PERENNIAL CEREAL CROPS FOR THE COLD TEMPERATE ZONE: AGRONOMY,
PHYSIOLOGY, SINK REGULATION AND DISEASE RESISTANCE
By
Nikhil Sankara Jaikumar

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Crop and Soil Sciences --- Doctor of Philosophy
Ecology, Evolution and Behavioral Biology --- Dual Major
2013

ABSTRACT

PERENNIAL CEREAL CROPS FOR THE COLD TEMPERATE ZONE: AGRONOMY,
PHYSIOLOGY, SINK REGULATION AND DISEASE RESISTANCE
By
Nikhil Sankara Jaikumar

Within the last few decade, renewed attention has been devoted to the development of
high yielding, herbaceous perennial plants that can meet the demand for food, forage and fiber in
environmentally sustainable ways. While woody plants and annual herbaceous plants have been
successfully selected to achieve high levels of reproductive allocation, herbaceous perennial
plants with very high levels of sexual reproduction do not presently exist. The lack of high
yielding herbaceous perennials is of intense interest to agronomists, horticulturalists,
evolutionary biologists, population ecologists, ecophysiologists, and breeders.
However, perennial cereal crops face challenges if they are to be economically feasible,
including genetic incompatibilities, resource tradeoffs between sexual reproduction and multiyear survival, and potential for disease buildup. In a series of studies between 2008 and 2012, I
explore these issues within the context of three perennial cereal species: perennial wheat
(Triticum aestivum x Thinopyrum spp.), perennial rye (Secale cereale x montanum) and
intermediate wheatgrass (Thinopyrum intermedium).
Chapter 1 of this dissertation is a literature review dealing with the history of perennial
cereal breeding, potential environmental benefits, and challenges that perennial cereals face.
Chapter 2 reports on a two-year agronomic study of four accessions of perennial wheat and one
accession of perennial rye. Chapter 3 describes a series of experiments exploring differences in

photosynthetic rates, and other traits affecting resource accumulation, between perennial and
annual cereals. Chapter 4 explores the extent to which photosynthesis and metabolism in
perennial cereals is responsive to changes in carbohydrate supply / demand ratio and whether this
responsiveness differs between perennial and annual cereals. Finally, Chapter 5 explores the
resistance of three perennial cereal species to the fungal disease Fusarium head blight (FHB).
In brief, we find the following.
Perennial wheat achieves grain yields of approximately 50% of annual wheat, while
perennial rye achieves 75% of annual rye. Both species show an ability to maintain these
yields into their second year, and show later flowering than annuals.
Perennial wheat, perennial rye and intermediate wheatgrass show 10-60% higher
photosynthetic rates than their annual analogues. Intermediate wheatgrass shows
declining photosynthetic rates with increasing plant age. These differences are driven
primarily by biochemical rather than hydraulic changes.
Perennial wheat and rye are more sink-limited than their annual analogues, and show an
ability to mostly compensate for moderate source/sink changes. The perennial species
appear to show a more conservative reproductive strategy than their annual relatives.
Intermediate wheatgrass shows high resistance to FHB, while perennial rye is moderately
susceptible and perennial wheat accessions vary in susceptibility.
Thus our studies provide novel contributions to the growing literature on perennial
cereals, illustrating some physiological traits of perennial cereals as well as some of the
problems they face. We hope that our results can contribute to advancing efforts to achieve
high yielding herbaceous perennials, as well as to improving our understanding of how life
history, source-sink balance and whole-plant age interact to affect resource acquisition rates.

Copyright by
NIKHIL SANKARA JAIKUMAR
2013

ACKNOWLEDGEMENTS

Sincere acknowledgements and thanks are due to all of the people who have helped me with this
research over the last several years. In particular I would like to acknowledge the following:

My committee members (Sieg Snapp, James Flore, Jen Lau, Doug Landis and Janet Lewis) for
invaluable guidance; John Green, and Mark Freeman for invaluable technical assistance and
project management; Lee Siler and Randy Laurenz for planting help, disease monitoring
assistance and technical advice; Arianna Pikus, Lacey Culbertson, Dan Kane, Emily May and
Iman Sylvain for field assistance and help with data collection and maintenance of experiments;
Randy Laurenz and Sue Hammar for helping me with my greenhouse experiment; Wayne
Loescher for help in designing protocols for carbohydrate measurements and for training me in
carbohydrate analysis; Wayne Loescher and Bert Cregg for the loan of critical gas exchange
equipment; Alex Eilts and Tom Sharkey for helping train me in gas exchange protocols and help
me design my gas exchange experiments; and a variety of people including Jeannine CavenderBares, N. Michelle Holbrook, Lawren Sack, Christian Körner, Peter Reich, Deborah Roach,
Richard Shefferson, Lee DeHaan, David Van Tassel and others for critical advice and comments
on my research topics and methods as they developed. I would like to express my deepest
gratitude to all of you.

v

TABLE OF CONTENTS

LIST OF TABLES

x

LIST OF FIGURES

xiii

CHAPTER 1. POTENTIAL BENEFITS AND CHALLENGES FACING
PERENNIAL CEREAL CROPS
Overview
Intermediate Wheatgrass: A Case Study in Domestication
Hybridization of Grain Crops (Poaceae) with Perennial Wild Relatives
Perennial Wheat and Rye: Case Studies in Grass x Cereal Hybridization
Potential Environmental Benefits of Perennial Agroecosystems
Nutrient Usage and Leaching
Improved Carbon Sequestration
Improved Soil Quality
Resistance and Tolerance to Herbivory and Insects
Challenges to Perennial Cereal Crops
Genetic Incompatibility
Disease Pressure
Perenniality vs. Productivity Tradeoffs
Reproductive Photosynthesis
Annual vs. Perennial Differences in Resource Acquisition
Source vs. Sink Limitation of Metabolism
Conclusion
REFERENCES

1
2
5
7
9
15
15
19
22
23
25
26
30
33
36
38
41
55
57

CHAPTER 2. AGRONOMY OF PERENNIAL WHEAT AND RYE
Introduction
Materials and Methods
Site Characteristics
Accession Selection
Experimental Design
Management
Measurements
Data Analysis
Results
Environment
Yield of First-year Perennials and Annuals
Effect of Plant Age on Yield and Yield Components
Plant Growth, PSCR and Phenology
Discussion
Conclusion

71
72
75
75
75
76
77
78
79
80
80
81
83
84
86
92

vi

Acknowledgements
APPENDIX
REFERENCES

CHAPTER 3. RESOURCE ACQUISITION TRAITS IN PERENNIAL
AND ANNUAL CEREALS
Introduction
Materials and Methods
Site and Goals
Selection of Accessions
Design and Layout
Mid-season Photosynthetic Rates
Diurnal Variation in Photosynthesis
Conductance, Transpiration and Ci / Ca
Triose Phosphate Utilization, Carboxylation, Electron
Transport and Stomatal Limitation
LMR and LMA
Experiment 3: Characterizing the Intermediate
Wheatgrass Population
Experiments 4 and 5
Statistical Analysis
Results
Photosynthetic Rates (Experiment 1)
Photosynthetic Rates (Experiment 2)
Biochemical Traits
Hydraulic Traits
Diurnal Variation in Photosynthesis
Morphological Traits
Experiment 3: Sampling the Intermediate Wheatgrass Population
Experiment 4: Replicating Wheat vs. Wheatgrass Comparisons
Experiment 5: Post-Sexual Cycle Regrowth and Photosynthetic Rate
Discussion
Conclusion
Acknowledgements
APPENDICES
APPENDIX A. Figures and Tables for Chapter 3
APPENDIX B. Modeling the A / Ci Curve
REFERENCES

CHAPTER 4. SOURCE VS. SINK REGULATION IN PERENNIAL
AND ANNUAL CEREALS
Introduction
Materials & Methods
Site and Plant Materials
vii

93
94
106

111
112
116
116
118
122
123
123
123
123
126
126
127
127
129
129
129
130
131
132
132
132
133
133
134
140
140
142
143
156
163

169
170
173
173

Experimental Design
Source / Sink Manipulations
Survival
Photosynthetic Rates
Nonstructural Carbohydrates
Seed Mass and PSCR
Fall Flowering and Fall Photosynthesis
Pilot Experiment
Statistical Analysis
Results
Survival
Spring Photosynthetic Rates
Nonstructural Carbohydrates
Seed Mass and Regrowth
Fall Flowering and Photosynthesis
Pilot Experiment
Discussion
Conclusion
Acknowledgements
APPENDICES
APPENDIX A. Tables and Figures for Chapter 4
APPENDIX B. Pictures of Perennial Cereal Experiments
REFERENCES

CHAPTER 5. RESISTANCE TO FUSARIUM HEAD BLIGHT IN
PERENNIAL CEREALS
Introduction
Materials & Methods
Experimental Design
Experiment 1: Plant Materials
Experiment 1: Planting Protocol
Experiment 1: Inoculation and Observation of Response
Experiment 1: Post-sexual Cycle Regrowth
Experiment 2: FHB Symptoms on Naturally Infected
Plants in the Field
Statistical Analysis
Results
Experiment 1: Overall ANOVA
Experiment 1: Resistant and Susceptible Lines
Experiment 1: Regrowth
Experiment 2: Natural Occurrence of FHB in the Field
Discussion
Acknowledgments
APPENDIX
REFERENCES
viii

173
175
176
176
177
180
180
181
181
182
182
183
184
186
187
187
188
194
195
196
197
214
221

227
228
230
230
231
233
234
235
236
237
238
238
238
239
240
240
244
245
253

CHAPTER 6: CONCLUSION. BROADER IMPACTS OF
PERENNIAL CEREALS

ix

258

LIST OF TABLES

Table 1. Accessions and planting year / age in 2008-2010 agronomy study

100

Table 2. Dates of field operations in 2008-2010 agronomy study

101

Table 3. Means, Fisher’s LSD values, and F-values for yield and yield
components, averaged over 2009 and 2010

102

Table 4. Means, Fisher’s LSD values, and F-values for
yield and yield components in one year old (1 y) and
two year old (2 y) perennial cereal plants, in 2010

103

Table 5. Means, Fisher’s LSD values, and F-values for
plant growth and phenology data, sampled from one year old
annual and perennial cereals in 2009 and 2010

104

Table 6. Means, Fisher’s LSD values, and F-values for
plant growth and phenology data in one (1 y) and
two year old (2 y) plants of perennial cereal accessions in 2010

105

Table 7. Means ( + SE) and F-values (p-values) for photosynthetic rates
(A) measured at multiple time points, in spring 2012,
on two annual cereal species and on first-year (1 y) and second-year
(2 y) plants of three perennial cereal species

152

Table 8. Means (+ SE) and F-values (p-values) for stomatal conductance
(gs), transpiration (E), ratio of leaf intercellular to ambient
carbon dioxide (Ci / Ca), and degree of stomatal limitation (LS)
measured on two annual cereal species and on first-year (1 y)
and second-year (2 y) plants of three perennial
cereal species in spring 2012
Table 9. Means ( + SE) and F-values (p-values) for maximum
carboxylation rate (VC), electron transport rate (J), and
triose phosphate utilization (TPU), on two
annual cereal species and on first-year (1 y) and second-year
(2 y) plants of three perennial cereal species in spring 2012
Table 10. Means (+ SE) and F-values (p-values) for leaf mass ratio
(LMR), percent decrease in photosynthetic rate between morning
and afternoon (ΔA) and leaf mass per area (LMA),
measured on two annual cereal species and on first-year plants
x

153

154

of three perennial cereal species in spring 2011

155

Table 11. Mean values for photosynthetic rate, stomatal conductance,
transpiration, and ratio of leaf intercellular to ambient
carbon dioxide, measured in spring 2012 on a population of 28
independent intermediate wheatgrass (Thinopyrum intermedium)
genotypes, and on the selected genotype TLC-1 which was
the subject of further experimental study

155

Table 12. Experiment 1 (2010) means (SE), F-values and
P-values for photosynthetic rates (Amax) following boot stage
and pre-anthesis manipulations

198

Table 13. Experiment 1 (2010) means (SE) for photosynthetic
rates, measured 5-8 days following post-anthesis
source/sink manipulations, for annual wheat,
perennial wheat, annual rye and perennial rye

199

Table 14. Experiment 2 (2011) means (SE), F-values and
P-values for photosynthetic rates (Amax) measured 5-8 days
following boot stage and pre-anthesis source/sink
manipulation, in annual wheat, perennial wheat,
annual rye and perennial rye and post-sexual cycle
regrowth (PSCR) in perennial wheat and rye

200

Table 15. Experiment 2 (2011) means (SE), F-values and P-values
for photosynthetic rates (Amax) measured 5-8 days following
post-anthesis source-sink manipulation in annual wheat,
perennial wheat, annual rye and perennial rye, and
post-sexual cycle regrowth (PSCR) in perennial wheat and rye

201

Table 16. Experiment 2 (2011) means for unmanipulated
plants (percent of dry wt + SE), F-values and P-values
for various carbohydrate pools (glucose, starch, fructose,
soluble fructans) measured around anthesis, on leaf
tissue from annual wheat, perennial wheat,
annual rye and perennial rye

202

Table 17. Experiment 2 (2011) means for unmanipulated plants
(percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (insoluble fructans, sucrose, and
total NSC) measured around anthesis, on leaf tissue
from annual wheat, perennial wheat, annual rye
and perennial rye

203

xi

Table 18. Experiment 2 (2011) means for unmanipulated plants
(percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (glucose, starch, fructose, soluble
fructans) measured around anthesis, on stem tissue
from annual wheat, perennial wheat, annual rye
and perennial rye

204

Table 19. Experiment 2 (2011) means for unmanipulated plants
(percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (insoluble fructans, sucrose, and
total NSC) measured around anthesis, on stem tissue
from annual wheat, perennial wheat, annual rye
and perennial rye

205

Table 20. Pilot Experiment (2009) analysis of variance results
for photosynthetic rates (Amax) and triose phosphate
utilization rates (TPU) measured 7-10 days following treatment
imposition, as well as seed yield and single seed mass measured
the following year, for three perennial cereals

206

Table 21. Pilot Experiment (2009) means (+ SE) for photosynthetic
rates (Amax) and triose phosphate utilization rates (TPU)
measured 7-10 days following treatment imposition,
as well as seed yield and single seed mass measured
the following year, for three perennial cereals.

207

Table 22. Genotypes involved in a 2010-2011 greenhouse study of
Type II resistance to Fusarium Head Blight (East Lansing, MI)

247

Table 23. Results from comparing percent of Fusarium-infected
spikelets in various perennial grass and grass-cereal
amphiploid lines to annual susceptible and resistant controls

248

Table 24. Results from comparing percent of Fusarium-infected
rachis segments in various perennial grass and
grass-cereal amphiploid lines to annual susceptible
and resistant controls.

249

Table 25. Means (+ SE) and F-values (p-values in parentheses) from
comparing incidence, severity and total disease index of
Fusarium graminearum infection on seed heads of perennial rye
(Rival Rye), intermediate wheatgrass (TLI-C1) and five
wheat / Thinopyrum amphiploids, growing in the field under natural
and non-inoculated conditions.

250

xii

LIST OF FIGURES

Figure 1. Precipitation recorded in 2009 and 2010 growing seasons

96

Figure 2. Seed yield for Triticum spp. in a two year agronomy study

97

Figure 3. Seed yield for Secale spp. in a two year agronomy study

98

Figure 4. Kernel mass for one and two year old perennial wheat
and perennial rye

99

Figure 5. Statistical power (1 - ) as a function of underlying
effect size for photosynthetic parameters in 2012 study

144

Figure 6. Photosynthetic rates (Amax) for first year and second
year perennial wheat and perennial rye plants, compared
to annual relatives (wheat and rye) in 2010

145

Figure 7. Photosynthetic rates (Amax) measured at three
time points in spring 2012, on annual wheat,
one year old wheatgrass (IWG- 1), and two year old
wheatgrass (IWG-2)

146

Figure 8. Photosynthetic rate (A) as a function of intercellular
CO2 (Ci) in annual wheat, one year old wheatgrass
(Thinopyrum intermedium: IWG – 1y) and two year old
wheatgrass (IWG –2y)

147

Figure 9. Photosynthetic rate (A) as a function of intercellular
CO2 (Ci) in annual rye, one year old perennial rye
(Secale cereale x montanum: 1y) and two year old
perennial rye (2y)

148

Figure 10. Photosynthetic rate (A) as a function of intercellular
CO2 (Ci) in annual wheat, one year old perennial wheat
(Thinopyrum elongatum x Triticum aestivum: 1y) and two year old
perennial wheat (2y)

149

Figure 11. Photosynthetic rate (Amax) in annual
wheat x Thinopyrum amphiploids, perennial
wheat x Thinopyrum amphiploids,
annual wheat cv. ‘Caledonia’,
and second year wheatgrass (Thinopyrum intermedium)

150

xiii

Figure 12. Photosynthetic rate (Amax) in ten wheat x Thinopyrum
amphiploids, as a function of post-sexual cycle regrowth (PSCR)

151

Figure 13. Graphical depiction of theoretical A / Ci curve

162

Figure 14. Light-saturated photosynthetic rate (Amax) as a function
of source/sink ratio in annual wheat (AW), perennial wheat (PW),
annual rye (AW), and perennial rye (PR), measured in Experiment 1 (2010)

208

Figure 15. Light-saturated photosynthetic rate (Amax) as a
function of source/sink ratio in annual wheat (AW),
perennial wheat (PW), annual rye (AR) and perennial rye
(PR), measured in Experiment 2 (2011)

209

Figure 16. Leaf nonstructural carbohydrate content (NSC)
as a function of source/sink ratio in annual wheat (AW),
perennial wheat (PW), annual rye (AR) and
perennial rye (PR), in a 2011 field experiment

210

Figure 17. Stem nonstructural carbohydrate content (NSC)
as a function of source/sink ratio in annual wheat (AW),
perennial wheat (PW), annual rye (AR) and
perennial rye (PR), in a 2011 field experiment

211

Figure 18. Kernel mass as a function of source/sink ratio in annual
wheat (AW), perennial wheat (PW), annual rye (AR) and
perennial rye (PR), in a 2010 field experiment

212

Figure 19. Light-saturated photosynthetic rate (Amax) in flowering
and non-flowering microplots of perennial wheat, during the
fall post-sexual cycle regrowth period, in a 2011 field experiment

213

Figure 20. Perennial wheat plots at late elongation stage,
in a spring 2011 study of ecophysiological differences
in annual and perennial cereals (Chapter 3, Experiment 2)

215

Figure 21. Perennial wheat plots at early elongation stage,
in a spring 2011 study of ecophysiological differences in
annual and perennial cereals (Chapter 3, Experiment 2)

216

Figure 22. Intermediate wheatgrass plants at early elongation stage,
in a spring 2011 study of ecophysiological traits in intermediate
wheatgrass (Chapter 3, Experiment 3)

217

xiv

Figure 23. “Low” treatment (source / sink ratio = 67% of control) in perennial wheat plants,
imposed at boot stage, in a 2011 study of physiological responses to source/sink
manipulation in annual and perennial cereals (Chapter 4, Experiment 2)
218
Figure 24. “High” treatment (source / sink ratio = 200% of control)
in perennial rye plants, imposed at boot stage, in a 2011
study of physiological responses to source/sink
manipulation in annual and perennial cereals
(Chapter 4, Experiment 2)

219

Figure 25. Photosynthetic measurements in perennial
wheat plants, in a 2011 study of physiological responses
to source/sink manipulation in annual and perennial cereals
(Chapter 4, Experiment 2)

220

Figure 26. Percent of Fusarium-infected spikelets in 13 perennial
grass and grass-cereal amphiploid lines, compared against
susceptible and resistant annual wheat controls.

251

Figure 27. Percent of Fusarium-infected rachis segments in 13
perennial grass and grass-cereal amphiploid lines, compared
against susceptible and resistant annual wheat controls

252

xv

xvi

CHAPTER 1.
POTENTIAL BENEFITS AND CHALLENGES FACING PERENNIAL CEREAL CROPS

1

Overview
In recent years, considerable and increasing attention has been paid to the goal of trying
to introduce greater perenniality into human-managed ecosystems (Crews, 2005). In both
tropical and temperate environments, techniques such as agroforestry, silviculture, planting
shrubs and perennial herbs, and moving to perennial-dominated pastoral systems have for many
years been promoted to achieve better functioning of ecosystem services within agricultural,
forest, and semi-managed ecosystems (Ewel, 1999). These strategies can be seen as examples of
a general trend towards increasing the amount of perennial vegetation within agroecosystems, as
there are numerous environmental benefits that are hypothesized to correlate with increased
perenniality.
However, while agroforestry is a widely practiced technique with tremendous ecosystemservice benefits as well as some economic costs, as yet relatively little attention has been paid to
trying to introduce perennial plants as main staple food crops. Woody perennials have an
important role in food production systems across the world: they include sources of oil-rich nuts
(walnut, hazelnut, apple, cashew, citrus), fruit trees (apple, cherry, pear, tamarind, mango and
many others), fruiting shrubs (raspberry, blackberry), and in a few cases trees harvested for green
leaves (e.g. Moringa). However, while these sources of fruits, vegetable products and oilseeds
are critical nutritional components of human diets, there are currently very few perennial species
which occupy the role of main food staples. The few perennial cereals or legumes that are widely
grown include forage legumes like alfalfa (Medicago), pigeon pea (Cajanus), and a handful of
others, yet their contribution to food production systems is marginal and/or limited to specific

2

regions of the world (Cajanus, for example, is a weak perennial mainly limited to India, the
Caribbean and parts of Africa, and even there is often grown as an annual).
Trees and shrubs, because of their slow growth and long establishment period, face some
inherent limitations in terms of their utility as staple food crops. While the role of woody
perennials in agriculture can be expected to grow in future, and there are good nutritional as well
as environmental reasons for shifting production land from staple grains to perennial fruits and
vegetables, it is almost certain that for the near future herbaceous plants like cereals, oilseeds and
legumes will occupy a key role in global food production systems. In recent years the case has
been made that unless we begin growing perennial grasses, forbs and legumes as replacements
for our staple food and forage crops, ultimately our ability to create sustainable agroecosystems
that carry out more ecosystem services is limited.
Since about the mid-1990s, plant breeders have been trying to develop perennial crops that
could supplement or replace grain, oilseed, and legume crops that we use today (Cox et al.,
2006). In some cases they are making efforts to domesticate existing perennial grasses or
legumes, such as intermediate wheatgrass (Thinopyrum intermedium), gamagrass (Tripsacum
dactyloides), Illinois bundleflower (Desmanthus) and others. In other cases they are attempting
to combine perennial traits with existing crops through hybridization. Millet, rice, wheat, rye,
sorghum, and other grains (as well as sunflower) have wild perennial relatives with which they
are interfertile, and potentially a successful hybrid cross might (as well as being reproductively
feasible and fertile) combine high yielding genes of one parent with the genes for perenniality
from the other. Some perennial plants are known to be highly productive (alfalfa, pigeon pea,
etc.) at least for a few years (relative to shrubs or trees, these are better called ‘weak perennials’,
and it is likely that perennial grains will be used as weak perennials as well.
3

The dearth of perennial cereals and legumes reflects an underlying reality of plant
population ecology: currently, there are virtually no perennial herbaceous plants which achieve
levels of reproductive allocation equivalent to the most productive perennial trees, perennial
shrubs or annual herbaceous plants. In other words, the combination of extremely high
reproductive allocation (e.g. over 20% of annual assimilation), herbaceous growth form and
perennial life history represents a ‘missing’ combination of traits that is extremely rare either in
nature or in cultivated agroecosystems (Van Tassel et al., 2010). Agroecosystems can be viewed
as a novel ecological niche, created by human-directed artificial selection, and characterized by
high fertility, regular disturbance, and management of herbivory.
Artificial selection has resulted in very high levels of reproduction in many cultivated annual
cereals: for example, barley can achieve harvest index of over 60% (Kemanian et al., 2007).
Wheat has achieved a harvest index of up to 55%, while corn has achieved over 50%. Some
woody cultivated perennials can also achieve very high levels of reproductive allocation, once
established. Though reproduction is typically low for the first few years, mature apple trees
(Malus domestica) can achieve reproductive allocation of up to 80% (Xia et al., 2009). Blueberry
(Vaccinium corymbosum) can achieve 55% (Pritts and Hancock, 1985) while olive (Olea
europaea) can achieve 60% (Villalobos et al., 2006) and oil palm (Elaeis guineensis) can achieve
54% (Corley, 1983).
However, no perennial herbaceous plants with comparable levels of reproductive allocation
are currently known. It is as yet unknown whether this reflects unknown metabolic and energetic
tradeoffs in the perennial herbaceous life history, or rather because their combination of traits did
not favor early domestication (Van Tassel et al., 2010). Thus, as plant breeders and agronomists
attempt to develop sustainable and productive cropping systems based on herbaceous perennials,
4

it is critical to gather evidence as to the current levels of productivity in perennial cereals, grasses
and legumes, as well as the potential for breeding to improve them. Equally, it is necessary to
gain a better understanding of what factors influence perenniality vs. productivity tradeoffs from
the ecological and physiological points of view, and to learn what physiological mechanisms
high yielding herbaceous perennials could use to mitigate and negotiate these tradeoffs.
This review will first consider some of the promising current prospects for perennial
cereal and legume crops, especially those adapted for the cold temperate, deciduous forest/prairie
zone, and some of the environmental benefits that might be associated with increased
perenniality in agroecosystems. We will then consider inherent ecological and physiological
challenges that perennial cereals might face, and what mechanisms perennial cereals might use to
reduce tradeoffs between seed production and long term survival. Finally, we will briefly outline
how some hypotheses regarding perennial cereal physiology might be tested, and introduce the
questions we will be exploring in the rest of this dissertation as well as methods we will be using.

Intermediate Wheatgrass: A Case Study in Domestication
Many of the world’s most important staple food crops are either C3 or C4 grasses: rice,
wheat, corn, barley, millet, sorghum, rye, and others. Thus, some of the initial efforts to develop
perennial cereal crops have focused on developing new perennial grasses for food production,
either through domestication of wild species or through hybridization between existing crops and
wild relatives. One of the most prominent candidates for domestication is the perennial grass
intermediate wheatgrass, Thinopyrum intermedium (formerly Elytrigia or Agropyron).
Intermediate wheatgrass is a relatively high-yielding, cross-pollinated, rhizomatous grass with an
autoallohexaploid genome, whose native range stretches from the Balkans across southern
5

Russia to Pakistan. It produces erect stems with erect basal leaves, of a bright green to bluishgreen color: plants begin growing fairly late in spring, but grow quickly to a maximum height of
1-1.5 m (Smoliak, 1981). It was introduced into the United States as a forage grass in 1932, and
grows well in moderately dry to mesic areas with temperate climates and at least 350 mm of
rainfall. It can survive over 50 years, although when grown as a dryland forage grass it is
encouraged not to be grown for more than 5 years.
Intermediate wheatgrass today grows as a naturalized grassland plant and forage/hay crop
throughout the western and Great Plains states, as well as in New York, Massachusetts and
Georgia (Berdahl and Redfearn, 2007; Malinowski et al., 2003; Moore et al., 1979). It has
intermediate vegetative competitiveness, able to maintain itself against weed pressure without
becoming a noxious weed itself; under moderate herbicide levels in areas infested with downy
brome, emergence is between 300-400% better with intermediate wheatgrass than with another
closely related forage, crested wheatgrass (Evans et al., 1967). Indeed both “Reliant” and
pubescent subspecies of intermediate wheatgrass have been used to reduce leafy spurge
prevalence by up to 72% three years after planting (Lym and Tober, 1997). Other perennial
wheatgrasses (e.g. T. ponticum, T. bessarabicum) have been found to have considerable salt
tolerance, being able to survive up to 2.5 to 7.5% NaCl solutions; introgressions of T. ponticum
chromosomes into bread wheat have been shown to improve salt tolerance, resulting in only 15%
declines in seed yield at salt levels which cause a 50% decline in normal bread wheat (Colmer et
al., 2006). It remains to be seen to what extent intermediate wheatgrass has similar salt tolerance
in the field; greenhouse experiments in Texas (Miyamoto et al., 2004) found that some cultivars
of T. intermedium were moderately salt-tolerant grass, registering only a 25% reduction in shoot
-1

growth at salt levels corresponding to about 6-8 dS m (roughly 3000-4500 ppm NaCl, or 0.36

0.45% salt). Finally, intermediate wheatgrass is a reasonably high-yielding grass. The older
-1

cultivars (e.g. ‘Greenar” in the 1930s ) averaged 200-400 kg ha of seed yield, but more recent
cultivars have achieved up to 900 kg ha

-1

under ideal conditions.

Although the benefits of intermediate wheatgrass as a forage crop are known (moderately
high biomass yield, large seed, competitiveness with weeds, salt tolerance, strong perenniality), it
has not yet been selected as a grain crop for high seed yield. As currently seed yields are under
-1

900 kg ha , much more selection needs to be undertaken in order to make it an economical grain
crop. Knowles (1977) was able to increase seed yield in a population by 10% per cycle by using
mass selection, picking the 5% highest yielding plants at each cycle; when pollination was
controlled by removing selected plants to the greenhouse during the winter, he was able to
achieve 20% gains in seed yield per cycle. Further efforts were made to select intermediate
wheatgrass for higher yields during the mid 1990s; it was selected as a promising candidate for
domestication after scanning about 100 perennial grasses on the basis of the following criteria:
palatability of seeds, threshability, medium to large seed size, synchronous seed maturity,
resistance to sharttering, tendency not to lodge, seedheads held above the foliage level, drydown
of seed stalks at maturity, high potential for mechanical harvest, and vigorous perennial growth
(Wagoner and Schauer, 1990). Wagoner (1995) cross-pollinated 20 of the highest yielding
accessions (out of 300) and eventually derived 14 accessions which had yields 25% higher than
the original population mean.

Hybridization of Grain Crops (Poaceae) with Perennial Wild Relatives

7

Improving productivity of perennial grasses through repeated selection will presumably
be a long and slow process. The development of modern, high-yielding annual crops from their
wild forebears took place over thousands of years, and few additional crops have been developed
in recent centuries. However, it has been argued that a ‘shortcut’ is available to the development
of high-yielding perennial grain and grass crops, through hybridizing existing annual crops with
perennial wild relatives (Cox et al., 2002). In this case, the theory is that since the hard work of
selecting for high yield has already been done, the introgression of part of the genome from the
perennial parent could introduce perenniality into succeeding generations while still maintaining
reasonably high yields. It is possible, however, that introducing perennial chromosomes into an
existing annual cereal could result in large declines in yield, if they result in the loss (or simply
lower levels of expression) of QTLs which contribute to high grain yield.
Many annual grain crops are known to have perennial wild relatives with which they are
interfertile, and in some cases these wild relatives have been used in hybridizations in the past
(usually to introduce some desirable trait such as reduced lodging, increased salt tolerance, pest
resistance, etc.). Some potential or actual crosses of crops and their wild relatives are as follows:
all are known to be interfertile but a * indicates that there are few problems with sterility or
embryo rescue.

Rice (Oryza sativa) and Oryza rufipogon*
Sorghum (Sorghum bicolor) and johnsongrass (Sorghum halepense)*
Sorghum (Sorghum bicolor) and Sorghum propinquum
Oats (Avena sativa) and Avena macrostachya
Corn (Zea mays) and eastern gamagrass (Tripsacum dactyloides)
8

Wheat (Triticum aestivum) and wheatgrass species (Thinopyrum intermedium, Th. elongatum or
Th. junceum)*
Millet (Pennisetum glaucum) and Napiergrass (Pennisetum purpureum)
Rye (Secale cereale) and Secale montanum*
(Cox et al., 2002).

Perennial Wheat and Rye: Case Studies in Grass x Cereal Hybridization
In parallel with the efforts to develop intermediate wheatgrass into a commercially viable
perennial cereal crop, efforts have been made in the last few decades to develop a hybrid
perennial wheat species, through hybridizing annual wheat (Triticum aestivum) with various
Thinopyrum species (Th. intermedium, Th. elongatum and Th. junceum). In addition, breeding
efforts have successfully developed perennial cultivars of rye through hybridizing annual rye
(Secale cereale) with S. montanum.
Annual wheat has varying degrees of cross-fertility with Thinopyrum species, and over the
last century breeding efforts have repeatedly crossed wheat with one or more Thinopyrum
species in order to introgress specific traits of interest into the wheat cultivar. Wheat has been
crossed with Th. junceum in order to achieve greater salt tolerance (Wang et al., 2003), while
crosses with Th. bessarabicum and Th. elongatum have been made to achieve combined
resistance to high salinity and low oxygen (Akhtar et al., 1994). Crosses with Th. bessarabicum
have also been made in order to introduce resistance to copper, aluminum or manganese toxicity.
Likewise, crosses with Th. intermedium have been made in order to achieve resistance to leaf
rust, stem rust, stripe rust and Fusarium head blight. The popular Chinese cultivar Xiao Yan 6 has
been developed through two chromosomal substitutions from Thinopyrum elongatum (a
9

congeneric of intermediate wheatgrass) and combines high yield potential with resistance to
oxidative stress and photoinhibition, resistance to Septoria tritici and stem rust, and tolerance of
high temperatures, and is considered a successful Chinese cultivar (Li et al., 2008).
The long period of development of cultivars like Xiao Yan 6 (the Thinopyrum-wheat cross
was made in 1960 and the variety finally released in 1977) illustrates that intergeneric
hybridization is a complex and challenging procedure (Li and Wu, 2006; Li et al., 2008).
However, the example of these past crosses does show that many wild relatives can be
successfully hybridized with wheat, and that there are few immediate barriers to intergeneric
hybridization. While such hybrids are usually self-sterile, fertility can be restored either through
back-crossing to wheat or to the wild parent (e.g. Murphy et al., 2010) or by using colchicines to
double the genome.
Until recently, however, little effort has been devoted to hybridizing wheat and its wild
relatives for the specific objective of developing a perennial cereal species. Perenniality is a
more difficult trait to achieve than resistance to a particular pathogen, as it is a complex life
history adaptation involving changes in many genes, and has profound effects on patterns of
allocation, senescence and growth. However, attempts to develop perennial wheat began with
Soviet breeding efforts in the 1900s through 1930s. Unfortunately, these efforts were later
abandoned and little literature on perennial wheat crosses was communicated to the scientific
establishment in western countries. In general, they showed poor perennial regrowth and yielded
good harvests in the first year (Tsitsin, 1965). In the 1960s, efforts to breed perennial wheat went
on at the University of California, Davis, but were abandoned by 1965 due to insufficient yields.
The genotype ‘MT-2’, developed through hybdirization between Th. intermedium and Triticum
turgidum, suffered from low kernel mass and unreliable persistence (Schulz-Schaeffer and
10

Haller, 1987) while similar crosses in Sweden also suffered from low yields. Most lines required
reseeding after 2-3 years, and suffered from reduced winter hardiness, high mortality rates, and
rapidly declining yields in the second and third years. Thus efforts to breed perennial wheat
remained of marginal interest to the broader scientific and agronomic communities (Wagoner,
1990). By the 1990s, however, it became clear that perennial wheat might be a feasible crop in
spite of achieving lower yields than annual wheat, if it could compensate for lower yields
through improved ecosystem services. In addition, breeding efforts on wild Triticeae grasses (for
forage purposes) had helped produce higher yielding, strongly perennial accessions that might be
good perennial parents for developing truly perennial hybrid cereals.
Since the late 1990s, efforts have been ongoing at Washington State University (Pullman,
WA) and the Land Institute (Salina, KS) to develop amphiploid hybrids of wheat with
Thinopyrum relatives, which combine acceptably high yields with perennial life history. Wild
grass species which have been used as perennial parents in crosses with wheat included
intermediate wheatgrass (Th. intermedium), Russian wheatgrass (Th. junceum) and tall
wheatgrass (Th. elongatum). Early attempts at perennialization of wheat included crosses with
Th. intermedium (Banks et al, 1993; Scheinost, 2001), as well as the cross ‘AgCS’ with Th.
elongatum, which displayed perenniality. A few years later, genetic analysis of ‘AgCS’ found that
the ability for perenniality was linked to chromosome 4E of Th. elongatum (Lammer et al.,
2004). A QTL on this chromosome allows the capacity for post sexual cycle regrowth to be
turned on or off, although loci on other chromosomes may contribute to the degree of
perenniality, and indeed the amount of post sexual cycle regrowth (PSCR) may be linked to the
amount of perennial ancestry.
Because of its smaller chromosome complement, which will be described in further detail
11

below, Thinopyrum elongatum generally is more likely to produce stable hybrids with wheat than
either Th. junceum or Th. intermedium. For this reason, beginning in the late 1990s, efforts were
undertaken at Washington State University (Pullman, WA) to hybridize Th. elongatum with
perennial wheat, with the goal of developing a robust, highly productive perennial hybrid. Both
for purposes of improving fertility and gaining more productivity-related traits from annual
wheat, it was decided to back-cross the amphiploids to annual wheat (Wagoner et al., 1990). In
1998, the spring wheat cultivar ‘Chinese Spring’ was crossed with Th. elongatum, and the
resulting hybrid line was backcrossed to the winter wheat ‘Madsen’. Approximately 500 single
plants were selected from the F2 generation of these hybrids, and were grown out (including self
fertilization) to increase seed for two more generations. Out of these 500 lines, 31 were selected
for field experiments in 2005-2006 (on the basis of showing acceptable regrowth). These lines
were tested for seed yield and regrowth, as well as nutritional characteristics, at Washington State
(Murphy et al., 2009, 2010) and also form the source of the lines used in studies to be described
later in this dissertation.
Simultaneously, efforts have been ongoing at the Land Institute (Salina, KS) to hybridize
various perennial wheatgrass species with wheat, as well as at various other locations in China,
Argentina, Russia, Mexico and India. Perennial parents have included Th. junceum, Th.
intermedium, Th. elongatum and the decaploid Agropyron elongatum, while annual parents
included Triticum aestivum, Triticum durum, and Triticum carthlicum. A range of over 150 of
these accessions was tested for seed production, disease resistance and perennial regrowth in
Australia between 2009 and 2011 (Hayes et al., 2012), compared against the annual cultivars
‘Wedgetail’, ‘Naparoo’ and ‘Ventura’ as well as perennial grasses Th. intermedium and Secale
montanum. A concurrent study in China examined morphological characters and perennial
12

regrowth in a hybrid of wheat x Elymus, backcrossed to Th. intermedium (Shi et al., 2011).
Similar breeding efforts, with somewhat greater success thus far, have been devoted to
breeding a perennial version of rye. Annual rye (Secale cereale) is an important cereal crop in
cold-temperate regions of the world, used for food, cover, and forage, and particularly important
as a food crop in parts of Europe. Since the early 1980s, efforts have been ongoing to hybridize
annual rye with a perennial wild relative, Secale montanum. S. montanum is known to have
desirable traits including cold tolerance, ability to tolerate high levels of heavy metals, weedcompetitiveness, and high protein content (Tang et al., 2011; Kubiczek et al., 1981). S.
montanum has been found to have smaller yields and kernel weights than annual rye, but also to
have 50% higher protein content, higher mineral content and adequate baking qualities (Füle et
al., 2005).
Various efforts have been made over the past decades to hybridize S. cereale with S.
montanum, and have generally been somewhat more successful than similar efforts with
perennial wheat. However, S. montanum itself has only 20% the grain yield of annual rye, and
thus it has been difficult to identify hybrid lines that are both strongly perennial and high
yielding (Cox et al., 2002). Previous experiments with breeding perennial rye found that
perennial growth habit required homozygosity for a few critical chromosomes from the perennial
parent; however, such homozygosity was rarely achieved, and those lines which were
homozygous for the desired chromosomes showed low yields due to their preponderance of S.
montanum genetics. Efforts were made to develop a new cultivar, “Black Mountain”, through
backcrossing the amphiploid parent twice to S. montanum. However, though perennial, this
cultivar showed declining grain yields over time, and was useful primarily as a forage (Oram
1996).
13

Reimann-Philipp (1995) took a new approach, using colchicine to double the chromosome
number of the S. cereale x montanum amphiploid, and develop a tetraploid rye, ‘Permontra’. This
cultivar achieved a better combination of high grain yield and perenniality than previous lines.
-1

Reimann-Philipp (1986) found yields of 2.2 in the first year and 2.4 tons ha in ‘Permontra”
perennial rye, thus demonstrating good ability for regrowth and yield stability over at least two
years. In a second study, however, yields declined over time due to weed pressure. A later study
involving both “Permontra” perennial rye and intermediate wheatgrass (Thinopyrum
-1

intermedium) found that pure stands of perennial rye yielded 1.7 – 2.7 tons ha in its first year
(Weik et al., 2002). By the second year, however, yields decreased by 73-88 %., most likely due
to weed pressure and nitrogen limitation (for example, in pure stands of perennial rye, weed
biomass increased 2500% between the first and second years). Yield stability was somewhat
higher in mixed stands of perennial rye and lupin, in which “Permontra” yields declined only
50% between the first and second year. These studies would seem to indicate that both nitrogen
limitation and weed competition play a role in the ability of perennial cereals to regrow.
More recently, intensive efforts to develop a commercially viable perennial rye species have
been ongoing in Canada. Acharya et al. (2003) developed the new forage cultivar ‘ACE-1’ from
a Secale cereael x montanum cross. In a multi-year study, the ACE-1 perennial rye cultivar
produced 32% more biomass than annual rye when averaged over 2 years, but 16% less biomass
than annual rye when averaged over three years, which is evidence of a rapid decline in forage
yields over time (Acharya et al, 2003). In a subsequent study, Secale cereale x montanum ‘ACE1’ produced around 55% of annual rye (‘Kodiak’) yields when averaging across 1-year old, 2year old, and 3-year old plants; biomass yields of the perennial rye were between 5% and 190%
higher than annual rye (Acharya et al., 2004).
14

Potential Environmental Benefits of Perennial Agroecosystems
One of the primary reasons for the interest in large-scale perennialization of agriculture,
is the belief that perennial ecosystems can maintain higher levels of ecosystem services. It has
been hypothesized that perennial systems might have numerous advantages in terms of
ecosystem services, which could include 1) lower nutrient leaching, 2) less soil erosion, 3) lower
use of agricultural chemicals, 4) higher populations of pollinator and predatory insects, and 5)
speculatively, more efficient and mutualistic relationships with mycorrhizae. Some key pieces of
evidence supporting some of these hypotheses is outlined below.

Nutrient Usage and Leaching
There is some evidence that perennial systems could greatly reduce nitrogen leaching to
almost zero (Dowdell and Webster, 1980; Crews, 2005) could effectively sequester carbon into
long-term, recalcitrant organic pools in the soil, and stabilize the soil. There is also evidence that
perennial grasses, if developed into grain crops, could be more efficient users of nutrients. Ridley
et al. (2001) compared Australian pastures sown with annual (clover/ryegrass) and perennial
(clover/perennial grass) mixtures, subjecting each pasture to rotational sheep grazing and
allowing annual plants to replace themselves over the subsequent two seasons through natural
seed dispersal. They found that annual nitrogen losses were 33% lower in unlimed perennial
pastures and 45% lower in limed perennial pastures as compared to annuals; this is, of course, a
highly conservative estimate of the difference since the “annual” pastures included a certain
percentage of clover cover which may have increased over the three years of the study. Tallgrass
15

prairie in Kansas, as an example of a largely perennial grassland system, showed nitrogen
-1

-1

leaching of less than 1 kg ha yr , while conventially grown continuous corn grown with 187
-1

-1

kg of N led to leaching losses of 60 kg ha yr (over 30%); under similar conditions barley
-1

-1

grown with 120 kg of N lost 8%, or 10 kg ha yr (Crews and Peoples, 2005; Crews, op. cit.).
Fescue and alfalfa systems are known to have 86% less nitrate leaching during the lifespans of
the plants, but nitrate leaching undergoes a temporary burst as soon as the land is tilled
(Bergstrom, 1987).
Potential mechanisms of reduced nutrient leaching in perennial cropping systems,
especially perennial polycultures, include nutrient retranslocation, and deeper, more long-lived
root systems (Crews, 2005). Comparisons of closely related annual and perennial grass species
(Roumet et al., 2006) and of ecologically similar intermountain perennial and annual grasses
(Arredondo et al., 1998) have found that perennial grasses tend to have thicker roots that are less
efficient at nutrient uptake but longer-lived. Some perennial species, for example intermediate
wheatgrass have much deeper and larger root systems than closely related annuals (Cox et al.,
2006), and in these cases the deeper root systems could recover nutrients from deeper soil levels.
Nevertheless, even when perennial grasses are relatively shallow-rooted, the year-round
presence of living roots enables more nutrients and water to be retained by the soil. A study of
perennial and annual grasses in grasslands in Spain found that perennial grasses took up about
25% more nitrogen over the course of a growing season. Interestingly, N uptake by perennials
was 82% of the total ‘potentially available’ nitrogen in the early part of the growing season
(October-January), while N uptake by annuals was only 52% of the potentially available N (the
potentially available N levels being similar for annuals and perennials). The authors took this as
16

evidence of leaching: “[t]he difference between the nitrogen effectively absorbed and the
nitrogen likely to be absorbed by the annuals was probably due to partial leaching of the nitrates
following the autumn rains….” (Joffre, 1990). Net mineralization in soils supporting perennial
grasses was higher throughout the autumn rainy period than in soils supporting annual grasses;
since net mineralization is negatively related to leaching, this is evidence that leaching was lower
as well.
Perennial grasses can also be expected to use nutrients more efficiently through nitrogen
recycling. In a study of fescue (Festuca rubra) and common bent (Agrostis capillaris), labeled
15

N was used to determine the provenance of the nitrogen in next year’s regrowth. It was found

that 70% and 82%, respectively, of the nitrogen used in early spring regrowth was from nitrogen
retranslocated and stored from the last year’s leaves. This declined to 35% and 45% respectively
by June (Bausenwein et al., 2001). Spring growth of perennial grasses under N-limited
conditions has been found to be related to levels of nitrogen availability in the previous fall
(Gloser, 2005).
The phenomenon of nitrogen retranslocation would reduce leaching if nitrogen is
trasnlocated out of senescing leaves instead of remaining in the leaves as they die, decay, and are
gradually mineralized: the amount of mineralized nitrogen in the soil would be lower, and less
could be vulnerable to leaching. Indeed, perennial grasses are able to control rates of nitrogen
loss through the litter: under low-nitrogen conditions, they can increase retranslocation of
nitrogen from senescing leaves to roots. In an aquatic (hydroponic) analogue of nitrogen-stress
conditions for perennial grasses, wheatgrass (Agropyron dasystachyum) were first supplied with
labeled nitrogen fertilizer and then transferred to either high- or low- N conditions. Under low
nitrogen conditions, plants had much more labeled N in new leaf growth, suggesting that the
17

labeled N had been effectively conserved and translocated to growing leaves: in the high-N
treatment, the lower N translocation to leaves corresponded to increasing N escaping into the
solution (presumably equivalent to leaching under terrestrial conditions). Low-N plants leaked
no measurable amount of labeled N into their surrounding solutions, while high N plants leaked
about 12% of their total stored nitrogen into the surrounding solution. If these results hold true in
natural, field conditions as well as in the greenhouse, they constitute powerful evidence that
perennial grasses are able to reduce leaching losses, when necessary, to almost zero through
effective nutrient recycling (Li et al., 1992). Perennial grasses under low nitrogen conditions
have been found to have lower N losses through leaf litter and higher N losses through root
senescence and exudation (suggesting increased translocation and storage of nitrogen in roots),
while overall nitrogen losses are lower than under high-nutrient conditions- in other words, the
decreased leaf losses more than counteract the increased root losses (Vasquez de Aldana et al.,
1996). Gusewell (2005) found similar evidence of nitrogen translocation, but in contrast to the
previously cited paper, found that rates of nitrogen resorption were linked to phosphorus, not
nitrogen concentrations: under low P conditions, wetland grasses increased their efficiency of
both nitrogen and phosphorus use by increasing resorption of both nutrients. Drought is yet
another stress which can increase nitrogen translocation from leaves to rhizomes and roots
(reducing leaf N concentrations by up to 40% and increasing rhizome concentrations by 20100%) in warm-season C4 perennial grasses (Heckathorn and DeLucia, 1994).
Beyond the reduced risk of leaching, however, improved nitrogen translocation means
that less applied fertilizer would be needed for early season regrowth: since only 20-30% of the
plant’s nitrogen needs are being met by taking up nitrogen from the soil, growers of perennial
grain forages or crops could potentially reduce nitrogen fertilizer applications in the early spring.
18

The effects of perenniality on nitrate leaching may interact with rainfall effects. In a study
of leaching under fully perennial and fully annual grass pastures (with no legume mixtures)
Ridley et al. (1999) found that perennial pastures had lower levels of nitrate leaching in dry
years, but not in wet years: apparently increased precipitation was enough to overcome the
ability of the deeper root systems to retain nitrate ions. The effects of perenniality may also
depend on how long the perennial crop is allowed to survive: studies on alfalfa indicate that
nitrate concentrations in deep soil layers are lowest after four years, and start increasing again in
the fifth and six years, as roots begin to senesce (Entz et al., 2001). In general, comparisons
between annual and perennial systems are fairly spare but seem to indicate that perennial systems
have great potential to minimize leaching: to optimize these benefits however, we need more
information on perennial grass life cycles and how long these crops should be left in the soil.

