD X loglo root depth), adjusted R2 = 0.44, n = 230, P < 0.0001 ................ 142

Figure 3.12. Multiple regression model of instantaneous water-use efﬁciency in relation

to photosynthetic photon ﬂux density (PPF D, pmol m'2 s") and leaf Narea (pg cm'z) at
OW 1 during very low soil moisture conditions. Each datum represents a species >< plot

mean. Regression model: WUE = -5.75 + 1.62 (PPFD) + 2.87 (leaf Narea), adjusted R2 =
0.28, n = 30, P = 0.0042 .................................................................................................. 144

Figure 3.13. Relationships between pre-dawn water potential (MPa) and speciﬁc root

area (cm2 g'l), total root surface area (cm2) and root depth (cm). Each datum represents
individual seedlings of all species across all study sites. Sites followed by a 1 are well-
drained, whereas sites followed by a 2 are sub-irrigated. Associated correlation statistics
are provided within each graph panel ............................................................................. 145

Figure 3.14. Relationships of seedling survival (%) versus Loglo PPFD across all species
within well-drained sites (OW 1, 1C 1, MOR 1). Seedling survival was estimated as the
percentage of the original seedling population (July 01) that was alive in October 02.
Each datum represents a plot-level PPF D average. Data were ﬁtted with a Gompertz
function with the general form: 01exp[-exp(02- 03—Log10 canopy openness)]. Each site-
speciﬁc function was solved for the best ﬁ function (i.e., minimized residual sums of
squares) iteratively using the nonlinear ﬁt platform within J MP (SAS Institute, Cary,
North Carolina). All ﬁts were signiﬁcant at P < 0.05 .................................................... 146

Figure 3.15. Relationships of SURVresid (i.e., residuals of survival vs. canopy openness
nonlinear functions) with leaf Nam, size (whole-plant mass) and morphological (root
area, SRA, RMR, root depth) and physiological characteristics (leaf-level photosynthesis,
Aarea; water-use efficiency, WUE). Relationships were examined within each of the

three well-drained Sites (OW 1, IC 2, MOR 1). Regression equations, adjusted R2, P
values and n for these relationships are presented in Table 3.10 .................................... 148

Figure 3.16. Conceptual diagram of factors inﬂuencing interspeciﬁc survival of ﬁeld
transplanted seedlings. Plus signs (+) indicate signiﬁcance in correlation analyses or
best-ﬁt linear models. Dashed line indicates that additional, unmeasured traits that are
associated with plant mass may have a positive effect on seedling survival .................. 151

XX

Appendix

Figure A.l. Correlation matrix of gravimetric soil moisture (%) for the driest sampling
date from the 2001, 2002 and 2003 growing seasons. Each datum represents a plot-level

average from the seedling transplant experiment. All values were loglo transformed prior
to analysis. Sample size, Pearson’s correlation coefﬁcients and signiﬁcance of
coefﬁcients are shown in each respective panel ............................................................. 190

Figure A.2. Correlation between seedling survival (%) recorded after the peak of the
drought in 2002 (6/02-10/02) and seedling survival (%) throughout the duration of the
seedling transplant experiment (7/01-10/02). Each datum represents a species-speciﬁc
average of seedling survival at the plot level. Sample size, Pearson’s correlation
coefficients and signiﬁcance of coefﬁcients are shown within the panel of the

graph ............................................................................................................................... 191

Figure A.3. Correlations of leaf area ratio (LAR), speciﬁc root area (SRA), root mass
ratio (RMR), root depth and root surface area with total plant mass. Sample size,
Pearson’s correlation coefﬁcients and signiﬁcance of coefﬁcients are shown next to each
respective panel ............................................................................................................... 192

Figure A.4. Species-level means (1 SD) of whole-plant mass (g) across ﬁeld sites from
the June 02 seedling harvest. Numbers above bars denote sample sizes (number of plots)
for each respective species. Sites are separated into well-drained (outwash = OW 1, ice
contact = IC 1, moraine = MOR l) and sub-irrigated categories (outwash = OW 2, ice
contact = IC 2, moraine = MOR 2) (see methods for details) and are organized top to
bottom from the most xeric to the most mesic. Species are arranged in order of their
drought tolerance. Species abbreviations are as follows: Qv = Quercus velutina, Qa =
Quercus alba, Qr = Quercus rubra, Ps = Prunus serotina, Ar = Acer rubrum, As = Acer
saccharum, Fa = Fraxinus americana, Ba = Betula alleghaniensis ............................... 194

Figure A.5. Species-level means (:l: SD) of root surface area (cmz) across ﬁeld sites from
the June 02 seedling harvest. Numbers within or above bars denote sample sizes (number
of plots) for each respective species. Sites are separated into well-drained (outwash =
OW 1, ice contact = 1C 1, moraine = MOR 1) and sub-irrigated categories (outwash =
OW 2, ice contact = IC 2, moraine = MOR 2) (see methods for details). Species are
arranged in order of their drought tolerance. Species abbreviations are as follows: Qv =
Quercus velutina, Qa = Quercus alba, QT = Quercus rubra, P3 = Prunus serotina, Ar =
Acer rubrum, As = Acer saccharum, Fa = Fraxinus americana, Ba = Betula
alleghaniensis ................................................................................................................. l 95

Figure A.6. Root mass ratio (g g_]) expressed as species-level means (:I: SD) and as
estimates at a common whole-plant mass (see materials and methods) across ﬁeld sites
from the June 02 seedling harvest. Numbers within or above bars denote sample sizes
(i.e., number of plots) for each respective species. (n.d.) indicates that the speciﬁed
common mass was beyond the range of individuals for a given species X site
combination. Sites are arranged top to bottom from the most xeric site to the most mesic

xxi

site. Species are arranged in order of their drought tolerance. Species abbreviations are
as follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis ...................................................................................................... 196

Figure A.7. Speciﬁc root area (cm2 g_l) expressed as Species-level means (i SD) and as
estimates at a common root mass (see materials and methods) across ﬁeld sites from the
June 02 seedling harvest. Numbers within or above bars denote sample sizes (i.e.,
number of plots) for each respective species. (n.d.) indicates that the speciﬁed common
mass was beyond the range of individuals for a given species X site combination. Sites
are arranged top to bottom from the most xeric site to the most mesic site. Species are
arranged in order of their drought tolerance. Species abbreviations are as follows: Qv =
Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus serotina, Ar =
Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba = Betula
alleghaniensis ................................................................................................................. 1 98

Figure A.8. Species-level means (i SD) of leaf-level photosynthesis (Ama) across ﬁeld
sites at four sampling dates that contrasted in volumetric soil moisture content (%) during
the 2002 growing season (left to right, very low = 3.3%; low = 4.2%; moderate = 6.8%;
high = 11.2%; and see also Appendix, Table A.4 for additional details). Numbers
contained within or above bars denote sample sizes (number of plots) for each respective
species. Sites are organized top to bottom from the most xeric to the most mesic (well-
drained sites, outwash = OW 1; ice contact = [C 1; moraine = MOR l; and sub-irrigated
sites, outwash = OW 2; ice contact = [C 2; moraine = MOR 2) (see methods for details).
Species are arranged in order of their drought tolerance. Species abbreviations are as
follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, PS = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis ...................................................................................................... 200

Figure A.9. Species-level means (i SD) of leaf-level water-use efﬁciency (WUE) across
ﬁeld sites (0W1, OW 2, IC 1, IC 2, MOR 1, MOR 2) at four sampling dates that
contrasted in volumetric soil moisture content (%) during the 2002 growing season (left
to right, very low = 3.3%; low = 4.2%; moderate = 6.8%; high = 11.2%; and see also
Appendix, Table A.4 for additional details). Numbers contained within or above bars
denote sample sizes (number of plots) for each respective species. Sites are organized
top to bottom from the most xeric to the most mesic (well-drained sites, outwash = OW
1; ice contact = IC 1; moraine = MOR 1; and sub-irrigated sites, outwash = OW 2; ice
contact = IC 2; moraine = MOR 2). Species are arranged in order of their drought
tolerance. Species abbreviations are as follows: Qv = Quercus velutina, Qa = Quercus
alba, Qr = Quercus rubra, PS = Prunus serotina, Ar = Acer rubrum, AS = Acer
saccharum, Fa = F raxinus americana, Ba = Betula alleghaniensis ............................... 201

Figure A.10. Relationships of survival deviations (i.e., species average plot survival —

overall average plot survival for all species) with leaf Narea, size (whole-plant mass) and
morphological (root area, SRA, RMR, root depth) and physiological characteristics (leaf-

xxii

level photosynthesis, Aarea; water-use efﬁciency, WUE). Relationships were examined
within each of the three well-drained sites (OW 1, IC 2, MOR 1). Regression equations,

adjusted R2, P values and n for these relationships are presented in Appendix,
Table A.38 ....................................................................................................................... 202

xxiii

INTRODUCTION
General Introduction

Plant ecologists have long sought the physiological mechanisms that account for
the distribution of species through time and space. Several lines of research highlight the
importance of resource availability in shaping spatial patterns in the distribution of tree
species. For example, numerous landscape-scale studies have documented associations
between the distribution of overstory tree species and variation in soil resources
(nitrogen, soil water) (Zak et al. 1989, Reich et al. 1997a, Bongers et al. 1999, Wang et
al. 2006, Engelbrecht et al. 2007). Among seedlings and saplings, growth and survival
generally increase with increasing levels of light (Pacala et al. 1994, Kobe et al. 1995)
and nitrogen availability (Walters and Reich 1997, Finzi and Canham 2000, Walters and
Reich 2000), but responses are species-speciﬁc. Furthermore, spatial and temporal
variation in soil water availability affect juvenile tree growth and survival (Walters and
Reich 1997) and similar to nitrogen, responses differ across species (Caspersen and Kobe
2001, Engelbrecht and Kursar 2003, Engelbrecht et al. 2005, Kobe 2006). Collectively,
these observations suggest that species distribution patterns may stem from differential
species performance across resource gradients.

It has been hypothesized that species differences in growth capacities may
underlie distribution patterns across resource gradients. For example, species
composition in high resource, competitive environments may reﬂect rapid growth
responses, whereas species with low inherent growth rates may have the ability to tolerate
harsh growing conditions (Grime 1977, Chapin 1980). This hypothesis suggests that

there is an unavoidable “trade-off” between rapid growth under high resources versus

survival in poor resource environments, which has been supported with experimental
evidence across gradients of light (Kobe et al. 1995, Poorter and Bongers 2006, Poorter et
al. 2006) and soil resources (Schreeg et al. 2005). Thus, both theory, and differential
species performance ranks and distribution patterns across resource gradients imply that a
single plant species cannot be a superior competitor in all resource environments (i.e.,
Jack of all trades is a master of none, Bradshaw 1965). So, why can’t a given species be
superior in all aspects? Plants allocate resources to contrasting functions (e.g., growth,
support, defense, storage, resource acquisition, reproduction) and these functions are
subject to opposing demands. For example, investment in a speciﬁc structure or function
that leads to enhanced survival in chronically low resource environments, may limit a
plant’s ability to acquire carbon or capture soil resources and ultimately reduce a plant’s
growth potential.

To date, studies have primarily focused on plant traits that underlie growth rates,
especially under optimal resource environments. For example, shade-tolerant and
intolerant species differ in their leaf-level photosynthetic responses (i.e., a proxy for
potential growth rates) to growth light intensity, which may provide a mechanism for
species sorting across successional gradients of light availability (Bazzaz 1979). In
addition, based on a plant competition model, Tilman (1988) presented predictions that
relative growth rates, which are hypothesized to be inﬂuenced by a plant’s proportional
allocation to leaves and roots, are a major determinant of grassland successional
dynamics across a soil N supply gradient. Furthermore, potted plant studies showed that
allocation to leaves, leaf surface area per leaf mass and whole-plant photosynthetic rates

(i.e., integration of leaf allocation and speciﬁc rates of photosynthesis) were positively

related to relative growth rates (Poorter et a1. 1990, Walters et al. 1993b). Although these
studies provide an integrative understanding of the linkages among growth, allocational
and physiological attributes and their potential role for the success of species in high
resource, competitive environments, there is a paucity of studies examining the
determinants of plant survival under poor resource conditions.

Variation in traits associated with resource economies likely contribute to the
growth versus survival trade-off. For example, three potential whole-plant mechanisms
thought to underlie plant survival in low resource environments include enhanced access
to limiting resources, storage of resources and greater use-efﬁciency of resources to
produce biomass. My global hypothesis is that plant traits associated with resource
access, storage and resource-use efﬁciency under low resources occur at a trade-off with
growth capacity under high resources. In order to gain a greater ability to explain the
mechanistic causes of differential species performance ranks across resource gradients, it
is necessary to isolate which plant functional traits (e.g., biomass allocation patterns,
morphological and physiological traits) underlie resource access, storage and resource-
use efficiency and how they relate to growth and survival in low versus high resource
environments. I have proposed an experimental framework with potted plant and ﬁeld-
based transplant studies that explicitly examines the effects of nitrogen (Chapter 1), light
(Chapter 2) and soil water availability (Chapter 3) on whole-plant
physiology/morphology, allocation programs, nutrient dynamics, species-speciﬁc growth

and survival, and their interactions.

Organization of dissertation

Nutrient resorption from senescing leaves is a well-documented nutrient conservation
strategy (Kobe et al. 2005), recycling ~ 50% of maximum foliar N content across a
variety of perennial life—forms (Aerts 1996). Fine roots may function similarly to leaves,
but evidence for root resorption is equivocal. For chapter 1, I carried out a potted plant
experiment that allowed me to investigate how ﬁne root dimensions and N concentrations
change during senescence and address the degree to which these changes inform the
unresolved issue of root N resoprtion for N fertilized and unfertilized seedlings of species
that differ in soil N afﬁnity. I hypothesized that species associated with sites that have
low N availability (Populus tremuloides Michx., Acer rubrum L.) would exhibit greater
root N conservation than species that typically occur on soils with high N status (Acer
saccharum Marsh. and Betula allegheniensis Britton).

Allocation to carbohydrate storage has been proposed as a low light carbon
conservation strategy that contributes to the trade-off between high resource growth
potential vs. low resource survival (Kobe 1997). This notion has been partly supported
by studies showing that carbohydrate storage is positively related to survival (Canham et
a1. 1999, lyer 2006, Myers and Kitajima 2007, Poorter and Kitajima 2007) and negatively
related to growth (lyer 2006, Myers and Kitijima 2007, but see, Poorter and Kitijima
2007). The carbohydrate storage vs. growth association is unlikely to be manifested by a
single trait (e.g., storage capacity), but rather by expressions of various morphological
and physiological growth-related traits. However, interrelationships between
carbohydrate storage and growth-related traits have received little attention and remain

incompletely understood. In chapter 2, I investigated the functional traits underlying

variation in whole-plant carbohydrate storage and relative growth rates, and potential
trait-based trade-offs between them for 36 temperate and boreal woody species
(angiosperm vs. gymnosperm) that were grown in a common low-light environment.
Traits potentially enhancing young seedling survival on drought-prone sites (e.g.,
greater proportional mass allocation to roots, deep roots, and conservative water use) may
compromise growth potential, and thus competitive ability, when water is predictably
plentiful. For example, increased allocation of biomass to root systems and/or the
production of deep rooted large diameter “taproots” may occur at the expense of
allocation to resource harvesting structures (e.g., proportional allocation of mass to leaf
and root area), which contribute to high growth capacities under optimal resource
conditions (Reich et al. 1998a, Poorter 1999, Walters and Reich 2000, Comas et al.
2002). Therefore, traits that confer survival during episodic drought events may occur at
a trade-off with traits enhancing growth potential when soil water is plentiful. In chapter
3, I examined the relationship of plant traits to tolerance (i.e., survival) of water deﬁcits

using a conceptual framework that classiﬁed traits into water-use efﬁciency (WUE) and

water access (Waccess) categories.

CHAPTER 1
MEASUREMENT BASES FOR FINE ROOT N CONCENTRATIONS:

IMPLICATIONS FOR SENESCENCE-RELATED N LOSS IN TEMPERATE
TREE SEEDLINGS

ABSTRACT

I investigated how ﬁne root dimensions and nitrogen (N) concentrations vary during
senescence and address the degree to which these changes inform the unresolved issue of
root N resorption in perennial plants. 1 estimated the difference in N between live and
dead ﬁne roots (AN) on mass, length, and calcium (N losszroot Ca) bases for fertilized (N
+ Ca) and unfertilized potted seedlings of four tree species. Compared to live roots, dead
roots had higher N on a mass basis, and lower N on length (-5 to -16%) and Ca (—14 to
-48%) bases. These differences could be partially ascribed to changes in non-N root
mass during senescence, which decreased substantially for all species (—23 to -40%).

AN on a mass basis, corrected for root mass loss ranged from —1 2 to —28%. For
Indivrdual seedlings, dead and live root N concentrations were posrtively correlated (R =

0.57, P < 0.0001), indicating that live root N is a major determinant of senesced root N.
Although leaching and microbial immobilization may partially obscure quantiﬁcation of
N in senesced roots, these results along with re-analyzed values from the literature
suggest that N resorption may occur in ﬁne roots to a greater degree than has previously
been reported, which may have implications for whole-plant resource economics and

ecosystem N cycling.

Introduction

Patterns in nitrogen (N) resorption from senescing leaves and their possible consequences
for ecological properties such as whole-plant resource economies and nutrient cycling
have been well described (Aerts 1996, Killingbeck 1996, Silla and Escudero 2004, Kobe
et al. 2005). Fine roots (here deﬁned as < 2 mm in diameter) may function similarly to
leaves, however, the extent of N resorption from ﬁne roots of perennial plants remains
unresolved (Gordon and Jackson 2000) for several reasons: (1) root studies are
methodologically challenging and labor-intensive; (2) assessing ﬁne root death and
senescence is often ambiguous and subjective (Pregitzer 2002); and (3) artifacts
associated with some methodologies could confound estimates of root resorption. For
example, estimating root N resorption efficiency (%) as the difference in mass-based root
N concentrations between live and dead roots (Nambiar 1987, Aerts 1990, Gordon and
Jackson 2000) implicitly assumes that all other non-N constituents of root mass remain
static. Although never quantiﬁed for roots, resorption of non-N mass from senescing
leaves underestimates actual N resorption by as much as 20% (van Heerwaarden et al.
2003). Fine roots contain mobile compounds (e.g., 2-20% non-structural carbohydrates,
Pregitzer et al. 2000, Kobe, unpublished data) that if resorbed would, as for leaves, lead

to underestimates of mass-based N resorption. I argue that estimated differences in mass-
based N concentrations of live versus dead roots (ANmass) are confounded by
physiological (i.e., carbohydrate resportion) and dimensional (i.e., root shrinkage)

changes during senescence, and thus, these estimates are likely biased.

As an alternative to mass-based expressions, I propose that ﬁne root N

concentrations based on unit root length (ANlength) and unit root calcium (ANCa) provide

more accurate measures of actual senescence-induced alterations in root N status. Root

mass loss during senescence can be described as the product of two dimensional

components: decreased mass per root length and decreased root length. Thus, ANlength,

which is insensitive to shifts in non-N mass per unit root length, should always be more

accurate than ANmaSS, and if alterations in root length during senescence are minimal,

then ANlength should closely estimate N loss from live roots. However, root length,

analogous to root mass per length, could change substantially during senescence. In this

circumstance, ANCa may better estimate changes in root N than either ANmass or ANlength

because Ca is phloem-immobile (McLaughlin and Wimmer 1999) and does not resorb
during leaf (van Heerwaarden et al. 2003) and presumably root senescence. Under these

conditions, Ca is likely stable as roots senesce and differences in N per unit root Ca

(ANCa) between live and dead roots would be insensitive to non-N root mass resorption.

In this paper, I focus on quantifying senescence-related changes in root mass and
N, reconciling mass-, length-, and Ca-based expressions of AN and discuss the possible
implications of these patterns for resorption of N from ﬁne roots. Speciﬁcally I ask: (1)
Do N concentrations differ between live and dead ﬁne roots? (2) How do AN estimates
compare among mass, length and Ca measurement bases? (3) Do ﬁne roots lose non-N
root mass during senescence, and, if so, can this account for AN differences among
measurement bases? (4) And, lastly, do AN estimates vary with species, N supply and/or
live root N content? To address these questions, I quantiﬁed and analyzed ﬁne root AN

on mass-, length-, and Ca-bases, and differences in root length and mass per length from

live to dead roots. for 3-year old potted seedlings of four broad-leaved tree species in

fertilized (N + Ca) and unfertilized treatments.

Materials and Methods

Plant Material, Growing Conditions and Experimental Design

The experiment took place at the Tree Research Center, Michigan State University, East
Lansing, MI (42°40' N, 84027' W). Seeds of Populus tremuloides Michx. (quaking

aspen), Acer rubrum L. (red maple), Acer saccharum Marsh. (sugar maple) and Betula
alleghaniensis Britton (yellow birch) were germinated in bench-top ﬂats ﬁlled with
potting soil (F affard 2 mix, Agawan, MA) beneath 50% neutral density shade cloth in a

temperature-controlled greenhouse (Mean daily minimum and maximum temperatures
were 23.6 and 184°C respectively). Populus tremuloides was germinated in mid-May

2000, and the other species in mid-May 2001. In early-June 2001, single seedlings of
each species were transplanted into plastic pots (17.15 cm width x 18.73 cm height) ﬁlled
with a homogenized low fertility ﬁeld soil mixture (Rubicon-Menominee, and Graycalm
and Grayling sands) and placed into randomly selected positions in two outdoor
hoophouses (4.6 m x 27.4 m). The ﬁeld soil was collected from the top 15-20 cm of sub-

organic soil with a backhoe at a forested sandy glacial outwash site in Roscommon, MI
(44°12' N, 84036' W). Hoophouses were covered with neutral density shade cloth to
achieve a targeted light environment of 35% full sun. Supplemental deionized water was

applied to seedlings every 3-5 days from early-June to mid-September throughout the

experiment. Within each hoophouse >< species group, half of the pots were fertilized with

a mixture of N delivered as 13.5 gm.2 of (N H4)7_SO.4 granules and Ca delivered as 150

g-m—2 of CaSO4 powder. The fertilizer was applied during late-July 2001 and in mid-

June during 2002 and 2003. Approximately 152 seedlings were allocated to each

hoophouse (2) X species (4) X fertilizer combination (2).

Root Measurements
Three to ﬁve seedlings were selected at random from each hoophouse >< species ><
fertilizer combination (total = 65 seedlings) over two weeks in late-September 2003 to

sample naturally-senescing roots. Soil was removed from root systems of individual

seedlings by gently rinsing with deionized water. I deﬁned ﬁne roots as non-woody 1“,

2nd and 3rd order lateral roots that were < 2.0 mm in diameter. More than 80% of the

total length of roots sampled (~ 1600 m) was < 0.5 mm in diameter and these
distributions were similar for live and dead collections (Figure 1.1). On an individual
seedling basis, I collected samples of root segments from live and dead root populations.
The total number of root segments per sample was determined by the mass needed for
root nutrient analyses. Classiﬁcation of live versus dead roots was based on color and
easily observed anatomical features. Live roots were translucent and white to tan,
whereas dead roots were dark gray to black, but showed no visible signs of decay
(McClaugherty et al. 1982, Steele et al. 1997) (e.g., Figure 1.2). All dead roots were
physically attached to the whole-root system and were disconnected with a slight pull on
individual dead root segments. Visual classiﬁcation was corroborated by removing the

root cortex and documenting the presence (live) or absence (dead) of an intact stele

10

(Spaeth and Cortes 1995) on a minimum of ﬁve randomly selected root segments per
sample. If any of the selected root segments were incorrectly classiﬁed, the entire sample
was rejected and a new sample was collected from the same seedling using reﬁned
selection criteria (i.e., based on a restricted range of root color).

Fine root collections were refrigerated S 2 days until fresh images of root samples
(5-9 images X species X fertilizer X root type combination) were acquired with a ﬂatbed

scanner at a resolution of 400 DPI (Epson Expression 1680, Nagano, Japan). Following

digitization, root samples were dried at 70°C for at least 48 hours, and then weighed.

Digitized images were manually edited with Adobe Photoshop 7.0 (Adobe Systems Inc.,
San Jose, California) with the goal to produce a black (roots) and white (background)
image that faithfully captured the original root image. Edited images were analyzed for

total root length with WinRhizo Pro 5.0 software (Regent Instruments, Blain, Quebec).
For a subset of edited images (3-5 each for live and dead roots of each species), 1St order

root length of individual roots was quantiﬁed (total n = 25 each for live and dead roots).
For each respective species, average ﬁrst order root length data (11 = 25) were used to

estimate changes in root length with senescence (i.e., root shrinkage) and was calculated

as: ((lengthLR — lengthDR)/lengthLR)) X 100. Root mass and total root length data from

individual samples were used to estimate live and dead root mass per root length.

Dried root samples were pulverized into a ﬁne powder with a ball mill (Kinetic
Laboratory Equipment Company, Visalia, California), or, for very small samples, with a
mortar and pestle. Nitrogen concentrations were measured with a CHN elemental
analyzer (Carlo-Erba, Milan, Italy). For root Ca measurements, sub-samples from

individual seedlings had to be aggregated over each combination of hoophouse X species

11

X fertilization X root status (live/dead) to obtain enough material for analysis.
Approximately 30-150 mg from each aggregated root sample was microwave digested in
a nitric acid-hydrogen peroxide mixture (Mars 5, CEM Corporation, Matthews, NC) and
Ca concentrations were measured with Direct Current Plasma Emission Spectroscopy
(DCP-AES, SMI Corporation). During microwave digestion for Ca analysis, several
composite samples were lost due to equipment failure, which resulted in no replication

for some treatment combinations. Ca and N concentrations were expressed on an oven-

. -l .
dry mass baSIS (Camass. Nmass_ mg g ) and N concentrations were also expressed on a
. —l . .
root length baSIs (N length, pg'cm ). Length-based N concentrations were estimated as
follows: (Nmass X root mass)/root length. Ca concentrations were used to express root N

on a Ca-basis (e.g., average NmaSS /aggregated Camass; NCa, unitless).

Calculations

Change in ﬁne root N during senescence (AN, %) was calculated from direct

measurements of live (LR) and dead roots (DR) as AN = ([N]LR — [N]DR)/ [N]LR X 100,
on three measurement bases: (1) per unit root mass (ANmaSS ); (2) per unit root length

(AN|ength); and (3) per unit Ca mass (ANCa). Note that none of these calculations

explicitly accounts for non-N root mass changes between live and dead roots.
Estimating non-N root mass change (Amass) between live and dead roots was
accomplished by combining changes in two dimensional components; root mass per root

length, and root length as:

12

 

tr w )
massDR
length len th
Amass=1- : DR; x[-l—g-h—DR—:| x100.
massLR eng‘ LR
\\ lengthLR _ j )

 

 

 

 

 

 

 

. . t
Changes In root mass per length and In 1S order root length can also be used to correct

ANmass for Amass. First, correcting AN only for changes in root mass per length yields '

the per root length based expression of AN:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

( . .
NDR(mg) x massDR(g) ( NDR(mg)
mass (g) length (cm) length (cm)
ANlengthzl— , DR 2 : DR x100=l— N DR x100.
NLR(mg) massLR(g) Lleg)
x length (cm)
I massLR(g) lengthLR(cm) ) K LR 2

Modifying ANlength with a correction for changes in total root length during senescence

 

 

 

 

 

 

 

 

 

 

 

 

yields:
{ \
NDR (mg) X lengthDR (cm)
length (cm) length (cm)
ANcorrected =1 — DR N ' LR x 100 -
LR (mg)
\ lengthLR (cm) )

Thus, ANconected accounts for both changes in root mass per length (i.e., ANlength) and

changes in length between live and dead roots. This calculation is based on the
assumption that measurements of root length changes did not include any tissue loss (e. g.,

belowground herbivory).
Unlike ANmaSS and to a lesser extent ANlength, ANCa may not require a correction

for mass loss because Ca is assumed to be immobile during senescence. To check the

13

assumption of Ca immobility, expected dead root Ca concentrations were calculated from
measured live root Ca and estimates of Amass:

1

measured Ca LR (mg-g )

)= (. ll
:length DR (cm); Length DR (cm)
_ length LR (cm)

1

 

expected Ca DR (mgg—

 

 

 

Length LR (cm)

 

 

 

 

 

Statistical analyses

I used J MP and its general linear models procedure for ANOVA for all analyses (SAS
Institute, Cary N. Carolina). Individual plants were considered experimental units for
most analyses. Before main analyses, ﬁne root N concentrations were analyzed with a
model that included main effects and interactions of hoophouses (i.e., blocks) (n = 2) and

root status (n = 2; live root vs. dead root). Preliminary ANOVA models indicated that

hoophouse and its interactions for Nmass and Nlength were not signiﬁcant (P 2 0.22); thus

we pooled these factors in the error term for subsequent analyses (Bancroft 1964). Fine

root N concentrations were analyzed with a model that included main effects and

interactions of species (n = 4), fertilizer (n = 2) and root status (n = 2). I analyzed ANmaSg,~

and ANlength with a model that included main effects and interactions of species (n = 4)

and fertilization treatment (n = 2). When main effects were found to be signiﬁcant (P S
0.05) in ﬁnal ANOVA models, I compared pairs of treatment means with tests of least
squares signiﬁcant difference (Tukey-Kramer HSD). Due to lack of replication for some

treatment combinations for root Ca concentrations and the similarity of fertilized and

14

unfertilized NCa values within live and dead categories, I present ANCa data as species

means without statistical comparisons.
I analyzed factors affecting dead root N with a mixed least squares linear model

that included main effects and interactions of species (n = 4) and fertilization (n = 2) as
nominal factors and live root Nlength as a continuous factor. The model excludes the

fertilization main effect and its interactions since P 2 0.25 in the preliminary model
(Bancroft 1964). In addition, we used simple linear regression to model the overall
relationship between live and dead root N concentrations, and in cases of signiﬁcant

species effects in the mixed model, species were analyzed individually.

Results

The basis on which N concentrations were expressed strongly inﬂuenced the direction
and magnitude of apparent changes in N between live and dead roots. On a mass basis, N

was actually higher in dead than live roots. In contrast, N was lower in dead than live

roots when expressed in terms of length, Ca or when ANmSS was corrected for changes in
mass during senescence (ANconected).

Overall, mass-based root N concentrations (N mass) varied with species,

fertilization, root status (live/dead) and species x root status interactions, but root status

effects dominated (Table 1.1a). NmaSS was higher in dead than live roots and in fertilized
versus non-fertilized treatments. For dead roots, NmSS was greater for A. saccharum and

A. rubrum than for P. tremuloides and B. alleghaniensis, whereas Nmass did not differ

15

among species within live roots (P < 0.01, Tukey HSD, Table 1.1). Averaged across

species, ANmass was 14.2% for the fertilized treatment and 27.8% for the unfertilized

treatments (Table 1.2). Among species x fertilization treatments, ANmass (Table 1.2)

ranged from a 6.6% increase in dead roots for fertilized P. tremuloides to a 40.4%

increase for unfertilized A. rubrum.

Length-based root N concentrations (Nlength) varied with species and root status,
but not with fertilization (Table 1.1a). Species rankings in Nlength were similar to those

for Nmass. In contrast to patterns for Nmass, ANlength values were approximately 9%

lower for dead than live roots and values were unaffected by species, fertilization

treatments, or their interactions (Table 1.2). Given the weak effects of fertilization on

ANlength and ANlength, the lack of fertilizer interactions for Nmass and ANmass and low
replication for NCa, I pooled fertilizer treatments for all subsequent summaries. Like
Nlengths calcium-based root N concentrations (NCa) were greater for live roots than for
dead, and species ranked similarly (Table 1.1). Values for ANCa were even lower than

those for ANlength and indicated that, averaged among species; NCa was 30% lower for

dead roots than live roots (Tables 1.1, 1.2).

Both root mass per root length and root length decreased from live to dead roots,
indicating substantial root mass loss during senescence for all species (Table 1.3).
Averaged among species, mean root mass per root length decreased 24% and mean 1St

order root length (cm) decreased by 13%, (Table 1.3), thus total mass loss was

approximately 34% (34% = 1- (0.76 x 0.87). ANmass values corrected for Aroot mass

16

(ANconected, mean of species = —21.0%) were closer to ANCa values (mean = —30.4%)
than were AN|ength (mean = -9.1%) or ANmass values (mean = 19.9%) (Tables 1.2, 1.3).

Expected Ca concentrations in dead roots (Expected C aDR). calculated from live root Ca,

changes in mass, and assuming stable Ca during senescence (see Methods) were similar
to measured dead root Ca values for 3 of the 4 study species, although expected values

were lower than measured dead root Ca in all cases (11% lower on average, Table 1.3).
Live root Nlength and species strongly affected dead root Nlength and their
interactions were marginally signiﬁcant (Figure 1.3 legend). However, a model including

live root Nlength. species, and their interactions explained only modest additional variation
in dead root Nlength (adjusted R2 = 0.69) over a model with only live root Nlength as a

predictor (adjusted R2 = 0.57). Within species, live Nlength vs. dead Nlength relationships

were signiﬁcant for A. rubrum and P. tremuloides, which had similar slopes and
intercepts (Figure 1.3 legend). Furthermore, intercepts were not signiﬁcantly different
from zero for either the species pooled data set (P = 0.16) or for individual species (P 2
0.84 in all cases). Collectively, these results indicate that: (I) live root N was the primary
determinant of dead root N and (2) given a zero intercept and a linear relationship, dead

root N was a constant proportion of live root N over the range of live root N examined.

17

Discussion

Comparing estimates of AN

Changes in N from live to dead roots varied markedly among measurement bases,
ranging from a 20% increase for ANmag,S to a 30% decrease for ANCa with ANlength values

intermediate (11% decrease). A major factor contributing to these differences was root
mass loss between live and dead roots which declined, on average, by 34%. Neither

mass-based nor length-based N concentrations completely account for root mass loss as
roots senesce. ANmass values corrected for root mass loss (ANconected) indicated a loss of

approximately 21% N (Table 1.3). These values likely represent the closest
approximation of actual N loss from senescing roots.

My results call into question the results of comparisons of live and dead root N
made on a mass-basis and not corrected for mass-loss during senescence. To my
knowledge, all other studies to date that have evaluated N in live vs. dead roots have done
so on an uncorrected mass-basis. These studies have found higher dead root N (N ambiar
1987, this study), no difference (Aerts 1990, Gordon and Jackson 2000), and 10% higher
N in live roots (Meier et al. 1985). If root mass changes during senescence, mass-based
measures are intrinsically biased and underestimate N loss between live and dead roots.

Since Ca is phloem-immobile, root Ca concentrations between live and dead ﬁne
roots may be more stable (McLaughlin and Wimmer 1999) than either ﬁne root mass per
length or length during senescence and thus I speculated that Ca should provide a more
accurate estimate of N loss than length- or mass-based estimates. In leaves, Ca moves
passively in the transpiration stream and accumulates in deciduous foliage throughout the

growing season, with the highest Ca concentrations in senescent leaves (Burton et al.

18

1993, Duchesne et al. 2001). The same mechanism may not occur in roots, but it is
notable that estimates of Ca in dead roots (i.e., calculated from live root Ca
concentrations and mass changes, and assuming constant Ca concentrations) slightly but
consistently underestimated measured values of dead root Ca (Table 1.3). This
underestimate may have occurred if root Ca increases with age, as live root samples
likely contained a wide range of root ages, from recently initiated to old, whereas dead
root collections were likely dominated by older roots. Thus, my assumption of constant
root Ca from live to dead roots may be wrong, and may result in an overestimate of N
loss when expressed on a Ca-basis. Despite this caveat, estimated Ca in dead roots was,

on average, only 11% less than actual Ca in dead roots. Furthermore, AN corrected for
Aroot mass (ANconected) was intermediate between ANlength and ANCa values.

Altogether, my results suggest that mass-, length- and Ca-based estimates of AN all have
their biases, and that mass-based N loss estimates corrected for root mass loss may
provide the best approximation of AN (mean of species = —21%, Table 1.3).
Nevertheless, corrected AN still has limitations given the methodology in this
study. First, there are several potential sources of root length loss unrelated to shrinkage,
such as herbivory, parasitism or decomposition, which could lead to overestimates of root

length shrinkage, and N loss. Furthermore, estimates of root length change were made on

t . . t
15 order roots, but shrinkage may vary among root orders. For example, if 15 order roots

shrink more than the higher order roots that also were included in live and dead root
samples, then length shrinkage and N loss would be overestimated. Related to this, an

additional potential source of error in AN calculations, regardless of the basis measured,

is that live and dead root samples might have contained different proportions of 1“, 2nd

l9

and 3rd order roots, with different diameter distributions. Root order (Pregitzer et al.

1997) and diameter (Gordon and Jackson 2000) are related to N concentrations, thus
differences in live and dead root collections could lead to differences in root N between
live and dead samples that are unrelated to senescence-related AN. However, the
diameter distributions of live and dead root samples were remarkably similar with > 85%
of root length being < 0.5 mm in diameter and none over 2 mm for either live or dead

root collections (Figure 1.1).

Whole plant and ecosystem implications

My data indicate that live root N concentration was a more important determinant of dead
root N concentrations than species and fertilization. Species effects were signiﬁcantly
independent of live root N, but species effects were weaker and fertilization effects were
not signiﬁcant (Figure 1.3 legend, and data not shown). These results suggest that
species and environmental differences in dead root N are mediated primarily by how
species and environment affect live root N and less so by species-speciﬁc or
environmentally induced variation in AN. My results for ﬁne roots are consistent with
those for leaves in a 297 species global dataset (Kobe et al. 2005).

If, as my limited data suggests, the dead root N vs. live root N relationship is
linear and with an intercept of zero, then dead root N is a constant proportion of live root
N at any live root N concentration. It is important to note however, that, although
proportional N loss may be constant, more N on an absolute basis is lost from the dead
roots of plants with high live root N. To reiterate, however, my data are limited, and

conclusions about the determinants of dead root N will require experiments that test these

20

relationships for a larger number of species and across a greater range of environmental
conditions than covered in this study.

Differences in mass-based root N between live and dead roots have been
interpreted as estimates of N resorption or lack thereof (Meier et al. 1985, Nambiar 1987,
Aerts 1990, Gordon and Jackson 2000). My results clearly indicate that it is erroneous to
conclude negligible N resorption based on studies that have used uncorrected mass-based

measures of live and dead root N. For example, using uncorrected mass-based measures

from published studies, I calculated ANmass and in 63% of the estimates, my calculations

suggested that resorption did not occur (Table 1.4). In contrast, when changes in mass

were accounted for (i.e., using estimates of mass loss from this study), my calculations of

ANconected implied that resorption occurred in 15 out of 16 estimates and values indicated

substantial resorption (range = -4.33 to -48.52%) (Table 1.4). Altogether, results from
this study and re-analysis of data from the literature further supports the notion that
previous estimates of root resorption are likely biased, depending on the extent of root
mass loss.

Unlike leaves, differences in root N between live and dead roots do not directly
measure N resorption because other processes, including leaching and microbial
immobilization, can also contribute to changes in N. At best, AN may serve as a crude
index of resorption. If, however, I can assume that AN is a crude estimate of N
resorption, then the moderate resorption values suggested by my results (e. g., 21% for
mass-corrected estimates) is considerably less than foliar resorption values (~ 60%, Aerts

1996, Kobe et al. 2005). For example, reported foliar resorption values for the species

21

included in this study are: P. tremuloides (43%, Killingbeck et al. 1990), B.
alleghaniensis (61%), A. rubrum (71%) and A. saccharum (66%) (Cote et al. 2002).

Post-senescent changes in root N that are independent of resorption pose obvious
challenges to accurately estimating N resorption from ﬁne roots. Although I adhered to
narrow condition criteria for selecting dead roots, senesced roots may have undergone
initial stages of decomposition. During decomposition of ﬁne root litter, N can initially
decrease then increase (John et al. 2002), a pattern that might be explained by leaching
(Chen et al. 2002) followed by microbial immobilization of N (Ostertag and Hobbie
1999). Unlike leaves, which lose negligible amounts of N to leaching (e.g., < 0.6 % of
total leaf N, Chapin and Kedrowski 1983), roots are in direct contact with the soil
solution, which likely facilitates N leaching during root death. Stress-induced loss of
membrane integrity during ﬁne root senescence has been shown to increase the leakage
of N in amino acids (Huang et al. 2005). Even for live, intact healthy roots in aqueous
solution N efflux can exceed inﬂux in some conditions (Lucash et al. 2005, McFarlane
and Yanai 2006) but it is unclear if this occurs for plants growing in soil (McFarlane and
Yanai 2006) and it is only relevant if large amounts are being lost relative to the total
amount of N in live roots.

N losses through leaching were not accounted for but would have been captured
by AN values and ultimately would have over-estimated resorption. Like leaching, N
immobilization would also be captured by AN estimates, but, unlike leaching,
immobilization would result in underestimates of N resorption. Unfortunately, there are

few data on leaching or immobilization per unit root during senescence, let alone studies

22

that simultaneously evaluate the contributions of leaching, immobilization, and resorption
to changes in root N.