Improved Carbon Sequestration
Perennial grain crops could potentially carry on photosynthesis later in the fall than
annuals, and begin photosynthesizing earlier in the spring (due to faster early-spring growth
fueled by belowground reserves). It is possible that due to extra photosynthesis early and late in
the growing season, more total CO2 could be assimilated on a yearly basis by perennials than by
annuals (Cox et al., 2006). For example, intermediate wheatgrass and Illinois bundleflower
maintain a large and photosynthetically active leaf area during August-September, a time when
wheat is not growing at all. The higher amounts of total seasonal assimilation in perennials
would be enhanced even more if perennial species maintained higher photosynthetic rates per
unit time, than their annual relatives. While this is not the case in many annual vs. perennial
species comparisons, some perennial species are known to maintain higher photosynthetic rates
19

than closely related annuals. These questions will be examined in detail, through both a review of
relevant literature and an experimental study of three perennial cereals, in Chapter 5 of this
thesis.
In addition, perennials may have larger and deeper root systems, and their root systems
can be expected to be longer-lived (since they must support the plant’s survival through each
winter). Indeed, Cox et al. (2006) found that only 25-40% of perennial wheatgrass roots die back
each winter as compared to 100% of annual wheat roots. The lower root turnover rates can be
expected to lead to a slower rate of organic carbon influx into the soil, and lower rates of soil
respiration. Perennial systems would also be incompatible with frequent tillage, which is a major
contributor to the breakdown and loss of soil organic carbon- e.g. West and Post (2002) found
that a switch from conventional tillage to no-till farming, across a variety of cropping systems,
resulted in an average of 57 g C m

-2

-1

yr . Thus it is reasonable to expect that, for one or more of

these reasons, carbon assimilation will be higher and/or soil respiration lower under perennial
cropping systems than among annual systems, therefore total carbon sequestration would be
higher. In an era of increased concern about climate change and rising CO2 levels, which are
expected to reach 560 ppm by 2050, perennial cropping systems could be a valuable means for
societies to reduce their total contribution to global climate change and for farmers to benefit
from receiving ‘carbon credits’.
To what extent is the hypothesis of increased carbon sequestration borne out in reality?
As yet, perennial grain crops are still in the process of development, so there is little assessment
of their contribution to CO2 fluxes in the field. Some evidence, however, is forthcoming from -1

perennial forage or energy crops sequester 320 to 440 kg ha of carbon per year, while the
20

annual crops studied fixed from 0-300. Zan et al., (2001) found that switchgrass crops grown for
bioenergy purposes resulted (after 4 years of production) in soil organic C concentrations
comparable to those of willow trees (the most carbon-negative system), while annual corn fields
had lower levels than willow plantations. Liebig et al. (2008) found that net carbon sequestration
-1

under switchgrass ranged from -0.6 (i.e. net CO2 production) to 4.3 Mg C ha over the course
of a year, thus generally this system was a carbon sink. Location was found to strongly influence
the differences in soil organic carbon that developed over eight years between a restored
perennial grassland (including various wheatgrass and needlegrass species) and a wheat-fallow
rotation; these differences varied between footslopes, midslopes and shoulder areas as well as
between sites (Nelson et al., 2008). Perennial grasslands sequestered up to 120% more carbon at
one location, and about 15% less carbon at another site; in general, however, since carbon
sequestration was higher in the grassland treatment at 6 of the 9 locations, we can tentatively
conclude that carbon sequestration may be improved in perennial grain systems.
One caveat to the goal of improved carbon sequestration under perennial systems, is
posed by the fact that buildup organic carbon is quickly erased by single tillage events. Thus in
order to optimize carbon sequestration in perennial cereal systems, it may be necessary to grow
these species in reduced-tillage systems.
Research is ongoing at Kellogg Biological Station of Michigan State University, to test both
of the above hypotheses: that perennial cereal species show lower rates of nutrient leaching, and
higher rates of carbon sequestration, than closely related annual species (Culman S. et al.,
unpublished data; Sprunger C., unpublished data). An ongoing four year experiment compares
annual wheat (Triticum aestivum cv. ‘Caledonia’) to a closely related, cool-season perennial
cereal species, intermediate wheatgrass (Thinopyrum intermedium). The wheatgrass accession
21

(‘TLI-C1) has been bred at the Land Institute (Salina, KS) for higher productivity, threshability
and palatability, while ‘Caledonia’ is a wheat variety well adapted to Michigan and considered
the ‘variety to beat’ in the mid-2000s (Lewis JM, personal communication). These systems are
being compared over a period of four years, with the wheatgrass growing for a four year cycle
and the wheat being replanted every year. Rates of nitrogen leaching are being monitored with
lysimeters, while rates of organic carbon buildup are being assayed through a permanganate
oxidation method (Culman S. et al., unpublished data).

Improved Soil Quality
Increased soil organic carbon, of course, would not merely be beneficial from the point of
view of offsetting greenhouse gas emissions, but also from the point of view of increasing soil
quality. Higher levels of soil organic matter could improve nutrient retention, water retention,
soil bulk density, soil structure, and soil microbe populations. If significant areas of land in, for
example, the Great Plains were dedicated in the long term to perennial cropping systems, soil
organic matter could increase over time to levels comparable to those before the advent of
modern agriculture. This could allow for eventual rotation of perennial and annual cropping
systems, with the increased levels of soil organic matter built up by the perennials able to support
high productivity by the annuals.
Islam and Weil (2000) found that perennial grasslands planted with Napier grass or wild
sugarcane, had comparable soil bulk density, aggregate stability, and N content to natural fores,
and were superior to cultivated fields including rice, cotton, and sugarcane. While this
experiment was carried out in a tropical environment, it is interesting in part because both
perennial grasses that they looked at are congeneric relatives of existing annual crops (millet and
22

sugarcane) and are currently under investigation as contributing to potential perennial crops. A
comparison of perennial kikuyugrass (Pennisetum clandestinum), native grassland and annual
ryegrass (Lolium multiflorum) in South Africa found similar results in a Mediterranean climate.
Soil organic C, potassium sulfate-extractable C, microbial biomass C, and soil aggregate stability
were lower under the annual than under the perennial treatments. In the case of aggregate
stability, values were higher under the perennial by 33%-100% depending on the site (Milne and
Haynes, 2004). Haynes and Francis (1993) found similar results to temperate environments, and
found that perennial ryegrass pastures produced higher levels of soil aggregate stability than
annual pastures.
The ability of perennial grass species to improve soil quality and soil physical properties
as well as improving the efficiency of nutrient use, suggests that perennial grain crops could do
similar things. These are essentially partly perennial, high-yielding grass or forbs, and they
follow perennial or weakly perennial growth habits. If the improvements in soil quality under
perennial pastures are related to perenniality itself, and not merely species effects, then it is
reasonable that perennial grain crops could potentially result in improved soil quality and be an
important part of soil restoration and sustainable agriculture efforts in future.

Resistance and Tolerance to Herbivory and Insects
Insect and pest pressure is an important environmental stress that most agricultural
systems have to deal with. It represents a particularly difficult stress to overcome since it is a
‘moving target’ unlike, say, nitrogen deficiency. Insects and other pests evolve to become
resistant to common pesticides, and have been found to develop resistance in as little as five
years. Because insect pressure is such an important stress to overcome, it is important to know
23

how perennial cropping systems might vary from annual ones in terms of the ability to resist
insect attack, tolerate herbivory, or facilitate biological control of pests.
In recent years, more attention has been paid to biological control, i.e. encouraging and
fostering natural enemy populations, as a means of crop protection. Natural enemies (primarily
predatory or parasitoid insects or arachnids) have many advantages over the use of many
chemical insecticides. Natural enemies are usually specific for the desired pest, and to some
extent self-limiting as their populations track those of their prey; they have the advantage of
biological reproduction, so that a population can be started from a very small seed population,
and they do not introduce broad-spectrum poisons into the environment. While in recent years
the use of genetically modified (Bt) crops has overcome some of the problems with older, less
specific insecticides, these varieties are only available for a few crops and it seems likely that
encouraging natural enemies will continue to be a valuable strategy in the future.
There is some reason to believe that diverse, perennial grain-based polycultures could
support larger, healthier and more diverse populations of natural enemies than current, lowdiversity and high-productivity annual systems. It has been argued that perennial crop systems
would favor conservation biological control through providing habitat to predators and
parasitoids through a longer portion of the year, and because they are subject to lower levels of
disturbance (Landis et al., 2000). Perennial grasses provide a habitat where natural enemies can
overwinter, and are able to re-colonizie rapidly in the spring, quickly establishing control over
the pest population; strips of perennial grasses have been used to increase predator habitat in
annual cereal fields (Thomas et al., 1992). Perennial grain crops may foster growth of predatory
beetle populations by providing continuous ground cover (Clark et al., 1997). There is evidence
that perennial grain crops could be more resistant to herbivory (Gutman et al., 2001), more able
24

to adapt to rapidly changing environments by altering the balance between aboveground and
belowground investment (ibid.), more able to produce anti-herbivore toxins, and more able to
attract natural enemies of herbivores (Tschartnke and Greiler, 1995; Bommarco, 1999; Fiedler
and Landis, 2007), and potentially more supportive of pollination (Tuell et al., 2008). Pollination
and biological control services are critical to agriculture as well as to the health of natural
ecosystems, and if perennial grain crops can support higher levels of these beneficial insects then
they could be an important aspect of sustainable managed and semi-managed ecosystems in
future, which could ensure higher levels of ecosystem services not merely on perennial grain
lands themselves but also on landscapes as a whole.

Challenges to Perennial Cereal Crops
Perennial cereal crops face a number of challenges, if they are to become viable crops
which are economically feasible for farmers to grow, while yielding significant environmental
benefits. For the remainder of this review we will be focusing on three major distinct challenges
which face perennial herbaceous crops in general, but we will specifically consider each topic
through the lens of three perennial cereals: perennial rye, perennial wheat and intermediate
wheatgrass. Each of these species is under development as an alternative food crop for the cold
temperate zone, and while the issues that they face are to some extent common to other
herbaceous perennial species, they can be most usefully considered with reference to the specific
challenges posed by the genetics, agronomy and physiology of these particular species. In this
review, we shall briefly outline issues of genetic incompatibility, disease pressure, and yield vs.
perenniality tradeoffs, with reference to our three perennial species. While the remainder of this
dissertation will not explicitly address genetic and breeding issues, Chapter 2 will address one
25

important aspect of pathogen resistance (specifically, resistance to Fusarium graminearum head
blight). Chapter 3 will address the remaining challenge (yield vs. perenniality tradeoffs).
Chapters 4 and 5 consider basic questions of resource acquisition and allocation in perennial
cereals, which should help cast further light on the extent to which high yield and perennial life
history might be compatible.

Genetic Incompatibility
Any wide cross between widely differing lineages faces potential issues of genetic
incompatibility. The fact that the offspring contain only half the genetic complement of each
parent could result in insufficient expression of some traits necessary to metabolism, defence,
reproduction or resource acquisition. This may be somewhat less of a problem in a polyploid
species, where homeologous genes on different chromosomes may fulfil the same or similar
roles, but it can still pose a challenge. Any cross between the two will probably result in a loss of
some yield potential, for several reasons: some genes affecting yield may be incompletely
dominant so that the heterozygote has lower yield, while some genes from the perennial parent
may somewhat downregulate allocation to seed production. It will be necessary to select for a
number of generations to get yield up to an acceptable level. Moreover, in the F2 and later
generations, there is potential for heterozygosity, which means that the offspring may segregate
for different traits and selection will be necessary to identify which offspring have the desired
trait. Finally, sometimes the chromosome number between parents may differ, which creates
problems for chromosome pairing and meiosis in the F1 generation unless chromosome doubling
occurs (e.g. with colchicine treatment: Jauhar, 1995; Anamthawat-Jonsson, 1996).

26

Challenges specific to the Triticum x Thinopyrum cross include difficulties in
chromosome pairing, unstable chromosome number, as well as potentially high genetic load.
Triticum aestivum (2n = 42) and Thinopyrum intermedium (2n = 42) have the same number of
chromosomes, but pairing can generally not occur because of the Ph locus in wheat, which
allows only homologous chromosomes (as opposed to homoeologous chromosomes) to pair
during meiosis. This means that the hybrids act functionally as diploid hybrids between two
completely different genomes, each comprising 21 chromosomes. Pairing cannot occur, which
usually results in the hybrids being self-sterile.
In making crosses with Thinopyrum spp., plant breeders have found varying ways to deal
with this problem. Some breeders have used Chinese Spring ph1 mutants, or other ph1-deficient
cultivars (e.g. Angas, Australian). The inhibition of pairing trait in wheat is recessive, thus only
one nonfunctional copy is needed (Marais, 1992; Han et al., 2003). Artificial doubling of the
genome, using agents like colchicine or dimethyl sulfoxide (DMSO) can also produce self-fertile
plants with twice the number of chromosomes (Anamthawat-Jonsson, 1996). This approach has
been successfully used with other Thinopyrum crosses but not yet with Th. intermedium. The
self-fertile plants are often only female-fertile, as aneuploidy tends to be more of a problem for
male gametes (complete male sterility) than for female gametes, This means that the F1 plants
can be used as female parents in backcrosses to either wheat or the perennial parent. In perennial
wheat breeding efforts at Washington State University (Murphy et al., 2009; Murphy et al., 2010)
the wheat x Thinopyrum crosses have been back-crossed to annual wheat.
Sometimes wheat x Thinopyrum hybrids may ‘naturally’ produce a certain number of
unreduced gametes, which contain a complete set of each haploid genome; these can combine
with gametes from either parent to produce a BC1 plant. There may also be ‘natural’ genome
27

reduction as certain chromosomes are eliminated; this is a common result of wheat-Thinopyrum
crosses. While this is not a problem for those breeders who are looking to introgress a specific
trait into wheat, since only one or two chromosomes are needed, it is a problem for people
interested in developing a new perennial wheat species, and these plants (with less than a full set
of Thinopyrum chromosomes) cannot be used in further breeding.
The unreduced gametes approach is the approach that breeders in Washington and Kansas
have used thus far in producing perennial wheat lines (Anamthawat-Jonsson et al., 1997; Cox et
al., 2002). The breeders have used cold temperatures during pollination and embryo development
to promote the formation of unreduced gametes- this results in gametes with twice the ‘normal’
number of chromosomes (n = 42) so that pairing and meiosis can occur. These plants, with
unreduced gametes, are self-fertile (which has been used to develop lines in Kansas). They can
also be used as the female parent in a back-cross with wheat (as has been done in Washington)
and produce self fertile offspring. However, the high chromosome number is unstable, and some
of the Thinopyrum chromosomes tend to be spontaneously eliminated so that after several
generations the chromosome number stabilizes around 2n = 56 (42 from wheat, 14 from
wheatgrass). This results in segregation for traits, including perenniality (as some of the plants
have eliminated the 4E chromosome which is linked to perenniality) and is clearly not ideal (Cox
et al., 2002). As a future strategy, breeders in Kansas are attempting to try to back-cross the F1
generation to the perennial parent (Th. intermedium) and develop a new crop species that will
have at least as much Thinopyrum ancestry as wheat (Anamthawat-Jonsson et al., 1997).
An additional issue is the unstable chromosome makeup in Thinopyrum x wheat hybrids
(Dvorak, 1976). Annual wheat has 42 chromosomes, while durum (tetraploid) wheat has 28:
Thinopyrum intermedium has 42, while Th. elongatum has 14 (Cox et al., 2002) and Th. junceum
28

has 42 (Wang et al., 2010). Combining the genomes, following doubling or lack of reduction of
gametes, would lead to 2n = 84 chromosomes for Th. intermedium x T. aestivum hybrids.
However, this chromosome number is unsustainably large (Schulz-Schaeffer and Haller, 1987;
Schulz-Schaeffer and Haller, 1988; ) and thus the chromosome number generally stabilizes
around 2n = 56, with the loss of some chromosomes (Jones et al. 1999; Banks et al., 1993),
though 2n = 42 hybrids are more stable meiotically. However, this chromosome loss takes place
somewhat randomly, with the elimination of both wheat and Thinopyrum chromosomes. Thus the
progeny of any given cross will be extremely variable, without even having the same
chromosomes, and many will suffer from extremely deleterious traits, since they may be missing
large amounts of critical genetic material. The cross ‘AT-3425’, for example, derived from a
wheat / Thinopyrum cross for the purpose of Cephalosporium resistance, had 36 wheat
chromosomes, 14 Thinopyrum chromosomes, and 6 derived from translocations between the two
(Cai et al., 1998). Another perennial amphiploid line, ‘AgCS’, with a stable chromosome
number of 2n = 56, was derived through hybridization of wheat and Th. elongatum (Cai et al.,
2001).
Because of the smaller genome of Th. elongatum compared to Th. intermedium or Th.
junceum, wheat hybrids with Th. elongatum are more stable, and can in fact stabilize at the
threshold of 2n = 56 without any chromosome loss at all (though as this is still above the ideal
for meiotic stability, some chromosome loss may take place in succeeding generations). For this
reason, in the mid-2000s, breeding efforts at Washington State University began to focus on
Thinopyrum elongatum as a source of perennial parentage in wheat x wheatgrass amphiploids
(Murphy et al, 2009: Murphy et al, 2010; Jauhar, 1992; Jauhar, 1995; Cox et al., 2002).

29

A third genetic challenge to developing wheat x Thinopyrum hybrids is that wheat is a
mostly self-pollinated, self-compatible crop (and thus has very few deleterious recessive alleles
and high tolerance for inbreeding; DeVries, 1974; Prakash and Singhal, 2003) while Thinopyrum
intermedium, Th. junceum and Th. elongatum are cross-pollinated and only slightly to
moderately self fertile (relying on wind or sometimes insects in nature: Jensen et al., 1990). Selfsterility in Thinopyrum is recessive, so the hybrids are self-fertile, but in the F2 generation (or in
the F2 generation following one backcross to wheat) deleterious recessive alleles may recombine
so that many of the progeny are homozygous for at least one of these deleterious traits. This
results in a drop in yield in the F2 generation, and extensive selection will be necessary to
eliminate unfavorable genotypes.
In perennial x annual Secale hybrids, inbreeding depression is less of an issue. Annual rye
is, itself, mostly an obligate outcrosser (Landes, 1939) and thus the Secale cereale x montanum
hybrid is outcrossing as well. While this protects the hybrid species against inbreeding
depression, it also exposes it to pollen limitation. Under conditions of pollen limitation, both
annual and perennial rye can become vulnerable to ergot (especially when weather conditions are
favourable for the pathogen) as the fungus preferentially colonizes unfertilized spikelets
(DeHaan LR, personal communication). The breeding syndrome of perennial rye means that
breeders, agronomists and farmers must be cognizant of the vulnerability of this species to ergot
attack.

Disease Pressure
A second challenge facing perennial cereals is the increasing buildup of pathogens over a
multi-year life cycle. Farmers commonly manage disease through crop rotation: as pathogens are
30

often specific to a particular species, genus or family, the buildup of a particular pathogen may
be avoided by tilling up that crop after harvest and replacing it with an unrelated crop the
following season. Crop rotation would not help suppress airborne pathogens such as wheat loose
smut, powdery mildew, sugarcane leaf rust or Southern corn leaf blight, which can disperse long
distances to colonize new areas (Nagarajan and Singh, 1990). These pathogens are able to spread
hundreds or thousands of miles on air currents, and can spread from one area of cultivation to a
spatially separated area. However, other pathogens have generally shorter dispersal patterns, and
normally overwinter either in the soil or on crop residues. Pathogens which overwinter on
residues or in the soil, and which spread through short distance dispersal events (e.g. through
water droplets) can be effectively suppressed through crop rotation.
Perennial crops inherently have a multi-year life cycle, and thus rotation cannot be used
as a method of disease control. This poses a problem for the control of any pathogen which
overwinters in the soil or on crop residues. Shortened rotations and reduced tillage are known to
increase the incidence of some fungal diseases, e.g. Fusarium head blight (FHB: Xu and
Nicholson, 2009), as this pathogen can overwinter on crop residues (Blandino et al., 2009). Thus
increased disease prevalence poses a major threat to the viability of perennial cereals. FHB, in
particular, could be a critical roadblock to developing perennial wheat, wheatgrass and other
perennial cereals for the cold temperate zone.
As cultural control (crop rotation, changing tillage frequency and planting time, removing
residues, etc.) cannot be used in a perennial system, options for controlling pests include either
biological control, chemical control, or breeding for resistance traits. Biological control would
involve bacterial or fungal agents antagonistic to disease-causing agents; chemical control could
involve the application of pesticides. Breeding for resistance traits could be an important strategy
31

in ensuring that perennial cereals do not pay a high disease cost for their perennial life history. If
a perennial cereal cultivar is highly resistant to a particular pathogen, then even if it is grown in
a single location for multiple years, the pathogen will not be able to colonize.
In Chapter 5 of this dissertation, we will consider a case study particularly relevant to
perennial cereals of the cold-temperate zone. Fusarium head blight, caused by the ascomycete
fungus Fusarium graminearum (teleomorph Gibberella zeae) is one of the major diseases of
wheat and other small grains in north America, as well as an important pest in Europe and China.
It not only reduces yields, but also renders the grain toxic to livestock and humans. F.
graminearum overwinters on crop residues, and thus the presence of over-wintering colonies is
considered to be the major source of inoculum under standard field conditions. Biological and
chemical efforts to control FHB have achieved partial control, at best, suggesting that if
perennial cereals are to become commercially feasible, it is essential to breed for genetic
resistance to FHB. For this reason, plant breeders developing perennial cereals have emphasized
identifying strongly disease resistant lines, that can maintain a perennial growth habit without
becoming a reservoir for disease (e.g. Cox et al., 2005).
Various wild Thinopyrum species have been used as sources of resistance to FHB in the
past, suggesting that perennial wheat x Thinopyrum crosses might show adequate levels of
resistance. In Chapter 5, we will attempt to assess FHB resistance in various accessions of three
perennial cereals (perennial wheat, perennial rye and intermediate wheatgrass) following
inoculation experiments in the greenhouse. We chose this species because of its economic
importance in North America, and because it was naturally present to some degree in initial
screenings of field-grown perennial wheat plants in 2010.

32

Some diseases, including important viral pathogens like barley yellow dwarf virus
(BYDV) are spread by insect vectors (in this case, aphids, as they move from plant to plant and
suck on the leaf sap). In the case of insect-borne viral diseases, the potential of perennial systems
to harbor more predatory insects and to experience lower levels of herbivory could potentially
help suppress herbivory and thus, the spread of these diseases. While this dissertation does not
address BYDV or other insect borne viral diseases, an observational study of three fungal and
one viral disease (BYDV, leaf rust, Septoria leaf blotch and Stagonospora glume blotch) is
ongoing at Kellogg Biological Station, and will be completed in 2013: we hope to eventually
report on observed susceptibility of selected accessions of perennial rye, perennial wheat and
intermediate wheatgrass to these diseases as well. All of these diseases are critical diseases in
North American small grain growing region. Although for reasons of space they are not
addressed in this dissertation, and although we treat FHB as a case study in disease resistance of
perennial cereals, it is important to gather information on these widespread diseases as well,
particularly since BYDV in particular affects a broad range of plant taxa and could be applicable
to other perennial agroecosystems as well.

Perenniality vs. Productivity Tradeoffs
The third great challenge to the idea of developing perennial grain crops lies in the
tradeoff between first year yield and longer-term survival and productivity. Farmers and land
managers generally hope for a plant that can give them moderate to high yields beginning as
soon as possible, ideally in the first year. Perennial plants are generally less well adapted to rapid
growth and yield than annuals- while they may produce more over the course of their lifetimes,
this is also not certain. It is often assumed that perennials would have lower production over their
33

lifespan than a series of annual crops over the same period, but especially in marginal
environments, this can be questioned. According to FAO statistics, the yield of bananas and
strawberries in Central America (two perennial crops), was an average of 8900 and 2400 kg ha

-1

respectively, while the yield of the three major dryland cereals (wheat, sorghum, rice and corn)
-1

were 4200, 2600, 3000 and 2100 kg ha respectively (DeHaan et al., 2005).
It is likely that production costs for a perennial crop would be lower (due to reduced
tillage, seed expenses, fertilizer expenditure and potentially insecticide as well) and that the
ecosystem service benefits would more than outweigh any foregone income: for example, a
modeling study in Australia found that in marginal lands, perennial wheat could be economically
feasible if it yielded only 60% of annual wheat, with no additional forage gains. It could
additionally be economically feasible in the most marginal areas if it yielded 40% of annual
-1

wheat yields, with an added 800 kg ha of forage (beyond the forage production of the annual
wheat: Bell et al., 2008). While some evidence to data suggests that perennial wheat, for
example, can achieve yields up to 70% of good annual wheat cultivars in the first year, more data
is needed on how and under what conditions perennial grains are able to achieve these yields,
and if they lose other benefits in the process (DeHaan et al., 2005).
The capacity of perennial cereals to achieve yields comparable to annuals depends, in
large part, on their capacity to simultaneously allocate resources to vegetative storage organs
which ensure multi-year survival, and to seed production (Bazzaz et al., 2000). These can be
viewed as alternative “sinks” (resource demanding tissues) that compete for a common resource
pool (Reekie and Bazzaz, 1987a). Much of the argument that perennial cereals cannot achieve
high seed yields competitive with annuals, depends on the idea that the perennial vegetative sink
34

will consume so many resources that the capacity of the plant to produce seed is greatly limited.
Tradeoffs between perennial life history and seed production have often been assumed in the
past, and have been shown in some taxa (Law, 1977; Aragon et al., 2009). However, if such
tradeoffs are mild enough that perennial cereals still maintain yields over the acceptable
threshold, the existence of somewhat lower yields would not be an insuperable problem.
Resource tradeoffs can be expressed in terms of several limiting resources including
carbon, nitrogen (Wheelwright and Logan, 2004; Witkowski and Lamont, 1995), phosphorus
(Zotz and Richter, 2006) and others. For example, Zotz and Richter (2006) found that
nonstructural carbohydrate concentrations did not decrease following a reproductive event;
concentrations of N decreased in some plant tissues but not others, while P, K, and Mg
concentrations decreased in all plant tissues. The authors concluded that in this species
phosphorus was the factor that limited reproduction, reflecting the fact that up to 60% of total
plant phosphorus was concentrated in the reproductive tissues. If nitrogen or phosphorus is the
most important limiting factor, then this could help reduce resource costs of perenniality, since
perennial storage organs can generally be expected to be fairly carbon-rich and nutrient-poor., If,
for example, nitrogen is the limiting resource, then plants could produce large quantities of lowN root tissue without much reducing its ability to invest nitrogen in N-demanding seed. This
dissertation will primarily focus on resource tradeoffs as mediated by carbon: it is important,
however, to realize that if these plants are in fact limited by other resources (nitrogen,
phosphorus, etc.) this could give a conservative estimate of the yields possible in a perennial
system. Perennial cereals which are nitrogen or phosphorus limited could potentially invest large
amounts of carbon in high carbon, low-nutrient vegetative storage organs, gaining a large
advantage in perennial survival and regrowth ability at only a small cost in seed yield.
35

Carbon is one of the most important currencies in which to assess resource tradeoffs:
carbohydrates form the basic energy currency of plants, while fixed carbon is also required for
nearly every other biological molecule. Carbohydrates are a versatile energy source that can be
moved around the plant and invested in acquiring other resources when they in turn become
limiting. For example, under water stress carbohydrates could be invested into key solutes used
for osmotic regulation: under nutrient stress they could be invested in root systems or fed to
phosphorus-harvesting mycorrhizae; under conditions of predation they could be invested in
defence. In addition, many plants are effectively under water stress, and carbon fixation
inevitably happens at the expense of water loss through the stomata. Thus for heavily waterlimited plants, carbon could be viewed as a limiting resource, as greater rates of carbon fixation
at a given level of ambient carbon dioxide would mean that stomata could remain open for fewer
hours of the day. For all these reasons, carbon has often been viewed as the most generally
relevant currency to assess costs of reproduction (Reekie and Bazzaz, 1987b).
In this dissertation we will consider resource tradeoffs between perennial and annual
cereals specifically with respect to their ability to accumulate carbon through photosynthesis, and
to allocate fixed carbon to various ‘sinks’. This is particularly of value when considering
resource costs of perenniality, as perenniating storage organs can generally be expected to be
carbon-rich and relatively nitrogen-poor.
The degree of the yield penalty in perennial cereals (amount by which they fall short of
annual analogues) will depend on several factors: 1) the the extent to which reproductive
structures can be self-supporting vs. resource drains on the plant, 2) the actual relative resource
costs of perennial survival and of reproduction, 3) the extent to which photosynthetic rates differ
in perennial vs. annual cereals, and 4) the extent to which increasing reproductive or vegetative
36

sink strength stimulates increased compensatory photosynthesis. While 1) and 2) will be briefly
dealt with here, Chapters 3 and 4 of this dissertation will consider photosynthetic differences
between annual and perennial cereals as well as how these species compensate metabolically for
changes in sink strength.

Reproductive Photosynthesis
Tradeoffs between perenniality and seed yield could also be reduced by reproductive
photosynthesis, by which reproductive structures themselves support a large portion of their own
carbohydrate demands (Obeso, 2002). Consistent evidence shows that in many grass and forb
species, reproductive structures can support from about 17% of their own growth in Ranunculus
to about 65% in Acer platanoides, through their own photosynthesis (Galen et al., 1993; Aschan
and Pfanz, 2003; Bazzaz and Carlson, 1979; Bazzaz et al., 1979) and in some cases can even be a
net exporter carbon to damaged root or leaf tissue. Jackson and DeWald found (1994) evidence
that in Tripsacum dactyloides, reproductive tillers can ‘rescue’ extremely stressed vegetative
tillers by exporting carbon. Plants with intact reproductive stalks and vegetative stalks clipped to
5 cm, had about 60% the final biomass, and 75% the reproductive biomass, of intact plants,
while plants with defoliated reproductive stalks and clipped vegetative stalks had only about half
the final biomass and 40% of the reproductive biomass. This was the case for high-yielding
(pistillate) mutants as well as for normal plants.
Salopek-Sondi et al., (2000) and Aschan et al. (2005) found that sepals of Helleborus
could photosynthesize at 58% to 80% of the rate of leaves despite having much lower levels of
chlorophyll, while in Lilium the anthers can fix carbon at 73% the rate of leaves (Clement et al.,
1997). Hoch (2005) found that in several European trees (beech, hornbeam and linden) fruiting
37

branches support all of their own carbon demand and do not constitute a sink for the plant.
Girdled branches, unable to import carbon and nutrients from the rest of the plant, experienced
changes in carbohydrate concentrations but generally did not suffer a yield decrease. He further
suggested that this pattern of complete independence of reproductive tillers would probably not
hold for commercially valuable fruit trees and herbaceous crops that have been bred for high
yield. Nevertheless, this does provide good evidence that at least in some perennial plants,
reproductive structures can be self-supporting. However, it can be argued that as branches are
specialized for carbon acquisition rather than fruit bearing, they are not good analogues of
reproductive tillers on a herbaceous plant.

Annual vs. Perennial Differences in Resource Acquisition
The ability of perennial cereals to allocate resources to both seed yields and perennial,
multi-year survival depends ultimately on the combination of resource allocation and resource
acquisition patterns. This is because total allocation to reproduction, for example, depends on the
product of total resource acquisition, and reproductive allocation. If a species had very high
photosynthetic rates, then it could produce large quantities of fruits or seeds even with a low
level of reproductive allocation. Likewise, a species with high reproductive allocation, if it had
low overall capacity to accumulate resources through photosynthesis, would have low
reproductive capacity.
Evolutionary arguments can be made that perennial species should have generally lower
photosynthetic rates than closely related annuals (Grime, 1977; Garnier, 1992; Ehrelinger, 1994).
Since perennials do not need to complete their entire life cycle in one season, they can devote
resources to defence and competition rather than to rapid resource accumulation. Longer lifespan
38

increases the likelihood of predation, pathogen attack, etc. and thus devoting resources to
defence might be a larger priority. This could lead to lower photosynthetic rates, as more durable
leaves generally have lower specific leaf area, and as allocation to nitrogenous defence
compounds might mean less nitrogen allocation to photosynthetic proteins. Perennials might also
generally adopt a more conservative water use strategy, particularly if they must survive harsh
periods of cold or drought, or if high conductance rates would put them at high risk for
cavitation, and this might also select for the production of leaves with lower maximal
photosynthetic capacity (e.g. leaves with lower stomatal conductance).
However, although such theoretical arguments can be made, it is unclear to what extent
they are supported empirically. Alternative arguments could be made, that perennial species
might in some situations have higher photosynthetic rates than closely related annuals. At least
three separate explanations could lead us to hypothesize higher photosynthetic rates in a
perennial compared to a closely related annual species.
Greater sink strength in perennials. Perennial species must allocate resources to
vegetative storage sinks as well to sexual reproductive sinks, which may create
greater sink strength overall. Sink demand for carbohydrates is known to affect
photosynthetic rates through feedback inhibition: carbohydrate buildup in
chloroplasts inhibits photosynthesis through a variety of mechanisms. If
perennials have greater demand overall for carbohydrates, this could promote
compensatory changes leading to higher photosynthetic rates (Zhao et al., 2008).
Nitrogen tradeoffs. Annuals generally produce more seed than perennials, and
seeds are highly nitrogen-demanding tissues. Perennials, by contrast, must
optimize seed production over their life cycle, not within a single season. By
39

reducing allocation to seeds in one year, they could increase allocation of nitrogen
to key photosynthetic proteins in leaves. This might increase the plants’ ability to
photosynthesize, which could allow them to produce more seed in later years. In
support of this claim, perennial Lesquerella mendocina is known to have higher
leaf N content than its annual relative L. fendleri (Ploschuk et al., 2005).
Moisture stress. If perennial species allocate more resources to root growth than
closely related annuals, or to feeding associated mycorrhizae, they could
accumulate more nitrogen, phosphorus, potassium and micronutrients, as well as
maintaining higher water uptake rates, than the annuals. This could allow them to
escape moisture stress under dry conditions, and thus to maintain a greater degree
of stomatal opening.
Any one of these non-exclusive hypotheses might lead to higher photosynthetic rates in perennial
cereals, and at this point little work has been done on how photosynthesis and resource
acquisition in perennial cereals compares to annual relatives. It is known that some wild Oryza
spp. (e.g. Oryza rufipogon) maintain photosynthetic rates 25-65% higher than commonly grown
rice cultivars (Zhao et al., 2008, 2010). Similarly, higher photosynthetic rates have been found in
perennial vs. annual Lesquerella (Gonzalez-Paleo and Ravetta, 2011), as well as in a number of
other perennial taxa compared to annual relatives (e.g. Taylor et al., 2010). Other experiments
with different genera or families have found higher rates in annual species compared to closely
related perennials (Ehrelinger, 1994). In order to understand fully the capacity for resource
acquisition in perennial cereals like perennial wheat, perennial rye and intermediate wheatgrass,
it is necessary to empirically compare photosynthetic rates, as well as to investigate related

40

parameters that might indicate whether differences in water stress, sink strength or nitrogen
allocation account for any observed species differences in photosynthetic rates.
Curves modeling photosynthetic rate (A) as a function of internal carbon dioxide
concentration (Ci) are a highly useful way to infer various parameters that effect rates of resource
acquisition (Sharkey et al., 2007; Wullschleger, 1993). In Chapter 3 of this dissertation, we shall
discuss how the A vs. Ci curves may be used to infer stomatal limitation, carboxylation rates,
electron transport and triose phosphate utilization. We will also describe the methods and results
of a three year physiological study comparing one- and two-year old perennial cereals to close
annual relatives.

Source vs. Sink Limitation of Metabolism
Even if photosynthetic rates are in general similar between perennial cereals and their
annual relatives, it is possible that resource flow in both annuals and perennials might be sink
limited, such that resource flow (in this case, carbohydrate flow) is controlled by the ability to
utilize carbohydrates rather than the ability to produce them through photosynthesis. While sink
limitation can refer to other resources, in this section we will restrict ourselves to discussing sink
limitation as it pertains to carbon metabolism. Sink limitation occurs when inadequate demand
from ‘sinks’ (carbohydrate demanding tissues such as roots, meristems, growing leaf buds,
flowers, seeds) leads to buildup of carbohydrates in leaf tissues (‘sources’) and downregulation
of photosynthesis. If perennial cereals are source-limited (i.e. limited by the capacity of
‘sources’, green leaves, to photosynthesize) then photosynthesis under current environmental
conditions in the field can be expected to be near its maximum, and increased sink demand (i.e.
by selecting for greater seed sinks or more allocation to perenniating storage organs) would not
41

have an effect on photosynthesis and metabolism. Increased allocation to one sink would cause
decreased allocation to the other, and there would be close to a one-to-one tradeoff between units
of carbohydrate allocated to perenniating storage organs and to seeds.
However, if annual and/or perennial cereals show sink limitation, then they would be able
to upregulate or downregulate photosynthetic rates as the ratio of sink tissue to source tissue
increases or decreases. The presence of greater sink strength (as a result of more allocation to
perennial storage tissues) would promote higher photosynthesis and allow perennial cereal
species to accumulate more carbohydrate resources, overall, then annuals. If full sink limitation
was seen, perennial cereals would be able to compensate fully for increased allocation to
vegetative sinks, by upregulated photosynthesis. If source limitation was seen, by contrast,
perennial cereals would already be at their maximal photosynthetic rate, and would not
experience photosynthetic regulation. Thus, each additional unit of carbohydrate allocated to
perenniating storage organs would mean one less unit allocated to seed. If perennial cereals were
co-limited, finally, they would be able to compensate for some but not all of the increase in
sink/source ratio. Co-limitation might be indicated by, for example, any of the following:
1) A change in photosynthetic rate that was too small to fully compensate for the increased sink
demand, e.g. a 20% increase in photosynthetic rate following a 33% increase in sink/source
ratio (or alternatively a 25% decrease in source/sink ratio).
2) A change in seed size that was too small to fully account for the increased sink demand, e.g.
a decrease in seed size of 10% following a 25% increase in sink/source ratio (or alternatively
a 20% decrease in source/sink ratio).
3) Changes in photosynthetic rate following alteration of source/sink ratio in one direction but
not the other. If decreasing sink strength led to decreased photosynthetic rate, but increasing
42

sink strength did not lead to increased photosynthesis, then this would be evidence that
photosynthesis was co-limited, i.e. concurrently limited by source strength and sink strength.
Under such conditions, photosynthesis would be at the point which balanced source and sink
strength, and decreasing either factor would lead to lower photosynthetic rates.

Source limitation, sink limitation and co-limitation of plant carbon metabolism are, of
course, only meaningful terms within the context of a particular set of environmental conditions
and patterns of allocation. For example, we may consider a plant that is sink-limited at a
particular level of seed production and environmental conditions. If this plant increased its level
of reproductive allocation, or experienced certain kinds of environmental stresses that decreased
photosynthesis, then it would eventually enter a state of source limitation. Likewise, lowering
the level of reproductive allocation or growth, increasing light or CO2 levels, or introducing
certain other types of environmental stress (e.g. moderate cold, nitrogen deficiency) would lead
a plant that was previously source limited to become sink limited. Generally low temperature,
nutrient deficiencies, high CO2 and high light would all tend to push a plant further towards
sink limitation; water stress, shade stress, warm temperatures, and low CO2 would push a plant
more towards source limitation. Plants can shift between source and sink limitation on a day-today basis, as has been seen in cherry (Flore and Layne, 1999), or on a medium term basis from
one part of the growing season to another.
Thus to study source and sink limitation in a species, and assess the relative importance of
these two processes, it is necessary to specify the environmental conditions, level of
reproductive allocation, and the degree of source/sink perturbation involved. Ideally, the
43

responses to source/sink manipulation should be studied at multiple time points during the
growing season, as phenology and ambient weather conditions can both profoundly affect the
degree of source or sink limitation at a particular time.
There is widespread evidence of sink limitations to photosynthesis and metabolism, in a
range of plant taxa. These include many cultivated woody perennials including apple (Zhou &
Quebedeaux, 2003), cherry (Flore & Layne, 1999; Layne and Flore, 1993), peach (Gucci et al.,
1991), citrus (Iglesias et al., 2002), and coffee (DaMatta et al., 2008), as well in grasses including
wheat (Borras et al., 2004) and sugarcane (McCormick et al., 2006). Sink limitation can be
inferred from two different lines of evidence. Sink limitation would be evidenced photosynthetic
rates responding to changes in sink strength, or conversely by seed size remaining unchanged
under conditions of changed sink strength. Alternatively, sink limitation would be evident if
photosynthetic rates did not respond to changes in light or CO2 concentration. Baysdorfer and
Bassham (1986) found that alfalfa plants are able to recover from defoliation by upregulating
photosynthesis, and ambient CO2 concentrations are ample for this purpose, indicating sink
limitation. Phosphorus limitation may be involved in this case, as these plants were grown
without phosphorus, but have the capacity to fix nitrogen; phosphorus is known to limit alfalfa
growth on some soils. Joshi et al. (2003) carried out experimental manipulation of source size in
pearl millet, by removing increasing numbers of leaves on plants. While he was not explicitly
looking for evidence of source vs. sink limitation, his data does fit best to a saturating Monod
function, which indicates that as carbon availability increases, seed production in pearl millet
becomes limited by some aspect of sink strength (whether the limiting factor is nitrogen,
phosphorus, available flowering heads, or something else).
44

Sink limitation has been studied in the context of a variety of broader ecophysiology and
population ecology questions. As indicated above, many economically important plant species,
both woody fruit trees and annual cereals, have been subjected to source/sink manipulation in
order to determine the factors regulating metabolism, which has applied value for horticultural
and agronomic purposes. Source vs. sink regulation has also been studied in the context of
climate change (Roumet and Roy, 1996) as a means to predict how various plant species will
respond to elevated CO2. Finally, sink regulation of plant metabolism under conditions of low
temperature has been studied in the context of elevational limits to plant ranges, and researches
have found evidence of sink limitation at high altitudes in many trees.
Hoch and Korner (2005) studied the carbon balance in Polylepis trees in Argentina at the
upper altitudinal limit of their distribution, and found that nonstructural carbohydrate
concentrations actually increased by about 10% at the highest elevations, as did nitrogen
concentrations in leaves. In spite of the lower partial pressure of CO2, these trees were not
limited at the highest elevations by a lack of photosynthesis. Rather, they were more likely
limited by the inhibitory effect of cold temperature on meristematic growth (sink limitation). In
Argentina, Kageneckia showed similar evidence of sink limitation with higher TNC
concentrations at higher altitudes (Piper et al., 2006). Neither do photosynthesis and metabolism
in these plants appear to be limited by nutrients; Korner (1989) found that nitrogen
concentrations are higher in plants at the treeline while concentrations of phosphate, K, Mg, N,
and Ca are not significantly different. This would tend to indicate that rather being limited by
either carbon availability or nutrient availability, resource flow in these plants is limited by
inadequate sink strength. The mechanism would be that cold inhibits growth and carbon
utilization processes more than it inhibits photosynthesis. Some support for this hypothesis
45

comes from the fact that in Ajuga reptans, cold treatment increases the concentration of total
nonstructural carbohydrates by a factor of 10. The most common soluble carbohydrate class, the
raffinose family oligosaccharides, are present in highest concentrations in the fall and winter and
lowest concentrations in the summer (Bachmann et al., 1994).
One possible mechanisms for stimulation of photosynthesis under high sink/source ratio
could involve a decrease in total nonstructural carbohydrate including triose phosphate, which
could stimulate increased carboxylation, A meta-analysis found that 23 of 109 plant species
showed triose phosphate utilization to be the primary limitation of photosynthesis under normal
conditions (Wullschleger, 1993). If this is true then increased sink demand should lead to a
drawdown of TNC and thus a reduction of feedback inhibition of the carbon reactions of
photosynthesis. Thus, a study on grapefruit trees (Citrus paradisi) found that shaded leaves had
higher asymptotes of the A vs. Ci curve, potentially because of greater sink strength (Jifon and
Syvertsen, 2003). Alternatively, sink vs. source regulation could proceed through hormonal
mechanisms that result in the downregulation of Rubisco activity or electron transport (Krapp
and Stitt, 1995).
Little effort, though, has been devoted to comparing the degree of source vs. sink
limitation in perennial species compared to close annual relatives. Perennials might tend, in
general, to be more source-limited than annuals, if they are characterized by much larger
vegetative sinks and lower photosynthetic capacity. The combination of these two traits would
set them closer to their maximal photosynthetic capacity under normal conditions, and they
would have less capacity to compensate for increases in sink strength (Arp, 1991). Conversely,
the fact that perennials generally allocate less to sexual reproductive sinks than related annual
species (i.e. seeds, flowers, etc.), particularly if they maintain high photosynthetic rates, could
46

mean that they generally maintain a more conservative strategy for resource allocation, and
should be more sink-limited (i.e. have greater capacity for compensatory changes in
photosynthesis and carbohydrate storage) than annuals (Masnatta and Ravetta, 2011; Gaines et
al., 1974).
There are evolutionary reasons why perennials might be expected to maintain a more
conservative reproductive strategy. If an annual plant ends the season with excess energy or
nutrient reserves, those reserves are ‘wasted’ since the plant will die; an annual plant should
therefore be selected to put the maximum amount of resources into reproduction, so as to
optimize short-term productivity. A perennial plant, however, must optimize reproduction over
the course of multiple years. In such cases, a conservative reproductive strategy could pay off: it
would buffer the plant against environmental stresses during the reproductive period (by
maintaining extra reserves) and if unused by the end of the season, those reserves could help
support the plant’s growth next year. However, as yet few studies have explicitly compared the
general degree of source vs. sink limitation in annuals compared to closely related perennials.
One of the few such studies is an experimental study by Roumet and Roy (1996) in which
source vs. sink limitation was assessed in a variety of cultivated perennial, wild perennial and
wild annual grasses (through measuring responses to elevated CO2). The study found that wild
annuals were more source-limited than wild perennials, but less so than cultivated perennials,
suggesting that both selection for high yield and annual vs. perennial life history can profoundly
affect the degree of source vs. sink limitation. As perennial cereals are characterized both by
perennial life history and by selection for high seed production, it is unclear to what extent they
should be expected to show source vs. sink limitation. In addition, no studies have yet, to the best
of our knowledge, quantified photosynthetic responses to source vs. sink manipulations in annual
47

rye, in spite of its agricultural importance. Thus, quantifying source vs. sink limitation in
perennial cereals as well as in annual rye is important to predicting the extent to which perennial
cereals may be able to support higher yields, as well as to understanding how photosynthetic
physiology and plant life history interact.
Various methods exist to manipulate source / sink ratios in plants. Briefly, they include
the following. It would be possible to carry out selection experiments, selecting for higher or
lower sink strength (e.g. rhizome size, floret number, seed number) or for higher or lower source
strength (e.g. leaf mass ratio). It would also be possible to carry out direct experimental
manipulations of the source/sink ratio. This might use any of the following techniques.
Source augmentation could be carried out by supplemental lighting (Flore and Layne,
1999), increased concentrations of carbon dioxide (Hunt, 1996; Ainsworth et al., 2003; Reich et
al., 2006, etc.), or direct feeding of soluble sugars to be absorbed through roots or leaves
(Hiyane, 2010). One commonly used method involves increasing available light to plants.
Increasing light intensity could easily lead to photoinhibition (and potentially to photodamage of
shade-adapted plants); thus a method sometimes used to increase source strength is to increase
daylength in greenhouse plants. In sour cherry (Prunus cerasus) continuous lighting led to a 40%
decrease in photosynthesis after one week, while 70% defoliation led to a 50% increase, and the
combination of both manipulations resulted in an increase comparable to defoliation alone. It
would seem that an augmentation of source size does lead to a decline in source activity, but only
if there is not a corresponding increase in sink size (Layne and Flore, 1993). Sawada et al. (1989)
found similar results in soybeans, where after 1 week at continuous lighting (compared to 10
hours of light under normal conditions) photosynthesis was 54% lower. This decline is
presumably attributable to the buildup of soluble carbohydrates and subsequent inhibition.
48

Similar results were found in Amaranthus cruentus; here there was a 50% decline in
photosynthesis after 12 days under continuous light, with a 40% increase in sucrose levels and a
200% increase in starch, thus providing a mechanism for the decline in photosynthesis (Sawada
et al., 1999). An additional method to increase source size is by physically spraying soluble
sugars on leaves (Hiyane et al., 2010).
One of the most common methods of increasing source/sink balance, and most important
in light of ongoing climate change, is carbon enrichment. Plants can be grown in the field under
a free-air carbon enrichment (FACE) system, or in greenhouse growth chambers; if temperature,
nitrogen, phosphorus, and other factors are at appropriate levels, then increased levels of carbon
dioxide in the air can allow plants to fix more carbohydrate. For example, McKee and Woodward
(1994) found strong evidence of sink limitation in dwarf spring wheat grown in growth
o

chambers. At high (24 / 16 C day/night) temperatures, a doubling of atmospheric CO2 levels
produced only a 16% in biomass, and biomass actually decreased under elevated CO2 at colder
temperature. The authors attributed this to sink activity being depressed more than source
activity, a common observation at low temperatures (e.g. Hoch and Korner, 2005). However,
Uprety et al. (2009) found contrasting results when wheat was grown in the field under a FACE
system in India. They found that spring wheat showed a 46% increase in biomass and a 22%
o

increase in yield when grown under 550 ppm (under hot conditions ranging from 16/6 to
o

37/20 day/night); this was much greater than the increase in diploid or tetraploid wheats (T.
monococcum or T. durum) and led them to conclude that wheat is still highly source-limited. In
Arizona, FACE experiments found that wheat plants showed a 28% increase in leaf
photosynthetic rate (and a 50% increase under drought conditions), but also found earlier
49

senescence and little effect on final grain yield, suggesting sink limitation under well watered
conditions (Pinter et al., 1996).
Some studies have also been done of perennial C3 grasses, as well as perennial forbs, and
their response to carbon dioxide elevation; since perennial wheat and perennial rye have a multiyear life history, these grasses could be more relevant models. A study of Lolium, Festuca and
Agrostis in greenhouse pots found 20% increases in dry mass of all three species when carbon
dioxide was raised to 700 ppm (Cotrufo and Gorissen, 1997). However, no effect of CO2 was
found on growth of Poa pratensis, a cool season grass, in an open top chamber experiment (Owensby et al., 1993). A study of Pascopyrum smithii (formerly Agropyron, a relative of intermediate wheatgrass), using pots in a greenhouse, found that photosynthetic rates increased under 700
ppm CO2, and final biomass increased by 19% (Hunt et al., 1996). Similar experiments with
Pascopyrum in open-top chambers found that such increases in photosynthesis were driven primarily by increased photosynthetic rates during drought (LeCain et al., 2003). Baxter et al.
(1994) found that two montane C3 grasses, Poa alpina and Agrostis capillaris, showed increased
biomass when grown under elevated CO2 in open-top chambers, but a Festuca species showed a
decline.
However, surprisingly, leaf-level differences in photosynthesis in response to elevated
CO2 were not found in the grassland studies either in Switzerland or Minnesota (Reich et al.,
2006). While photosynthetic rates are usually higher at elevated CO2, the A / Ci curve is also
shifted left (i.e. the carbon saturated photosynthetic rate is lower), indicating acclimation to elevated CO2 and subsequent downregulation of photosynthesis. This effect is stronger at low lev50

els of nitrogen, illustrating that nitrogen limitation can contribute to carbon sink limitation and
that determining A / Ci curves can be a good way to test for sink limitation (Isopp et al., 2000;
Ainsworth and Long, 2005).

Source reduction might involve clipping leaves (Handa et al., 2005; Macedo et al., 2007;
Horibata et al., 2007), partially shading leaves (Jifon and Syvertsen, 2003; Araus et al., 1993;
Peet and Kramer, 1980; Rosas et al., 1976; Ralphs et al., 1998), or painting leaves with opaque
coverings (Niva et al, 2003; Karlsson, 1994). Defoliation is perhaps the most common method of
source reduction. A major drawback of this technique, however, is that it could lead to hormonal
changes in the plant that could confound the effects of the source size reduction itself. For
example, defoliation could promote regrowth, formation of defence compounds, changed
allocation to seed, production of volatile chemicals, changes in water status and so forth. An
alternative method is to wrap shade cloth sheaths around the selected leaves. McCormick et al.
(2006) used this technique in sugarcane. It has the advantage that shade cloth sheaths would
allow some ventilation, which would reduce the risk of heat and humidity being trapped.
Some researchers have used aluminum foil to shade individual leaves. Aluminum foil
has been used in poplar (Davis et al., 1991), cassava (Rosas et al., 1976) and larkspur (Ralphs et
al., 1998) among many others. Peet and Kramer (1980) used an adhesive-backed aluminized
mylar film applied to the upper surfaces of leaves to shade them. The basic problem with
aluminum foil coverings, besides that they can fall off, is that they could trap heat and alter the
metabolism of the plant through heat-mediated stress. Though aluminum foil is known to reflect
95% of incident light, the remaining 5%, if converted to heat energy, could do substantial
damage to a plant.
51

A further alternative is painting the leaves to be “shaded” with a nontoxic white paint
coating; this was done in Rhododendron lapponicum (Karlsson, 1994), and in Linnea borealis
(Niva et al., 2003), both times using Tipp-Ex correction fluid. The advantage is that white paint
is opaque but also highly reflective: it would absorb little radiation, and the leaf would not heat
up too much. It is necessary that the paint involved be non-toxic and hopefully also water-soluble
so that it does not build up in the soil and is not absorbed into the plant. Tipp-Ex matches these
conditions. An advantage of painting leaves, as opposed to designing structures to cover them, is
that it is easier to brush small leaves with paint than to delicately design shading treatments to fit
them. If they are too large, aluminum foil or shade cloth sheaths could slip off and it would be
hard to make them small enough. We conclude that using white paint on individual leaves would
seem to be an appropriate method for our perennial grain plants.
Paint treatments, of course, must be applied only to the top surfaces of leaves, so as not to
interfere with evaporation or with whatever residual photosynthesis can occur even after the
opaque coating (most likely this will be very small). In the dicots that Karlsson (1994) and Niva
et al. (2003) studied, this worked well since in most dicots, water loss and light absorbption are
somewhat spatially separated: most dicot stomata are on the underside (abaxial surface) of
leaves. Wheat and wheatgrass are monocots, and thus are amphistomatic, having stomata on both
surfaces: in fact, wheat stomata are more common (15-40%) on the upper surface (adaxial) of the
flag leaves (Yousufzai et al., 2009), and also in fourth (5-20%) and first node (25%) in first node
leaves (Araus et al., 1989). Thus unlike with Rhododendron, this method when used in grasses
cannot physically perfectly separate the effects of painting on evapotranspiration and on
photosynthesis. The situation is also a bit difficult because wheat and grass leaves often are
relatively upright and exposed to light from all around. However, since light is absorbed mostly
52

through the upper surface, and water loss presumably happens roughly equally in proportion to
the number of stomata on the adaxial and abaxial surfaces, painting the upper surfaces should
allow the leaves to transpire and shed heat while also not absorbing much light.
Sink augmentation can be carried out by altering the photoperiod (Reekie and Bazzaz,
1987a), by increasing hand pollination (Snow and Whigham, 1989; Primack and Hall, 1990) or
by extending the flowering period (Laporte and Delph, 2006). Artificially augmented pollination
can lead to an increased sink size. For example, hand pollination of pink lady’s slipper orchid led
to greatly increased seed set, but decreased leaf area (by about 10-13%) and seed set in the
following year; these effects were compounded over four years in a multi-year experiment
(Primack and Hall, 1990), illustrating a large cost of reproduction. A study in willow,
contrariwise, covered calyces to prevent pollination, and studied whether there was an increase
in next year’s growth and seed set (Fox and Stevens, 1991). Hand pollination has also been used
in sticky catchfly (Jennersten, 1991) and in Tipularia orchids (Snow and Whigham, 1989); the
last mentioned, for example, found that each fruit cost the plant about 2% of its mass. However,
pollination manipulation will only allow cost of reproduction or compensatory responses to be
estimated if the species under discussion is actually pollen limited (Dudash and Fenster, 1997).
Shading reproductive structures is specifically a way to increase sink size in plant species
where there is considerable ‘reproductive photosynthesis”, i.e. photosynthesis by flowers and
reproductive support structures. Some studies have carried out a combination of these
manipulations. For example, Gehring and Delph (2006) shaded calyces of Silene latifolia and
delayed pollination (leading to initiation of more floral buds), both of which effectively increased
sink size. This led to more floral abortion and smaller fruit mass, indication that reproductive

53

output was controlled by carbohydrate availability and not the other way around (i.,e. Silene is
source limited).
Sink reduction involves decreasing the size of reproductive sinks. The most common
way of reducing sink size is simply to clip and remove flowering heads or fruits. For example,
Jackson and DeWald (1994) removed seed stalks to reduce reproductive sinks in Tripsacum
dactyloides, and found a small cost of reproduction (although seed stalks were partially self
sufficient and could even in extreme circumstances serve as a carbohydrate source). Similar
manipulations have been done in Viscaria (Jennersten, 1991) among other species. Alternative
methods of altering sink size include looking at female vs. male or female vs. hermaphroditic
plants (in dioecious or gynodioecious species respectively). Generally, female plants are
expected to have higher reproductive allocation. Still another is to alter flowering through
altering the photoperiod of the plants (Reekie and Bazzaz, 1987a). This has risks, however, of
hormonal alterations on plant metabolism caused by the photoperiod, and is in any case
inappropriate for a field experiment where photoperiod cannot be easily manipulated.
Tracing experiments to follow the fate of carbohydrates in the plant can offer
supplementary information about the underlying basis of source vs. sink limitation. For example,
C

13

tracing experiments in Adonis ramosa, a spring ephemeral, found that current photosynthate

fixed by leaves during the flowering period went mostly to belowground storage, and very little
went to the fruits: the fruits were to a large extent supported by their own photosynthesis and by
long-term stored carbohydrate from the roots. This helped explain why defoliation did not lead to
a decrease in fruit production (Horibata et al., 2007). Conversely, in apples (Hansen, 1967) about
90% of photosynthate produced by leaves is transferred to leaves nearby.

54

In Chapter 4 of this dissertation, we shall describe the results of altering source/sink ratios
through source reduction and sink reduction.

Conclusion
Perennial cereals may have the potential to satisfy human needs for food, fuel, and fiber
in more environmentally sustainable ways, and provide improved levels of ecosystem services
within agroecosystems and their surroundings. However, several challenges must be overcome in
order for perennial cereals to become economically feasible to be produced on a significant
scale. In the remainder of this dissertation, we will consider various aspects of perennial cereal
production from the point of view of agronomy, physiology and disease resistance. The studies
detailed in Chapters 2-5 will answer questions about the grain and biomass productivity of
perennial cereals over time, the ability of perennial cereals to withstand an important North
American cereal disease, the physiological differences in resource acquisition rates between
perennial and annual cereals, and the extent to which resource flow in perennial cereals is source
vs. sink limited. The goal of this literature review and the following experimental studies are to
contribute to the growing literature on perennial cereals, and to address both basic and applied
questions regarding a new and highly unusual plant growth form.

55

REFERENCES

56

REFERENCES

Acharya SN, Mir Z, Moyer JR, Orshinsky BR, Thomas JE. 2003. Effect of row spacing and
seeding date on forage yield and quality of perennial cereal rye (Secale cereale L.). Canadian
Journal of Plant Science 83: 363-369.
Acharya SN, Mir Z, Moyer JR. 2004. ACE-1 perennial cereal rye. Canadian Journal of Plant
Science 84: 819-821.
Ainsworth EA, Long SP. 2005. What have we learned from 15 years of free-air CO2 enrichment
(FACE)? A meta-analytic review of the responses of photosynthesis, canopy properties, and plant
production to rising CO2. New Phytologist 165(2): 351-371.
Akhtar J, Gorham J, Qureishi RH. Combined effect of salinity and hypoxia in wheat (Triticum
aestivum L.) and wheat-Thinopyrum amphiploids. Plant and Soil 166: 47-54.
Anamthawat-Jonsson K. 1996. Wide-hybrids between wheat and Lymegrass : breeding and
agricultural potential. Buvusindi 10:101–113.
Anamthawat-Jonsson K, Bodvarsdottir SK, Bragason BT, Gudmundsson J, Koebner RMD.
1997. Wide hybridization between wheat (Triticum L.) and lymegrass (Leymus Hochst.). Euphytica 93:293–300
Aragón CF, Méndez M, Escudero A. 2009. Survival costs of reproduction in a perennial plant:
live hard, die young. American Journal of Botany 96: 904-911.
Arp WJ. 1991. Effects of source-sink relations on photosynthetic acclimation to elevated CO2.
Plant, Cell and Environment 14: 869-875.
Araus JL, Tapia L, Alegre L. 1989. The effect of changing sowing date on leaf structure and gas
exchange characteristics of wheat flag leaves grown under Mediterranean climate conditions.
Journal of Experimental Botany 40: 639-646.
Araus JL, Brown HR, Febrero A, Bort J, Serret MD. 1993. Ear photosynthesis, carbon isotope
discrimination and the contribution of respiratory CO2 to differences in grain mass in durum
wheat. Plant, Cell and Environment 16(4): 383-392.
Arredondo JT, Jones TA, Johnson DA. Seedling growth of intermountain perennial and weedy
annual grasses. Journal of Range Management 51(5): 584-589.
Aschan G, Pfanz H. 2003. Non-foliar photosynthesis- a strategy of additional carbon acquisition.
Flora 198: 81-97.
57

Aschan G, Pfanz H, Vodnik D, Batic F. 2005. Photosynthetic contribution of vegetative and
reproductive structures of green hellebore (Helleborus viridis L. agg.) Photosynthetica 43(1): 5564.
Bachmann M, Matile P, Keller F. 1994. Metabolism of the raffinose family oligosaccharides in
leaves of Ajuga reptans L. (Cold acclimation, translocation, and sink to source transition:
discovery of chain elongation enzyme.) Plant Physiology 105: 1335-1345.
Banks PM, Xu SJ, Wang RRC, Larkin PJ. 1993. Varying chromosome composition of 56chromosome wheat x Thinopyrum partial amphiploids. Genome 36: 207-215.
Bausenwein U, Millard P, Raven JA. 2001. Remobilized old-leaf nitrogen predominates for
spring growth in two temperate grasses. New Phytologist 152(92): 283-290.
Baxter R, Ashenden TW, Sparks TH, Farrar JF. 1995. Effects of elevated carbon dioxide on three
montane grass species: I. Growth and dry matter partitioning. Journal of Experimental Botany
45:305-315.
Baysdorfer C, Bassham JA. 1986. Photosynthate supply and utilization in alfalfa: a
developmental shift from a source to a sink limitation of photosynthesis. Plant Physiology 77:
313-317.
Bazzaz, F.A., Ackerly, D.D., and Reekie, E.G. 2000. Reproductive allocation in plants. In:
Fenner, M. Seeds: the ecology of regeneration in plant communities. Wallingford, UK: CABI
Publishing, 2000. pp. 1-30.
Bazzaz FA, Carlson RW. 1979. Photosynthetic contribution of flowers and seeds to reproductive
efforts of an annual colonizer. New Phytologist 82(1): 223-232.
Bazzaz FA, Carlson RW, Harper JL. 1979. Contribution to reproductive effort by photosynthesis
of flowers and fruits. Nature 279: 554-555
Bell LW, Byrne F, Ewing MA, Wade LJ. 2008. A preliminary whole-farm economic analysis of
perennial wheat in an Australian dryland farming system. Agricultural Systems 96: 166-174.
Berdahl JD, Redfearn JD. 2007. Grasses for arid and semiarid areas. In Barnes, R.F., Nelson,
C.J., Moore, K.J., and Collins, M. eds. Forages: The science of rangeland agriculture.
Blackwell, 2007.
Bergstrom L. 1987. Nitrate leaching and drainage from annual and perennial crops in tile drained
plots and lysimeters. Journal of Environmental Quality 16:11-18.
Blandino M, Reyneri A, Scudellari D. 2009. Crop residues and rain bring together risk of
Fusarium disease. Informatore Agrario 65: 44-46.
Bommarco R. 1999. Feeding, reproduction and community impact of a predatory carabid in two
agricultural habitats. Oikos 87(1): 89-96
58

Borras L, Slafer GA, Otegui ME. 2004. Seed dry weight response to source-sink manipulations
in wheat, maize and soybean: a quantitative reappraisal. Field Crops Research 86: 131-146.
Cai X, Jones SS, Murray TD. 1998. Molecular cytogenetic characterization of Thinopyrum and
wheat-Thinopyrum translocated chromosomes in a wheat-Thinopyrum amphiploid. Chromosome
Research 6: 183-189.
Cai X, Jones SS, Murray TD, Weber WE. 2001. Molecular cytogenetic characterization of
Thinopyrum genomes conferring perennial growth in wheat-Thinopyrum amphiploids. Plant
Breeding 1: 20-26.
Clark MS, Gage SH, Spence JR. 1997. Habitats and management associated with common
ground beetles (Coleoptera: Carabidae) in a Michigan agricultural landscape. Environmental
Entomology 26: 519-527.
Clement C, Mischler P, Burrus M, Audran JC. 1997. Characteristics of the photosynthetic
apparatus and CO2 fixation in the flower bud of Lilium.II. Anther. International Journal of Plant
Sciences 158(6): 801-810.
Colmer TD, Flowers TJ, Munns, R. 2006. Use of wild relatives to improve salt tolerance in
wheat. Journal of Experimental Botany 57: 1059-1078.
Corley RHV. 1983. Oil palm and other tropical tree crops. In Smith WH, Banta SJ., eds.
Symposium on Potential Productivity of Field Crops under Different Environments. pp. 383-399.
International Rice Research Institute, Los Baños, Philippines.
Cotrufo MF, Ineson P, Scott A. 2002. Elevated CO2 reduces the nitrogen concentration of plant
tissues. Global Change Biology 4(1): 43-54.
Cotrufo MF, Gorissen A. 1997. Elevated CO2 enhances below ground C allocation in three
perennial grass species at different levels of N availability. New Phytologist 137(3): 421-431.
Cox TS, Bender M, Picone C, Van Tassel DL, Holland JB, Brummer EC, Zoeller BE, Paterson
AH, Jackson W. 2002. Breeding perennial grain crops. Critical Reviews in Plant Sciences 21(2):
59-91.
Cox TS, Glover JD, Van Tassel DL, Cox CM, DeHaan LR. 2006. Prospects for developing
perennial grain crops. BioScience 56(8): 649-659.
Cox, C. M., Garrett, K. A., Cox, T. S., Bockus, W. W., and Peters, T. 2005. Reactions of
perennial grain accessions to four major cereal pathogens of the Great Plains. Plant Disease
89:1235-1240.

59

Crews TE. 2005. Perennial crops and endogenous nutrient supplies. Renewable Agriculture and
Food Systems 20:25-37
Crews TE, Peoples MB. 2005. Can the synchrony of nitrogen supply and crop demand be
improved in legume- and fertilizer-based agroecosystems? A review. Nutrient Cycling in
Agroecosystems 72(20: 101-120.
DaMatta FM, Cunha RL, Antunes WC, Martins SCV, Araujo WL, Fernie AR, Moraes GABK.
2008. In field-grown coffee trees source-sink manipulation alters photosynthetic rates,
independently of carbon metabolism, via alterations in stomatal function. New Phytologist 178:
348-357.
Davis JM, Gordon MP, Smitt BA. 1991. Assimilate movement dictates remote sites of woundinduced gene expression in poplar leaves. Proceedings of the National Academy of Sciences,
USA. 88: 2393-2396
DeHaan LR, Van Tassel DL, Cox TS. 2005. Perennial grain crops: a synthesis of ecology and
plant breeding. Renewable Agriculture and Food Systems 20: 5-14.
De Vries AP. 1974. Some aspects of cross-pollination in wheat (Triticum aestivum L.). 4. Seed
set on male sterile plants as influenced by distance from the pollen source, pollinator: Male
sterile ratio and width of the male sterile strip. Euphytica 23:601–622.
Dowdell RJ, Webster CP. 1980. A lysimeter study using nitrogen-15 on the uptake of fertilizer
nitrogen by perennial ryegrass swards and losses by leaching. European Journal of Soil Science
31(1): 65-75.
Dudash MR, Fenster CB. 1997. Multi-year study of pollen limitation and cost of reproduction in
the iteroparous Silene virginica. Ecology 78(2): 484-493.
Dvořak J. 1976. The cytogenetic structure of a 56–chromosome derivative from a cross between
Triticum aestivum and Agropyron elongatum (2n = 70). Canadian Journal of Genetics and
Cytology 18:271–279.
Ehrlén, J., and van Groendendael, J. Storage and the delayed costs of reproduction in the
understorey perennial Lathyrus vernus. Journal of Ecology 89(2): 237-246.
Ehrelinger JR. 1994. Variation in gas exchange characteristics among desert plants. In: Schulze
EM, Caldwell MM. 1994. Ecophysiology of Photosynthesis. Berlin: Springer Verlag, 1994, pp.
361-392.
Entz MH, Bullied WJ, Forster DA, Gulden R, Vessey JK. 2001. Extraction of subsoil nitrogen
by alfalfa, alfalfa-wheat and perennial grass systems. Agronomy Journal 93: 495-503.
Evans RA, Eckert RE, Kay BL. 1967. Wheatgrass establishment with paraquat and tillage on
downy brome ranges. Weeds 159: 50-55.
60

Ewel JJ. 1999. Natural systems as models for the design of sustainable systems of land use.
Agroforestry Systems 45(1-3): 1-21.
Fielder AK, Landis DA. 2007. Attractiveness of Michigan native plants to arthropod natural
enemies and herbivores. Environmental Entomology 36(4): 751-765.
Flore JA, Layne DR. 1999. Photoassimilate production and distribution in cherry. HortScience
H34: 1015-1019.
Fox JF, Stevens GC. 1991. Costs of reproduction in a willow: Experimental responses vs. natural
variation. Ecology 72(3): 1013-1023.
Füle L, Hódos-Kotvics G, Galli Z, Ács E, Heszky L. 2005. Grain quality and baking qualities of
perennial rye (cv. ‘Perenne’) of interspecific origin (Secale cereale x montanum). Cereal
Research Communications 33: 809-814.
Gaines MS, Vogt KJ, Hamrick JL, Caldwell J. 1974. Reproductive strategies and growth-patterns
in sunflowers (Helianthus). American Naturalist 108: 889-894.
Galen C, Dawson TE, Stanton ML. 1993. Carpels as leaves: meeting the carbon cost of
reproduction in an alpine buttercup. Oecologia 95(2): 187-193.
Garnier E. 1992. Growth analysis of congeneric annual and perennial grass species. Journal of
Ecology 80: 665-675.
Gehring JL, Delph LF. 2006. Effects of reduced source-sink ratio on the cost of reproduction in
females of Silene latifolia. International Journal of Plant Sciences 167(4): 843-851.
Gloser V. 2005. The consequences of lower nitrogen availability in autumn for internal nitrogen
reserves and spring growth of Calamagrostis epigejos. Plant Ecology 179(1): 119-126.
González-Paleo L, Ravetta D. 2011. Relationships between reproductive output, morphophysiological traits, and lifespan in Lesquerella (Brassicaceae). Industrial Crops and Products
34: 1372-1380.
Grime JP. 1977. Evidence for the existence of three primary strategies in plants and its relevance
to ecological and evolutionary theory. American Naturalist 111: 1169-1194.
Gucci R, Xiloyannis C, Flore JA. 1991. Gas exchange parameters, water relations and
carbohydrate partitioning in leaves of field-grown Prunus domestica following fruit removal.
Physiologia Plantarum 83(3): 497-505.
Gutman M, Noy-Meir I, Pluda D, Seligman NA, Rothman S, Sternberg M. 2001. Biomass
partitioning following defoliation of annual and perennial Mediterranean grasses. Conservation
Ecology 5(2): 1. [online] URL: http://www.consecol.org/vol5/iss2/art1/
61

Gusewell S. 2005. High nitrogen:phosphorus ratios reduce nutrient retention and second-year
growth of wetland sedges. New Phytologist 166(2): 537-550.
Han FP, Fedak G, Benabdelmouna A, Armstrong K, Ouellet T. 2003. Characterization of six
wheat x Thinopyrum intermedium derivatives by GISH, RFLP, and multicolor GISH. Genome
46: 490-495.
Handa IT, Körner C, Hattenschwiler, S. 2005. A test of the treeline carbon limitation hypothesis
by in situ CO2 enrichment and defoliation. Ecology 86(5): 1288-1300.
Hansen P. 1967. Studies on apple trees: I. The effect of the fruit on the translocation and
distribution of photosynthates. Physiologia Plantarum 20(2): 382-391.
Hayes RC, Newell MT, DeHaan LR, Murphy KM, Crane S, Norton MR, Wade LJ, Newberry M,
Fahim M, Jones SS, Cox TS, Larkin PJ. 2012. Perennial cereal crops: An initial assessment of
wheat derivatives. Field Crops Research 133: 68-89.
Haynes RJ, Francis GS. 1993. Changes in microbial biomass C, soil carbohydrate composition
and aggregate stability induced by growth of selected crop and forage species under field
conditions. European Journal of Soil Science 44(4): 665-675.
Heckathorn SA, DeLucia EH. Drought-induced nitrogen retranslocation in perennial C4 grasses
of tallgrass prairie. Ecology 75(7): 1877-1886.
Hiyane R, Hiyane S, Tang AC, Boyer JS. 2010. Sucrose feeding reverses shade-induced kernel
losses in maize. Annals of Botany 106: 395-403.
Hoch G. 2005. Fruit bearing branchlets are carbon autonomous in mature broad leaved temperate
forest trees. Plant, Cell and Environment 28(5): 651-659.
Hoch G, Körner C. 2005. Growth, demography and carbon relations of Polylepis trees at the
world’s highest treeline. Functional Ecology 19(6): 941-951.
Horibata S, Hasegawa SF, Kudo G. 2007. Cost of reproduction in a spring ephemeral species,
Adonis ramosa (Ranunculaceae): Carbon budget for seed production. Annals of Botany 100: 565571.
Hunt HW, Elliott ET, Detling JK, Morgan JA, Chen DX. 1996. Response of a C3 and a C4 grass
to elevated CO2 and temperature under different water regimes. Global Change Biology 2(1):
35-47.
Iglesias DJ, Lliso I, Tadeo FR, Talon M. 2002. Regulation of photosynthesis through source: sink
imbalance in citrus is mediated by carbohydrate content in leaves. Physiologia Plantarum 116:
563-572.
62

Islam KR, Weil RR. 2000. Land use effects on soil quality in a tropical forest ecosystem of
Bangladesh. Agriculture, Ecosystems and Environment 79(1): 9-16.
Isopp H, Frehner M, Long SP, Nosberger J. 2000. Sucrose-phosphate synthase responds
differently to source-sink relations and to photosynthetic rates: Lolium perenne L . growing at
elevated pCO2 in the field. Plant, Cell, and Environment 23: 597-607.
Jackson L, DeWald C. 1994. Predicting evolutionary consequences of greater
reproductive effort in Tripsacum dactyloides, a perennial grass. Ecology 75(3): 627-641.
Jauhar PP. 1992. Synthesis and cytological characterization of trigeneric hybrids involving
durum wheat, Thinopyrum bessarabicum, and Lophopyrum elongatum. Theoretical and Applied
Genetics 84:511–519.
Jauhar PP. 1995. Meiosis and fertility of F1 hybrids between hexaploid annual wheat and
decaploid tall wheatgrass (Thinopyrum ponticum). Theoretical and Applied Genetics 90: 865871.
Jennersten O. 1991. Cost of reproduction in Viscaria vulgaris (Caryophyllaceae): A field
experiment. Oikos 61(2): 197-204.
Jensen KB, Zhang YF, Dewey DR. 1990. Mode of pollination of perennial species of the
Triticeae in relation to genomically defined genera. Canadian Journal of Plant Science 70: 215225.
Jifon JL, Syvertsen JP. 2003. Moderate shade can increase net gas exchange and reduce
photoinhibition in citrus leaves. Tree Physiology 23: 119-127.
Joffre R. 1990. Plants and soil nitrogen dynamics in Mediterranean grasslands: a comparison of
annual and perennial grasses. Oecologia 85(1): 142-149.
Jones TA, Zhang XY, Wang RRC. 1991. Genome characterization of MT-2 and OK-906 annual
wheat x intermediate wheatgrass hybrids. Crop Science 39: 1041-1043.
Joshi AK, Pandya JN, Buhecha KV, Dave HR, Pethani KV, Dangaria CJ. 2003. Grain yield in
pearl millet in relation to source size and proximity to sink. Photosynthetica 41(1): 157-159.
Karlsson PS. 1994. The significance of internal nutrient cycling in branches for growth and
reproduction of Rhododendron lapponicum. Oikos 70(2): 191-200.
Kemanian AR, Stöckle CO, Huggins DR, Viega LM. 2007. A simple method to estimate harvest
index in grain crops. Field Crops Research 103: 208-216.
Knowles RP. 1977. Recurrent mass selection for improved yields in intermediate wheatgrass.
Crop Science 17: 51-54.
63

Körner C. 1989. The nutritional status of plants from high altitudes. Oecologia 81(3): 379-391.
Krapp A, Stitt M. 1995. An evaluation of direct and indirect mechanisms for the "sink
regulation" of photosynthesis in spinach: Changes in gas exchange, carbohydrates, metabolites,
enzyme activities and steady state transcript levels after cold girdling source leaves. Planta 195:
315-323.
Kubiczek R, Łuczak W, Molski B. 1981. Protein resources of wild Secale species. Kulturpflanze
29: 159-167.
Laporte MM, Delph LF. 1996. Sex-specific physiology and source-sink relations in the dioecious
plant Silene latifolia. Oecologia 106(1): 63-72.
Landes, M. 1939. The causes of self-sterility in rye. American Journal of Botany 26(8): 567-571.
Landis DA, Wratten SD, Gurr GM. 2000. Habitat management to conserve natural enemies of
arthropod pests in agriculture. Annual Review of Entomology 45: 175-201.
Law R, Bradshaw AD, Putwain PD. 1977. Life history variation in Poa annua. Evolution 31:
233-246.
Layne DR, Flore JA. 1993. Physiological responses of Prunus cerasus to whole plant source
manipulation. Leaf gas exchange, chlorophyll fluorescence, water relations and carbohydrate
concentrations. Physiologia Plantarum 88: 44-51.
LeCain DR, Morgan JA, Mosier AR, Nelson JA. 2003. Soil and plant water relations determine
photosynthetic responses of C3 and C4 grasses in a semi-arid ecosystem under elevated CO2.
Annals of Botany 92:41-52.
Lammer D, Cai XW, Arterburn M, Chatelain J, Murray S, Jones S. 2004. A single chromosome
addition from Thinopyrum elongatum confers a polycarpic, perennial habit to annual wheat.
Journal of Experimental Botany 55: 1715-1720.
Li MJ, Wu SY. 2006. Assessment to adult plant resistance to stripe rust of for quality
improvement of wheat germplasms parent materials in Yunnan. Journal of Triticeae Crops 26(1):
113- 116 (in Chinese with an English abstract).
Li YS, Redmann RE, van Kessel C. 1992. Nitrogen budget and
wheatgrass. Functional Ecology 6(2): 221-225.

15

N translocation in a perennial

Li Z, Li B, Tong Y. The contribution of distant hybridization with decaploid Agropyron
elongatum to wheat improvement in China. Journal of Genetics and Genomics 35: 451-456.
Liebig MA, Schmer MR, Vogel KP, Mitchell RB. 2008. Soil carbon storage by switchgrass
grown as bioenergy. Bioenergy Research 1(3-4): 215-222.
64

Lym RG, Tober DA. 1997.Competitive grasses for leafy spurge (Euphorbia esula) reduction.
Weed Technology 11(4): 787-792.
Macedo TB, Peterson RKD, Weaver K. 2006. Photosynthetic responses of wheat, Triticum
aestivum L. plants to simulated insect defoliation during vegetative growth and at grain fill.
Environmental Entomology 35(6): 1702-1709.
Malinowski DP, Hopkins A A, Pinchak WE, Sij JW, Ansley RJ. 2003. Productivity and survival
of defoliated wheatgrasses in the rolling plains of Texas. Agronomy Journal 95: 614-626
Marais GF. 1992. The modification of a common wheat-Thinopyrum distichum translocated
chromosome with a locus homoallelic to Lr19. Theoretical and Applied Genetics 85: 73-78.
Masnatta WJ, Ravetta DA. 2011. Seed-yield and yield components response to source sink ratio
in annual and perennial species of Lesquerella (Brassicaceae). Industrial Crops and Products 34:
1393-1398.
McCormick AJ, Cramer MD, Watt DA. 2006. Sink strength regulates photosynthesis in
sugarcane. New Phytologist 171: 759-770.
McCormick AJ, Cramer MD, Watt DA. 2008. Changes in photosynthetic rates and gene
expression of leaves during a source-sink perturbation in sugarcane. Annals of Botany 101: 89102
McKee IF, Woodward FI. 1994. CO2 enrichment responses of wheat: Interactions with
temperature, nitrate and phosphate. New Phytologist 127(3): 447-453.
Milne RM, Haynes RJ. 2004. Soil organic matter, microbial properties,
and aggregate stability under annual and perennial pastures. Biology and Fertility of Soils.
39:172-178.
Miyamoto S, Martinez I, Padilla M., Portillo A. 2004. Landscape plant lists for salt tolerance
assessment. Texas A&M University Agricultural Research and Extension Center at El Paso,
Texas Agricultural Experiment Station.
Moore KJ, Vogel KP, Klopfenstein TJ, Masters RA, Anderson BE. 1995. Evaluation of four
intermediate wheatgrass populations under grazing. Agronomy Journal 87: 744-747
Murphy KM, Lyon SR, Balow KA, Jones SS. 2010. Post-sexual cycle regrowth and grain yield
in Thinopyrum elongatum x Triticum aestivum amphiploid. Plant Breeding 129: 480-483.
Nagarajan S, Singh DV. 1990. Long distance dispersal of rust pathogens. Annual Review of
Phytopathology 28: 139-153.
65

Nelson JDJ, Schoenau JJ, Malhi S. 2008. Soil organic carbon changes and distribution in
cultivated and restored grassland soils in Saskatchewan. Nutrient Cycling in Agroecosystems
82(2); 137-148.
Nippert JB, Knapp AK. 2007. Soil water partitioning contributes to species coexistence in
tallgrass prairie. Oikos 116(6): 1017-1029.
Niva M, Svensson BM, Karlsson PS. 2003. Nutrient resorption from senescing leaves of the
clonal plant Linnea borealis in relation to reproductive state and resource availability. Functional
Ecology 17(4): 438-444.
Obeso JR. 2002. The costs of reproduction in plants. New Phytologist 155(3): 321-348.
Oram RN. 1996. Secale montanum: a wider role in Australasia? New Zealand Journal of
Agricultural Research 39:629–633.
Owensby CE, Coyne PI, Ham JM, Auen LM, Knapp AK. 1993. Biomass production in a
tallgrass prairie ecosystem exposed to ambient and elevated CO2. Ecological Applications 3(4):
644-653.
Peet MM, Kramer PJ. 1980. Effects of decreasing source/sink ratio in soybeans on
photosynthesis, photorespiration, transpiration and yield. Plant, Cell and Environment 3(3): 201206.
Pinter PJ, Kimball BA, Garcia RL, Wall GW, Hunsaker DJ, LaMorte RL. 1996. Free air CO2
enrichment: responses of cotton and wheat crops. In Koch GW, Mooney HA. eds. Carbon
Dioxide and Terrestrial Ecosystems. Physical Ecology Series. San Diego, CA: Academic Press,
1996. pp. 215-250.
Piper FI, Cavieres LA, Reyes-Diaz M, Corcuera LJ. 2006. Carbon sink limitation and frost
tolerance control performance of the tree Kageneckia angustifolia D. Don (Rosaceae) at the
treeline in central Chile. Plant Ecology 185(1): 29-39.
Ploschuk EL, Slafer GA, Ravetta DA. 2005. Reproductive allocation of biomass and nitrogen in
annual and perennial Lesquerella crops. Annals of Botany 96: 127-135.
Prakash S, Singhal NC. 2003. Natural cross-pollination in bread wheat (Triticum aestivum L.).
Seed Research 31: 22-26.
Primack RB, Hall P. 1990. Costs of reproduction in the pink lady’s slipper orchid: a four year
experimental study. American Naturalist 136(5): 638-656.
Pritts MP, Hancock JF. 1985. Lifetime biomass partitioning and yield component relationships in
the highbush blueberry, Vaccinium corymbosum. American Journal of Botany 72: 446-452.
66

Ralphs MH, Manners GD, Gardner DR. 1998. Influence of light and photosynthesis on alkaloid
concentration in larkspur. Journal of Chemical Ecology 24(1): 167-182.
Reekie EG, Bazzaz FA. 1987a. Reproductive effort in plants. 1. Carbon allocation to
reproduction American Naturalist 129: 876-896.
Reekie EG, Bazzaz FA. 1987b. Reproductive effort in plants. 2. Does carbon reflect the
allocation of other resources? American Naturalist 129: 897-906.
Reich PB, Hobbie SE, Lee T, Ellsworth DS, West JB, Tilman D, Knops JMH, Naeem S, Trost, J.
2006. Nitrogen limitation constrains sustainability of ecosystem response to CO2. Nature 440:
922-925.
Ridley AM, Simpson RJ, White RE. Nitrate leaching under phalaris, cocksfoot, and annual
ryegrass pastures and implications for soil acidification. Australian Journal of Agricultural
Research 50(1): 55-64.
Ridley AM, White RE, Helyar KR, Morrison GR, Heng LK, Fisher R. 2001. Nitrate leaching
loss under annual and perennial pastures with and without lime on a duplex (texture contrast) soil
in humid southeastern Australia. European Journal of Soil Science 52(2): 237-252.
Reimann-Philipp, R. 1986. Perennial spring rye as a crop alternative. Journal of Agronomy and
Crop Science 157: 281-285.
Reimann-Philipp R. 1995. Breeding perennial rye. Plant Breeding Reviews 13:265–292.
Robertson GP, Paul EA, Harwood RR. Greenhouse gases in intensive agriculture: contributions
of individual gases to the radiative forcing of the atmosphere. Science 289(5486): 1922-1925.
Rosas C, Cock JH, Sandoval G. 1976. Leaf fall in cassava. Experimental Agriculture 12: 395400.
Roumet C, Roy J. 1996. Prediction of the growth response to elevated CO2: A search for
physiological criteria in closely related grass species. New Phytologist 134: 615-621.
Roumet C, Urcelay, C, Díaz, S. 2006. Suites of root traits differ between annual and perennial
species growing in the field. New Phytologist 170(2): 357-368.
Salopek-Sondi B, Kovac M, Ljubesic N, Volker M. 2000. Fruit initiation in Helleborus niger L.
triggers chloroplast formation and photosynthesis in the perianth. Journal of Plant Physiology
157(4): 357-364.
Sawada S, Hasegawa Y, Kasai M, Sasaki M. 1989. Photosynthetic electron transport and carbon
metabolism during altered source/sink balance in single rooted soybean leaves. Plant and Cell
Physiology 30: 691-698.
67

Sawada S, Arakawa A, Muraki I, Echigo H, Miyashita M, Iwafune M, Kasai M, Usuda I. 1999.
Photosynthesis with single-rooted Amaranthus leaves. I. Changes in the activities of RuBP-1,5bisphosphate carboxylase and phosphoenolpyruvate carboxylase and the amounts of
intermediates in photosynthetic metabolism in response to changes in the source-sink balance.
Plant and Cell Physiology 40: 1143-1151.
Scheinost P, Lammer D, Cai X, Murray TD, Jones SS. 2001. Perennial wheat: a sustainable
cropping system for the Pacific Northwest. American Journal of Alternative Agriculture 16:147–
151.
Schulz-Schaeffer J, Haller SE. 1987. Registration of Montana-2 perennial x Agrotriticum intermediodurum Khizhnyak. Crop Science 27: 822-823.
Schulz-Schaeffer J, Haller SE. 1988. Alien chromosome addition in wheat: Advanced progeny.
Genome 30: 303-306.
Sharkey TD, Bernacchi CJ, Farquhar GD, Singsaas EL. 2007. Fitting photosynthetic carbon
dioxide response curves for C3 leaves. Plant, Cell and Environment 30: 1035-1040.
Shi C, Tian C, Li J, Zhang Y. 2011. Prospect of perennial wheat in agroecological systems.
Procedia Environmental Sciences 11: 1574-1579.
Smoliak S. 1981. Hay and pasture crops: Kinds of forage crops. 7-37. In Alberta Forage Crops
Advisory Committee, Alberta Forage Manual, Alberta Agriculture, Edmonton.
Snow AA, Whigham DF. 1989. Costs of flower and fruit production in Tipularia discolor
(Orchidaceae). Ecology 70(5): 1286-1293.
Tang ZX, Ross K, Ren ZL, Yang ZJ, Zhang HY, Chikmawati T, Gustafson JP. 2011. Secale.
Wild Crop Relatives: Genomic and Breeding Resources. Chittaranjan K. ed. Berlin: Springer
Verlag, 2011. pp. 367-396.
Taylor SH, Hulme SP, Rees M, Ripley BS, Woodward FI, Osborne CP. 2010. Ecophysiological
traits in C3 and C4 grasses: a phylogenetically controlled screening experiment. New Phytologist
185: 780-791.
Thomas MB, Wratten SD, Sotherton NW. 1992. Creation of island habitats in famrmland to
manipulate populations of beneficial arthropods: Predator densities and species composition.
Journal of Applied Ecology 29(2): 524-531.
Tscharntke T., Greiler HJ. Insect communities, grasses and grasslands. Annual Review of
Entomology 40: 535-558.
Tsitsin NV. 1965. Remote hybridisation as a method of creating new species and varieties of
plants. Euphytica 14:326–330.
68

Tuell JK, Fiedler AK, Landis D, Isaacs R. 2008. Visits by wild and managed bees
(Hymenoptera: Apoidea) to eastern U.S. native plants for use in conservation programs.
Environmental Entomology 37(3): 707-718.
Uprety DC, Dwivedi N, Raj A, Jaiswal S, Paswan G, Jain V, Maini HK. 2009. Study on the
response of diploid, tetraploid and hexaploid wheat to elevated CO2. Physiology and Molecular
Biology of Plants 15: 161-168.
Van Tassel DL, DeHaan LR, Cox TS. 2010. Missing domesticated plant forms: can artificial
selection fill the gap? Evolutionary Applications 3: 434-452.
Vazquez de Aldana BR, Geerts RHEM, Berendse, F. 1996. Nitrogen losses from perennial grass
species. Oecologia 106: 137-143.
Villalobos FJ, Testi L, Hidalgo J, Pastor M, Orgaz F. 2006. Modelling potential growth and yield
of olive (Olea europaea L.) canopies. European Journal of Agronomy 24: 296-303.
Wagoner P. 1995. Intermediate wheatgrass (Thinopyrum intermedium): development of a
perennial grain crop. In: Williams JT ed(s). Cereals and pseudocereals. (Underutilized crop
series; 2). London: Chapman and Hall, pp.247-259.
Wagoner P, Schaeffer JR. 1990. Perennial grain development: past efforts and
potential for the future. Critical Reviews in Plant Sciences 9(5): 381-408.
Wagoner P, Schauer A. 1990. Intermediate wheatgrass as a perennial grain crop. In Janick, J. and
Simon, J.E. eds. Advances in new crops : proceedings of the First National Symposium NEW
CROPS, Research, Development, Economics, Indianapolis, Indiana, Oct 23-26, 1988. Portland:
Timber Press, 1990.
Wang RRC, Larson SR, Jensen KB. 2010. Analyses of Thinopyrum bessarabicum, T. elongatum
and T. junceum chromosomes using EST-SSR markers. Genome 53: 1083-1089.
Wang RRC, Li X-M, Hu Z-M, Zhang J-Y, Larson SR, Zhang X-Y, Grieve CM, Shannon MC.
2003. Development of salinity tolerant wheat recombinant lines from a wheat disomic addition
carrying a Thinopyrum junceum chromosome. International Journal of Plant Sciences 164: 2533.
Weik L, Kaul H-P, Kubler E, Aufhammer W. 2002. Grain yields of perennial grain crops in pure
and mixed stands. Journal of Agronomy and Crop Science 188(5): 342-349.
West TO, Post WM. Soil organic carbon sequestration rates by tillage and crop rotation: a global
data analysis. Soil Science Society of America Journal 66: 1930-1946.
Wheelwright NT, Logan BA. 2004. Previous year reproduction reduces photosynthetic capacity
and slows lifetime growth in females of a neotropical tree. Proceedings of the National Academy
of Sciences 101: 8051-8055.
69

Witkowski ETF, Lamont BB. 1995. Disproportionate allocation of mineral nutrients and carbon
between vegetative and reproductive structures in Banksia hookeriana. Oecologia 105: 38-42.
Wullschleger SD. 1993. Biochemical limitations to carbon assimilation in C3 plants: a
retrospective analysis of the A / Ci curves from 109 species. Journal of Experimental Botany 44:
907-920.
Xia G, Cheng G, Lakso A, Goffinet M. 2009. Effects of nitrogen supply on source-sink balance
and fruit size of ‘Gala’ apple trees. Journal of the American Society for Horticultural Science
134: 126-133.
Xu X, Nicholson P. 2009. Community ecology of fungal pathogens causing wheat head blight.
Annual Review of Phytopathology 47: 83-103.
Yousufzai M-N, Siddiqui KA, Soomroo AQ. 2009. Flag leaf stomatal frequency and the
interrelationship with yield and yield components in wheat (Triticum aestivum). Pakistani
Journal of Botany 41(2): 663-666.
Zan CS, Fyles JW, Girouard P, Samson RA. 2001. Carbon sequestration in perennial bioenergy,
annual corn and uncultivated systems in southern Quebec. Agriculture, Ecosystems and
Environment 86(2): 135-144.
Zhao M, Botwright Acuña TL, Lafitte HR, Dimayuga G, Sacks E. 2008. Perennial hybrids of
Oryza sativa x O. rufipogon: Carbon exchange and assimilate partitioning. Field Crops Research
106: 214-223.
Zhao M, Ding Z, Lafitte R, Sacks E, Dimayuga G, Holt D. 2010. Photosynthetic characteristics
in Oryza species. Photosynthetica 48: 234-240.
Zhou R, Quebedeaux B. 2003. Changes in photosynthesis and carbohydrate metabolism in
mature apple leaves in response to whole plant source-sink manipulation. Journal of the
American Society for Horticultural Science 128: 113-119.
Zotz G, Richter A.2006. Changes in carbohydrate and nutrient contents throughout a
reproductive cycle indicate that phosphorus is a limiting nutrient in the epiphytic bromeliad,
Werauhia sanguinolenta. Annals of Botany 97: 745-754.