Given that previous work generally indicates that mass-based N concentrations
are similar in live and dead roots (Nambiar 1987, Aerts 1990, Gordon and Jackson 2000),
numerous investigations covering a broad spectrum of ecological processes have assumed
that ﬁne root N is not resorbed during senescence. These processes include: ﬁne root N
and decomposition dynamics (Dilustro et a1. 2001, Ludovici and Kress 2006, Valverde-
Barrantes et al. 2007); covariance of foliar and ﬁne root nutrient concentrations (Newman
and Hart 2006); whole-plant and stand-level nutrient-use efficiencies (Silla and Escudero
2004, Norby and Iversen 2006, Silla and Escudero 2006); and stand-level N cycling (Will
et al. 2006). I recognize the strong contributions these and other studies have made
towards understanding these processes and that progress in ecological research often
requires making pragmatic assumptions about processes that are poorly quantiﬁed. By
clearly showing that the assumption of no resorption from roots is erroneous, my aim is
to stimulate new investigations on the N dynamics of senescing roots. I believe that such
investigations will be strengthened by considering N dynamics on bases that are

insensitive to mass changes that occur in roots as they senesce, as I have identiﬁed in this

paper.

23

Table 1.1. Results from ANOVA and summary of ﬁne root nitrogen (N) concentrations.

 

(a) Results of a standard least squares linear model for main effects and interactions of species (n = 4),
fertilization (unfertilized, fertilized) and root status (live, dead) on mass- and length-based ﬁne root N
concentrations.

Tukey-Kramer HSDc

 

ANOVA effects d.f. SS F P
N (mg 5"
Species 3 133.46 10.9 < 0.0001 Pt (ac), Ba (3), Ar (be), As (b)
Fert 1 56.69 13.89 0.0003 fertilized > unfertilized
Species X Fert 3 16.22 1.32 0.2704
Root status 1 168.52 41.28 < 0.0001 dead > live
Species X Root status 3 52.36 4.28 0.0069
Fert X Root status 1 3.12 0.77 0.3836
Species X Fert X Root status 3 16.93 1.38 0.2521
N (mg-cm l)b
Species 3 41.37 67.99 < 0.0001 Pt (a), Ba (b), Ar (c), As (c)
F ert 1 0.29 1.42 0.2400
Species X Fert 3 0.19 0.31 0.8220
Root status 1 1.81 8.92 0.0035 live > dead
Species X Root status 3 0.44 0.72 0.5400
Fert X Root status 1 0.1 1 0.57 0.4500
Species X Fert X Root status 3 0.74 1.22 0.3100

 

2 b 2
Note: aOverall model: adjusted R = 0.44, P < 0.0001; Overall model: adjusted R = 0.64, P < 0.0001

cMeans among species without a common letter are signiﬁcantly different (P < 0.05, Tukey-Kramer
HSD). Species abbreviations are as follows: P. tremuloides (Pt), 3. allegham‘ensis (Ba), A. rubrum
(Ar), A. saccharum (As).

(b) Mean nitrogen (N) concentrations (:t one SE) in live and dead ﬁne roots collected from unfertilized and

-1
fertilized (N + Ca) potted seedlings of four tree species as expressed on root mass (mg N-g root), length

-1
(mg N-cm root) and Ca bases (unitless).

 

 

 

Nmass Nlength
(mg-g ]) (mg-cm 1) NCa (unitless)
Species Unfertilized Fertillized Overall Overall
P. tremuloides
Live 12.9 i 1.3 (6) 14.6 i 0.9 (7) 1.4 i 0.1 (12) 1.4 :t 0.1 (4)
Dead 13.1 i0.8 (7) 16.1i0.6 (7) 1.3i0.1(13) 1.0i0.1(4)
B. alleghaniensis
Live 12.8 :t 0.6 (6) 13.7 i 0.6 (8) 2.3 :1: 0.1 (14) 1.5 :t 0.1 (2)
Dead 14.3i1.1 (7) 15.3:t0.3 (9) 1.8:I:0.1(l6) 0.8:l:0.1 (2)
A. rubrum
Live 12.8 at 0.4 (8) 13.8 :1: 0.5 (8) 3.0 i 0.1 (16) 2.1 :t 0.2 (4)
Dead 17.8 i 0.4 (8) 17.6 i 0.5 (7) 2.8 :t 0.1 (15) 1.4 :1: 0.1 (4)
A. saccharum
Live 13.7 d: 0.6 (8) 16.8 i 0.8 (8) 2.8 i 0.1 (16) 1.8 :1: 0.2 (3)
Dead 17.7 i 0.5 (9) 18.1 at 1.3 (8) 2.6 i 0.1 (17) 1.4 :t 0.2 (2)

 

Species group means were calculated ﬁ'om samples of individual seedlings for Nmass (mg-g_ ) and

-l , . .
Nlength (mg~cm ). Overall specres means for NCa (unitless) were calculated from composrte samples.
Sample sizes are in parentheses.

24

Table 1.2. Results from ANOVA and summary of changes in ﬁne root N (AN) during
senescence.

(a) Results of a standard least squares linear model for main effects and interactions of
species (n = 4) and fertilizer (unfertilized, fertilized) on estimates of AN during

senescence. ANCa was not evaluated with ANOVA due to lack of replication for some
treatments (see methods).

 

 

ANOVA 6
effects d.f. SS F P Tukey-Kramer HSD
ANm... (%)"
Species 3 4519.78 4.99 0.0043 Pt (a), Ba (a), Ar (b), As (ab)
Fert 1 1824.64 6.05 0.0176 fertilized > unfertilized

SpecieSXFert 3 1100.43 1.22 0.3142

b
ANlengtll (0%)
Species 3 1137.99 0.92 0.4403
Fert 1 17.78 0.04 0.8367

Species X Fert 3 1833.23 1.48 0.2331
Note: aOverall model: adjusted R2 = 0.26, P = 0.0023; bOverall model: adjusted R2 =

0.005, P = 0.4165 c:Means among species without a common letter are signiﬁcantly
different (P < 0.05, Tukey-Kramer HSD). Species abbreviations are as follows:
P. tremuloides (Pt), B. alleghaniensis (Ba), A. rubrum (Ar), A. saccharum (As).

(b) Means of AN (i one SE) for unfertilized and fertilized (N + Ca) potted seedlings of
four tree species. Fine root AN estimates were expressed on root mass (mg N-g—lroot),

length (pg N-cm—l root) and Ca bases (unitless). Means of ANnrlass (%) and ANlength (%)
represent the mean of all individual seedlings within a species (overall) or species X

fertilizer group. Species means for ANCa (%) represent the mean of composite samples.
Sample sizes are in parentheses.

 

 

ANmass (%) ANlength (%) ANCa (%)
Species Unfertilized Fertilized Overall Overall Overall

10.0 d: 3.9 -9.1 d: 5.5 -27.2 :l: 1.8

P. tremuloides 13.9 i 8.0 (5) 6.6 i 3.1 (6) (11) (10) (4)
13.9 a: 4.7 —l6.1 :t 7.1 —47.6 i 2.2

B. allegham'ensis 14.8 i 8.8 (6) 13.2 i 5.4 (8) (14) (14) (2)
34.6 i 4.1 -6.3 i 3.4 —32.4 d: 3.7

A. rubrum 40.4 :1: 5.6 (8) 27.9 i 5.2 (7) (15) (15) (4)
21.3 :t 6.0 -5.0 :1: 3.7 —l4.4 i 7.5

A. saccharum 33.7 :1: 5.9 (8) 8.8 a: 8.6 (8) (l6) (l6) (2)

 

25

do .«0 803382 0: $8838 2E3 $08 “005 80¢ 08230—006

.308 000.0 8 02. 80¢ $88 “005 .8“ $500800 223 $6524 80¢ 08250309
.2: x 2&8: 8 omega Eeozoat x ﬁwcﬂ EB Ba 32: E owedzo 1805085 -_ n 3% 8&8 00853.02
.88 080 n MG :08 03. u M‘— .momoﬁeocmm 8 0.8 38m 038mm .masouw 360% 853»

8388 87.8800 M0 8808 88080.— 8038800800 8828 ~08 .«0 .8002 .3008 .580 1 132208 m0 Swan: :88 05 83.082
338:3 ﬁweﬂ 80M .mnsem 360% 853» @8603 8:03:08 .«0 8008 08 06.6qu624 98 Ewen.— HoohmmmE H00.— mo mo=_a>

 

n:
523.:
5 2 an»

6: 31.va

wdml

3T
$3 :3 a E
as :3 u 0..

2:
3 so a we
3 ed a we

R: as a EST

fowl

ET
as :3 a M:
as :3 u. 2

S;
S to u we
C so u 3

Av: 0N H mel

wdml

0.2:
as :3 u 2.
as :3 a 3

3:
3 ed a 3:
:0 to u 2:

2: Z a 3.?

odml

we:
as :3 a we
as :3 a 3

owouooaxnxrlwm8v memo

008038A _ 1w.w8v menu

BSQoEATWwEV «.50
83.339 802

.8 3.2.324
8.8 2 38.09.39

%.XL 308 H0084
as; 582 .88

A885 «Bums—
A88V 559—2

 

2:: we? 0.le mm? 3.0 emeuwaag
:3 ho a a: G: we a of G: :0 u 02 a: no u a.» p-580 85862888
G: :0 u 2: 6: we u «as 0.: 2 a a: a: no u 2: Araaev 5505.386

.8385 3:88.338 30%
szeﬁuaw .V :33»: .V ﬂatﬁxetwmza .m 330359: .& 360mm

 

 

88038800800 «U H08 was 00800888 maize A23 2 ~08 new 8 098:0 0800800 .muwﬁﬁo 7828086 80% .m.~ 030,—.

26

 

0000: :00

 

 

 

 

 

 

 

00ww0_
000w
00002 20.0%-NN 0000000000 008 £05 0000-0000 00008 000— ._0 00 8.0—0000082
00002000 00080.:-
0000 20-0053
000% 20-2-3 0¢ 00082003 000002000 mwa— ._0 00 00.02
000000800.
000% 20-3-3 000030000 00008
00000-0000 000
0000808 20- 0?? 080 000 000000000002 00002000 00000083- N03 ._0 00 3.0—300.002
00000 20 th-mm 000000
00000 20 00-02 00000 80_w_0m 00000-0 0080,0000 mg— ._0 00 M09..— 00>
0005000 00000083-
0000803 003 0200030 500m 33 003802
00000 008 m 00¢ 0:00
0080 008 < 00¢ 2.00 00000 0.0 00800 000026 2.0000 00000-0 00000-0000 «03 ._0 00 80.00.04“
00000083-
00ﬂm\0000% 8 005 0908mm 00008 0:000:08 ”we 0:00—02 000*5—02
00000080,.-
0030 >000m 0008000.; 0030mm 020000-— 080:— 0005040

 

 

.Soom .00300m 000 000000 80¢ 0000003 020000 0002500 80¢ 00.55% 800 00¢ 220000023 000— 0008 00¢
00000000 2 0000 00¢ 8 m0w0000 000 @3223 Z 0000 00¢ 8 m0w0000 00000-0008 a000000000000 Z 0000 00¢ 0000 000 024 .04 030._.

27

acorn—=23 Sm $5508 8mm 38 Z

 

 

 

 

 

 

 

 

«do? a: 25 Rd _ v 25

”3? SN Po who ﬂ v 2%

8.3.. 25 So 85 _ v 2.:

3.21 Ram 21 So :3 g V 2E 39: .3 .0 5223.....5

397 $.01 3:. 3o 858: o o.~-o 2a

3.? «new 2 and E58: 0 93 2e Am:— ._a .9 3E):

3.2: a: S E 2.3. no v 2%

on»? 8.? of 8; so: 328 no v 25

$2.. 3. T a? S 2.3 3 v 2E

3.8: 3.2 R; 3 so: “88.“ 2 v 2E 5:: .3 .0 Qogwsaﬁozv

38: NS £2 :2 _ v 2a

N? o _ .a. 3. 83 mo _ v 2e :33 .3 .9 ”semﬁz

22' who 3; a: no 3 v 26 9:: :3...an

8.3T ST we; N: N v 2%

~23: mm. an 8; 3 :5 3. o. BEE was 538? N v 2E 53 .3 .0 522.3

3.97 3.”: 2... mm; 85-86 SEN v 2E 8%:
5.3.3: as“ age—2v

E; 3.» 3.; Eu 55.8.. 3 3:5 2:9 22::
masochhooZQa mmﬂEZﬂa Emu—u on: AEV Sun—0A— houoaama— 25M—

 

 

.APEOQV v; 033.

28

 

Live Roots: p = 0.36 mm

 

 

 

Dead Roots: p = 0.29 mm
0.8 -

Root length (proportion)

0.6 -

0.4 -

0.2 -

 

 

 

0.0

 

50.5 0.5-1.0 1.0-1.5 1.5-2.0
Root diamter class (mm)

Figure 1.]. Proportion of ﬁne root length across root diameter classes for both live and
dead root categories. Root length data from individual seedlings were combined for all
species and fertilizer treatments.

29

 

 

Live root Dead root

 

 

 

 

 

Figure 1.2. Representative examples of live and dead ﬁne root segments for A.
saccharum. Roots were procured from individual seedlings in late-September 2003 at the
Tree Research Center, Michigan State University. Images were acquired with a ﬂatbed
scanner (Epson Expression 1680, Nagano, Japan) at a resolution of 400 DPI and later
edited for inherent root color differences and shadowing with Adobe Photoshop 7.0
(Adobe Systems Inc., San Jose, California). All roots in this image are < 2.0 mm in
diameter.

30

 

-e- P. tremuloides

"O- A. rubrum 0
Cl A. saccharum

3 I B. alleghaniensis [33%

— all data

  
  

 

 

Dead Root N (pg-cm'1)
N

 

 

 

1 2 3 4
Live Root N (pg-cm'1)

Figure 1.3. Live (LRN) versus dead (DRN) ﬁne root Nlength for four tree species. Data
represent individual seedlings. In a mixed linear least squared model partial P values for
LRN, species and their interaction as predictors of DRN were P = 0.0007. 0.0023, and
0.0614, respectively. Summary statistics for signiﬁcant (P < 0.05) regressions are:
Overall relationship, DRN = 0.332 + 0.765(LRN), R2 = 0.57, P < 0.0001, n = 55; P.
tremuloides, DRN = -0.025 + 0.924(LRN), R2 = 0.50, P = 0.0218, n = 10; A. rubrum,

DRN = -0.162 + 0.993(LRN), R2 = 0.51, P = 0.0029, n = 15.

31

CHAPTER 2

PLANT TRAIT CORRELATES OF WHOLE-PLANT CARBOHYDRATE
STORAGE AND RELATIVE GROWTH RATE IN TEMPERATE AND BOREAL
TREE SEEDLINGS: ARE THERE TRADEOFFS?

ABSTRACT

If interspeciﬁc variation in whole-plant total non-structural carbohydrates (TNpr) is a
major trait underlying trade-offs between growth vs. survival adaptive strategies, then

TNpr must negatively covary with traits that enhance growth potential. I explored this
hypothesis by comparing interspeciﬁc relationships of TNpr, relative growth rates

(RGR), and functionally related allocational, morphological and C02 exchange traits for

seedlings of temperate and boreal tree species grown in low light (2.8% of open sky).

Consistent with their evergreen leaf habit and lack of sprouting ability, gymnosperms (n
= 8) had lower TNpr, and also lower RGR, speciﬁc leaf area and root mass ratio, and

greater leaf mass ratio and leaf production rates than angiosperms (n = 28). Trait
interrelations also differed between groups, so I analyzed interrelations among all species

and for the larger angiosperm group separately. Across all species, over three orders of

magnitude, variance in seed and seedling mass were positively correlated with TNpr

and negatively related to RGR. RGR and TNCw p were negatively correlated but the

relationship was weak and could be driven largely by covariation in plant size. In

contrast, the residuals of the RGR vs. seedling mass regression correlated positively with

TNpr; i.e., when plant size effects on RGR are removed, RGR is positively related to

TNpr. Among physiological and morphological traits, across all species and within the

32

angiosperm group, leaf area ratio correlated most strongly to RGR, root mass ratio was

most strongly related to TNpr, and root mass ratio and leaf area ratio were themselves

negatively correlated indicating a possible necessary trade-off between leaves for
productivity vs. roots for storage over all species. However, lower leaf light
compensation points for photosynthesis, lower leaf production rates and lower whole-

plant respiration rates contributed to greater RGR independent of plant size and/or greater

TNpr, but were only weakly related to RGR. RGR independent of size was also

positively related to seedling survival thus providing further support for the positive
interrelations between growth-survival and carbon conservation traits in low light. In
summary, under low light conditions, carbon conservation traits that increase growth
independent of plant size also increase stored carbohydrates. Trade-offs between growth
and storage (i.e, survival) related traits are only evident when variation in plant size is not

taken into account.

33

Introduction

Allocation to carbohydrate storage (total nonstructural carbohydrates, TNC) has been
proposed as a key plant trait that underlies perennial plant survival (Chapin et al. 1990,
Kobe 1997, Machado and Reich 2006, Myers and Kitajima 2007, Poorter and Kitajima
2007). INC may enhance survival because it can be mobilized in response to tissue loss
(e.g., from ﬁre, Kruger and Reich 1997b); herbivory (Canham et al. 1999, Myers and
Kitajima 2007) and in response to resource shortfalls that limit carbon gain (e. g., drought,
Busso et al. 1990, Volaire I995); dormant seasons (Loescher et al. 1990, Kozlowski
1992); and periods of deep shade (Kobe 1997, Veneklaas and den Ouden 2005, Myers
and Kitajima 2007). Long-term TNC pools can constitute as much as 45% of root mass,
but are highly variable among species (range = <2%-45%) and environments (Marquis et
al. 1997, Gansert and Sprick 1998, Canham et al. 1999, Newell et al. 2002, Gaucher et al.
2005, lyer 2006, Myers and Kitajima 2007, Poorter 2007). In this paper, I explore some
potential sources of interspeciﬁc variation in TNC including (1) phylogeny, (2) seed and
plant size; and (3) growth vs. survival adaptive strategies. Related to (3), I investigate the
popular supposition that there are necessary trade-offs between trait expressions favoring
growth potential vs. storage, and that these deﬁne an axis along which species growth vs.
survival (storage) adaptive strategies vary.

There could be relatively simple and general allocation based trade-offs between
storage and growth related traits, such as allocation to leaves (growth) vs. roots (storage
and survival), but it is possible that other traits, including ones that vary phylogenetically,
may impact the form of these interrelations and as such different patterns could emerge

for phylogenetically broad compared to phylogenetically narrow comparisons. For

34

example, general differences in the leaf habit between gymnosperm conifers (generally
long-lived leaves) and winter deciduous angiosperm species could promote differences in
TNC, their distribution among foliage, stems and roots, and their relationship with
growth-related traits. Winter deciduous species must develop a full canopy of leaves
each spring which may necessitate, compared to evergreen conifers, higher TNC reserves
overall but little TNC in their ephemeral, thin and poorly defended leaves. At the other
end of the comparative spectrum, TNC-growth related trait interrelations among closely
related species may reveal nuanced trade-offs that could be obscured by phylogeny in
comparisons of more distantly related species.

In general, seed size correlates positively with young seedling survivorship
(Leishman and Westoby 1994, Saverimuttu and Westoby 1996, Walters and Reich 2000,
Moles and Westoby 2004) and negatively with relative growth rates (Walters et al.
l993a, Leishman and Westoby 1994, Reich et al. 1998a, Poorter and Rose 2005).
Positive survival-seed size relations have been hypothesized to occur via greater reserves,
lower growth rates (dilution of reserves) and/or greater seedling size for larger seeded
species and some evidence exists for all three (Green and Juniper 2004, Quero et al.
2007). Although not a test of any one of these alternative hypotheses, a logical extension
is that seedlings of larger seeded species have greater TNC. There is evidence that,
within species, larger seedlings have greater TNC (Lusk and Piper 2007), but little
attention has been paid to interspeciﬁc relationships.

Interspeciﬁc variation in TNC has been proposed as a key trait underlying an
adaptive strategy axis deﬁned by trade-offs between high resource growth potential vs.

low resource survival (Kobe et al. 1995, Schreeg et al. 2005, Poorter and Bongers 2006).

35

This supposition has been partially supported by studies showing that that TNC is
positively related to survival (lyer 2006, Myers and Kitajima 2007, Poorter and Kitajima
2007) and negatively associated with growth rates (lyer 2006, Myers and Kitajima 2007,
but see, Poorter and Kitajima 2007). Given these empirical results and supporting theory,
simple trade-offs in the expression of morphological and physiological growth-related
traits may underlie growth vs. storage/survival trade-offs (e.g., allocation to roots
(storage) vs. leaves (growth)) but, to date, these interrelationships have received little
attention.

The survival-growth rate relationship was been most convincingly shown for low
light survival versus high light growth rate across seedlings of species that vary in a wide
range of traits including seed size and phylogeny (Walters and Reich 2000). But, the
shape and direction of growth-survival relations, growth-TNC relations and the traits that
underlie them may depend on the sources of variation examined. For example, within
species, greater light availability can lead to greater growth and survival (Kobe et al.
1995, Walters and Reich 2000) and greater growth and storage (Iyer 2006). Across
species under high light, the TNC-growth relationship could be strongly negative because
variation in growth and storage potentials and their underlying traits could be fully
expressed. Conversely, across species under low light, the TNC-growth relationship
could be less clear because the manifestation of growth and storage capacities are likely
muted (e.g., Walters and Reich I999, Portsmuth and Niinemets 2007) and traits
underlying these patterns (e. g., allocation programs) may be altered in unexpected ways.

If interspeciﬁc variation in TNC is a major trait underlying trade-offs between

growth vs. survival adaptive strategies, then TNC must negatively covary with traits that

36

enhance growth potential. If so, then what might some of these candidate traits be? High
growth potential, here deﬁned as high potential relative growth rate (RGR), requires
allocation to leaves and roots with high resource acquisition capacities, which in turn
requires high surface areas (as indexed by high leaf mass ratios, leaf area ratios, speciﬁc
leaf areas and speciﬁc root areas (Poorter and Remkes 1990, Walters et al. 1993b», and
high metabolic potentials (as indexed by nitrogen concentrations, and respiration rates
(Reich et al. 1998a, Reich et al. 1998b, Reich et al. 2003a, Tjoelker et al. 2005)). In
contrast, the metabolic costs of synthesis (Poorter and Villar 1997) and maintenance
(Kobe 1997) of TNC are low. Furthermore, TNC is not used for resource capture until
mobilized to produce growth-related structural tissue (Chapin et al. 1990), and as such
TNC storage in protected perennial organs with little chance of being consumed or
damaged should be favored over storage in other areas, including those used for resource
acquisition. Given these differences in growth- and storage-related traits, I hypothesize

that allocation to TNC storage in roots and stems occurs at the expense of allocation to

leaves and ﬁne roots, which will lead to lower nitrogen concentrations and C02 exchange

rates (photosynthesis and respiration).

In this paper, I quantiﬁed interrelations of whole-plant total non-structural

carbohydrates (TNpr), RGR and associated morphological and physiological traits for

seedlings of 36 temperate and boreal woody species that were grown in a common low

light environment. I focused my comparisons of these interrelations on the following

questions: (1) Does TNpr and its distribution differ between species groups of

contrasting phylogenies and leaf habits (i.e., winter deciduous angiosperms vs. evergreen

gymnosperms), (2) How do TNpr, RGR, seed and seedling size, and growth rates

37

covary?; (3) Which morphological, allocational and physiological traits are associated

with TNpr and/or with RGR? Based on these comparisons can I identify trade-offs

between characteristics that enhance growth vs. storage?

Materials and Methods

Study species, growing conditions and experimental design

A total of 36 temperate and boreal woody species, mostly from North America, but some
with Eurasian distributions were used in this study (Table 2.1). Species differed in seed
mass, taxonomic orders and leaf habit (broad-leaved, winter deciduous angiosperms and
evergreen (except Larix laricina) gymnosperms), shade tolerance, and drought tolerance.
Seeds used in the experiment were collected from the Beal Arboretum at Michigan State
University (MSU), East Lansing, M1, or purchased from commercial sources (Ontario
Tree Seed, Angus, ON; Lawyer Nursery, Inc., Plains, MT; Ministry of Forests, Surrey,
B.C.; Shefﬁeld's Seed Co., Inc., Locke, NY). Seeds were pre-treated and stratiﬁed
according to Young and Young (1992) and germinated in bench-top trays ﬁlled with
potting soil (Faffard 2 mix, Agawan, MA) underneath a 50% shade lathe house at the
Tree Research Center (TRC), MSU in mid-May 2004. Over a two week interval starting
in mid-June 2004, germinant seedlings were planted into individual poly-coated bleached
board plant bands (7.6 cm X 7.6 cm X 25.4 cm; Zipset Plant Bands, Monarch
Manufacturing, Inc., Salida CO). Plant bands in groups of 8 (large-seeded species) or 16
(small-seeded species) were inserted into milk crates (30.5 cm X 30.5 cm X 27.9 cm) and
species positions within milk crates were randomly selected. Between 40 and 200

seedlings per species were planted resulting in a total of ~2275 seedlings. Seedlings were

38

grown in a 60/25/15% homogenized mixture of a ﬁeld soil mix, silica sand and pea
gravel. The ﬁeld soil mix was collected from the top 15-20 cm of sub-organic soil with a

backhoe at a forested sandy glacial outwash site and a moraine site in Roscommon, MI
(44012' N, 84°36' W) and later combined in equal amounts. Milk crates were randomly
distributed to ﬁxed positions within a 30 m X 50 m area in the understory of a closed

canopy, self-thinning, 39 year old Pinus strobus plantation at the TRC (42°40' N, 84027'

W). The experimental area was fenced with 5 cm mesh welded wire to 1.5 m height and
1.25 cm wire mesh to 1 m height to prevent browsing by mammals. Supplemental water
was only applied to seedlings during extended dry periods (i.e., more than 7 days without
rain). A controlled release fertilizer (Osmocote® Plus by Scotts Fertilizer, Marysville,
Ohio, USA) was applied on 13 July 2004 at the rate of 200 kg N/ha to the soil surface
layer within each plant band. Canopy openness, an index of light availability, was
estimated in mid-August 2004 with paired LAI-2000 plant canopy analyzers (LI-COR,
Inc. Lincoln, NE). Brieﬂy, measurements above each milk crate (n = 232) were obtained
when the sky was uniformly overcast with one LAI-2000 unit, while an identical remote
unit was placed on a tripod in a nearby clearing and simultaneously recorded open-sky
values. Data from each unit were combined later to calculate canopy openness values
and the mean light environment (t SE) for the experimental area was 2.81 d: 0.05 % of
open-sky. Air temperature was recorded with Hobo Tidbit v2 data loggers (Onset
Computer Corporation, Bourne, MA) from 30 June through 27 September 2004 and mean

daily minimum, maximum and average temperatures (i SE) were 13.9 i: 0.1, 24.3 i 0.1

and 18.4 i 0.10C, respectively.

39

Measurements

Seedlings were harvested at three stages during the experiment, with harvest day varying
slightly among species. Harvests, time since transplant and number of seedlings
harvested were: (I) germinant harvest, ~ ﬁve seedlings (mean 5.4, range 3-9, total 194),
just prior to transplant; (2) harvest 1, ~ 8 seedlings (mean 8.1, range 4-16, total 293), 49-
65 days, and; (3) harvest 2, ~7 seedlings (mean 7.4, range 2-18, total 267), 85-102 days.
In this report, germinant harvest mass is sometimes used as a proxy for seed mass, which
can be justiﬁed since seedlings were harvested within a couple days of germination and
germinant mass was strongly correlated with published values of seed size (P < 0.0001, r
= 0.98, data not shown). For the germinant harvest, entire seedlings were dried. For
harvests 1 and 2, seedlings were partitioned into leaves/needles, stems and roots and
dried. Since most cotyledons had started to detach from seedlings by harvest 1, they

were excluded from estimates of whole-plant biomass for harvests 1 and 2. Seedlings
were dried in a forced air-oven at 100°C for 1 hour to quickly stop respiration and then at
70°C for 48 hours, after which the dry mass of each sample was obtained.

Leaf net photosynthesis was measured 25 August through 19 September 2004
from 8:30 to 16:00 local time with a C02 analyzer operating as a closed system (LI-6200,

LI-COR, Inc., Lincoln, NE). The C02 infrared gas analyzer was calibrated daily against

C02 standards. For angiosperm species, photosynthesis was measured on the second or

third fully-expanded leaf as close to their natural orientation as was possible within either
a 0.25 or 4 liter gas-exchange chamber. The small seedling size of gymnosperm species

precluded photosynthesis measurements on intact seedlings. Since photosynthetic rates

40

of isolated foliage do not differ from whole-shoot measures for small conifer seedlings

(Reich et al. 1998b), photosynthesis was measured on conifer shoots. Individual shoots
(stem and needles) or multiple shoots (several individuals) were clipped along the stem
and photosynthesis was measured on intact needle canopies within the 0.25 liter gas-

exchange chamber. All photosynthesis measurements were expressed on a dry mass

basis (nmol C02 g—l s_l). During measurements, chamber air temperature was 26.3 :1:

0.10C (mean d: SE), relative humidity was 57.4 :t 0.4% and ambient C02 was 372.8 i 0.4

ppm. Prior to photosynthesis measurements, photosynthesis was induced by placing
seedlings in naturally occurring sunﬂecks for 5 to 10 minutes. Individual seedlings were
placed within a three-sided enclosure (1 m x 1 m X 2 m) that was covered on the top and

three sides with black shade cloth (~5% ﬁrll sun) and located within the experimental

. . . . . -2
area. This structure marntarned light levels that were consrstently lower than 1 pmol m

-l . . . .
s and a small house fan was used to crrculate arr to maintain temperatures and C02

concentrations within the enclosure that were nearly identical to ambient experimental
conditions. An incandescent lamp was equipped with a dimming device and placed

directly above seedlings to produce photosynthetic photon ﬂux densities (PPF D) that
were < 50 pmol m-2 s—1 at leaf level. The light source was used to develop
photosynthetic light response curves that consisted of 5-10 PPF D levels, starting at values
slightly < 50 pmol m-2 s.l and ending at values > 2 pmol m_2 3*]. This particular light

range was used to provide data for the estimation of leaf-level quantum yield, light
compensation points and photosynthetic rates at a common PPF D (see Parameter

calculations subsection for details). Two or three replicate light response curves were

41

obtained from randomly selected individuals for each species (overall: 10-30 points per
species).

At the end of each day following photosynthesis measurements, seedlings were
harvested and partitioned into leaves/needles, stems and roots to acquire: (1) biomass of
seedling components; (2) images of individual leaves/needles used for gas-exchange; (3)
images of whole-plant leaf/needle canopies; and (3) images of whole-plant root systems.
Digitized root images were manually edited with Adobe Photoshop 7.0 (Adobe Systems
Inc., San Jose, California, see Chapter 1). Leaf images were analyzed for projected leaf
area with WinFolia Pro software (Regent Instruments, Blain, Quebec) for angiosperm
species and WinSeedle software (Regent Instruments, Blain, Quebec) for gymnosperm
species. Edited root images were analyzed for total root surface area with WinRhizo Pro
5.0 software (Regent Instruments, Blain, Quebec).

Prior to sunrise during harvest 2 (24 to 27 September 2004), seedlings randomly
selected for whole-plant respiration measurements were moved to a dark room in a
laboratory on the campus of MSU. Intact seedlings were harvested and root systems

were rinsed with deionized water to remove soil prior to respiration measurements.
Whole-plant dark respiration (pr) was measured at 250C i 0.03 SE with a C02 analyzer
operating as a closed system (LI-6200, LI-COR, Inc., Lincoln, NE). In order to obtain
adequate changes in C02 concentrations over the course of a measurement interval, 1 to 4

intact seedlings were placed in either a 0.25 or a 4 liter gas-exchange chamber, depending
on individual plant size. Seedlings were allowed to stabilize for 2-5 minutes before

measurements were recorded. Between 2 and 4 replicate measurements were obtained

42

. . . . -l
for each specres and respiration rates were expressed on a dry mass basrs (nmol C 02 g

—l

5).

Total non-structural carbohydrate and nitrogen concentrations

Individual seedling tissue samples from harvest 2 were aggregated by tissue type (i.e.,
leaves, stems, roots) for each species. Aggregated tissue samples were pulverized into a
ﬁne powder with a ball mill (Kinetic Laboratory Equipment Company, Visalia,
California) prior to TNC and N analyses. TNC was quantiﬁed using a modiﬁcation of
Roper et al. (1988) and Marquis et al. (1997). This procedure involved a two-stage
analysis with an extraction of soluble sugars from the plant tissue followed by starch

analysis of the extraction residues. Approximately 15-20 mg of each aggregated tissue

sample was extracted three times at 75°C using 2ml of 80% ethanol and then centrifuged

at 1900g for 5 minutes. The supematants were collected and diluted with 6 ml of
deionized water. Concentration of soluble sugars (i.e., glucose equivalents) in extracts
was measured at 490 nm with a visible spectrophotometer (Spectronic 20D+, Therrno
Scientiﬁc, Waltham, MA) using a phenol-sulfuric acid colorimetric assay (Dubois et al.

1956). The pellet remaining after ethanol extraction was dried and then gelatinized by
autoclaving at 125°C for 10 minutes along with 2 ml of 0.1 M sodium acetate buffer, pH
4.8. After cooling, samples were incubated with ~60 units of amyloglucosidase from
Aspergillus niger (Sigma-Aldrich, St. Louis, M0) at 55°C for 3 hours. The extract was

analyzed colorimetrically for starch using a glucose-speciﬁc trinder reagent (Pointe

Scientiﬁc, Inc., Canton, MI). Absorbance was measured at 505 nm with a UV

43

spectrophotometer (Lambda 20 scanning spectrophotometer, Perkin-Elmer, Waltham,

MA). TNC concentrations (mg g_l dry mass) for aggregated tissue samples were

calculated as the sum of glucose equivalent measures for soluble sugars and starch.
Lastly, mass-based N concentrations of aggregated tissue samples were assessed with dry
combustion gas-chromatography (NA 1500 elemental analyzer, Carlo-Erba, Milan, Italy)

for each species.

Parameter selection and calculations
A priori, I selected six morphological traits (leaf mass ratio, LMR; speciﬁc leaf area,
SLA; leaf area ratio, LAR; stem mass ratio, SMR; root mass ratio, RMR; speciﬁc root

area, SRA), two allocational traits (leaf partitioning ratio, LPR; root partitioning ratio,

RPR), ﬁve C02 exchange traits (leaf-level photosynthesis, A301,; whole-plant
photosynthesis, A30wp; leaf-level light compensation point, LCP; quantum yield, QY;
whole-plant dark respiration rate, pr), and one physiochemical trait (whole-plant N

concentration, pr) to relate to TNpr and RGR (described following and Table 2.1).

Traits chosen were ones that have been theoretically and empirically related to RGR
and/or TNC (Walters et al. 1993b, Reich et al. 1998b, Poorter 2001, Kobe et al.

unpublished manuscript). Hereafter, to simplify presentation, this group of traits, and

TNpr and RGR will be referred to collectively as functional traits.
LMR (leaf mass/total plant mass, in g g_l), SMR (stem + petiole mass/total plant

mass, in g g-l) and RMR (root mass/total plant mass, in g g_') were calculated for both

44

harvest 1 and harvest 2 data. SLA (leaf area/leaf mass, in cm2 g-]) was determined on
plants harvested during C02 exchange measurements (between Harvests 1 and 2). These

values were used to calculate leaf area ratio (LAR; leaf area/total plant mass, in cm2 g—l).
for both harvest 1 and harvest 2 mass data as LMR x SLA = LAR. Similarly, root
surface area determined during gas exchange measurements was combined with biomass
data from this harvest to calculate SRA (root area/root mass, cm2 g_I ).

In contrast to LMR and RMR, which are static descriptors of biomass fractions,
the allocational traits LPR and RPR ( leaf partitioning ratio, A leaf mass/A total plant
mass; root partitioning ratio, A root mass/ A total plant mass, respectively, in %) capture
the dynamics of newly produced biomass fractions during deﬁned growth intervals
(Poorter 2001). LPR and RPR were calculated for the harvest 1 to harvest 2 interval.
Since our harvest interval was from early August to late September, the negative LPR
values we calculated for eight of the 36 species likely resulted from the initiation of leaf

senescence in some angiosperm species or from the loss of cotyledons in conifer species.

In these cases, species with negative LPR values were assigned a LPR value of 0.
The leaf-level C02 gas exchange vs. PPFD relationships were within the linear

portion of the response curve (<2 to <50 PPF D), so they were ﬁtted with least squares

simple linear regression (JMP 4.0, SAS Institute, Cary, North Carolina, USA). The

resulting species-level ﬁts from these models (P < 0.0001, for all; R2 range = 0.869 -
0.992, see Appendix, Table A. I) were used to estimate photosynthesis at a PPFD of 30

-2 -l . .
pmol m s , a level commonly observed in temperate forest understorres (Weber et al.

1985, Sipe and Bazzaz 1995). Photosynthesis at 30 pmol m—2 s—I PPFD was expressed

45

on both a leaf mass basis (A30L, nmol C02 (g leaf l) s") and a plant mass basis (A30wp,

nmol C02 (g whole-planf') s"), where Awp30 = A30L x LMR). The slope of species-

level A-PPF D ﬁts is the apparent quantum yield of photosynthesis (QY, unitless), and the

Y-intercept of this ﬁt is the leaf-level light compensation point for photosynthesis (LCP,
PPFD at which net photosynthesis = 0). Since whole plant respiration rate (pr, in nmol
C02 g.1 s") was measured on whole seedlings it was calculated from whole-plant C02
exchange and dry mass data.

Average relative grth rate (RGR, mg g—1 d_l) was calculated as: (ln[mean
biomass at harvest 2) — ln[mean germinant biomassD/days (Evans 1972), where days
ranged among species from 85-102. From TNC concentrations for leaves (TNC|eaf),
stems (TNCstem) and roots (TNCroot), we calculated whole-plant TNC concentrations
(TNpr) as: TNpr = (TNcleaf x LMR) ‘+ (TNCstem x SMR) + (TNCroot x RMR).
Similarly, whole-plant N concentrations (N wp) were calculated as: Nw p = (N|eaf >< LMR)

+ (Nstem >< SMR) + (Nroot >< RMR). Whole-plant TNC pool distribution (%) among

organs was calculated as: (organ-level TNC pool size/whole-plant TNC pool size) X 100.

Statistical analysis
All analyses were completed with JMP statistical software (SAS Institute, Cary, North

Carolina, USA). For all analyses, species means were considered experimental units.

Due to distribution characteristics all traits except RGR and TNpr required Logo

46

transformation to normalize distributions in order to satisfy the assumptions of least

squares methods.
Differences in functional traits, TNpr, RGR and size between angiosperm and

gymnosperm groups were compared with t-tests (Table 2.1). Many of the plant traits
were related to mass at the time of measurement for both angiosperm and all species data
sets (Table 2.2), and, on average, gymnosperms seedlings were smaller than angiosperms
(Table 2.1). These factors led me to compare gymnosperm and angiosperm groups
normalized for mass by comparing partial P-values for taxonomic order (angiosperms,
gymnosperm) in models also including mass, and presenting least squares means adjusted
by mass. Based on the results of gymnosperm-angiosperm comparisons above (see Table
2.1 for results) subsequent analyses were conducted for three data sets, angiosperms (n =
28), gymnosperms (n = 8) and all data (n = 36). Data were also analyzed for Quercus
spp. , the most well represented genus (n = 9). For the sake of brevity, analyses of groups
with limited sample sizes (gymnosperm and Quercus spp.) are only presented when they
provide unique insight to the overall analysis. To analyze differences in the distribution
of TNC between gymnosperms and angiosperms, ANOVA was used to evaluate TNC as
a function of order, organ type (leaves, stems, roots) and their interactions. For
signiﬁcant nominal effects (P S 0.05), treatment means were compared with Tukey-
Kramer HSD for organ types within orders and Student’s t test for orders within organ

type.

TNpr, RGR, plant sizes and plant trait interrelations were ﬁrst examined with

Pearson correlations. Due to the potential inﬂuence of plant size on both TNpr and

RGR (MacFarlane and Kobe 2006), the relationship between residual values for

47

regressions of RGR and TNpr vs. plant mass were generated for both all data and

angiosperm data sets. Both of these residuals and raw values of RGR and TNpr were

correlated with plant traits. Correlations between these residuals and plant traits can be

interpreted as the correlation between TNpr or RGR with the plant trait independent of

plant mass effects. I compared correlations of residuals for TNpr and RGR vs. mass

for each harvest with plant traits, and patterns were similar among harvests (data not

shown). For further analyses with plant traits, I used the residuals of germinant mass

with RGR and the residuals of Harvest 2 mass for TNpr. Justiﬁcation for this

approach includes: (1) brevity, (2) residual RGR values are thus expressed as
independent of germinant mass and thus independent of seed size and size at the

beginning of the interval used to calculate RGR, and (3) Harvest 2 mass was the harvest
at which TNpr was determined.

Based on the results of correlation analyses, I developed multiple regression
models of RGR and TNpr using combinations of plant traits as predictors. Models

were developed by ﬁrst including the strongest bivariate predictor, then adding the

variable with the second strongest bivariate predictor, and its interaction. If the added
variable and its predictor did not both improve the adjusted R2 and have a signiﬁcant
partial P (at P < 0.10) then it was removed. This process was continued iteratively until

all plant traits were examined. Models were developed for RGR and TNpr both with

and without mass (germinant harvest mass and harvest 2 mass, respectively) as the ﬁrst

added predictor so as to provide the multiple regression equivalents of correlations with

48

residuals of regressions of TNpr and RGR with plant mass. In addition to multiple

regressions with RGR and TNpr as predictor variables, I also developed multiple
regression models of RMR at harvest 2 with RMR at harvest 1 as the ﬁrst added
predictor. I did this because RMR was overwhelmingly the best predictor of TNpr

with no other trait contributing extra explained variance, and RMR increased between

harvest 1 and harvest 2. Thus, modeling increases in RMR as a function of plant traits

can provide additional insight on the contribution of plant traits to increases in TNpr.

Results

Functional trait comparisons for gymnosperms and angiosperms

Among the 18 traits measured (Table 2.1), germinant mass varied the most (2,874—fold)

and pr the least (two-fold). Reﬂecting the low light growth environment, RGR was

low overall but only one species Quercus phellos, had negative RGR. TNpr varied 13-

fold across all species but varied less within angiosperm (ﬁve-fold) and gymnosperm

groups (two-fold) as the orders formed distinct groups (Student’s t-test) with
gymnosperms having 1/5 the TNpr of angiosperms. Compared to angiosperms,
gymnosperms also had, on average, similar RGR, higher LMR and LPR, lower RMR and
RPR, lower SLA and a slightly lower LAR, slightly higher SRA, lower A30L and slightly
lower A30wp, similar QY but higher LCP, higher pr and slightly higher pr.

However, gymnosperms had, on average, lower mass than angiosperms and several

functional traits varied strongly with mass (Tables 2.2, 2.3) such that differences in

49

ﬁmctional traits between angiosperms and gymnosperms could be driven by differences
in mass. Normalized by covariation in mass, the general differences (but often not the

magnitude of differences) between gymnosperms and angiosperms were preserved for

LMR, LPR, RMR, RPR, SLA, LAR, A30L, A30wp, LCP, and TNpr. Particularly for

LAR and A30wp. the magnitude of differences at a common mass were much greater than

for comparisons of raw data means between groups with angiosperms having much
greater values. Differences in direction and/or signiﬁcance for gymnosperm vs.

angiosperm comparisons at a common mass as compared to raw data include: for

gymnosperms, lower RGR, lower SRA, modestly lower QY, and no difference in pr or

pr.

In a mixed model, plant order and plant organ (leaves, stems roots) had strong
interacting effects on TNC (P < 0.0001) revealing differences in TNC distribution among
organs between plant orders. For angiosperms, TNC concentrations ranked leaves <
stems < roots, whereas for gymnosperms TNC concentrations were much lower and did
not differ among organs (Figure 2.1a). Calculated as the product of organ-based TNC
concentrations and mass fractions, TNC pool partitioning differed between plant orders
for roots and leaves but not stems with angiosperms having a greater proportion of the
total TNC pool in roots (> 70% vs. 35%) and gymnosperms having a greater proportion

in leaves (> 50 % vs. ~7%) (Figure 2.1b).

50

Interrelations of plant size, T NC in: and RGR

Seedling mass right after germination at the beginning of the experiment (Germinant

mass) correlated strongly with mass (Final) approximately three months later at the end
of the experiment (Table 2.2), a pattern explained by large variation in germinant mass
combined with low RGR (Table 2.1) resulting from the low light environment in which

seedlings were grown. Across all species, RGR was strongly negatively related to mass,
especially germinant mass, and TNpr was positively related to mass, especially ﬁnal
mass which was when TNpr was determined (Table 2.2, Figure 2.2a,b). However,

despite a generally strong relationship overall, angiosperms and gymnosperms had
different relationships with lower RGR for a given germinant mass for gymnosperms
(Table 2.1, Figure 2.2a). Despite these differences, RGR consistently declined with

germinant mass for angiosperms, gymnosperms and Quercus spp. (Figure 2.2a). For

TNpr, the inclusion of gymnosperms (i.e., all species data) strengthened the positive
relationship between TNpr and ﬁnal mass, but this was due to low ﬁnal mass for
gymnosperms as TNpr was unrelated to ﬁnal mass within the gymnosperm group

(Figure 2.2b). TNpr was negatively related to RGR for the angiosperm group, but-the
relationship was weak and could have been driven by the combination of positive size-
TNpr covariation and negative size-RGR covariation (Figure 2.2c). Furthermore,
TNpr-RGR correlations were insigniﬁcant for gymnosperms and strongly positive for

Quercus spp. seedlings which varied little in size (Table 2.1, Figure 2.2c). To remove

plant size effects from the relationship between RGR and TNpr, I correlated the

51

residuals of the germinant mass—RGR relationship with TNpr and found a positive

signiﬁcant relationship for all data and a positive, but insigniﬁcant (P = 0.106)

relationship for angiosperms (Table 2.2, Figure 2.2d, inset).

Relationships of functional traits with size, RGR and T NC in)

Most functional traits were strongly related to seedling size (Table 2.3). Relationships
were generally a little stronger for ﬁnal mass when most of the traits were measured than
for germinant mass, but overall relationships were similar, likely due to the strong
correlation between germinant mass and ﬁnal mass (Table 2.2). Differences between
correlations for all data and angiosperms are due to fundamentally different interrelations
for some characteristics and/or differences in average mass between the two groups
(Table 2.1 and data not shown). For angiosperms, negative relationships with mass were
strong for morphological traits including biomass fraction traits (LMR, RMR), leaf

morphology (SLA), and especially surface area to mass ratios (LAR, SRA). They were

also negative for indices of metabolism including pr and pr and photosynthetic traits

(QY, LCP, A3013 A3OWP)-

For all species, angiosperm and gymnosperm data sets, LAR was the single trait
most strongly related to RGR and the form of the relationship was similar for all groups

(Table 2.3, Figure 2.3a). SLA, a component of LAR, and SRA, like SLA and LAR a

measure of surface area per unit mass, were also closely and positively related to RGR
(and negatively related to mass). Whole plant photosynthetic rate (A3owp), a

physiological manifestation of LAR, was also strongly related to RGR (Table 2.3).

52

For all species, angiosperm and gymnosperm data sets, RMR was the single trait
most strongly related to TNpr and the positive relationship was consistent for all
groups (Table 2.3, Figure 2.4a). Many of the traits negatively related to size and

positively related to RGR were negatively related to TNpr. Those traits that followed

this pattern for both angiosperm and all species groups included LMR, SMR, RMR,

SRA, RWP, and QY. LPR, RPR, and LCP were related to TNpr, but unrelated to RGR

and more weakly related to mass than to TNpr (Table 2.3, Figure 2.4b,c). Thus, lower
relative allocation to leaves, greater allocation to roots and maintaining positive
photosynthesis at lower PPF D contributed to higher TNpr, but not to lower RGR. For

both angiosperm and all data groups, RGR independent of mass (i.e., the residuals of the
RGR vs. germinant mass relationship) correlated negatively with LPR and LCP and

positively with RPR (Table 2.3, Figure 2.3b,c). The only variable signiﬁcantly correlated

with TNpr independent of whole-plant mass effects on TNpr was RMR (Table 2.3).

Thus, at any given size, species with greater RMR had greater TNpr.

In multiple regressions, the only trait that contributed to explained variance in
RGR over and above LAR was SRA and this was only for the all data group (Table 2.4).
For models of RGR with germinant mass as the ﬁrst term in the model (i.e., RGR
independent of size effects), SLA was most important additional predictor for the all
species data group and the addition of either LCP or LPR to SLA and germinant mass
explained additional variance in RGR. For angiosperms, SLA was unimportant. Instead,
LPR and LCP added to models already including germinant mass explained additional

variance in RGR. For both all data and angiosperm data groups, no functional trait

53

explained variance in TNpr over and above that explained by RMR. RMR increased

for all species from harvest 1 to harvest 2 (data not shown). Modeling harvest 2 RMR by

ﬁrst including harvest 1 RMR as a predictor provides insight into the factors responsible
for increases in RMR (and thus TNpr). The factors that increase RMR between
harvests 1 and 2 were generally the same that contributed to increased RGR independent
of mass; LPR and LCP. In addition, pr explained additional variance in harvest 2
RMR for the angiosperm group. Substituting RPR for LPR in models where LPR was
signiﬁcant yielded similar, but slightly weaker results. Thus, TNpr increases with
RMR, and RMR and RGR independent of mass increase with lower LCP and LPR, and

for angiosperm RMR, lower pr.

Discussion
Angiospenn and gymnosperm seedlings differed strongly in TNpr and TNC

distribution among leaves, stems and roots. TNpr was markedly higher in angiosperms
than gymnosperms, which is consistent with patterns found for root TNC concentrations
of four temperate species (Kobe 1997) and TNpr of three cold-temperate species

(Machado and Reich 2006). Root systems dominated whole-plant TNC pools for
angiosperms due to a combination of high root TNC concentrations and high RMR. In
contrast, for gymnosperms a majority of the overall low whole-plant TNC pools were in
leaves due to high LMR, but not higher leaf TNC concentrations as concentrations were

similar among organs. Strikingly different TNC patterns for gymnosperms likely reﬂect

54

two general differences between the species representing these groups; differences in leaf
habit, and differences in sprouting ability. Except for Larix laricina, gymnosperms were
all evergreen and all angiosperms were winter deciduous. Following complete leaf
senescence in the autumn and winter dormancy, angiosperms have to mobilize TNC
reserves from stems and roots early in the spring in order to initiate carbon gain through
new leaf production and subsequent photosynthesis (Teng et al. 1999). By contrast, the
existing evergreen needle cohorts of gymnosperms are able to support early growing
season photosynthetic carbon gain for later needle production obviating the need to store
large amounts of TNC to develop a canopy. In this study the higher LPR of
gymnosperms and the negative relationship between LPR and storage across all species
indicates that gymnosperms continue to develop a leaf canopy late in the growing season
at the expense of allocating carbon to TNC. Resprouting of lost aboveground stern tissue
is nearly ubiquitous among angiosperms but rare among gymnosperms (Del Tredici 2001,
Bond and Midgley 2003). Perhaps especially for angiosperms associated with
environments with a high probability of aboveground damage or death (e.g., ﬁre prone
and herbivore modiﬁed ecosystems ), there may be a selective premium placed on TNC

storage, especially in roots, which would allow vigorous resprouting following

aboveground tissue loss. Similar TNpr for winter deciduous Larix Iaricina and the

evergreen gymnosperms is somewhat surprising, given that this species must completely
replace its leaves on an annual basis, but it may reﬂect a low sprouting capacity which

has been reported for the congener Larix kaempferi (Shibuya et al. 2007).
In addition to differences in TNpr between plant orders, differences in leaf

habit and related traits led to fundamentally different relationships between functional

55

traits and RGR for gymnosperms and angiosperms. Gymnosperrns had lower A30L and

SLA, but a higher LMR than angiosperms with the patterns in leaf traits consistent with

the leaf lifespan differences between these groups (Reich et al. 1999). This resulted in
gymnosperms having higher RGR at a given A30L and SLA given their higher LMR and

vice versa. However, because LAR integrates SLA and LMR, LAR relationships with
RGR had the same form across groups. At a similar mass, despite two-fold greater LMR
for gymnosperms, SLA was > three-fold greater for angiosperms resulting in greater
LAR and RGR for angiosperms. Differences in LAR between similar sized ﬁrst-year
seedlings of angiosperms and gymnosperms may dissipate as seedlings get older as
evergreen gymnosperms will continue to accrue new foliage cohorts while retaining at
least one older cohort (Reich et al. 1999), whereas angiosperms will not.

Across all species, LAR was the single trait most closely related to RGR, and in
combination, only SRA described additional variance in RGR. It should be noted that the
strength of the relationship between LAR and RGR could be artiﬁcially inﬂated due to a
statistical artifact because the same whole-plant mass values were used to calculate LAR
(X axis) and RGR (Y axis) (Prairie and Bird 1989, but see, Berges 1997). However, this
statistical difficulty does not invalidate the importance of this relationship because LAR

is one of the theoretical determinants of RGR (i.e., RGR = LAR X net assimilation rate,
Evans 1972). Whole-plant photosynthetic rate (A30wp) a physiological manifestation of

LAR (Walters et al. 1993b, Kruger and Volin 2006), was also strongly related to RGR.
These functional traits and RGR were also among the most strongly negatively related to
mass. Collectively, the strong negative relationship between RGR and size can be

explained by the necessary decline in resource acquiring surfaces (root surface area and

56

leaf area) as a proportion of total mass as structural and support tissue (and storage) in
stems, and higher order roots increases (Givnish 1988). It was not surprising that
variation in RGR independent of size effects (i.e., residuals of RGR vs. germinant mass)
was unrelated to the functional traits that were themselves strongly size dependent and
strongly related to raw RGR data as there was little residual variation in these traits
independent of mass. In other words, at any given mass, there was little variation in LAR
to explain variation in RGR at the same given mass. Instead, light compensation point
(LCP) and leaf and root partitioning ratios (LPR, RPR) which were unrelated to raw RGR
values and more weakly related to mass than many other functional traits, were the most

strongly related functional traits with RGR-mass residuals.
While I show that TNpr is related to mass, it was the stronger relationship

between mass and RMR that drove this relationship, as RMR was the strongest single
variable related to TNpr and the only trait that explained additional variation in
TNpr over and above mass effects. Thus, bigger seedlings tended to have greater

fractions of total mass in roots and greater TNpr, but RMR and TNpr varied

independent of size. TNpr, RMR and RGR independent of mass were all associated

with mostly the same set of carbon conservation traits (lower values for new leaf

production, light compensation points for photosynthesis and whole plant respiration).

This combined with the weak positive relationships between TNpr and RGR

independent of mass (and the strong positive relationships within Quercus spp which

varied little in initial mass) suggests that for young tree seedlings in low light; (1) growth

independent of mass and TNpr are positively related, (2) carbon conservation traits

57

lead to greater growth independent of mass and TNpr, (3) these carbon conservation

traits are not the same ones driving growth in high light (Walters et al. 1993b, Kruger and
Volin 2006) and/or driving growth when differences in initial mass are not taken into

account. Thus, seedlings with high storage capacity (i.e., large germinant size, high
TNpr) were still “growing” and accumulating biomass, but newly acquired

photosynthates were preferentially allocated to TNC in stems and roots, instead of to

growth-related structural components (e. g., leaves). The particular relationships I found
among TNpr, RGR and functional traits might be restricted to low light environments

like the one I used (mean = 2.81 % canopy openness). Traits such as LAR have been
shown to have a diminishing effect on growth as light decreases which might result in the
increased importance of other traits to RGR (Walters and Reich 1999, Portsmuth and
Niinemets 2007). However, even in my low light environment, when initial mass was
unaccounted for, LAR was the single most important driver of RGR. F urtherrnore, these
results suggest that indeterminate (i.e., continuous production of leaves, high LPR) vs.
determinate growth patterns (low or zero LPR values later in the growing season) may
distinguish species that store a little vs. a lot of TNC (Kays and Canham I991, Kobe
1997, but see Canham et al. 1999).

A strikingly strong general result of this study is the predominant inﬂuence that
seed size has on plant characteristics. While some of these relations may represent
necessary allometric constraints (e.g., biomass fractions) it is also possible that strong
relationships between functional traits and seed mass represent selection for combinations
of traits including seed size that confer greater ﬁtness in a given set of environmental

conditions (Wright and Westoby 1999, Reich et al. 2003b). For example, large seed mass

58

was associated with large seedlings, with greater RMR, lower LAR and growth rates and

higher TNpr. Attaining a large stature as a juvenile tree is a function of initial

germinant size and growth rates and growth rate and seed size are often inversely related
(Walters and Reich 2000, Green and Juniper 2004, this study). Yet, after nearly three
months of growth, ﬁnal mass was strongly related to germinant mass (and seed mass),
with very similar rankings among species (Spearman’s p = 0.97, data not shown). In
another study (Chapter 3), large seeded species had larger seedlings even after over two
years of growth independent of resources and that these seedlings have deeper roots and
during drought have greater access to water and greater survival. Other studies show
increased representation of large seeded species as aridity increases (Wright and Westoby
1999). Thus, the combination of traits large seedlings have may confer greater survival
under low resource conditions and thus may be under similar selection pressure rather
than merely being allometrically constrained.

My results may help to reconcile some of the equivocal data on TNC-plant size

relations that have been reported in previous studies. The positive TNpr-germinant

mass (and ﬁnal mass) relationship I found contrasts with Myers and Kitajima (2007),
who found that across a more limited number of species (n = 7) TNC concentrations and

pool sizes in stems and roots were not correlated with seed mass for tropical tree

seedlings. The positive TNpr-germinant mass association in our study may provide an
explanation for the positive relationship between seed size and early seedling survival
(Walters and Reich 2000), as it has been hypothesized that over the short term storage

reserves from large seeds could be mobilized to sustain metabolic activity and may

replace tissues that are lost to herbivores or pathogens (Leishman and Westoby 1994,

59

Green and Juniper 2004). However, this is merely speculation as I did not determine

TNpr for young gerrninants and TNpr was more a function of RMR than mass.

Within a narrow light range (2-5% of open sky), Lusk and Piper (2007), found that
TNpr in large seedlings (400-600mm) was higher (22% of dry mass) than those of

small seedlings (40-60mm, 14% of dry mass) of six broad-leaved evergreen species, but

this difference was driven by light demanding taxa (Aristotelia chilensis, Northofagus

dombeyi, Eucryphia corditolia). Based on the positive association between TNpr and

growth rates in our study, it is probable that the light demanding species in Lusk and
Piper (2007) had higher growth rates, which over time led to the accumulation of more
TNC in larger seedlings. In a complementary study with some overlapping species and
similar growing conditions (2-5% of open sky), Lusk (2004) found that two of the more
light demanding species (Aristotelia chilensis and Eucryphia corditolia) sustained higher
growth rates than species with higher shade tolerance early in ontogeny, but this pattern
reversed in later stages of ontogeny (i.e., size X species interaction). A model of
carbohydrate allocation predicts that small plant size is an outcome of allocation to TNC
(Kobe, 1997) and Machado and Reich (2006) found that whole-plant and tissue-level
TNC concentrations generally decreased with increasing size within three cold-temperate
sapling species. However, in Machado and Reich (2006), plant mass was associated with
age and saplings varied widely in age (range = 6-24 years), which may partly explain the

negative relationship between TNpr and plant mass. Unlike TNpr, mass-based

whole-plant and tissue-level respiration rates increased with plant size/age, presumably

due to higher costs of protein turnover in older saplings (Machado and Reich 2006).

60

Therefore, it is possible that larger saplings in this experiment actually allocate similar or
higher amounts of photosynthates to TNC, but higher respiration rates may ultimately
reduce TNC concentrations.

Collectively, our observations along with empirical data from other studies imply
that there is a trade-off between high storage capacity (and associated carbon
conservation traits) and allocation of carbohydrates to structural components for the
interception of light (i.e., leaves). It is hypothesized then that allocation to TNC enhances
survival in stressful environments but likely compromises growth capacity and thus
competitive ability in high resource environments. The evaluation of the TNC-survival
versus growth capacity trade-off requires one important caveat and that is I did not
explicitly examine this life history trade-off because I did not measure growth rates
and/or competitive ability in a comparable high light treatment. However, I did ﬁnd a

positive relationship between survival (over the three months of the experiment) and

TNpr (Figure 2.5) thereby providing additional support for the positive interrelations

between growth-survival and carbon conservation traits in low light environments.

61

Table 2.1 Summary of seedling characteristics for angiosperm and gymnosperm species.

 

 

 

 

 

 

seed mass ﬁnal mass RGR LMR SMR
-1 -1 -| -1 -1 -1
(mg) (mg) (mg g d ) (g g ) (g g )
Angiospenns
Acer negundo 21.5 174.5 20.3 0.25 0.36
Acer rubrum 5.5 127.8 31.1 0.33 0.28
Acer saccharinum 73.9 1067.9 25.2 0.25 0.35
Acer saccharum 45.7 498.6 23.7 0.25 0.21
Aesculus glabra 2796.6 3419.4 2.0 0.1 l 0.15
Aesculus hippocastanum 4023.7 5301.0 2.6 0.17 0.26
Ailanthus altissima 12.1 121.6 22.9 0.27 0.21
Alnus incana 2.3 26.0 25.1 0.40 0.33
C arya tomentosa 1435.1 1689.0 1.6 0.23 0.08
Catalpa speciosa 13.7 277.7 30.1 0.30 0.30
C ornus amomum 6.0 96.5 28.0 0.29 0.27
Cornus sericea 4.2 81.2 30.6 0.33 0.21
Gleditsia triacanthos 96.6 754.9 19.4 0.20 0.35
Juglans cineraea 1285.1 4592.3 12.6 0.20 0.31
Lindera benzoin 43.6 380.0 21.6 0.27 0.13
Platanus occidentalis 1.8 19.6 24.7 0.48 0.25
Quercus alba 557.0 1 126.3 6.6 0.21 0.1 1
Quercus bicolor 21 13.4 2776.9 2.9 0.27 0.20
Quercus coccinea 1079.9 2088.3 6.5 0.31 0.12
Quercus macrocarpa 732.5 1302.3 5.8 0.28 0.1 1
Quercus phellos 491.4 489.7 0.0 0.33 0.20
Quercus prinus 1906.9 2203.6 1.4 0.33 0.13
Quercus robur 880.8 1 138.0 2.6 0.20 0.1 l
Quercus rubra 1863.1 2184.1 1.6 0.31 0.16
Quercus velutina 1327.9 1894.2 3.6 0.34 0.13
Rhus typhina 6.8 74.8 24.2 0.33 0.31
Robinia pseudoacacia 27.7 136.6 15.8 0.25 0.34
Ulmus americana 10.1 84.0 21.0 0.36 0.25
Mean 1 l 1.6 492.2 14.8 0.27 0.20
LS mean (for mass) 17.1 0.29 0.22
Gymnospenns
Abies amabilis 20.4 39.3 6.5 0.49 0.20
Abies concolor 13.1 42.5 1 1.8 0.53 0.21
Larix laricina 1.6 13.2 21.5 0.55 0.25
Picea stichensis 1.4 9.0 18.6 0.48 0.29
Pinus nigra 13.3 59.2 15.0 0.52 0.22
Pinus ponderosa 25.0 84.1 12.1 0.57 0.22
Pinus strobus 8.7 31.6 12.9 0.58 0.25
Pseudotsuga menziesii 7.5 29.4 14.1 0.53 0.21
Mean 7.8 31.3 14.0 0.53 0.23
LS mean (for mass) 5.8 0.42 0.18
P, t-test 0.007 <0.001 0.86 <0.001 0.458
Partial P, model incl. mass <0.001 0.001 0.268

 

62

Table 2.1 (cont’d).

 

 

 

 

 

 

RMR LPR RPR SLA LAR SRA
-1 -1 2 -l 2 -1 2 -1
(g g ) (%) (%) (cm 3 ) (cm g ) (cm s )
Angiosperms
Acer negundo 0.39 0.0 101.5 431.6 148.4 567.6
Acer rubrum 0.38 9.1 50.0 365.4 154.8 373.2
Acer saccharinum 0.40 12.0 51.6 440.0 137.9 286.2
Acer saccharum 0.54 9.5 69.9 368.4 124.0 326.8
Aesculus glabra 0.74 0.0 100.3 347.4 35.5 51.3
Aesculus hippocastanum 0.57 13.1 65.1 256.6 41.7 1 16.6
Ailanthus altissima 0.52 9.2 80.8 657.5 193.4 422.3
Alnus incana 0.27 20.7 28.5 554.9 228.1 711.2
Carya tomentosa 0.70 0.0 0.0 376.5 91.3 100.2
Catalpa speciosa 0.40 0.0 67.0 589.5 206.8 450.7
Camus amomum 0.44 0.0 60.7 515.4 191.6 460.8
Camus sericea 0.45 16.0 56.5 446.7 159.1 380.4
Gleditsia triacanthos 0.45 10.0 54.4 4 14.4 82.2 194.6
Juglans cineraea 0.50 0.0 241.0 542.6 138.2 98.7
Lindera benzoin 0.59 16.9 71.7 525.4 164.2 409.0
Platanus occidentalis 0.28 40.2 29.8 575.7 347.0 834.9
Quercus alba 0.69 17.0 73.5 273.5 55.2 75.7
Quercus bicolor 0.54 16.2 65.4 274.9 78.8 172.7
Quercus coccinea 0.57 18.3 73.2 253.2 80.5 106.3
Quercus macrocarpa 0.62 19.1 72.1 299.1 82.0 161.2
Quercus phellos 0.47 26.5 59.2 259.2 1 17.7 317.7
Quercus prinus 0.53 22.9 63.0 230.4 56.6 128.7
Quercus robur 0.68 3.6 89.2 230.8 58.3 185.0
Quercus rubra 0.53 1 1.7 65.5 291.7 95.9 169.9
Quercus velutina 0.53 22.8 65.4 239.4 77.9 139.4
Rhus typhina 0.36 25.4 41.7 740.2 318.9 790.8
Robinia pseudoacacia 0.40 0.0 58.4 540.9 307.2 405.6
Ulmus americana 0.39 34.2 0.0 389.3 1 16.6 342.2
Mean 0.48 14.0 66.3 385.0 1 17.5 245.3
LS mean (for mass) 0.45 -------------- 423.0 141.2 313.3
Gymnosperms
Abies amabilis 0.32 31.2 64.4 137.7 62.1 420.2
Abies concolor 0.26 60.5 0.0 150.2 81.1 475.1
Larix laricina 0.20 45.5 28.2 316.2 154.1 636.6
Picea stichensis 0.23 47.4 28.0 195.5 96.7 398.7
Pinus nigra 0.26 0.0 90.3 149.8 90.6 454.5
Pinus ponderosa 0.21 63.8 4.8 178.7 94.1 376.6
Pinus strobus 0.17 59.2 21.3 150.1 101.8 428.6
Pseudotsuga menziesii 0.26 41.0 34.1 194.4 1 14.4 467.1
Mean 0.23 44.0 33.9 178.0 96.4 451.8
LS mean (for mass) 0.30 ------------- 129.0 50.8 192.3
P, t-test <0.001 <0.001 0.049 <0.001 0.037 0.03 1
Partial P, model incl. mass <0.001 ------ 0.799 <0.001 <0.001 <0.0001

 

63

Table 2.1 (cont’d).

 

 

 

 

 

 

A301. A30wp LCP QY
(nmol g I s 1) (nmol g I s l) (pmol m 2 s 1) (unitless)
Angiosperms
Acer negundo 42.0 15.1 10.2 0.072
Acer rubrum 36.0 17.3 7.3 0.062
Acer saccharinum 38.5 13 .2 5 .5 0.049
Acer saccharum 56.6 22.5 3.3 0.091
Aesculus glabra 28.8 4.0 6.0 0.042
Aesculus hippocastanum 30.4 5.8 4.1 0.05
Ailanthus altissima 79.3 26.5 9.8 0.073
Alnus incana 64.5 34.9 10.3 0.079
C arya tomentosa 52.6 10.6 4.6 0.049
C atalpa speciosa 69.5' 29.4 5.8 0.069
Cornus amomum 53.1 22.3 6.8 0.064
Camus sericea 68.8 29.3 8.9 0.093
Gleditsia triacanthos 3 1 .5 9.3 8.3 0.052
Juglans cineraea 29.9 9.8 7.4 0.040
Lindera benzoin 57.5 20.1 2.9 0.052
Platanus occidentalis 79.3 44.5 1 1.6 0.088
Quercus alba 23.9 6.1 7.0 0.046
Quercus bicolor 27.4 9.9 4.5 0.053
Quercus coccinea 24.9 9.4 6.4 0.051
Quercus macrocarpa 28.9 10.6 5.6 0.053
Quercus phellos 27.8 8.8 8.2 0.047
Quercus prinus 27.7 10.2 7.4 0.058
Quercus robur 18.2 5.1 7.6 0.048
Quercus rubra 29.5 12.4 6.8 0.060
Quercus velutina 29.2 12.0 4.6 0.058
Rhus typhina 94.4 45.2 9.6 0.091
Robinia pseudoacacia 36.7 16.1 1 1.4 0.063
Ulmus americana 65.0 24.9 9.1 0.086
Mean 40.6 14.2 6.8 0.060
LS mean (for mass) 47.0 17.7 7.3 0.064
Gymnosperms
Abies amabilis 21.8 1 1.4 9.9 0.084
Abies concolor 23.0 12.0 12.4 0.085
Larix laricina 27.3 17.1 15.0 0.065
Picea stichensis 26.5 12.8 1 1.0 0.072
Pinus nigra 7.5 4.9 20.2 0.063
Pinus ponderosa 4.0 2.3 25.7 0.051
Pinus strobus 23.5 13.9 10.7 0.082
Pseudatsuga menziesii 12.0 6.9 17.2 0.052
Mean 15.3 8.6 14.5 0.068
LS mean (for mass) 9.2 4.0 1 1.4 0.053
P, t-test <0.001 0.0642 <0.001 0.203
Partial P, model incl. mass <0.001 <0.001 0.01 0.021

 

64

Table 2.1 (cont’d).

 

 

 

 

 

 

RWP 1 NWPl TNer
(nmolg ls ) (mgg ) (mgg )
Angiospenns
Acer negundo 16.6 26.1 121.1
Acer rubrum 10.6 23.9 129.5
Acer saccharinum 7.7 15.4 1 18.2
Acer saccharum 4.0 20.6 192.8
Aesculus glabra 3.5 28.3 180.7
Aesculus hippocastanum 4.6 16.4 147.1
A ilanthus altissima 5.6 24.3 130.3
Alnus incana 19.5 29.7 52.0
Carya tomentosa 4.9 27.0 163.9
C atalpa speciosa 9.8 19.8 94.5
Cornus amomum 12.3 18.2 103.9
Camus sericea 13.9 20.4 132.3
Gleditsia triacanthos 5.3 22.6 125.3
Juglans cineraea 8.4 20.3 177.8
Lindera benzoin 8.4 21.7 165.8
Platanus occidentalis 18.7 28.3 55.7
Quercus alba 3.5 16.9 245.3
Quercus bicolor 5.9 14.9 1 12.6
Quercus coccinea 4.2 14.5 161.8
Quercus macrocarpa 4.6 16.5 169.7
Quercus phellos 4.8 15.0 1 14.1
Quercus prinus 6.8 18.0 138.2
Quercus robur 4.4 18.4 1 13.3
Quercus rubra 5.6 15.7 107.0
Quercus velutina 4.3 15.7 157.6
Rhus typhina 23.7 23.3 1 10.4
Robinia pseudoacacia 15.6 31.5 122.1
Ulmus americana 12.6 24.9 81.2
Mean 7.5 20.4 133.3
LS mean (for mass) 8.8 21.4 124.0
Gymnosperms
Abies amabilis 14.4 26.4 37.1
Abies concolor 15.0 23.0 20.3
Larix laricina 18.0 22.9 26.0
Picea stichensis 9.7 21 .3 38.2
Pinus nigra 33.0 29.3 20.0
Pinus ponderosa 16.7 30.3 19.5
Pinus strobus l 1.4 28.0 18.8
Pseudotsuga menziesii 19.3 26.7 25.5
Mean 16.1 25.8 25.7
LS mean (for mass) 9.2 21.9 57.0
P, t-test 0.001 0.012 <0.001
Partial P, model incl. mass 0.84 0.788 <0.001

 

65

Table 2.2. Correlation matrices for germinant mass, ﬁnal mass, relative growth rate
(RGR), residuals of RGR vs. germinant mass, whole-plant total non-structural
carbohydrates (TNpr) and residuals of TNpr vs. ﬁnal mass. The top number is the

correlation coefﬁcient for angiosperms only and the bottom number represents the
coeﬁicient for all species. * P < 0.05, ** P < 0.01, *** P < 0.0001.

 

 

 

 

 

 

 

 

 

 

Germinant
mass
096*" Final Angiosperms
095*“ mass All species
-0.92*** -0.77*** RGR
-0.81*** -0.58***
0.00 0.28 0.39* Resid. RGR
0.00 0.32“ 0.58*** vs. germ mass
0.54“ 0.61 *** -0.38* 0.31 TNCwP
0.64*** 077*" -0.21 0.53***
-0.05 0.00 0.12 0.18 0.79*** Res. TNCwP
-0.15 0.00 0.37* 0.44“ 064*" vs. final mass

 

 

 

66

 

Table 2.3. Correlation statistics for interrelationships between plant functional traits and
germinant mass, ﬁnal mass, relative growth rate (RGR), residuals of RGR vs. germinant

mass, whole-plant total non-structural carbohydrates (TNpr) and residuals of TNpr
vs. ﬁnal mass. The top number is the correlation coefﬁcient for angiosperms only and
the bottom number represents the coefﬁcient for all species. * P < 0.05, ** P < 0.01, ***
P < 0.0001.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Initial Final mass RGR Residuals Residuals
TNCWP
mass RGR vs Mass TNpr vs Mass
LMR All -0.72*** -0.78*** 0.44“ -0.24 -0.83*** -0.35*
Ang -0.56** -0.64*** 0.58" -0.08 -0.58*** -0.24
SMR A11 -0.59*** -0.46** 064*" 0.29 -0.43** -O.11
Ang -0.59*** -0.48*"‘ 0.65“” 0.27 -0.54** -0.31
RMR All 076*“ 084*“ -0.47** -0.25 090*" 038*
Ang 079*" 078*" -0.73*** -0.11 077*" 0.37"
LPR All -0.39* -0.56*** -0.03 -0.59*** -0.64*** -0.32
Ang -0.17 -0.30 -0.02 -0.45* -0.29 -0.14
RPR A11 0.41 * 0.50" -0.13 035* 0.48“I 0.15
Ang 0.34 0.44“I -O.15 0.41 * 0.40“ 0.18
SLA All -0.17 0.07 0.57*** 074*" 0.40* 054*"
Ang -0.76"'** -0.68*** 0.77*** 0.19 -0.36 0.07
LAR A11 -0.62*** -0.52** 080*" 036* -0.21 0.30
Ang -0.84"* -0.82*** 082*" 0.08 -0.55** -0.05
SRA A11 -0.89*** -0.86*** 069*" 0.00 -0.69*** -0.13
Ang -0.89*** -0.89*** 077*" -0.14 -0.70*** -0.19
A 3 0L A11 -0.20 0.00 0.48” 055*" 0.35* 0.56***
-0.80 -0.76*** 0.75*** 0.03 -0.43* 0.05
Ang
A3OWP A11 -0.67*** -0.40* 068*" 0.40* 0.11 0.42“
-0.87 -0.84"”""I 080*" -0.01 -0.56** -0.05
Ang
LCP All -0.56*** -0.69*" 0.16 -0.49** -0.75*** -0.35*
Ang -0.50"'* -0.58" 0.27 -0.46* -0.58"”" -0.20
QY A11 -0.71*** -0.69*** 055*" 0.04 -0.45** 0.14
Ang -0.78*** -0.79*** 066*" -0. 14 -0.50** -0.02
RWP A11 -0.78*** -0.79*** 0.54*** —0.16 -0.75*** -0.21
-0.80"* 080*" 069*" -0.1 1 -0.67*** -0.24
Ang
NWP A11 -0.54*"‘* -0.59*** 0.29 -0.26 -0.48** -0.04
-0.53** -0.58** 0.39" -0.26 -0.29 0.09
Ang

 

67

Table 2.4. Multiple regression models of relative growth rate, relative growth rate with

initial mass as a covariate, whole-plant total non-structural carbohydrates, and root mass
ratio. Models were developed by ﬁrst including the strongest bivariate predictor (Table

2.1), then adding the variable with the second strongest bivariate predictor, and its

interaction. Additional variables were left in the model if adjusted R values and Pratt
indices indicated that their inclusion explained additional variance in the predicted term.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predicted Predictor Standard. b Pvalue F M. 8. Mt R?
All data
RGR LAR -. ----- <0.0001 60.3 2208 0.63
RGR LAR 0.605 <0.0001 36.3 1186 0.67
SRA 0292 0.0300
RGR Germ. mass -........... <0.0001 65.45 0.65
RGR Germ. mass -0.737 <0.0001 93.48 1468 0.84
SLA 0.444 <0.0001
RGR Germ. mass -0.874 <0.0001 74.6 1007 0.86
SLA 0.355 <0.0001
LCP -O.218 0.0165
RGR Germ. mass -O.826 <0.0001 69.5 998 0.85
SLA 0.337 0.0003
LPR -O.182 0.0520
TNpr RMRSept ------- <0.0001 139 94885 0.80
RMRSept RMRAug ---.... <0.0001 316 0.903 0.90
0.809 <0.0001 277 0.473 0.94
RMR RMR
Sept LPR Aug -0250 <0.0001
0.720 <0.0001 239.00 0.319 0.95
RMR RMR
Sept LPR Aug 0230 <0.0001
LCP -0.150 0.0076
Alliosperms
RGR LAR --.--- <0.0001 53.2 2221 0.66
RGR Germ. mass ------- <0.0001 147.1 2809 0.84
RGR Germ. mass -1.030 <0.0001 101.4 1472 0.88
LCP 0.230 0.0055
RGR Germ. mass -0.953 <0.0001 93.14 1457 0.87
LPR —O-l8l 0.0156
Germ. mass -1050 <0.0001 87.07 1009 0.91
LCP 0.212 0.0046
TNpr RMRSept ------ <0.0001 38.07 27313 0.58
RMRSept RMRAu <0.0001 135 0.28 0.83
0.880 <0.0001 87 0.146 0.86
RMR RMR
569‘ LPR Aug 0190 0.0132
0.625 <0.0001 88 0.102 0.91
RMR RMR
5“" LPR Aug -0.180 0.0062
-0110 0.0019
RWP
0.810 <0.0001 66 0.099 0.88
RMR RMR
Sept LPR Aug -O.180 0.0148
LCP -0.1 10 0.0630

 

68

 

 

leaves -

 

 

 

 

 

 

 

 

 

 

stemsEZD
roots -
375
Organ'“ 1.) (a) 100 P
Order*** A
300 ~ °\°
A |nter*** ' ‘g 80
.— 2
IO) o—o
o, 225 . I g 60 .
e b 25
3 150» f l 1; 4o.
1- . g
o
O
75- f E 20-
o . 9 a a
5% m i a
. 0

 

Lf St Rt Lf St Rt

An ‘0 rms G mnos rms
Angiospenns Gymnosperms 9' spe y pe

Figure 2.1. Box plots of tissue-level (Lf = leaves, St = stem, Rt = roots) non-structural
carbohydrate concentrations of Angiosperms (n = 28) and Gymnosperm (n = 8) species

(a). Lower and upper ends of the boxes represent the 25th and 75th percentile, lower and

upper whiskers represent the 10th and 90th percentile and the horizontal lines within the
boxes represent the median. Tissue-level TNC concentration means that do not share a
common letter are signiﬁcantly different (P < 0.05, Tukey-Kramer HSD). Total non-
structural carbohydrate (TNC) partitioning for angiosperms and gymnosperms (b).
Signiﬁcant t-test statistics (* P < 0.0001) indicate differences between plant orders
(angiosperms, gymnosperms). In TNC partitioning to speciﬁc organs (leaves, stems,
roots). Results of ANOVA for TNC as a function of order, organ (leaves, stems, roots)
and their interactions are indicated as: *** P < 0.0001.

69

'7). 35
(0
,P 30
'co
m 25
E 20
0
g 15
g 10
O
5, 5
d.)
.2
% -5
0:
TO)
01
g 300
8
E 250
9.
.c 200
0
e
g 150
a
g 100
3
0
g 50
it
8 o
C
Te
0
'—

0 Gymnosperms
o Angiospenns
v Quercus spp.

 

 

 

 

 

 

 

 

 

1
0 1 2 3 4
Germinant mass (10910 mg)

(C)

r O

1- , O

5~ . .
\\€5‘Q3~O Q
P- M
o
. Q)
o ‘ O .
-5 0 5101520253035

Relative growth rate (mg 9'1 day‘1)

-1

Total non-strucutural carbohydrates (mg 9'1) Total non-strucutural carbohydrates (mg 9

300
250
200
1 50
1 00
50
0

 

 

 

 

 

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

300
250
200
1 50
1 00

50

Final mass (10910 mg)

 

Angiospeorms only (d)

 

 

 

 

 

 

0
-20 -15 -10 -5 0 5 10 15

Residuals (RGR vs. Germinant mass
(m9 9'1 day")

Figure 2.2. Relationships between relative growth rate and germinant mass (a), total non-
structural carbohydrates and ﬁnal mass (b), total non-structural carbohydrates and relative
growth rate (c), and total non-structural carbohydrates and the residuals of the regression
of relative growth rate vs. germinant mass ((1). The inset on (d) is for the residuals of the
regression of relative growth rate vs. germinant mass for angiosperms only and has the
same axis scales as the larger ﬁgure panel. In the larger panels, solid lines are regression
ﬁts of all data, and hatched lines are for angiosperms, and in (a) and (c) for Quercus spp.
Correlation statistics for these relationships are in Table 2.2.

70

35
30 ~
25
20
15
10 ~

 

 

 

 

_5 1 1 1 1
1.6 1.8 2.0 2.2 2.4 2.6

 

Relative growth rate (mg 9'1 d'1)

Leaf area ratio (10910 (leaf area/total plant mass)

 

 

20 b T
is!) i

10—

 

 

 

o
8
o
8

 

 

 

initial mass (mg 9'1 day-1)

 

 

 

 

 

Residuals, relative growth rate vs.
initial mass (mg 9'1 day'1)
Residuals, relative growth rate vs.

5
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

leaf light compensation point for
photosynthesis (Iog10pppo)

010 20 30 40 50 60 70

Leaf partitioning ratio (%)

Figure 2.3. Relative growth rate vs. leaf area ratio (a), and residuals of the regression of
relative growth rate vs. germinant mass vs.leaf partitioning ratio (b), and vs. leaf light
compensation point (c). Corresponding correlation statistics are in Table 2.3. See Figure
2.2 legends for other details.