70

CHAPTER 2. AGRONOMY OF PERENNIAL WHEAT AND RYE

71

Introduction

In recent years, increasing attention has been paid to developing perennial alternatives to
annual food crops (Cox et al., 2006; Sacks et al., 2005; Piper and Kulakow, 1994; DeHaan et al.,
2005; Jackson and DeWald, 1994). Perennial cereals are of interest due to potential for improved
environmental services (e.g. Ridley et al., 2001; Crews, 2005) and lower production costs. Since
the majority of cropping systems worldwide are based on staple cereals, developing perennial
cereals will be critical to large-scale perennialization of agriculture. Two promising perennial
cereal candidates for cold-temperate regions, developed by hybridizing annual cereals with
perennial wild relatives, include perennial wheat (Triticum aestivum x Thinopyrum spp.) and
perennial rye (Secale cereale x S. montanum). The current scarcity of plant life-forms combining
herbaceous growth form, perennial life history, and very high reproductive allocation (Van Tassel
et al., 2010) makes these two cereals interesting ecologically as well as agronomically.
It is thought that grain yield penalties may occur in perennials due to resource tradeoffs
between reproduction on the one hand, and winter survival and regrowth on the other. A
modeling study (Bell et al., 2008) concluded that perennial wheat would be economically viable
in Australia if it achieved 40% of annual wheat grain yields, combined with at least 800 kg ha

-1

of forage per year (or alternatively if it could achieve at least 60% of annual wheat yield with no
additional forage production). This prompts the question: can perennial cereals, at their current
stage of development, achieve the threshold level of 40-60% yield level relative to annuals?
Forage production, particularly early in the season, may be an important additional contribution
of perennial cereals, which makes it critical to measure biomass production as well as grain
yield. Finally, an agronomic assessment of perennial cereals will need to include an
72

understanding of their phenology. Differences in phenology between perennial cereals and their
annual relatives could affect early season grazing potential, susceptibility to pathogens (Emrich
et al., 2008), and vulnerability to extreme weather, but little information is available on the
phenology of emerging perennial cereal species.
To understand the agronomic potential of perennial cereals, they must be studied over
multiple years, as grain production may change with stand age. Perennial cereals might display
delayed reproductive investment, similarly to many woody perennials: in this case, plants would
show low reproductive investment the first year (the establishment year), but in future years the
established root and crown reserves would allow plants to begin growth earlier, grow larger, and
show marked increases in seed production (Jackson and Jackson, 1999). Alternatively, yield
decreases might be observed due to the buildup of soil pathogens, short plant life span, or the
proliferation of weak and unproductive tillers. This is a pattern that is observed in grasses such as
Bromus inermis, Festuca arundinacea, and Poa pratensis (Loch et al., 1999; Fulkerson, 1980;
Fairey and Lefkovitch, 2001), although a number of other cool season forage grasses show
stability in seed yields over two to six seasons (Chastain et al., 1997; Canode and Law, 1978;
Mueller-Warrant and Rosato, 2002; Fulkerson, 1980) or even increases as demonstrated for
Agropyron cristatum and Festuca rubra (Canode, 1968). In order to test the hypothesis that
yields will decrease over time, it is necessary to study perennial cereals over at least two years.
Perennial wheat and rye are still in the process of development, and there are relatively
few studies of grain yield under agronomic conditions. Early work on perennial wheat was done
in the former Soviet Union, but little of this data has been easily accessible to researchers in
other countries. More recently, first year grain yield was measured in a study of 31 perennial
wheat genotypes in eastern Washington (Murphy et al., 2010), finding yields of up to 93% of
73

annual wheat in the highest yielding accession. Grain yield was studied in approximately 90
perennial wheat derivatives over two years in Australia (Hayes et al., 2012) and with some
earlier cultivars of perennial rye over two- or three-year periods: “Permontra” (Reimann-Philipp,
1986; Weik et al., 2002) and the newly developed ACE-1 (Acharya et al., 2003; Acharya et al.,
2004). These studies generally found perennial rye yielded approximately 55-60% of annual rye,
and that grain yield of perennial wheat lines were highly variable, ranging between 2% and
135% of annual wheat among those lines that showed appreciable regrowth. However, these
multi-year studies did not explicitly separate the effects of calendar year (reflecting year to year
weather fluctuations) from plant age (reflecting possible effects of senescence, metabolic
tradeoffs, changes in allocation, changing energy and nutrient status of the plants, and other agerelated phenomena). Our study makes a novel contribution to the growing literature on perennial
cereals by observing one year-old plants in two different years (achieving replication in time), in
a new and different environment, and by comparing one and two year-old plants within a single
year (thus separating the effects of plant age and calendar year). We were thus able to consider
how genotype / cultivar, calendar year, and plant age each affected plant yield, biomass and
phenology.
Between 2008 and 2010, we assessed the agronomic potential of perennial wheat and
perennial rye in southwest Michigan using a multi-year field experiment. We measured
parameters relating to biomass, yield, and phenology to document how these plants compare,
over time, to their annual relatives. Our guiding hypotheses were the following: 1) One year-old
perennial wheat and rye will yield lower than their annual analogues, 2) Two year-old perennial
wheat and rye will produce lower biomass and lower seed yields, compared to first-year plants,
and 3) Two year-old perennial wheat and rye will show more early season growth, and earlier
74

flowering, than one year-old perennial wheat and rye. We also hoped to establish whether current
accessions of perennial wheat and rye could achieve the threshold for economic feasibility set by
modeling studies (Bell et al., 2008).

Materials and Methods

Site Characteristics
We conducted this study at the W. K. Kellogg Biological Station (KBS) located in
southwest Michigan, USA, 50 km east of Lake Michigan (42° 24′ N, 85° 24′ W, elevation 288
m) on soils developed from glacial outwash deposited 12000 years ago. Soils are mainly of the
Kalamazoo (fine-loamy, mixed, mesic Typic Hapludalfs), Oshtemo (coarse-loamy, mixed, mesic
Typic Hapludalfs) and Miami (fine-loamy, mixed mesic Typic Hapuldalfs) series, which all occur
on our site (Crum and Collins, 1995). The area receives ca. 897 mm of precipitation annually
(based on a 24-year average), with approximately half as snow, and the mean annual temperature
is 9.7° C. The study was laid out in a field approximately 20 m wide by 73 m long, oriented
roughly north-south on level ground. Further site and soils descriptions for KBS are available at
http://lter.kbs.msu.edu/Data/LTER_Metadata.jsp/Dataset/KBS042.

Accession Selection
Eight accessions (either breeding lines or named varieties) were involved in the study.
The annual winter wheat (Triticum aestivum) ‘checks’ included Frankenmuth (PVP 8000165),
and Pioneer 25R37 (PVP 0020232: Pioneer Hi-Bred Ltd.). Frankenmuth is an older cultivar
75

which was released in 1971 from MSU. It was commonly grown in the 1980s and 1990s (e.g.,
Huebner et al., 1999) and has been used as a benchmark for yield and quality in breeding studies
at Michigan State University. It lacks the semi-dwarfing genes common in modern cultivars, and
has somewhat greater biomass and lower grain yield. In a 2007 comparison of grain yields across
29 varieties and six sites, Frankenmuth yielded approximately 13% lower than the mean, and
11% lower than Pioneer 25R37. Pioneer 25R37 is a newer, high yielding variety that was widely
grown in the early to mid-2000s and considered well adapted for Michigan. In the
aforementioned study from 2007, Pioneer 25R37 yielded the same as the mean across all 29
lines. Pioneer 25R37 is a soft red variety while Frankenmuth is a soft white.
Four perennial wheat accessions were included (Triticum aestivum cv. ‘Chinese Spring’ x
Thinopyrum elongatum // T. aestivum cv. ‘Madsen’), all of which were drawn from a larger
group of accessions contributed by Washington State University in 2005, and had been involved
in previous studies in Washington (Murphy et al., 2009; Murphy et al. 2010). We selected four
accessions out of the 20 based on observed differences in harvest index and morphology
displayed in a 2008 pilot observational trial at Kellogg Biological Station. P3 showed a relatively
low harvest index (as well as more ‘grass-like’ spikelet morphology) while P11 had the highest
harvest index combined with spikelets resembling annual wheat: P15 and P19 had intermediate
harvest index and morphology. Also included was a cover annual rye cultivar (Secale cereale cv.
‘Wheeler’), and a perennial hybrid rye cultivar (Secale cereale x montanum cv.‘Rival’).

Experimental Design
This study was laid out as a split plot RCBD with five blocks, using ‘planting year’ as the
76

whole-plot factor and ‘accession’ as the split-plot factor. ‘Planting year’ was an indication of
when a particular plot was first planted (2008 or 2009). When comparing one year-old plants in
2009 to one year-old plants in 2010, ‘planting year’ thus reflected year-to-year weather variation
(and will be referred to in the analysis as ‘year’); when comparing older vs. younger perennial
plants in 2010, ‘planting year’ reflected the age of the plants, and will be referred to as ‘age’
(since one year-old plants had been planted in 2009, and two year-old plants had been planted in
2008). Thus we were able to consider separately the effects of both year-to-year weather
variation, as well as plant age (Shefferson and Roach, 2010). A very similar approach (comparing
a single cohort across years, and multiple age cohorts within years) has been used to study a
close relative of an agricultural plant, Beta maritima, in the greenhouse (Van Dijk, 2009). In the
analysis, tables, and graphs, the term ‘2009’ will be taken to refer to the entire Oct. 2008 - Sept.
2009 growing season, and ‘2010’ to the entire Oct. 2009 – Sept. 2010 growing season.
One full set of plants, with each of the eight accessions each replicated five times, was
first planted in fall 2008. Following harvest in Aug. 2009, perennial accessions were allowed to
regrow into a second season, while the annual plots were tilled under and replanted with the
same annual cultivars in Nov. 2009. Also during Nov. 2009, a full second set of perennial plants
was planted. Thus in 2010 the study included two year-old perennial plants, one year-old
perennial plants, and one year-old annual plants. The subplots were each approximately 5.47 m
long, and included six rows of plants with 19 cm between rows; there were 57 cm alleys between
subplots and 90 cm between whole plots (Table 1).

Management
Dates of field operations are given in Table 2. Plants were seeded at 1.5 million
77

-1

seeds ha about 1.9 cm deep, using an Almaco small plot planter. The field was fertilized in fall
2008 with composted dairy manure (at the rate of 90 kg ha
with approximately 102 kg ha

-1

-1

N), while in fall 2009 we fertilized

N in the form of pig blood meal (13% N). Plots were irrigated

once in summer 2009, applying approximately 3.2 cm of water. In Oct. 2010, regrowth was
extremely low and variable, with no accession showing more than 10% regrowth, and over half
the plots showing zero regrowth. This might have been due to the warm spring coupled with hot
and dry weather in the late summer, and intense weed pressure. Because of very poor regrowth,
the experiment was ended in Oct. 2010. During each season, weeding took place four times and
was carried out by hand.

Measurements
On 20 March 2009 and 16 March 2010, about one week after snowmelt, the number of
growing plant stems per meter in each plot were counted (which estimated spring emergence for
annuals and one year-old perennials, and regrowth for the two year-old perennials). Plant height
was estimated by sampling six plants from each plot, at the same time that we counted plant
stems. To estimate flowering date, plants were observed once every four days from 20 May to 23
July 2009 or from 1 May to 6 July 2010. We estimated the percent having flowered at each date
and calculated a 50% flowering date through linear interpolation. In the analysis, “first-year
plants” denotes plants during their first growing season. “Second-year plants” denotes plants
during their second full growing season.
Harvest during the 2009 and 2010 growing seasons took place six to seven weeks
following flowering for each accession. Harvest dates in 2009 were 23 July for perennial rye,
78

annual rye and annual wheat, and 13 August for perennial wheat. Harvest dates in 2010 were 10
July for annual wheat, 17 July for annual and perennial rye, and 9 August for perennial wheat.
Hand harvests were taken for yield estimation on 4 m lengths for each of the two central rows:
o

seed heads, leaves, and stems were dried at 65 C for four days (until weight did not change) and
re-weighed. Because a pilot study had found that perennial wheat had low threshability in
standard equipment, we used a customized tabletop seed cleaner at Michigan State University to
estimate the grain percentage of the total seed head mass. These estimates were based on the
grain recovery from 70 g subsamples of unthreshed seed heads for each accession. Our
calculated grain mass / seed head mass ratios obtained from the tabletop seed cleaner were
multiplied by the measured total seed head mass for each plot, to give an estimate of threshed
seed yield. At the time of harvest, we counted reproductive tillers on the harvested rows:
following seed cleaning we measured the thousand-kernel weight on bulked seed samples of
each accession and divided by 1000 to obtain kernel mass. In Oct. of each year, four 20 cm

2

microplots in each plot were sampled for regrowth. Regrowth was estimated as (number of
visibly regrowing green leaves) x (number of reproductive tillers at harvest)

-1

x 100.

Data Analysis
To determine the effect of year-to-year variation on each parameter of interest, we
compared first-year plants in 2009 to first-year plants in 2010 (including annual wheat and rye
checks). Our model was a split plot RCBD using calendar ‘year’ as the main plot factor,
‘accession’ as the subplot factor, calendar ‘year x accession’, and block. Following the overall
analysis, we then held year constant to determine the effect of accession differences within each
79

year. To determine the effect of plant age on each parameter, we compared first-year plants
(perennial accession only) in 2010 to second-year plants within the same year. Here our model
was a split plot RCBD using plant ‘age’ as the main plot factor, ‘accession’ as the subplot factor,
‘age x accession’, and block. To determine differences between accession within each age class,
we held age constant and considered the effect of accession. All analyses were done using the
MIXED procedure in SAS 9.2 (SAS Institute, 2008) followed by planned contrasts using
protected LSD.
Analysis of kernel mass did not include a block effect since these were measured after
bulking seed from all blocks; we used five replicates for each variety. Mass-related parameters
were log-transformed to improve normality, threshability and harvest index were logittransformed to meet assumptions of ANOVA, and other variables were untransformed.
The overall model for one of the response variables (threshed yield) was, e.g.:
ln (yield) = block + year + block x year + accession + year x accession.

Results

Environment
The pattern of precipitation in the fall of 2008 and 2009 was similar, with a dry
November (50% below the 24 year average), but otherwise well distributed precipitation (Figure
1). Winter temperatures may have affected survival, as evidenced by the lower springtime
emergence in 2010 following a very cold January). Early spring precipitation was wet in 2009
(54% above the 24-year average) and dry in 2010 (25% below the 24-year average). However,
80

there was no difference between the years in terms of early spring growth of perennial cereals, as
indicated by March plant height measurements (Table 6). The later spring (May and June) was
much wetter in 2010 than in 2009, with 120% higher precipitation; overall, the spring was
o

consistently warmer in 2010 than in 2009, with an average difference of 1.8 C over the MarchJune period. Thus the 2010 spring was warmer, drier early on but much wetter in the later spring,
as compared to 2009.

Yield of First-year Perennials and Annuals
Threshed grain yield of first-year plants was similar between 2009 and 2010, with no
overall main effect of year (Table 3). The yield of the annual cereals was low: averaging across
-1

both years, grain yields were 2.41 Mg ha for Pioneer 25R37, 2.94 Mg ha

-1

for Frankenmuth

-1

and 1.83 Mg ha for Wheeler rye. These yields are consistent with those achieved at our site in
the past (Smith et al., 2008) reflecting the sandy soils and the fact that our southwest Michigan
location is at the fringes of the wheat belt. Two accessions (P19 and P15) yielded 40% higher in
2010 compared to 2009, while all other accessions did not differ between the two years.
Perennial wheat and perennial rye plants consistently yielded lower than their annual
counterparts (Table 3). The first-year perennial wheat plants, averaging across accession, yielded
45% of annual wheat in 2009 (t = 10.16, p < 0.001), and 54% of annual wheat in 2010 (t = 8.86,
p < 0.0001; Figure 2). Similarly, perennial rye, achieving 76% of annual rye yields in 2009 and
69% of annual rye in 2010 (t = 3.06, p = 0.003; Figure 3). All four perennial wheat accessions
performed the same in 2009, but in 2010 the first-year P19 and P15 plants achieved
approximately 81% higher yields than the lowest yielding line P3.
81

The lower yields in perennial cereals as opposed to their annual relatives reflected lower
ratio of grain to chaff, lower kernel mass, lower harvest index, and a lower density of
reproductive tillers (Table 3): they did not, by contrast, appear to reflect lower biomass
production or lower number of seeds per tiller. The percentage of grain in unthreshed seed heads
was lower in perennial cereals than in annuals: when using the tabletop seed cleaner, our seed
recovery ranged from 66%-67% (in the case of P3 and perennial rye) to 78% (in the case of
P19): seed recovery was 84-86% for the three annual lines. Kernel mass also explained much of
the perennial vs. annual yield differences. Kernels of first year perennial wheat plants (averaging
across accessions) were 31% smaller than annual wheat in 2009 (t = 10.92, p < 0.0001), and 56%
smaller in 2010 (t = 8.53, p < 0.0001). Similarly, kernel mass of first year perennial rye was 24%
lower than annual rye in 2009, and 20% lower in 2010 (p < 0.0001).
Lower number of reproductive tillers per meter contributed to lower seed yield in
perennial wheat, but not in perennial rye. Perennial rye had 50% more reproductive tillers than
annual rye in 2009 (t = 2.87, p = 0.006), and equivalent number of reproductive tillers in 2010,
illustrating that lower grain yields in perennial rye were not explainable by a lower number of
reproductive tillers. Harvest index differed among accessions: first year perennial wheat had a
harvest index 43% lower than annual wheat, while first year perennial rye had a harvest index
23% lower than annual rye. Total biomass, by contrast, was actually higher in perennial rye
compared to annual rye in 2010, while perennial wheat biomass was equivalent to annual wheat
that year. This suggests that differences in biomass were not an important contributor to the
lower grain yields in perennials: rather, annuals appear to allocate a greater fraction of biomass to
seed reproduction and thus had a higher harvest index than perennials (Table 3).
82

Similarly, lower grain yields in perennials do not seem to reflect lower number of seeds
per reproductive tiller. Perennial rye, in both years, had equivalent number of seeds per
reproductive tiller as annual rye. Perennial wheat did not have a consistently lower number of
seeds per tiller, and in fact in 2010 three of the four lines had equivalent number of seeds per
reproductive tiller as Frankenmuth annual wheat, and 60-89% more than Pioneer 25R37. Thus
lower grain yields in the perennial species are not explained by fewer grains per tiller. This is,
importantly, not equivalent to saying that they have equal fertility to annual wheat and rye, since
we did not count the number of sterile spikelets: if perennial wheat and rye have many more
spikelets than their annual analogues, it is possible that lower fertility might coexist with a
comparable or higher number of seeds per tiller. Further work is needed to determine whether
perennials in fact display lower fertility, in the sense of a lower ratio of mature seeds to total
spikelets, as some preliminary observations suggest (DeHaan, L. R., personal communication).

Effects of Plant Age on Yield and Yield Components
Overall, first-year and second-year perennial plants in 2010 showed consistent yields,
with no effect of plant age or age x accession interaction observed on grain yield (Table 4).
Second-year perennial wheat lines yielded 53% of annual wheat (t = 5.08, p < 0.0001),
equivalent to the first-year plants. Grain yield of second-year perennial rye plants, at
-1

approximately 1.44 Mg ha , was not significantly different from either annual rye (p = 0.45), or
first year perennial rye plants (p = 0.10). First-year perennial wheat plants showed accessionrelated differences in yield, with P19 and P15 producing 81% higher yields than P3; in the
second year, however, P19 and P15 yielded only 24% higher than P3, suggesting a decline in
their relative yield advantage.
83

The overall consistent grain yields reflected the lack of strong age-related trends in yield
components, although some individual accessions did show changes in yield components
between first and second year plants (Table 4). Kernel mass, for example, showed a strong
accession effect (F = 12.00, p < 0.0001), an accession x age interaction (F = 4.91, p = 0.03),
reflecting the fact that some lines did show a decrease in kernel mass with age. Second-year
perennial rye and P19 plants had 15% smaller kernels than first year plants, while the other three
perennial wheat lines showed no differences (Figure 4). Similarly, there were no consistent agerelated trends in terms of biomass or harvest index. Comparing across generations in 2010, we
found that biomass was significantly affected by accession, but not by age, nor was there an age
x accession interaction observed (Table 4). Harvest index, similarly, showed strong age x
accession interactions. Specifically, older plants of P11 and perennial rye had roughly a 16%
lower harvest index than younger plants, while P3, P15 and P19 had a 16 – 33% higher harvest
indices. Thus there appeared to be few overall clear trends in yield components between the two
years, which helped to explain the overall lack of change in grain yield.

Plant Growth, PSCR and Phenology
Interestingly, a plant age effect was observed for early season growth: plant height
achieved by mid-March differed between first and second year perennials, with the older plants
being on average 110% taller (Table 6). In mid-March 2010, first-year perennial wheat was 22%
shorter than annual wheat and two year-old perennial wheat was 59% taller. A similar pattern
was observed in perennial rye, where one year-old perennial plants were 27% shorter than annual
rye and two year-old plants were 75% taller. ). Early season height differences were, to some
extent, maintained later in the season: two year-old perennials were overall 10% taller at anthesis
84

than one year-old perennials, with a 26% difference in the case of perennial rye (Table 6).
Second-year perennial wheat and rye were 21% and 11% taller than annual wheat and rye in
2010, respectively. Height at flowering showed effects of calendar year as well as plant age, with
first-year perennial wheat and rye both being shorter in 2010 compared to 2009.
The number of plant stems present in March 2010 was similar for both one year-old and
two year-old perennial plants (Table 6) and there was no main effect of age nor age x accession
interaction. This suggests that each perennial stand, as a whole, was able to fully replenish itself
through either regrowth or re-seeding. Regrowth in the fall of 2009 (Table 5) was vigorous, with
perennial rye and P19 perennial wheat both showing over 100% regrowth, and with the most
poorly regrowing line, P3, showing over 50% regrowth. In the fall of 2010, however, regrowth
was extremely poor: no accession achieved over 10% regrowth, and 55% of plots showed zero
detectable regrowth. Regrowth in the fall of 2010 was equally poor when one and two year-old
plots were compared.
Flowering dates of one year-old plants also showed strong effects of accession (with
perennials generally showing later flowering) as well as effects of calendar year, and the
interaction of calendar year x accession (Table 5). The one year-old perennial wheat generally
flowered on 31 June in 2009 (26 days after annual wheat, p < 0.0001) and 19 June in 2010 (15
days after annual wheat, p < 0.0001), while the average flowering date was 11 June for perennial
rye in both years (11 – 12 days after annual rye). The tendency of plants to flower earlier in
2010, due to the difference in weather, was one of the strongest and most distinct trends between
the years, and was much more marked in the perennial species than in the annual species.
Flowering date in 2010 showed strong differences between first- and second- year
perennials (Table 6). When comparing the one and two year-old perennial plants in 2010, we
85

found that flowering date was significantly affected by accession and by age, but not by the
interaction of the two; thus all second-year perennials showed a similar shift in flowering date
relative to first-year plants (approximately 7 – 11 days earlier). Two year old perennial wheat
plants flowered on 13 June , nine days after annual wheat (t = 7.75, p < 0.0001), while two yearold perennial rye plants flowered at approximately the same time as annual rye, on 3 June.
Perennial wheat accessions did not significantly differ in terms of flowering date. To sum up, the
strong age effects on early-season plant height and flowering date reflect an overall “shift’ in
phenology towards earlier in the year. It is interesting that plant age and perennial vs. annual life
history both seem to have effects of similar magnitude on flowering date (shifting them by one to
two weeks in either direction).

Discussion

To the best of our knowledge, this is one of the first reports outside the former Soviet
Union on a two-year test of yield potential for perennial wheat and rye that separates effects of
plant age from effects of year-to-year weather variation. This allowed testing of our hypotheses
that first-year grain yields in perennial cereals would be lower than annuals, and that perennial
grain yields would decline with increasing plant age.
Consistent with our first hypothesis, first-year yield in perennial wheat and perennial rye
were consistently lower than their annual analogues: perennial rye yielded 72% of annual rye,
while perennial wheat yielded 50% of annual wheat. These results are somewhat higher than the
normal range found for perennial wheat lines tested in Australia or in the Pacific Northwest, but
86

comparable to previous results on these particular lines. A previous study of 31 perennial wheat
lines in Washington (Murphy et al., 2010) included all four of the lines included in this study,and
reported yields ranging from from 28% of annual wheat (in the case of P15) to 51% of annual
wheat (in the case of P19). A more recent study (Hayes et al., 2012) compared over 90 perennial
cereal lines at a plot scale and an additional 86 in single rows. Among the 40 lines which showed
evidence of post-sexual cycle regrowth, grain yield was generally 26 – 42% of annual wheat at
the 95% confidence interval; among the pool of Washington State University (WSU) breeding
lines from which our four lines were selected, grain yield was 23 – 43% of annual wheat, but in
absolute terms higher than achieved in our study (possibly reflecting adverse environmental and
-1

soil conditions at our site). P19, for example, produced approximately 2.2 t ha grain yield,
which was about 50% of ‘Wedgetail’ annual wheat at the site. Additionally, there was no
correlation between vigor of regrowth and grain yield. Annual wheat yields in our study were
higher than in Washington but lower than in Australia, reflecting the marginal nature of our site.
Our estimates of perennial rye yield were 1.3 to 1.4 Mg ha
60% of previously reported values (e.g., 2.2 to 2.4 Mg ha

-1

-1

in first year plants, about

in central Europe; Reimann-Philipp,

1986; Weik et al., 2002). However, perennial rye yields relative to annual rye were actually
higher in our study than in previous studies, indicating that in this relatively low-yielding
environment perennial rye suffers less of a yield reduction than annual rye. For example, studies
-1

of the perennial rye ACE-1 in western Canada found yields of 2.5 Mg ha , averaging over 1 –
3 year old plants, approximately 55% of annual rye yield at that site (Acharya et al., 2004).
The lower yields in perennial cereals reflected lower reproductive allocation and kernel
mass. This is consistent with previous studies of perennial sorghum, which showed 35% lower
87

kernel mass and 16% lower reproductive allocation than annual sorghum (Piper and Kulakow,
1994) as well as ACE-1 perennial rye which had lower harvest index but greater biomass than
annual rye (Acharya et al., 2004). Lower grain yields in perennials may also have reflected lower
initial plant population density. In both years the perennial lines generally had fewer growing
plant stems in the early spring, as well as (in perennial wheat) fewer reproductive tillers.
Interestingly, the shorter period for pre-winter growth in the fall of 2009, does not appear to have
reduced yields in 2010 relative to 2009. Growing conditions prior to vernalization can have a
very strong effect on subsequent year yields in cool season perennial grasses (Chastain and
Young, 1998), but such effects were not seen in our study.
We did not find support for our second hypothesis: grain yield in perennial wheat and rye
showed no effect of plant age. This is consistent with the results of Hayes et al. (2012) who
found that second-year perennial wheat grain yield was highly variable between the perenniating
lines, ranging from about 1% to over 1000% of first-year yield (depending on the accession). The
data from this previously published experiment suggests that some perennial wheat accessions do
show a strong pattern of delayed reproductive investment, while others are close to being fully
annual with very little second year seed production. The four lines in our study were in the
middle of their range of observations, with approximately equal first- and second-year yields, as
well as slightly higher yields overall. We found that perennial rye also maintained equivalent
yields in the first and second years: this is consistent with an early study of ‘Permontra’
(Reimann-Philipp, 1986) but conflicts with a later study that found a 73 – 88% decline in yield
(Weik et al., 2002). However, yield declines in the 2002 study were attributed to heavy weed
pressure and possibly nitrogen limitation as weed cover increased 2500% in the second year. In
particular, our study highlights the potential that perennial rye might have under high-nutrient,
88

intensively managed conditions. The lack of yield declines in our two perennial species is
encouraging.
Previous studies of perennial forage grasses supplement the sparse literature on perennial
wheat and rye, and provide conflicting evidence concerning the potential for second-year seed
yield. A study of Pennisetum polystachyon and Andropogon gayanus, involving spaced plants,
found increases in seed yield between years 1 and 2, followed by decrease (Mishra and
Chatterjee, 1968). Seed yields doubled from the first to the second year in desert wheatgrass
(Agropyron desertorum: Canode and Law, 1978) and also increased in red fescue (Festuca rubra:
Canode, 1968). In contrast, yields of Kentucky bluegrass (Poa pratensis: Chastain et al., 1997),
orchardgrass (Dactylis glomerata: Fulkerson, 1980) and perennial ryegrass (Lolium perenne:
Mueller-Warrant and Caprice Rosato, 2002) stay generally stable over the first few years. Still
other grasses show yield declines starting from the first year, e.g. Russian wildrye (Elymus
junceus: Lawrence and Ashford, 1964) and timothy (Phleum pratense; Fulkerson, 1980). Thus
changes in yield are highly dependent on the species, and our study provides important new
information about two individual species of interest.
As yet, little is known about the biomass production potential of perennial cereals, in
spite of the fact that increased forage production could be an important supplementary benefit of
these species. In our study we found that biomass was equivalent or higher in perennial wheat
and rye, compared to their annual analogues, while perennials allocate a lower fraction of
biomass to reproduction. No significant differences in biomass were seen between one and
twoyear-old perennial plants, indicating no decline or increase as plants ages (Table 5). This
contrasts with a field study which found that perennial rye biomass yields declined from around
-1

14 to around 7.5 tons ha between their first- and second-years (Acharya et al., 2003). Because
89

perennial rye biomass was harvested twice in their experiment, these results are not directly
comparable to our own. It is suggestive, however, that in the western Canada environment the
authors found substantially higher first-year biomass yields of perennial rye than we found in our
study. Results from studies of forage grasses over multiple years vary between species. Grasses
often produce increasing biomass yields during the first few years of their life: for example,
Miscanthus biomass production tripled between the first and second year (Clifton-Brown and
Lewandowski, 2002) and similar increases in yield over the first three years have been found in
Arundo donax (Mantineo et al., 2009) and in switchgrass (Schmer et al., 2010). However, a
different study of newly planted forage grasses, including close wheat relatives, found strong
declines in biomass yields over five years of study (White, 1985). We found that reproductive
allocation, reflected by harvest index, remained constant between one year-old and two year-old
plants. This is consistent with the hypothesis that age has little effect on seed production or
reproductive allocation in these perennial cereals. Perennial forage grasses may show increases
in harvest index between the first and second years (e.g. Festuca arundinacea) or decreases (Poa
pratensis), depending on the species (Fairey and Lefkovitz, 2001) while woody agricultural
perennials tend to show increases.
The high rate of regrowth in fall 2009 is encouraging, given that previous work on these
lines in Washington had found only 40% PSCR in the fall (Murphy et al., 2010). Hayes et al.
(2012) found that only a minority of lines showed strong regrowth, and specifically that of the
four perennial wheat lines involved in our study, two showed no regrowth and two showed very
poor regrowth (generally scoring between 0 – 1.9, compared to 7.2 for perennial rye). Similarly
a field study of Oryza sativa x rufipogon found that perenniality could quickly be lost following
backcrossing to the annual parent (Sacks et al., 2006). However, in keeping with the findings of
90

these previous studies, we did find strong year-to-year variation in the ability of these perennial
plants for PSCR. In 2010 we found very low regrowth in all perennial lines, possibly due to the
combination of weed pressure, a hot spring, and very dry weather in late summer. The extreme
variability of regrowth in perennial wheat and rye indicates that further breeding efforts should
prioritize vigorous regrowth and reliable perenniality. As yet perenniality in these lines is not
reliable enough for them to be used as economically viable perennial crops.
Our findings regarding phenology were consistent with the general trend that perennial
grasses tend to have delayed phenology compared to closely related annuals (Garnier et al.,
1997). Interestingly, older perennial plants flowered substantially earlier than younger ones in
our study. There is contrasting evidence from the literature regarding the effect of plant age on
phenology, which differs across species. For example, a study of Beta maritima, a close relative
of the agriculturally important common beet, found that with each increase of one year in age,
plants flowered approximately 1.3 days later (Van Dijk, 2009). Conversely, older plants flower
earlier in perennial Lupinus populations (Bishop and Schemske, 1998) as well as in other species
(Torimaru and Tomaru, 2006). Weik and colleagues (2002) found that two year-old perennial rye
plants flowered 2 – 7 days later than one year-old plants, in contrast to our results where older
plants flowered earlier. However, in the Weik (2002) experiment there was no explicit separation
of calendar year and plant age effects. Our study suggests that calendar year, perennial growth
habit, and plant age can all exert effects of similar magnitude on flowering date (ca. 1 – 2 weeks)
which makes separating age and year effects important for future studies. Flowering date is an
important characteristic of perennial cereal accessions as it could affect harvest time as well as
vulnerability to weather stresses and pathogen attack.
Perennial cereal plants appear to grow more slowly than annuals in their first spring, as
91

illustrated by the lower height achieved early in the season, although we found that plants in their
second spring are larger than annual wheat and rye. This could reflect the ability to draw on
stored root and crown reserves early in the spring, and under certain environmental conditions
may potentially contribute to perennials growing larger than annuals and compensating for lower
reproductive allocation, although this was not observed in our study. While Ward et al. (2011)
found no difference in root or shoot growth over the first month in potted annual and perennial
wheat plants, this does not appear to reflect performance of the plants in the following spring
under natural conditions. Increased growth early in the spring found for second-year perennials
may be one factor contributing to earlier flowering when compared to first-year plants.

Conclusions

Yield potential of perennial cereals did not decline over multiple years, and perennial
wheat and rye consistently produced approximately 50% and 75% of their respective annual
counterparts. This exceeds the 50% minimum threshold of economic feasibility proposed for
perennial cereals to succeed as a single-use grain crop in the Australian context. For perennial
wheat to be economically feasible in places other than extremely marginal lands, however, these
modest seed yields will probably need to be coupled with increased biomass production for
forage, and our experiment found little evidence that perennial cereals can as yet produce higher
biomass than annuals. This may be partly due to insufficient planting density, and future studies
should experiment with higher planting densities. Clearly, North American ecological and
economic situations are different than those of the Australian wheat belt, and analogous
92

longitudinal studies in diverse ecological and socioeconomic contexts are needed. Despite this,
our results are promising as they provide some of the first field-based evidence that growth and
yield potential of novel perennial cereal accessions can be substantial relative to annual
counterparts. Importantly, our results were consistent over multiple years, indicating that firstand second-year plants can provide earlier biomass growth, equivalent or higher biomass, and
approximately 50 – 75% of annual cereal yield potential. Improved nutrient content (Murphy et
al., 2009; Fule et al., 2005) may also compensate for some of the reduced yield of perennial
cereals.
Perennial wheat and rye thus appear to be nearing the threshold of being viable crops.
Further breeding should focus on increasing biomass as well as grain yield, and selecting for
larger kernel size will likely be an important aspect of improving grain yields. The consistency of
yield and yield components over multiple ages and years indicates that perennial wheat and rye
are reaching the point of reliable perenniality and that perennial rye, in particular, could be
grown as a grain crop in marginal areas.

Acknowledgements

Acknowledgements are due to Janet Lewis, Lee Siler, and Randy Laurenz for assistance
with planting and harvesting and advice about accession selection; to John Green and the KBS
Long Term Ecological Research Site team for assistance with field operations; to Arianna Pikus
for assistance with field measurements; to Sasha Kravchenko for help with statistical analysis
and to Danielle Zoellner-Kelly for help with editing and preparation.

93

APPENDIX

94

Figures and Tables for Chapter 2

95

Figure 1. Precipitation recorded in 2009 and 2010 growing seasons. Data were taken at
Kellogg Biological Station, Hickory Corners, MI. 2009 covers the Oct 2008-Sept 2009 growing
season and 2010 covers Oct 2009-Sept 2010.

96

Figure 2. Seed yield for Triticum spp. in a two year agronomy study. Accessions in the study
include two annual checks and four perennial test lines. Data were taken from a 2009-2010 field
study at Kellogg Biological Station, Hickory Corners, MI. Means and standard errors for each
accession in 2009 and 2010 are shown separately.

97

Figure 3. Seed yield for Secale spp. in a two year agronomy study. Accessions in the study
include one annual check and one perennial test line. Data were taken from a 2009-2010 field
study at Kellogg Biological Station, Hickory Corners, MI. Means and standard errors for each
accession in 2009 and 2010 are shown separately.

98

Figure 4. Kernel mass for one and two year old perennial wheat and perennial rye. Data
include four perennial wheat lines (Triticum aestivum x Thinopyrum elongatum) and one
perennial rye (Secale cereale x montanum), including one- and two- year old plants, in a 2010
field study at Kellogg Biological Station, Hickory Corners, MI. Means and standard errors for
one-year old (1 y) and two- year old plants (2 y) are shown separately.

99

Table 1. Accessions and planting year / age in 2008-2010 agronomy study. Perennial
accessions are indicated with italics.

Planting
year

Measurement
year

Species

Accession

Plant age at
measurement

2009

2009
2009
2009
2009
2009
2009
2009
2009
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010

Annual wheat
Annual wheat
Perennial wheat
Perennial wheat
Perennial wheat
Perennial wheat
Annual rye
Perennial rye
Perennial wheat
Perennial wheat
Perennial wheat
Perennial wheat
Perennial rye
Annual wheat
Annual wheat
Perennial wheat
Perennial wheat
Perennial wheat
Perennial wheat
Annual rye
Perennial rye

Frankenmuth
Pioneer 25R37
P3
P11
P15
P19
Wheeler
Rival
P3
P11
P15
P19
Rival
Frankenmuth
Pioneer 25R37
P3
P11
P15
P19
Wheeler
Rival

1
1
1
1
1
1
1
1
2
2
2
2
2
1
1
1
1
1
1
1
1

2010

100

Table 2. Dates for field operations in 2008-2010 agronomy study. Field operations carried out
at Kellogg Biological Station, southwest Michigan, USA.

Field operation
Tilled
Fertilized
Soil finished
Planted
Irrigation
Weeding 1
Weeding 2
Weeding 3
Weeding 4
Post-harvest tillage

2009
13 Oct. 2008
22 Oct. 2008
24 Oct. 2008
25 Oct. 2008
3 Aug. 2009
15 – 20 May 2009
7 – 14 June 2009
29 June – 7 July 2009
30 Aug – 5 Sept. 2009
7 Oct 2009

101

2010
7 Oct. 2009
20 Oct. 2009
19 Oct. 2009
9 Nov. 2009
5 Sept. 2010
1 – 5 May 2010
1 – 7 June 2010
25 June – 2 July 2010
28 Aug – 1 Sept. 2010
23 Oct 2010

Table 3. Means, Fisher’s LSD values and F-values for yield and yield components, averaged over 2009 and 2010. Means in the
same column followed by a different letter differ significantly according to LSD post-hoc tests. Perennial accessions are indicated with
-1
italics. Threshed yield and biomass are expressed in Mg ha , kernels per tiller are unitless, kernel mass is expressed in mg,
-2

threshability and harvest index in %, and tiller number in tillers m .

Accession
Pioneer 25R37
wheat
Frankenmuth
wheat
P3 wheat
P11 wheat
P15 wheat
P19 wheat
Wheeler rye
Rival rye
LSD
ANOVA
results+
Accession
Year
Accession x
year

Threshed
yield
2.41a

Threshability

Kernels/tiller

Kernel mass

Biomass

83.8a

Tiller
number
60.1a

21.4a

36.7a

4.43ad

Harvest
index
54.4a

2.94b

85.6b

57.4a

30.5b

32.8b

6.07b

48.5b

0.99c
1.34d
1.36d
1.46d
1.83e
1.33d
0.28

67.0c
71.7d
69.2e
75.4f
85.2g
68.1h
0.1

43.8b
42.1b
40.9b
44.4b
40.7b
51.8a
11.4

18.9a
26.5b
30.1b
29.6b
21.8a
27.3b
4.8

24.6c
23.3cd
22.1d
22.3cd
34.0b
26.4e
2.1

3.94c
3.44c
4.40acd
4.29d
5.17d
5.02ad
0.87

26.6c
39.0d
16.8ef
18.4ef
18.7e
14.3f
4.3

31.12***
1.65
2.25*

15.19***
-----------

3.88**
2.08
3.15**

5.72***
18.62*
2.38*

113.6***
111.5***
18.09***

6.99***
6.98
3.30**

41.89**
10.33*
1.85

+F-values are presented for each independent variable and interaction term
* Denotes that treatment effects are significant at P < 0.05
** Denotes that treatment effects are significant at P < 0.01

102

Table 4. Means, Fisher’s LSD values, and F-values for yield and yield components in one year old (1 y) and two year old (2 y)
perennial cereal plants, in 2010. Means in the same column followed by a different letter differ significantly according to LSD post-1
hoc tests. Threshed yield and biomass are expressed in Mg ha , kernels per tiller are unitless, threshability and harvest index are
-2

expressed in %, and tiller number in tillers m .

Threshed yield
1y
2y
0.91c
1.29
1.38d
1.35
1.63d
1.44
1.67d
1.76
1.29de
1.44
0.36
0.52

Accession
P3 wheat
P11 wheat
P15 wheat
P19 wheat
Rival rye
LSD
ANOVA
results+
Accession
Plant age
Plant age x accession

5.08**
0.23
1.66

Tiller number
1y
2y
32.4b
43.4ab
41.4bc
36.6a
42.2bc
35.8a
44.8c
41.6a
37.6b
58.4b
10.8
15.3

Kernels/tiller
1y
2y
22.8b
26.8a
31.1ac
34.6b
36.5a
37.3b
32.4ac
41.7b
28.0c
24.3a
6.5
7.7

Biomass
1y
2y
4.29a
4.33ab
4.29a
3.70a
4.40a
5.75b
4.88a
5.31ab
6.33b
4.76ab
1.20
1.83

Harvest index
1y
2y
0.30c
0.23a
0.31c
0.37b
0.33cd 0.28ab
0.36d
0.32ab
0.23e
0.27ab
0.04
0.09

1.82
0.50
3.66*

12.00***
3.40
2.01

3.08*
0.01
2.14

13.07***
1.58
6.42***

+F-values are presented for each independent variable and interaction term
* Denotes that treatment effects are significant at P < 0.05
** Denotes that treatment effects are significant at P < 0.01
*** Denotes that treatment effects are significant at P < 0.001.