71

.238".
0:8 com mucowﬂ Nd oSwE com .m.~ 2an 5 2a 353on coca—oboe wcmccoamotou .on mﬁoﬁcmmouoﬁ 8m “Eon 8:359:06
Em: .2633— 28 A3 2.8 $683353 .32 A3 038 32: ~08 3:3 336.3386 358255.15: :33 mo maimconﬂom 4mm oSwE

$86 :63 3853.: .00: ago:
0:9 mmmE .03.

amino Fac: 2855885
no. .58 cozmmcanoo Em: MRoi. 36v one. ascoatmaumoo

V... N... 0... ad 0.0 v.0 ow ov cm 0

— — q a

 

 

A”: By 3

 

 

 

 

 

(L-ﬁ 6w) sateipAqoqieo ieintnomis-uou 19101

72

 

 

 

 

 

A 100-
°\°
I 80*
m
.2
g 60+
fl)
,3 4o-
5
in 20r
U)
o 1 1 1

 

 

 

-10 0 10 20
Residuals, relative growth rate vs.

initial mass (mg 9'1 day'1)

Figure 2.5. Seedling survival vs. the residuals of the regression of relative growth rate vs.
germinant mass for all data (larger panel) and of seedling survival vs. the residuals of the
regression of relative growth rate vs. germinant mass for angiosperms for the inset panel.
Fits are for all data (larger panel) and angiosperms only (inset). Pearson correlations are
r = 0.52, P = 0.001 for all data and r = 0.55, P = 0.002 for angiosperms.

73

CHAPTER 3

ASSOCIATION OF MORPHOLOGICAL AND PHYSIOLOGICAL TRAITS
WITH NORTHERN TEMPERATE TREE SPECIES LANDFORM AFFINITY

ABSTRACT

Greater water use efﬁciency (WUE) and access to water (Waccess) may be two general

adaptive mechanisms to low soil water availability. Contrasting glacial landforms with
differences in water holding capacity (outwash-low, ice contact-moderate, moraine-high),
and dominant vegetation, in northwestern lower Michigan provide an ideal system to
develop a mechanistic understanding of the association between plant traits and plant
performance. F irst-year seedlings of eight tree species (in order of increasing site
moisture afﬁnity, (Quercus velutina, Quercus alba, Quercus rubra, Prunus serotina.

A cer rubrum, Acer saccharum, F raxinus americana, Betula alleghaniensis)) were
transplanted across these landforms. In their third year, leaf gas-exchange and soil
moisture were measured monthly, and seedlings were completely excavated early-
summer and fall to obtain growth, size and morphological characteristics. 2002 had a dry
growing season with July-September precipitation of 11.2 mm vs. a 30—year average of

31.8 cm. Soil moisture decreased from moraine to outwash sites and all sites were lowest
in July. The ability to maintain positive photosynthetic rates (i.e., Waccess) and high
photosynthesis per unit water loss (i.e.,WUE) during drought enhanced seedling survival.
Across species, increased Waccess was realized via deeper rooting which was positively
related to seed and seedling size. Interspeciﬁc variation in WUE was positively related to

area-based leaf N (leaf Narea). Quercus spp, which were generally the most xeric adapted

of the species, had greater leaf Nana, root depth, seed and seedling size and survival on

74

all sites. Conversely, more mesic-associated species had lower survival, but had the
highest values of traits related to surface area for resource capture and growth potential
under optimal resource conditions, especially on the most mesic sites. Thus, these
contrasting suites of traits may partly underlie interspeciﬁc differences in growth and
survival responses, which likely contribute to the observed species distribution patterns

across glacial landforms in northwestern Michigan.

75

Introduction

Composition of temperate forest communities depends on the species-speciﬁc responses
of individual trees, particularly seedlings and saplings, to spatial and temporal variation
in resource availability (Kobe 1996, Pacala et a1. 1996). To date, most studies examining
interspeciﬁc variation in plant growth and survival in relation to natural and experimental
resource variation have mostly focused on nitrogen and light (Kobe et al. 1995, Canham
et al. 1996, Walters and Reich 1997, Carlton and Bazzaz 1998, Fahey et a1. 1998, Finzi
and Canham 2000, Walters and Reich 2000). In contrast, much less is known about
water despite the fact that tree species distribution patterns are associated with regional
rainfall gradients (Swaine 1996, Bongers et a1. 1999, Wang et a1. 2006), regional
variation in potential evpotranspiration (Gholz 1982) and species-speciﬁc differences in
drought sensitivity (Engelbrecht et al. 2007). Species-speciﬁc variation in young
seedlings’ sensitivity to water deﬁcits may function as an important ecological ﬁlter by
controlling species composition via mortality (Haeussler et al. 1995). For example, soil
water availability is an important limiting resource to juvenile tree growth and survival
(Coomes and Grubb 2000, Tanner and Barberis 2007) and species vary markedly in these
responses (Caspersen and Kobe 2001, Sack 2004, Engelbrecht et al. 2005, Kobe 2006, de
Gouvenain et al. 2007). Suites of plant traits likely underlie these responses; however,
our understanding of the physiological mechanisms and associated plant traits that govern
plant performance across gradients of soil water availability is limited.

Two general mechanisms thought to underlie adaptations to low water availability

are greater efﬁciency in using water to produce biomass (WUE) and enhanced access to

water (Waccess). Greater WUE occurs primarily by maximizing photosynthetic gain per

76

unit water lost through transpiration, especially in extreme xeric environments (Cowan

and F arquhar 1977, Cowan 1986). Evidence for WUE as an adaptation to soil water
deﬁcits remains equivocal. Based on interpretation of carbon isotope ratios (l3CzlzC),

WUE has been found to be greater in xeric than mesic habitats in some studies
(Gurevitch et al. 1986, Ehleringer and Cooper 1988, Dudley 1996), but not others
(Schulze et a1. 1996, Schulze et al. 1998). For seedlings of four temperate tree species,
using instantaneous gas-exchange measurements as an index of WUE (i.e.,
photosynthesis/stomatal conductance to water vapor), Ni and Pallardy (1991) found no
clear trend towards increased WUE for more xeric species. Higher WUE has been found
for wheat cultivars adapted to drought prone habitats (VandenBoogaard and Villar 1998).
However, an annual crop completes its life cycle in a single season, placing a selective
premium on rapid growth. Rapid growth is characterized by high leaf nitrogen, and high

photosynthetic rates (Reich et al. 1998a, Reich et al. 1998b), resulting in greater WUE
mostly as a consequence of higher photosynthetic rates drawing down intercellular C02

concentration rather than stomatal regulation minimizing water loss (Field et al. 1983).
Furthermore, higher WUE via lower stomatal conductance could increase soil water
availability to competing plants, depending on how transpiration is manifested on a

whole-plant basis (Cohen 1970) and several studies demonstrated that the preemption of
water (i.e., presumably through greater Waccess) is more important than WUE to growth

and survival in drought-prone environments (Bunce et al. 1977, Delucia et al. 1988,
Delucia and Heckathom 1989, Royce and Barbour 2001). Thus, empirical evidence

suggests that high WUE may not be a fundamental component of drought adaptation.

77

Reich and Hinckley (1989) showed that greater Wacccss, as evidenced by higher

pre-dawn leaf water potential, was highly correlated with daily maximum leaf

conductance and presumably photosynthesis. Thus, greater Waccess may sustain

photosynthesis during drought events. Similar to the positive association between sapling

survival and photosynthetic rates under waterlogged conditions (Pennington and Walters

2006), the maintenance of positive carbon balance via enhanced W,ccess could increase

seedling survival during prolonged water deﬁcits. Greater Waccess may be realized by

attaining a large size and/or through increased proportional allocation of mass to roots
(root mass ratio, RMR), surface area (Givnish 1986, VandenBoogaard and Villar 1998)
or rooting depth (Nepstad et al. 1994, Canadell et al. 1996, Jackson et a1. 1996, Schulze et
al. 1996, Jackson et a1. 1999). Since water availability typically increases with soil depth
during extended dry periods (Landsberg 1986), deep roots are likely the primary location
of water uptake. Variation in root depth may be related to species differences in seed size
(Kohyama and Grubb 1994, Guerrero-Campo and Fitter 2001) or species associated with
xeric environments may produce deeper tap-roots for a given investment in root mass
(Yamada et al. 2005). Although there is evidence of greater rooting depth in extreme
environments (Canadell et al. 1996) and modest increases in root mass allocation in
response to water limitation (Poorter and Nagel 2000), information on rooting depth
patterns and root morphology of species grown together across soil moisture gradients
under realistic ﬁeld conditions is scant. However, one recent study showed that variation
in rooting depth among ﬁrst-year seedlings of ﬁve Mediterranean woody species growing

in a common garden were strongly related to survival during a prolonged drought event

78

(Padilla and Pugnaire 2007). This ﬁnding suggests that rooting depth is an important
species-level trait with potential consequences for seedling establishment in dry

environments and community dynamics across gradients of soil water availability.

Traits potentially enhancing young seedling survival on drought-prone sites (e.g.,
greater proportional mass allocation to roots, deep roots, and conservative water use) may
“compromise growth potential, and thus competitive ability, when water is predictably
plentiful. For example, increased allocation of biomass to root systems and/or the
production of deep rooted large diameter “taproots” may occur at the expense of
allocation to resource harvesting structures (e.g., proportional allocation of mass to leaf
and root area), which contribute to high growth capacities under optimal resource
conditions (Reich et al. 1998a, Poorter 1999, Walters and Reich 2000, Comas et a1.
2002). Therefore, traits that confer survival during episodic drought events may occur at
a trade-off with traits enhancing growth potential when soil water is plentiful.
Quantifying the interrelationships of plant performance and speciﬁc plant traits during
drought events will contribute signiﬁcantly to the efforts to understand current species
distribution patterns and aid in predicting the future outcome of climate change (e.g.,

altered precipitation regimes IPCC 2001) on landscape-level forest composition patterns.

Striking differences among dominant forest communities are apparent among
contrasting glacially derived landforms in northern lower Michigan that differ in soil
texture (Host et al. 1988), nitrogen availability (Zak et a1. 1989) and calcium availability
(Schreeg et al. 2005). For example, slow-growing oak-dominated stands on low-fertility
drought-prone outwash plains are adjacent to productive mesic hardwood forests on high

fertility, mesic moraines (Table 3.1). This landscape provides an ideal model system to

79

develop a more complete mechanistic understanding of the plant traits that underlie
species-speciﬁc responses to variation in soil water availability as site differences in soil

water and nutrient availability are not confounded by variation in regional climate.

In this study, eight species (Acer rubrum, Acer saccharum, Betula alleghaniensis,
F raxinus americana, Prunus serotina, Quercus alba, Quercus velutina, and Quercus
rubra) differing in soil resource afﬁnity were transplanted across six sites, two on each
post-glacial landfonn within this regional landscape (outwash = 2, ice contact = 2,
moraine = 2). One site on each of the landforms was chosen to be well-drained and the
other to have an elevated water table in order to try to minimize covariation in nutrients
and water availability across sites. For multiple seedling plots on each site we quantiﬁed
soil resource availability (soil N and water) species-speciﬁc rooting depth, root
morphology, gas-exchange, seedling water status and survival in response to natural

seasonal variation in soil moisture. Speciﬁc predictions were:
H1. Across sites and gas-exchange sampling dates, at higher soil moisture, interspeciﬁc
variation in leaf-level photosynthesis will be most strongly associated with leaf N status

as opposed to traits associated with Waccess.

H2. Across sites and gas-exchange sampling dates, at lower soil moisture, interspeciﬁc

variation in leaf-level photosynthesis will be most strongly associated with some
combination of traits that confer Waccess (e.g., whole-plant mass, RMR, root surface area,

root depth).

80

H2a. The association between leaf-level photosynthesis and root depth and
photosynthesis during the peak of a drought will be strongest on the most xeric

site and weakest on more mesic sites.

H2b. Compared to species associated with mesic sites, species associated with
drought prone sites will have greater rooting depths.
H3. Following a prolonged drought event, species with the greatest survival will have

greater expressions of traits associated with Waccess and higher photosynthetic rates under

low soil moisture, whereas WUE will be generally unimportant.

H3a. WUE will be strongly associated with leaf N.

H4. High allocation to traits that promote Waccess during drought events will occur at a

trade-off with traits associated with high growth potential under optimal resources (e. g.,
proportional allocation to leaf and root area) and this trade-off may partly underlie

current overstory species distributions across this post-glacial landscape.

Materials and Methods

Research Sites and Plot Layout
This study was conducted in the Manistee National Forest (MN F), Wexford and Manistee

counties, in the northern lower peninsula of Michigan (Figure 3.1). The MNF’s glaciated
landscape results in wide landscape scale variation in forest composition (Table 3.1) that
is associated with post-glacial landforrn variation in soil nutrients (Nitrogen, N and

Calcium, Ca) and soil water holding capacity. The forests are second growth stands that

established after extensive logging around the turn of the 20th century. Mean annual

81

precipitation totals 81 cm (Albert 1994) and is distributed, on average, evenly throughout
the year; however, year to year growing season precipitation is variable due to stochastic
drought events. The MNF provides an ideal natural soil moisture gradient without the
confounding effects of climate, elevation and latitude. In order to span a gradient in soil
mineral nutrients and water, ﬁeld sites were established in six forest stands across the
glaciated landscape in the MNF, including two on outwash plains (OW, low water), two
on ice contact landforms (IC, intermediate water) and two on moraines (MOR, high
water). Sites were selected to achieve some variation in soil water independent of
nutrients by choosing three well—drained (OW, IC, MOR) and three sub-irrigated sites
(OW, IC, MOR). Sites were chosen with the aid of the ecosystem classiﬁcation systems
of Cleland et a1. (1993) and all three well-drained sites were previously used as reference
sites in the development of this system.

Based on visual estimates, seedling transplant plots were positioned across a
continuous light gradient (~1-32 % open sky) within each site. However, due to species-
and site-speciﬁc differences in the openness of overstory canopies (e. g., open oak
canopies on xeric sites versus closed sugar maple canopies on moraine sites), the lowest
light levels were not present at ice contact and outwash sites. Plots were established
within ﬁve targeted light levels on the two outwash sites, six levels on the two ice contact
sites and six or seven levels on moraine sites. Plots were weeded as necessary throughout
the experiment to maintain consistent light levels at the seedling level, but this
maintenance was minor.

Seedling plots consisted of four subplots for two separate seedling harvests, which

were used to quantify species-speciﬁc morphology, rooting depth, physiology and

82

survivorship in relation to natural variation in aboveground and belowground resources,
especially soil water availability. Plots were fenced with 2 inch welded wire to
approximately 1.5 min height and 1.25 cm wire mesh to a height of 1.0 m to prevent
mammalian herbivory. Subplots were 120 cm by 140 cm with a 40 cm buffer zone
between subplots and between the fence and outer subplots. Individual seedlings were
randomly placed into subplots within a 7 x 8 grid system with 20 cm spacing between
seedlings. Three to eight seedlings (depending on germination success) of each species
(A cer rubrum, Acer saccharum, Betula alleghaniensis, F raxinus americana, Prunus
serotina, Quercus alba, Quercus velutina, and Quercus rubra) were transplanted into

ﬁeld plots.

Seedling establishment and transplanting

For all eight species, we obtained seed from a commercial source (Shefﬁeld's Seed Co.,
Inc., Locke, NY) and all seed originated from USDA Hardiness Zone 4 or 5. Seeds were
pre-treated and stratiﬁed according to Young and Young (1992) throughout late winter
and early spring 2000 at MSU. Recent germinants were planted into seedling ﬂats
(individual root plugs were 12 cm deep by 5.5 cm in diameter) at the Department of
Forestry’s Tree Research Center (TRC), MSU. Seedlings were grown in low fertility
ﬁeld soil obtained from a sandy glacial outwash site in Roscommon County, MI and
watered with deionized water in order to minimize nutrient carryover effects to ﬁeld
plots. Starting in April 2000, seedlings were initially grown in a whitewashed,
temperature and humidity controlled greenhouse at the TRC and transferred in mid-May

to an outdoor lathe house and grown under 25% of full sun until mid-July.

83

In mid-July, seedlings were transported in vans to MNF, where they were kept
outdoors under moderate light conditions and watered with tap water until transplanting.
Due to low germination, young naturally established germinants of F. americana and P.
serotina were excavated and directly transplanted into ﬁeld plots. Overall, I transplanted
approximately 7600 seedlings into ﬁeld plots between July 20 and October 15, 2000.
Before transplanting, soil from seedling ﬂats was gently rinsed from seedling root
systems and seedlings were stored on trays with wet newspaper. Given the large scale of
this transplant experiment, only enough seedlings were prepared in this manner that could
be manageably planted in a single day. This process was repeated each day throughout
the duration of the transplanting. To minimize the potential confounding factors
associated with transplant shock, seedlings received supplemental water for

approximately two weeks after initial transplant.

Resource measurements

Canopy openness (%), an index of light availability, was estimated at the top of the
seedling canopy for each sublot across all six sites during late summer 2002 with paired
LAl-2000 plant canopy analyzers (LI-COR, Inc. Lincoln, NE). Brieﬂy, measurements
above each subplot were obtained when the sky was unifome overcast with one LAI-
2000 unit, while an identical remote unit was placed on a tripod in a nearby clearing (< 1
km away from each site) and simultaneously recorded open-sky values. Data from each
unit were combined later to calculate canOpy openness values and subplot estimates were

averaged to obtain plot-level means (it = 35).

84

To characterize landforrn variation in soil N availability, we measured standing
extractable pools and mineralization rates in the upper 20 cm of mineral soil with in situ
incubation of soil cores (Raison et al. 1987). Separate incubations took place over three
intervals (May l6—June 12, July 9-August 13, August 20—September 18) during the
2002 growing season. Within each plot, closely spatially paired PVC cores (5.08 cm
diameter) were placed in the buffer zone near each respective subplot. One core from
each pair was bulked (time 0 = initials) at the plot-level, placed in polyethylene bags
inside an ice-ﬁlled cooler and transported to the laboratory for analysis. The remaining
core from each pair was covered with a loose ﬁtting cap to prevent leaching and was
allowed to incubate for approximately 30 days (ﬁnals). Like initial cores, incubated
cores were bulked at the plot-level and transported to the laboratory for processing and
analysis. In the laboratory at MSU, approximately 20 g fresh weight soil from bulked

samples was sieved (4 mm sieve), homogenized and extracted with 50 ml of 2M KCl.
Nitrate (N 03-) and ammonium (N H4+) in solutions were measured colorimetrically with
a continuous ﬂow ion autoanalyzer (01 Analytical, College Station, Texas, USA).

Differences in NO3--N and NH4+-N between initial and ﬁnal extracts were used to

. . . . . —i —1 . . .
calculate net rates of N-mrneralization (mg g 5011 d ). Calculated N mineralization

rates from the three incubation intervals were averaged to estimate integrated growing

season variation in soil N availability within and between sites.

To assess landforrn variation in soil water availability, sub-samples (20 g) of soils

from bulked plot-level samples used for initial N extracts (May 16, July 9, August 13,

September 18) were dried in a forced-air oven at 105°C for 48 hours to determine

85

gravimetric soil water (%). To determine vertical proﬁles in soil moisture, additional
samples were collected with a bucket auger at 0-20 cm, 20-40 cm, 40-100 cm depths
from each subplot on June 25, July 25 and September 9, 2002. For all samples, sub-plot
values were averaged to obtain plot-level means (n = 35) for each respective depth

interval.

Seedling physiology measurements

At all six sites, we measured leaf-level C02 and H20 exchange at four different sampling

intervals throughout the 2002 growing season. Measurement periods were: (1) June
9-25, (2) July 11-23, (3) July 24-August 8, and (4) August 12-30. For each site and
measurement period, measurements were collected on a single cloudless day from 8:30 to
18:00 h local time (24 days of measurements total). Due to a variety of logistical
constraints, measurement times differed slightly between sites and measurement periods,
but importantly for species comparisons, measurement times did not differ between
species, and species X site and species X measurement period interactions were not
signiﬁcant (Appendix, Tables A.2, A.3). On each measurement day, gas exchange was
measured multiple times in all plots (range = 4—8) and plots were sampled in an order
that would approximately distribute sampling times evenly throughout the day for each
respective plot. Upon arrival at each plot, a subplot was randomly selected to start
measurements and when I returned to each plot later in the day, a second subplot was
selected. Sampling within plots alternated between each of the selected subplots
throughout the day. Within subplots, one leaf from a seedling of each existing species

was sampled for gas exchange and to the degree possible leaves were not re-sampled as

86

sampling was dispersed throughout the seedlings and leaves of a given species within

each respective subplot over the course of the day.

1 simultaneously measured leaf-level photosynthesis, stomatal conductance to
water vapor and transpiration and photosynthetic photon ﬂux density (PPFD) with two
Ll-COR 6400 portable infrared gas analyzers (LI-COR Inc., Lincoln, NE). Gas-
exchange was measured under ambient conditions with the 2 X 3 cm leaf chamber and the
LI-COR 6400 was operated as an open system. The inlet air stream was attached to a

buffer volume consisting of a 20 gallon garbage bag with air containing 377.0 (i 20.1
SD) ppm of C02 that was attached to the LI-COR 6400 with a 2 m long (3 mm inside
diameter) section of Bev-A-Line® IV plastic tubing (Thermoplastic Processes, Inc.,
Stirling, NJ, USA). This approach was used to minimize ambient C02 induced variation

in photosynthesis. For measurements at each subplot, a new buffer volume of air was

collected at approximately seedling level. Depending on the light environment, ﬂow
rates were adjusted in order to maximize C02 differentials between the reference and
sample infrared gas analyzers. Measurements were collected only after stable
photosynthetic photon ﬂux density (PPF D, pmol m.-2 s-l) values and C02 differentials
had been maintained for at least 10 s. For leaves that were too small to completely cover
the chamber, a transparent grid system (150; 0.04 cm2 blocks) was used to estimate the

amount of leaf area that was sampled during gas-exchange measurements. Leaf area
estimates were used to re-calculate gas-exchange rates. Gas-exchange was measured
concurrently with volumetric soil moisture measurements. Volumetric soil moisture was

measured to a depth of 20 cm with a time domain reﬂectrometor (TDR, Environmental

87

Sensors, Inc., Victoria, British Columbia) and based on variation in volumetric soil
moisture between the four gas-exchange measurement periods, periods were classiﬁed

into very low, low, moderate and high soil moisture categories (Appendix, Table A.4).

Photosynthesis measurements provide an index of Waccess for a given soil moisture and

light availability. Water use efﬁciency of photosynthesis (WUE, mmol C02 mol—l H20)

was calculated from instantaneous gas-exchange measurements as
photosynthesis/transpiration. A total of 3516 instantaneous gas-exchange measurements
were made during 2002. At all six sites, for each of the measurement periods, individual
gas-exchange measurements within subplots were averaged at the species-level and

subplot values were averaged to determine plot-level species means.

Seedling Survivorship and Harvests

Throughout the three year experiment I monitored seedlings for survivorship at ﬁve
different census dates (July and September 2001, June, August and October 2002). For a
particular census interval (e.g., June-Aug 2002), species-speciﬁc seedling survival at the
plot-level was calculated as: (number individuals surviving at end of interval/number
individuals surviving at beginning of interval) X 100. There was a prolonged dry-period
during each year seedling survival was tracked and in the year following the experiment
(August 1 2001, July 25 2002, August 19 2003). Furthermore, gravimetric soil moisture
during these drought events was strongly correlated (Appendix, Figure A. l) and species-
speciﬁc seedling survival during a single drought event (June-Aug 2002) scaled well with

survival over the entire experiment (July 2001-October 2002) (Appendix, Figure A2).

88

Due to the frequent occurrence of drought events in this landscape and the strong
association between subsets of seedling survival data, we believe that seedling survival
over the duration of the experiment more effectively integrates species- and subplot-
speciﬁc survival responses to drought than during a single census interval. Thus,
hereafter, survival (%) will refer to survivorship over the entire experiment.

Within each plot, all surviving seedlings in two of four subplots were harvested
June 18-30 2002 (harvest 1, n = 1559) and in the remaining two subplots from October
5-November 2002 (harvest 2, n = 1380). Pre-dawn on the mornings of harvest 1, leaf
xylem water potential was determined for a sub-sample of seedlings and plots (n = 223

individual seedlings) with a pressure bomb (PMS Instruments, Corvallis OR, USA).
These values provide an additional index of Waccess as seedlings should be in equilibrium

with soil water potential at this time of the day, and predawn water potentials have been
found to be closely related to leaf conductance during the photoperiod (Reich and
Hinckley 1989). At each harvest, seedlings were completely excavated and sandy soils
enabled high recovery of ﬁne roots. Maximum rooting depth was measured for each
individual seedling with a meter stick (to the nearest 0.1 cm). Seedlings were placed in
polyethylene bags inside ice-ﬁlled coolers and transported to a nearby ﬁeld laboratory
where they were stored in refrigerators (0—24 hours) until processed. Seedlings were
gently rinsed with deionized water to remove excess soil and were partitioned into root,

stem (including petioles) and leaf fractions. Plant fractions were dried in a forced air-

oven at 100°C for 1 hour to quickly stop respiration and then at 70°C for 24 hours. After

preliminary drying, samples were transported to MSU, dried at 70°C for another 24

89

hours, and then weighed. From harvest 1 primary biomass data, we calculated root mass
ratio (RMR; root mass/total plant mass, in g g—l).

Prior to drying at the ﬁeld laboratory, images of whole-plant root systems and
leaves were acquired with a ﬂatbed scanner at resolutions of 400 and 200 DPI,
respectively (Epson Expression 1680, Nagano, Japan) and archived for image analysis.
Digitized root images were manually edited with Adobe Photoshop 7.0 (Adobe Systems
Inc., San Jose, California) with the goal to produce a black (roots) and white
(background) image that faithfully captured the original root image. Edited root and leaf
images were analyzed for total surface area with WinRhizo Pro 5.0 and WinFolia Reg
2003b, respectively (Regent Instruments, Blain, Quebec). From harvest 1, primary

biomass data and root and leaf surface area data were used to calculate speciﬁc root area
. 2 -1 . 2 -l
(SRA, cm g ) and leaf area ratios (LAR, cm g ).

Within subplots from harvest 1, leaves from individual seedlings of each
respective species were bulked and pulverized into a ﬁne powder with a ball mill (Kinetic
Laboratory Equipment Co., California). To assess plant nitrogen (N) status, sub-samples
(2—4 mg) of bulked leaf samples (n = 426) were measured at MSU with dry combustion
gas-chromatography (NA 1500 elemental analyzer, Carlo-Erba, Milan, Italy). Species-

speciﬁc leaf N values for bulked subplots were averaged to obtain plot-level means,

which were expressed on a leaf area basis (leaf Nam, 11g cm—Z).

Statistical analysis
Unless noted otherwise, all statistical analyses were carried out with JMP 4.0 statistical

software (SAS Institute, IN C., Cary, North Carolina, USA). Gravimetric soil moisture

90

was ﬁrst analyzed with a mixed linear model that included main effects and interactions
of site (n = 6) and sampling date (n = 6) as nominal factors and canopy openness (%) as a
continuous factor. Based on the signiﬁcant effect of sampling date on gravimetric soil
moisture (Appendix, Table A5), linear models were also developed that included main
effects and interactions of site (n = 6) and canopy openness (%) for each respective
sampling date and for growing season averages. For the July 25 sampling date (i.e., the
peak of the drought), vertical gravimetric soil moisture proﬁles were evaluated with
linear models that included main effects and interactions of site (it = 6) and depth interval
(n = 3; 0—20 cm, 20—40 cm, 40-100 cm). Variation in seasonal averages of N-
mineralization rates was tested with models that included main effects and interactions of
site (n = 6) and canopy openness (%). Analyses of soil characteristics were based on
plot-level means (11 =35) of canopy openness and soil resources. When main effects of
site were found to be signiﬁcant (P S 0.05), I compared pairs of site means with Tukey-
Kramer HSD.

Plant trait values were compared among species and other sources of variation on
two bases: (1) as means of seedlings at a common harvest time and (2) as estimates at a
common mass based on trait-mass allometric functions. I decided to use estimates at a
common mass because plant traits scaled non-proportionally with whole-plant mass
(Appendix, Figure A3) and plant mass varied among species, sites and light
environments. Preliminary analyses indicated that site-speciﬁc allometric relations by
species were appropriate functions for these estimates. My justiﬁcations for this decision
were (1) within each site, mixed models for the main effects and interactions of species (n

= 8), mass (whole-plant or root mass, depending on the trait) and canopy openness on

91

plant traits showed that species and mass effects dominated, whereas canopy openness
and its interactions were generally unimportant (data not shown). (2) Mixed models of
the main effects and interactions of mass, site (n = 6) and species on traits indicated
strong mass and species effects, but also signiﬁcant site main effects and interactions
(Appendix, Tables A.6-A.9). Using data from individual seedlings from harvest 1 (n =
1520), standardized major axis (SMA) linear regression (after Warton et a1. 2006) was
used to estimate allometric relationships of whole-plant mass with root mass, rooting
depth and leaf area (SMARTR, Version 2.0) and of root mass with root area for every site
by species combination. SMA ﬁtting techniques were considered appropriate, as there
was error associated with both the X and Y variables. These regressions (Tables 10-33)
were used to estimate RMR, root depth and LAR at a common whole-plant mass of 0.5 g
and SRA at a common root mass of 0.3 g. These values were selected to maximize the
degree of overlap among data. Species traits were not estimated at a common mass if
regression equations were not signiﬁcant (P > 0.05) or species were not within :5 0.1 g of
the common whole-plant (0.5 g) or root mass (0.3 g). The inﬂuence of interspeciﬁc
variation in seed size on whole-plant size and rooting depth patterns of transplanted
seedlings was explored with linear regression. Published values of seed mass were used
for these analyses (Young and Young 1992). SMA techniques were not applied to seed
mass interrelationships because seed mass was estimated without associated error as a

published value.
In this study, indices of WacceSS included pre-dawn water potential for a sub-
sample of individual seedlings from harvest 1 (i.e., beginning of the drought event) and

AElma during measurement periods that varied in soil moisture status. Several plant traits

92

were examined for their associations with Waccess and/or WUE, including leaf Nam,

whole-plant mass, total root area, SRA, RMR, and maximum root depth. We assessed
these associations with two complementary approaches. First, Pearson’s correlations
were used to test for associations between pre-dawn water potential and plant

characteristics at the individual seedling level. Secondly, leaf-level photosynthesis was
analyzed as a function of PPF D (pmol m2 5.1) with simple linear regression and as a
function of PPFD in combination with plant traits with multiple linear regression. Values
used for regressions were species-speciﬁc plot-level means (11 = maximum of 280).
These models were evaluated for each measurement period (very low, low, moderate and

high soil moisture conditions) and then compared to identify plant characteristics driving

plant gas-exchange responses to increases and decreases in soil moisture.
Preliminary models of WUE as functions of PPFD, leaf Narea, site and their
interactions indicated signiﬁcant site effects (Appendix, Table A.34), thus, PPFD and leaf

Narea effects on WUE were assessed for sites separately. Models were generally weaker

or not signiﬁcant at higher moisture status (data not shown) so I only present data from
the measurement period with very low soil moisture.

Within the three well-drained sites (OWl , IC 1, MOR l) which tended to have
lowest soil moisture during drought, canopy openness data were ﬁtted to seedling
survival data (%) with a Gompertz growth function (i.e., general form:

01exp[-exp(02- 03—Log10 canopy openness)]). The Gompertz growth ﬁmction has been

used previously to predict survival as a ﬁmction of seedling relative growth rates

(Walters and Reich 2000) and due to the strong relationship between light availability

93

and RGR (Walters and Reich 2000), we expected canopy openness to model survival
with a similarly shaped function. Within each well-drained site, the function was solved
iteratively for the best ﬁt (i.e., minimized residual sum of squares) using the nonlinear
platform within JMP. Due to an inability to ﬁt a function to survival data at MOR 1, data
from the plot with the highest light availability were excluded and the analysis was
repeated. All ﬁts were signiﬁcant at P < 0.05. Based on model estimates of seedling

survival from nonlinear model ﬁts, species-plot residuals of light-survival functions

(SURVresid) were calculated for each plot as follows: observed species survival — overall

plot-level survival estimate (for all species combined). Alternatively, species-speciﬁc
survival deviations from overall plot-level average survival were calculated as: average

plot-level species survival — overall average plot-level survival (for all species
combined). The advantage of this approach was that unlike the calculation of SURVresid,
which excluded high light data from MOR, all data were used for estimates of survival
deviations. Species-speciﬁc SURVresid and survival deviations represent variation in

survival that was unexplained by canopy openness (i.e., removing the effects of canopy

openness in seedling survival) and my goal was to account for this unexplained variance.

As a result, I used linear regression to examine relationships between SURVmsid, survival

deviations and size (whole-plant mass), morphological (root area, SRA, RMR, root

depth) and physiological characteristics (leaf-level photosynthesis,WUE).

94

Results

Resource availability

Across all six landforrn study sites, canopy openness within seedling transplant plots
ranged from 0.7 to 46 % (Table 3.2) and generally ranked as follows across landforms:
OW > IC > MOR. Variation in canopy openness was greatest in MOR sites (43 and 18
fold) and lowest within OW sites (6 and 3 fold). Measures of canopy openness were
highly correlated with averages of instantaneous PPFD obtained during gas-exchange
measurements both within (Table 3.2) and across sites (P < 0.0001, r = 0.92, data not
shown).

The 2002 growing season was marked by a prolonged drought in which July
through September precipitation was approximately 1/3 of the 30-year mean (11.2 vs.
31.8 cm average, 1971-2000 period, Wellston Tippy Dam NOAA Climatic Station,
Figure 3.2). Soil moisture varied markedly among sampling dates, with averages across
sites ranging from a low of 5.3 % at the peak of the drought on 25 July to a high of 13.5
% on 16 May (Appendix, Table A5, Figure 3.2). Site effects explained most of the
variation in soil water (Appendix, Table A.35), whereas canopy openness effects
independent of site were weak. Soil moisture varied across sites for all sample dates, but
only moderately on July 9, and was generally highest at MOR sites and lowest at OW 1
(Table 3.3, Figure 3.2). Furthermore, for MOR and OW sites, but not IC sites, the well-
drained site had lower soil water than the corresponding sub-irrigated site. Surprisingly,
soil moisture decreased with increasing soil depth (0-20 cm = 20-40 cm > 40-100 cm) on
ﬁve of the six sites for all measurement dates (height of the drought, Figure 3.3; other

data not shown).

95

Similar to soil water, site effects explained most of the variation in average
growing season N-mineralization rates (Appendix, Table A35), and canopy openness
effects independent of site were weak (canopy openness: P > 0.1276, Appendix, Table
A.36). Among sites, N-mineralization rates varied 3-5 fold, with MOR sites generally

having higher N-mineralization rates than the others (Figure 3.4).

Seedling characteristics

Averaged across species, site-level leaf Narea varied from 56.7 to 66.8 pg cm_2, with

landforms generally ranked OW = MOR > 1C. In general, leaf Narea was highest for the

Quercus species, intermediate for B. alleghaniensis and lowest among P. serotina, A.

rubrum, A. saccharum and F. americana (Figure 3.5). Across all sites, intraspeciﬁc
variation in leaf Narea was low for most species, except for F. americana, which varied

1.5-fold (Figure 3.5).

Among species, average whole plant mass, root surface area, RMR and root depth
were highest for the three Quercus species, intermediate for A. saccharrum and F.
americana, and lowest for P. serotina, A. rubrum and B. alleghaniensis across most sites
(Appendix, Figures A.4-A.6; Figure 3.6). When compared at a common mass, species
rankings were similar for RMR, however, patterns of root depth were the reverse of the
general; root depths were actually lowest for the Quercus species at a common mass of
0.5 g (Figure 3.6). Across landforms, whole-plant mass and root surface area generally
ranked as follows: MOR > OW > IC (Appendix, Figures A.4, A.5), whereas root depth

ranked: OW > [C > MOR (Figure 3.6). Despite two years of post-germination growth,

96

species variation in whole-plant mass and root depth was strongly positively associated
with seed size for all sites (Figure 3.7).

Across species, SRA and LAR varied 2.3—fold and 5-fold, respectively. For
most sites, species values of SRA were highest for B. alleghaniensis and A. rubrum,
intermediate for F. americana, A. saccharum and P. serotina and lowest for Quercus
species (Appendix, Figure A.7). When species were compared at a common mass,
Quercus species maintained the lowest SRA and LAR values and this trend was most
evident at moraine sites (Appendix, Figure A.7; Figure 3.8). However, differences in
SRA among the other species were more subtle than species comparisons at a common
harvest, whereas differences among estimates of LAR were more pronounced. Across

landforms, SRA and LAR generally ranked as follows: MOR > IC > OW.

Gas-exchange interrelationships

Throughout the 2002 growing season, Aarea varied considerably among species, sites and
measurement periods that differed in soil moisture status (Appendix, Figure A.8). Aarea

was positively related to PPF D and the slopes and the amount of variance in Aarea

explained by PPFD increased with soil water (Tables 3.4-3.6, Figure 3.9), except for the
date with the highest soil water (Table 3.7). At this early growing season date leaves of
Quercus species had not fully developed, especially within the MOR sites and this may
explain the weaker relationship for this date (J. Kunkle and M. Walters, personal

observation).

97

Models with PPFD, leaf Narea and their interaction explained more total variation

in Aarea than models with just PPFD and their overall effects increased as soil moisture

status increased (e. g., 3% additional variance explained at very low water vs. 11% at

moderate soil water), again, except for the highest, and earliest soil water sampling date
(Tables 3.4-3.6, Figure 3.10). The PPFD x leaf Narm terms were signiﬁcant for low and
moderate soil water conditions (Tables 3.5, 3.6). The lack of a signiﬁcant independent
effect of leaf Narea on Aarea at high water (Table 3.7) may be due to leaves that were not
fully developed by this early season sampling period.

In comparison to PPF D and leaf Nma, which had the largest inﬂuence on Aarea as
soil moisture increased, variation in root depth, whole-plant mass and root surface area

had the greatest effect on Aarea on the lower soil water measurement dates. For example,

at the highest soil water, root depth was not a signiﬁcant predictor of Ama independent of
PPFD (Table 3.7), whereas at moderate soil water, the interaction of PPFD and root depth

was signiﬁcantly related to Ama (Table 3.6). Under low and very low moisture levels,

root depth was positively related to Aarea and the interaction of PPFD and root depth was

also signiﬁcant, but only under low moisture conditions (Tables 3.4, 3.5, Figure 3.11)
Notably, in comparisons of models at low and very low soil moisture, the inﬂuence of

PPFD was stronger under low moisture (i.e, F -value: 99.22 vs. 47.24), whereas variation
in root depth showed a greater effect on Ame,l under very low soil moisture (i.e., F-value:
21.49 vs. 13.15). In models with PPF D, whole-plant mass, root area and their

interactions with PPFD showed effects similar to that for root depth on Aarea (Tables 3.4-

98

3.7). Unlike whole-plant mass, root surface area or root depth, RMR was generally not a

strong positive predictor of Aarea, and especially at low soil water. SRA was a signiﬁcant

positive predictor of Aarea independent of PPF D during high soil water conditions (Table

3.7), whereas under moderate and low soil moisture conditions, SRA and the interaction

of PPFD and SRA were negatively associated with Aarea (Tables 3.5, 3.6).

As root surface area and root depth displayed similar effects on Aarea as soil

moisture levels changed throughout the growing season (Table 3.4-3.7) and because they

are themselves highly correlated (P < 0.0001, r = 0.73, data not shown) it is difﬁcult to
ascertain if they are independently important for Waccess. In order to examine this issue,
root area was added as a main factor to a model that included root depth and PPF D as
predictors of Aarea under very low moisture (i.e., peak of the drought). The addition of
root area to the model explained almost no additional variation and the model containing
only PPF D and root depth explained more variance in Aarea than the model containing
only PPF D and root area (Table 3.8). Collectively, these models indicate that root depth
was more important for increasing WI,ccess and maintaining photosynthesis during the
peak of the drought than root surface area.

In mixed models of Aarea that included main effects and interactions of site, PPFD

and root depth, Aarea differed across sites under very low moisture conditions (P =

0.0036, Appendix, Table A3 7). Thus, in an effort to examine the inﬂuence of root depth

on Aarea during the peak of the drought across sites that differed in soil water status (very

low moisture dataset), Am,l was also analyzed separately within the three well-drained

99

sites (OW 1, IC 1, MOR 1) with models that included PPFD, root depth and their

interaction. In OW l, the driest site during the drought, Aarea did not vary with PPF D,

but Aarea was positively and strongly related to root depth and the PPFD X root depth
interaction (Table 3.9). In contrast, on IC 1 and MOR 1, sites with greater soil water,
PPFD was positively related to Aarea and explained the most variation in the overall
model. In addition, in both of these sites, the PPFD X root depth term was signiﬁcant and

negatively associated with Aarea (parameter estimates —2.14 and -3.19, respectively,

Table 3.9). A negative interaction term indicates that A3,rea decreased more with PPFD at
deeper rooting depths.
In a mixed model with PPFD, leaf Nam, site and their interactions, WUE differed

across sites under very low moisture levels (Appendix, Table A34) and WUE tended to

be highest at OW 1 (Appendix, Figure A.9). Thus, in site-speciﬁc models for well-

drained sites, leaf Narea was positively related to WUE and PPFD was a moderate

predictor, but only for OW 1 (Figure 3.12).

Pre-dawn water potential

During the onset of the drought, root depth and pre-dawn water potential were positively
correlated and this association was consistent in both well-drained (P < 0.0001, r = 0.47,
data not shown) and sub-irrigated (P < 0.0001, r = 0.52, data not shown) sites and within
ﬁve out of the six study sites (P range = 0.0077-0.0001, r range = 0.38-0.63, data not
shown). The association between root depth and pre-dawn water potential was stronger

than for whole-plant mass (P < 0.0001, r = 0.45, data not shown), root surface area

100

(Figure 3.13), RMR (P = 0.72, r = 0.02, data not shown), or SRA (Figure 3.13).
lntraspeciﬁc variation in root depth and pre-dawn water potential were also positively
associated for 5 of the study species (P < 0.01 for all, r = 0.46-0.65, data not shown), but
not for A. saccharum, F. americana and B. alleghaniensis (P > 0.05), the most mesic

species.

Seedling survival interrelationships

Across species, seedling survival was signiﬁcantly associated with canopy openness (P <
0.05 within each site), but responses differed among sites (Figure 3.14), depending on
site-level soil moisture status and variation in light availability. For example, within OW
l the site with the lowest soil moisture status, seedling survival was similar (estimate =
36%) across plots that ranged from 11 to 16 % canopy openness and survival decreased
markedly under 30 % canopy openness (estimate = 5%). In comparison to OW 1, overall
seedling survival was generally higher within IC 1, which tended to have slightly higher
soil moisture, but survival responses as a function of canopy openness were similar for
both sites (Figure 3.14). For instance, in plots that ranged from 3 to 10% canopy
openness, survival estimates were invariable (estimate = 53%) whereas survival declined
appreciably under 43% canopy openness (estimate = 32%). Due to the lack of ﬁt for the
nonlinear survival function across all data within MOR 1 (P > 0.05), data from the plot
with the highest canopy openness (46%) were excluded from the ﬁnal model ﬁt (Figure
3.14). Similar to OW 1 and IC 1, under the highest light environment at MOR 1,
seedling survival decreased considerably (mean = 18%) to a level that was the same as

the estimate at the lowest light level (estimate = 18%) (Figure 3.14). In contrast to the

101

other two well-drained sites, MOR 1 had highest overall survival and survival showed a
positive relationship with canOpy openness in plots that spanned from 1 to 18% of full
sunlight (survival estimate range = l8-81%).

The residuals from the nonlinear survival versus canopy openness relationships
(SURVresid) were most strongly related to whole-plant mass, root surface area and root
depth for all well-drained sites (Table 3.10, Figure 3.15). The amount of variation in

SURVmSid explained by these linear models and the slopes all increased from the most

mesic site (MOR l) to the most xeric site (OW 1), indicating that these traits had
increasingly positive effects on seedling survival as site-level soil moisture status
decreased. However, covariance among whole-plant mass, root depth and root area
(Appendix, Figure A3, and data not shown) make it difﬁcult to determine which

predictor had functional importance for survival. In an effort to tease apart the relative
importance of these predictors for survival, linear models of SURVresid for OW 1 were

developed as follows: (1) root area added as a predictor with whole-plant mass, (2) root
depth added as a predictor with whole-plant mass and (3) root depth added as a predictor

with root area. When root area was added as a predictor along with whole-plant mass,
the model explained less variation in SURVresid than a model with only whole-plant mass

(Table 3.11). In contrast, when root depth was added as a factor with whole-plant mass,

the model explained greater variation in SURVmsid than in a model with just whole-plant

mass (Table 3.1 1). Furthermore, a model with root depth and root area as main factors
explained 10% more variation in SURVmid than a model with only root area, but only

5% more variation was explained than a model with only root depth (Table 3.11).

102

Collectively, results from a variety of models suggest that root depth is more strongly

related to seedling survival than root surface area. Leaf Narea, Aarea and WUE were

positively correlated to SURVresid, but only within OWl (Table 3.10, Figure 3.15). For
seedlings at IC 1 and MOR 1, RMR was weakly and positively associated with
SURVresid (Table 3.10, Figure 3.15). In contrast to the other morphological and

physiological characteristics, across all well-drained sites, SRA showed a negative

relationship with SURVresid (Table 3.10, Figure 3.15).

In a complementary analysis to the one using residuals of PPF D vs. survival
model ﬁts, species-speciﬁc survival deviations from plot-level averages were related to

the same set of morphological and physiological traits. The advantage of this approach
was that unlike the analysis for SURVmsid, all of the transplant seedling survival data
could be utilized in this alternative analysis. Results for the survival deviation versus

species trait interrelationships were consistent with SURVmsid versus trait

interrelationships, with the exception of leaf Nam, which also showed a weak relationship

with survival deviation at MOR 1 (Appendix, Table A3 8, Figure A.lO). The striking

similarity in results between both indices of seedling survival was not surprising,
considering that the SURVmsid were highly correlated with the species-speciﬁc deviations

from plot-level averages of survival within all sites (P < 0.0001, r range = 0.96-0.99, data

not shown).

103

Discussion

Watemccess vs. W UE as a basis for drought tolerance

I found that both the ability to maintain positive photosynthetic rates (i.e.,increased
Waccess) during the peak of the drought and high photosynthesis per unit water loss (i.e.,
water use efﬁciency, WUE) contributed to tolerance of drought for tree seedlings
common in northern temperate forests (Figure 3.16). Increased Waccess was achieved via

deeper rooting, which varied among species with seed and seedling size and not with

interspeciﬁc variation in root-whole plant allometry. Interpspeciﬁc variation in WUE
was positively related to area-based leaf N content (leaf Nam). Although direct positive

relationships of survival with photosynthetic rates and WUE were only evident at the
driest site at the height of the drought, several lines of indirect evidence suggest the

general importance of these mechanisms.

This study demonstrates that the drivers of leaf-level photosynthesis (Ama) are

highly dependent on soil water status and these relationships have important implications

for seedling survival during extended drought events. For example, photosynthetic

photon ﬂux density (PPF D) and leaf Narea became increasingly important predictors of

Aarea as soil water increased (supporting H1). This should be expected under high water

availability given the dependence of instantaneous photosynthetic rates on light
(Bjorkman 1981), and of photosynthetic capacity on leaf N concentrations both within
(Walters and Reich 1989) and across (Field and Mooney 1986, Reich et al. 1997b)

species. However, at low water availability, this relationship changed somewhat. Light

104

and leaf Narea were still signiﬁcant drivers of photosynthesis, but both were weaker. At
lower soil water availability increased rooting depth became an increasingly important

driver of Aarea, especially on the most xeric site (supporting H2, H2a).

As expected, species associated with drought prone sites had greater rooting
depths than mesic species when compared at a common harvest, which is consistent with
H2b. Surprisingly, after accounting for differences in plant mass with the use of
allometric approaches, mesic species actually had deeper roots than xeric species. My
results conﬂict with those of Yamada et a1. (2005), who found that within two genera
(Dryobalanops, Scaphium) in Malaysian tropical forests, sandy-soil specialists (i.e., drier
soils) had deeper taproots than the clay-rich-soil specialists (i.e., wetter soils).
Alternatively, in my study, differences in root depth were associated with variation in
seed size (Figure 3.16), a result that is consistent with shade-tolerant seedlings in a warm-
temperate rainforest (Kohyama and Grubb 1994) and with more than 300 adult woody
plant species from Britain and northeast Spain (Guerrero-Campo and Fitter 2001).
Furthermore, other studies showed that the abundance of large seeded species increases
as environments become increasingly xeric (Wright and Westoby 1999). Thus, it appears
that initial root depth advantages via larger seed sizes could be preserved until later life-
history stages, but this relationship should be evaluated for a broader size and age range
of individuals for the same species.

I found that variation in whole-plant mass, associated differences in root depth
and the maintenance of higher rates of Aarea during the peak of the drought were all
positively related to seedling survival (Figure 3.16), which is consistent with H3 that

increased Waccess and associated traits enhances survival during water shortages.

105

Collectively, my results may provide a physiological basis for the strong positive
relationship between root depth (but not whole-plant mass) and ﬁrst year seedling
survival that Padilla and Pugnaire (2007) found for ﬁrst-year seedlings of ﬁve

Mediterranean woody species. In my study, the root depth—survival relationship was

more robust than the association between Aarea during the peak of the drought and

seedling survival. This result was not surprising because Aarea was only measured once
at each site during the peak of the drought (on different days), whereas root depth, which
was positively and strongly associated with Ama, serves as a quantitative plant trait that
integrates potential carbon gain during extreme water shortages.

Interpretation of the interrelationships between root depth, Waccess and seedling

survival necessitates one key caveat. During drought events, soil water availability
typically increases with root depth (Landsberg 1986, Padilla and Pugnaire 2007), but I

did not observe this pattern within my study sites. This pattern appears to be in conﬂict

with the notion that deeper roots enhance W1,ccess and the positive relationships that I

found between root depth and Aarea and survival. However, the positive association

between root depth and pre—dawn water potential provides compelling evidence that deep
anchored roots increase the water status of seedlings during water shortages. If moisture
is lower at increasing depths then how can I reconcile the greater water status of
seedlings that have roots deployed within these “drier” soil strata? Soil organic matter
has been shown to increase water holding capacity; however, water may be held more
tightly within soil organic matter and may not be completely available to plants. Within

my study system, I speculated soil organic matter decreased with increasing soil depth, a

106

pattern that has been documented in other systems (Don et al. 2007). If this pattern
occurs, even though gravimetric soil water was lower at deeper soil strata, plant available
water might have been higher than in the upper soil horizons, which had the highest
levels of gravimetric soil moisture.

Within the most xeric site whole-plant mass explained 6% greater variation in
seedling survival than for root depth. This suggests that whole-plant mass may be
interrelated with additional unmeasured traits besides root depth that confer survival
during drought events (Figure 3.16). For example, I found a strong association between
whole-plant mass and whole-plant carbohydrate storage for 36 temperate and boreal
woody species (Chapter 2). Carbohydrate reserves may provide a carbon source for
maintenance respiration and growth during drought events when photosynthesis is
severely limited. There is direct, but fragmentary evidence for the importance of
carbohydrate storage for drought tolerance. For instance, several species have greater
carbohydrate pools in drought, than well-watered treatments (Dina and Klikoff 1973,
Busso et al. 1990, Oosthuizen and Snyman 2001) and Busso et al. (1990) demonstrated
that post-drought biomass production was associated with TNC pools in cool season
grasses. However, we are not aware of any studies that directly examined the importance
of interspeciﬁc variation in carbohydrate storage for drought tolerance in woody plants.

Contrary to my expectations (H3), WUE during the peak of the drought was
positively related to seedling survival on the most xeric site. Although stomatal

conductance was reduced considerably during the peak of the drought, my results suggest

that leaf Nama contributed to increases in WUE by enhancing photosynthetic rates and

drawing down intercellular C02 concentrations (Field et al. 1983). Altogether, these

107

results imply that the physiological basis of the relationship between WUE and seedling

survival stems from the maintenance of relatively high photosynthetic rates via high leaf

Nana rather than from primarily limiting water loss.

Drought tolerance vs. competitive ability at high water availability

Collectively, my observations along with empirical data from other studies imply that
there is a trade-off between traits that enhance Waccess and characteristics that are related

to surface area for the interception of light and acquisition of soil resources, which likely
compromises growth capacity under optimal resource conditions (H4). These contrasting
suites of traits may underlie interspecifc differences in growth and survival responses,
which likely contribute to observed species distribution patterns across glacial landforms
in northwestern Michigan. For example, in comparison to species associated with mesic

sites, Quercus species had the largest seed sizes, whole-plant mass and root depths, which

enhanced Waccess (as indexed by pre-dawn water potential and Aarea) and survival of

transplanted seedlings during acute water shortages (Figure 3.16). These results suggest
that Quercus species possess collections of traits that enable these species to persist on
ice contact and outwash landforms, which tend to have lower soil water status than
moraine sites. Conversely, when compared at a common harvest or a common mass,
more mesic species displayed the highest values of SRA and LAR and these expressions
of traits likely maximize growth rates under high resource conditions. Due to the
extreme drought event during the 2002 growing season, which resulted in severe water
limitations across all sites and greatly diminished growth rates (data not shown), our

ability to assess the contribution of SRA and LAR to growth capacity in this experimental

108

framework was extremely limited. Therefore, any examination of trade-offs can be no
more than speculative since this notion could not be directly evaluated with data from this
ﬁeld transplant study. However, empirical data from the literature and from my study in
Chapter 2 allowed me to further explore this notion. For example, SRA scales well with
speciﬁc root length (SRL, P < 0.0001, r = 0.93, Chapter 2, data not shown), which has
been found to be positively related with mass-based N uptake rates (Reich et al. 1998b)
and growth rates (Reich et al. 1998a). Numerous multi-species studies have
demonstrated that LAR is the morphological trait that is most strongly related to growth,
especially in moderate to high light environments (Walters et al. 1993b, Lusk et al. 1997,
Reich et al. 1998a, Poorter 1999, Walters and Reich 1999). Furthermore, in combination,
LAR and SRA explained more variance in growth rates than LAR alone for 36 temperate
and boreal woody seedling species (Chapter 2). In this study, differences in SRA and
LAR between xeric and mesic species suggest that mesic species may realize higher
growth rates when soil water is plentiful. Thus, over time, higher growth potential may
enable mesic species to overtop xeric species on moraine sites, which typically have
higher N-mineralization rates and soil water, especially during non-drought years.
Interpretation of the seedling trait data requires one important caveat. Although
comparisons of plant traits at a speciﬁc common whole-plant mass (0.5 g) or a common
root mass (0.3 g) were selected to maximize overlap in the masses of study species,
common masses were considerably lower than the published values of seed mass for the
three Quercus species (> 1.85 g). Thus, at a common mass, it appears that estimates of
plant traits for the Quercus species were based on individuals that were experiencing

negative carbon balance and potentially near death. As a result, estimates of Quercus

109

traits may have been biased (RMR, SRA, LAR), but the consistency in species ranks at a
common mass and averages at a common harvest (Appendix, Figures A.6, A.7; Figure
3.8), which include all seedlings, suggests that biases associated with these estimates
were negligible. Furthermore, although species rankings of rooting depth were reversed
when compared at a common mass and a common harvest, these rankings were consistent
at progressively higher common plant masses, with the exception of the species that were

excluded from these analyses due to non-overlap in masses (data not shown).

If there is a trade-off between traits that enhance Wg,ccess and characteristics that

are related to surface area for garnering resources, then why is Q. rubra, a species
characterized by low growth capacity as a young seedling (Walters et al. l993a), one of
the dominant species on more mesic moraine landforms (Table 3.1)? The present
abundance may, in part, reﬂect legacy effects from disturbance histories. For example, in
the early 19003, extensive logging, subsequent ﬁres and mass dieback of competing
vegetation may have played a role in the proliferation and current dominance of this basal
sprouting species (Host et al. 1988), a notion that is supported by experimental evidence
in mesic hardwood stands in southwestern Wisconsin (Kruger and Reich 1997a).
Additionally, why is A. saccharum, a species characterized by low photosynthetic and
growth rates (Walters et a1. l993a), the most dominant species on moraine sites (Table
3.1)? The success of A. saccharum suggests that additional unmeasured physiological
traits may also contribute to the success of some mesic species. I speculate that these
traits may be associated with low-light carbon balance because moraine sites without
frequent or severe disturbances are typically dominated by low light regeneration niches

(Table 3.2). In these environments, seedling success is dependent on enduring shade for

110

prolonged periods of time as advance regeneration in a “seedling bank” while
maintaining the capacity to respond rapidly to increased light availability from canopy
openings created by the death of overstory trees (Marks 1975, Canham 1985). Based on
a compilation of data from unpublished studies and from the literature, leaf-level light
compensation points and respiration rates, two traits likely to play an important role for
tolerance to low light availability (Lusk and Del Pozo 2002), varied considerably among
our study species (Table 3.12). For example, for one of the most mesic species, A.
saccharum tended to have a relatively low light compensation point and respiration rates,
whereas Q. alba, one of the most xeric species had the highest values. These carbon
conservation traits may help A. saccharum to persist in the understory until eventual
canopy recruitment on moraine sites. Therefore, variation in these physiological traits

may also contribute to species sorting across glacial landforms.

111

Table 3.1. Mean species basal area across glacial landforms in Manistee
National Forest, near Cadillac, MI (condensed from Host and Pregitzer 1992).
Outwash has the lowest water holding capacity and rich moraines the highest.

 

 

 

Species Common Outwash Ice contact Rich Moraines
name (n = 22) (n = 22) (n = 8)
Quercus velutina Black oak 9.3 7.8 -
Quercus alba White oak 8.4 6.3 -
Acer rubrum Red maple 0.3 1.2 0.1
Prunus serotina Black cherry - - 2.4
Quercus rubra Red oak 1.1 7.3 4.7
Acer saccharum Sugar maple - - 1 1.6
F raxinus americana White ash - - 2.3

Betula alleghaniensis

Yellow birch Not reported, assoc/w mesic/hydric sites

112

Table 3.2. Mean, standard deviation, ranges and Pearson’s correlation for the different
indices of light availability used in this study.

 

 

 

Canopy -2 -1

Site n openness (%) PPFD (pmol m s ) Correlation

OW-l 5 Mean (SD) 17.1 (7.6) 324.7 (282.5) 0.91**
Range 1 1.3-30.4 170.4-829.2

OW—2 5 Mean (SD) 19.3 (14.4) 180.2 (86.5) 0.85
Range 7.7-44.0 84.1-310.5

IC-l 6 Mean (SD) 12.2 (15.0) 172.4 (177.8) 0.88*
Range 3.2-42.5 43.7-517.1

IC-2 6 Mean (SD) 15.0 (12.2) 174.3 (181.4) 0.93**
Range 3.8—36.5 35.8-506.3

MOR-1 6 Mean (SD) 12.2 (17.5) 107.0 (155.1) 0.96“
Range 1.1-45.6 17.7-419.5

MOR-2 7 Mean (SD) 5.5 (4.1) 77.9 (58.8) 0.91"
Range 0.7-12.5 12.2-186.0

 

Note: *P < 0.05, **P < 0.01.

113

.Emm SEEMSBEH .36 v a: ﬁesta. base—baa 03 032 .8888 a 50:23 3% .83 mam—9:8 some 8m "8oz

 

 

 

 

0 cm m< m< m< < Rd 885 v 2:: as m 35 56:
=So>O
m < < < < < «2 :53 v a; a: m an seesaw w
m m< < < < < 23 88¢ 23 82 m an 533. ON
0 cm m< m< om < one 88¢ v a: 69° m an bi R
< < < < < < as 3.35 9% ”do m 6% b3 6
< < < < < < 86 8:; 63 and m an 6:3 mm
m m 5. m< m< < moo ~83 6.4 93 m an a: E
352 352 N9 72 $50 730 Nm .5... a .82 mm .3 56% 38
8255560 25 <>oz<

 

 

demmom wEBBw 65 $28 9&8on can 350 mesa-hem EoBb€
mmocom AR; 0.5568 :8 03083me 5 G H 5 86 mo gumbo ES: co.“ :62: 80:: 38:3 “we“: @32me a («o 9:53— .m.m 03m...

114

Table 3.4. Linear relationship of leaf-level photosynthesis (Ama) with photosynthetic
photon ﬂux density (PPF D) at very low soil water availability (See methods and
Appendix, Table A.4. for more details about soil moisture categories). Multiple linear

regression models of Aarea as a function of PPFD in combination with plant traits (leaf
nitrogen, whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth).

115

Table 3.4.

 

 

 

 

Par. Whole-model
, 2
ANOVA effects d.f. 53 F P Est. P Ad1.R
PPFD
-2 -1
L0810PPFD(umo|m 5) 1 38.73 91.73 <0.0001 1.04 <0.0001 0.28
LeafNitrogen
-2 -1
Log10PPFD(umolm s) 1 29.74 72.66 <0.0001 0.95 <0.0001 0.31
-2
L0g1oleafN(ugcm) 1 4.92 12.02 0.0006 1.29
Whole-plant mass
-2 -1
Log10PPFD(umolm 5) 1 26.21 63.7 <0.0001 0.92 <0.0001 0.30
Log10 whole-plant mass (g) 1 4.42 10.74 0.0012 0.32
Rootarea
-2 -1
L<>g1oPPFD(11molm s) 1 28.01 66.67 <0.0001 0.95 <0.0001 0.29
2
Log1orootarea(cm) 1 2.49 5.92 0.0158 0.34
Rootmass ratio
-2 -1
L0810PPFD(umolm s) 1 35.12 81.49 <0.0001 1.02 <0.0001 0.27
-1
RMR(gg) 1 0.05 0.12 0.7294 0.12
Speciﬁc rootarea
-2 -1
L0810PPFD(umo|m s) 1 30.86 73.54 <0.0001 0.97 <0.0001 0.29
2 -1
LoglOSRMcmg) 1 2.58 6.14 0.0139 —0.41
Rootdepth
-2 -1
Log10PPFD(umo|m s) 1 18.59 47.24 <0.0001 0.81 <0.0001 0.33
Log10 root depth (cm) 1 8.46 21.49 <0.0001 1.03

 

Note: Models exclude interaction term when P > 0.25 in preliminary model (Bancroft,

1964)

116

Table 3.5. Linear relationship of leaf-level photosynthesis (Aarea) with photosynthetic
photon ﬂux density (PPFD) at low soil water availability (See methods and Appendix,
Table A.4. for more details about soil moisture categories). Multiple linear regression
models of Aaml as a function of PPFD in combination with plant traits (leaf nitrogen,
whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth).

117

Table 3.5.

 

 

 

 

Par. Whole-model
ANOVA effects d.f. SS F P Est. P Adj. R2
PPFD
Log10 PPFD (me1 m-2 s") 1 91.54 147.70 <0.0001 1.74 <0.0001 0.39
Leaf Nitrogen
L0810 PPFD (nmol m'2 s4) 1 68.88 133.22 <0.0001 1.56 <0.0001 0.49
Log1o leaf N (118 ma) 1 13.03 25.22 <0.0001 2.1 1
L9810 PPFD " Log1o leaf N 1 9.05 17.52 <0.0001 5.30
Whole-plant mass
Log10 PPFD (pmol m.2 s!) 1 64.53 119.55 <0.0001 1.53 <0.0001 0.47
Log10 whole-plant mass (g) 1 13.87 25.69 <0.0001 0.54
Loglo PPFD X Loglo WP
mass 1 7.74 14.35 0.0002 1.08
Root area
Log10 PPFD (nmol m'2 5-1) 1 69.84 120.63 <0.0001 1.58 <0.0001 0.43
L9810 root area (cmz) 1 9.20 15.89 <0.0001 0.62
Logm PPFD X Log“) root
area 1 3.55 6.14 0.014 1.03
Root mass ratio
Log10 PPFD (mmol 111'2 5-1) 1 88.99 141.28 <0.0001 1.76 <0.0001 0.38
RMR (g g") 1 0.58 0.92 0.3396 -0.40
Speciﬁc root area
L0810 PPF D (nmol m.2 s") 1 75.03 126.98 <0.0001 1.63 <0.0001 0.42
Log10 SRA (cm2 8'1) 1 5.42 9.17 0.0027 -0.57
Log1o PPFD x Log10 SRA 1 5.54 9.38 0.0025 -1.49
Root depth
L0810 PPF D (nmol m-2 sl) 1 56.18 99.22 <0.0001 1.48 <0.0001 0.44
L9810 root depth (cm) 1 7.45 13.15 0.0004 0.91
Log") PPFD X Loglo root
depth 1 6.69 1 1.81 0.0007 2.41

 

Note: Models exclude interaction term when P > 0.25 in preliminary model (Bancroft,

1964)

118

Table 3.6. Linear relationship of leaf-level photosynthesis (Ama) with photosynthetic
photon flux density (PPFD) at moderate soil water availability (See methods and
Appendix, Table A.4. for more details about soil moisture categories). Multiple linear

regression models of Aarea as a function of PPF D in combination with plant traits (leaf
nitrogen, whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth).

119

Table 3.6.

 

 

 

 

Par. Whole-model
ANOVA effects d.f. SS F P Est. P Adj. R2
PPFD
Logto PPFD (pmol m'2 s']) 169.24 3 17.36 <0.0001 2.26 <0.0001 0.58
Leaf Nitrogen
Log10 PPFD (mmol m.2 5-1) 1 144.06 356.44 <0.0001 2.17 <0.0001 0.69
Log10 leaf N (118 ma) 1 9.60 23.75 <0.0001 1.83
L0810 PPFD >< Logto leafN 1 19.03 47.08 <0.0001 7.31
Whole-plant mass
L0810 PPFD (pmol m'2 s!) 158.14 352.21 <0.0001 2.28 <0.0001 0.65
Log10 whole-plant mass (g) 1 4.49 10.00 0.0018 0.31
Loglo PPFD >< Loglo WP
mass 1 15.07 33.56 <0.0001 1.42
Root area
Logm PPFD(14mo|m'2 s") 1 162.41 338.54 <0.0001 2.30 <0.0001 0.63
Logloroot area (cmz) 1 2.09 4.35 0.0382 0.30
Log”) PPFD X Loglo root
area 1 11.66 24.30 <0.0001 1.80
Root mass ratio
LOSIO PPFD(1umo|m'2 s1) 1 165.33 309.08 <0.0001 2.30 <0.0001 0.58
RMR (g g") 1 0.62 1.16 0.2822 —0.43
Loglo PPFD x RMR 1 0.75 1.39 0.2392 1.16
Speciﬁc root area
L0810 PPFD (nmol m-2 s!) 158.62 323.92 <0.0001 2.26 <0.0001 0.62
Loglo SRA (em2 g") 1 2.80 5.71 0.0177 -0.42
Loglo PPFD x Loglo SRA 1 8.79 17.95 <0.0001 -1.85
Root depth
L0810 PPFD (mmol m-2 s']) 153.04 320.10 <0.0001 2.32 <0.0001 0.63
Log] 0 root depth (cm) I 0.59 1.24 0.2676 0.26
Loglo PPFD x Loglo root
depth 1 11.63 24.32 <0.0001 2.76

 

120

Table 3.7. Linear relationship of leaf-level photosynthesis (Aarea) with photosynthetic
photon ﬂux density (PPFD) at high soil water availability (See methods and Appendix,
Table A.4. for more details about soil moisture categories). Multiple linear regression

models of Aarea as a function of PPFD in combination with plant traits (leaf nitrogen,
whole-plant mass, root area, root mass ratio, speciﬁc root area, root depth).

121

Table 3.7.

 

 

 

 

Par. Whole-model
2
ANOVA effects d.f. SS F P Est. P Adj. R
PPFD
-2 -1
L91310 PPFD(11mo|m s) 1 90.46 130.99 <0.0001 1.63 <0.0001 0.35
Leaf Nitrogen
-2 -1
LongPFDwmolm 5) 1 89.02 128.43 <0.0001 1.67 <0.0001 0.35
-2
Log10|eafN(ugcm ) 1 0.19 0.28 0.5993 0.25
Whole-plant mass
-2 -1
Log10PPFD(umolm s ) 1 89.18 128.96 <0.0001 1.70 <0.0001 0.35
Loglo whole-plantmass (g) 1 0.58 0.84 0.3592 —0.11
Rootarea
-2 -1
L0810PPFD(1-lmo|m s ) 88.64 128.07 <0.0001 1.70 <0.0001 0.35
2
Logloroot area (cm ) l 0.44 0.63 0.4285 ' -0.l3
Root mass ratio
-2 -1
Log10PPFD(umolm 5) 1 103.92 171.52 <0.0001 1.88 <0.0001 0.43
-1
RMR(gg ) 1 21.53 35.53 <0.0001 -2.55
Log10PPFD><RMR 1 2.10 3.47 0.0638 —2.19
Speciﬁc rootarea
-2 -1
Log10PPFD(11molm s) 1 97.84 147.78 <0.0001 1.78 <0.0001 0.38
2 .1
L0810 SRA (cm 8 ) 1 7.83 11.83 0.0007 0.677
L0810PPFDXLog105RA 1 0.93 1.41 0.2369 0.62
Rootdepth
-2 -1
Log1o PPFD(11mo|m s ) 1 88.42 128.46 <0.0001 1.74 <0.0001 0.35
L0810 root depth (cm) 1 1.34 1.94 0.1649 —0.37

 

Note: Models exclude interaction term when P > 0.25 in preliminary model (Bancroft,

1964)

122

Table 3.8. Multiple linear regression models of leaf-level photosynthesis (Ama) during

very low soil moisture at OW 1 with PPFD as a covariate, root area, root depth or both
root depth and root area.

 

 

 

 

Predicted Parameter Whole-model
, 2
variable Predictor 53 F P Estimates P AdJ- R
Aarea log10 PPFD 28.0 66.7 < 0.0001 0.95 < 0.0001 0.28
LO810 root area 2.5 5.9 0.0158 0.34
Aarea logio PPFD 18.6 47.2 < 0.0001 0.81 < 0.0001 0.33
I0810 root depth 8.5 21.5 < 0.0001 1.03
Aarea Iog1o PPFD 18.2 46.7 < 0.0001 0.80 < 0.0001 0.34
logio root depth 7.0 17.8 < 0.0001 1.44
logio root area 1.0 2.5 0.1 142 -0.33

 

123

 

if 885 Rd 8; _ eaaeaeeeaﬁxoaaaoaﬁ

25 an; NS :3 _ Aeavsaaeaeeéwﬁ
:3 823v ado 585v 8.8 9% _ Armageeeonaaegq
:52

2.? 88¢ a? MS _ adaeaeeewegxaaaaeaﬁ

3.? 8:3 :3 85 _ Ase ease .ee 33
mg 32:. mod 386 8.2 New _ Ara as 353 93a 33
_ o—

2}: 33d mvd on; _ ﬁned 82 Smog x Dana Smog

 

 

m2 385 3.2 Noe _ A83 .38 82 33

”no :85 :3 Be? as :6 _ A7... N.5 683 max: 33

_ 3o

ax ..€< a Begum a a mm .2. maﬁa <>oz<
3on “8089.3

-2055

 

$536.80 05688 :8 >5— bo> war—6
: mo: ._ B ._ 30V 8:... 858333 e25 2: 55E dosages :2: 2a A83 :38 82 Ara we .25. .977: base as:
c993 ozoﬁcmmouoﬁ mo case—Sm a ma ATm $8 351 33$ mmmoﬁcxmouoﬁ .32.me— mo 228:. 2238on 30:: 2&st dd 2an

124

.390 A A: m: 3
080205 0.8 22323: 0::0h:w_mr:0z .2255 v a: .12.. .Cod v a: a... Amod v 5 ... ”mm 00:00:05 08 307.8%“: 0::0E:w_m “802

 

 

 

 

 

a: an we 2. 08 1002: u an at: a ”.5?
a: R a: 8 a: 10%.: n a .25 : 85<
3:. 1303. n a tad em 2% 103.8 n a :85 t. 3.: 10033 u a :32 an 59% 82 33
0% + 0002.. n a .3 .o R :8 + 08.91 n a :53 S 32 + 08.3.1 n a 1%.: am 55 ewe:
”.8 103.: u a .2 .c 3 0% 102.8 n a :20 t. a: an :2:
«.8 t 30.8 n a .2 .o 3 ea .. 08.9. n » ream; S 22 10008 u a 4:30 3. 83 80?. Be 33
3 + 0%.: n a. an; 2 are + 08.2 n a 1.2:. S _.N_ + 03.3. n a tamed em was 053.22; sue:
a: 2 me 2. am: t 000% .1. a: L 3 mm 2:8. ewe:
5:260 5.30.31 1 z 5:260 ~— ..€< 2 cot—:50 gimp—mom x ..€< 2 020:? 300530:—
..w0< :0_mm2wom m N
_ :02 _ 0_ _ 30

 

 

.AONI 7.0:: NOU BEE .538: 05 m0 030: 05 wﬁSw 55650 cm: :8va m5?

ATm ~-E NOD BE: .Ew:0€ 05 .«0 x00: 05 wEBc Emoﬁgnmoaona 35:00: 83¢. A88 5:00 :00: ATw 80 .008 :00: 05025

N
«Em ATw w 6:8 30:: :00: M32 .ANES we: oommsm :00: .va 30:: 32920:? $-80 m1 .Zv :0w0E: .30— um>>0=0m m: 0:: 830:?
20200005 :0: SE: 0:: 305030500 :8 m:0E:m0Q .3 .m 0::me 80¢ 208:3 Suﬁ—:0: 30::000 30:3 0&2 .m> Azae>~5mv

32.2% :0 $0323: ﬂ 0338.» 0:00:03: 0:: £0.80 :0 E .m_.m 033m 5 800 :0“ woman—0:: 563%»: 30:: m0 bagm .o_.m 033.

1.25

 

 

126

 

3m :0 :3 S 33: 83 :8: 202
3o :83. v 0.8 28.0 2: N88 580 :8: 202 232%:
one 586 v 33 586 v 0% @082 53:. 88 202 2328:
$0 885 v 35 :83: v 0.3 :93 as :8: 202 523:
o. 3 £80 pm :08 5:2: :8: 202
$5 585 v 3: :83. :6 2:: a2: 2323:
3030—065 0&0:
N0: $8.0 cod 0.: 83 :8: 202
v0.0 585 v 33 «.83 3 :52 was 2323:
€030.93, Smo—
Sd :83: v :3 585 v o. 3 3?: was 2328:
3.03-0.9? Swo—
: :3 a 8385: m I: mm .9202: 05%?
N 8862:
3608-20: :oHoEmhwm

 

 

0231088050 w:_=000m 80:: (:0 5:050:80
a :0 5:00 :00: £03 :00: 63:: 2030—095 :33 _ 30 :0: 3:528: :0 20:00:: :0_wm0:w0: 80:: 032:8 0:: 295m A :.m 030%

Table 3.12. Leaf-level light compensation points (LCP) and respiration rates (R1) of
study species from shaded understory or greenhouse conditions. Data compiled from
unpublished studies and from the literature. Species values with multiple superscripts

represent averages from cited studies.

 

 

 

Species LCP (mmol m-2 s']) RL (umol C02 m'2 s")
Quercus velutina 466 05°

Quercus alba 8.26 ()jh’i

Acer rubrum 6.8e ()_4b'd’e’f’h
Prunus serotina _ 0.10b
Quercus rubra 7.3e 0.38b'c’d’f

Acer saccharum 3.3e 02¢,ng

F raxinus americana 1.13 Q44a

Betula alleghaniensis 5. 1‘ 016

 

dNote: a(Bazzaz and Carlton 1982), b(Jurik et a1. 1988), c(Kloeppel et al. 1993),
d(Kubiske and Pregitzer 1996), e(Kunkle and Walters, unpublished data), f(Loach 1967),
g(Lusk and Reich, unpublished data), h(Teskey and Shrestha 1985), (Walters and Reich,

unpublished data), (- ) no data.

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.1. Map of study area with arrow pointing to Lake, Wexford and Manistee
counties, in the northern lower peninsula of Michigan.

128

35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 30 d
“T: 25 - . —o— OW-1
2 20 .. ‘ —o— ow-2
.3 ‘ —-I— IC-1
'5 15 - E , —c1—IC-2
2 I‘ \ . ""_ MOR-1
E 10 ‘ \ - -\ . + MOR-2
(D *— a‘l /A
5 ‘ \./ \-
0 r l I T l T 1'
E July-September 30 yr average
E’ precipitation (1971-2000): 31.8 cm
_c_> 3‘ 2002:11.2cm
E
E
g 2
CL
2
"3 11
N
8 ill—L JIL Lil l I
“‘0 1 “t'r'i 1"‘i"'
May Jun Jul Aug Sep Oct

Figure 3.2. Mean gravimetric soil water availability at 0-20 cm depth for study sites
located on different glacial landforms (OW = outwash, IC = ice contact, MOR =
moraine) and daily precipitation from May 1 to Septmeber 15, 2002 (Wellston-Tippy

Darn Weather Station). Sites followed by a 1 are well-drained, whereas sites followed by
a 2 are sub-irrigated.

129

 

 

12

 

[:1 0-20 cm site (P < 0.0001)
1222 20-40 cm depth (P < 0.0001) -
9 m 40-100 cm site x depth (ns)

 

 

 

 

      

   

Gravimetric soil moisture (%)
O)
l

 

 

 

 

 

 

 

\\\\\\\\ ‘

| n“
.............
.0...0.0.0.0.0.0.0.0.0.0‘ O

 

 

 

‘ S
K
_0:0:0:9:0:0:0:0:0:0:0:0:0:0 :'
Ooooooooooooooa

 

 

0W1 0W2 |C1 ICZ MOR1MOR2

Figure 3.3. Vertical proﬁles (0-20 cm, 20-40 cm, 40-100 cm) of mean gravimetric soil
water availability on different glacial landforms (OW = outwash, IC = ice contact, MOR
= moraine) on July 25, 2002 (i.e., peak of the drought). Sites followed by a 1 are well-

drained, whereas sites followed by a 2 are sub-irrigated. Results of ANOVA for soil
water with site, depth and their interaction as factors.

130

 

0.6
Site (P < 0.0001) c

0.5

I

T
I

0.4

 

l

0.3

 

 

 

0.2

ID
0'

 

 

 

 

0.1 -

1—Hm
H

i—Hm

+——1m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0

 

Growing season average in situ nitogen
mineralization rates (ug N 9'1 soul day'1)

0W1 0W2 IC1 I02 MOR1MOR2
Study Sites

Figure 3.4. Means (d: 1 SD) of in situ nitrogen mineralization rates across landform sites
for different measurement dates and averaged across the growing season. For pairwise
comparisons of sites, means followed by a different letter are signiﬁcantly different at P
< 0.05 according to Tukey HSD.

131

Well-drained Sub-irrigated

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

90 OW1 ' ow2 Xeric
6o - - * , _ i f - , l
0 5 5 5 4 4 4 5 3 5 5 5 5 5 5 5 5
NA 90 161 102
I l. '- 2,
E ”"1 . _‘_ o -
o __ : _
a, 60 - - j '7 - ~‘ -
3 _ - j ' , . 'T
Z 30- ﬂﬂi—iﬂﬂ - Hﬂﬂﬂﬂ
“5
3 0 6 6 6 6 6 6 6 5 6 6 6 5 5 5 6 5
MOR1 MOR2
90 '- ‘v - I. f ‘T
””‘a—é— f '7‘-—~— I
60 _ f" , __ ,. 1 " . . : r—é—
30 b HUI—i H h V
0 6 6 6 5 5 5 2 6 6 6 5 Mesrc

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

Qv Qa Qr Ps Ar As Fa Ba Qv Qa Qr Ps Ar As Fa Ba

. Drought , Drought
H'gheﬁTolerance—a‘ Low High<—-To|erance——-> Low

Figure 3.5. Species-level means (i SD) of leaf nitrogen content (N, ug cm'z) across ﬁeld
sites from the June 02 seedling harvest. Numbers within or above bars denote sample
sizes (number of plots) for each respective species. Sites are separated into well-drained
(outwash = OW 1, ice contact = [C 1, moraine = MOR 1) and sub-irrigated categories
(outwash = OW 2, ice contact = [C 2, moraine = MOR 2) (see methods for details).
Species are arranged in order of their drought tolerance. Species abbreviations are as
follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, Ps = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = Fraxinus americana, Ba =
Betula alleghaniensis.

132

Figure 3.6. Root depth (cm) expressed as species-level means (:1: SD) and as estimates at
a common whole-plant mass (see materials and methods) across ﬁeld sites from the June
02 seedling harvest. Numbers within or above bars denote sample sizes (i.e., number of
plots) for each respective species. (n.d.) indicates that the speciﬁed common mass was
beyond the range of individuals for a given species X site combination. Sites are arranged
top to bottom from the most xeric site to the most mesic site. Species are arranged in
order of their drought tolerance. Species abbreviations are as follows: Qv = Quercus
velutina, Qa = Quercus alba, Qr = Quercus rubra, P3 = Prunus serotina, Ar = A cer
rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba = Betula alleghaniensis.

133

Figure 3.6. Common harvest time

Rooting depth (cm)

30

20

10

30

20

30

20

30

20

30

20

30

20

Common whole-plant mass: 0.5 g

 

 

 

 

 

 

: F— OW1-OW1

— 5 i1 5 mmﬂmgz mil—1m ndnd nd &
: V . :

’ U Hnﬂﬂg’gn... .

_ sis—i 6 mmﬂmﬁ hlcijl’lﬂ nd ndﬂi—i m

Ummﬂna

.H

mm nd nd nd nd nd nd

 

MOR 1

unit

6
6

 

 

 

m

6 6

_ MOR1

'nﬂm. ﬂﬂ.

 

 

MOR 2

.HHMmmﬂm

_MOR2

 

 

 

iﬂﬂﬂ mﬂﬂ

 

Qv Qa Qr Ps Ar As Fa Ba
Drought

Hi9h<—Toleranoe —> Low

Qv Qa Qr Ps Ar As Fa Ba

. Drought
Hughe— Tolerance—> Low

134

Xeric

ll

 

V

Mesic

 

 

 

 

 

 

 

 

 

 

 

0.5
:93
8
m 0.0—
E
E
L“
9 -o.5 —
2
O
.C
3
S -1.0 r
O)
O
_|
.15 1 1 L 1 —0—OW1
—4 -3 -2 —1 0 1 -e-OW2
Log 10 seed mass (9) + Icz
16 ﬁ-MOR1
' —¢‘—MOR2
’E‘ 1.4 ~
3
5
o. 1.2 -
Q)
"U
*6
9 1.0 -
$2
8’
.1 0.8 —
0.6 l l 1 l

 

 

 

Log 10 seed mass (9)

Figure 3.7. Relationships between whole plant mass, root depth and published values of
seed mass. Relationships were examined within each of the study sites.