103

Table 5. Means, Fisher’s LSD values, and F-values for plant growth and phenology data, sampled from one year old annual
and perennial cereal plants in 2009 and 2010. Means in the same column followed by a different letter differ significantly
according to LSD post-hoc tests. Perennial accessions are indicated with italics. Emergence is expressed in stems per meter, flowering
date in terms of day of year, early season and flowering heights in cm, and regrowth as a %.

Emergence
Accession
2009
2010
Pioneer
59.1a
40.5a
25R37 wheat
Frankenmuth
76.0b
56.5b
wheat
P3 wheat
31.2c
12.6c
P11 wheat
38.0ac
13.9c
P15 wheat
31.9c
11.4c
P19 wheat
34.5c
15.1c
Wheeler rye
33.8c
32.0a
Rival rye
56.4a
9.4c
LSD
11.2
15.3
ANOVA results+
Accession
23.95***
Year
65.39**
Accession x year
1.30

Flowering date

Early season
height
2009
2010
7.9
8.0a

2009
157.6a

2010
157.9a

153.4b

151.2b

6.7

179.8c
180.2c
182.0d
184.4e
150.4f
162.2g
2.2

169.2cd
171.2c
170.6cd
168.4cb
151.8b
162.4e
2.4

7.8
6.9
7.7
7.0
7.5
5.9
2.3

254.28***
143.77***
22.54***

104

Regrowth

2009
49.5a

2010
67.3a

2009
---

2010
---

7.5ab

66.9b

86.9b

---

---

6.3bc
6.5bc
6.2bc
5.5c
10.1d
7.4ab
1.4

80.7c
72.3bc
71.9bc
73.5bc
143.5e
142.0e
9.1

93.3b
84.9b
91.1b
90.4b
163.5c
142.9d
8.85

53.6
84.7
80.9
104.8
--116.1
23.4

0.0
1.4
4.0
0.9
--0.9
11.7

3.06**
0.01
2.71*

+F-values are presented for each independent variable and interaction term
* Denotes that treatment effects are significant at P < 0.05
** Denotes that treatment effects are significant at P < 0.01
*** Denotes that treatment effects are significant at P < 0.001.

Flowering height

230.68***
103.51***
2.47*

2.05
80.65***
2.26

Table 6. Means, Fisher’s LSD values, and F-values for plant growth and phenology data in one (1 y) and two year-old (2 y)
plants of perennial accessions in 2010. Means in the same column followed by a different letter differ significantly according to LSD
post-hoc tests. Emergence is expressed in stems per meter, flowering date in terms of day of year, early season and flowering heights
in cm, and regrowth as a %.

Emergence
Accession
P3 wheat
P11 wheat
P15 wheat
P19 wheat
Rival rye
LSD
ANOVA results+
Accession
Age
Age x accession

1y
12.6
13.9
11.4
15.1
11.4
15.3

2y
12.9ab
14.3ab
12.9ab
11.7a
16.3b
4.1
3.08**
0.86
1.46

Flowering date
1y
169.2a
171.2a
170.6a
168.4a
162.4b
4.2

2y
179.8a
180.2a
182.0a
184.4a
162.0b
4.0

Early season
height
1y
2y
6.3
11.9a
6.5
13.3a
6.2
12.07a
5.5
12.07a
7.4
18.1b
2.3
3.1

14.69***
33.62**
1.79

+F-values are presented for each independent variable and interaction term
* Denotes that treatment effects are significant at P < 0.05
** Denotes that treatment effects are significant at P < 0.01
*** Denotes that treatment effects are significant at P < 0.001
‡ Denotes that treatment effects are marginally significant at P < 0.06.

105

6.65**
61.57***
2.61‡

Flowering height
1y
93.3a
84.9a
91.1a
90.4a
142.9b
8.9

2y
103.1a
80.1b
91.8c
97.0c
179.7d
11.7

151.00***
18.01*
9.66***

Regrowth
1y
0.0a
1.4a
4.0b
0.9a
0.9a
7.7

2y
0.7
1.2
1.0
1.1
0.8
11.5
1.46
0.01
0.97

REFERENCES

106

REFERENCES

Acharya SN, Mir Z, Moyer JR, Orshinsky BR, Thomas JE. 2003. Effect of row spacing and
seeding date on forage yield and quality of perennial cereal rye (Secale cereale L.). Canadian
Journal of Plant Science 83: 363-369.
Acharya SN, Mir Z, Moyer JR. 2004. ACE-1 perennial cereal rye. Canadian Journal of Plant
Science 84: 819-821.
Bell LW, Byrne F, Ewing MA, Wade LJ. 2008. A preliminary whole-farm economic analysis of
perennial wheat in an Australian dryland farming system. Agricultural Systems 96: 166-174.
Bishop JG., Schemske DW. 1998. Variation in flowering phenology and its consequences for
lupines colonizing Mount St. Helens. Ecology 79: 534-546.
Canode CL. 1968. Influence of row space and nitrogen fertilization on grass seed production.
Agronomy Journal 60: 263-267.
Canode CL, Law AG. 1978. Influence of fertilization and residue management on grass seed
production. Agronomy Journal 70:543-546.
Chastain TG, Kiemnec GL, Cook GH, Garbacik CJ, Quebbeman BM, Crowe FJ. 1997. Residue
management strategies for Kentucky bluegrass seed production. Crop Science 37:1836-1840.
Chastain TG, Young WG. 1998. Vegetative plant development and seed production in cool season
perennial grasses. Seed Science Research 8: 295-301.
Clifton-Brown JC, Lewandowski I. 2002. Screening Miscanthus genotypes in field trials to
optimize biomass yield and quality in southern Germany. European Journal of Agronomy 16: 97110.
Cox TS, Glover JD, Van Tassel DL, Cox CM, DeHaan LR. 2006. Prospects for developing
perennial grain crops. BioScience 56: 649-659.
Crews TE. 2005. Perennial crops and endogenous nutrient supplies. Renewable Agriculture and
Food Systems 20: 25-37.
Crum JR, Collins HP. 1995. KBS soils. [Online]. Available at
www.lter.kbs.msu.edu/soil/characterization (verified 21 July 2006). W. K. Kellogg Biol. Stn.,
Long-Term Ecological Research Project, Michigan State University, Hickory Corners, MI.
DeHaan LR, Van Tassel DL, Cox TS. 2005. Perennial grain crops: a synthesis of ecology and
plant breeding. Renewable Agriculture and Food Systems 20: 5-14.

107

Emrich K, Wilde F, Miedaner T, Piepho HP. 2008. REML approach for adjusting the Fusarium
head blight rating to a phenological date in inoculated selection experiments of wheat.
Theoretical Applied Genetics 117: 65-73.
Fairey NA, Lefkovitch LP. 2001. Effect of post harvest management on seed production of
creeping red fescue, tall fescue and Kentucky bluegrass in the Peace River region of northwestern Canada. Canadian Journal of Plant Science 81: 693-701.
Füle L, Hódos-Kotvics G, Galli Z, Ács E, Heszky L. 2005. Grain quality and baking qualities of
perennial rye (cv. ‘Perenne’) of interspecific origin (Secale cereale x montanum). Cereal
Research Communications 33: 809-814.
Fulkerson RA. 1980. Seed yield responses of three grasses to post harvest stubble removal.
Canadian Journal of Plant Science 60: 841-846.
Garnier E, Cordonnier P, Guillerm J-L, Sonié, L. 1997. Specific leaf area and leaf nitrogen
concentration in annual and perennial grass species growing in Mediterranean old fields.
Oecologia 111: 490-498.
Hayes RC, Newell MT, DeHaan LR, Murphy KM, Crane S, Norton MR, Wade LJ, Newberry M,
Fahim M, Jones SS, Cox TS, Larkin PJ. 2012. Perennial cereal crops: An initial assessment of
wheat derivatives. Field Crops Research 133: 68-89.
Huebner FR, Bietz JA, Nelsen T, Bains GS, Finney PL. 1999. Soft wheat quality as related to
protein composition. Cereal Chemistry 76: 650-655.
Jackson LL, Dewald CL. 1994. Predicting evolutionary consequences of greater reproductive
effort in Tripsacum dactyloides, a perennial grass. Ecology 75: 627-641.
Jackson W, Jackson LL. 1999. Developing high seed yielding perennial polycultures as a mimic
of mid-grass prairie. In: Agriculture as a Mimic of Natural Systems pp. 1-37. Lefroy EC, Hobbs
RJ, O’Connor MH, Pate JS. Eds., Kluwer Academic Publishers, Dordrecht, Netherlands.
Lawrence T, Ashford R. 1964. Seed yield and morphological development of Russian wild
ryegrass as influenced by grazing, Canadian Journal of Plant Science 44: 311-320.
Loch DS, Rámirez Áviles L, Harvey GL. 1999. Crop management: Grasses. In Loch DS,
Ferguson JE. (eds). Forage Seed Production Vol. 2: Tropical and Subtropical Species. 159-176.
(CAB International: Oxon, UK).
Mantineo M, D’Agosta GM, Copani V, Patanè C, Cosentino SL. 2009. Biomass yield and energy
balance of three perennial crops for use in the semi-arid Mediterranean environment. Field Crops
Research 114: 204-213.
Mishra ML, Chatterjee BN. 1968. Seed production in the forage grasses Pennisetum
polystachyon and Andropogon gayanus in the Indian tropics. Tropical Grasslands 2: 51-56.
108

Mueller-Warrant GW, Caprice Rosato S. 2002. Weed control for stand duration of perennial
ryegrass seed production. II: Residue Retained. Agronomy Journal 94: 1192-1203.
Murphy KM, Hoagland LA, Reeves PG, Baik B-K, Jones SS. 2009. Nutritional and quality
characteristics expressed in 31 perennial wheat breeding lines. Renewable Agriculture and Food
Systems 24: 285-292.
Murphy KM, Lyon SR, Balow KA, Jones SS. 2010. Post-sexual cycle regrowth and grain yield
in Thinopyrum elongatum x Triticum aestivum amphiploids. Plant Breeding 129: 480-483.
Piper JK, Kulakow PA. 1994. Seed yield and biomass allocation in Sorghum bicolor and F1 and
backcross generations of S. bicolor x S. halepense hybrids. Canadian Journal of Botany 72: 468474.
Reimann-Philipp, R. 1986. Perennial spring rye as a crop alternative. Journal of Agronomy and
Crop Science 157: 281-285.
Ridley AM, White RE, Helyar KR, Morrison GR, Heng LK, Fisher R. 2001. Nitrate leaching
loss under annual and perennial pastures with and without lime on a duplex (texture contrast) soil
in humid southeastern Australia. European Journal of Soil Science 52: 237-252.
Sacks EJ, Dhanapala MP, Tao DY, Sta Cruz MT, Sallan R. 2005. Breeding for perennial growth
and fertility in an Oryza sativa x longistaminata population. Field Crops Research 95: 39-48.
Sacks EJ, Dhanapala MP, Sta Cruz MT, Sallan R. 2006. Clonal performance of perennial Oryza
sativa / O. rufipogon selections and their combining ability with O. sativa for survival, stolon
production and yield. Field Crops Research 100: 155-167.
SAS Institute. 2008. The SAS System for Windows. v. 9.2. SAS Institute, Cary, NC.
Schmer MR, Mitchell RB, Vogel KP, Schacht WH, Marx DB. 2010. Spatial and temporal effects
on switchgrass stands and yield in the Great Plains. Bioenergy Research 3: 159-171.
Shefferson RP, Roach DA. 2010. Longitudinal analysis of Plantago: adaptive benefits of
iteroparity in a short-lived, herbaceous perennial. Ecology 91: 441-447.
Smith RG, Gross KL, Robertson GP. 2008. Effects of crop diversity on agroecosystem function:
Crop yield response. Ecosystems 11: 355-366.
Torimaru T, Tomaru N. 2006. Relationships between flowering phenology, plant size, and female
reproductive output in a dioecious shrub, Ilex leucoclada (Aquifoleaceae). Canadian Journal of
Botany 84: 1860-1869.
Van Dijk, H. 2009. Ageing affects in an iteroparous plant species with a variable life span.
Annals of Botany 104: 115-124.
109

Van Tassel DL, DeHaan LR, Cox TS. 2010. Missing domesticated plant forms: can artificial
selection fill the gap? Evolutionary Applications 3: 434-452.
Ward PR, JA Palta, Waddell HA. 2011. Root and shoot growth by seedlings of annual and
perennial medic, and annual and perennial wheat. Crop Pasture Science 62: 367–373.
Weik L, Kaul H-P, Kübler E, Aufhammer W. 2002. Grain yields of perennial grain crops in pure
and mixed stands. Journal of Agronomy and Crop Science 188: 342-349.
White LM. 1985. Stand age, precipitation, and temperature effects on forage yield. Journal of
Range Management 38: 39-43.

110

CHAPTER 3.
RESOURCE ACQUISITION TRAITS IN PERENNIAL AND ANNUAL CEREALS

111

Introduction

Annual monocarpy and perennial polycarpy represent conflicting plant life history
strategies, which are commonly seen across the plant kingdom and which require radically
different patterns of resource allocation. In one, plants completely sacrifice long-term survival to
maximize current reproduction: in the other, plants must balance allocation to current
reproduction and future survival (Aragon et al., 2009; Reekie and Bazzaz, 1987a). Because
plants face tradeoffs between allocation to various life history functions (seed size vs. seed
number, reproduction vs. survival, growth vs. defence) perennials are generally expected to
show lower levels of reproductive allocation than annuals, which can afford to spend as much as
possible of the resources accumulated by mid-season, on reproduction. While some woody
perennials do show very high levels of reproductive allocation in a given year, these are usually
compensated for by lower allocation during the establishment phase.
An ongoing focus of research, however, is the extent to which perennials and annuals
differ not only in patterns of resource allocation, but resource acquisition as well (Van
Noordwijk and de Jong, 1986). As reproduction depends on the product of resource acquisition
and reproductive allocation (Houle, 1991), differences between annuals and perennials in terms
of reproductive output as well as growth and survival could be critically affected by changes in
the ability to accumulate resources. It is unclear how much of the decreased seed production and
increased investment in storage in perennials is explainable by simple shifts in allocation from
one function to another, as opposed to differences in resource acquisition ability. For example, if
shifting to perennial life history involved a decline in resource acquisition, then costs of
perenniality in terms of reduced reproduction would be greater than strict one-to-one tradeoffs

112

predicted. Conversely, if the shift to a perennial life history was associated with a decline in
photosynthetic rates and resource acquisition capability, the cost of perenniality would be bigger
than might be predicted based on one-to-one tradeoffs in carbohydrate allocation would predict.
In order to more accurately understand costs of reproduction and costs of perenniality, it would
be ideal to determine how photosynthesis and resource acquisition traits vary as a function of life
history.
Newly developed perennial cereals (DeHaan et al., 2005) provide an ideal novel system
to address these questions. Perennial cereals developed for cold temperate environments include
perennial wheat (Triticum aestivum x Thinopyrum spp.), perennial rye (Secale cereale x
montanum) and intermediate wheatgrass (Thinopyrum intermedium). All three have been
developed as potential alternative food crops for the cold temperate zone. The first two are
hybrids of annual wheat and rye with related perennial grasses, while the third is a steppe grass
closely related to wheat. These species, unusually, combine herbaceous growth form,
perenniality and relatively high reproductive allocation (Culman S. et al., unpublished data;
Jaikumar et al., 2012; Murphy et al., 2010), a trait combination rare in nature or agroecosystems
(Van Tassel et al., 2010). While some woody perennials, once established, (e.g. apple) can
achieve reproductive allocation levels of 80%, exceeding annual crops like wheat or barley (Xia
et al., 2009; Kemanian et al., 2007), no herbaceous perennial currently achieves these levels of
reproductive allocation. By comparing resource acquisition traits of these species to close annual
relatives (annual wheat and annual rye) we can better understand how shifts between annual and
perennial life history affect resource acquisition traits.
Costs of reproduction and perenniality can be expressed in terms of various limiting
resources: carbon, nitrogen (Wheelwright and Logan, 2004), phosphorus (Zotz and Richter,

113

2006), and others. Reekie and Bazzaz (1987b) argued that carbon is the most broadly relevant
currency to assess costs of reproduction, since plants can increase access to other nutrients by
investing carbon in roots. In this paper, we consider how perennial and annual cereals differ with
respect to their ability to accumulate carbon through photosynthesis.
Perennials might be expected to have lower photosynthetic rates (Amax) on the basis of
evolutionary theory: plants with longer lifespans should adopt more conservative strategies,
trading off rapid resource accumulation and high seed yield for increased lifespan and stress
tolerance (Grime, 1977; Garnier, 1992). While this is a plausible argument it may not be always
true: empirical comparisons of closely related perennials and annuals indicate photosynthetic
rates in perennials can be higher (Zhao et al., 2010; Gonzalez-Paleo and Ravetta, 2011b; Taylor
et al., 2010) or lower (Sobrado, 2011; Van Auken and Bush, 2011) than annual relatives.
Perennial cereals might achieve higher Amax than annual relatives if they allocated more
nitrogen to photosynthetic proteins at the expense of seeds (the ‘nitrogen allocation’ hypothesis),
if their deeper root systems allowed greater access to moisture (the ‘moisture stress’ hypothesis),
or if the presence of larger vegetative sinks removed feedback inhibition of photosynthesis (the
‘sink strength’ hypothesis: Zhao et al., 2008). As source/sink balance, nitrogen availability and
rooting depth may change with plant age, it is necessary to study different age classes of
perennial cereals to understand the effects of both species identity and age on photosynthesis and
associated traits. Furthermore, it would be desirable to measure other physiological traits which
might elucidate which of these three non-exclusive hypotheses (more nitrogen allocation to
leaves, less moisture stress, higher sink capacity) might explain higher photosynthetic rates in
perennials, if such trends are observed.
Greater nitrogen allocation to leaves, and higher sink capacity, would affect
114

photosynthetic rates primarily by affecting biochemical traits that affect photosynthesis. In other
words, they would contribute to increased photosynthetic rates through increased enzymatic
content or activity. This might be reflected in higher ribulose bisphosphate carboxylation
capacity (VC), higher electron transport capacity (J) or higher triose phosphate utilization
capacity (TPU). These three processes are all biochemical processes which limit photosynthesis
at a particular level of internal carbon dioxide (Ci) and therefore, would contribute to increasing
or decreasing photosynthesis at a given Ci. In contrast, lower moisture stress would contribute to
higher photosynthetic rates through increased stomatal opening, which would effect the degree of
stomatal limitation, and the ratio of intercellular to ambient CO2. These factors would contribute
to increased Ci within the leaf, rather than increasing photosynthesis at a given Ci.
In a three year study in southwest Michigan, we compared first- and second-year plants
of three perennial cereals (perennial wheat, intermediate wheatgrass, and perennial rye) to close
annual relatives (annual wheat in the first two cases, annual rye in the third). Mid-season Amax,
and various biochemical and hydraulic traits related to photosynthesis, were measured. We
hypothesized: 1) Perennial cereals will show higher mid-season photosynthetic rates than their
annual analogues. 2) Photosynthetic rates will be similar between one- and two-year old
perennial cereal plants. 3) The higher photosynthetic rates in perennial cereals are explained by
differences in biochemical traits rather than hydraulic traits. associated with by higher stomatal
conductance (gs) and transpiration (E) rates, less stomatal limitation (Ls), higher ratio of
intercellular to ambient carbon dioxide (Ci / Ca), higher capacity for ribulose bisphosphate
carboxylation (VC) and electron transport (J), and greater sink strength (as measured by triose
115

phosphate utilization, TPU).
Thus this study should help elucidate some plant traits that might be associated with any
photosynthetic differences that we observe, and help test various explanations proposed above
(lower moisture stress, greater sink strength or more nitrogen allocation to leaves) for why
perennial and annual species might differ. If perennials achieve higher photosynthetic rates
because deeper root systems allow them to avoid water stress (the water stress hypothesis) then
we would expect perennial cereals to show higher gs, higher E, higher Ci / Ca and lower Ls than
annual relatives. If perennial cereals achieve higher photosynthetic rates than annuals because of
greater allocation of nitrogen to photosynthetic machinery as opposed to seeds (the nitrogen
tradeoff hypothesis), or because of higher sink strength (the sink strength hypothesis) we should
see higher VC , J, and TPU in perennial cereals compared to annual relatives. If the nitrogen
tradeoff hypothesis is correct, we would also expect higher leaf protein and Rubisco content in
perennials, and lower total N allocation to seeds, but these parameters were not tested in our
study.

Materials and Methods

Site and Goals

We conducted this study at W. K. Kellogg Biological Station, located in southwest Michigan, USA, 50 km east of Lake Michigan (42° 24′ N, 85° 24′ W, elevation 288 m) on soils developed from glacial outwash deposited 12000 years ago. Soils are mainly fine to coarse loamy,

116

mixed mesic Typic Hapludalfs. The area receives ca. 890 mm of precipitation annually (based on
a 24-year average), about half as snow, and the mean annual temperature is 9.7° C. Further site
and soils descriptions for KBS are available at
http://lter.kbs.msu.edu/Data/LTER_Metadata.jsp/Dataset/KBS042.

Our study included four field experiments, to identify differences between perennial and
annual cereals, as well as between first- and second-year perennials. The first study (Experiment
1), in 2008-2010, considered two perennial species, and measured only one parameter, early- to
mid-season Amax. A second study (Experiment 2), in 2010-2012, replicated the results from
Experiment 1, added a third perennial species, and measured more parameters. We measured
early- to mid-season Amax as well as other parameters linked to photosynthesis: conductance
(gs), ratio of intercellular to ambient carbon dioxide (Ci/Ca ratio), leaf mass per area (LMA),
diurnal variation in photosynthesis (ΔA) and leaf mass ratio (LMR). We also used gas-exchange
data to infer carboxylation capacity (VC), electron transport rate (J), triose phosphate utilization
(TPU) and degree of stomatal limitation (LS). Species was the main factor of interest, with plant
age nested within species.
Three smaller, supplementary studies were included to give extra information about
photosynthetic rates in intermediate wheatgrass and perennial wheat. Experiment 3 sampled
photosynthetic rates, conductance and Ci / Ca across a population of intermediate wheatgrass
genotypes, to determine how the wheatgrass accession in Experiment 2 compared to the
Thinopyrum intermedium population as a whole. In Experiment 4, we compared second-year

117

wheatgrass photosynthetic rates to a second annual wheat cultivar (cv. ‘Caledonia’). The purpose
of this experiment was to determine whether the trend towards higher photosynthesis in
wheatgrass was maintained when it was compared to a different wheat variety. Finally, the
purpose of Experiment 5 was to determine whether the association between perennial life history
and higher photosynthetic rate held across a range of wheat x Thinopyrum amphiploid crosses. It
is possible that any differences between our perennial and annual wheat lines could be a result of
having chosen an unusual perennial cultivar. To rule out this possibility and ensure that any
association between perenniality and photosynthetic rate that we found was real, we attempted to
test whether interspecific crosses which displayed perennial regrowth differed from those which
did not, in terms of key resource acquisition traits.

Selection of Accessions
Experiment 1 included four species: perennial rye (PR-Rival), perennial wheat (PW-P19),
annual rye (AR-Wheeler), and annual wheat (AW-Frank). Experiment 2 added another species,
intermediate wheatgrass (IWG-TLC1). The perennial rye (Secale cereale x montanum cv.
‘Rival’: Peters Seed Company, OR) had demonstrated perenniality in Michigan (Jaikumar et al.,
2012). The annual rye (Secale cereale L. ‘Wheeler’: Helsel and Thomas, 1987), was a forage
cultivar developed in southern Michigan. The perennial wheat accession, ‘P-0019’, obtained
from Washington State University, represented the F5 generation of interspecific hybrids
backcrossed once to annual wheat (Thinopyrum elongatum x Triticum aestivum ‘Chinese Spring’
// T. aestivum ‘Madsen’: Murphy et al., 2010). It was chosen because it had displayed
moderately high yields, and reasonable ability to perenniate, in both Michigan and Washington
(Murphy et al., 2010; Jaikumar et al., 2012). The annual wheat accession (Triticum aestivum cv.

118

‘Frankenmuth’, PVP 8000165: Huebner et al., 1999), is used as a benchmark for yield and
quality in breeding studies at Michigan State University. Both wheatgrass and perennial wheat
were compared against annual wheat. The wheatgrass accession, TLI-C1, was a breeding
population that had undergone two cycles of selection for higher grain yield and larger seed size
at The Land Institute (Salina, KS).
The two supplementary experiments considered a broader range of accessions, with the
goal of gaining more information about variability within our perennial cereal species.
Experiment 3 included a different annual wheat variety, cv. ‘Caledonia’. This is a highly
productive variety, well adapted to Michigan, which was considered the variety ‘to beat’ for
breeding efforts in the mid 2000s (Lewis, Janet M., personal communication), and we hoped to
determine how intermediate wheatgrass photosynthetic rates compared to a more modern wheat
line as well as to the older line ‘Frankenmuth’. Finally, in Experiment 4 we considered a range of
28 breeding lines of Thinopyrum intermedium, currently under selection at the Land Institute for
various desirable agronomic traits including larger seeds and greater palatability, to determine the
normal range of mid-season photosynthetic rates for this species.

Design and Layout
Experiment 1 was laid out as a nested RCBD (n = 5 blocks), each including one 8.30 m

2

plot for each species and age-class. Each species was planted on Oct 25, 2008 at 1.5 million
-1

seeds ha about 2.5 cm deep. Following the harvest, perennial plants regrew into a second
season, while a second set of plots of each species was planted on November 7, 2009 at 1.5
-1

-1

million seeds ha . The field was fertilized in Oct 2008 with 90 kg ha N (2% dairy manure) and

119

-1

in Oct 2009 with 102 ha N (13% blood meal). Thus in 2010 the study included annuals,
second-year perennials and first-year perennials. Weeding was done by hand, four times during
the season.
Experiment 2 was laid out as a nested RCBD, (n = 6 blocks) each including one 2.30 m

2

plot for each species and age-class. Each species was planted in Oct 2010 at 1.75 million seeds
-1

ha approximately 2.5 cm deep. This change in planting density was made because we had
experienced weed pressures in Experiment 1, indicating that plant density might be inadequate.
Following the harvest perennial plants regrew into a second season, while annual plots were
tilled under and a second set of plots of each species (annuals and perennials) was planted at the
same density in November 2011. The field was fertilized in Oct 2010 and Oct 2011 with 91 kg
ha

-1

N (4% poultry manure). Thus in 2012 the study included annuals, second-year perennials,

and first-year perennials. In both studies perennial wheat and wheatgrass (if present) were
compared against annual wheat, while perennial rye was compared against annual rye: first- and
second-year plants of each perennial species were also compared with each other. Weeding was
done by hand, four times during each season. Pictures of Experiment 2 are shown in Figures 2021 (Appendix 2 to Chapter 4).
Experiment 3 sampled photosynthetic and hydraulic traits across a broad population of
Thinopyrum intermedium genotypes. As only one wheatgrass accession (TLI-C1) was used in
this experiment, we wished to gain a sense for how this accession compared to the broader
population of Th. intermedium accessions, as this could help strengthen or qualify our
conclusions about intermediate wheatgrass as compared to wheat. In September 2010, individual
first-year plants of 28 Thinopyrum intermedium accessions were obtained from the Land Institute
120

(Salina, KS): these accessions represented a range of genetic variation and had been selected for
seed size, resistance to shattering, and other agronomic traits. Each of these accessions, including
1-3 plants, was transplanted in an unreplicated variety trial at Kellogg Biological Station, and
fertilized with 3.4 liters of 24:8:16 water soluble fertilizer solution over six weeks: the following
spring they were fertilized again with 150 kg ha

-1

N (urea). Weeding was carried out by hand,

three times during the season. A picture of Experiment 3 is shown in Figure 22 (Appendix 2 to
Chapter 4).
Experiment 4 was laid out as a RCBD split plot design with n = 4 blocks arranged east to
west. Two genotypes were used: ‘Caledonia’ annual wheat, and TLI-C1 intermediate wheatgrass
(the same population used in Experiments 1 and 2). Each species was planted on Nov 12, 2009 at
-1

the rate of 1.25 million seeds ha about 2.5 cm deep. The field was fertilized in Oct 2009 with
about 90 kg ha

-1

N in the form of poultry manure (4% N), as well as with 200 kg ha

(K2O) and 2240 kg ha

-1

-1

potash

-1

lime (CaMgO3): in October 2010 90 kg ha N was applied again, but

no further potash or lime were added. Following the first year’s harvest, annual wheat plots were
tilled under, fertilized again, and replanted, while wheatgrass plots were allowed to grow into a
second season. Thus in 2011 the experiment included annual wheat and second-year intermediate
wheatgrass.
In Experiment 5, to supplement our study of photosynthetic rates in perennial cereals, a
broad range of perennial cereal genotypes was sampled in a correlational study to determine
whether there was an association between life history and photosynthetic rate. In October 2010,
50 different amphiploid lines (obtained by various crosses between annual wheat, T. aestivum
and perennial Thinopyrum elongatum, Th. intermedium or Th. junceum) were obtained from The
121

Land Institute (Salina, KS). These were planted in unreplicated rows at Kellogg Biological
-1

Station, following tillage and fertilization with 91 kg ha N in the form of poultry manure.
These 50 amphiploid lines were derived from 20 distinct Triticum x Thinopyrum crosses, and
thus formed 20 independent observations. Weeding was done by hand, three times during the
season.

Mid Season Photosynthetic Rate
Early to mid season Amax was measured, in Experiment 1, on May 10 – 12 (first- and
second-year plants). In Experiment 2 measurements were taken on April 28 – 30, 2012 (first- and
second-year plants). At these points all four species were in late vegetative or mid-elongation
phase, with annual wheat 25 days before anthesis in 2010 and 30 days before anthesis in 2012. In
addition, Amax measurements were taken on annual wheat and wheatgrass on April 20, 2012 (40
days before annual wheat anthesis). As wheatgrass is the most vigorously perennial of the three
perennial cereals, and was not represented in Experiment 1, we wished to have additional data
for this annual vs. perennial species comparison.
Photosynthesis was measured on flag leaves of three visibly green, apparently healthy
plants per plot, using a Licor LI-6400, over a period between 8:30 and 12:00 a.m., sequentially
by block. We measured photosynthetic rates in the morning as this is expected to be the time of
greatest photosynthetic activity. Measurement took place at ambient CO2 (394 ppm), ambient air
-2 -1

temperature, and saturating PAR (1200 µmol m s ).
To determine whether photosynthetic differences were maintained throughout the late
spring, all plants were sampled again (only one subsample per plot) during spring 2012.
122

Measurements were taken when annual wheat was 16 days before anthesis, on May 12 – 15.

Diurnal Variation in Photosynthesis
In spring 2011, additional measurements were taken in both morning and afternoon, on
plants of all five species, to determine whether the species showed a similar degree of diurnal
variation in photosynthesis. Morning measurements were taken between 8:30 a.m. and 12 noon
on May 6-9, 2011; afternoon measurements were taken over the same time period, between 2:30
and 4:00 in the afternoon. If perennial species showed less of a decline over the course of the
day, this could be suggestive of either less moisture stress, or less feedback inhibition due to
greater sink strength. Temperature, irradiance and humidity settings were as specified above.

Conductance, Transpiration and Ci / Ca
At the May 12 – 15 sampling (16 days before annual wheat anthesis) we also measured
intercellular carbon dioxide concentration (Ci) using the LI-6400 XT, and calculated the ratio of
ambient and intercellular carbon dioxide (Ci / Ca). Stomatal conductance and transpiration were
measured on each sampled plant, on May 22-24, with the LI-6400 XT. The Ci / Ca ratio is a
measure of the degree of CO2 depletion within the leaf, and is associated with water use
efficiency, with lower values of the ratio corresponding to greater water use efficiency (Ehrlinger
and Cerling, 1995).

Triose Phosphate Utilization, Carboxylation, Electron Transport and Stomatal Limitation
To gain more information about biochemical and hydraulic limitations of photosynthesis,
123

A / Ci curves (i.e. relationships showing how light saturated photosynthetic rate, A, changes with
intercellular CO2 concentration, Ci) were generated for each species and age class through
intensive field measurements. From these curves, five parameters were estimated: TPU (triose
phosphate utilization, an estimate of CO2 – saturated photosynthesis), initial ribulose
bisphosphate carboxylation rate (VC, indicative of the maximal capacity of Rubisco to catalyze
carboxylation), electron transport rate (J, indicative of the ability to regenerate ribulose
bisphosphate), carbon compensation point (x0) and degree of stomatal limitation (Ls).
TPU capacity sets the upper limit of photosynthesis, and measures the rate at which
carbohydrates can be phosphorylated and exported from chloroplasts. It is affected by
temperature and phosphate availability, but also by sink strength, as the strength of carbohydrate
sinks throughout the plant affects rates of carbohydrate export from leaves. In our study, it is
unlikely that phosphate availability was involved, since test analyses of the soil at our site
indicated ample soil phosphorus. VC depends on both the quantity and activation state of the key
carboxylation enzyme Rubisco (ribulose 1,5-bisphosphate carboxylase-oxygenase), as well as on
*

Γ , its relative affinity for CO2 relative to O2 (though the relative affinity generally varies little
across species). VC is approximated by the initial slope of the A / Ci curve at low values of Ci,
and can alternatively be inferred by curve-fitting and optimization of sums of squares. Inference
of VC from the A / Ci curve has been found to closely correlate with estimates of VC from
chlorophyll fluorescence (Manter and Kerrigan, 2004). J is the electron transport rate, which
controls the rate at which reductant (NADPH) can be formed and the substrate for carboxylation

124

(RuBP, ribulose 1,5-bisphosphate) can be regenerated. All three parameters are temperature
dependent and must be standardized to a single temperature to allow comparison across species
and age-classes.
To estimate these parameters, A at a range of 12 to 14 Ca concentrations from 0-1600
ppm was measured between 8:30 a.m. and 12:30 p.m. on May 10 – 20, 2012, using the LI-6400
XT under temperature, humidity and irradiance conditions specified above. Ca concentrations
used for this experiment were 0, 50, 100, 150, 200, 250, 300, 400 (twice), 600, 800, 1200, 1400,
and 1600 ppm. TPU, VC and J were calculated using the curve-fitting method of Sharkey et al.
(2007), which models photosynthesis as a piecewise function sequentially limited by ribulose
bisphosphate (RuBP) carboxylation, RuBP regeneration controlled by electron transport (both
represented as saturating Monod functions) and triose phosphate utilization (represented as a
horizontal asymptote). Formulas for the three portions of the A / Ci curve, as well as a more
detailed explanation of the procedure, is given in Appendix 3 following this chapter. All three
biochemical parameters are reported below. Stomatal limitation was calculated as follows:
Ls = 1 – Ao / Ai
where Ao corresponds to A at Ca = 400, and Ai corresponds to A at Ci = 400, determined from
the A / Ci curve through linear interpolation (Farquhar and Sharkey, 1982). Degree of stomatal
limitation indicates how much of the limitation of A is due to resistance between the air and the
cell interior, rL. While rL technically represents leaf resistance as a whole (comprising both
boundary-layer and stomatal resistance), it can be used as a rough proxy for stomatal resistance,
which is expected to be much greater than boundary layer resistance. If perennial cereals show
125

less stomatal limitation than annuals, this could indicate that differences in photosynthetic rates
are due to differences in access to soil water and subsequent stomatal closure.
Carbon compensation point (x0) was defined as the level of Ci for which A = 0 (i.e. the
level at which respiration + photorespiration cancelled out photosynthesis) and was estimated by
linear interpolation.

LMR and LMA
LMR (leaf dry mass / total dry mass) was measured by clipping four plants of each
o

species at anthesis (in 2011), drying at 60 C to constant weight, separating into heads, leaves,
and stems, and then weighing each tissue. It is a useful parameter in this study since it resource
acquisition rates depend not merely on the rate of photosynthesis per unit leaf area, but also on
the total size of the leaf mass pool, i.e. how much of the total plant tissue is allocated to leaves.
LMA (leaf mass per area) was measured by clipping four flag leaves of each species (on June 4,
o

2011), scanning them through a Licor LI-3100 leaf area scanner, drying at 60 C to constant
weight, and then weighing the leaves.

Experiment 3: Characterizing the Intermediate Wheatgrass Population
On May 12 – 14, 2011, between 8:30 – 11:30 a.m. we measured Amax on flag leaves of
each plant in the variety trial. To better understand how TLC-1 compared to the broader
intermediate wheatgrass gene pool, we also measured Amax on the first year wheatgrass plants
growing in Experiment 2 (n = 6). On May 17 – 18 2011, we measured gs, E and Ci / Ca on 13

126

randomly selected genotypes from the variety trial and used these values to characterize the
population.

Experiments 4 and 5
In Experiment 4, Amax was measured on May 11, 2011, following the above procedure. In
Experiment 5, on May 12 – 14, 2011, approximately at mid-boot stage (Zadoks rating = 45),
Amax was measured on each line following the above specified procedure. In September 2011,
following harvest, post sexual cycle regrowth (PSCR) was estimated on each line by visual
inspection. Eight families showed no regrowth in any of the constituent lines, and were
considered non-perennial. Eight other families included at least one line with more than 50%
plants regrowing: we selected only the strongest regrowing lines within these eight families (with
more than 50% regrowth), and averaged Amax for the strongly regrowing lines within each
family. Amax was then compared between the strongly perennial and non-perennial families.

Statistical Analysis
The model for all parameters in Experiments 1 and 2 was a nested mixed-model ANOVA with
species as a fixed factor, block as a random factor, and plant age as a nested, fixed factor within
the perennial species:
Amax = µ + Block + Species + Age (Species).
If an age effect existed, individual comparisons of first- and second-year plants were made using
paired t-tests to determine which of the three species showed an age effect. Where no age effect
was found, measurements from first- and second-year plants were averaged. For parameters
127

measured in 2011, all plants were in their first year, so age was not a factor. Analyses were run
using PROC MIXED in SAS 9.2 (SAS Institute: Cary, NC). Species comparisons (e.g. first-year
wheatgrass vs. annual wheat or second-year wheatgrass vs. annual wheat) were made by paired ttests following the Bonferroni correction for k multiple comparisons. Since our experiments
included two or three perennial species and two age classes, we used k = 4 in Experiment 1 and k
= 6 in Experiment 2. For purposes of fitting the photosynthetic model of Sharkey et al. (2007), to
determine TPU, VC and J, mesophyll conductance (gm) was constrained between 0 and 10, day
-2 -1

respiration (Rd) was constrained between 0 and -3.00 µmol m s , and the value of each
o

parameter was standardized to 25 C, based on the measured temperature on that day and
exitsting literature on temperature response curves. Parameters were estimated so as to minimize
the total sum of squares for the model, and the single data point closest to Ci = 250 ppm was
omitted from the model (as it was assumed to be the zone of transition between Rubisco and
electron transport limitation).
A retrospective power analysis for Experiment 2 was made, using the following
assumptions: 1) observations for each species and age class were normally distributed with a
coefficient of variation of 8% (based on average coefficients of variation across the
photosynthetic and biochemical parameters); 2) half this variation could be removed through
blocking, which was suggested by inspection of the data; 3) comparisons were made using the
Bonferroni procedure with six comparisons. The power analysis depicted in Figure 5 indicates
that a 10% difference between treatments would be detectable with only 7% probability, but a
17.5% difference with 50% probability, a 21.2% difference with 80% probability, and a 32%
difference with 99% probability.
128

In Experiment 4, intermediate wheatgrass was compared to ‘Caledonia’ annual wheat
through a paired t-test. In Experiment 5, photosynthetic rate was compared between annual and
perennial families, using an unpaired t-test.

Results

Photosynthetic Rates (Experiment 1)
Mid-season Amax was elevated in perennial wheat and rye compared to annual relatives
(F = 25.06, p < 0.0001, df = 24), but did not differ between first- and second-year plants of the
same species (F = 0.18, p = 0.83): there was no block effect (F = 0.74, p = 0.23: Figure 6).
Perennial wheat had 18% higher Amax than annual wheat (t = 5.68, p = 0.02, n = 5). Perennial
rye had 16% higher Amax than annual rye (t = 5.02, p = 0.029, n = 5).

Photosynthetic Rates (Experiment 2)
In April, perennial wheat and perennial rye showed higher photosynthetic rate (Amax)
than annual wheat and rye respectively, while first- and second-year individuals of these species
did not differ (Table 7). In late April, perennial wheat had 21% higher Amax than annual wheat (t
= 8.35, p = 0.0024, n = 6) and perennial rye had 11% higher Amax than annual rye (t = 4.90, p =
0.027, n = 6). Similar trends were maintained in mid-May: 37% higher in perennial wheat than in
annual wheat (t = 5.47, p = 0.017, n = 6) and photosynthetic rates were 47% higher in perennial

129

rye compared to annual rye (t = 4.56, p = 0.036, n = 6).
Wheatgrass, unlike perennial wheat or perennial rye, showed pronounced differences in
photosynthetic rate between first- and second-year plants, maintained across all three sampling
dates, although both age classes still exceeded annual wheat (Table 7, Figure 7). First-year
wheatgrass showed Amax rates 29% higher than annual wheat in mid-April (t = 4.48, p = 0.039,
n = 6), 61% higher in late April (t = 6.11, p = 0.011, n = 6), and 47% higher in mid-May (t =
6.23, p = 0.01, n = 6). By contrast, second-year wheatgrass showed Amax rates only 11% higher
than annual wheat in mid-April (t = 4.68, p = 0.032, n = 6), 28% higher in late April (t = 7.23, p
= 0.05, n = 6) and 33% higher in mid-May (t = 5.29, p = 0.019, n = 6). In mid-May, wheatgrass
-2 -1

-2 -1

Amax was 25 – 28 µmol m s for first-year plants and 22 – 26 µmol m s

for second-year

plants, at the 95% confidence interval.
The carbon compensation point (x0) did not vary between species or age-classes, and was
generally reached at a Ci between between 41 and 46 ppm (Table 8).

Biochemical Traits
Representative A vs. Ci curves for individual leaves are depicted in Figures 8-10 below.
Figure 8 depicts A / Ci curves for annual wheat, and one- and two-year old wheatgrass; Figure 9
depicts annual rye and one- and two-year old perennial rye; Figure 10 depicts annual wheat, and
one- and two-year old perennial wheat. Overall, TPU capacity was higher in the perennial cereals
than in the annuals (Table 9) but showed no effect of plant age. TPU was 40% higher in
wheatgrass (t = 4.46, p = 0.04, n = 6), and 29% higher in first-year perennial wheat than in
130

annual wheat (t = 4.43, p = 0.041, n = 6). In perennial rye, TPU was 35% higher than in annual
-2 -1

rye (t = 4.66, p = 0.033, n = 6). TPU capacity was approximately 14 – 30 µmol m s for
-2 -1

wheatgrass, approximately 16 – 28 µmol m s

for first-year perennial wheat, and

-2 -1

approximately 22 – 28 µmol m s for perennial rye.
Initial carboxylation rate (VC) was affected by both species and, in wheatgrass, by plant
age (Table 9). VC was 31% higher in first year wheatgrass (t = 7.71, p = 0.004, n = 6), 17%
higher in second-year wheatgrass (t = 5.26, p = 0.02, n = 6), and 24% higher in perennial wheat
than in annual wheat (t = 5.82, p = 0.013, n = 6): however, VC did not differ between annual and
perennial rye (t = 2.32, p = 0.41, n = 6). Electron transport rate (J) was higher in perennial
cereals than annuals, and also showed age effects (Table 9). J was 53% higher in first year
wheatgrass (t = 4.84, p = 0.028, n = 6) and 24% higher in second-year wheatgrass (t = 5.11, p =
0.022, n = 6) compared to annual wheat, while J was 37% higher in perennial rye compared to
annual rye (t = 4.44, p = 0.04, n = 6). While first year perennial wheat showed 41% higher J
than annual wheat (t = 5.36, p = 0.018, n = 6), second-year perennial wheat did not differ from
annual wheat.

Hydraulic Traits
Stomatal conductance (gs) did not differ between annual and perennial rye, or between
annual and perennial wheat (Table 8). However, while second-year wheatgrass did not differ
from annual wheat, first-year wheatgrass had 110% higher stomatal conductance (t = 4.47, p =
0.04, n = 6). Transpiration rates were 42% higher in first year wheatgrass than in annual wheat (t
131

= 4.34, p = 0.044, n = 6), while second-year wheatgrass was no higher than annual wheat:
perennial and annual wheat did not differ, nor did perennial and annual rye. Conductance values
-2 -1

on May 20 ranged from approximately 300 mmol m s

for perennial rye, to approximately

-2 -1

850 mmol m s for first-year wheatgrass. The ratio of intercellular to ambient [CO2], Ci /Ca,
did not vary between species or age classes, and Ci was generally in the range of 70-85% of Ca.
Stomatal limitation (LS) did not vary between species and was generally between 5% and 20%.

Diurnal Variation in Photosynthesis
In 2011, diurnal variation in photosynthetic rate (change between morning and midafternoon) varied between species, but showed no effect of perennial vs. annual life history. All
species showed a 13 – 30% decline in photosynthetic rates in the afternoon relative to morning.
Decline in photosynthetic rate was greater in annual and perennial rye was greater than in annual
wheat, perennial wheat or wheatgrass: there was no difference between annual and perennial
species (Table 10).

Morphological Traits
LMR did not vary between the species, and generally was between 25% and 40% (Table
10). LMA varied between species: wheatgrass and perennial wheat had 26% higher and 28%
higher LMA than annual wheat, respectively. Perennial and annual rye did not differ in terms of
LMA.

Experiment 3: Sampling the Intermediate Wheatgrass Population
132

Our sampling of the broader intermediate wheatgrass gene pool in spring 2011 indicated
-2 -1

that Amax for the population ranged between 19 – 32 µmol m s , while Ci / Ca was between 7478%. TLI-C1 was within an average range (65th and 63rd percentile, respectively) for the
wheatgrass population in terms of Amax and Ci / Ca, but above average for conductance and
transpiration (Table 11). This suggests that our data for TLI-C1 can be taken as indicative of the
broader population rather than an unusually productive outlier.

Experiment 4: Replicating Wheat vs. Wheatgrass Comparisons
The same trend, of higher Amax in wheatgrass compared to annual wheat, was observed
when first- and second- year intermediate wheatgrass were separately compared to a different
wheat cultivar, Triticum aestivum cv. ‘Caledonia’. Second-year intermediate wheatgrass had 31%
higher photosynthetic rates than ‘Caledonia’ annual wheat (t = 9.90, p = 0.004, df = 3: Figure
11).

Experiment 5: Post-Sexual Cycle Regrowth and Photosynthetic Rate
Among the sixteen wheat x Thinopyrum amphiploid families sampled in Experiment 3, strongly
perennial families showed 13% higher Amax at mid-boot stage than non-perennial families (t =
3.32, p = 0.004, n = 19; Figure 11.). Among perennial families, Amax was moderately correlated
2

with the degree of post-sexual cycle regrowth (r = 0.53, p = 0.004, n = 8: Figure 12.). Average
-2 -1

Amax among strongly regrowing lines was 22.7 – 27.7 μmol m s compared to 20 – 24 μmol
-2 -1

m s for non-perennial lines.
133

Discussion

Our study found substantially higher Amax at mid-season in all three perennial cereals
than their annual analogues. In perennial wheat and perennial rye these differences (17 – 20%)
were consistent between first- and second- year plants, and across calendar years. By contrast, in
wheatgrass, Amax was lower in second-year as compared to first-year plants at certain time
points, with first-year wheatgrass showing the highest Amax in our study. These trends appeared
to be fairly robust, holding true when intermediate wheatgrass was compared to a second annual
wheat cultivar, and when wheat x Thinopyrum amphiploids that showed perennial regrowth were
compared with those which did not. The wheatgrass line used in our study and compared against
annual wheat, proved to be fairly typical of the broader wheatgrass population.
These results, while surprising, are consistent with studies that have found higher Amax
in some perennial plants compared to annual relatives. The perennial rice relative Oryza
rufipogon, for example, shows Amax 25% – 57% higher than annual rice checks (Zhao et al.,
2010). Perennial Lesquerella species maintained 42% higher Amax than annual Lesquerella spp.
(Brassicaceae: Gonzalez-Paleo and Ravetta, 2011b) while perennial Panicum and Aristida spp.
have 36 – 44% higher Amax than annual P. miliaceum and A. adscensionis respectively (Poaceae:
Taylor et al., 2010). Higher Amax is also found in perennial compared to annual races of
Machaeranthera gracilis (Asteraceae: Monson and Szarek, 1981). By contrast, some
134

phylogenetically controlled annual vs. perennial contrasts show higher rates in annuals, e.g. in
Polygonum (Zangerl and Bazzaz, 1983), Malvaceae (Van Auken and Bush, 2011), and Zea
(Sobrado, 2011). Our results suggest that these three perennial species match trends found in
perennial Oryza and Lesquerella, in which higher allocation to seed in annuals may come at the
expense of lower photosynthetic rates.
At certain time points, older wheatgrass plants had lower Amax and gs than younger ones,
while older perennial wheat plants had lower TPU than younger plants. Woody perennials
sometimes show higher Amax during the seedling establishment phase compared to older plants.
For example, current-year seedlings of giant sequoia had 150% higher maximum Amax, and
higher gs, than two-year-old saplings (Grulke and Miller, 1994): similar results were also found
in black cherry (Prunus serotina: Fredricksen et al., 1996), although many deciduous trees show
opposite trends (Rijkers et al., 2000; Steppe et al., 2011). To the best of our knowledge, changes
in physiology with whole-plant aging have rarely been studied in grasses or other herbaceous
perennials, although Urtica dioica shows declines in some photosynthetic parameters between 4month and 12-month plants (Oñate and Munné-Bosch, 2009) and sugarcane (Amaya et al., 1995)
also shows photosynthetic declines with age. Our study provides evidence that such age-related
effects exist, at certain time points, in a perennial grass (intermediate wheatgrass). Some authors
have suggested evolutionary explanations for the frequently observed higher Amax in seedlings.
Very young plants face hazardous environments and lower probabilities of survival: therefore, a
more risk-taking strategy, associated with morphological differences in leaf traits resulting in
higher Amax, could be selectively advantageous (Bond, 2000). Further research should

135

investigate whether changes in leaf composition and morphology, consistent with this
hypothesis, explain age-related physiological changes in wheatgrass.
Numerous hypotheses might explain the observed elevated Amax in our perennial species.
First, perennial cereals might, for example, invest more in roots than in annuals, allowing them
greater access to water, thereby leading to less stomatal closure. Secondly, perennial cereals
might maintain higher Amax due to greater vegetative sink strength (Zhao et al., 2008). Buildup
of carbohydrates in leaf tissue can lead to reductions in photosynthetic rate (Flore and Layne,
1999; McCormick et al., 2006) and greater sink capacity may remove such feedback inhibition.
In support of this claim, selection for high yield led to lowered Amax during the vegetative
period in annual but not in perennial Lesquerella, which could reflect the effect of lower
vegetative sink strength (Gonzalez-Paleo and Ravetta, 2011a). Perennial cereals may, finally,
allocate more nitrogen to leaves as opposed to seeds, resulting in higher concentration of key
photosynthetic leaf proteins; for example, perennial Lesquerella mendocina has higher N
allocation to leaves than the annual L. fendleri (Ploschuk et al., 2005). This would be an
adaptation which might optimize seed production over the course of a multi-year lifespan, but
result in lower seed production on an annual basis. If this is true, higher Amax in these perennial
cereals would reflect underlying reproduction vs. survival tradeoffs based around allocation of
nitrogen, rather than carbon (McGinley and Charnov, 1988).
Our study failed to find annual vs. perennial differences in stomatal limitation, or Ci / Ca,
indicating that differences in access to soil moisture is unlikely to explain the higher Amax in
these perennial species. Stomatal conductance and transpiration were much higher in the younger

136

wheatgrass plans than in annual wheat, but similar trends were not found in two-year old
wheatgrass or in either of the other perennial species. Thus while increased access to soil water,
and subsequent increases in gs may explain part of the observed differences in photosynthesis in
young wheatgrass plants, other factors are most likely involved. In particular, the lack of species
differences in Ci / Ca suggests that rather than maintaining higher Ci through avoiding stomatal
closure, these three perennials may be more efficient photosynthesizers than their annual
relatives at a given Ci. The low values for stomatal limitation indicate that 80 – 95% of the
limitation on photosynthetic rate is due to biochemical factors rather than to stomatal resistance,
and is an indication that these plants, under the conditions of our study, were not water stressed.
This is a further indication that the photosynthetic increases in perennial cereals are not primarily
due to improved water relations.
We did find evidence that coordinated increase in all three biochemical determinants of
photosynthesis contribute to elevated photosynthetic rates in the perennials. In particular,
increases in Amax are at a similar scale to increases in TPU, indicating that total carbon-saturated
photosynthetic capacity, and photosynthetic rate under ambient conditions in these species, are
fairly closely coupled. It is unclear, however, whether the nitrogen tradeoff hypothesis or the sink
strength hypothesis (or both, as these are non exclusive) might explain the coordinated increase
in biochemical determinants of photosynthesis in our perennial species.
Concurrent study of photosynthetic responses to source-sink manipulation (see Chapter 5
of this dissertation) shows that perennial wheat and rye do appear to show some degree of sink
regulation of photosynthesis. Furthermore, ongoing research at Kellogg Biological Station
(Sprunger, C. unpublished data) indicates that wheatgrass shows much more root growth and

137

more total biomass production on a yearly basis than annual wheat, indicating that they have
substantially larger vegetative sinks. It is possible, therefore, that increased sink strength could
be a factor. Since these species, under ambient [CO2], do not appear to be experiencing TPU
limitation, it is likely that increased sink strength contributes to higher Amax indirectly through
changes in Rubisco activation, electron transport rate or stomatal opening (Pieters et al., 2001). It
is possible, in theory, that deeper root systems might allow the perennials increased access to soil
phosphorus, which could overcome TPU limitation due to inadequate internal phosphate pools.
We did not, in this study, explicitly test the third hypothesis explaining higher
photosynthetic rates in perennials (i.e. greater allocation to key photosynthetic enzymes and
proteins). We did find evidence for higher activity of key enzymatic processes that limit
photosynthesis. Greater capacity for ribulose bisphosphate carboxylation and electron transport
was found in the perennial species, with these trends more strongly pronounced in younger
wheatgrass plants. This is consistent with the idea that they allocate more nitrogen to
photosynthetic machinery and less to seeds. In support of this idea, ongoing research at KBS
(Sprunger, C. unpublished data) has found that intermediate wheatgrass shows much less
nitrogen leaching than annual wheat, which is consistent with the wheatgrass plants taking up
nitrogen more efficiently than annual wheat. Perennial wheat appears to have approximately 50%
more protein per unit seed mass than annual wheat (Lewis, J. A. unpublished data; Murphy et al.,
2009) while intermediate wheatgrass has approximately 75% more protein (Becker et al., 2007).
At the same time, first- and second-year perennial wheat (of the same cultivar in this study)
achieves only around 50% of annual wheat grain yields (see Chapter 1 of this study) while
intermediate wheatgrass, in its second year, has only demonstrated an ability to produce
approximately 33% of annual wheat yields (Culman et al., unpublished data). Thus both
138

perennial wheat and intermediate wheatgrass appear to be absorbing large quantities of nitrogen
but allocating less nitrogen to seed production than annual wheat (40% less in the case of
wheatgrass, 25% less in the case of perennial wheat). This suggests that perennial wheat and
intermediate wheatgrass may, in fact, be shifting their nitrogen budgets to allocate more nitrogen
to leaf proteins, as the nitrogen tradeoff hypothesis would predict (and as observed in
Lesquerella).
However, as yet we have no direct evidence for the hypothesis of shifting nitrogen
allocation to leaves. The observed higher carboxylation and electron transport rates are also
consistent with the sink strength hypothesis, as higher sink strength is known to contribute to
increases in Rubisco activity and electron transport. Increases in carboxylation and electron
transport are not necessarily linked to higher concentrations of key photosynthetic proteins, as
they may reflect changes in the activation state of photosynthetic enzymes such as Rubisco,
rather than in their amount.
We thus find no support for the water-stress hypothesis, and indications that higher
photosynthetic rates in perennial cereals are due to biochemical rather than hydraulic factors.
Rather than increased access to soil moisture, perennial cereals maintain higher photosynthetic
rates through coordinated increases in carboxylation rates, electron transport and triose
phosphate utilization. As yet, it is unclear whether higher sink capacity, increased access to soil
nitrogen and phosphorus, shifting allocation of nitrogen budgets, or all three factors are involved
in explaining higher photosynthetic rates in perennial cereals. Further research is needed to
determine whether perennial and annual cereals show differences in photosynthetic protein
levels, as well as in total nitrogen allocation to seeds and leaves, in order to test the nitrogen
tradeoff hypothesis.

139

Conclusion

Our study provides new information about an unusual plant growth form, and suggests
that high-yielding herbaceous perennials may balance multi-year survival with high reproductive
effort through maintaining higher resource acquisition rates. Elevated Amax appears to be
associated with higher stomatal conductance (in one species), and with a coordinated increase in
all three biochemical determinants of photosynthetic rate. Perennials show higher carboxylation
capacity and electron transport rates, and with higher TPU capacity, possibly reflecting higher
sink strength as well as greater allocation of nitrogen to leaf proteins at the expense of seeds.
This could be part of a strategy to optimize seed production over the long term, through
increasing the ability to accumulate carbohydrate reserves, at the cost of lower seed production
in the first season. Further research is necessary to identify factors contributing to elevated Amax
in our three perennial cereal species, as well as in perennial Oryza, Lesquerella, Aristida and
other species which maintain higher Amax than close annual relatives, and to understand in more
detail the role that aging plays in photosynthetic physiology of herbaceous plants.

Acknowledgements

Acknowledgements are due to Janet Lewis for assistance with planting and cultivar
selection; to Kevin Murphy, Stephen Jones and Lee DeHaan for providing seed materials; to
John Green, Steve Culman and Mark Freeman for field management; to Bert Cregg for planning

140

and advice; to Wayne Loescher for equipment; and to Deborah Roach, Jeannine Cavender-Bares
and Christian Körner for helpful comments.

141

APPENDICES

142

APPENDIX A. Figures and Tables for Chapter 3

143

Figure 5. Statistical power (1 – β) as a function of underlying effect size for photosynthetic
parameters in 2012 study. Power calculations make the assumptions that the coefficient of
variation for each species / age combination is approximately 8%, that half of this variation can
be removed through blocking, that observations are normally distributed and that comparisons
are made using the Bonferroni procedure for six comparisons. For interpretation of the references
to color in this and all other figures, the reader is referred to the electronic version of this
dissertation.

144

Figure 6. Photosynthetic rates (Amax) for first year and second year perennial wheat and
perennial rye plants, compared to annual relatives (wheat and rye) in 2010. Data were taken
from a 2008-2010 study of perennial cereal physiology at Kellogg Biological Station (Hickory
Corners, MI). The symbol ‘*’ represents significant differences ( p < 0.05) following the
Bonferroni correction for 4 comparisons. Red represents perennial rye, pink annual rye, dark
green perennial wheat, light green annual wheat.

145

Figure 7. Photosynthetic rates (Amax) measured at three time points in spring 2012, on
annual wheat, one year old wheatgrass (IWG-1), and two year old wheatgrass (IWG-2).
Data were taken from a 2010-2012 study of perennial cereal physiology at Kellogg Biological
Station (Hickory Corners, MI). The symbols ‘*’ and ‘**’ represent significant differences ( p <
0.05 and p < 0.01 respectively) following the Bonferroni correction for 6 comparisons. Green
represents annual wheat, light blue represents first-year wheatgrass, and dark blue second-year
wheatgrass.

146

Figure 8. Photosynthetic rate (A) as a function of intercellular CO2 (Ci) in annual wheat, one year old wheatgrass (Thinopyrum
intermedium: IWG – 1 y) and two year old wheatgrass (IWG – 2 y). Data are taken from a 2012 study of ecophysiological
differences in annual and perennial cereals at Kellogg Biological Station (Hickory Corners, MI). Each of the three curves is based on
measurements from a single leaf, all within one block of the experiment.

147

Figure 9. Photosynthetic rate (A) as a function of intercellular CO2 (Ci) in annual rye, one year old perennial rye (Secale
cereale x montanum: 1 y) and two year old perennial rye (2 y). Data are taken from a 2012 study of ecophysiological differences in
annual and perennial cereals at Kellogg Biological Station (Hickory Corners, MI). Each of the three curves is based on measurements
from a single leaf, all within one block of the experiment.

148

Figure 10. Photosynthetic rate (A) as a function of intercellular CO2 (Ci) in annual wheat, one year old perennial wheat
(Thinopyrum elongatum x Triticum aestivum: 1 y) and two year old perennial wheat (2 y). Data are taken from a 2012 study of
ecophysiological differences in annual and perennial cereals at Kellogg Biological Station (Hickory Corners, MI). Each of the three
curves is based on measurements from a single leaf, all within one block of the experiment.

149

Figure 11. Photosynthetic rate (Amax) in annual wheat x Thinopyrum amphiploids,
perennial wheat x Thinopyrum amphiploids, annual wheat cv. ‘Caledonia’, and secondyear wheatgrass (Thinopyrum intermedium). Data are taken from a spring 2011 study of
ecophysiological differences in annual and perennial cereals at Kellogg Biological Station
(Hickory Corners, MI). Light green represents Thinopyrum / wheat amphiploids with annual life
history, dark green amphiploids with a perennial life history, brown represents annual wheat and
blue represents wheatgrass.

150

Figure 12. Photosynthetic rate (Amax) in ten wheat x Thinopyrum amphiploids, as a
function of post-sexual cycle regrowth (PSCR). Data are taken from a spring 2011 study of
ecophysiological differences in annual and perennial cereals at Kellogg Biological Station
(Hickory Corners, MI).

151

Table 7. Means ( + SE) and F-values (p-values) for photosynthetic rates (A) measured at multiple time points, in spring 2012,
on two annual cereal species and on first-year (1 y) and second-year (2 y) plants of three perennial cereal species. Data were
-2 -1
taken from a 2011-2012 study. Units for Amax are µmol m s of CO2.

Species / Age

Amax (Apr 20)

Annual Wheat
Perennial Wheat (1 y)

15.82 + 1.85
---

Perennial Wheat (2 y)
Int. Wheatgrass (1 y)
Int. Wheatgrass (2 y)
Annual Rye
Perennial Rye (1 y)
Perennial Rye (2 y)
Sources of Variation
Species
Age
Block

Amax (Apr 30)

Amax (May 12)

--20.44 + 2.65**
17.60 + 1.58**
-------

17.02 + 1.78
1
22.87 + 2.46**
20.46 + 2.05**
26.53 + 1.42**
21.72 + 2.06**
17.70 + 1.97
19.51 + 1.73*
20.18 + 2.55*

22.99 + 1.03*
26.35 + 0.55**
23.85 + 0.78*
14.58 + 1.95
21.07+ 1.38*
24.02 + 1.43*

14.67 (0.003)
8.64 (0.015)
1.52 (0.064)

11.20 (0.0001)
3.71 (0.023)
1.50 (0.066)

16.84 (< 0.0001)
5.55 (0.003)
0.69 (0.25)

1

18.19 + 1.33
26.56 + 1.57*

Symbols ‘*’ and ‘**’ represent significant differences when compared to annual wheat (in the case of perennial wheat and
wheatgrass) or annual rye (in the case of perennial rye) at p = 0.05 and p = 0.01 respectively (following the Bonferroni correction for 6
comparisons).

152

Table 8. Means (+ SE) and F-values (p-values) for stomatal conductance (gs), transpiration (E), ratio of leaf intercellular to
ambient carbon dioxide (Ci / Ca), and degree of stomatal limitation (LS) measured on two annual cereal species and on first-

1

year (1 y) and second-year (2 y) plants of three perennial cereal species in spring 2012. Data were taken from a 2011-2012 study .
-2 -1

Units for gs and E are mmol m s of CO2, values for Ci / Ca and LS are given as %, and x0 is expressed in ppm.
Species / Age

1

Int. Wheatgrass (2 y)
Annual Rye
Perennial Rye (1 y)
Perennial Rye (2 y)

gs
403 + 56
465 + 42
377 + 56
2
848 + 114*
340 + 56
286 + 30
352 + 80
233 + 26

Sources of Variation
Species
Age
Block

6.11 (0.001)
13.23 (< 0.0001)
0.01 (0.99)

Annual Wheat
Perennial Wheat (1 y)
Perennial Wheat (2y)
Int. Wheatgrass (1 y)

1

E
6.12 + 0.54
7.14 + 0.89
4.99 + 0.31
8.71 + 0.72*

Ci / Ca
73.0 + 4.4
78.2 + 1.8
75.3 + 2.2
77 + 3.0

LS
14.9 + 5.0
9.4 + 2.0
19.7 + 4.3
14.3 + 3.2

x0
44.2 + 2.0
48.0 + 5.0
42.3 + 2.1
43.6 + 3.8

4.12 + 0.43
4.17 + 0.29
5.64 + 0.50
3.62 + 0.41

77.3 + 1.9
78.2 + 1.4
77.0 + 2.2
81.5 + 1.7

18.7 + 4.4
13.5 + 4.4
4.9 + 2.0
10.0 + 3.5

46.6 + 4.6
41.9 + 1.6
39.0 + 2.5
43.1 + 4.1

4.95 (0.003)
16.72 ( < 0.0001)
0.01 (0.99)

1.34 (0.28)
0.98 (0.42)
1.16 (0.12)

2.27 (0.082)
2.43 (0.083)
1.18 (0.12)

0.58 (0.68)
1.08 (0.37)
1.03 (0.15)

All parameters measured in May 2012.

2

Symbols ‘*’ and ‘**’ represent significant differences when compared to annual wheat (in the case of perennial wheat and
wheatgrass) or annual rye (in the case of perennial rye) at p = 0.05 and p = 0.01 respectively (following the Bonferroni correction for 6
comparisons).

153

Table 9. Means ( + SE) and F-values (p-values) for maximum carboxylation rate (VC), electron transport rate (J), and triose
phosphate utilization (TPU), on two annual cereal species and on first-year (1 y) and second-year (2 y) plants of three perennial
cereal species in spring 2012. All parameters are standardized to their value at 25o C based on observed values, observed temperature
-2 -1
at the time of measurement, and established temperature-response curves. Units for VC, J and TPU are µmol m s of CO2, electrons
and triose phosphate respectively.

Species / Age

VC

1

J

TPU

Annual Wheat
Perennial Wheat (1 y)
Perennial Wheat (2 y)
Int. Wheatgrass (1 y)
Int. Wheatgrass (2 y)
Annual Rye
Perennial Rye (1 y)
Perennial Rye (2 y)

102.3 + 6.2
136.0 + 5.1*
117.9 + 8.5*
139.2 + 7.9**
120.2 + 4.76*
93.2 + 8.7
114.7 + 3.8
131.5 + 6.0

191.3 + 17.9
271.0 + 24.1*
222.5 + 20.5
293.3 + 31.8*
237.5 + 24.6*
204.5 + 14.1
279.2 + 11.4*
281.8 + 11.8*

15.78 + 1.76
21.92 + 2.23*
18.68 + 2.35*
23.43 + 2.83*
20.95 + 3.51*
18.62 + 1.28
25.8 + 1.23*
24.25 + 1.53*

Sources of Variation
Species
Age
Block

10.42 ( < 0.0001)
4.36 (0.011)
1.15 (0.12)

5.66 (0.001)
3.38 (0.029)
1.26 (0.10)

4.47 (0.005)
1.60 (0.21)
1.35 (0.089)

1

Symbols ‘*’ and ‘**’ represent significant differences when compared to annual wheat (in the case of perennial wheat and
wheatgrass) or annual rye (in the case of perennial rye) at p = 0.05 and p = 0.01 respectively (following the Bonferroni correction for 6
comparisons).

154

Table 10. Means (+ SE) and F-values (p-values) for leaf mass ratio (LMR), percent decrease
in photosynthetic rate between morning and afternoon (ΔA) and leaf mass per area (LMA),
measured on two annual cereal species and on first-year plants of three perennial cereal
1
species in spring 2011. Data were taken from a 2011-2012 study at Kellogg Biological Station.
2

LMR and ΔA are expressed as %, while LMA is expressed in cm / mg.
Species / Age
Annual Wheat
Perennial Wheat

ΔA
18.2 + 3.9
43.5 + 6.6

Int. Wheatgrass
Annual Rye
Perennial Rye

39.5 + 7.7
34.8 + 1.5
33.3 + 1.7

36.5 + 5.8
0.8 + 7.8
11.0 + 5.9

LMA
5.69 + 0.08
2
7.24 + 0.24**
7.18 + 0.35*
6.74 + 0.25
7.60 + 0.65

Sources of Variation
Species
Block
1

LMR
31.0 + 2.5
25.1 + 3.3

1.71 (0.19)
0.04 (0.98)

9.16 (0.0002)
0.53 (0.30)

4.27 (0.01)
0.30 (0.38)

LMR, LMA and ΔA were measured in May 2011, thus only first year plants are present.

2

Symbols ‘*’ and ‘**’ represent significant differences when compared to annual wheat (in the
case of perennial wheat and wheatgrass) or annual rye (in the case of perennial rye) at p = 0.05
and p = 0.01 respectively (following the Bonferroni correction for 6 comparisons).

Table 11. Mean values for photosynthetic rate, stomatal conductance, transpiration, and
ratio of leaf intercellular to ambient carbon dioxide, measured in spring 2012 on a
population of 28 independent intermediate wheatgrass (Thinopyrum intermedium)
genotypes, and on the selected genotype TLC-1 which was the subject of further
1
experimental study. Data were taken from a 2011-2012 study at Kellogg Biological Station
-2 -1

-2 -

(Hickory Corners, MI). Units for Amax, gs, E, and Ci / Ca are µmol m s of CO2, mmol m s
1

-2 -1

, mmol m s , and % respectively.

Trait
Photosynthetic Rate (Amax)

Population
25.39 + 3.11

TLI-C1
28.61 + 1.29

Stomatal Conductance (gs)
Transpiration (E)
Ci / Ca

529 + 32

848 + 114

6.14 + 1.12
76.3 + 2.2

8.71 + 0.72
77.0 + 3.0

1

Photosynthetic rates were measured in spring 2011, while other parameters were measured in
spring 2012. Means are shown + standard deviation for the population, + standard error (n = 6)
for the experimental line TLI-C1.

155

APPENDIX B. Modeling the A / Ci Curve

156

C3 plant photosynthesis, as a function of intercellular carbon dioxide (Ci) has been
modeled according to several different functional forms. For example, Bethke and Drew (1992)
modeled photosynthesis at levels of Ci < 500 ppm, as a Michaelis-Menten (rectangular
hyperbola) function of Ci :
A = b (Ci – h) / (Ci – h + k )

with b, h, k constants.

By contrast, Marenco et al. (2001) and McCormick et al. (2006) modeled the A / Ci curve as a
monomolecular function:
X

A = a (1 – exp (– bCi ) – c) where exp (X) = e

and a, b, and c are constants.

The model we used, because of its explicit mechanistic justification, its extremely
widespread use in the literature (e.g. Wullschleger, 1993), and its ability to provide estimates for
multiple biochemically important parameters, conceptualizes photosynthesis as a piecewise
function defined by the minimum of three separate rate-limiting processes, which are in
operation at different ranges of internal cellular concentration (Ci: Farquhar et al., 1980). As Ci
increases, a shift from one limiting process to another takes place (see Figure 13).
Canonically, ribulose bisphosphate (RuBP) carboxylation is expected to be limiting at
low values of Ci, RuBP regeneration (which depends on the supply of reductant and therefore on
light energy absorption and electron transport rate) at intermediate values of Ci, and triose
phosphate utilization which depends on sink demand for carbohydrates as well as on supply of
free inorganic phosphate (Pi) at high values of Ci. In any given situation, one or more of these

157

limitations may be absent. Under conditions of very low light, RuBP carboxylation limitation
may be absent: under conditions of low temperature RuBP regeneration limitation may be
absent; under conditions of high sink demand, adequate free Pi and high temperature, triose
phosphate utilization (TPU) limitation is often absent (Sage and Kubien, 2007). RuBP
carboxylation and RuBP regeneration are both modeled as saturating Michaelis-Menten
functions, so even if the other two limiting processes were theoretically entirely absent, A would
face inherent limits from any of the three processes when Ci became sufficiently high (Sharkey
et al., 2007).
The rate of RuBP carboxylation can be represented as:
1. A = VC [ CC – Γ*] / [CC + KC (1 + [O2] / KO )] – Rd.

Where:
2. CC = Ci – A / gm.

Here VC represents the maximum rate of RuBP carboxylation (Long and Bernacchi, 2003),
CC = concentration of carbon dioxide within the chloroplast, Ci = concentration of carbon
dioxide within the leaf, KC is the Michaelis constant for carboxylation at Rubisco, KO is the
Michaelis constant for oxygenation at Rubisco, [O2] is the concentration of oxygen at Rubisco,
Γ* is the critical carbon dioxide concentration at which photorespiration is cancelled out by
carboxylation (von Caemmerer, 2000), Rd = light adapted respiration (i.e. excluding
photorespiration), and gm = mesophyll conductance. Generally, RuBP carboxylation, under well

158

lit and moderate temperature conditions, is expected to prevail when Ci < 200 ppm.
The rate of RuBP regeneration can be represented as:
3. A = J [ CC – Γ*] / [4CC + 8 Γ*] – Rd.

Here J = the rate of electron transport, assuming that four electrons are required per
carboxylation and oxygenation. The precise quantitative relationship between electron transport
rate and photosynthesis is somewhat uncertain, and while our study used this equation, it has
alternatively been suggested that a better form may be:
4. A = J [ CC – Γ*] / [4.5CC + 10.5 Γ*] – Rd.
Here we chose the former equation (Eq. 3).
Electron transport is critical to RuBP regeneration since reducing power in the form of
NADPH is necessary to form RuBP. As electron transport is ultimately dependent on light
absorption, light availability will affect J. RuBP regeneration limitation is usually seen at Ci
values > 300 ppm.
Finally, at very high values of Ci, TPU limitation is observed (Harley and Sharkey, 1991),
such that:
5. A = 3TPU – Rd.
Here TPU represents the rate of use of triose phosphate molecules. In these equations, VC is
-2 -1

-2 -1

expressed in terms of µmol m s of carbon dioxide molecules, J in terms of µmol m s
-2 -1

of electrons, and TPU in terms of m s of triose phosphate molecules.
Ideally, with enough data points, it should be possible to fit curves to accurately estimate

159

five parameters: VC, J, TPU, Rd and gm. However, Rd and gm can both be difficult to measure (in
the case of Rd, partly because measurement accuracy of A at very low values of Ci becomes
compromised due to chamber leaks, and partly because at extremely low levels of Ci, photorespiration interferes with the estimate of light-adapted respiration under ambient conditions). In order to increase accuracy of estimation for the other three parameters, Rd and gm can be constrained in advance, on the basis of realistic ranges determined from previous experiments or the
literature. VC, J and TPU all show strong exponential temperature dependence, so it is necessary
o

to standardize them to a given temperature, generally 25 C, to allow comparisons. This can be
done using the observed temperature, and temperature response relationships from the literature
(Harley et al., 1992; Bernacchi et al., 2001; Bernacchi et al., 2003). Alternatively, Rd could be
separately measured: either by limiting O2 to minimize photorespiration and carefully correcting
for chamber leaks, or by measuring respiration at normal, ambient carbon dioxide concentration,
and zero light. For improved estimation of other parameters, VC and J can also be separately estimated, in the first case through assays of Rubisco activity and in the second by fluorescence
measurements.
Two additional parameters can be estimated from the A vs. Ci curve: the carbon
compensation point (x0) and the degree of stomatal limitation (Ls). The carbon compensation
point is the level of Ci at which A = 0, i.e. at which gross photosynthesis exactly equals
respiration + photorespiration. This can be estimated through linear interpolation between the
nearest two measured data points to A = 0 (the method we followed). It can also be calculated on

160

the basis of Eq. 1 and Eq. 2 once VC has been calculated:

6. x0 = { Rd [ KC (1 + [O2] / KO )] + VC (Γ*)} / (VC – Rd) .

The degree of stomatal limitation is an index of how much photosynthesis is under hydraulic
limitation (the ability of carbon dioxide to diffuse to the chloroplast through the leaf boundary
layer, stomata and mesophyll) versus biochemical limitation (the ability to fix carbon dioxide at
the chloroplast). It can be calculated as follows (Farquhar and Sharkey, 1982):
7. Ls = 1 – Ao / Ai .
Here Ai is the rate of net gas exchange when Ci = 400, and Ao the rate of net gas exchange
when Ca = 400. Since Ci cannot be controlled exactly, Ai can be estimated from the curve
through linear interpolation between the nearest two data points to Ci = 400, or alternatively
calculated from Eq. 3.

161

Figure 13. Graphical depiction of theoretical A / Ci curve. Three idealized functions model the three rate-limiting processes that
determine photosynthetic rate in C3 plants. RuBP carboxylation (blue curve) represents the rate of Rubisco-mediated reaction of free
CO2 with its substrate, ribulose-1,5-bisphosphate (RuBP). RuBP regeneration (red) represents the rate of regeneration of substrate, a
function of light absorption, NADPH generation and subsequent electron transport. Finally, triose phosphate utilization (green)
represents the rate of synthesis and export of triose phosphate. Photosynthesis is defined as the minimum of these three functions.

162

REFERENCES

163

REFERENCES

Aragón CF, Méndez M, Escudero A. 2009. Survival costs of reproduction in a perennial plant:
live hard, die young. American Journal of Botany 96: 904-911.
Baker NR, Harbinson J, Kramer DM. 2007. Determining the limitations and regulation of
photosynthetic energy transduction in leaves. Plant, Cell and Environment 30: 1107-1125.
Becker R, Wagoner P, Hanners GD, Saunders RM. 1991. Compositional, nutritional and
functional evaluation of intermediate wheatgrass (Thinopyrum intermedium). Journal of Food
Processing and Preservation 15: 63-77.
Bernacchi CJ, Singsaas EL, Pimentel C, Portis AR, Long SP. 2001. Improved temperature
response functions for models of Rubisco-limited photosynthesis. Plant, Cell and Environment
24: 253-259.
Bernacchi CJ, Pimentel C, Long SP. 2003. In vivo temperature response functions of parameters
required to estimate RuBP-limited photosynthesis. Plant, Cell and Environment 26: 1419-1430.
Bethke PC, Drew MC. 1992. Stomatal and nonstomatal components to inhibition of
photosynthesis in leaves of Capsicum annuum during progressive exposure to NaCl salinity.
Plant Physiology 99: 219-226.
Bond BJ. 2000. Age-related changes in photosynthesis of woody plants. Trends in Plant Science
5: 349-353.
DeHaan LR, Van Tassel DL, Cox TS. Breeding perennial grain crops: a synthesis of ecology and
plant breeding. Renewable Agriculture and Food Systems 20: 5-14.
Ehrlinger JR, Cerling TE. 1995. Atmospheric CO2 and the ratio of intercellular to ambient CO2
levels in plants. Tree Physiology 15: 105-111.
Farquhar GD, von Caemmerer S, Berry JD. 1980. A biochemical model of photosynthetic CO2
assimilation in leaves of C3 species. Planta 149: 78-90.
Farquhar GD, Sharkey TD. 1982. Stomatal conductance and photosynthesis. Annual Review of
Plant Physiology 33: 317-345.
Flore JA, Layne DR. 1999. Photoassimilate production and distribution in cherry. HortScience
34: 1015-1019.
Fredericksen TS, Steiner KC, Skelly JM, Joyce BJ, Kolb TE, Kouterick KB, Ferdinand JA. 1996.
Diel and seasonal patterns of leaf gas exchange and xylem water potentials of
164

different sized Prunus serotina Ehrh. trees. Forest Science 42: 359-365.
Garnier E. 1992. Growth analysis of congeneric annual and perennial grass species. Journal of
Ecology 80: 665-675.
Grime JP. 1977. Evidence for the existence of three primary strategies in plants and its relevance
to ecological and evolutionary theory. American Naturalist 111: 1169-1194.
González-Paleo L, Ravetta D. 2011a. Indirect changes associated with a selection program for
increased seed yield in wild species of Lesquerella (Brassicaceae): Are we
developing a phenotype opposite to the expected ideotype? Industrial Crops and Products 34:
1372-1380.
González-Paleo L, Ravetta D. 2011b. Relationships between reproductive output, morphophysiological traits, and lifespan in Lesquerella (Brassicaceae). Industrial Crops
and Products 34: 1372-1380.
Grulke NE, Miller PR. 1994. Changes in gas exchange characteristics during the life span of
giant sequoia: implications for response to current and future concentrations of atmospheric
ozone. Tree Physiology 14: 659-668.
Harley PC, Sharkey TD. 1991. An improved model of C3 photosynthesis at high CO2: reversed
O2 sensitivity explained by lack of glycerate reentry into the chloroplast. Photosynthesis
Research 27: 169-178.
Harley PC, Thomas RB, Reynolds JF, Strain BR. 1992. Modelling photosynthesis of cotton
grown in elevated CO2. Plant, Cell and Environment 15: 271-282.
Helsel ZR, Thomas JW. 1987. Small grains for forage. Journal of Dairy Science 70: 2330-2338.
Houle D. 1991. Genetic covariance of fitness correlates: what genetic correlations are made of
and why it matters. Evolution 45: 630-648.
Huebner FR, Bietz JA, Nelsen T, Bains GS, Finney PL. 1999. Soft wheat quality as related to
protein composition. Cereal Chemistry 76: 650-655.
Jaikumar NS, Snapp SS, Murphy KM and Jones SS. 2012. An agronomic assessment of some
accessions of perennial wheat and rye. Agronomy Journal. In press.
Kemanian AR, Stöckle CO, Huggins DR, Viega LM. 2007. A simple method to estimate harvest
index in grain crops. Field Crops Research 103: 208-216.
Long SP, Bernacchi CJ. 2003. Gas exchange measurements: what can they tell us about the
underlying limits to photosynthesis? Procedures and sources of error. Journal of
Experimental Botany 54: 2393-2401.
165

Manter DK, Kerrigan J. 2004. A / Ci curve analysis across a range of woody plant species:
influence of regression analysis parameters and mesophyll conductance. Journal
of Experimental Botany 55: 2581-2588.
Marenco RA, Goncalves JFD, Vieira G. 2001. Leaf gas exchange and carbohydrates in tropical
trees differing in successional status in two light environments in central Amazonia. Tree
Physiology 21: 1311-1318.
McCormick AJ, Cramer MD, Watt DA. 2006. Sink strength regulates photosynthesis in
sugarcane. New Phytologist 171: 759-770.
McGinley MA, Charnov EL. 1988. Multiple resources and the optimal balance between size and
number of offspring. Evolutionary Ecology 2: 77-84.
Monson RK, Szarek SR. 1981. Life cycle characteristics of Machaeranthera gracilis
(Compositae) in desert habitats. Oecologia 49: 50-55.
Murphy KM, Hoagland LA, Reeves PG, Baik B-K, Jones SS. 2009. Nutritional and quality
characteristics expressed in 31 perennial wheat breeding lines. Renewable Agriculture and Food
Systems 24: 285-292.
Murphy KM, Lyon SR, Balow KA, Jones SS. 2010. Post-sexual cycle regrowth and grain yield
in Thinopyrum elongatum x Triticum aestivum amphiploid. Plant Breeding 129: 480-483.
Obeso JR. 2002. The costs of reproduction in plants. New Phytologist 155: 321-348.
Oñate M, Munné-Bosch S. 2009. Influence of plant maturity, shoot reproduction and sex in the
dioecious plant Urtica dioica. Annals of Botany 104: 945-956.
Pieters AJ, Paul MJ, Lawlor DW. 2001. Low sink demand limits photosynthesis under Pi
deficiency. Journal of Experimental Botany 52: 1083-1091.
Ploschuk EL, Slafer GA, Ravetta DA. 2005. Reproductive allocation of biomass and nitrogen in
annual and perennial Lesquerella crops. Annals of Botany 96: 127-135.
Reekie EG, Bazzaz FA. 1987a. Reproductive effort in plants. 1. Carbon allocation to
reproduction American Naturalist 129: 876-896.
Reekie EG, Bazzaz FA. 1987b. Reproductive effort in plants. 2. Does carbon reflect the
allocation of other resources? American Naturalist 129: 897-906.
Rijkers T, Pons TL, Bongers F. 2000. The effect of tree height and light availability on
photosynthetic leaf traits of four neotropical species differing in shade tolerance.
Functional Ecology 14: 77-86.

166

Sage RF, Kubien DS. 2007. The temperature response of C3 and C4 photosynthesis. Plant, Cell
and Environment 30: 1086-1106.
Sage RF, Sharkey TD. 1987. The effect of temperature on the occurrence of O2 and CO2
insensitive photosynthesis in field grown plants. Plant Physiology 72: 409-413.
Sharkey TD, Bernacchi CJ, Farquhar GD, Singsaas EL. 2007. Fitting photosynthetic carbon
dioxide response curves for C3 leaves. Plant, Cell and Environment 30: 10351040.
Taylor SH, Hulme SP, Rees M, Ripley BS, Woodward FI, Osborne CP. 2010. Ecophysiological
traits in C3 and C4 grasses: a phylogenetically controlled screening experiment. New Phytologist
185: 780-791.
Sobrado MA. 2011. Leaf gas exchange and fluorescence of two teosinte species: Zea mays ssp.
parviglumis and Z. diploperennis. Journal of Tropical Agriculture 41: 91-94.
Steppe K, Niinemets U, Teskey RO. 2011. Tree size- and age-related changes in leaf physiology
and their influence on carbon gain. Tree Physiology 4: 235-253.
Van Noordwijk AJ, DeJong G. 1986. Acquisition and allocation of resources: their influence on
variation in life history tactics. American Naturalist 128: 137-142.
Van Auken OW, Bush JK. 2011. Photosynthetic rates of two species of Malvaceae, Malvaviscus
arboreae var. drummondii (wax mallow) and Abutilon theophrasti (velvetleaf). The
Southwestern Naturalist 56: 325-332.
Van Tassel DL, DeHaan LR, Cox TS. 2010. Missing domesticated plant forms: can artificial
selection fill the gap? Evolutionary Applications 3: 434-452.
Von Caemmerer S. 2000. Biochemical Models of Leaf Photosynthesis pp. 1-165. CSIRO
Publishing, Collingwood, Victoria, Australia.
Wheelwright NT, Logan BA. 2004. Previous year reproduction reduces photosynthetic capacity
and slows lifetime growth in females of a neotropical tree. Proceedings of the National Academy
of Sciences 101: 8051-8055.
Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin
T, Cornelissen JHT, Diemer M, etc. 2004. The worldwide leaf economics spectrum. Nature 428:
821-827.
Wullschleger SD. 1993. Biochemical limitations to carbon assimilation in C3 plants: a
retrospective analysis of the A / Ci curves from 109 species. Journal of Experimental Botany 44:
907-920.
167

Xia G, Cheng G, Lakso A, Goffinet M. 2009. Effects of nitrogen supply on source-sink balance
and fruit size of ‘Gala’ apple trees. Journal of the American Society for Horticultural Science
134: 126-133.
Zangerl AR, Bazzaz FA. 1983. Plasticity and genotypic variation in photosynthetic behavior of
an early and a late successional species of Polygonum. Oecologia 57: 270-273.
Zhao M, Botwright Acuña TL, Lafitte HR, Dimayuga G, Sacks E. 2008. Perennial hybrids of
Oryza sativa x O. rufipogon: Carbon exchange and assimilate partitioning. Field Crops Research
106: 214-223.
Zhao M, Ding Z, Lafitte R, Sacks E, Dimayuga G, Holt D. 2010. Photosynthetic characteristics
in Oryza species. Photosynthetica 48: 234-240.
Zotz G, Richter A.2006. Changes in carbohydrate and nutrient contents throughout a
reproductive cycle indicate that phosphorus is a limiting nutrient in the epiphytic bromeliad,
Werauhia sanguinolenta. Annals of Botany 97: 745-754.

168

CHAPTER 4.
SOURCE VS. SINK REGULATION IN PERENNIAL AND ANNUAL CEREALS

169

Introduction

The extent to which resource flow in plants is regulated by source strength (capacity to
generate carbohydrates) as opposed to sink strength (capacity to accumulate carbohydrates) is an
ongoing focus of research. Sink regulation occurs when changes in carbohydrate demand by
flowers, fruits, or other ‘sinks’ lead to compensatory upregulation or downregulation of
photosynthesis (Reekie and Bazzaz, 1987). Sink-regulated photosynthesis exists in many
cultivated woody perennials including apple (Zhou and Quebedeaux, 2003), peach (Gucci et al.,
1991), cherry (Layne and Flore, 1995; Flore & Layne, 1999), citrus (Iglesias et al., 2002),
mango (Urban and Lechaudel, 2005) and coffee (DaMatta et al., 2008), as well in grasses
including wheat (Borras et al., 2004) and sugarcane (McCormick et al., 2006). Other species,
such as peach (Grossman et al., 2006) and maize (Hiyane et al., 2010) can be source limited or
colimited (where both sink and source strength play a role in regulating resource flow). Source
and sink limitation are only relevant terms within a given environment and level of allocation: a
plant that is sink-limited at a particular level of seed production, or in a particular environment,
could become source-limited at higher levels of seed production or under certain environmental
stresses (Flore and Layne, 1999).
There has been little consideration, however, of how source vs. sink limitation may vary
as a function of annual vs. perennial life history. There is some evidence that perennials can be
more sink-limited than annuals, which could be due to longer periods for resource accumulation,
and to past selection for more conservative reproductive strategies (Gaines et al., 1974; Masnatta
and Ravetta, 2011). Alternatively, Arp (1991) hypothesized that perennials, due to larger
vegetative sinks, will be more source-limited than closely related annuals. Recently developed

170

perennial wheat and rye are genetically closely related to their annual analogues, and offer an
ideal system to consider interactions between life history and source/sink dynamics (Jaikumar et
al., 2012). These species combine high reproductive allocation, perennial life history and
herbaceous growth form, a combination of traits that is rare either in nature or in cultivated
agroecosystems (Van Tassel et al., 2010). The findings of Roumet and Roy (1996) suggest that
both life history and selection for high seed yield could profoundly affect the degree of source vs.
sink limitation (as measured by sensitivity to elevated CO2). For example, annual rye might be
less source-limited than annual wheat, since it has not experienced equally intense selection for
high harvest index.
Our goal in this study was to determine whether carbon metabolism in perennial cereals
is sink-limited under low-stress, high-fertility conditions, subjected to modest alterations in
source/sink ratio (+ 50% alterations in spikelet to leaf ratio) from their current levels, and
whether sink limitation in these two perennials is greater than in annual relatives. While past
studies have considered effects of source-sink manipulation in annual wheat, our study makes a
novel contribution by examining source vs. sink limitation in high yielding herbaceous
perennials, as well as in the economically important annual rye, and by considering how life
history affects source vs. sink limitation. Parameters indicating source vs. sink limitation
include photosynthetic rate (Amax), seed size, nonstructural carbohydrate content (NSC), and
subsequent ability for regrowth. Changes in NSC often precede sink-regulated changes in
photosynthetic rate, and serve as a mechanism of feedback inhibition (Franck et al., 2006).
However, these effects are not always observed (DaMatta et al., 2008; Yan et al., 2011) and other
mechanisms may be involved. By measuring NSC we can assess the involvement of this
mechanism in sink regulation. We estimated source/sink balance using the ratio of the leaf area to
171

the number of spikelets (Usenik et al., 2010).
In field experiments, we experimentally manipulated the source/sink ratio in two
perennial cereals (perennial wheat and perennial rye) and their annual analogues, at three time
points, to determine the extent of source vs. sink limitation. Responses measured included
photosynthetic rate, NSC, seed size, and fall regrowth. We hypothesized: 1) perennial and annual
cereals will exhibit enhanced Amax under low source/sink ratios, and reduced Amax under high
ratios, with perennials responding more strongly than annuals; 2) levels of NSC in all species
will decrease under low source/sink ratios and increase under high ratios, and will generally be
higher in perennials than in annuals; and 3) perennial cereals, following the hypothesis of
Masnatta and Ravetta (2011) should fully compensate for changes in source/sink ratio and thus
show no changes in single seed mass or fall regrowth. Each hypothesis is equivalent to assuming
that perennial cereals are sink-limited. Lack of change in seed size, in response to changed
source/sink ratio, would indicate that plants are able to fully compensate for the decrease in
carbohydrate availability (or have no need for excess carbohydrates beyond those achieved by
current levels of photosynthesis). The ability to compensate for part, but not all, of the change in
source/sink ratio would indicate colimitation (Alvaro et al., 2008), while inability to compensate
(e.g. no change in photosynthetic rate, change in seed mass proportionate to change in
source/sink ratio) would indicate source limitation.