135

Figure 3.8. Leaf area ratio (cm2 g4) expressed as species-level means (i SD) and as
estimates at a common whole-plant mass (see materials and methods) across ﬁeld sites
from the June 02 seedling harvest. Numbers within or above bars denote sample sizes
(i.e., number of plots) for each respective species. (n.d.) indicates that the speciﬁed
common mass was beyond the range of individuals for a given species >< site
combination. Sites are arranged top to bottom. from the most xeric site to the most mesic
site. Species are arranged in order of their drought tolerance. Species abbreviations are
as follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P3 = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis.

136

Figure 3.8.

LAR (cm:2 9'1)

Common harvest time Common whole-plant mass: 0.5 g
OW1 OW1
225 . -
150 - ~
2
75 _ g - 4 _5 -
,mmmHMmﬂﬂ mﬂ_ﬂmﬂﬂm
owz owz
225 - -
150 - » _
0 Bis—immmmm 5 ,_,[—|,_, ndﬂﬂ
IC1 I01
225 ~ .
150 - . -
75 " "I .- 1 ; ,3 _
Tﬂ4ﬂﬂﬂﬂ n
Om6m66665 .—.l—l._.nd 7nd—
IC2 IC2
225 - -
150 — -
0 [—ll—jr-ﬁ "d nd
MOR1 MOR1
225 ~ -
150 - . .1 —
MOR2
225 - - __
I '_ i—
I l
150 '- 1 [7 i-
75 - ; M Q 4 1 -
0 . 667.12!” mm ,..
Qv Qa Qr Ps Ar As Fa Ba Qv Qa Qr Ps Ar As Fa Ba
Drought Drought

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

High e—Toleranoe —> Low

137

Xeric

 

Mesic

Figure 3.9. Leaf-level photosynthesis (Ama) as a ﬁmction of photosynthetic photon ﬂux

density (PPFD, mmol m'2 s") for sampling periods that varied in volumetric soil water
content (very low, low, moderate, high; see methods and Appendix, Table A.4 for more
details). Linear regression summary statistics are provided within each respective graph
panel. See Tables 3.4-3.7 for parameter estimates.

138

Figure 3.9.

Log10 Aarea (pmol 002 m'2 5'1)

 

 

n = 232 Very low
’ P < 0.0001

Adj. R2 = 0.29

 

 

n = 235 Low
” P < 0.0001

2 O
Adj.R =0.39° o

 

 

n=227 °

 

 

n = 243 High

 

 

 

Log1O PPFD (pmol m'2 s'1)

139

Figure 3.10. Multiple regression model of leaf-level photosynthesis (Ama, pmol m.2 s")
in relation to photosynthetic photon ﬂux density (PPFD, pmol m'2 s") and leaf Nam.a (pg

cm'z) across all sites during (1) very low, (2) low and (3) moderate soil moisture
conditions. Each datum represents a species X plot mean. Regression models are as

follows: (1) very low, Aarea = -2.978 + 0.949 (loglo PPFD) + 1.294 (loglo leaf Nana),
adjusted R2 = 0.31, n = 227, P < 0.0001; (2) low, Aarea = -5.176 + 1.555 (loglo PPFD) +
2.115 (loglo leaf NW) + 5.308 (loglo PPFD x loglo leafNarea), adjusted R2 = 0.49, n =
230, P < 0.0001; and (3) moderate, Aarea = -5.999 + 2.174 (logm PPF D) + 1.833 (loglo

leaf Narea) + 7.307 (loglo PPFD X loglo leaf Nam), adjusted R2 = 0.69, n = 230, P <
0.0001.

140

 

Figure 3.10.

-2 5:1)

Aarea (pmol 002 m

 

 

 

  
 

'0) 1.
0“ 601‘
E 1
N 45i
O l
E.’ 3.01
O t
E .
E:

(U

2

(D

<

141

Figure 3.11. Multiple regression model of leaf-level photosynthesis (Aarea, pmol m-2 s")

in relation to photosynthetic photon ﬂux density (PPFD, umol m.2 54) and root depth
(cm) across all sites during (1) very low and (2) low soil moisture conditions. Each
datum represents a species X plot mean. Regression models are as follows: (1) very low,

Aarea = -1.612 + 0.81 (loglo PPFD) + 1.028 (loglo root depth), adjusted R2 = 0.33, n =
227, P < 0.0001; (2) low, Ama = —2.380 + 1.482 (loglo PPFD) + 0.911 (loglo root depth)
+ (loglo PPFD X logo root depth), adjusted R2 = 0.44, n = 230, P < 0.0001.

142

 

 

:6 N-

 

E «00 .95; 6294

 

111:9 o
0 .
. 1 0

(11,111..
5
4“ 3

a
EN 0.0E3m9 <
f.» N- 0

Figure 3.11.

143

SN: 5.

 

 

 

 

.08 N00 .955 m3>>

-l

-2

Figure 3.12. Multiple regression model of instantaneous water-use efﬁciency in relation
to photosynthetic photon ﬂux density (PPFD, pmol m s ) and leaf Narea (ug cm'z) at

OW 1 during very low soil moisture conditions. Each datum represents a species X plot
mean. Regression model: WUE = -5.75 + 1.62 (PPFD) + 2.87 (leaf Nam), adjusted R2

0.28, n = 30, P = 0.0042.

144

Pre-dawn water potential (MPa)

 

 

 

 

0 - A A A
D «ammo
m (D. 0 OD
mmaA A
O m
A DAB
0 IMAOAM A D
O .0 D D
O .- D
I C DA
MED
° n=218
I I P=0.01
-8* 0 r=-0.17
1.2 16 2.0 2.4 2.8

Log1O SRA (cm2 g")

3.2

 

 

 

 

 

D n = 218
I I P < 0.0001
0 r= 0.48
0.4 0.8 1.2 1.6 2.0 2.4

Log1o root area (cm2)

 

 

 

 

 

 

 

 

 

a? o —
a.
E
3g '2 h o OW1
.9 o owz
o I IC1
a _
3 '4 a lC2
a; A MOR1
3 _6 _ A MOR2
c
3 c1 n=223
g _ _ P < 0.0001
9'3 -8 _ a r= 0.49
a. 1 l l

0.4 0.8 1.2 1.6

Log10 root depth (cm)

Figure 3.13. Relationships between pre-dawn water potential (MPa) and speciﬁc root

area (cm2 g'l), total root surface area (cm2) and root depth (cm). Each datum represents
individual seedlings of all species across all study sites. Sites followed by a 1 are well-
drained, whereas sites followed by a 2 are sub-irrigated. Associated correlation statistics
are provided within each graph panel.

145

Figure 3.14. Relationships of seedling survival (%) versus Loglo PPFD across all species
within well-drained sites (OW 1, IC 1, MOR 1). Seedling survival was estimated as the
percentage of the original seedling population (July 01) that was alive in October 02.
Each datum represents a plot-level PPF D average. Data were ﬁtted with a Gompertz

function with the general form: Blexp[—exp(92- 93—Log10 canopy openness)]. Each site-
speciﬁc function was solved for the best ft function (i.e., minimized residual sums of
squares) iteratively using the nonlinear ﬁt platform within JMP (SAS Institute, Cary,
North Carolina). All ﬁts were signiﬁcant at P < 0.05.

146

 

 

 

 

 

 

100

75

50

25

100

75

50

25

Seedling survival (%)

100

75

50

25

Figure 3.14.

 

 

 

 

 

 

 

 

 

 

r0W1 IQv OQa lQr DPs
AAr AAs OFa <>Ba
I
- u
I
- 000
I
_ .A I
6.51
20
- J l 1 001 l
5101 A
O O
A
_ . I
g . G I
a o
O O .A
~ PDT
A I R
I D ‘0
- E] 0 Cl
<>
0 A O
- l l o l o 1 lo
”MOR1 O a
Q I
8
- O
- 5‘ O
A 3
- 3 n
O
o
_ IO 0 l l l l0
0.0 0.4 0.8 1.2 1.6

Logw canopy

147

openness (%)

Xeric

 

Mesic

Figure 3.15. Relationships of SURVIeSid (i.e., residuals of survival vs. canopy openness

nonlinear functions) with leaf Narea, size (whole-plant mass) and morphological (root
area, SRA, RMR, root depth) and physiological characteristics (leaf-level photosynthesis,

Amea; water-use efﬁciency, WUE). Relationships were examined within each of the

three well-drained sites (OW 1, IC 2, MOR 1). Regression equations, adjusted R2, P
values and n for these relationships are presented in Table 3.10.

148

 

 

 

 

 

OW1 o 09
O O
O O OO
O
_ o o
O
Cboo
I go
0 CO
I l l

 

 

1.4 1.6 1.8 2.0

Log10 leaf N (pg cm'z)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l'lC10 Coo
O
o ((3)90
0

d9

0

O 000
00%0

0o
o O0

o
008
oo
o
(D 08
' <0

 

1 J L

 

 

1.4 1.6 1.8 2.0

Log10 leaf N (ug cm'2)

 

 

 

 

 

 

 

 

 

 

 

 

 

(DO

0
6830
c1008
. €0.05
- O O

O
OO

 

 

l 1 l

1.4 1.6 1.8 2.0
Log10 leaf N (pg cm'z)

 

 

 

 

 

 

 

 

 

 

50 'OW1 "'01 C?) o o MOR1
25 r - 5 0 00 <3;
0 O
O OCDOC%O%
A 0 _ - i- OO./O/8
o\° 04’ O o
. v _25 __ _ _ 0
g g -50 L - L
-E g 1 l l l l L L l l l l l l l l
a 8' -1.5 -1.0 -0.5 0.0 0.5 -1.5 -1.0 -O.5 0.0 0.5 -1.5 -1.0 -0.5 0.0 0.5
2' a L091owhole-plant mass (9) L°91O whole-plant mass (9) L°910 whole-plant mass (9)
g 2 5° ”OW1 [IC1 <20 FMOR1
:9 8 CD 00
g 2 25 -
3 0 -
-25 _
-50 _
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
Log10 root surface area (cmz) Log1o root surface area (cmz) Log1o root surface area (cmz)
50 “CW 1 g) IC 1 g 8? MOR ‘1
25 - 8 - ° 0 ~ 0 O
Q) CD (CDC) 0
0 t o 060 ‘ Q) r f)
(SD 0 o
-25 r o 008 O ” OO O 00 F g) 0
oo 8 O o o
m
-50 d d o o _

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

Figure 3.15.

RMR (9 9'1)

 

 

RMR (g g")

149

 

 

 

0.0 0.2 0.4 0.6 0.8 1.0

RMR (9 9'1)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50: OW1 ’00 O IC1 * MOR1
0Q) OO O
25 _ h 90 r 0080
0 _ _ 0%) Q) r- 0 (Dog
25 - 0° (60% - 000 O
- _ O O 00 O
0 O O
-50 _ _ O O _
1.5 1.8 2.1 2.4 2.7 3.0 1.5 1.8 2.1 2.4 2.7 3.0 1.5 1.8 2.1 2.4 2.7 3.0
Log10 SRA (cm2 g“) 1.6910 SRA (cm2 9'1) Log1o SRA (CD12 9'1)
50 *ow1 9%,) "IC1 o MOR1
O
’3 25 I w o - - <2, 8 8
o\ B 6 O
05 V 0 5 o o o 5 5 00 o
> (D O O O
a 8 -25 i- O0 0 L- l- O O
.2 E o o 8> 0
Z a) -50 I t .
a 8- _L A L 1 l l L l J l l J
" >
-g 8- 0.6 0.9 1.2 1.5 0.6 0.9 1.2 1.5 0.6 0.9 1.2 1.5
g g Log1o root depth (cm) Log1o root depth (cm) Log1o root depth (cm)
3 ‘39 5°“0W1 80 5 0d? IC1 * MOR1
m 0’ 25 ~ - 00 0 ~ 0
2 00 (£3
0 ‘D O o 0 E 00
.. - O O 1.0 O O
0% CD00
-25 - _ CD0 0 _ o o
0 0 C230 0
-50 - _ _
0 1 2 3 4 0 1 2 3 4 0 1 2 3 4
50 Aarea (meI 002 rn'2 6'1) Aarea (pmol rn'2 6'1) Aarea (pmol rn'2 s'1)
”OW1 O 8 I @g IC1 * MOR1
O o O
25 h 00 a) _ O i- O O O O
OO O 00 000
o 0 £90 0
0 - h (8)0 *0 O o o
(2238080 O 000
-25 r o 000 00 ~ 0 O I o 0800
-50 _ - _

 

 

 

 

 

 

 

 

 

-1.5 0.0 1.5 3.0 4.5 -1.5 0.0 1.5 3.0 4.5 -1.5 0.0 1.5 3.0 4.5
WUE (mmol coz moi“1 H20) WUE (mmol 002 mol'1 H20) WUE (mmol C02 mor1 H20)

Figure 3.15 (cont’d).

150

 

Species

\__.J

 

 

Variation in seed
mass

(+)

 

 

 

[ Whole-plant maSSJ _____________________ 1

 

 

 

 

 

 

 

 

 

 

 

 

(+)
[ VariatiNoane:n Leaf ] Root depth ]
(+) (+) (+) l
WUE A”... (+) Waccess
during drought
(+) (+) l
Seedling survival }_ _ _ _ _ _ _ - - "(1'). ........

 

Figure 3.16. Conceptual diagram of factors inﬂuencing interspeciﬁc survival of ﬁeld
transplanted seedlings. Plus signs (+) indicate signiﬁcance in correlation analyses or
best-ﬁt linear models. Dashed line indicates that additional, unmeasured traits that are
associated with plant mass may have a positive effect on seedling survival.

151

APPENDIX

152

 

 

 

 

153

 

 

A? 3.2 ~86 83 i 0885 u A :55 .65 3.6% 5,266.6 66E
9: Ed ~85 ~26 1 cog; u A :35 6:65 A6 2.66.6 666::
A3 8.2 83. ﬁg 1383 n A 63 68A a Adm 5% 6.6.25 323
:3 52 oz; 562 1 383 u A :65 1A 6662.6 5.2663.
A: 3..» ~86 £2. 1 3&3 u A :85 q 6566.5 6.5.666
A3 6M; 32. $3 1 333. n A :65 {A 66.26 658
A: $6 $3 58 1 333 u A :85 .32 .5 5552.6 655
63 :3 $2. 553 1 3585 u A :35 .5655 6 65.66 6.66% 66.66% 6665
52 A3 523 3.8 1 353.5 u A .1. 35 6:2 265 6 .66.: 6.2656 668
62 62: 5:; on; 13585 u A :65 6665 «.5 =8 666: .56 666: 3C 6666 .336
A: 62. A85 83 1 ES; 1 A :35 665% «.3: E 65.536 65263
OS 2 .5 03° 82. 1 333 u A :63 .5 55626655: .3665.
:2 86 $3 $3 1 369° n A :35 33A 6666. .3665.
A3 53 52. 32 1 383 u A :86 is: 256666. 6%
on; Em 523 $3 I 3635 u A :35 q 56,366.52 6.6
N1 63 $3 $3. 1 383 n A :53 q 5566 66w
:2 N2: N85 5.0 1 ES: 1 A :66 4 66:56.. 66...
a: 2.2 god m3 1333 u A :55 66:: 6 .65 €65 6 669 666:8 6.55.
of 8.5 336 5:. 1 3355 u A :35 6566A 6 .665 :33 6 6.6% £3656 636
6.3 d3 .5 8:65 mm

868886938555 mowed—69¢. :Emmewum 863m

.83
.o n EmoﬁcxmoAS—Q £023 8 .26— Qmmm rod ESQ 8:359:00 Em: AG 28 ”90 u ego—m rod Born 8353: A3 x66<v Tm ~-E

BE: cm 3 mmmonAimvoﬁ 86:53 A: “8 wow: 803 32:33 commmodwoa _o>o_-mo_ooam mat—awe 2:. .moﬂg Nd use vacuumimﬁ mo
$6: 33683 can 28:33 commmouwe 80:: ATm ~-E 383 Our; .m> owcacox9mmw N00 3538. .Ao $883 36on ._.< 2an

.Sood v .7... hSod v m... "Sol

 

 

 

of Ed 255 Ed 1 3235 1. N :36 .6 6566696 653
S._ in: $3 :3 1 3%: u a 3.6% q 366663 Siam
a: $6 83 83 1853 u .A :66 q 623% 25
2: 8... ”mod 82 1 3%: u .A :36 .53 ESE §§m
of 3% 83. 2:. 1 383 u x :85 .6 665 666$
8; m3 ”:3 $2 1 $35 1 k :35 q :38 266%
N2 :Q was 33 1 gamed n z :35 q 62E 266$
2: .3 :55 £2 1 35.3 1 » ...... 85 q 6:23 666$
”2 $6 $3. 32. 1 3mm: 1 a :85 .665 658265 6663
8.. as 53 ”NS 1 9:85 1 x :36 £663: 6568 266$
3.. 5. mod :3 1 $3; 1 .A :§ 62:. 636.6 266$
2: as $3 82 1 333 u a :86 q 33 666$
$5 8.: ~85 83 1 335 u a 1.2.5 36665 666666
NE 6.: ”so 8." 1 339° n z :85 q £36666 6:65
$4 92 Sod and 1 ES? 1 z :86 4 66,5. 65
mg 8.2 58 NS 1 353 u x :35 $3 6 ~68 286:3 65
So m _ .om Sod 2.. 1 some: n x :85 365: Ems 3E
6.3 “5.— >0 8:35 mm

220:8th 3383mm mombmc< commawom mﬂooqm

 

 

.8283 ._.< 2an

154

Table A2. Means, standard deviations and ranges of measurement times (0.0-24.00 h

local time) across sites and measurement periods during the 2002 growing season.

 

 

Measurement Periods

 

 

1 2 3 4
Site (June 9-25) (July 1 1-23) (July 24-August 8) (August 12-30)
OW-l n 134 122 142 102
Mean (SD) 13.4 (2.5) 13.6 (2.1) 13.8 (2.3) 13.3 (2.3)
Median 13.8 14.1 14.2 13
Range 9.2-17.6 9.9-17.0 9.5—17.0 9.8—17.1
OW-2 n 1 10 l 14 144 148
Mean (SD) 14.7 (2.7) 13.3 (2.1) 13.6 (2.2) 12.8 (2.4)
Median 15.3 13.1 13.8 12.4
Range 9.5—18.2 9.5-16.8 9.8-16.9 9.0-17.2
10] n 169 201 126 138
Mean (SD) 13.8 (2.2) 13.8 (1.8) 13.4 (2.1) 13.3 (2.5)
Median 13.9 14.1 13.2 13.1
Range 9.8—17.4 10.8-16.9 10.2-16.9 9.4—17.3
[02 n 222 134 124 1 11
Mean (SD) 13.0 (2.3) 13.3 (2.2) 13.2 (2.6) 13.4 (2.3)
Median 12.9 13.2 13.6 13.6
Range 9.0—17.6 9.8-16.9 9.0—17.2 9.8—16.9
MOR-1 n 246 133 182 94
Mean (SD) 13.5 (2.3) 13.2 (2.5) 12.8 (2.5) 14.4 (1.9)
Median 13.7 13 12.9 14.7
Range 9.3-17.4 9.2-17.0 8.8-16.8 11.0—17.5
MOR-2 n 219 104 145 109
Mean (SD) 13.7 (2.1) 13.8 (2.5) 13.7 (2.4) 13.7 (2.1)
Median 13.9 13.7 14.1 13.9
Range 9.8-17.4 9.6-17.7 9.3—17.2 10.3—17.1

 

155

Table A3. Results of a standard least squares linear model for main effects and
interactions of measurement periods (n = 4), site (n = 6) and species (n = 8) on
measurement times of leaf-level gas-exchange.

 

 

 

ANOVA effects d.f. SS F P

Measurement period 3 45.42 2.89 0.034
Site 5 74.88 2.86 0.014
Measurement period X Site 15 424.36 5.40 P < 0.0001
Species 7 3.88 0.1 1 0.998

 

Note: Overall model, P < 0.0001; Adjusted R2 = 0.022442. Weak
interactions (P > 0.25) were removed from the model (Bancroft, 1964).

156

Table A.4. Summary of soil moisture categories for gas-exchange analyses. Categories
were based on variation in volumetric soil moisture which was measured concurrently
with gas-exchange measurements during the 2002 growing season across seedling
transplant plots.

 

 

Volumetric soil moisture (%)

 

Soil moisture

 

category Sampling dates n mean SD Min Max
Very low 11-23 July 280 3.3 2.2 1.1 12.1
Low 24 July-8 August 280 4.2 2.4 0.9 12.4
Moderate 12-30 August 224 6.8 4.1 1.6 18.1
High 9—25 June 248 11.2 5.2 4.5 32.1

 

157

Table A5. Results of a standard least squares mixed linear model for main effects and
interactions of canopy Openness (%), site (n = 6) and sampling date (n = 6) on
gravimetric soil moisture (%) across seedling transplant plots.

 

 

Whole-model

 

 

ANOVA effects d.f. 35 F P P Adj. R2
site 5 2.5 19.4 < 0.0001 < 0.0001 0.64
date 5 5.5 43.4 < 0.0001
site x date 25 1.0 1.6 0.0387

loglo canopy openness (%) 1 0.1 4.0 0.0461

Note: Models exclude interaction terms when P > 0.25 in preliminary model (Bancroft,
1964)

158

Table A6. Results of a standard least squares mixed mode] for the main effects and

interactions of log”) (whole-plant mass) (g), site (n = 6) and species (n = 8) on loglo (root
mass) (g). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002.

 

 

 

 

Whole-model
ANOVA effects d.f. SS F P P AdJ- R2
Site 5 0.54 13.63 < 0.0001 < 0.0001 0.98
Species 7 2.33 42.02 < 0.0001
Site x species 35 0.69 2.48 < 0.0001
loglowhole-plant mass 1 66.46 8399.10 < 0.0001
Site >< loglo whole-plant mass 5 0,03 0.68 0.6389
Species X loglo whole-plant mass 7 0.15 2.65 0.0102
Site >< species X logo whole-plant mass 35 0,5] 1,83 0,0024

 

Table A.7. Results of a standard least squares mixed model for the main effects and
interactions of logto (root mass) (g), site (n = 6) and species (n = 8) on logo (root area)

(cmz). The model is based on data are from all individual seedlings that were harvested
from transplant plots in June 2002.

 

 

 

 

Whole-model
ANOVA effects d.f. 55 F P P Adj. R2
Site 5 0.56 5.56 < 0.0001 < 0.0001 0.87
Species 7 7.70 54.41 < 0.0001
Site X species 35 1.77 2.50 < 0.0001
IOgIO 1‘00t mass 1 47.51 2350.07 < 0.0001
Site x loglo root mass 5 0.1 1 1.08 0.3669
Species x loglo root mass 7 0.80 5.65 < 0.0001
Site x species x loglo root mass 35 1,01 1.43 0.0499

 

159

Table A8. Results of a standard least squares mixed model for the main effects and

interactions of loglo (whole-plant mass) (g), site (n = 6) and species (n = 8) on logo (root
depth) (cm). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002.

 

 

 

 

Whole-model
ANOVA effects d.f. 55 F P P Adj. R2
Site 5 0.52 3.18 0.0074 < 0.0001 0.55
Species 7 0.53 2.30 0.0250
Site X species 35 1.87 1.63 0.0116
Ioglowhole-plant mass 1 10.13 309.32 < 0.0001
Site X loglo whole-plant mass 5 0,19 1.16 0.3259
Species X loglo whole-plant mass 7 1,43 623 < 0,000]

Site X species X loglo whole-plant mass 35 1,85 1.61 0.0135

 

Table A.9. Results of a standard least squares mixed model for the main effects and
interactions of loglo (whole-plant mass) (g), site (11 = 6) and species (n = 8) on loglo (leaf

area) (cmz). The model is based on data are from all individual seedlings that were
harvested from transplant plots in June 2002.

 

 

 

 

Whole-model
ANOVA effects d.f. 88 F P P Adi. R2
Site 5 5.50 14.15 < 0.0001 < 0.0001 0.75
Species 7 15.16 27.85 < 0.0001
Site X species 35 4.46 1.64 0.01 10
loglowhole-plant mass 1 102.86 1322.70 < 0.0001
Site X Ioglo whole-plant mass 5 0.69 1.79 0.1 129
Species X loglo whole-plant mass 7 2,07 3.81 0.0004
Site X species X loglo whole-plant mass 35 3,27 1,20 0,197]

 

160

. 292mg 2: 89¢ 38:83. 3.5 3885332: 533%
.3 v 5 08m 0388 28338 8 25 .wmod M m .5545 H 058% $3 ”360% $83 ego—m 80888 8.“ “mob 3ch

 

 

 

 

3.3 88.5 v 231 .42.? 8251 8: .8: $3 3: n 5 .208... E.
a; 885 v 831 .321 :31 ES .32: 33 A: u 5 38.5.8 2.53m
m .3 88.5 v 331 .531 $31 $8 .23 em: a... u 5 3.23.83. .8...
a... 885 v So... .821 and: 83 4.3.5 32 E n 5 see: as.
8.3 885 v 831 .831 £31 £2 .82 2.3 E u 5 829...... 35.8
~35 585 v ”:51 .32.. N8? 2 2 .82 83 5. u 5 Se: 2888
3.2 88.0 v :3: .231 831 E: .33 on: AS 1 5 use 32.3
e53 885 v 8on $31 :21 ES .25... 33 an n 5 355.... 2356

Na 2.35 £0 $3 $3 20 $3 a 868m

.88 2.2 5 32..

8.2388 80¢ @8825: 203 35 $8608 86368 =a 88m 0.8 8% no women 0.8 womb“: .Aozm £8380 38964—035 _ 30
8.33 38988 36on =w 28 8603 3:239: Sm A3 mmm8 88 28 g mmmE Enid—0:3 88253 82823—2 36:: mo_-wo_
Sm 8835ch mo 3m“: 33688 can 28288 8:08.88 .85 823.63 .63 mono—m 8282on “88 8.88 382385 .o_.< 2an

161

.386 v 8 .vaNB 1 888m “m8 8:888 @335 main 88on 83253
80333 8 855 How 881—. .ood n m 4.38:” 0338 88 828% $88 one? 8888 8m “8% 3ch

 

 

 

 

$08 888 v 2.81 .881 881 83 .83 ”8.. 8m 1 5 838m .2
888 888 v 881 .8881 :81 88 .23. 88 3 1 5 88.5288 55%
88 888 v 881 83.81 0881 23 .88 3.3 8 1 5 383.5 8555
8:8 888 v 281 .881 381 mm: .288 :3 6m 1 5 53.888 .8“.
$8 888 v 888 .8881 88.81 E: .88 83 3 1 5 .53.: see.
58 888 v 881 .881 8.81 83 .28 83 A8 1 5 552% .355
88 888 v 881 .8881 $81 83 .88 888 E. 1 5 as... .888
88 88.8 v 231 .881 881 23 .888 :3 E 1 5 ese 83.8
838 888 v 8281 .881 881 82 .28 83 am 1 5 3.259. 2.228

mm e234 20 :8 a mod 20 £8 a 838m

.88 82 5 .63

88388.: 80¢ 3803: 803 85 88:38 .82sz :8 80¢ 08 88 co 3mg 08 8928?. A88 £8258 3888-83 N30
823 38388 88on :8 28 880% 186388 8m 85 $88 88 98 A3 888 883-283 :8an 83803—8 80:: woﬁwo.
8m 888$ch mo 388 38888 88 28388 882.88 A»: 8038.0 .93 mono—m 88888 88 888 388885 ._ _.< 33,—.

162

.39: n m .mm~.§ H 888m :88 ”86on $0.88 82m 8888 8m “8% 382
$2 :82 v 821 .821 :21 $2 .82 :2 8m 1 5 830% 8..
82 :52 v 821 .821 821 22 .82 82 A: 1 5 8:83:88 558
22 882 v 321 .321 821 83 .82 88.: A? 1 5 888:6 3.535

 

 

 

 

82 .82 v 221 .821 821 2 2 £2 83 8 1 5 53385. .8:
882 882 v :21 .821 #21 83.83 a: an 1 5 :38: as:
882 882 v 8821 .821 821 82 .83 :1 ._ 5. 1 5 353...: 2.5::
82 882 v 821 .821 £21 82 .82 82 5. 1 5 Ba: 8&8
82 882 v 821 .821 E21 82 .82 83 a: 1 5 2% 8&8
w _2 :52 v 821 .821 £21 83 .82 9.2 E. 1 5 2.258 8:8

N: 02“; m8 :8 a 8.: m8 :3 a 868m

.88 as: a $2.:

88888 80¢ 8:88: 803 88 88:88 88368 :2 88m 2: 88 :o woman 08 mama—84¢ .85 88:8 8_ 8:881:25 w
2 8885 8:588 880% =n 8m 8808 838368 :8 EV 8:8 woo: 8:“ EV 8:8 82:12:85 80258 838252 :88— onwo—
:8 8:858? mo 88: 8:888 85 88.88 8:08.88 .85 8888—0 .93 8:0? 88888: min 8.88 BEE—wcgm .m_.< 28...

163

.wood H K .893 n oumuﬂm $8 “$6on $88 ago—m .8888 8% “mob “8:2

 

 

 

 

mad 88.5 v oo~.o:.2~.o: ES: 3... $2 $2 5n n 5 860% =<
23 88.5 v 555 .22: $3: 5: .33 33 A E n 5 53323323 55%
:3 88.5 v ~25: ._ So: :3: $2 .535 83 A? u 5 5.82%: §5§é
M33 885 v Sod: .83: 33: Na: .23 $2 5 u 5 53.38% 3%
S2 885 v 33 4.8.5: 83: 82 68.5 g: E u 5 5:33 5%
$3 585 v 2&5: .325: $2: $2 .82 83 8m : 5 353$, ,355
83 585 v 3:5: .33: 83: a: .235 ME: 9% u 5 5&2 $226
58.5 88.5 v :25: :22: 83: 2 2 .83 8.3 :m u 5 2% £2§®
$3 885 v ES: .23: 22.5: on: .235 on: 3 u 5 355% “Sega

Na 335 £0 £3 a we; 20 $8 a 860mm

.88 22 5 3%

28328.: Set Bantu..— 203 35 $5608 1323?: =m 82m 03 Saw :0 woman 98 mug—mg .35 88:8 02 woﬁwmbmbsmv N
2 £55, 3:588 330% =m was 860% :32sz 8m Amy mama «ooh was A3 mmnE «Swab—0:3 5933 35328—2 80:: onmo—
8m oogomﬁmﬁ mo Smog 8330me can £5285 oococucoo .55 gouge—u .03 mono—m .863qu mg .8me voﬁcavcﬁm .m_.< 2an

164

.3335. on. :8: 38288 3.5 8.3833392.» 333%
.G v 5 05m 0383 03:33.5 8 25 .258 H m 8.3 u 03.33... 3.8 ”$8on $083 ego—m 88:80 8.: 30,—. 382

 

 

 

 

8... 58... v $2.: .32.: $2.: 22 .3... a... 6mm : 5 8.8... =<
3... 2.8... v Ed: .32.: o. 2.: E... .8... o... G... n 5 382...... £53...
.3... 88... v 22.: :85: $2.: 2.... an... 3... cm H 5 53.3.32... 3....
2.... 2.2.... v 22.: «3...: 32.: mm. .32. 8.. an n 5 5.2.... .8...
mm... 58... v :55: .82.: :3: 8m. .3... am. g n 5 use... .55....
a... 2.8... v Ed: 58...: .52.: at... .82. 8... a... u 5 2...: 2.22.8
Na... 88... v 82.: .32.: 82.: mm... .82. a... a. u 5 3... 2.22.8
.5... 88... v 8...: .Ma. .5: Ed: 2 ... .98... 8.. 2 m n 5 s........... 2.22.8

NM. 08...... 2.0 .me .33 £0 :8 a 3.8mm

.88 2.3 a. .3...

23388. :8: 3.2323: 083 3:. 35:88 32:39: _3 :8: 03 m3“. 8 woman 93 8335.. 8:... 0:388 80:38:33. _ M02
53:? 3:568 330% _3 33 3:8on 33:39: 8.: Aw. 83:. 88 93 Am. 938 233-2933 .8038: 3:58:33. .32:— onwo—
8.: 8.32.:in 8 3%... 3388.3 33 3338.: 8:08:80 .8. 8233.3 .68 mono—m 8638...... 3.8 8.88 conﬁumunﬁm .3 .< 23,—.

165

.83: n K .mmmdm n 0:333 38 328% 3883 283 80888 8.3 amok 882

 

 

 

 

33 883 v 33: .23: 88: G: .33 a 2 a: u 5 3:23 5..
$3 585 v 38: .33: m5? 3 _ ._ .33 38.— SM n 5 ”3833:3323 553
Rod 38: v 33: .58: 53: E 2 .83 B: an n 5 3.85:5 33¢
23 885 v 83: .32: 83: on: .53 82 an n 5 55383 33
was 88¢ v 83: .82: $3: 82 .83 2 2 :3 u 5 55:: as...
3.3 88.3 v 53: .33.: 23: a: .33 £3 :3 n 5 3.222. 33$
83 823 v 83: .32.: $2: 33 .83 53 as u 5 Es: 32%:
ES 88.0 v 3.8: .32: 23: $2 .53 :3 G. u 5 2% 3:228
:3 885 v RS: .32: 825: 83 .33 825 A? u 5 3:55 2.2%:

N: 8.3: 20 £3 a Me: 2.0 $3 a 838m

.88 us: 5 :2:

83383: 8:: 9:83: 033 35 38:33 3:239: _3 88.: 03 33: :o 933: 03 mama—38¢. ASE 0:388 :o3wEm-::mv N :02
82:? 98388 83on _3 93 338% 3:939: :8 A3 938 88 93 A3 338 83333:? 800250: 33:80:33: 32:— onwo.
.8 832.933 .8 mama“ 33383 93 m3>383 3:09.88 .85 82353 63 mono—m 8:38on 33 838 wouﬁugaﬁm .m_.< 33:.

166

.mmmbmcm 05 Set vows—88 mes ﬂatmES‘Muza 55mm

.3 v 5 0N6 295% 33322: 3 can ._ Ed u m .393 H 2693 “m3 630on awoken ego—m :oEEoo Sm amok "8:2

 

 

 

 

53 88.5 V S3 484 on: 83 “Ed 23 E: u 5 360% =<
$3 885 v «3% .2 3 33 ES .83 S 2 A: u 5 sausage 335
35 88.5 v 32 $3 8mm 83 $3 33 an n 5 EEEBS. 5%
53 v85 5% if $3 $3 $35 $3 5 n 5 53$: 3%
SS 885 v £3 .2: 89m $2 .225 33 As H 5 353:: 35m
32 585 v 23 .23 :2 98d .83 £2 5 u 5 ES: 3:33
a; 885 v 23 .8: an: an. .88 82 5 u 5 as: .2853
$25 885 v on: .8: mm: mm: .83 82 em u 5 3.25% §§m

Na 835 2.0 $3 a we; 30 :3 a 332$

.88 as:

167

E 303 Enigma 88,“ 38032 203 35 3:558 3:232: =m Eoﬁ Ba 8% :o woman 08 mug—mg ABE £3330 BEE—IEBV

_ 30 £53» 3:388 38on =m EB 36on 7323?: 8m $83 3.3 88 98 35 39b 88 5233 823252 80:: wOTwo—
8m oogocEwE mo 88. 3.2083 98 28,82: oocowmcoo .85 £89320 Adv mono? :33“:on mus 5.38 wonmvacaﬁm .o _ .< Bomb

.585 v m .838 u 23%; a2 555% 335

$5: 360% 50253 cargo—o 5 3:5 Sm mach domd u m :26 n ozmnﬁm 68 ”momooam mmohom ego—m .8883 8% «web 382

 

 

 

 

SS 88.5 v 83 we: 32 $2 .555 55.5 :8 n 5 830% =<
33 58.5 v 33 .83 23 $3 .Ed 33 a: n 5 E§=§§B 55%
was 58.5 v 83 .83 £3 £2 .93 32 Ga u 5 882%: 33¢
53.5 88.5 v 83 $2 ”Nam SS .23 mm: as. u 5 £35583 3%
3:5 88.5 v 33 .33 82 am? .m :5 :55 A2 u 5 5:55 53
:3 38.5 v M32 .53 83. $2 .88 $2. am u 5 3:23. §§£
5.2 88.5 v as: .3: N3: 32 .33 2 2 :v u 5 2%: 322$
$3 :55 E: .8: 3: oz: .825 23 c. n 5 2% £2.55
223 88.5 v 23 $2 on: 23 £55 M: 2 am n 5 $.55: ”322$

% 85¢ 20 $3 5 mm: 2U £3 a 360mm

.88 22

N .30 £53» 3:388 36on =« can 360% 1323?: 8a A
8.5 852.2%? mo mama“ @3300me was 28:85 vacuum—Bu .85 3285—0 .68 memo—m ~868qu 2x5 8.88 voﬁcauzﬁm .24 2an

N

E 803 “cm—awash Sea @3822 22$ :2: $5803. $3365 =m 80¢ 03 Sat :0 woman 08 mowing .onm £3330 quwETnsmv
E3 88 88 98 35 $9: 88 2833 3298332 30:: onm2

168

.mood H & .mmm.: H otmsﬂm «m8 528% $88 82m 8888 8% amok 382
a; 885 v at: .83 3: a; «$5 :3 8% u 5 89% =<
82 x85 “:3 at: 88 3: .22 83 An N 5 8232332: 55mm
83 58.5 v :3 .E _ .N $3 $2 .83 82 A? u 5 3882: 325$

 

 

 

 

a; 88.5 v 82 .22 $2 32 .8: 32 am u 5 235388 3%
83 88.5 v 32 ax: 8: ”So .585 88 GN u 5 535: Law
:2 88.5 v $2 .83 Rod 28 .35 32. E. u 5 552% @355
$2 58.5 v E; .33 a: 88 .22 83 5 n 5 Es: 23:58
28 885 v :2 .33 E: 2.2 .88 8.3 a: u 5 33 £88
83 58.5 v £2 .8: $2 $3 .83 38 as u 5 555% £228

Na 2.338, 20 :3 d wﬁ 2.0 :3 a 8626

.88 as: a

32m 828888 88¢ @388; 083 85 mwczwoum 883:8 =m 89a 08 98 8 .8an Pa mug—84¢ .85 88:8 8_ 88884335
_ m: 8.33 88388 38on =m 88 88on .8239: 88m $83 88 88 can 3 $88 88 :8an 83828—2 88: 878.

.8 882.886 .8 Ema 88888 93 28238 888580 A8 2888—0 .08 88? 86882 88 888 wﬁuﬁwcﬁm .w _ .< 2an

169

.wmod u & .mmﬁm— H 0585 68 ”88on amp—om ego—m .8888 .8,“ “mob "8:2

 

 

 

 

32 88.5 v 63 .55 E: and “N55 255 GR u 5 35% .2
mm; 58.5 v $3 .33 $3 22 .83 $25 A z n 5 E§=§w§u 55%
0:3 58.5 v 33 .03.. 83 $2 £3 83 A? n 5 $32523 35¢
08.5 58.5 V SS .33 :2 $2 .5: ES 5 n 5 $3585. 5%
$2 885 v 82 .3: $2 95.. .32 SE A: u 5 555: Law
mm; 58.5 v 28 .83 :2 mm: .83 Sad am n 5 3522. 555
5:3 88.5 v £2 £2 83 $2 a”: 8: am u 5 Se: 522%
33 88.5 v 2:: .9: ma: 33 $23 as ._ :m u 5 2% £226
82 88.5 V cm: .2: E; 035 .35 ~23 g u 5 E252 3:26

mm new: 30 $3 a m3 £0 :3 a 86on

.88 83 5

32a Enigma“ 80¢ 833.3: 825 35 mwczuoom 3:232: =m 88m 08 Saw :0 8me 2w mug—22‘ ABE 83:8 03 8335-933

N n: GEES 8:388 880% =m 28 38on 3:239: 8m ANEQV 88 88 98 Amv 3%: 88 8353 32288.2 80:: onwo—
Sm came—mama mo mama“ 838me 98 $382: 882.28 A9 gouge—o Adv mono—m =23»:on mus 8qu voﬂpaucﬂm .3 .< 3an

170

. .mmmbﬁa 05 Bot wows—88 mm? ﬂwmﬂzapxwmzc 55mm
.6 v 5 onmm 295% 8382ng 3 gm .256 u & $3.3 u owmuﬁm $8 $20on 888 ego—m .5888 “on “me... 332

 

 

 

 

33 885 v 33 .83 2.3 $3 .83 39o 9.8 u 5 830% =<
:3 88.5 v 25$ $3 32 83 .23 825 3. u 5 genius.» 35¢
25.5 58.5 v 33 .32“ 2.3 E: .82 and cm H 5 533583. .8...
22. 885 v 5% .Qod 33 N8; .32. and :m n 5 SEE sow
”3.5 88.5 v on: .3..— SE SE .82. ~de AR u 5 3523 2.52m
23 88.5 v we: .2: SE at: .33 $2 8*. u 5 ES: 3:55
on: 88.5 v a: .8: £2 a? £3 ms.— az n 5 33 2.2.26
”85 88.5 v $3 .3: 8.: M: 2 .23 33 am n 5 8.3% 2.25m

mm 2.35 2.0 :8 a MS 20 :3 a .36on

.88 as;

171

E 32m 28328: Set 33253 80>» 85 $5608 3:339: =m 80¢ 03 83 no woman 0.8 3ng .Aozm 0589: 32884—255
_ MOE £55, 3:588 360% :5 was 862% 152365 8m $83 was 88 Ba A3 39: 88 5233 82223—2 30:: wo_-wo_
Sm oogomima mo memo“ 38383 98 mug—85 3:09.28 .85 8285.0 .98 mono—m commmewa was SEE woﬂuumucﬂm .o~.< 033.