172

Materials and Methods

Site and Plant Materials
We conducted this study at the W. K. Kellogg Biological Station in southwest Michigan,
USA, 50 km east of Lake Michigan (42°24′ N, 85°24′ W, elevation 288 m) on soils developed
from glacial outwash deposited 12000 years ago. Soils are fine to coarse loamy, mixed mesic
Typic Hapludalfs. The area receives ca. 897 mm of precipitation annually (based on a 24-year
average up to 2010), with approximately half occurring as snow. Mean annual temperature is
9.7° C. Four species were involved: annual wheat, perennial wheat, annual rye, perennial rye.
The annual wheat cultivar (Triticum aestivum cv. ‘Frankenmuth’, PVP 8000165: Huebner et al,
1999) is used as a benchmark for yield in breeding studies at Michigan State University. The
perennial wheat line, ‘P19’ (Triticum aestivum cv. “Chinese Spring” x Thinopyrum elongatum //
T. aestivum cv. “Madsen”), was developed by Washington State University researchers in 2005
(Murphy et al., 2010) and has been tested for agronomic performance in Michigan (Jaikumar et
al., 2012). Also included was an annual rye cultivar adapted to Michigan (Secale cereale cv.
‘Wheeler’: Helsel and Thomas, 1987), and a perennial hybrid rye (Secale cereale x montanum
cv. ‘Rival’: Peters Seed Company, Oregon); these genotypes have shown good yield potential in
southwest Michigan (Jaikumar et al., 2012).

Experimental Design
A field study was repeated in time, with Experiment 1 carried out in 2009-2010 and
Experiment 2 in 2010-2011. Experimental design, species, and statistical model were the same in
both experiments, though each included measurements not present in the other. Both experiments
173

followed a RCBD split-split plot design, with species the main plot factor, growth stage the
subplot factor, and treatment (source/sink ratio) the sub-subplot factors. In Experiment 1, each
2

-1

species was planted in October 2009 in 6.9 m plots (n = 5 blocks), at 1.5 million seeds ha , 3.8
-1

cm deep, following fertilization with 102 kg ha N. In Experiment 2, each species was planted
2

-1

in November 2010, in 2.25 m plots (n = 6 blocks) at 1.75 million seeds ha , 3.75 cm deep,
-1

following fertilization with 90 kg ha N. In 2010, dates of anthesis were June 1 (annual rye),
June 7 (annual wheat), June 11 (perennial rye) and June 17 (perennial wheat). In 2011, dates of
anthesis were May 31 (annual rye), June 7 (annual wheat), June 8 (perennial rye), and June 19
(perennial wheat). Weeding was done by hand, four times during the season in Experiment 1 and
four times during the season in Experiment 2.
To quantify TPU and second year seed production in response to source/sink
manipulation, a separate pilot experiment was carried out in 2008-2009. This experiment was
carried out before our main source/sink experiment, and did not include annual species; it is of
supplementary interest, however, as it includes two parameters not measured in subsequent
experiments: effects of source/sink manipulation on triose phosphate utilization (TPU) and on
seed production the following year. This experiment involved two perennial wheat genotypes
only, and no annual wheat or rye. One perennial wheat genotype (P19) was mentioned in the
description of Experiment 1 above. The other (P3) was a genotype with generally more ‘grasslike’ spikes. Analysis of grain yields and other agronomic parameters showed no statistical
differences between the genotypes in terms of yield or regrowth. This experiment was laid out as
an RCBD split plot with genotype as the whole plot factor and treatment as the sub plot factor.
There were 5 blocks, each with one plot of P19 and one plot of P3. Each species was planted on

174

2

-1

Oct 25, 2008 in 6.9 m plots at the rate of 1.5 million seeds ha about 2.4 cm deep, using an
Almaco small plot planter. The field was fertilized in Oct 2008 with about 102 kg ha

-1

N in the

form of composted dairy manure. Weeding was done by hand, four times during the season in the
pilot experiment.

Source / Sink Manipulations
Manipulations of source/sink ratio involved “Low” and “High” treatments, each
compared against the control. “Low” and “High” will for the rest of this chapter, refer to sourcereduced and sink-reduced treatments respectively. In the Low treatment, leaf area was reduced
2

by painting the distal 33% of all leaves within a 0.28 m microplot with a nontoxic, solvent-free
white paint (BioShield: Santa Fe, NM). This rendered painted portions photosynthetically
inactive (Karlsson, 1994; Niva et al., 2003). In the High treatment, reproductive sinks were
reduced by manually reducing the number of spikelets on each plant, within an 0.28 m

2

microplot, by 50%. Plants in the control treatment were left intact, and monitored within a
2

corresponding 0.28 m microplot. Following DaMatta et al. (2008) and Masnatta and Ravetta
(2011), leaf area/spikelet ratio (i.e., source/sink ratio) was 67% of the control in the Low
treatment and 200% of control in the High treatment. Source / sink manipulations are shown in
Figures 23-25 (Appendix 2 of this chapter).
Treatments were imposed at three time points: boot stage (12-14 days before anthesis),
pre-anthesis (one week before anthesis) and post-anthesis (one day after anthesis), or
alternatively at approximately stages 45, 55 and 65 on the Zadoks scale (Zadoks et al., 1974). In
Experiment 1, Amax was measured on the perennial species during all three stages and on annual
175

rye during the final two stages: due to adverse weather we were able to measure annual wheat
only during the middle stage. In Experiment 2, Amax was measured during all three stages for all
species. These three sets of manipulations will be referred to as “Boot”, “Pre-Anthesis” and
“Post-Anthesis” stages respectively. During the final two stages, all three treatments (High, Low,
and Control) were imposed; at boot stage, since plants had not flowered, only Low and Control
treatments were imposed.

Survival
In 2011 (i.e., Experiment 2) we monitored “survival to harvest” as a measure of plant
health. In early May, 2011, we counted the number of living plants within each microplot prior
to the first set of manipulations. We then counted the number of living plants again in July, 2011,
to determine the percentage of plants in each microplot that survived until harvest

Photosynthetic Rates
Photosynthetic rates (Amax) were measured on a sunny day 5-8 days after treatments
were imposed for each growth stage. Plants were chosen through a stratified random method,
choosing healthy sections of the plot (with deep green leaves) and selecting three leaves
randomly within this section. Photosynthetic rate (Amax) was measured between 2:30 and 4:30
p.m., as afternoon is often the time at which feedback inhibition of photosynthesis is most
evident (Flore and Layne, 1999). Measurements were taken with a Licor 6400 gas-exchange
system (Licor Biosciences: Lincoln, NE), using approximately ambient CO2 (392 ppm), PAR =
-2 -1

o

1200 µmol m s , at ambient air temperature (ranging from 10 C for boot stage treatments to
176

o

30 C for some post-anthesis treatments) and at 45-55% humidity.

Nonstructural Carbohydrates
To provide insights into photosynthetic feedback processes, NSC of leaves and stems
were monitored in Experiment 2. Enzymatic assays for individual carbohydrates were used to
assess NSC (Andersen & Sorensen, 1999). We opted to use an enzymatic approach, converting
each main type of carbohydrate separately to glucose and assaying the glucose with a
spectrophotometer, in preference to a gas chromatography approach due to the greater ease and
lower cost of the procedure.
In June 2011, (i.e., Experiment 2) stem and leaf tissue samples were collected for NSC
measurements. Four plants from each treatment, for each species, were clipped 10 days
following the pre-anthesis treatment. Flowering heads were removed, leaves (totaling at least 3 g
per sample) and stems (totaling at least 5 g per sample) were separated, and all painted portions
o

of leaves discarded. Samples were flash frozen at -70 C to inactivate invertase and glycosidases
o

(which could destroy or polymerize carbohydrates), and stored in -20 C freezer. In September
2011 samples were removed, lyophilized for two days, and ground through a Christy-Turner mill
with a 0.5 mm screen. To quantify total source strength, leaf mass ratio (leaf dry mass / whole
plant dry mass), was measured on the basis of these samples.
Samples of up to 200 mg of powdered leaf or stem tissue were used for analysis, bulking
across all four plants (in some cases, there was less than 200 mg available). Soluble
o

carbohydrates were extracted four times in hot water at 85 C (3 ml of hot water each time) with
five minutes of centrifugation (3000 rpm) following each extraction. The pellets remaining after

177

-1

the four sequential extractions were digested with 50 units mL alpha-amyloglucosidase (from
-1

Aspergillus niger) and 40 units mL fructanase (Megazyme Co: Dublin, Ireland) in 6 mL total
o

solution at 60 C for 30 min. Following digestion, each carbohydrate was assayed using the
2+

following six enzyme combinations: 1) hexokinase, glucose 6-P dehydrogenase, NADP, Mg

on

an aliquot from the supernatant to assay glucose; 2) hexokinase, glucose 6-P dehydrogenase,
2+

NADP, Mg

and phosphoglucose isomerase on an aliquot from the supernatant to assay
2+

fructose; 3) hexokinase, glucose 6-P dehydrogenase, NADP, Mg

and invertase on an aliquot
2+

from the supernatant to assay sucrose; 4) hexokinase, glucose 6-P dehydrogenase, NADP, Mg ,
phosphoglucose isomerase and fructanase on an aliquot from the supernatant to assay soluble
2+

fructans; 5) hexokinase, glucose 6-P dehydrogenase, NADP, Mg , and phosphoglucose
isomerase on an aliquot from the pellet to assay insoluble fructans; and 6) hexokinase, glucose 62+

P dehydrogenase, NADP, and Mg

on an aliquot from the pellet to assay starch. Each aliquot

was about 1.5 mL total volume, and included generally 10-40 microliters of leaf extract solution.
Absorbance of NADPH at 340 nm for each fraction was measured using a Shimadzu UV-2600
spectrophotometer (Shimadzu Corporation: Kyoto, Japan). Summed together, these form the
NSC fraction.
This method is relatively new, and constitutes a novel adaptation of existing enzymatic
approaches to assay plant NSC. Previous enzymatic assay procedures, particularly in studies of
source/sink balance, have generally focused on estimating starch and soluble sugars (e.g. Flore
and Layne, 1999). Cool-season grasses (Poaceae) pose a problem for such procedures, as they

178

may maintain large amounts of their total carbohydrate reserves in the form of small to large size
fructose polymers built on a sucrose substate (fructans). Fructans may range in size from as few
as 8 to as many as 100,000 fructose units (Vijn and Smeekens, 1999) and fructan production is a
trait cool season grasses share along with certain other important plant families including
Asteraceae (the sunflower family), Amaryllidaceae (onion family), Agavaceae (agave family),
Asparagaceae (asparagus family), Iridaceae (iris family) and Boraginaceae (borage family). It is
believed that fructans play an important role in cold tolerance, cryoprotection of membranes, and
osmotic adjustment, and they may also play a role in helping to store carbohydrates within
vacuoles, in a way that reduces feedback inhibition of photosynthesis (Kim S-H, personal
communication).
Standard methods for assaying NSC are poorly adapted for assays of fructan
concentration, as fructans show a range of water solubilities (high for small chains, low for large
chains) and thus partition partly within the supernatant and partly within the pellet, depending on
the conditions of extraction. They are also acid-labile and can break down under the acidic
conditions used in some protocols for assaying starch. Thus in order to accurately estimate
fructan concentrations it is necessary to include a step which specifically breaks fructans down to
fructose, to avoid strongly acidic starch digestion conditions, and to assay fructans both within
aliquots from the supernatant as well as the pellet. Our modification of previous enzymatic NSC
assay procedures consists of the incorporation of a fructan digestion step. Our method was
validated using samples of sunchoke (Helianthus tuberosus) tubers and Kentucky bluegrass (Poa
pratensis) leaves, comparing our estimate of each type of carbohydrate to published literature.

179

Seed Mass and PSCR
In Experiment 1, plants were harvested on July 21 (annuals) and on August 15
(perennials). Seed heads were removed and cleaned using a customized tabletop thresher.
Subsamples of 20 seeds per treatment were randomly selected and weighed to determine single
seed mass. Harvest was conducted on July 21 (annuals) and August 10 (perennials) for
experiment 2, while post-sexual cycle regrowth (PSCR, percent of plants regrowing) was
visually estimated for each microplot on Oct 15.

Fall Flowering and Fall Photosynthesis
One unusual feature of perennial wheat and rye, which may be indicative of the fact that
these species derive from an annual x perennial cross, is that they may experience a second
round of unproductive, abortive flowering in the fall, if they have been able to grow large enough
during the post-sexual cycle regrowth (PSCR) period. This may occur if weather conditions are
favorable. Fall flowering is generally unproductive, as it is interrupted by the onset of cold
weather which kills the flowers. However, as only a minority of the perennial cereal plants in our
experiment initiated fall flowering in 2011, this allowed us to conduct a natural experiment to
determine how closely photosynthetic rate was coupled with flowering. By comparing
photosynthetic rates in flowering and non-flowering plants at the same moment in time, we could
determine whether flowering was associated with higher photosynthetic rates. We carried out
photosynthetic measurements on October 15 – 17, 2011, on both perennial wheat and rye,
between 9:00 – 11:30 a.m., following the same protocol above.

180

Pilot Experiment
In the pilot experiment, the Low, High and Control treatments were imposed only at one
of the three growth stages (pre-anthesis). Photosynthetic rates (Amax) were measured on P19 and
P3, using the methods specified above, 1 week after treatment impositions. In addition, triose
phosphate utilization (TPU) was measured 10 days following treatment imposition. TPU is a
measure of the total capacity of plants to process and export triose phosphate molecules: it
represents the maximum rate at which photosynthesis can proceed, as plants which generate
carbohydrates faster than they can be exported will experience depletion of free phosphate and
feedback inhibition of photosynthesis. TPU was modeled as Asat – Rd, where Asat is defined by
observed net photosynthesis at saturating concentrations of CO2, and Rd represents day
-2 -1

respiration (Sharkey et al., 2007). Here day respiration was assumed to be - 2.00 µmol m s ,
and 1600 ppm CO2 was assumed to be saturating.
Following imposition of treatments, the microplots where treatments had been imposed
were labeled with colored flags and tracked into the following year. In July 2011, seed heads
from each of the labeled microplots were harvested, threshed through a customized tabletop
seed cleaner, and weighed. The purpose was to determine whether source-sink ratio manipulation
had delayed effects on the reproductive fitness of plants in the following year.

Statistical Analysis
The model for photosynthetic rates and seed size was an RCBD split-split plot ANOVA:
Amax = µ + Species + Block + Species x Block +Species x Block x Stage + Stage + Treatment +
Species x Stage + Species x Treatment + Species x Stage x Treatment.

181

Stage (i.e., growth stage at the time of treatment imposition), species and treatment were
fixed factors and block was considered a random factor. Contrasts of High and Low treatments
against the control, for a given species and/or species x stage combination, were made using
Dunnett’s post-hoc test. Experiments 1 and 2 were examined separately, using PROC MIXED in
SAS 9.2 (SAS Institute, 2008). Post sexual cycle regrowth (PSCR) was measured on all six
source/sink manipulation treatments at the same time, so we did not include ‘stage’ as a variable.
Rather, we treated each stage x manipulation combination as a separate treatment for this
analysis, (e.g., “Pre-Anthesis Low” and “Post-Anthesis Low”) were considered different
treatments. Our model for regrowth (logit-transformed to meet assumptions of ANOVA) was
thus:
log [(0.1 + PSCR) / (1.1 – PSCR)] =
µ + Species + Block + Species x Block + Treatment + Species x Treatment.

NSC was analyzed separately for stem tissues and leaf tissue samples. For each, we used
the same model as for PSCR, but with only three treatments (i.e., Low, High and Control).
The overall model in the pilot experiment was as follows:
A = µ + Genotype + Block + Genotype x Block + Genotype x Treatment .

Results

Survival
No differences in survival to harvest were seen between different treatments, indicating

182

that treatments did not increase immediate mortality (data not shown).

Spring Photosynthetic Rates
In Experiment 1, photosynthetic rates (Amax) were affected by source-sink ratio: there
was also a species x treatment interaction, meaning that species responded differently to
source/sink manipulation (Table 12). As there was not a three-way interaction, we averaged
Amax for each species x treatment combination across growth stages (Figure 14). In perennial
rye, the Low treatment had 17.5% higher Amax than the control, while the High treatment was
14% lower. In perennial wheat, the Low treatment had 11% higher Amax than the control, while
the High treatment did not differ from controls. Neither annual species showed treatment effects.
-2 -1

Photosynthetic rates (Amax) ranged between 21-26 µmol m s for perennial wheat, 18-20 µmol
-2 -1

-2 -1

-2 -1

m s for annual wheat, 15-23 µmol m s for perennial rye, and 10-23 µmol m s for
annual rye. Responses were similar between the boot and pre-anthesis stages (means shown in
Table 12) and the post-anthesis stage (Table 13).
Photosynthetic responses in Experiment 2 were similar (Table 14). Due to the three-way
interaction we considered treatment responses separately for each growth stage. Photosynthetic
rate (Amax) for control plants, averaged across the three growth stages (Figure 15) were between
-2 -1

-2 -1

13-19 µmol m s for perennial wheat, 11-17 µmol m s for annual wheat, 16-19 µmol m
-1

-2

-2 -1

s for perennial rye, and 15-17 µmol m s for annual rye. Leaf mass ratio (leaf dry mass /
total dry mass) did not vary between perennial and annual wheat (23 – 33%), or perennial and

183

annual rye (32 – 36%).
Source/sink manipulation, at boot stage, influenced Amax for every species except annual
wheat (Table 14). The Low treatment showed 10-12% higher Amax than the control in annual and
perennial rye, and 18% higher Amax than the control in perennial wheat. At the pre-anthesis
stage, Amax responded to source/sink manipulation only in perennials (Table 14). The Low
treatment showed 27% and 17% higher Amax than the control, in perennial rye and perennial
wheat, respectively. The High treatment showed no response in perennial rye and 12% lower
Amax in perennial wheat. Finally, source/sink manipulations at the post-anthesis stage affected
photosynthetic rate in all species (Table 15). The Low treatment showed 13-14% higher Amax
than the control in the perennials, and 8-11% higher in the annuals. The High treatment showed
14% lower Amax than the control in perennial rye, 24% lower in perennial wheat, 12% lower in
annual rye, and did not differ from control in annual wheat.

Nonstructural Carbohydrates
In all four species, leaf NSC showed similar responses to source/sink manipulations, with
NSC concentrations increasing in ‘High’ source/sink treatments (14 – 84%), and decreasing in
‘Low’ treatments (12 – 33%). Although treatment responses in leaf tissue were more pronounced
in the two Triticum species than in the Secale species, there were no consistent differences in
treatment response between annual and perennial species, and the perennials generally showed
similar responses to annuals (Table 17: Figure 16). In the Low source/sink treatment, leaf NSC
was reduced by 33% in annual wheat, and 12-14% in all other species (Figure 16). The High
184

treatment increased leaf NSC by 23% in annual wheat, 84% in perennial wheat, 22% in annual
rye and 14% in perennial rye. Treatment had an effect on concentrations of sucrose (P = 0.036)
and soluble fructans (P < 0.0001) but not on other carbohydrates: thus changes in NSC appear to
be mostly driven by changes in soluble fructans (32-42% of the total: Table 16) and in sucrose
(14-19% of the total: Table 17). This could reflect the fact that these two carbohydrates are
quantitatively the largest pools, accounting for over half the total: alternatively, it could be
related to the fact that sucrose and fructan metabolism are closely related, with sucrose serving as
the initial substrate for fructan formation (Vijn and Smeekens, 1999).
Stem NSC showed the same trend for annual and perennial wheat, increasing in the High
treatment (43% -45%) and decreasing in the Low treatment (11% in annual wheat, 38% in
perennial wheat): however, stem NSC was not affected by treatment in annual or perennial rye
(Table 19: Figure 17). As with the leaf tissues, buildup of NSC under high source/sink ratio and
depletion of NSC under low source/sink ratio could be explained entirely by changes in soluble
fructan and sucrose concentration (the two largest carbohydrate pools). Stem NSC was 34 –
43% soluble fructans (Table 18), and 16 – 23% sucrose (Table 19). Starch, fructose, glucose and
insoluble fructan concentrations were not affected by source-sink treatments.
Overall, total NSC concentrations in unmanipulated (control) plants did not differ
between stem and leaf tissues, ranging from 13 – 19% of dry matter in leaves (Table 17; Figure
16) and 15 – 25% in stems (Table 19; Figure 17). Leaf NSC was 28% higher in perennial
compared to annual wheat (t = 3.09, p = 0.027), but did not differ between perennial compared to
annual rye (Table 13). Likewise, stem NSC in control plants was 31% higher in perennial
compared to annual wheat (t = 3.26, p = 0.022), but was no different in perennial and annual rye
(Table 14). Thus the wheat species, but not the rye species, support our hypothesis that

185

carbohydrate reserves would be greater in the perennial species. This is consistent with the
results of Gonzalez-Paleo and Ravetta (2011) who found higher leaf NSC in perennial as
compared to annual Lesquerella. It also suggests an overall trend for a more conservative
resource strategy in perennial wheat: perennial wheat, relative to annual wheat, appears to be
maintaining reproductive allocation at a fairly low level, with additional photosynthetic capacity
and additional stored carbohydrates that are not being utilized for reproduction.
Differences in leaf NSC between species appear to be driven by changes in glucose
content (15 – 26% of total NSC in leaves; Table 16) rather than by other carbohydrate pools. The
same is true of stems, where glucose constitutes 11 – 25% of total NSC (Table 18). The glucose
contents were unusually high, while starch and sucrose contents were low; these results may be
indicative of some breakdown of starch and sucrose, through residual amylase and invertase
activity, during the transferring and processing of our samples.

Seed Mass and Regrowth
In Experiment 1, seed mass was affected by stage x treatment interaction (F = 7.82, P =
0.001), and by species x treatment interaction (F = 3.35, P = 0.005; Figure 18). Since there was
no three-way interaction we averaged across growth stages for each species x treatment
combination. Seed mass of control plants was on average 36 mg in annual wheat, 27 mg in
perennial wheat, 36 mg in annual rye and 28 mg in perennial rye. Source/sink ratio affected seed
size for annual wheat, and annual rye, perennial rye, but not perennial wheat. In annual rye,
pooled across the three growth stages, the Low treatment reduced seed size by 12%, while the
High treatment increased it by 12%. In perennial rye, the Low treatment reduced seed size by
14%, while the High treatment did not increase it. In annual wheat, the Low treatment reduced

186

seed size by 17%, while the High treatment increased seed size by 7% (Figure 18). Much of the
treatment effect was driven by effects at boot stage, when annual wheat, annual rye and perennial
rye showed large responses. For the annual species examined here, increased source/sink ratio
led to a 30% decline in seed size. Regrowth percentage was not affected by treatment in either
perennial species although, perennial wheat did tend to have a lower percentage of regrowth
(63 – 83%) than perennial rye (90 – 96%: Table 14).

Fall Flowering and Photosynthesis
Fall flowering in perennial rye was very low (across the six blocks, only 2.5% of
perennial rye plants flowered on average). Because of the low flowering rates, we could not
compare photosynthesis in nonflowering and flowering individuals. In perennial wheat, between
6 – 60% of plants produced flowering heads in the fall, although no mature seed heads were
produced due to the onset of frost in late fall. Flowering individuals in the perennial wheat plots
maintained 35% higher photosynthetic rates than non-flowering individuals (t = 8.26, p = 0.0016
following the Bonferroni correction for k = 2 comparisons: Figure 19). This does not
demonstrate causality, since it is unclear whether higher rates of resource acquisition drive
flowering or vice versa. However, it does provide further evidence of close coupling between
resource supply and demand in these plants.

Pilot Experiment
In the 2009 pilot experiment, photosynthetic rate was measured only on the two perennial
wheat genotypes. In one of the plots the Low treatment was heavily damaged by rodents and no
measurements could be taken. Photosynthetic rate (Amax) and TPU were both strongly affected
by treatment. There was no main effect of genotype on either parameter, nor an effect of

187

treatment x genotype, and treatment effects were significant for both species (Table 20). To
simplify the model for Amax and TPU we pooled both genotypes (since there was little
difference between the genotypes). Treatment affected Amax when pooled across genotypes (F =
6.33, p = 0.04): the Low treatment had 20% higher photosynthetic rates than the control (t =
5.51, p < 0.01, n = 9), while the High treatment did not differ from the control (Table 21).
Treatment also affected TPU when pooled across genotypes (F = 53.78, p < 0.0010). The Low
treatment had 18% higher TPU than the control (t = 6.25, p < 0.01, n = 9) while the High
treatment was 11% lower than the control (t = 3.79, p < 0.01).
In the pilot experiment, neither total second-year seed yield from treated plants (i.e. seed
yield in 2010, in the year following treatments), nor single seed mass, was affected by treatment,
genotype or the interaction of genotype and treatment (Table 20). Thus the manipulations of
source/sink ratio had no effect on kernel mass or on seed yield in the following year (Table 21).

Discussion

Annual cereals showed colimitation rather than sink limitation during flowering and grain
filling, and compensated for only part of the source/sink change (e.g., seed size declined by only
16% when source/sink ratio declined by 33%, indicating the lack of complete source limitation).
This is suggested by the lack of consistent compensatory photosynthesis, as well as by changes
in seed size. When manipulations were implemented during boot stage, both species showed
clear source limitation, with seed mass reflecting source/sink ratio. These results match previous
studies, in which annual wheat varies between being sink-limited, and colimited, during grain

188

filling (Borras et al., 2004; Reynolds et al., 2005). Wheat has also been found to be sink limited
during vegetative stage (Harrison et al., 2010) but source limited shortly before anthesis, as
illustrated by stimulatory effects of elevated CO2 on yield (Amthor, 2001). The mechanism of
switching between source and sink limitation is that carbohydrate concentrations at anthesis
affect the number of florets fertilized, and plants fertilize only as many florets as can be filled
largely on the basis of stored reserves (Cruz-Aguado et al., 1999). The effect sizes we found in
annual wheat are consistent with existing literature on changes in photosynthetic rate (Zhenlin et
al., 1998; Reynolds et al., 2005), seed size (Borras et al., 2004; Alvaro et al., 2008), and NSC
content (Kuhbauch and Thome, 1989; Ma et al., 2009).
To the best of our knowledge, no studies have explicitly considered source vs. sink
limitation of photosynthesis in annual rye. However, a previous study found that defoliation
during reproductive period caused 7 – 17% decreases in seed mass, with greater effects when
manipulations took place earlier in the season (Rho et al., 1990); in the wheat-rye hybrid triticale,
overall seed mass decreases were greater than in rye. These data give some indication that rye
shows colimitation before and after anthesis, that the degree of source limitation is less than in
wheat (since triticale, with its introduction of wheat genetics, shows more source limitation than
rye) and that source limitation becomes less important later during reproductive development.
Our results, which found greater compensatory changes in Amax and stability of seed mass in
annual rye compared to wheat, are consistent with these conclusions.
In contrast to the source/sink colimitation seen in annual wheat and annual rye, perennial
wheat showed complete sink limitation during all three growth stages. This was indicated by
larger changes in Amax relative to annuals, stability of seed size and large buildup of NSC.

189

Increased photosynthesis and NSC depletion allowed perennial wheat to tolerate low source/sink
ratio with no change in seed mass. Perennial rye showed sink limitation during flowering and
grain filling, but colimitation during boot stage where Amax increased. Seed size also declined in
response to source/sink reduction in perennial rye, although less so than in the annuals. Perennial
rye did not show changes in stem NSC in response to changed source/sink ratio. This could be
because perennial rye shows both greater upregulation of photosynthetic rate (including the
highest increase in Amax in our study) and small declines in seed size in response to decreased
source/sink ratio, which could allow it to avoid depleting stem NSC reserves (Figure 17).
Similarly, increased source/sink ratio results in interrelated compensatory responses, including
(in some cases) reductions in photosynthesis, buildup of TNC in leaves, and no change in seed
mass. In neither species was fall regrowth affected by moderate source/sink perturbations,
suggesting that perennial cereals could tolerate selection for moderately increased sink strength
without losing the ability to perenniate. Thus our hypotheses were generally supported, with
perennial cereals showing more sink limitation than annuals.
Perennial wheat and rye, with their high Amax, small seeds and lower grain yield
compared to annuals, demonstrated a stronger form of the conservative reproductive strategy
observed previously in annual wheat (Reynolds et al., 2009). This stands in contrast to the study
by Masnatta and Ravetta (2011) which found equal responsiveness to source-sink manipulations
in annual and perennial Lesquerella. A conservative strategy appears to act as a partial (in
perennial rye) or full (in perennial wheat) buffer against reductions in effective leaf area. In
general, wild perennial grasses that pursue a ‘conservative’ strategy face fewer risks than annual
cereals, as low reproductive output in one year can be compensated for in the future. Perennials

190

also face risks from a ‘liberal’ strategy, since they are not subjected to human interventions that
can ameliorate effects of environmental stresses after anthesis. Increased sink capacity, coupled
with sink limitation, could be a means for some perennial cereals to achieve higher Amax and
resource acquisition rates than annuals. Previous research has indicated that some perennial rice
species maintain higher Amax than annual rice (Zhao et al., 2008). If the findings from this study
indicating sink limitation hold for other perennial cereals, this supports the hypothesis that
increased allocation to vegetative structures could promote increased photosynthesis relative to
annuals (Zhao et al., 2008).
Not only do perennial cereals appear to maintain relatively stable seed size during the
harvest immediately following source/sink ratio manipulation, but stability of seed size appears
to be maintained during the following year as well. Our pilot experiment indicated that
increasing or decreasing the source sink ratio in perennial cereal plants did not affect the seed
yield of those plants in the following year. Thus there do not appear to be effects of carbohydrate
depletion persisting post harvest, and during the year following source sink manipulation plants
appear to be able to compensate fully for changes in source sink ratio and carbohydrate reserves,
both in the immediate and the longer term. Changes in source sink ratio do not appear to affect
immediate survival, fall regrowth, or seed production the following year, and have only modest
(in perennial rye) or no effects (in perennial wheat) on seed size within the same year.
Physiological responses to sink reduction and source reduction in perennial fruit trees and
grasses are often larger than the effect sizes observed in our study (DaMatta et al., 2008; Park
and Kim, 2011). For example, heavy shading of sugarcane (source reduction) caused a 27%
decrease in soluble sugar content in unshaded leaves, with concomitant 50-65% increases in
Amax (McCormick et al., 2008a). Conversely, cold-girdling of sugarcane (sink reduction)
191

increased soluble sugar concentrations by 80% and Amax by 40% (McCormick et al., 2008b).
Crested wheatgrass (Agropyron cristatum) showed an even larger reponse: photosynthesis
increased by 190% and 160% in response to 70% and 85% biomass removal, respectively (Peng
et al., 2007). In a study on apple trees (Malus domestica) source reduction led to a 33% increase
in photosynthesis, whereas root sink removal through cold-girdling of apple trees resulted in a
50% decrease in Amax (Zhou and Quebedeaux, 2003; Fan et al., 2010). Similarly, in cherry,
increased source/sink ratios lead to 30-50% increases in Amax (Layne and Flore, 1995). Thus the
physiological responses to source-sink ratio in our study were modest compared to many
perennials.
The specific mechanisms by which source-sink ratio manipulation affects photosynthetic
rate in perennial cereals is unclear. Buildup of assimilates and subsequent feedback inhibition of
photosynthesis, possibly through depletion of free phosphates and declines in ATP/ADP ratio, is
a mechanism that has been widely proposed (Sharkey, 1985), and indeed many plants do show
buildup of soluble carbohydrates in advance of the decline in photosynthetic rates (Iglesias et al.,
2002; Flore and Layne, 1999). It is also known that changes in sink/source ratio can affect
phosphate pools in leaf tissue, which could be a mechanism involved in feedback inhibition.
However, buildup in carbohydrate concentrations prior to declines in photosynthesis are not
always seen, and other mechanisms are almost certainly involved in many cases. In addition, as
around 75% of plants, under normal environmental conditions, are not at the Ci level where they
are limited by TPU (Wullschleger, 1993) it is likely that in most cases effects of sink limitation,
while they may most directly and immediately reduce TPU, also have indirect effects on other
determinants of photosynthetic rate. Buildup of carbohydrates may lead to stomatal closure (Yan
et al., 2011), or decreases in Rubisco activity and electron transport rate (Lewis, James D.,
192

personal communication). For example, Ainsworth et al. (2004) found declines in Rubisco
activity under elevated CO2 in soybean, while Krapp and Stitt (1995) found a 45% decrease in
Rubisco activity in cold girdled spinach leaves, which was not attributable to declines in total
amount of Rubisco but most likely to changes in the activation state. A study in soybean (Kasai,
2008) found that high source/sink ratios and buildup of soluble carbohydrates were directly
associated with declines in the activation state of Rubisco. Schaefer et al. (1992) found similar
results in Chenopodium rubrum, where buildup of carbohydrates through glucose feeding led to
decreases in Rubisco activity. Foyer et al. (1990) also hypothesized that electron transport could
be affected by sink-regulated changes in free phosphate, through regulation of the redox state of
intermediate electron carriers.
In the case of these perennial cereals, little work was done on elucidating the mechanisms
of sink regulation of photosynthesis. We did find in our pilot experiment, that at least in
perennial wheat, changes in source/sink ratio affect TPU as well as photosynthetic rate, and that
the changes in TPU are of similar magnitude to changes in Amax. However, we also observed
that TPU was much higher than Amax at ambient CO2. This suggests that downregulation of TPU
and sink activity decrease photosynthetic rate indirectly, and that declines in TPU have
secondary effects on other processes regulating photosynthesis (either stomatal closure, electron
transport rate or Rubisco activity). Our examination of the impacts of source-sink ratio separately
on TPU and on A illustrate that sink regulation of photosynthetic activity appears to be a
complex process that involves coordinated changes in several parameters, and that regulates A so
that it is maintained well below maximal TPU capacity.

193

Conclusion

Perennial wheat and rye show more sink regulation of metabolism than annual
equivalents, and demonstrate compensatory responses that help buffer seed size against
environmental fluctuation. This indicates that perennial cereals currently operate further below
their maximum photosynthetic capacity than annual wheat and rye, and follow a conservative
reproductive strategy. Our results suggest that higher sink capacity could help explain why some
perennial species maintain higher photosynthetic rates than closely related annuals, and illustrate
how life history and physiology affect each other in these emerging representatives of a novel
plant growth form.
Further work is necessary to determine the precise mechanisms by which source/sink
ratio regulates photosynthetic rate and metabolism in perennial cereals, as well in other species
where the mechanisms are not fully clear. One approach towards elucidating mechanisms of sink
regulation, that could build upon the work done in this study, would be to generate A vs. Ci
(photosynthetic rate vs. intercellular carbon dioxide concentration) curves for perennial and
annual cereal plants subjected to higher and lower source/sink ratios. These curves allow
estimation of stomatal limitation, electron transport rates, TPU and Rubisco-catalyzed
carboxylation of ribulose bisphosphate carboxylation (Sharkey et al., 2007; Farquhar and
Sharkey, 1982). By observing which specific parameters differed between High, Low and
Control treatments, as well as observing overall intercellular/ambient carbon dioxide ratios and
stomatal conductance, it would be possible to more closely elucidate the mechanisms of sink
regulation.
A further continuation of this study could be to look at how source/sink ratio affects not
simply the rate of photosynthesis, but its duration. It is known that in some plants, changes in
194

source/sink ratio affect the timing of leaf senescence: for example, in wheat, higher source/sink
ratios associated with elevated CO2 cause earlier leaf senescence (Zhu et al., 2009) while buildup
of carbohydrates is associated with later leaf senescence in the spring geophyte Erythronium
americanum (Gandin et al., 2011). If plants with low source/sink ratios keep their leaves green
and photosynthetically functional for longer, this could serve as an additional strategy by which
they could compensate for the increased carbohydrate demand relative to leaf area. In our study
it was observed that the changes in photosynthetic rate, while significant, were relatively modest
compared to some woody perennials or wild grasses, and were quantitatively too small to fully
compensate for the changes in carbohydrate demand relative to supply. This was reflected,
among other things, in the fact that high sink strength did lead to carbohydrate depletion, and
high source/sink ratio led to carbohydrate buildup. Changes in the timing of leaf senescence,
such that plants with low source/sink ratios maintain green leaves for longer, could serve as an
additional compensatory mechanism, and further studies are needed to determine the extent to
which senescence and length of growing season respond to source/sink ratio.

Acknowledgements

Acknowledgements are due to Mark Freeman and John Green for field management: to
Bert Cregg for equipment and advice: and to P. Staffan Karlsson of the Swedish Research
Council for Statistics and Analysis for helpful suggestions. The authors acknowledge financial
support from the USDA Organic Agriculture Research and Extension Initiative (OREI) project
on ‘Practical Perennials: Partnering with Farmers to Develop a New Type of Wheat Crop’

195

APPENDICES

196

APPENDIX A. Tables and Figures for Chapter 4

197

Table 12. Experiment 1 (2010) means (SE), F-values and P-values for photosynthetic rates (Amax) following boot stage and
pre-anthesis manipulations. Photosynthetic rates were measured 5-8 days following treatment imposition, for annual wheat,
perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” = 67%
of control, “Control”, and “High” = 200% of control) at two or three growth stages. Only the means for the first two growth stages are
shown in this table, but F-values are reported for the experiment as a whole. Symbols ‘*’ and ‘**’ represent significant differences
-2 -1
from the control, following Dunnett’s post-hoc testing, at P = 0.05 and P = 0.01 respectively. Units for Amax are µmol m s of CO2.

Species
Annual wheat
Perennial wheat
Annual rye
Perennial rye

Boot
Low
----25.3 (0.9)**
----23.3 (1.2)**

Boot
Control
----23.0 (0.4)
----20.6 (0.9)

Pre
Low
19.9 (1.1)
26.1 (0.8)**
22.5 (1.0)
22.6 (1.6)**

Species
Treatment
Stage
Sp x Trt
Sp x Stage
Trt x Stage
Sp x Trt x Stage

F-ratio
38.22
14.93
50.61
2.41
23.64
0.89
0.61

ANOVA results

198

Pre
Control
18.0 (1.0)
22.8 (0.8)
19.7 (1.4)
19.3 (1.6)
P-value
< 0.001
< 0.001
< 0.001
0.036
< 0.0007
0.50
0.69

Pre
High
18.3 (0.7)
22.9 (1.0)*
20.0 (1.0)
17.1 (1.6)

Table 13. Experiment 1 (2010) means (SE) for photosynthetic rates, measured 5-8 days following post-anthesis source/sink
manipulations, for annual wheat, perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of
source/sink ratio treatments (“Low” = 67% of control, “Control”, and “High” = 200% of control) at 1 day post-anthesis. This
experiment was conducted at Kellogg Biological Station, Hickory Corners, MI. Symbols ‘*’ and ‘**’ represent significant differences
-2 -1
from the control, following Dunnett’s post-hoc testing, at P = 0.05 and P = 0.01 respectively. Units for Amax are µmol m s of CO2.

Species
Annual wheat
Perennial wheat
Annual rye
Perennial rye

Post
Low
----23.9 (0.7)**
10.3 (1.4)**
20.8 (0.9)**

Post
Control
----21.7 (0.6)
10.0 (1.1)
16.9 (0.4)

Post
High
----21.0 (0.8)**
10.2 (1.2)
15.5 (1.0)**

199

Table 14. Experiment 2 (2011) means (SE), F-values and P-values for photosynthetic rates (Amax) measured 5-8 days following
boot stage and pre-anthesis source/sink manipulation, in annual wheat, perennial wheat, annual rye and perennial rye and
post-sexual cycle regrowth (PSCR) in perennial wheat and rye. All plants were subjected to three levels of source/sink ratio
treatments (“Low” = 67% of control, “Control”, and “High” = 200% of control) at two or three growth stages (“Boot” = boot stage,
“Pre” = 1 week pre-anthesis, and “Post” = 1 day post-anthesis). Only means for the first two stages, and analysis of variance results
for photosynthetic rates, are shown in this table. The experiment was conducted at Kellogg Biological Station, Hickory Corners, MI.
Symbols ‘*’ and ‘**’ represent significant differences from the control, following Dunnett’s post-hoc testing, at P = 0.05 and P = 0.01
-2 -1
respectively. Units for Amax are µmol m s of CO2, while PSCR is expressed as a %.

Boot
Low

Boot
Control

Pre
Pre
Low
Control
Photosynthetic Rate

Pre
High

Annual wheat
Perennial wheat
Annual rye
Perennial rye

15.1 (1.3)
18.0 (1.5)**
21.4 (1.4)**
22.6 (1.0)**

14.0 (1.2)
15.4 (1.8)
19.6 (1.4)
20.3 (1.2)

11.7 (1.0)
21.7 (2.0)**
13.3 (1.0)
20.0 (2.1)**

12.3 (1.5)
18.7 (1.8)
13.8 (0.9)
15.8 (1.5)

10.7 (1.3)
16.5 (1.8)*
13.4 (1.0)
16.6 (1.9)

Perennial wheat
Perennial rye

66 (14)
97 (2)

65 (14)
93 (2)

PSCR
69 (6)
93 (3)

65 (14)
93 (2)

85 (8)
89 (2)

F-ratio

ANOVA results (Amax)
Species
Treatment
Stage
Sp x Trt
Sp x Stage
Trt x Stage
Sp x Trt x Stage

200

5.92
69.20
2.81
4.36
4.63
2.89
2.07

P-value
0.007
< 0.001
0.072
< 0.001
0.001
0.039
0.04

Table 15. Experiment 2 (2011) means (SE), F-values and P-values for photosynthetic rates (Amax) measured 5-8 days following
post-anthesis source-sink manipulation in annual wheat, perennial wheat, annual rye and perennial rye, and post-sexual cycle
regrowth (PSCR) in perennial wheat and rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” =
67% of control, “Control”, and “High” = 200% of control) at two or three growth stages (“Boot” = boot stage, “Pre” = 1 week preanthesis, and “Post” = 1 day post-anthesis). Only means for the post-anthesis growth stage, and analysis of variance results for PSCR,
are shown in this table. The experiment was conducted at Kellogg Biological Station, Hickory Corners, MI. Symbols ‘*’ and ‘**’
represent significant differences from the control, following Dunnett’s post-hoc testing, at P = 0.05 and P = 0.01 respectively. Units
-2 -1
for Amax are µmol m s of CO2, while PSCR is expressed as a %.

Annual wheat
Perennial wheat
Annual rye
Perennial rye
PSCR
Perennial wheat
Perennial rye

Post
Post
Low
Control
Photosynthetic Rate
17.0 (0.8)**
15.8 (0.7)
14.9 (2.2)**
13.2 (2.2)
20.6 (1.1)**
18.6 (0.8)
18.8 (0.8)**
16.5 (0.6)

Post
High
14.5 (1.0)
10.4 (2.3)**
16.4 (1.9)
14.2 (0.5)**

%
68 (13)
93 (3)

ANOVA results
Species
Treatment
Sp x Trt

65 (14)
93 (2)
PSCR
F-ratio
0.15
4.85
7.07

86 (6)
94 (3)
P-value
0.44
0.001
0.001

201

Table 16. Experiment 2 (2011) means for unmanipulated plants (percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (glucose, starch, fructose, soluble fructans) measured around anthesis, on leaf tissue from annual wheat,
perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” =
67% of control, “Control”, and “High” = 200% of control) at approximately 1 week pre-anthesis. The experiment was conducted at
Kellogg Biological Station, Hickory Corners, MI. Only means for unmanipulated (“Control”) plants are shown. The symbol ‘*’
indicates significant differences compared to the annual species (annual wheat or rye) at α = 0.05.

Species

Glucose
%
2.54 + 0.42
4.62 + 1.29*
2.34 + 0.36
3.34 + 0.49*

Starch
%
0.40 + 0.08
0.41 + 0.15
0.23 + 0.06
0.28 + 0.09

Fructose
%
2.80 + 0.62
2.24 + 0.51
3.68 + 0.83
4.82 + 1.05

Sol. Fructans
%
5.11 + 1.15
7.27 + 1.86
4.70 + 1.51
7.01 + 1.99

Source of Variation

Glucose

Starch

Fructose

Sol. Fructans

Species
Treatment
Species x Treatment
Block
Block x Treatment

5.77 (0.008)
0.31 (0.73)
0.44 (0.84)
0.43 (0.33)
0.85 (0.20)

2.34 (0.11)
1.07 (0.35)
1.01 (0.43)
0.24 (0.40)
0.01 (0.99)

4.40 (0.021)
1.04 (0.36)
0.45 (0.84)
0.01 (0.99)
0.01 (0.99)

1.75 (0.20)
17.60 (< 0.0001)
0.59 (0.74)
0.01 (0.99)
1.24 (0.11)

Annual Wheat
Perennial Wheat
Annual Rye
Perennial Rye

202

Table 17. Experiment 2 (2011) means for unmanipulated plants (percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (insoluble fructans, sucrose, and total NSC) measured around anthesis, on leaf tissue from annual wheat,
perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” =
67% of control, “Control”, and “High” = 200% of control) at approximately 1 week pre-anthesis. The experiment was conducted at
Kellogg Biological Station, Hickory Corners, MI. Only means for unmanipulated (“Control”) plants are shown. The symbol ‘*’
indicates significant differences compared to the annual species (annual wheat or rye) at α = 0.05.