.mmod H & .mom.m_ n 0593 “m8 533% $28 ego—m cos—Eco Sm “mob Baez

 

 

 

 

82 888 v 83 .83 33 £3 .88 :2. a: u 5 860% =<
”3.8 88.8 v 33 .8; 8mm 23 .83 88.8 SM n 5 E55533 55%
Ed 88.8 v :3 .82 EN 83 .988 03.8 § n 5 €855.» §§§m
82 88.8 v «.88 ﬁg ”:8 a? .888 8: an n 5 53.38% .8...
5.8 88.8 v 82 .33 83 83 «88 M88 :3 n 5 £33.: Ex
88 88.8 v 23 .83 $2 $3 .82 88.8 :8 u 5 sees. 2.55
was 88.8 v 3: .EE 8: new; .83 :3 cm H 5 Ba: 25.25
83 58.8 v 83 .8: 8M: 82 .268 28.8 § u 5 2% 322$
885 :58 $3 .83 m8; 23 .888 88; 8 u 5 3.55.: 322$

Na 029$ 30 £2. a mﬁ 2d :3 a 362%

.88 83

N M02 553» 3:588 $6on E» was 8603 3:229: Sm A

N

5 303 Enigma €on 3833: 803 “m5 mwczcuom .3232: =m 80¢ 2m Emu so 3me 2m mombﬁa A81... 2:808 woameTnsmv
E8 33 88 28 A3 SS: 88 5953 3528538 and: wo_-wo_

8m oogomﬁwﬁ .8 3m“: 38683 98 2.838": ooaowmcoo .85 gouge—o .68 mono—m :23»:on mg 8.38 vuﬁvuawcﬁm Ame. 03m...

172

Ema—5:5 05 80¢ vows—86 33 5325355343 £33m
.6 v 5 86 2588 85539.5 8 can .de n K .53.? H osmuﬁm 558 528% $085 one? 22:88 8.5 “mob 3ch

 

 

 

 

5:5 585 v 53 .22 555.. 555.. .5555 555 E: n 5 8.8% 5.
:_5 $55 585 .55.... a: 555 5...: 255 A...” u 5 33.5.5. 5.5.8....
525 555 25.. .SE 55.. $.55 .32 35.. am u 5 55.58.... s...
55 £55 5555 .855: 555 $55: .355- .52: G. n 5 5.3... .8...
2:5 5555 55: 5:: .53 52 55.5 5.55 5 u 5 55.2.. 2.5....
E .5 855 5555 .355 £55 £55 .55.... 535 5. u 5 35... 2.22.5
$55 :55 5?... .5555 $2 83 .535 BE 5 n 5 .5... 3......5
$5 585 552 .2 2 85. 552 .55 58.. SN u 5 55...... 2.82.5

.5 02.3. 2.0 .555 5 53 £0 .55 a .2855

.58 2.3 a 325

5:535:55 5on 58525.: 203 555 mmE—woom 3:239: :5 80¢ 25 855 no woman can 585553.. ABE €535.30 35564—255 H 30
52:3 3:383 538% :5 can 86on 3:229: 85 955 550.5 88 55... A3 5558 353.223 :8an 32.048522 30:: mono—
85 8:525:56 no 385 525683 was 25285 8:05:28 .85 28355—0 .63 $50? 52.0.5235 mg 52.58 woﬁuawcﬁm :86. 035,—.

173

.306 u & .Nwm.m_ u oumsﬂm “mo. 620on $0.88 32m .8888 Em and. 3ch

 

 

 

 

$2 585 v 32 .22 22 ~85 .92 $2 88 n $862; .2
RS 22. m a: .52.. $8 2.2 .2 no 23 a: u 5 Eﬁsﬁméa 55mm
82. $3 :3 .32 3: £3 .23 E: Ax n 5 ES§§ 2.5.8.8
205 885 V SS .32 $3 $8 .83 83 am u 5 538.3%... :3
9:5 3:. $3 .22 32 $2 .33 $3 5 u 5 25.: Law
:3 $3 82 .28 $2 33 .82 32 am u 5 3222. 2.55
”$5 885 v 82 .5: $8 8: .965 23 E. u 5 8&2 322$
5. _ no m 85 KS .2 2 $8 82 .33 83 E u 5 2% §§m
$3 83. $2 .82 MR; om? Km; 83 a: u 5 355% §§®

mm 33¢ 20 :3 a mag £0 £3 a 8.08m

.88 08: 8 32a

174

Exam—ab 80¢ 88328— 225 85 @8808 132388 =a 88m Ba 83 :o 883 98 89:92. .386 53380 uoaawmtmbsmv N 30
8:23 3:588 36on =a new 36on 832388 88 A895 53c goo.“ 83 @ mmm8 82920:? 83252. 3280822 80:: onmo.
88 852.3%? go 3.0.8 3660me was 29:38 3:08:88 .85 8885—0 .08 mono—m 8238on v.88 8.88 wﬁcﬁvcﬁm .mm.< 033.

.Sood v K 65.3 n ocmcﬁm “m3 ”266.3 max—<3

man: 36on 50253 meagre—0 5 3:5 8m 93% .036 n m .392 n 0593 $2 ”$50on $88 32m .8888 H8 “no... 332

 

 

 

 

Es 885 v :2 .33 22 £3 .82 53 GE u 5 360% .2
9:5 .23 E: .354 $2 82 .525 mm; 5 u 5 532:5...53 53%
go... 3.5 3.3 .82 32 God .82 83 5. u 5 $3.32: 2.5555
S2 585 v £3 if a: 83 .53 82 a... n 5 53.588 5%
$2. ”8.5 an: .82 83 mm: .23 $3 €~ u 5 ESE 3%
8.3 a; a? .23 £3 82 .55 $3 an n 5 uses... 3:5
”:5 v8.5 32 .82 E: mom: dad 82 65. u 5 Es: 322%
Sod 88.5 v :2 .E: 3.2 33 £3 82 a: u 5 2% 3.5m
83 835 $2 .32 32 $2 623 $3 5. u 5 33.: 2.22%

Na 2:95 20 $3 a mag 2.0 $3 a scam

.88 22 5 3%

175

Enigma 8on 363.8: 0.83 35 3:288 3:232: =m 82.5 2a 8% no woman can mug—«Sc. .35 83:8 com BEANSAEBV _ 2
£53» 3:588 880% =m new 36on 3523?: 8a A88 Snow Hoe can Amy 32: “Swab—on? 5232. 3:35:22 80:: wo_-wo_
he oogocﬁmi mo $8“ 36683 28 £358.: 8:03:80 .85 3235.0 .38 mono—m =23»:on 38 8.88 voﬁcbﬁcﬁm 4&4 2an

.88... v a. .898 u 2.2.9... .8. 5.3.3.. SS...

mg... 8.3:... 503.2. gage—o E 3.3—m .8 £3... .32. H a. .thd— H £8.38 .8. 828% $83. ego—m 58:80 .8 80,—. 3ch

 

 

 

 

m3... 8...... v 82 .32 82 Now... .5... m . m... :8 n 5 3.02.. ..<
:2. 88... 83 .82 E... 8.... .28... 88... . E u 5 5.5.33.5... 55mm
88... 8.... 8%.. .82 8.... mm... .28... .8... 2.. u 5 2.8.5.5.. 2.5.8...
.8... 8.... 82 .88.... En. R... .88... 32 A... u 5 555.8% E...
8.... 8...... 82 .82 on... 22 .82. 82. .2 n 5 5.28.: .8:
88... 5.... 55.. .22 3... 8.... .82. 8.... as u 5 952% 2.5::
88... 82... 8... .28.. 8.... 82 .82. 3.... A... u 5 8...: 322$
22. 88... .52 .8... 82 3.2 .88... So... 2 m n 5 8... $226
3.... .8... 22 .m. . .. $2 82 .22. 3.8... .3. n 5 $25.... 2.22:.

N: 02?-.. 3.0 :2. a mo. 20 :3 a 8.8mm

.88. 25. a. 32..

176

.5185... Eo... 383...: 0.83 .2. $5603 .9636... :8. So... 0.: 8% :o 68.3 o... mama—«:4... A88 88:8 3. 38955:... N 0.
E589 3:588 3.8% :5 :5. 8.8% 12636:. .8 2:3 5:2. .8. 6:: Am. 3:... 283-2023 :85»... 3532.20. .32.: onmo—
.8 oogomﬁwi mo 8%. 3868mm 9... £5,855 3:02.38 .8. 2.2853 58 mono—m :ommmoao. mt... 8.8.: 3:63:55 .mm.< 2an

8.9.88 0... 80... 68628.... 83 8.88.8888...» 8.8%
.3 v 5 8.... 8.888. 08:86.8. 6. 25 .mod n & .mmw.m. H 28.8.. .8. ”8.8.... 80.8 ego—m 8.888 .8 .8 H 8.97.

 

 

 

 

8m... 8...... v 82 .82 82 .8... 8...... 8.... .8. u 5 8.8.... ..<
8.... 8.... $2 .8... 82 8.... 8.... 8m... .8 n 5 3.3.8.5. 2.5.8....
8.... 8...... v 82 .82 82 .8... .88... 8.... .8 u 5 559.83. .8...
.8... 8.... 8.2 .82 82 .8... 8...... 88... .8 n 5 5.28. .8...
..8.. 8.... 8.... .88... 8.... 2.8: .88...- 88.... .8 u 5 see... 852..
88... 8.... 8.... .8... .8... 3.2 .8... .8... .3. n 5 Es: .322...
8.... 8.... 82 .8... m. . .. m8... .8... .8... .8. u 5 8... .8228
8...... 8...... v 82 .8... 3.2 8.2 .8... 88.. .. m u 5 353...... 828m

N. 8.8-. 2.6 .88 a mo. 2... :8 a 8.8.8

.88 2.... a. 22..

.8388. 80.. 68.88... 8.83 8... 886.8 8.6.268 =8 80.... 08 8.86 :o 688.. 88 88.8.5. 8.8 8:888 68.86.83. . 80.).
88.3 68.288 8.8% :8 68 8.8% $563.68 .8 A80. 886 .8. 68 Am. .88 88.929...» 883.2. 8.88.8.8. .88. mo.-wo.
.8 8:88.88 .0 88. 688.888 68 £8.88. 886...:8 .8. 80.8.8.0 do. 8...... 888.8. .68 888 68.6.8688 .6N.< 838...

177

.Sod H m :36». M 0:283 38 528% $83 ego—m =9:an Sm “8% "3:2

 

 

 

 

E2 882 v 23 at; SS 32 .22 $2 a: n 5 360% =<
2.2 :52 v :2 .82 83 $2 £2 £2 8m u 5 ”55:33? 55%
2 2 882 v 83 .SZ 82 $2 $2 82 am n 5 EEEEc 35¢
32 82 $2 .82 E2 82.. .33: 22: as u 5 23585. :3
2:2 :2 82 £2 a: £3 .32 33 SN u 5 ESE 3%
:2 882 v 2:: 222 32 mm? .82 82 so u 5 358$, 555
3.2 882 v 83 .32 32 S 2 “$2 :2 cm H 5 Se: 9.3.:me
:2 382 v 82 .3: $2 23 .32 22 § u 5 2% §2§®
2.2 582 v a: .33 R: 83 $2 $2 av u 5 $252 285$

Na 02“: £0 :8 a wﬁ 2.0 $8 5 332a

.88 22 a $25

178

Edamcmb Eoﬁ 359:2 29$ 85 3:238 3:222: =m 89¢ 08 8% so woman 8a 823: .35. 05308 3335553 N MOE
5:23 3:588 86on =m can 86on 3323?: 8m 9.3 58¢ 88 can A3 368 “Swab—0:3 52,83 82223—2 80:: onwo.
Sm ounce—mama we v.32 38682 98 29:25 8:02.28 .85 22353 68 mono—m :238w2 2.8 8.38 wﬁwﬁwcﬁm .34 033.

29288 05 88a wows—axe 2.3 ﬂwzmszxwmza 55mm
.8 v 5 0N6 038% 083338 2 25 .Nood H m .N_N.mm N 93888 83 883% $28 ego—m 80888 88 “mob 3ch

 

 

 

 

$2 .82 v 3.: $2 :2 £2 .82 $2 9: n 5 W289. =<
$2 882 v 82 an: 83 82 .53 E: :5 u 5 888% 3.5.8.8
82 :52 v 3.3 .82 8: $2 .83 32 an n 5 53883. 83..
82 S2 :2 .3: 3.3 2: .22 a: A: n 5 .525: 8%
£2 82 as .N .62 9: $3 .22 83 S n 5 2.22% 33$
2.2 882 v :2 .52 $2 Rod .83 32 5. u 5 8%: 322$
82 82 :2 .82 E: 23 .3: 82 S u 5 2% 2.2%
E2 S2 85 .32: :2 23% .33 SS Gm u 5 3533. 2.25m
mm as“: 2.0 $8 a we. 20 :8 a 362%
.88 as; a. 322

179

Enigma 80¢ @8838: 225 85 38:53 8:229: =a 80¢ 2d 8% co comma 0.8 mama—m2 ABE £2380 voEEuAEBV _ 30

853» 8.8583 860% =m 98 330% 8:238: 8.5 $83 88 .82 88 Amy 28 Saab—0:3 8353 83828.2 .808. onwo—
8.“ 02.85%; no 38. 682082 can mEZBE 3:08:88 .85 meow—«>20 .Cov mono—m 8238on min 8qu voﬁmzmccﬁm .w~.< 03$.

.Sod H m 63. _ m H 058% “m8 838% $83 one—m .8888 Sm amok 382

 

 

 

 

mono 88.5 v 25 .83 3H.— 32 .33 $2 :8 u 5 860% 3.
:3 38.5 v 82 .83 83 mm: .82 a: a: u 5 535:.5823 55.5
on: 58.5 v 83 as: 83 .82 .58 $3 :5 u 5 3.8.5.5. 3.5.8.8
53 885 v :3 .55.. £3 33 .82 3.2 Gm u 5 EEEBS 3%
was Bod 89m .85 $2 83 .92 £8 5 u 5 255.: .85..
$2 885 v 83 .3: :3 NS; .33 32 am n 5 852% .355
255 38.5 v 32 .52. E: comm .83 £3 8... u 5 95.: .5226
an; 58.5 v a: .22 $3 $3 $2 mm: 3 u 5 2% £25m
$3 :55 32 .32 mm: a; .33 83 an n 5 35.5.: 35$

Na magi 20 :3 n as 20 :3 a 860mm

.88 us: a $25

180

“SN—Q88“ 80¢ BBQ/.8: 82$ 85 mwﬁﬁoom 8:222: :5 80¢ can 9% no woman 08 3928?. A86 £3330 woumwthbsmv N 30

855, 38388 86on :8 98 360% 332388 Sm $83 38 .82 98 Amy 38: 8292083 8838: 328232 808. onwo.
8m oogomei .8 3mm“ 888083 was £9..ch 8:08.88 .85 8885—0 .98 memo—m 8832on mca 8.3.8 coﬂwhmccﬁm .om.< 2an

.256 H m Katmm .I. cumuﬁm 38 $20on awoken ego—m coEEoo 8m amok 832

 

 

 

 

82. 88.5 v a: .33 $5 32 .E: a: :8 n 5 830mm =<
55 38.5 v 32 .3: 8mm 85 .33 an? 5 n 5 £33553? 33%
SS 38.5 v :3. .83 2.3 E: .5: $3 3 u 5 ESEE 33¢
~85 585 v 83 .2: £3 83 .32 $3 am u 5 £5583 5%
35 88.5 v 82 .5: 3.3 83 i: 33 GN u 5 5:33 5%
$3 885 v 93 .32 8mm 53 .83 a: A? u 5 352% .525
5.5 88.5 v 52 ,3: 32 35a 53 3% § n 5 Es: ”83$
55 83. a: .32 $3 33 .33 we: a: n 5 2% §2§®
52 88.5 v SS ,8? 33 8S .5: S3 A? u 5 3.25% £2§®

Na 8.3m 3.0 :3 a v.3 2.0 :3 a 832$

.88 22 a 3%

181

Enamcab 28¢ @8822 203 85 $5608 :32sz =a 80.: 2m 83 no woman— 08 mama—Ea .35 88:8 3m Esau—£335 _ OH
553, 3:388 36on =w 28 380% .3229: Sm $88 33 awo— ucm va mama E2922?» :3qu 338238 .805— moTwo—
8m nonwomﬁwmm ,«o 33“ 33683 98 mag—BE “zinc—Mace .85 98396.0 .38 mono—m commmoume mg 5.88 Bummzmvcﬂm .oM.< £an

.256 n a: .womém n 05:88 :8: 528% 30:8 ego—m :08800 8.: Bob "8:2

 

 

 

 

52 58.5 v 22 .32 a: $2 .22: 33: 8% u 5 862w .2
SS 885 v 83 .32 $2 32 .565 $8 A E u 5 Eﬁséwmaa 55mm
33 58.5 v 32 an: «2.: NE: .32 32 a: u 5 3.835: 2.53,:
$25 58.5 v as: .32 £2 92 .35.: £2 E. n 5 5.238% .8...
~25 85.5 no; .82 £3 :32 .33 S2 A: n 5 $3.: .8...
32 58.5 v 62 .23 $2 3.2 .3: :2 an n 5 352% 2.5::
3:5 88.5 v 82 .82 El 3% .22 82 2% u 5 ES: .5228
82: 58.5 v £2 .:2 E2 5%.: .32 v3.0. 5 n 5 use 28:8
32 88.5 v 22 .32 .52 Em: .8: $2 a: u 5 325.: 32:6

N: 2:33: 2.0 :3 a mo: 2.0 :3 a scam

.88 25: a $2:

182

338:8: 82.: 8:822 825 :2: 8:588 32:36:: E: 89.: 0:: 9:: :o 8mg 0:: mug—mg ABE 83:8 02 Bawthbsmv N U—

:233 8:388 86on =: :5: 860% 3:22:85 :0: $88 8:: (82 9:: 35 mag: 28.90355 8233 828320: 80:: wo_-wo_
:8 858$:me go $8: 8860mm: :5: 2:385 8:25:00 .9: 80396—0 .33 mono—m 868:3: ﬁx: :o_.::: vuﬁvacgﬁ .34 033.

.8388 0:: 808 8:308 83 M888§M0§0 038%
.G v 2: 08m 0388 08:00:88 0: 0:3 .:oo.o H & ._o_.om H 05:88 :8: ”8:00am 80:08 0:08 :0800 :8 :8 P "0:02

 

 

 

 

82: :82: v 82 .32 a2 :2 .88 $2 :8 u 5 88% 5..
was 58:: v 23 .32 82.: S2 .22 ~22 :3 n 5 $8.55.: 2.55::
88 82:: v 8:: .83 33 E2 .62 32 G: n 5 53883. 8:.
8:: 88.0 v 82 .33 82 S2 .82: 82 :8 u 5 53:3 $8
88:: 58.: v 38 .32 83 3%. 23¢ 82 GN n 5 3.22.: 2.5::
82: :83: v 22 .82.: 82 8.: i2 :3: 5. u 5 use 3:26
83 88.0 v 82 .82 :52 3mm .32 2.2 a: u 5 33 2.22%
32: :82: V SS .32 R2 83 .82 32 :8 n 5 058% 2.2%:

N: 28?: 2.0 :3 a mo: 2.0 $3 a 88%

.88 25: a 32:

183

82:88:: 80:: 8:88: 0:03 8:: 88:88 8:23:08 :8 808 08 88: :0 :88: 08 888:2. A08: 0880:: 00:88:03: : 80:2

88:3 8:580”: 8:00am :8 :08 8:008 8:23:08 :8 $80: 808 .80: :08 Am: 88:: 883-2085 803:0: 85888—0: 80:: w0_-w0_
:8 00808::me .:0 38: 886088 :08 88:08: 00:08.88 .8: 8288—0 .30: 8:0? :0_m8:w0: 0:8 :088 88:88:88 .Nm.< 038,—.

.256 H & £0vo u oumzﬁm “m8 ”museum mmohoa ego—m cos—Eco 8m H3.—L "8:2

 

 

 

 

:3 885 v E: .93 £3 22 £3 82 a _ m u 5 860% =<
was 885 v 33 .53 83 ~33 .83 3; 8m n 5 Eﬁsaﬁmﬁ 55mm
mm; 885 v 82 £3 33 N3: .32 32 am u 5 sautﬁa 353m
085 :2. $3 at: £3 a? at: owed an n 5 55.58% 53
53 885 v $3 .83 was if ,8? 5: am u 5 55%: saw
83 88.0 v 32 $3 22 Nam .33 :2 As H 5 322% .355
33 88.5 v 3: .32 33 a: .32 :2 cm H 5 Es: “522$
o2; ~85 m8." .2? 33 83 «NS «:2 § n 5 2% §§m
m3; 885 v $3 at; as: mm: “mm: 32 av u 5 ESE ”Esau

mm 02“: 20 $3 a we; 3.0 $3 a 33%

.88 2:; a 3%

184

Exams“: 80¢ @2839— 825 85 mwczuoom 126365 :w 80¢ 08 8% no woman 08 mama—«5‘ .356 2:89: BEEF—795$ N MOE
5:23 @2588 86on =m can 360% 3:222: Sm $83 88 .32 cam va mmmE «Swab—0:3 5953 amp—28:22 30:: onwo.
8.5 858$in mo Emu“ “583083 38 ﬂatch vacuumcoo .85 28396—0 .93 mono—m 2232on mg 8.38 woﬂnﬁcgm .mm.< 293.

Table A.34. Results of a standard least squares mixed model for the main effects and
interactions of logic (photosynthetic photon ﬂux density, PPFD) (pmol m'2 5.1), loglo

(leaf Nana) (ug cm'z) and site (n = 6) on water-use efficiency (WUE) under very low soil
moisture (see methods for soil moisture classiﬁcation scheme).

 

 

 

 

Whole-model
Adj.
ANOVA effects d.f. ss F P P R2
Site 5 52.7 16.4 <0.0001 <0.0001 0.34
Loglo PPF D (umol rn_2 s-1) 1 11.3 17.7 <0.0001
Loglo leaf N (ug cm-z) 1 0.5 0.8 0.3853

Site X logo leaf N (pg cm-z) 5 8.5 2.7 0.0237
Note: Weak interactions (P > 0.25) were removed from the model (Bancroft, 1964).

185

Table A.35. Results of a standard least squares mixed linear model for main effects and
interactions of canopy openness (%) and site (11 = 6) on gravimetric soil moisture (%)
across seedling transplant plots. Models were evaluated for ﬁve different sampling dates
(16 May, 25 June, 9 July, 25 July, 20 August, 8 September) and averaged across the 2002
growing season.

 

 

 

 

Whole-model
, 2
Date ANOVA effects d.f. SS F P P AdJ- R
16 May LOgIO canopy openness (%) 1 0.08 5.04 0.0328 0.0010 0.43
Site 5 0.20 2.45 0.0582
25 June Loglo canopy Openness (%) l 0.01 0.44 0.51 14 0.0281 0.24
Site 5 0.36 3.24 0.0197
9 July Loglo canOpy Openness (%) 1 0.03 1.22 0.2793 0.0635 0.19
Site 5 0.19 1.67 0.1737
25 July Log1o canOpy Openness (%) 1 0.01 0.81 0.3764 0.0002 0.50
Site 5 0.54 6.1 l 0.0006
20 August LOg1o canOpy Openness (%) l 0.11 3.81 0.0611 0.0001 0.51
Site 5 0.64 4.57 0.0036
8 September Loglo canopy openness (%) l 0.00 0.00 0.9538 0.0002 0.51
Site 5 l .49 6.48 0.0004
Overall Mean Loglo canOpy Openness (%) l 0.02 1.73 0.1996 < 0.0001 0.58
Site 5 0.42 6.83 0.0003

 

Note: Models exclude interaction term when P > 0.25 in preliminary
model (Bancroft, 1964).

186

Table A.36. Results of a standard least squares mixed linear model for main effects and
interactions of canopy openness (%) and site (n = 6) on in situ nitrogen mineralization
rates averaged across the 2002 growing season.

 

 

 

 

Whole-model
ANOVA effects d.f. ss F P P Adj. R2
Logto canopy openness (%) 1 0.06 2.47 0.1276 < 0.0001 0.58

Site 5 0.77 6.18 0.0006
Note: Model excludes interaction term when P > 0.25 in preliminary
model (Bancroft, 1964).

 

187

 

 

38.5 v $6 m? m can: 33 x 58.6 so. 33 x 25
88.5 Se 3d _ 9:: 33 x 59% see 263
$8.5 own NS m 5.: 33 x 25
:85 ”E N: _ A.-.“ «.8 355 aura SS
6685 62 :e m 5.5 82 33 x 25
£35 85 8.5 _ €55 ease as 33
93 58.5 v 38.5 mom 3% m 25
am ..€< m m k mm .8. seems <>oz<

 

385-6355

 

 

.36on @3838“

some do.“ 59% 88 can Guam mo mowﬂgm .2678:— 5 women m_ 3on 2? .8838 Sumo—mama? 05208 :8 8m €052:
oomv 2:59: :8 Bo. be.» $95 ﬂmoﬁczmouonq maoamﬂ :o G .l. 5 8mm 28 A83 23% good Swo— ATm ~-E BEE REA—m 53:26

x3: :Sosa ocoﬁcxmoaose Smo— mo 32880:: was ﬂooto BE: 2: no.“ 3608 @828 3853 ammo— Emccﬁm a we 338M .bm.< 038.

188

 

 

 

 

2 S 2 8 EN u 33: n a .25 E was

2. S 2 S 3.13.... n» .26 a «2?

3m x 33m u x :25 S Sm: 3»? n a :2d S S: I312. u » 2:3,: Sm 59% so. 33

ode + 32? u h .12 S 23 + EmSI u a :35 S 22 + 337 n a 2:25 Sm $5

23 n 356 u x :2... S 23 u EMS. n x : _ 3 S 2 Sn E5.

3% u 3.2 u h 52 S 2m u can: u a 526 S 3:: .. 33» u » 3S2 E 8: 8&5 .8. 2%:

pa + 398 n a :SS S E. + 30. _ m u » tam... S _.N_ + 3a.: u % ...:S.o Sm 3.2. $3-223 33

3m . 325 u x .85 S m: S 3.: I 329 u a .25 mm 3358. 33

c2853 :o_mm2mox ~— ..€< = Harman :oEmonum mm ..€< 2 9.23:3 cemmohwom «a ma; 2 £355 E02639:
_ mo: _ B _ Bo

 

 

ACNE 738 N00 .088 mama—0.6 05 .«o 33¢ 05 wESv @5650 8: .8323

m5? ATm ~-E NOD BE: .2996 05 mo xnoa 05 mat—6 mmmoﬁimouona 35:8: 8:3 A88 59% 82 ATw N80 .83 88 05025
gm ATw w .23. 3.2: “cob M32 $83 38 08.23, 88 .va mmmE ~§_a-o_o_._>> .A~.Eo w: .5 88:8 comeba— maB ”mac—BM mm 08
mow—par? Econoaowﬁ 8m SE: was 32332an Sm maoﬁcmoa .1 .m oSwE 80¢ 33:28 29:3 Swo— .m> Amoﬁoqm =m Sm 3333

BE omega =Eo>o I 3333 SE awash“ museum ”mm Baa—:28 rod .5323 owEo>m «2a Boa 3323 86on E gauged
m_ 033.? E23596 05 £88 =a E .o _ .< 235 .5959? 5 5% 8m mug—£8 .868qu 30:: .«o 353m .wm.< 2an

189

Log10 8/19/03 soil water (%)

Figure A. 1. Correlation matrix of gravimetric soil moisture (%) for the driest sampling
date from the 2001, 2002 and 2003 growing seasons. Each datum represents a plot-level

average from the seedling transplant experiment. All values were loglo transformed prior

1.2

0.9

0.6

0.3

 

 

 

 

 

O o
o o o
o0 O
0
Q59 Q) 0
00 £00 00 O
o
@0000 CD 8
o (53
O O n = 34 O O n = 34
p < 0.0001 P < 0.0001
0 r= 0.76 o r= 0-75
0.3 0.6 0.9 1.2
1.2 - O
Log1o7/25/02 soil water (%) 0'9 " O o O
o
O O O 00
0% O
0.6 +- (D
g) 8 O n = 35
o 00 P < 0.0001
0.3 _o r= 0.84
0.3 0.6 1.2

Log-10 8/1/01 soil water (%)

to analysis. Sample size, Pearson’s correlation coefﬁcients and signiﬁcance of
coefﬁcients are shown in each respective panel.

190

 

 

100 __ n = 257
P < 0.0001 y

80 _ r=0.67 §
0
O

Survwal (%) 7/01-10/02
A O)
O O
o
o 80
53000
ocm o
‘ 6123
o
CDCCDDO o 030 CDCDGIIOCCD o

 

 

 

Survival (%): 6/02-10/02

Figure A.2. Correlation between seedling survival (%) recorded after the peak of the
drought in 2002 (6/02-10/02) and seedling survival (%) throughout the duration of the
seedling transplant experiment (7/01—10/02). Each datum represents a species—speciﬁc
average of seedling survival at the plot level. Sample size, Pearson’s correlation
coefﬁcients and signiﬁcance of coefficients are shown within the panel of the graph.

191

Figure A.3. Correlations of leaf area ratio (LAR), speciﬁc root area (SRA), root mass
ratio (RMR), root depth and root surface area with total plant mass. Sample size,
Pearson’s correlation coefﬁcients and signiﬁcance of coefﬁcients are shown next to each
respective panel.

192

Figure A3.

2.8

- 2 -1
Log108RA(cng1) Log1o LAR (cm 9 )

1.2
0.9

RMR (g 9")

Log") root area (cmz) Log1o root depth (cm)

 

1.50 - '
0.75 ~

0.00 .

225 . .:".'.-"'

 

 

2.0 -
1.6 -

2.4 . '- “'11.. 4*

n = 1476
P< 0.0001
r= —0.14

n = 1441
P < 0.0001

"-“._ r=-0.46

 

0.6 -

n = 1498
P < 0.0001

I' r=0.41

0.3- "4

 

1.50 .

0.75- » ' ' H

 

 

2.25 -

1.50 ~

 

 

0'75 _ h ,.

 

 

-1.50 -0.75 0.00 0.75

n = 1451
P < 0.0001
r= 0.70

n = 1468
P < 0.0001
r= 0.89

Log1o whole-plant mass (9)

193

5.00
3.75
2.50
1.25

0.00
5.00

3.75
2.50
1.25
0.00

Whole -plant mass (9)

3.75
2.50
1.25
0.00

Well-drained

Sub-irrigated

 

OW1

_ P

b 5 -

5 H
.— 7 4 '
4 g 5
NH mrimrﬁ;

 

5.00 F

IC1

l— I—

_ 6
. .
l

6
6.
.‘JH 66
WW £éﬁﬂi

 

 

MOR 1

" 6

_

 

 

.— 6
66' "6 7‘5
’Wﬂﬂaaééé’WMHégémm

 

 

 

QanQrPsArAsFaBa

. Drought
H|9h<——— Tolerance— ’ Low

Qv Qa Qr Ps Ar As Fa Ba

. Drought
Hugh<—— Tolerance-—> Low

Xeric

 

Mesic

Figure A.4. Species-level means (3: SD) of whole-plant mass (g) across ﬁeld sites from
the June 02 seedling harvest. Numbers above bars denote sample sizes (number of plots)
for each respective species. Sites are separated into well-drained (outwash = OW 1, ice
contact = 1C 1, moraine = MOR l) and sub-irrigated categories (outwash = OW 2, ice
contact = IC 2, moraine = MOR 2) (see methods for details) and are organized top to
bottom from the most xeric to the most mesic. Species are arranged in order of their
drought tolerance. Species abbreviations are as follows: Qv = Quercus velutina, Qa =
Quercus alba, Qr = Quercus rubra, Ps = Prunus serotina, Ar = Acer rubrum, As = Acer
saccharum, Fa = F raxinus americana, Ba = Betula alleghaniensis.

194

Well-drained Sub-irrigated

 

 

 

 

 

  

 

 

 

 

0W1 0W2 Xeric
120- ~ A
80- _ _
‘“ﬂﬂﬂ4mﬂ2'ﬂﬂ Hﬂ*
.20 555mm45l’1'l sssmﬁssl—a
N
E IC1 ICZ
8.120- -
(U
9
a so» .
a) _ 5 .
8 ‘ .. -_ f ‘
m ‘* 6 ‘ 5 5W1;
3 o eeeﬂl’i‘lmﬂx—gﬁ 666%rﬁ56-
m° MOR1 MOR2 l
120- . - 7 ,
- i . .7
80'. . i 5 _ - l E
; 2 2 I : ‘ l ' v
40" i f b ' t 4' E
o 6 6 6FT“ll—:l 6 l 6 ‘ 6mri—I6 l m esnc

 

Qv Qa Qr Ps Ar As Fa Ba Qv Qa Qr PsArAsFa Ba

_ Drought Drou ht
H'ghé—Tolerancee Low Highe—Toleragnce—é Low

Figure A.5. Species-level means (i SD) of root surface area (cm?) across ﬁeld sites from
the June 02 seedling harvest. Numbers within or above bars denote sample sizes (number
of plots) for each respective species. Sites are separated into well-drained (outwash =
OW 1, ice contact = IC 1, moraine = MOR 1) and sub-irrigated categories (outwash =
OW 2, ice contact = IC 2, moraine = MOR 2) (see methods for details). Species are
arranged in order of their drought tolerance. Species abbreviations are as follows: Qv =
Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P3 = Prunus serotina, Ar =

Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba = Betula
alleghaniensis.

195

Figure A.6. Root mass ratio (g g_]) expressed as species-level means (:t SD) and as
estimates at a common whole-plant mass (see materials and methods) across ﬁeld sites
from the June 02 seedling harvest. Numbers within or above bars denote sample sizes
(i.e., number of plots) for each respective species. (n.d.) indicates that the speciﬁed
common mass was beyond the range of individuals for a given species >< site
combination. Sites are arranged top to bottom from the most xeric site to the most mesic
site. Species are arranged in order of their drought tolerance. Species abbreviations are
as follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis.

196

Figure A.6. Common harvest time Common whole-plant mass: 0.5 g

 

 

 

0.9 OW1 _Ow1 Xeric
0.6. - ' _ _ A
0.0 5 5 5 4 4 4 5 2 , nd nd
0.9-0W2 -ow2

0.6— - ., , , l—

”-llllll 1mm

33» IC51 5 5 5 5 5 5 l—5—i - |c17 7 nd nd [—i

 

 

 

0.3- ﬂ”
6066666665 pm pm

 

 

 

 

 

 

 

 

 

 

 

TO)

5?)

o: 09 IC2 _IC2

2

m .
0.6» fr 7 . ___
0.3- Hﬂﬂﬂﬂ _ UN
00 6 6 6 5 5 5 e 5 nd nd
0:9 ' MOR1_MOR1 —
0.6- "'— > _ __
03— Hﬂﬂﬂ _ NH
00 6 6 6 5 5 5 6m , 7nd
oj9_M0R2 -MOR2
0.6

O
(a)

 

WWW UMUBUD .1.

. Drought _ Drought
Hugh E Tolerance——) Low “'9'“:— Toleranoe—‘> Low

 

 

 

 

197

Figure A.7. Speciﬁc root area (cm2 g_l) expressed as species-level means (i SD) and as
estimates at a common root mass (see materials and methods) across ﬁeld sites from the
June 02 seedling harvest. Numbers within or above bars denote sample sizes (i.e.,
number of plots) for each respective species. (n.d.) indicates that the speciﬁed common
mass was beyond the range of individuals for a given species X site combination. Sites
are arranged top to bottom from the most xeric site to the most mesic site. Species are
arranged in order of their drought tolerance. Species abbreviations are as follows: Qv =
Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, P5 = Prunus serotina, Ar =
Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba = Betula
alleghaniensis.

198

Figure A.7.

Speciﬁc root area (cm2 9'1)

525

350

175

525

350

175 '-

D

525 -

350

175-

0
525

350

175-

O

525

350

175 -

O

525

350

17

0

Common harvest time

Common root mass: 0.3 g

 

-OW1

aaamﬂmﬂﬂ

lmw1

l

l-

F‘indr—ir'ﬁ

 

l

0W2

-OW2

 

 

 

 

01
l

 

mammUmDU mMMHMWﬂm
|01 _IC1
6 § 6 mini—aim 5 r—Iljl—ll—H—H—Iﬂ ﬂd__
_IC _m2
Wﬂmﬂﬂﬂﬂ 5 F—Wr—H‘ﬁﬂ ndl—lﬂ nd
FMOR1 -MOR1
.éaﬂﬂﬂﬂg mmﬂmﬂnﬂm
-MOR2 ‘ _MOR2

e . H 4 , -

H%[;H:Jﬁligi6 r—ﬂ_7r—J—1nd[—Hj]{1

 

 

Medc

 

 

Qv Qa Qr Ps Ar As Fa

Dor ught
H'9h< Tolerance_ _>

Ba

Low

Qv Qa Qr Ps Ar As Fa Ba

. Drought
HIgh<—— Tolemnce—_> Low

199

Very low Low Moderate High

 

5n50W1
awii 3

1.75 i ’ I a]
33% one L i
sm~ - f5? §?T 5

 

 

 

 

51
1.75;?! .1 A '. .7 «in. i1

;‘si‘3 ‘52 53‘15243 45
0'2‘3IC1

 

 

 

 

1.75 >

om ‘
5.25 | 2 6

3% § 6
in i

om
nsMO

am '6 g 5
in ??

EB?»MOR

am 656 5 6 T 6 6
6 6 7

in i6 6457? i116i5 ii [i

o o i i I l l i I 1 I I

QﬂﬂrPsArAsFaBa QﬂdDrPsArAsFaBa QxQﬂorPsArAstBa QﬂaQrPsArAsFaBa

Drought Drought ought Drought
HigheToleram+LowHigh<-T°leranoe->Law High<- T oieranoe-H'ow High<- T 0| eran oe->Low

3.50 r a 6
i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.—.—«c,
a) »-—.
\f—ﬁ

0. .—
31

5'

9

O

 

Figure A.8. Species-level means (i SD) of leaf-level photosynthesis (Anna) across ﬁeld
sites at four sampling dates that contrasted in volumetric soil moisture content (%) during
the 2002 growing season (left to right, very low = 3.3%; low = 4.2%; moderate = 6.8%;
high = 11.2%; and see also Appendix,Table A.4, for additional details). Numbers
contained within or above bars denote sample sizes (number of plots) for each respective
species. Sites are organized top to bottom from the most xeric to the most mesic (well-
drained sites, outwash = OW 1; ice contact = [C 1; moraine = MOR 1; and sub-irrigated
sites, outwash = OW 2; ice contact = IC 2; moraine = MOR 2) (see methods for details).
Species are arranged in order of their drought tolerance. Species abbreviations are as
follows: Qv = Quercus velutina, Qa = Quercus alba, Qr = Quercus rubra, Ps = Prunus
serotina, Ar = Acer rubrum, As = Acer saccharum, Fa = F raxinus americana, Ba =
Betula alleghaniensis.

200

Very low Low Moderate High

 

 

 

 

610W1 ’ l ‘ 'Xeric
4E E ‘ 5 4 i A
+-fL '2. 1 i ‘
H‘El‘lﬁipi 5344 3 i554.
pinsﬂj WE 5QWD+ mi Eﬂmému m{mrﬂmhﬂi

5

g: ‘1 5+

66 5466
6 66 5 6 671 rhiﬁrgirhrﬁiﬁﬂi

 

WUE (mmol coz moI'1 H20)

 

 

 

EEEEE3 iEEmmi géiéasi

 

 

 

 

 

 

Siihiéni EEEEEEEE EEEEE

QﬂdersArAsFﬁa QﬂdersArAsFﬁa QQaDrPsArAsFﬂa QﬂdersArAfﬁa

Drought Drought Drought Drought
Hark-Tolerance I“owH'gh‘Tderance +Low ngh"Tolerance —>Low ngh"Tolerance +Low

 

Figure A.9. Species-level means (:1: SD) of leaf-level water-use efﬁciency (WUE) across
ﬁeld sites (OW1, OW 2, IC 1, IC 2, MOR 1, MOR 2) at four sampling dates that
contrasted in volumetric soil moisture content (%) during the 2002 growing season (leﬁ
to right, very low = 3.3%; low = 4.2%; moderate = 6.8%; high = 11.2%; and see also
Appendix, Table A.4 for additional details). Numbers contained within or above bars
denote sample sizes (number of plots) for each respective species. Sites are organized
top to bottom from the most xeric to the most mesic (well-drained sites, outwash = OW
1; ice contact = IC 1; moraine = MOR 1; and sub-irrigated sites, outwash = OW 2; ice
contact = IC 2; moraine = MOR 2). Species are arranged in order of their drought
tolerance. Species abbreviations are as follows: Qv = Quercus velutina, Qa = Quercus
alba, Qr = Quercus rubra, Ps = Prunus serotina, Ar = Acer rubrum, As = Acer
saccharum, Pa = Fraxinus americana, Ba = Betula alleghaniensis.

201

Figure A.lO. Relationships of survival deviations (i.e., species average plot survival -

overall average plot survival for all species) with leaf Narea, size (whole-plant mass) and
morphological (root area, SRA, RMR, root depth) and physiological characteristics (leaf-
level photosynthesis, Aarea; water-use efﬁciency, WUE). Relationships were examined
within each of the three well-drained sites (OW 1, IC 2, MOR 1). Regression equations,

adjusted R2, P values and n for these relationships are presented in Appendix, Table
A.38.

202

Species survival deviations from
mean plot values (%)

 

 

 

 

5° "OW1 <6
25 ~ 033%
o o 00
0’ O o O o
o 00
-25 l- 8000 O OO
-50 _
1.4 1.6 1.8 2.0

Log1o whole-plant mass (9)

50

50
25

Log10 leaf N (pg cm'z)

 

OW1

 

 

 

 

-1.5 -1.0 -0.5 0.0 0.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' IC 1 (P o 0 "MOR 1
O O 0o o
.. go _ O O Q
o 0 0 o 0 o 0000 © 0
0090 C5 M
o 09 (9(9) 0 l- 00
_ o a O _ 00000 o
(I) o O O o
l O 0,9 '
1.4 1.6 1.8 2.0 1.4 1.6 1.8 2.0
Log1o leaf N (pg cm'2) Log10 leaf N (pg cm'2)
-IC 1 00 00 0 MOR 1 o
b b 0000 ($99
000(1) gEQSQ
.— 1— O8 8)
_ __ 0 0OO (9
00 o

 

 

 

 

 

 

-1.5 -1.0 -0.5 0.0 0. -1.5 -1.0 -0.5 0.0 0.5
Log1o whole-plant mass (9) L091o whole-plant mass (9)

 

 

 

 

 

 

 

 

 

 

 

 

 

"OW1 'IC 1 O OO 'MOR 1
o o
_ _ o 08 0
O O OO (O O
.. l— 069 O O
o o O
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
Log1o root surface area (cmz) Log10 root surface area (cmz) Log1o root surface area (cm2)
“CW 1 0 IC 1 MOR 1
C» 0 (ho O
_ 9 p _ (9 00
O 0 C83 00 O) O 00
l— 0%6 _ _
- 0 go o 6 0% O
O <96 r- h o 0 O
L l l 1 i 1 1 _1 l 1 l 1 l

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

 

 

 

RMR (9 9'1)

Figure A.lO.

 

 

 

 

 

0.0 0.2 0.4 0.6 0.8 1.0

RMR (9 9'1) RMR (9 9’1)

203

 

-5o [ ‘

1.5 1.8 2.1 2.4 2.7 3.0

 

l°o§ 0° ~
1‘13 3% l
L 63
iqﬁogo
l 0% ‘b
i O 00 ’

iL MOR 1
O O
o o o
00
O O 8
W250 0 i
00 O ‘
l

 

1.51.8 2.12.4 22713.0

1.51.8 2.12.42.127 3.0

 

 

 

 

 

 

 

 

 

 

 

Log1oSRA1cm2g'1) L091o SRA (cmg1) L0910 SRA (cm 91)
. LMOR101
E l 2% l
9
2 g 0 ,
.9 v i
‘6 m
'3 3 . . l
%E 0.6 09 15 0.6 0.9 1.2 1.5 0.6 09 1.2 1.5
.E % Log-lo root d1epth (cm) Log1o root depth (cm) Log1o root depth (cm)
a 8 5° ‘ow1 o 008 IC1 L O MOR1
m d) _ Gog) _ 0 (Q
2 E 25 00 o 0 8&830 o
o o o ‘ o (8;,
0 o O: _ o o o O 90
o. O O 8 0% o o
(I) O (5% 0 Q;
8 O 00 l O O O
-25 _ o oo o _ o 0
CD 0 OO O
‘50 L—A—A¥ _ E
0 1 2 3 4 0 1 2 3 4 O 1 2 3 24
Aarea(l1m0| 002 nfz s'1) Aarea (pmol 002 m "12s ) Aarea (pmol 002 m 25 1)
50 W O 083 101, MOR1
O O 8
25 l- o o
8 00%0 I ‘ O (D:QDO g) o
0 o 0 (3b 00
_ O 0%089 O O (b O O
OO O 0 (DO 0 O O (b O
- 1. O __
25 00 DC? ‘ o 0(5)
-50 — i
-1.5 0.0 1.5 3.0 4.5 -1.5 0.0 1.5 3.0 4.5 -1.5 0.0 1.5 3.0 4.5

WUE (mmol 002 mol‘1 H20) WUE (mmol 002 mol'1 H20) WUE (mmol 002 mol'1 H20)

Figure A. 10. (cont’d).

204

REFERENCES

205

REFERENCES

Aerts, R. 1990. Nutrient-use efﬁciency in evergreen and deciduous species from
heathlands. Oecologia 84:391-397.

Aerts, R. 1996. Nutrient resorption from senescing leaves of perennials: Are there general
patterns? Journal of Ecology 84:597-608.

Ahlstrom, K., H. Persson, and I. Borjesson. 1988. Fertilization in a mature scots pine
(Pinus sylvestris L.) stand-effects on ﬁne roots. Plant and Soil 106:179-190.

Albert, D. A. 1994. Regional Landscape Ecosystems of Michigan, Minnesota, and
Wisconsin: A working Map and Classiﬁcation. US. Department of Agriculture,
Forest Service, North Central Experiment Station, St. Paul, MN.

Arunachalam, A., H. N. Pandey, R. S. Tripathi, and K. Maithani. 1996. Fine root
decomposition and nutrrient mineralization patterns in a subtropical forest
following tree cutting. Forest Ecology and Management 86:141-150.

Bancroft, T. A. 1964. Analysis and inference for incompletely speciﬁed models involving
use of preliminary test(s) of signiﬁcance. Biometrics 20:427-442.

Bazzaz, F. A. 1979. Physiological ecology of plant succession. Annual Review of
Ecology and Systematics 10:351-371.

Bazzaz, F. A., and R. W. Carlton. 1982. Photosynthetic acclimation to variability in the
light environment of early and late successsional plants. Oecologia 54:313-316.

Berges, J. A. 1997. Ratios, regression statistics, and "spurious" correlations. Limnology
and Oceanography 42: 1006-1 007.

Bjorkman, O. 1981. Responses to different quantum ﬂux densities. Pages 57-107 in O. L.
Lange and P. S. Nobel, editors. Encyclopedia of Plant Physiology I. Springer,
Berlin.

Bond, W. J., and J. J. Midgley. 2003. The evolutionary ecology of sprouting in woody
plants. International Journal of Plant Sciences 164:8103-8114.

Bongers, F., L. Poorter, R. Van Rompaey, and M. P. E. Parren. 1999. Distribution of
twelve moist forest canopy tree species in Liberia and Cote d'Ivoire: response
curves to a climatic gradient. Journal of Vegetation Science 10:371-3 82.

Bradshaw, A. D. 1965. Evolutionary signiﬁcance of phenotypic plasticity in plants.
Advances in Genetics 13:115-155.

206

Bunce, J. A., L. N. Miller, and B. F. Chabot 1977. Competitive exploitation of soil water
by 5 eastern North American tree species. Botanical Gazette 138:168-173.

Burton, A. J ., K. S. Pregitzer, and N. W. Macdonald. 1993. Foliar nutrients in sugar
maple forests along a regional pollution-climate gradient. Soil Science Society of
America Journal 57:1619-1628.

Busso, C. A., J. H. Richards, and N. J. Chatterton. 1990. Nonstructural carbohydrates and
spring regrowth of 2 cool-season grasses: Interaction of drought and clipping.
Journal of Range Management 43:336-343.

Canadell, J., R. B. Jackson, J. R. Ehleringer, H. A. Mooney, O. E. Sala, and E. D.
Schulze. 1996. Maximum rooting depth of vegetation types at the global scale.
Oecologia 108:583-595.

Canham, C. D. 1985. Suppression and release during canopy recruitment in Acer
saccharum. Bulletin of the Torrey Botanical Club 112:134-145.

Canham, C. D., A. R. Berkowitz, V. R. Kelly, G. M. Lovett, S. V. Ollinger, and J.
Schnurr. 1996. Biomass allocation and multiple resource limitation in tree
seedlings. Canadian Journal of Forest Research 26:1521-1530.

Canham, C. D., R. K. Kobe, E. F. Latty, and R. L. Chazdon. 1999. Interspeciﬁc and
intraspeciﬁc variation in tree seedling survival: effects of allocation to roots
versus carbohydrate reserves. Oecologia 121:1-11.

Carlton, G. C., and F. A. Bazzaz. 1998. Resource congruence and forest regeneration
following an experimental hurricane blowdown. Ecology 79:1305-1319.

Caspersen, J. P., and R. K. Kobe. 2001. Interspeciﬁc variation in sapling mortality in
relation to growth and soil moisture. Oikos 92:160-168.

Chapin, F. S. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and
Systematics 11:233-260.

Chapin, F. S., and R. A. Kedrowski. 1983. Seasonal changes in nitrogen and phosphorus
fractions and autumn retranslocation in evergreen and deciduous taiga trees.
Ecology 64:376-391.

Chapin, F. S., E. D. Schulze, and H. A. Mooney. 1990. The ecology and economics of
storage in plants. Annual Review of Ecology and Systematics 21:423-447.

Chen, H., M. E. Harmon, J. Sexton, and B. F asth. 2002. Fine-root decomposition and N

dynamics in coniferous forests of the Paciﬁc Northwest, USA. Canadian Journal
of Forest Research 32:320-331.

207

Cleland, D. T., J. B. Hart, G. E. Host, K. S. Pregitzer, and C. W. Ramm. 1993. Field
Guide: Ecological Classiﬁcation and Inventory System of the Huron-Manistee
National Forests. Huron-Manistee National Forest, Cadillac, MI.

Cohen, D. 1970. Expected efﬁciency of water utilization in plants under different
competition and selection regimes Israel Journal of Botany 19:50-54.

Comas, L. H., T. J. Bouma, and D. M. Eissenstat. 2002. Linking root traits to potential
growth rate in six temperate tree species. Oecologia 132:34-43.

Coomes, D. A., and P. J. Grubb. 2000. Impacts of root competition in forests and
woodlands: A theoretical framework and review of experiments. Ecological
Monographs 70:171-207.

Cote, B., J. W. Fyles, and H. Djalilvand. 2002. Increasing N and P resorption efﬁciency
and proﬁciency in northern deciduous hardwoods with decreasing foliar N and P
concentrations. Annals of Forest Science 59:275-281.

Cowan, I. R. 1986. Econmics of carbon ﬁxation in higher plants. Pages 133-170 in T. J.
Givnish, editor. On the Economy of Plant Form and Function. Cambridge
University Press, Cambridge, UK.

Cowan, I. R., and G. D. F arquhar. 1977. Stomata function in relation to leaf metabolism
and environment. Pages 471-505 in D. H. Jennings, editor. Symposia of the
Society for Experimental Biology. Cambridge University Press, Cambridge, UK.

de Gouvenain, R. C., R. K. Kobe, and J. A. Silander. 2007. Partitioning of understorey
light and dry-season soil moisture gradients among seedlings of four rain-forest
tree species in Madagascar. Journal of Tropical Ecology 23:569-579.

Del Tredici, P. 2001. Sprouting in temperate trees: A morphological and ecological
review. Botanical Review 67: 121-140.

Delucia, E. H., and S. A. Heckathom. 1989. The effect of soil drought on water-use

efﬁciency in a contrasting Great Basin desert and Sierran montane species Plant
Cell and Environment 12:935-940.

Delucia, E. H., W. H. Schlesinger, and W. D. Billings. 1988. Water relations and the
maintenance of Sierran conifers on hydrothermally altered rock. Ecology 69:303-
311.

Dilustro, J. J ., F. P. Day, and B. G. Drake. 2001. Effects of elevated atmospheric C02 on
root decomposition in a scrub oak ecosystem. Global Change Biology 7:581-589.

Dina, S. J ., and L. G. Klikoff. 1973. Effect of plant moisture stress on carbohydrate and
nitrogen content of big sagebrush Journal of Range Management 26:207—209.

208

Don, A., J. Schumache, M. Scherer-Lorenzen, T. Scholtern, and E. D. Schulze. 2007.
Spatial and vertical variation of soil carbon at two grassland sites-implication for
measuring soil carbon stocks. Geoderma 141:272-282.

Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F. Smith. 1956. Colorimetric
method for determination of sugars and related substances. Analytical Chemistry
28:350-356.

Duchesne, L., R. Ouimet, C. Camire, and D. Houle. 2001. Seasonal nutrient transfers by
foliar resorption, leaching, and litter fall in a northern hardwood forest at Lake
Clair Watershed, Quebec, Canada. Canadian Journal of Forest Research 31:333-
344.

Dudley, S. A. 1996. The response to differing selection on plant physiological traits:
Evidence for local adaptation. Evolution 50:103-1 10.

Ehleringer, J. R., and T. A. Cooper. 1988. Correlations between carbon isotope ratio and
microhabitat in desert plants. Oecologia 76:562-566.

Engelbrecht, B. M. J ., L. S. Comita, R. Condit, T. A. Kursar, M. T. Tyree, B. L. Turner,
and S. P. Hubbell. 2007. Drought sensitivity shapes species distribution patterns
in tropical forests. Nature 447:80-U82.

Engelbrecht, B. M. J ., and T. A. Kursar. 2003. Comparative drought-resistance of

seedlings of 28 species of co-occurring tropical woody plants. Oecologia
136:383-393.

Engelbrecht, B. M. J ., T. A. Kursar, and M. T. Tyree. 2005. Drought effects on seedling
survival in a tropical moist forest. Trees: Structure and Function 19:312-321.

Evans, G. C. 1972. The Quantitative Analysis of Plant Growth. University of California
Press, Berkeley, CA.

Fahey, T. J ., J. J. Battles, and G. F. Wilson. 1998. Responses of early successional
northern hardwood forests to changes in nutrient availability. Ecological
Monographs 68:183-212.

Field, C., J. Merino, and H. A. Mooney. 1983. Compromises between water-use
efﬁciency and nitrogen-use efficiency in 5 species of califomia evergreens.
Oecologia 60:384-3 89.

Field, C., and H. A. Mooney. 1986. The photosynthesis-nitrogen relationship in wild
plants. Pages 25-55 in T. J. Givnish, editor. On the Economy of Plant Form and
Function. Cambridge University Press, Cambridge.

Finzi, A. C., and C. D. Canham. 2000. Sapling growth in response to light and nitrogen

availability in a southern New England forest. Forest Ecology and Management
131:153-165.

209

Gansert, D., and W. Sprick. 1998. Storage and mobilization of nonstructural
carbohydrates and biomass development of beech seedlings (F agus sylvatica L.)
under different light regimes. Trees: Structure and Function 12:247-257.

Gaucher, C., S. Gougeon, Y. Mauffette, and C. Messier. 2005. Seasonal variation in
biomass and carbohydrate partitioning of understory sugar maple (A cer

saccharum) and yellow birch (Betula alleghaniensis) seedlings. Tree Physiology
25:93-100.

Gholz, H. L. 1982. Environmental limits on aboveground net primary production, leaf
area, and biomass in vegetation zones of the Paciﬁc Northwest. Ecology 63:469-
481.

Givnish, T. J. 1986. Biochemical constraints on crown geometry in forest herbs. Pages
525-583 in T. J. Givnish, editor. On the Economy of Plant Form and Function.
Cambridge University Press, New York, NY.

Givnish, T. J. 1988. Adaptation to sun and shade: A whole-plant perspective. Australian
Journal of Plant Physiology 15:63-92.

Gordon, W. S., and R. B. Jackson. 2000. Nutrient concentrations in ﬁne roots. Ecology
81:275-280.

Green, P. T., and P. A. Juniper. 2004. Seed mass, seedling herbivory and the reserve
effect in tropical rainforest seedlings. Functional Ecology 18:539-547.

Grime, J. P. 1977. Evidence for existence of 3 primary strategies in plants and its

relevance to ecological and evolutionary theory. American Naturalist 111:1169-
1194.

Guerrero-Campo, J ., and A. H. Fitter. 2001. Relationships between root characteristics
and seed size in two contrasting ﬂoras. Acta Oecologica 22:77-85.

Gurevitch, J ., J. A. Teeri, and A. M. Wood. 1986. Differentiation among populations of
Sedum wrightii (Crassulaceae) in response to limited water availability: Water
relations, CO; assimilation, growth and survivorship. Oecologia 70:198-204.

Haeussler, S., J. C. Tappeiner, and B. J. Greber. 1995. Germination, survival, and early
growth of red alder seedlings in the Central Coast Range of Oregon. Canadian
Journal of Forest Research 25:1639-1651.

Host, G. E., K. S. Pregitzer, C. W. Ramm, D. P. Lusch, and D. T. Cleland. 1988.
Variation in overstory biomass among glacial landforms and ecological land units

in northwestern lower Michigan. Canadian Journal of Forest Research 18:659-
668.

210

Huang, X. M., A. N. Lakso, and D. M. Eissenstat. 2005. Interactive effects of soil
temperature and moisture on Concord grape root respiration. Journal of
Experimental Botany 56:2651-2660.

IPCC. 2001. Third Assessment Report of the Intergovernmental Panel on Global Climate
Change. Cambirdge University Press, New York, NY.

lyer, M. 2006. Root carbohydrate storage in temperate and tropical forest tree seedlings:
implications for species coexistence. Dissertation. Michigan State University,
East Lansing.

Jackson, R. B., J. Canadell, J. R. Ehleringer, H. A. Mooney, O. E. Sala, and E. D.
Schulze. 1996. A global analysis of root distributions for terrestrial biomes.
Oecologia 108:389-411.

Jackson, R. B., L. A. Moore, W. A. Hoffmann, W. T. Pockman, and C. R. Linder. 1999.
Ecosystem rooting depth determined with caves and DNA. Proceedings of the
National Academy of Sciences of the United States of America 96:11387-11392.

John, B., H. N. Pandey, and R. S. Tripathi. 2002. Decomposition of ﬁne roots of Pinus
kesiya and turnover of organic matter, N and P of coarse and ﬁne pine roots and

herbaceous roots and rhizomes in subtropical pine forest stands of different ages.
Biology and Fertility of Soils 35:238-246.

Jurik, T. W., J. A. Weber, and D. M. Gates. 1988. Effects of temperature and light on
photosynthesis of dominant tree species of a northern hardwood forest. Botanical
Gazette 149:203-208.

Kays, J. S., and C. D. Canham. 1991. Effects of time and frequency of cutting on
hardwood root reserves and sprout growth. Forest Science 37:524-539.

Killingbeck, K. T. 1996. Nutrients in senesced leaves: Keys to the search for potential
resorption and resorption proﬁciency. Ecology 77: 1716-1727.

Killingbeck, K. T., J. D. May, and S. Nyman. 1990. Foliar senescence in an aspen
(Populus tremuloides) clone: The response of element resorption to interramet

variation and timing of abscission. Canadian Journal of Forest Research 20:1156-
1164.

Kloeppel, B. D., M. D. Abrams, and M. E. Kubiske. 1993. Seasonal ecophysiology and
leaf morphology of four successional Pennsylvania barrens species in open versus
understory environments. Canadian Journal of Forest Research 23: 1 81-1 89.

Kobe, R. K. 1996. Intraspeciﬁc variation in sapling mortality and grth predicts
geographic variation in forest composition. Ecological Monographs 66:181-201.

Kobe, R. K. 1997. Carbohydrate allocation to storage as a basis of interspeciﬁc variation
in sapling survivorship and growth. Oikos 80:226-233.

211

Kobe, R. K. 2006. Sapling growth as a function of light and landscape-level variation in
soil water and foliar nitrogen in northern Michigan. Oecologia 147:119-133.

Kobe, R. K., C. A. Lepczyk, and M. Iyer. 2005. Resorption efficiency decreases with
increasing green leaf nutrients in a global data set. Ecology 86:2780-2792.

Kobe, R. K., S. W. Pacala, J. A. Silander, and C. D. Canham. 1995. Juvenile tree
survivorship as a component of shade tolerance. Ecological Applications 5:517-
532.

Kohyama, T., and P. J. Grubb. 1994. Below- and above-ground allometries of shade-
tolerant seedlings in a Japanese warm-temperate rain forest. Functional Ecology
8:229-236.

Kozlowski, T. T. 1992. Carbohydrate sources and sinks in woody plants Botanical
Review 58:107-222.

Kruger, E. L., and P. B. Reich. 1997a. Responses of hardwood regeneration to ﬁre in

mesic forest openings. 1. Post-ﬁre community dynamics. Canadian Journal of
Forest Research 27:1822-1831.

Kruger, E. L., and P. B. Reich. 1997b. Responses of hardwood regeneration to ﬁre in
mesic forest openings. III. Whole-plant growth, biomass distribution, and nitrogen
and carbohydrate relations. Canadian Journal of Forest Research 27:1841-1850.

Kruger, E. L., and J. C. Volin. 2006. Reexamining the empirical relationship between
plant growth and leaf photosynthesis. Functional Plant Biology 33:421-429.

Kubiske, M. E., and K. S. Pregitzer. 1996. Effects of elevated C02 and light availability
on the photosnyhtetic response of trees of contrasting shade tolerance. Tree
Physiology 16:351-358.

Landsberg, J. J. 1986. Physiological ecology of forest production. Academic Press,
London, UK.

Leishman, M. R., and M. Westoby. 1994. The role of large seed size in shaded
conditions: Experimental evidence. Functional Ecology 8:205-214.

Loach, K. 1967. Shade tolerance in tree seedlings I. Leaf photosynthesis and respiration
in plants raised under artiﬁcial shade New Phytologist 66:607-621.

Loescher, W. H., T. McCamant, and J. D. Keller. 1990. Carbohydrate reserves,
translocation, and storage in woody plant roots Hortscience 25:274-281.

Lucash, M. S., J. D. Joslin, and R. D. Yanai. 2005. Temporal variation in nutrient uptake
capacity by intact roots of mature loblolly pine. Plant and Soil 272:253-262.

212

Ludovici, K. H., and L. W. Kress. 2006. Decomposition and nutrient release from fresh
and dried pine roots under two fertilizer regimes. Canadian Journal of Forest
Research 36:105-111.

Lusk, C. H. 2004. Leaf area and growth of juvenile temperate evergreens in low light:
species of contrasting shade tolerance change rank during ontogeny. Functional
Ecology 18:820-828.

Lusk, C. H., O. Contreras, and J. Figueroa. 1997. Growth, biomass allocation and plant
nitrogen concentration in Chilean temperate rainforest tree seedlings: Effects of
nutrient availability. Oecologia 109:49-58.

Lusk, C. H., and A. Del Pozo. 2002. Survival and growth of seedlings of 12 Chilean
rainforest trees in two light environments: Gas exchange and biomass distribution
correlates. Austral Ecology 27:173-182.

Lusk, C. H., and F. I. Piper. 2007. Seedling size inﬂuences relationships of shade
tolerance with carbohydrate-storage patterns in a temperate rainforest. Functional
Ecology 21:78-86.

MacFarlane, D. W., and R. K. Kobe. 2006. Selecting models for capturing tree-size

effects on growth-resource relationships. Canadian Journal of Forest Research
36:1695-1704.

Machado, J. L., and P. B. Reich. 2006. Dark respiration rate increases with plant size in
saplings of three temperate tree species despite decreasing tissue nitrogen and
nonstructural carbohydrates. Tree Physiology 26:915-923.

Marks, P. L. 1975. Relation between extension growth and successional status of
deciduous trees of northeastern United States. Bulletin of the Torrey Botanical
Club 102:172-177.

Marquis, R. J., E. A. Newell, and A. C. Villegas. 1997. Non-structural carbohydrate
accumulation and use in an understorey rain-forest shrub and relevance for the
impact of leaf herbivory. Functional Ecology 11:636-643.

McClaugherty, C. A., J. D. Aber, and J. M. Melillo. 1982. The role of ﬁne roots in the
organic matter and nitrogen budgets of 2 forested ecosystems. Ecology 63: 1481-
1490.

McFarlane, K. J ., and R. D. Yanai. 2006. Measuring nitrogen and phosphorus uptake by
intact roots of mature Acer saccharum Marsh, Pinus resinosa Ait., and Picea
abies (L.) Karst. Plant and Soil 279:163-172.

McKay, H. M., and D. C. Maclcolm. 1988. A comparison of the ﬁne root component of a

pure and a mixed coniferous stand. Canadian Journal of Forest Research 18: 141 6-
1426.

213

McLaughlin, S. B., and R. Wimmer. 1999. Tansley Review No. 104 - Calcium
physiology and terrestrial ecosystem processes. New Phytologist 1422373-417.

Meier, C. E., C. C. Grier, and D. W. Cole. 1985. Below- and aboveground N and P use
by Abies amabilis stands. Ecology 66:1928-1942.

Moles, A. T., and M. Westoby. 2004. Seedling survival and seed size: a synthesis of the
literature. Journal of Ecology 92:372-383.

Myers, J. A., and K. Kitaj irna. 2007. Carbohydrate storage enhances seedling shade and
stress tolerance in a neotropical forest Journal of Ecology 95:3 83—395.

Nambiar, E. K. S. 1987. Do nutrients retranslocate from ﬁne roots? Canadian Journal of
Forest Research 17:913-918.

Nepstad, D. C., C. R. Decarvalho, E. A. Davidson, P. H. Jipp, P. A. Lefebvre, G. H.
Negreiros, E. D. Dasilva, T. A. Stone, S. E. Trumbore, and S. Vieira. 1994. The

role of deep roots in the hydrological and carbon cycles of Amazonian forests and
pastures. Nature 372:666-669.

Newell, E. A., S. S. Mulkey, and S. J. Wright. 2002. Seasonal patterns of carbohydrate
storage in four tropical tree species. Oecologia 131:333-342.

Newman, G. S., and S. C. Hart. 2006. Nutrient covariance between forest foliage and ﬁne
roots. Forest Ecology and Management 236:136-141.

Ni, B. R., and S. G. Pallardy. 1991. Response of gas-exchange to water stress in seedlings
of woody angiosperms Tree Physiology 8:1-9.

Norby, R. J ., and C. M. Iversen. 2006. Nitrogen uptake, distribution, turnover, and
efﬁciency of use in a COz-enriched sweetgum forest. Ecology 87 :5-14.

Oosthuizen, I. B., and H. A. Snyman. 2001. The inﬂuence of water stress on non-
structural carbohydrate concentration in T hemeda triandra. South African Journal
of Botany 67:53-57.

Ostertag, R., and S. E. Hobbie. 1999. Early stages of root and leaf decomposition in
Hawaiian forests: effects of nutrient availability. Oecologia 121:564-573.

Pacala, S. W., C. D. Canham, J. Saponara, J. A. Silander, R. K. Kobe, and E. Ribbens.
1996. Forest models deﬁned by ﬁeld measurements: Estimation, error analysis
and dynamics. Ecological Monographs 66:1-43.

Pacala, S. W., C. D. Canham, J. A. Silander, and R. K. Kobe. 1994. Sapling growth as a
function of resources in a north temperate forest. Canadian Journal of Forest
Research 24:2172-2183.

214

Padilla, F. M., and F. I. Pugnaire. 2007. Rooting depth and soil moisture control

Mediterranean woody seedling survival during drought. Functional Ecology
21:489-495.

Pennington, M. R., and M. B. Walters. 2006. The response of planted trees to vegetation

zonation and soil, redox potential in created wetlands. Forest Ecology and
Management 233:1-10.

Poorter, H., and O. Nagel. 2000. The role of biomass allocation in the growth response of

plants to different levels of light, C02, nutrients and water: a quantitative review.
Australian Journal of Plant Physiology 27:595-607.

Poorter, H., and C. Remkes. 1990. Leaf area ratio and net assimilation rate of 24 wild
species differing in relative growth rate. Oecologia 83:553-559.

Poorter, H., C. Remkes, and H. Lambers. 1990. Carbon and nitrogen economy of 24 wild
species differing in relative growth rate. Plant Physiology 94:621-627.

Poorter, H., and R. Villar. 1997. The fate of acquired carbon in plants: chemical
composition and construction costs. Pages 37-72 in F. A. Bazzaz and J. Grace,
editors. Plant Resource Allocation. Academic Press, London, UK.

Poorter, L. 1999. Growth responses of 15 rain-forest tree species to a light gradient: the
relative importance of morphological and physiological traits. Functional Ecology
13:396-410.

Poorter, L. 2001. Light-dependent changes in biomass allocation and their importance for
growth of rain forest tree species. Functional Ecology 15:113-123.

Poorter, L. 2007. Are species adapted to their regeneration niche, adult niche, or both?
American Naturalist 169:433-442.

Poorter, L., and F. Bongers. 2006. Leaf traits are good predictors of plant performance
across 53 rain forest species. Ecology 87:1733-1743.

Poorter, L., L. Bongers, and F. Bongers. 2006. Architecture of 54 moist-forest tree
species: Traits, trade-offs, and functional groups. Ecology 87: 1289-1301.

Poorter, L., and K. Kitajima. 2007. Carbohydrate storage and light requirements of
tropical moist and dry forest tree species. Ecology 88:1000-1011.

Poorter, L., and S. Rose. 2005. Light-dependent changes in the relationship between seed

mass and seedling traits: a meta-analysis for rain forest tree species. Oecologia
142:378-387.

Portsmuth, A., and U. Niinemets. 2007. Structural and physiological plasticity in

response to light and nutrients in ﬁve temperate deciduous woody species of
contrasting shade tolerance. Functional Ecology 21:61-77.

215

Prairie, Y. T., and D. F. Bird. 1989. Some misconceptions about the spurious correlation:
Problem in the ecological literature. Oecologia 81:285-288.

Pregitzer, K. S. 2002. Fine roots of trees: A new perspective. New Phytologist 154:267-
270.

Pregitzer, K. S., J. A. King, A. J. Burton, and S. E. Brown. 2000. Responses of tree ﬁne
roots to temperature. New Phytologist 147:105-115.

Pregitzer, K. S., M. E. Kubiske, C. K. Yu, and R. L. Hendrick. 1997. Relationships
among roof branch order, carbon, and nitrogen in four temperate species.
Oecologia 111:302-308.

Quero, J. L., R. Villar, T. Maranon, R. Zamora, and L. Poorter. 2007. Seed-mass effects
in four Mediterranean Quercus species (Fagaceae) growing in contrasting light
environments. American Journal of Botany 94:1795-1803.

Raison, R. J ., M. J. Connell, and P. K. Khanna. 1987. Methodology for studying ﬂuxes of
soil mineral-N in situ Soil Biology & Biochemistry 19:521-530.

Reich, P. B., C. Buschena, M. G. Tjoelker, K. Wrage, J. Knops, D. Tilman, and J. L.
Machado. 2003a. Variation in growth rate and ecophysiology among 34 grassland

and savanna species under contrasting N supply: a test of functional group
differences. New Phytologist 157:617-631.

Reich, P. B., D. S. Ellsworth, M. B. Walters, J. M. Vose, C. Gresham, J. C. Volin, and W.
D. Bowman. 1999. Generality of leaf trait relationships: A test across six biomes.
Ecology 80:1955-1969.

Reich, P. B., D. F. Grigal, J. D. Aber, and S. T. Gower. 1997a. Nitrogen mineralization
and productivity in 50 hardwood and conifer stands on diverse soils. Ecology
78:335-347.

Reich, P. B., and T. M. Hinckley. 1989. Inﬂuence of pre-dawn water potential and soil-
to-leaf hydraulic conductance on maximum daily diffusive conductance in 2 oak
species. Functional Ecology 3:719-726.

Reich, P. B., M. G. Tjoelker, M. B. Walters, D. W. Vanderklein, and C. Bushena. 1998a.
Close association of RGR, leaf and root morphology, seed mass and shade

tolerance in seedlings of nine boreal tree species grown in high and low light.
Functional Ecology 12:327-338.

Reich, P. B., M. B. Walters, and D. S. Ellsworth. 1997b. From tropics to tundra: Global
convergence in plant functioning. Proceedings of the National Academy of
Sciences of the United States of America 94:13730-13734.

Reich, P. B., M. B. Walters, M. G. Tjoelker, D. Vanderklein, and C. Buschena. 1998b.
Photosynthesis and respiration rates depend on leaf and root morphology and

216

nitrogen concentration in nine boreal tree species differing in relative growth rate.
Functional Ecology 12:395-405.

Reich, P. B., I. J. Wright, J. Cavender-Bares, J. M. Craine, J. Oleksyn, M. Westoby, and
M. B. Walters. 2003b. The evolution of plant functional variation: Traits, spectra,
and strategies. International Journal of Plant Sciences 164:Sl43-Sl64.

Roper, T. R., J. D. Keller, W. H. Loescher, and C. R. Rom. 1988. Photosynthesis and
carbohydrate partitioning in sweet cherry: Fruiting effects. Physiologia Plantarum
72:42-47.

Royce, E. B., and M. G. Barbour. 2001. Mediterranean climate effects. I. Conifer water
use across a Sierra Nevada ecotone. American Journal of Botany 88:911-918.

Sack, L. 2004. Responses of temperate woody seedlings to shade and drought: do trade-
offs limit potential niche differentiation? Oikos 107:110-127.

Saverimuttu, T., and M. Westoby. 1996. Seedling longevity under deep shade in relation
to seed size. Journal of Ecology 84:681-689.

Schreeg, L. A., R. K. Kobe, and M. B. Walters. 2005. Tree seedling growth, survival, and
morphology in response to landscape-level variation in soil resource availability
in northern Michigan. Canadian Journal of Forest Research 35:263-273.

Schulze, E. D., H. A. Mooney, O. E. Sala, E. Jobbagy, N. Buchmann, G. Bauer, J.
Canadell, R. B. Jackson, J. Loreti, M. Oesterheld, and J. R. Ehleringer. 1996.
Rooting depth, water availability, and vegetation cover along an aridity gradient
in Patagonia. Oecologia 108:503-511.

Schulze, E. D., R. J. Williams, G. D. Farquhar, W. Schulze, J. Langridge, J. M. Miller,
and B. H. Walker. 1998. Carbon and nitrogen isotope discrimination and nitrogen

nutrition of trees along a rainfall gradient in northern Australia. Australian Journal
of Plant Physiology 25:413-425.

Shibuya, M., T. Yajima, M. Yoshida, and E. Sasaoka. 2007. Different biomass allocation
patterns among four tree species in heavily distrubed sites on a volacanic
mountain in Hokkaido, northern Japan. Ecological Research 22: 177-182.

Silla, F., and A. Escudero. 2004. Nitrogen-use efﬁciency: trade-offs between N

productivity and mean residence time at organ, plant and population levels.
Functional Ecology 18:51 1-521 .

Silla, F ., and A. Escudero. 2006. Coupling N cycling and N productivity in relation to
seasonal stress in Quercus pyrenaica Willd. saplings. Plant and Soil 282:301-311.

Sipe, T. W., and F. A. Bazzaz. 1995. Gap partitioning among maples (Acer) in central
New England: Survival and growth. Ecology 76:1587-1602.

217

Spaeth, S. C., and P. M. Cortes. 1995. Root cortex death and subsequent initiation and
growth of lateral roots from bare steles of chickpeas. Canadian Journal of Botany
73:253-261.

Steele, S. J., S. T. Gower, J. G. Vogel, and J. M. Norman. 1997. Root mass, net primary
production and turnover in aspen, jack pine and black spruce forests in
Saskatchewan and Manitoba, Canada. Tree Physiology 17:577-587.

Swaine, M. D. 1996. Rainfall and soil fertility as factors limiting forest species
distributions in Ghana. Journal of Ecology 84:419-428.

Tanner, E. V. J ., and I. M. Barberis. 2007. Trenching increased growth, and irrigation
increased survival of tree seedlings in the understorey of a semi-evergreen rain
forest in Panama. Journal of Tropical Ecology 23:257-268.

Teng, Y. W., K. Tanabe, F. Tamura, and A. Itai. 1999. Translocation of C-13-assimilates
in the spring following fall assimilation of (C02)-C-1 3 by 'Nijisseiki' pear (Pyrus
pyrifolia Nakai). Journal of the Japanese Society for Horticultural Science
68:248-255.

Teskey, R. O., and R. B. Shrestha. 1985. A relationship between carbon dioxide,
photosynthetic efﬁciency and shade-tolerance. Physiologia Plantarum 63:126-
132.

Tilman, D. 1988. Plant strategies and the dynamics of plant communities. Princeton
University Press, Princeton, NJ.

Tjoelker, M. G., J. M. Craine, D. Wedin, P. B. Reich, and D. Tilman. 2005. Linking leaf
and root trait syndromes among 39 grassland and savannah species. New
Phytologist 167:493-508.

Valverde-Barrantes, O. J ., J. W. Raich, and A. E. Russell. 2007. Fine-root mass, growth
and nitrogen content for six tropical tree species. Plant and Soil 290:357-3 70.

van Heerwaarden, L. M., S. Toet, and R. Aerts. 2003. Current measures of nutrient
resorption efﬁciency lead to a substantial underestimation of real resorption
efﬁciency: facts and solutions. Oikos 101:664-669.

VandenBoogaard, R., and R. Villar. 1998. Variation in grth and water-use-efﬁcieny: a
comparsion of Aegilops L. species and Triticum aestivum L. cultivars. Pages 289-
308 in H. Lambers, H. Poorter, and M. M. 1. Van Vuuren, editors. Inherent
Variation in Plant Growth: Physiological Mecahnisms and Ecological
Consequences. Backhuys Publishers, Leiden, Netherlands.

VanPraag, H. J ., S. Sougnezremy, and F. Weissen. 1988. Root turnover in a beech and a
spruce stand of the Belgian Ardennes. Plant and Soil 105:87-103.

218

Veneklaas, E. J ., and F. den Ouden. 2005. Dynamics of non-structural carbohydrates in

two Ficus species after transfer to deep shade. Environmental and Experimental
Botany 54:148-154.

Volaire, F. 1995. Growth, carbohydrate reserves and drought survival strategies of

contrasting Dactylis glomerata populations in a Mediterranean environment.
Journal of Applied Ecology 32:56-66.

Walters, M. B., E. L. Kruger, and P. B. Reich. 1993a. Growth, biomass distribution and
C02 exchange of northern hardwood seedlings in high and low light:
Relationships with successional status and shade tolerance. Oecologia 9427-16.

Walters, M. B., E. L. Kruger, and P. B. Reich. 1993b. Relative growth rate in relation to
physiological and morphological traits for northern hardwood tree seedlings:
Species, light environment and ontogenic considerations. Oecologia 96:219-231.

Walters, M. B., and P. B. Reich. 1989. Response of Ulmus americana seedlings to

varying nitrogen and water status. 1 Photosynthesis and growth. Tree Physiology
5:159-172.

Walters, M. B., and P. B. Reich. 1997. Growth of Acer saccharum seedlings in deeply
shaded understories of northern Wisconsin: Effects of nitrogen and water

availability. Canadian Journal of Forest Research-Revue Canadienne De
Recherche F orestiere 27:237-247.

Walters, M. B., and P. B. Reich. 1999. Low-light carbon balance and shade tolerance in
the seedlings of woody plants: do winter deciduous and broad-leaved evergreen
species differ? New Phytologist 143:143-154.

Walters, M. B., and P. B. Reich. 2000. Seed size, nitrogen supply, and growth rate affect
tree seedling survival in deep shade. Ecology 81:1887-1901.

Wang, X. R, Z. Y. Tang, and J. Y. Fang. 2006. Climatic control on forests and tree
species distribution in the forest region of northeast China. Journal of Integrative
Plant Biology 48:778-789.

Warton, D. I., I. J. Wright, D. S. F alster, and M. Westoby. 2006. Bivariate line-ﬁtting
methods for allometry. Biological Reviews 81:259-291.

Weber, J. A., T. W. Jurik, J. D. Tenhunen, and D. M. Gates. 1985. Analysis of gas-
exchange in seedlings of Acer saccharum: Integration of ﬁeld and laboratory
studies. Oecologia 65:338-347.

Will, R. E., D. Markewitz, R. L. Hendrick, D. F. Meason, T. R. Crocker, and B. E.
Borders. 2006. Nitrogen and phosphorus dynamics for 13-year-old loblolly pine
stands receiving complete competition control and annual N fertilizer. Forest
Ecology and Management 227: 1 55-168.

219

Wright, I. J ., and M. Westoby. 1999. Differences in seedling growth behaviour among
species: trait correlations across species, and trait shifts along nutrient compared
to rainfall gradients. Journal of Ecology 87:85-97.

Yamada, T., E. Suzuki, T. Yarnakura, and 8. Tan. 2005. Tap-root depth of tropical
seedlings in relation to species-speciﬁc edaphic preferences. Journal of Tropical
Ecology 21:155-160.

Young, J. A., and C. G. Young. 1992. Seeds of woody plants in North America.
Dioscorides Press, Portland, Oregon.

Zak, D. R., G. E. Host, and K. S. Pregitzer. 1989. Regional variability in nitrogen
mineralization, nitriﬁcation, and overstory biomass in northern lower Michigan.
Canadian Journal of Forest Research 19:1521-1526.

220

   

uagilliiglilgili1131111171