Species
Annual Wheat
Perennial Wheat
Annual Rye
Perennial Rye

Insol. Fructans
%
0.41 + 0.17
0.11 + 0.06
0.30 + 0.08
0.43 + 0.15

Sucrose
%
1.51 + 0.42
2.13 + 0.54
3.40 + 0.54
3.92 + 1.07

Total
%
12.76 + 1.44
16.78 + 0.81*
14.64 + 1.71
19.82 + 1.36*

Source of Variation

Insol. Fructans

Sucrose

Total

Species
Treatment
Species x Treatment
Block
Block x Treatment

1.02 (0.41)
2.32 (0.11)
1.36 (0.25)
0.21 (0.42)
0.01 (0.99)

1.71 (0.21)
3.47 (0.04)
3.73 (0.005)
0.01 (0.99)
0.01 (0.99)

7.29 (0.003)
41.19 (< 0.0001)
4.11 (0.003)
0.72 (0.24)
1.72 (0.04)

203

Table 18. Experiment 2 (2011) means for unmanipulated plants (percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (glucose, starch, fructose, soluble fructans) measured around anthesis, on stem tissue from annual wheat,
perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” =
67% of control, “Control”, and “High” = 200% of control) at approximately 1 week pre-anthesis The experiment was conducted at
Kellogg Biological Station, Hickory Corners, MI. Only means for unmanipulated (“Control”) plants are shown.
Species
Annual Wheat
Perennial Wheat
Annual Rye
Perennial Rye

Source of Variation
Species
Treatment
Species x Treatment
Block
Block x Treatment

Glucose
%
2.45 + 0.13
5.89 + 0.40**
2.90 + 0.22
2.21 + 0.55

Starch
%
0.29 + 0.11
0.60 + 0.14
0.27 + 0.05
0.20 + 0.06

Fructose
%
4.73 + 0.62
3.90 + 1.01
4.42 + 1.27
4.24 + 1.00

Sol. Fructans
%
10.83 + 1.31
10.39 + 0.71
6.14 + 1.73
2.72 + 1.15

Glucose

Starch

Fructose

Sol. Fructans

13.04
(0.0002)
0.09 (0.91)
0.36 (0.90)
0.01 (0.99)
0.23 (0.41)

6.75 (0.004)

2.11 (0.14)

6.55 (0.005)

1.30 (0.28)
1.52 (0.18)
0.56 (0.28)
0.37 (0.36)

0.78 (0.48)
1.54 (0.17)
0.01 (0.99)
0.01 (0.99)

19.38 (< 0.0001)
1.25 (0.28)
0.01 (0.99)
0.01 (0.99)

204

Table 19. Experiment 2 (2011) means for unmanipulated plants (percent of dry wt + SE), F-values and P-values for various
carbohydrate pools (insoluble fructans, sucrose, and total NSC) measured around anthesis, on stem tissue from annual wheat,
perennial wheat, annual rye and perennial rye. All plants were subjected to three levels of source/sink ratio treatments (“Low” =
67% of control, “Control”, and “High” = 200% of control) at approximately 1 week pre-anthesis The experiment was conducted at
Kellogg Biological Station, Hickory Corners, MI. Only means for unmanipulated (“Control”) plants are shown.
Species
Annual Wheat
Perennial Wheat
Annual Rye
Perennial Rye

Insol. Fructans
%
0.49 + 0.10
0.88 + 0.24
0.33 + 0.04
0.67 + 0.20

Sucrose
%
2.05 + 0.33
5.06 + 2.02
2.85 + 0.63
4.25 + 1.00

Total
%
20.85 + 1.46
26.66 + 2.17*
16.91 + 1.36
15.94 + 2.47

Source of Variation

Insol. Fructans

Sucrose

Total

Species
Treatment

0.40 (0.76)
0.46 (0.63)

Species x Treatment
Block
Block x Treatment

0.78 (0.59)
0.01 (0.99)
0.01 (0.99)

3.62 (0.04)
10.48 (<
0.0001)
5.89 (< 0.0001)
0.29 (0.39)
0.63 (0.26)

11.65 (0.0003)
24.27 (<
0.0001)
2.60 (0.02)
0.51 (0.30)
0.01 (0.99)

205

Table 20. Pilot Experiment (2009) analysis of variance results for photosynthetic rates (Amax) and triose phosphate utilization
rates (TPU) measured 7-10 days following treatment imposition, as well as seed yield and single seed mass measured the
following year, for three perennial cereals. Perennial accessions include ‘Rival’ perennial rye, ‘WSU-P19’ perennial wheat and
‘WSU- P3’ perennial wheat. All plants were subjected to three levels of source/sink ratio treatments (“Low” = 67% of control,
“Control”, and “High” = 200% of control) at approximately 1 week pre-anthesis. The experiment was conducted at Kellogg Biological
Station, Hickory Corners, MI.

Sources of Variation
TPU

Amax

Seed Yield

Single Seed Mass

Accession

F
18.02

p
< 0.0001

F
0.76

p
0.45

F
0.25

p
0.78

F
1.18

p
0.36

Treatment
Accession x Treatment
Block
Accession x Block

0.51
1.90
0.65
0.44

0.52
0.19
0.25
0.33

63.86
2.48
0.06
0.80

< 0.0001
0.13
0.47
0.21

1.41
0.68
0.01
1.07

0.26
0.61
0.99
0.14

1.50
1.46
0.01
0.79

0.25
0.25
0.99
0.21

206

Table 21. Pilot Experiment (2009) means (+ SE) for photosynthetic rates (Amax) and triose phosphate utilization rates (TPU)
measured 7-10 days following treatment imposition, as well as seed yield and single seed mass measured the following year, for
three perennial cereals. Perennial accessions include ‘Rival’ perennial rye, ‘WSU-P19’ perennial wheat and ‘WSU- P3’ perennial
wheat. All plants were subjected to three levels of source/sink ratio treatments (“Low” = 67% of control, “Control”, and “High” =
200% of control) at approximately 1 week pre-anthesis. The experiment was conducted at Kellogg Biological Station, Hickory
Corners, MI. Symbols ‘*’ and ‘**’ represent significant differences from the control, following Dunnett’s post-hoc testing, at P = 0.05
-2 -1
and P = 0.01 respectively. Units for Amax and TPU are µmol m s of CO2 and triose phosphate respectively, while seed yield is
-1

expressed in grams m and seed mass in mg.

Accession

Low

Perennial Wheat P3
Perennial Wheat P19

21.9 + 1.1**
21.3 + 0.5**

Perennial Wheat P3
Perennial Wheat P19

14.0 + 1.5**
13.5 + 0.5**

Perennial Wheat P3
Perennial Wheat P19
Perennial Rye

21.8 + 2.0
19.8 + 1.4
28.5 + 5.0

Perennial Wheat P13
Perennial Wheat P19
Perennial Rye

23.7 + 0.4
16.5 + 0.1
18.2 + 0.2

Control
---Amax --17.4 + 0.9
18.3 + 0.4
---TPU--11.3 + 1.4
12.3 + 0.2
---Seed Yield--21.8 + 0.9
19.4 + 1.5
20.0 + 2.2
---Single Seed Mass --16.7 + 0.2
16.2 + 0.1
18.5 + 0.1

207

High

17.2 + 0.5
19.1 + 1.4
10.5 + 0.33**
10.7 + 0.4**
21.8 + 0.9
20.0 + 1.9
21.1 + 2.4
17.6 + 0.2
16.7 + 0.2
18.2 + 0.1

Figure 14. Light-saturated photosynthetic rate (Amax) as a function of source/sink ratio in
annual wheat (AW), perennial wheat (PW), annual rye (AR), and perennial rye (PR),
measured in Experiment 1 (2010). Photsynthetic rates (Amax) was measured 5-8 days after
treatment imposition on two different dates, and averaged across dates. Three treatments are
compared: Low (source/sink ratio = 67% of control), Control (100% of control) and High (200%
of control). The symbols “*” and “**” represent differences from the control with P < 0.05 or P
< 0.01 respectively. Data are from Experiment 1 (2010).

208

Figure 15. Light-saturated photosynthetic rate (Amax) as a function of source/sink ratio in
annual wheat (AW), perennial wheat (PW), annual rye (AR), and perennial rye (PR),
measured in Experiment 2 (2011). Photosynthetic rates (Amax) was measured 5-8 days after
treatment imposition on two different dates, and averaged across dates. Three treatments are
compared: Low (source/sink ratio = 67% of control), Control (100% of control) and High (200%
of control). The symbols “*” and “**” represent differences from the control with P < 0.05 or P
< 0.01 respectively.

209

Figure 16. Leaf nonstructural carbohydrate content (NSC) as a function of source/sink
ratio in annual wheat (AW), perennial wheat (PW), annual rye (AR) and perennial rye
(PR), in a 2011 field experiment. Low (source/sink ratio = 67% of control), Control (100% of
control) and High (200% of control). * and ** represent differences from the control with P <
0.05 or P < 0.01 respectively.

210

Figure 17. Stem nonstructural carbohydrate content (NSC) as a function of source/sink
ratio in annual wheat (AW), perennial wheat (PW), annual rye (AR) and perennial rye
(PR), in a 2011 field experiment. Low (source/sink ratio = 67% of control), Control (100% of
control) and High (200% of control). * and ** represent differences from the control with P <
0.05 or P < 0.01 respectively.

211

Figure 18. Kernel mass as a function of source/sink ratio in annual wheat (AW), perennial
wheat (PW), annual rye (AR) and perennial rye (PR), in a 2010 field experiment. Kernel
mass is compared between three treatments: Low (source/sink ratio = 67% of control), Control
(100% of control) and High (200% of control), each averaged over treatments imposed at 2-3
time points. * and ** represent differences from the control with P < 0.05 or P < 0.01
respectively.

212

Figure 19. Light-saturated photosynthetic rate (Amax) in flowering and non-flowering
microplots of perennial wheat, during the fall post-sexual cycle regrowth period, in a 2011
field experiment. Pink and red represent flowering and non-flowering plants respectively.

213

APPENDIX B. Pictures of Perennial Cereal Experiments

214

Figure 20. Perennial wheat plots at late elongation stage, in a spring 2011 study of
ecophysiological differences in annual and perennial cereals (Chapter 3, Experiment 2).

215

Figure 21. Perennial wheat plots at early elongation stage, in a spring 2011 study of
ecophysiological differences in annual and perennial cereals (Chapter 3, Experiment 2).

216

Figure 22. Intermediate wheatgrass plants at early elongation stage, in a spring 2011 study
of ecophysiological traits in intermediate wheatgrass (Chapter 3, Experiment 3).

217

Figure 23. “Low” treatment (source / sink ratio = 67% of control) in perennial wheat
plants, imposed at boot stage, in a 2011 study of physiological responses to source/sink
manipulation in annual and perennial cereals (Chapter 4, Experiment 2).

218

Figure 24. “High” treatment (source / sink ratio = 200% of control) in perennial rye plants,
imposed at boot stage, in a 2011 study of physiological responses to source/sink
manipulation in annual and perennial cereals (Chapter 4, Experiment 2).

219

Figure 25. Photosynthetic measurements in perennial wheat plants, in a 2011 study of
physiological responses to source/sink manipulation in annual and perennial cereals
(Chapter 4, Experiment 2).

220

REFERENCES

221

REFERENCES

Ainsworth EA, Rogers A, Nelson R, Long SP. 2004. Testing the “source sink” hypothesis of
downregulation of photosynthesis in elevated CO2 in the field with single gene substitutions in
Glycine max. Agricultural and Forest Meteorology 122: 85-94.
Alvaro F, Royo C, del Moral LFG, Villegas D. 2008. Grain filling and dry matter translocation
responses to source-sink modifications in a historical series of durum wheat. Crop Science 48:
1523-1531.
Amthor JS. 2001. Effects of atmospheric CO2 concentration on wheat yield: review of results
from experiments using various approaches to control CO2 concentration. Field Crops Research
73: 1-34.
Andersen R, Sorensen A. 1999. An enzymatic method for the determination of fructans in foods
and food products - Comparison of the results by high performance anion exchange
chromatography with pulsed amperemetric detection. European Food Research and Technology
210: 148-152.
Arp WJ. 1991. Effects of source-sink relations on photosynthetic acclimation to elevated CO2.
Plant, Cell & Environment 14: 869-875.
Borras L, Slafer GA, Otegui ME. 2004. Seed dry weight response to source-sink manipulations
in wheat, maize and soybean: a quantitative reappraisal. Field Crops Research 86: 131-146.
Cruz-Aguado JA, Reyes F, Rodes R, Peres I, Dorado M. 1999. Effect of source-to-sink ratio on
14
partitioning of dry matter and C-photoassimilates in wheat during grain filling. Annals of
Botany 83: 655-665.
DaMatta FM, Cunha RL, Antunes WC, Martins SCV, Araujo WL, Fernie AR, Moraes GABK.
2008. In field-grown coffee trees source-sink manipulation alters photosynthetic rates,
independently of carbon metabolism, via alterations in stomatal function. New Phytologist 178:
348-357.
Fan PG, Li LS, Duan W, Li WD, Li SH. 2010. Photosynthesis of young apple trees in response to
low sink demand under different air temperatures. Tree Physiology 30: 313-325.
Flore JA, Layne DR. 1999. Photoassimilate production and distribution in cherry. HortScience
34: 1015-1019.

Foyer C, Furbank R, Harbinson J, Horton P. 1990. The mechanisms contributing to
photosynthetic control of electron transport by carbon assimilation in leaves. Photosynthetic
222

Research 25: 83-100.

Franck N, Vaast P, Genard M, Dauzat J. 2006. Soluble sugars mediate sink feedback downregulation of leaf photosynthesis in field-grown Coffea arabica. Tree Physiology 26: 517-525.
Gaines MS, Vogt KJ, Hamrick JL, Caldwell J. 1974. Reproductive strategies and growth-patterns
in sunflowers (Helianthus). American Naturalist 108: 889-894.
Gandin A, Gutjahr S, Dizengremel P, Lapointe L. 2011. Source sink imbalance increases with
growth temperature in the spring geophyte Erythronium americanum. Journal of Experimental
Botany 62: 3467-3479.
Grossman JM, Sheaffer C, Wyse D, Bucciarelli B, Vance C, Graham PH. 2006. An assessment of
nodulation and nitrogen fixation in inoculated Inga oerstediana, a nitrogen-fixing tree shading
organically grown coffee in Chiapas, Mexico. Soil Biology and Biochemistry 38: 769-784.
Gucci R, Xiloyannis C, Flore JA. 1991. Gas exchange parameters, water relations and
carbohydrate partitioning in leaves of field-grown Prunus domestica following fruit removal.
Physiologia Plantarum 83(3): 497-505.
Harrison M, Kelman WM, Moore AD, Evans JR. 2010. Grazing winter wheat relieves
plant water stress and transiently raises photosynthesis. Functional Plant Biology 37: 726-736.
Helsel ZR, Thomas JW. 1987. Small grains for forage. Journal of Dairy Science 70: 2330-2338.
Hiyane R, Hiyane S, Tang AC, Boyer JS. 2010. Sucrose feeding reverses shade-induced kernel
losses in maize. Annals of Botany 106: 395-403.
Huebner FR, Bietz JA, Nelsen T, Bains GS, Finney PL. 1999. Soft wheat quality as related to
protein composition. Cereal Chemistry 76: 650-655.
Iglesias DJ, Lliso I, Tadeo FR, Talon M. 2002. Regulation of photosynthesis through source: sink
imbalance in citrus is mediated by carbohydrate content in leaves. Physiologia Plantarum 116:
563-572.
Jaikumar N, Snapp SS, Murphy K, Jones S. 2012. A field assessment of the agronomic potential
of two novel perennial cereal crops. Agronomy Journal In press.
Kasai M. 2008. Regulation of leaf photosynthetic rate correlating with leaf carbohydrate
status and activation state of Rubisco under a variety of photosynthetic source/sink balances.
Physiologia Plantarum 134: 216-228.
Karlsson PS. 1994. The significance of internal nutrient cycling in branches for growth and
reproduction of Rhododendron iapponicum. Oikos 70: 191-200.

223

Krapp A, Stitt M. 1995. An evaluation of direct and indirect mechanisms for the "sink
regulation" of photosynthesis in spinach: Changes in gas exchange, carbohydrates, metabolites,
enzyme activities and steady state transcript levels after cold girdling source leaves. Planta 195:
315-323.
Kuhbauch W, Thome U. 1989. Nonstructural carbohyudrates of wheat stems as infuleced by
sink-source manipulations. Journal of Plant Physiology 134: 243-250.
Layne DR, Flore JA. 1995. End-product inhibition of photosynthesis in Prunus-cerasus L. in
response to whole-plant source-sink manipulation. Journal of the American Society for
Horticultural Science 120: 583-599.
Ma Y, Li J, Li X, Tang X, Liang Y, Huang S, Wang B, Liu H, Yang X. 2009. Phosphorus
accumulation and depletion in soils in wheat–maize cropping systems: Modeling and validation.
Field Crops Research 110: 207-212.
Masnatta WJ, Ravetta DA. 2011. Seed-yield and yield components response to source sink ratio
in annual and perennial species of Lesquerella (Brassicaceae). Industrial Crops and Products 34:
1393-1398.
McCormick AJ, Cramer MD, Watt DA. 2006. Sink strength regulates photosynthesis in
sugarcane. New Phytologist 171: 759-770.
McCormick AJ, Cramer MD, Watt DA. 2008. Changes in photosynthetic rates and gene
expression of leaves during a source-sink perturbation in sugarcane. Annals of Botany 101: 89102.
McCormick AJ, Cramer MD, Watt DA. 2008. Regulation of photosynthesis by sugars in
sugarcane leaves. Journal of Plant Physiology 165: 1817-1829.
Murphy KM, Lyon SR, Balow KA, Jones SS. 2010. Post-sexual cycle regrowth and grain yield
in Thinopyrum elongatum x Triticum aestivum amphiploids. Plant Breeding 129: 480-483.
Niva M, Svensson BM, Karlsson PS. 2003. Nutrient resorption from senescing leaves of the
clonal plant Linnaea borealis in relation to reproductive state and resource availability.
Functional Ecology 17: 438-444.
Park SJ, Kim YK. 2011. Defruiting effect of young Fuyu persimmon (Diospyros kaki) on
assimilate partitioning in-season and early growth the next season. Scientia Horticulturae 130:
197-202.
Peng Y, Jiang GM, Liu XH, Niu SL, Liu MZ, Biswas DK. 2007. Photosynthesis, transpiration
and water use efficiency of four plant species with grazing intensities in Hunshandak Sandland,
China. Journal of Arid Environments 70: 304-315.
Reekie EG, Bazzaz FA. 1987. Reproductive effort in plants. 3. Effect of reproduction on
224

vegtative activity. American Naturalist 129: 907-919.
Reynolds M, Foulkes MJ, Slafer GA, Berry P, Parry MAJ, Snape JW, Angus WJ. 2009. Raising
yield potential in wheat. Journal of Experimental Botany 60: 1899-1918.
Reynolds MP, Pellegrineschi A, Skovmand B. 2005. Sink-limitation to yield and biomass: a
summary of some investigations in spring wheat. Annals of Applied Biology 146: 39-49.
Rho CW, Cha YH, Kim TS, Kwon KC, Park YK. 1990. Effect of defoliation on seed
development of rye and triticale cultivar. Korean Journal of Crop Science 35: 211-217.
Roumet C, Roy J. 1996. Prediction of the growth response to elevated CO2: A search for
physiological criteria in closely related grass species. New Phytologist 134: 615-621.
SAS Institute. 2008. The SAS System for Windows. 9.2. Cary, NC: SAS Institute, Inc.
Schaefer C, Simper H, Hofmann B. 1992. Glucose feeding leads to coordinated changes
in chlorophyll content, ribulose 1,5-bisphosphate carboxylase-oxygenase activity, and
photosynthetic activity in photoautotrophic suspension cultured cells of Chenopodium rubrum.
Plant, Cell and Environment 15: 343-350.
Sharkey TD, Bernacchi CJ, Farquhar GD, Singsaas EL. 2007. Fitting photosynthetic carbon
dioxide response curves for C3 leaves. Plant, Cell and Environment 30: 10351040.
Urban L., Léchaudel M. 2005. Effect of leaf-to-fruit ratio on leaf nitrogen content and net
photosynthesis in girdled branches of Mangifera indica L. Trees- Structure and Function 19(5):
564-571.
Usenik V, Orazem P, Stampar F. 2010. Low leaf to fruit ratio delays fruit maturity of 'Lapins'
sweet cherry on Gisela 5. Scientia Horticulturae 126: 33-36.
Van Tassel DL, DeHaan LR, Cox TS. 2010. Missing domesticated plant forms: can artificial
selection fill the gap? Evolutionary Applications 3: 434-452.
Vijn I, Smeekens S. 1999. Fructan: More than a reserve carbohydrate? Plant Physiology 120:
351-359.
Wullschleger SD. 1993. “Biochemical limitations to carbon assimilation in C3 plants: a
retrospective analysis of the A / Ci curves from 109 species.” Journal of Experimental Botany
44: 907-920.
Yan ST, Li XD, Li WD, Fan PG, Duan W, Li SH. 2011. Photosynthesis and chlorophyll
fluorescence response to low sink demand of tubers and roots in Dahlia pinnata source leaves.
Biologia Plantarum 55: 83-89.
225

Zadoks JC, Chang TT, Konzak CF. 1974. A decimal code for the growth stages of cereals. Weed
Research 14: 415-421.
Zhao M, Acuna TLB, Lafitte HR, Dimayuga G, Sacks E. 2008. Perennial hybrids of Oryza sativa
x Oryza rufipogon - Part II. Carbon exchange and assimilate partitioning. Field Crops Research
106: 214-223.
Zhenlin W, Yanping Y, Mingrong H, Hongming C. 1998. Source-sink manipulation effects on
postanthesis photosynthesis and grain setting on spike in winter wheat. Photosynthetica 35: 453459.
Zhou R, Quebedeaux B. 2003. Changes in photosynthesis and carbohydrate metabolism in
mature apple leaves in response to whole plant source-sink manipulation. Journal of the
American Society for Horticultural Science 128: 113-119.
Zhu C, Zhu J, Zeng Q, Liu G, Xie Z, Tang H, Cao J, Zhao X. 2009. Elevated CO2 accelerates
flag leaf senescence in wheat due to ear photosynthesis which causes greater ear nitrogen sink
capacity and ear carbon sink limitation. Functional Plant Biology 36: 291-299.

226

CHAPTER 5.
RESISTANCE TO FUSARIUM HEAD BLIGHT IN THREE PERENNIAL CEREAL
CROPS

227

Introduction

Novel perennial cereals are in the process of being developed as alternatives to annual
grain crops (Cox et al. 2006). Promising candidates for the cold temperate zone include perennial
wheat (Triticum aestivum x Thinopyrum spp.), perennial rye (Secale cereale x montanum), and
intermediate wheatgrass (Thinopyrum intermedium). Perennial wheat and perennial rye have
been developed by crossing annual wheat and rye with wild relatives (Acharya et al. 2004;
Murphy et al. 2010). Intermediate wheatgrass is currently being domesticated through selection
for high yield (Cox et al. 2006).
Because of their perennial life history, these emerging species face the challenge of
pathogen accumulation, since these plants live for multiple years and could potentially represent
a reservoir for disease. Some existing strategies to manage disease in annuals (e.g. crop rotation)
cannot be used with perennials. One serious disease threat to perennial cereals is Fusarium head
blight (scab, FHB) caused by Fusarium graminearum Schwabe (teleomorph Gibberella zeae:
Xue et al. 2009). Winter survival of Fusarium on crop residues is the primary inoculum source
(Zhang L. et al. 2011), which underscores the importance of selecting for FHB resistance in
perennial cereal crops. FHB resistance can be categorized into Type I (resistance to initial
infection), Type II (resistance to spread within an infected spike), Type III (resistance to kernel
infection), Type IV (yield tolerance) and Type V (ability to degrade mycotoxins: Mesterházy
1995; Schroeder and Christensen 1963).
Previous greenhouse studies of FHB resistance in Thinopyrum and in Thinopyrum / wheat
crosses indicate that the genus can contribute Type II resistance, though not all Thinopyrum /
wheat crosses display resistance and there is high variability within the F1 generation (Cai et al.

228

2005). For example, approximately 25% of wheat-alien species derivatives (mostly Thinopyrum /
wheat) in one study grouped with the ‘resistant’ checks ‘Sumai 3’, ‘Wangshuibai’ and ‘Alsen’
(Oliver et al. 2005) while in a later study 75% of derivatives showed equivalent resistance to
‘Sumai 3’ (Oliver et al. 2006). A further study of Th. intermedium / wheat derivatives found that
one out of the seven studied, had better type II resistance than ‘Sumai 3’ (Han et al. 2003). Th.
intermedium addition lines into a background of ‘Vilmorin’ wheat reduced average FHB disease
ratings to 3.6 (on a 0-9 scale) compared to 4.6 for ‘Vilmorin’ and 2.8 for Thinopyrum (Gilbert et
al. 1996). Similarly, a reduction of 75% to 90% in FHB as compared with the annual wheat
parent followed the additions of Th. elongatum chromosomes within a background of susceptible
‘Chinese Spring’ (Wang et al. 2010).
Most studies of FHB resistance have not considered whether the crosses studied
displayed perenniality or high seed yield: this makes them helpful for improving FHB resistance
in annual wheat, but less so for understanding the potential for FHB resistance in a new perennial
cereal species. Until now, no FHB resistance studies have been carried out on the set of
Thinopyrum and Thinopyrum / wheat lines which have undergone intense selection over the last
ten years for perennial regrowth as well as high seed production. Furthermore, to the best of our
knowledge, none have been conducted on perennial rye. Because FHB resistance in Thinopyrum
/ wheat amphiploids has been shown to vary greatly among genotypes, it is important for
breeding purposes to determine whether these specific lines show resistance, or whether
selection for high yield in a low rainfall environment may have caused the loss of FHB
resistance, through narrowing the genetic base and subsequent genetic drift. It is also possible, if
Thinopyrum-derived FHB resistance is genetically linked to Thinopyrum QTLs that depress
yield, that selecting for higher yield could result in lowered FHB resistance; it is also possible

229

that there may be direct metabolic tradeoffs between resistance and yield.
Our main goal in this study, which distinguishes it from previous studies of FHB in
Thinopyrum / wheat amphiploids, was to determine whether a set of perennial cereal lines (
Thinopyrum / wheat amphiploids, perennial rye, and Thinopyrum intermedium), that had
undergone several cycles of selection for high yield and regrowth, would show the high levels of
resistance comparable to those previously shown in some Thinopyrum / wheat amphiploids and
in forage accessions of Th. intermedium. Secondarily, our goal was to identify lines that are
highly resistant to FHB, for breeding purposes. Our guiding hypotheses were that some of the
perennial cereal lines would show type II resistance equal to the “Resistant” checks.

Materials and Methods

Experimental Design
We conducted two experiments to determine FHB resistance and susceptibility in a range
of perennial cereal accessions. Experiment 1 examined responses to FHB inoculation in the
greenhouse. Experiment 2 considered natural occurrence of FHB on plants growing in the field,
during a year characterized by a cool and wet spring. Experiment 1 included more accessions
than Experiment 2, and compared against both positive and negative controls (susceptible and
resistant wheat lines): Experiment 2 included a subset of lines from Experiment 1, observed the
prevalence of FHB on non-inoculated plants, and compared against a single, moderately
susceptible annual wheat cultivar adapted to Michigan climate.

230

Experiment 1: Plant Materials
We used 17 genotypes in our study (Table 22). Four annual wheat checks were included,
two resistant to FHB and two susceptible. Resistant checks included winter wheat ‘Truman’ (PI
634824; McKendry et al. 2005) and spring wheat ‘Sumai 3’ (PI 481542). Sumai 3 is currently
considered a highly resistant wheat cultivar and the most widely used source of resistance traits
for breeding purposes (Zhou et al. 2010). Susceptible checks included winter wheat
‘Ambassador’ (PI 656845; Lewis et al. 2010) and spring wheat ‘Wheaton’ (PI 469271: Busch et
al. 1984). ‘Ambassador’ was a white wheat cultivar bred at MSU, which yielded 7% higher than
the mean across 29 lines in a 2007 variety trial (Lewis J. M. unpublished data). ‘Truman’ was a
soft red variety, relatively late maturing, bred at the University of Missouri and released in 2003;
it yielded 6% lower than the mean across 29 lines in the aforementioned 2007 Michigan State
variety trial.
Eleven Thinopyrum / wheat amphiploids, one perennial rye and one intermediate
wheatgrass line were chosen to represent a wide range of ancestry.
Four of these were chosen, in fall 2010, out of a preliminary screening of 15
Thinopyrum / wheat amphiploids developed at Washington State University (S.S.
Jones and K.M. Murphy, personal communication, 2009). We selected four lines
out of the 15 on the following basis: P10-2009 demonstrated the highest yield and
P15-2009 the most vigorous PSCR, while P3-2009 and P5-2009 were moderate
yielding lines. The parentage of these lines was Thinopyrum elongatum / Triticum
aestivum cv. ‘Chinese Spring’ // T. aestivum cv. ‘Madsen’ /3/ T. aestivum /
Thinopyrum elongatum cv. ‘Spitzer’ (Oliver et al. 2008).

231

We included four Thinopyrum / wheat amphiploids from earlier materials obtained
from WSU in 2005, numbered P3-2005, P11-2005, P15-2005, and P19-2005.
These four were part of a set of 31 F5 lines tested at Pullman, WA in 2005
(Murphy et al. 2010) that had demonstrated the ability to perenniate in Michigan
(Jaikumar et al., unpublished data). The parentage of these lines was Thinopyrum
elongatum / T. aestivum cv. ‘Chinese Spring’ // T. aestivum cv. ‘Madsen’.
Three perennial wheat accessions (LI-19, LI-47 and LI-48) were picked from a set
of accessions developed at the Land Institute, (Salina, KS) in 2009. LI-19 was a
hybrid of Triticum durum ‘Afuwan’ (PI 634318) x Thinopyrum junceiforme (PI
414667), while LI-47 and LI-48 were both derived from crosses between wheat
(PI 386154) and pubescent wheatgrass (Thinopyrum intermedium barbulatum:
DeHaan, L.R. personal communication). All three were selected from a larger set
of lines, on the basis of demonstrated PSCR in a pilot trial. Thus the Thinopyrum /
wheat amphiploids contained ancestry from three different Thinopyrum species.
We included one perennial rye cultivar, Secale cereale x montanum, cv. ‘Rival’
(Peters Seed Company, Oregon, USA).
Finally, we included one intermediate wheatgrass line (Thinopyrum intermedium).
The wheatgrass, ‘TLI-C1’ was from a breeding population at the Land Institute
following a cycle of selection for larger grain size (DeHaan LR, personal
communication).

232

Experiment 1: Planting Protocol
All accessions were planted on 7 planting dates at the MSU East Lansing campus, every
4 days from November 1 through December 7, 2010. Every accession was planted on each
planting date: the purpose was to allow a reasonable number of inoculation dates for each
accession, so that each inoculation date could be treated as a replicate. Fifteen seeds of each
accession were planted into an approximately 500 mL container filled with Perlite potting mix,
and fertilized with around 10 g Osmocote slow-release pellets (14-14-14 NPK). After 10 days the
pots were transported into a cold chamber for vernalization, and were maintained for 8 weeks at
o

6 C with 10.5 hours/day lighting.
o

Following vernalization, plants were removed and kept at room temperature (18 C) for 3
days, before being transplanted. Ten plants from each accession were selected and transplanted
into 700 mL conical pots. These pots were fertilized with approximately 10 g of Osmocote
pellets and then placed in the Michigan State University glasshouses (East Lansing, MI). Plants
were subject to natural light through the sun roof and supplemented with artificial lighting for a
total 16 hour day: they were watered daily and fertilized weekly with 300 ppm Peters
Professional Water Soluble Fertilizer (NPK = 20:20:20) added to the water. Insects were
controlled as needed following Michigan State University recommended protocol: tetraconazole
was sprayed in early February, sulfur in early April and mid-April, and pyrometrazine in early
May.
Overall plants grew between 40 – 90 cm tall and had multiple flowering stems. As an
indication of overall plant vigor, chlorophyll readings (measured using the Minolta SPAD-502
chlorophyll meter) were comparable to plants growing in the field (readings of 50 – 55 on the
233

SPAD).

Experiment 1: Inoculation and Observation of Response
Inoculum was generated by culturing Fusarium graminearum isolate PH1 (NRRL31084)
in Carboxymethyl Cellulose liquid medium (Cappelli and Peterson 1965). Ten microliters of
stock culture was added to autoclaved media (100ml in 250 ml flasks). The culture was shaken at
200-250 rpm and 25° C for four days. Mycelia were removed by filtration with a sterile
cheesecloth. Spores were counted using a haemocytometer and the culture was diluted to a final
5

concentration of 5 x 10 spores/ml.
We used the single floret inoculation method to inoculate the plants (Schroeder and
Christensen 1963; Mackintosh et al. 2007). Here a ‘spike’ refers to the entire inflorescence on a
single wheat stalk; a ‘spikelet’ refers to one of the multiple smaller bundles of florets which
make up a spike, protected by glumes; and a floret refers to an individual flower within a
spikelet. Twelve inoculation dates were used, between mid-March and mid-May. On each date,
plants at mid-anthesis were selected and one floret per plant (generally midway along the spike)
5

was infected with 10 µL of inoculum containing Fusarium graminearum spores (5 x 10 spores /
mL) using a micropipette. The inoculated spikes were then covered with plastic sandwich bags
and tied shut to maintain a humid environment for 48 hours, after which the bags were removed
and the plants monitored for disease expression.
After 14 days, plants were assessed for disease symptoms. The two parameters measured
were spikelet infection (proportion of diseased spikelets/ total spikelets) and rachis infection
(proportion of diseased rachis segments/ total rachis segments). Here a ‘rachis segment’ refers to

234

the length along the rachis between one spikelet and the adjacent spikelet. A spikelet, or a
segment of rachis, was considered diseased if it showed brownish discoloration (this varied from
light tan color to dark brown, compared with a greenish color for healthy tissue), or visible
fungal growth. Any significant sign of brown discoloration was interpreted as infection. Infection
was scored as a proportion between 0 and 1, e.g. number of diseased spikelets / number of total
spikelets. Because only plants at mid-anthesis were selected, not every genotype is represented at
each inoculation date. In general, between 0-5 plants per genotype were inoculated at each
planting date (generally 2-3) and each plant was represented at between 6-11 inoculation dates
(generally 7-9).

Experiment 1: Post-sexual Cycle Regrowth
In addition to quantifying FHB resistance among the accessions in our study, we also
carried out a field trial to measure post-sexual cycle regrowth (PSCR). The purpose of this trial
was to determine whether the highly FHB-resistant accessions in our study also showed good
regrowth. Part of the goal of this experiment was to identify useful, FHB-resistant accessions for
use in future breeding efforts in perennial wheat. However, in order to determine that these lines
would be well suited for breeding efforts, it is necessary to establish that they show adequate
levels of regrowth. We planted eight lines (those for which sufficient seed existed: Table 17) in a
field trial to quantify PSCR. The area for planting was chisel plowed on Sept 29, 2010 and
-1

fertilized on Oct 11 with 100 kg ha nitrogen in the form of poultry manure. A soil finisher was
-1

used to prepare plots on Oct 12, and on Oct 13 each line was planted at 1.75 million seeds ha ,
2

in 2.30 m plots, replicated 4 times, with a small plot planter (Allen Machine Company: Nevada,

235

IA). Weeding was carried out three times. Plots were harvested on Aug 1, 2011. On September
29, 2011, PSCR was visually evaluated by estimating the percentage of plants that showed green
regrowth in the central two rows of a plot.

Experiment 2: FHB Symptoms on Naturally Infected Plants in the Field
In addition to our greenhouse experiment investigating Type II resistance, we also sought
to observe the susceptibility of perennial cereals to FHB infection under non-treated, ‘natural’
field conditions. This was a smaller, observational study, supplementary to Experiment 1. Here
we did not explicitly distinguish between Type I and Type II resistance, but rather observed the
incidence, severity and disease index of FHB on non-inoculated, naturally infected plants.
In this experiment, we used the following subset of seven genotypes: the four lines
received from Washington State University in 2009 (P3-2009, P5-2009, P10-2009, P15-2009),
one line received in 2005 (P19-2005), the perennial rye (PR-Rival) and the intermediate
wheatgrass accession (TLI-C1). As described above, they were chisel plowed in September
2010, fertilized with 100 kg ha

-1

-1

N, soil finished and then planted at 1.7 million seeds ha in a

IBD design. The four lines P3-2009, P5-2009, P10-2009 and P15-2009 were replicated four
times, while the other three test lines were replicated six times. The annual wheat cultivar against
which perennial rye, intermediate wheatgrass and the wheat / Thinopyrum accessions were
compared was the cultivar ‘Frankenmuth’, used as a benchmark in breeding studies at Michigan
State University (Huebner et al., 1999; Lewis JM, personal communication). This was the same
annual wheat cultivar used in Chapters 2-4 of this dissertation, and is considered to be
moderately susceptible to FHB; it was replicated six times. Plots were weeded three times during
the 2011 season. Anthesis date for each plot was observed through regular observations during

236

late spring and early summer, and generally varied between June 6 (for Frankenmuth) to July 4
(for intermediate wheatgrass).
FHB symptoms were assessed 21 days after anthesis on each plot. We randomly chose
two ‘typical’ microplots within each plot, generally located towards the center, and assessed FHB
symptoms on 10 plants in each microplot. Characteristic whitish to tan discoloration, or dark
brown streaking similar to those seen on FHB inoculated plants in the greenhouse, were
considered symptomatic of FHB. We recorded three parameters:
Number of plants showing symptoms / number of total plants (incidence)
Number of infected spikelets / number of total spikelets on each plant (severity)
Incidence x average severity per plant (disease index).

Statistical Analysis
In Experiment 1, disease symptoms were analyzed as an incomplete block design with
PROC MIXED in SAS, using a two-way ANOVA with genotype as a fixed factor and date as a
random factor. Disease symptoms and PSCR were logit-transformed:
x’ = log [(0.1 + x) / (1.1 – x) ].
Pairwise comparisons were made individually against the four annual wheat checks,
using the Dunnett procedure. To increase power, the three other checks were omitted from the
analysis during each comparison to a check. Statistical significance was assessed at α = 0.05.
We considered a genotype “Resistant” if its levels of rachis and spikelet infection were
both statistically similar to ‘Truman’ and ‘Sumai 3’ and different from ‘Ambassador’ and
‘Wheaton’. “Susceptible” genotypes had levels of rachis and spikelet infection similar to both
‘Ambassador’ and ‘Wheaton’ and different from ‘Truman’ and ‘Sumai 3’. Genotypes which were
237

statistically equivalent to a resistant check in one test and to a susceptible check in another, or
were consistently different than both, were considered “Intermediate”. To determine whether
PSCR was correlated with spikelet or rachis infection, we treated each genotype as one data
point, and carried out a regression of first logit-transformed average spikelet infection and then
logit-transformed average rachis infection (y’) against logit-transformed PSCR (x’). Values
reported in Table 1 and Fig. 1-2 are untransformed means.
In Experiment 2, disease index, severity and incidence were transformed using a logit
transformation, and analyzed as an incomplete block design ANOVA, with genotype as a fixed
factor and block as a random factor. Comparisons of each genotype against ‘Frankenmuth’ were
made using Dunnett’s test, with α = 0.05.

Results

Experiment 1: Overall ANOVA
Genotype strongly affected the percentage of infected spikelets in the greenhouse (F =
15.35, p < 0.0001) but inoculation date did not (F = 1.44, p = 0.075). Similarly the percentage of
infected rachis was significantly affected by genotype (F = 10.12, p < 0.0001), but not by
inoculation date (F = 1.55, p = 0.060).

Experiment 1: Resistant and Susceptible Lines
P3-2005, P11-2005, P15-2005 and P19-2005 all demonstrated a similar level of
resistance to ‘resistant’ checks (Sumai 3 and Truman) in terms of spikelet infection (Table 23,
Figure 26) and rachis infection (Table 24, Figure 27), as did Thinopyrum intermedium TLI-C1,

238

and LI-19. Two other lines, P10-2009 and P3-2009, grouped with the ‘susceptible’ checks. The
other lines (Rival Rye, LI-47, LI-48, P15-2009 and P5-2009) generally showed intermediate
levels of spikelet infection (Fig. 26) and rachis infection (Fig. 27), with P15-2009 being the
lowest and P5-2009 the highest, The annual ‘resistant’ checks showed very low levels of rachis
and spikelet infection (7-14%), while the ‘susceptible’ checks were in the 80-90% range. The one
perennial rye genotype was not significantly different from the perennial wheat lines. The Th.
intermedium line bred at the Land Institute for high grain production (TLI-C1) had extremely
low FHB infection, comparable with the resistant wheat lines, and displayed excellent type II
resistance (around 12% of spikelets infected). P15-2009 grouped with the resistant checks in
terms of spikelet infection and with the intermediate group in terms of rachis infection, but
otherwise the two parameters always agreed.

Experiment 1: Regrowth
PSCR ratings ranged from zero regrowth in three WSU-2009 lines to 100% in TLI-C1
(Table 22). As an exploratory procedure, we ran linear regressions of spikelet and rachis
infection against PSCR to determine if there was a correlation. We found that spikelet infection
was negatively correlated with PSCR, i.e. less disease symptoms was associated with greater
PSCR. A 10% increase in the odds of a plant regrowing corresponded to a 4.5% decrease in the
2

odds of a particular spikelet on that plant being infected (y’ = – 0.454x’ – 0.098; r = 0.788, df
= 6, p = 0.004). Similarly, percentage rachis infection was negatively correlated with PSCR,
with a 10% increase in the odds of a plant regrowing being associated with a roughly 5%
decrease in the odds of a particular rachis segment being infected. For the 8 lines in our study,
2

the observed relationship was y’ = – 0.479x’ – 0.095 (r = 0.723, df = 6, p < 0.01). Lines P15239

2005, P19-2005 and TLI-C1 combined very high PSCR with high levels of resistance.

Experiment 2: Natural Occurrence of FHB in the Field
Incidence varied between genotypes (p < 0.0001) as did severity (p = 0.049) and disease
index (p = 0.0007). Intermediate wheatgrass (TLI-C1) showed 86% lower incidence (t = 3.87, p
< 0.01), 74% lower severity (t = 4.74, p < 0.01) and 79% lower disease index (t = 4.02, p < 0.01)
than ‘Frankenmuth’, and was the only genotype to outperform ‘Frankenmuth’ on all three
parameters. Perennial wheat accession P19-2005 had 86% lower incidence than ‘Frankenmuth’,
but equal severity and disease index. The line P5-2009 had 360% greater incidence (t = 7.35, p <
0.01) and 380% greater disease index than ‘Frankenmuth’ (t = 3.16, p < 0.05), while perennial
rye, P3-2009 and P10-2009 generally had similar disease symptoms. Thus, in partial
corroboration of our results from the greenhouse, perennial rye was found to be moderately
susceptible to FHB, while intermediate wheatgrass was highly resistant and perennial wheat
accessions P19-2005 and P15-2009 were moderately resistant (Table 25).

Discussion

Our study found that six of the accessions studied (Th. intermedium ‘TLI-C1’, wheat /
Th. junceiforme, and four wheat / Thinopyrum elongatum F5 lines, from 2005) showed resistance
equivalent to the resistant checks. Other lines were highly or moderately susceptible. Our
findings were consistent with studies that showed that addition or substitution of Thinopyrum
chromosomes into annual wheat could confer FHB resistance (Oliver et al. 2005; Gilbert et al.
240

1996). We found high levels of FHB resistance in 6/11 of the hybrids, despite having undergone
intense selection for high seed yield and perenniality in the field, within a low-FHB environment.
Selection for yield, up to 2005, did not cause loss of FHB resistance.
Perennial rye showed intermediate resistance rather than grouping with the resistant
checks. While no previous work has considered impacts of FHB on perennial rye, or of other
diseases on this particular cultivar, studies do suggest the perennial rye cultivars ‘ACE-1’ and
‘Permontra’ have shown little susceptibility to disease in the field, with the notable exception of
ergot (Claviceps purpurea: Acharya et al. 2004; Cox et al. 2005). In contrast to other common
regional pathogens, FHB (along with ergot) appears from our study to be an important threat that
perennial rye breeders must overcome.
Thinopyrum-derived resistance to FHB is strongly influenced by QTLs on chromosome
7el2 of Th. elongatum and Th. intermedium, as well as by chromosome 2 of Th. intermedium.
The QTL on 7el2 is located between markers XBE445653 and Xcfa2240, shows partial
dominance, and explains 30% of variation in the F2 generation (Shen et al. 2004; Shen and Ohm
2007; Zhang X. et al. 2011). There is evidence that QTLs on the homoeologous annual wheat
chromosomes 7A, 7B or 7D condition susceptibility to FHB, based on substitution experiments
(Fu et al. 2011). Chromosome 2D of Th. intermedium has also been found to contribute to FHB
resistance in wheat / Thinopyrum amphiploids (Fedak and Han, 2005).
Neither of these two QTLs appears to be directly linked to the ability to perenniate, which
is controlled by genes on the short arm of chromosome 4E from Thinopyrum (Lammer et al.
2004). It is unlikely, therefore, that FHB resistance has strong genetic links to the ability for
post-sexual cycle regrowth. Resistance to FHB spread in wheat is influenced to some extent by
QTLs on chromosomes 4A , B, and D, homoeologous to the 4E chromosome that confers

241

perenniality (Buerstmayr et al. 2009; Liu et al. 2009). Thus is it possible there might be some
weak links between perenniality and regrowth, as our study indicated, that might be linked to the
presence of chromosome 4E from Thinopyrum.
FHB resistance has, in the past, been linked to perennial growth habit in a range of wild
Triticeae. Perennial Hordeum spp. ranked 0 or 1 (completely or highly resistant) on a 0-4 scale,
while annual wild Hordeum spp. were all highly susceptible, and most cultivated Hordeum
accessions tended to rank 2-3 on the scale. Perennial genera (e.g. Agropyron) predominated
among the most resistant accessions, while the least resistant tended to be annual genera. Rather
than propose a direct mechanistic link, however, the authors suggested that perenniality had been
selected for under the same cool moist conditions that favored FHB, while annual Triticeae,
associated with drier climates, had not been subject to selection for FHB resistance (Wan et al.
1997). Similarly, it is likely that in our study, the strongly perennial lines showed greater FHB
resistance simply because they had more overall genetic material from Thinopyrum: as yet there
is no data or theoretical reason to believe that genes controlling PSCR are closely linked to the
most important QTLs involved in FHB resistance. It is unlikely that the traits are closely linked,
and therefore selection for high PSCR in future may not necessarily result in high FHB
resistance. Explicit screening for FHB resistance will be necessary to ensure good resistance in
perennial wheat lines in future. We did find that intermediate wheatgrass accession TLI-C1, as
well as the perennial wheat lines P15-2005 and P19-2005 all showed a combination of high FHB
resistance (comparable to the ‘Resistant’ annual wheat checks) and high PSCR (above 80%).
Thus these three accessions would seem to provide good parents for breeding efforts directed
towards developing FHB-resistant, strongly perennial cereal lines. The perennial ‘Rival’ rye,
however, shows moderately high susceptibility to FHB, thus using ‘Rival’ for future breeding

242

efforts may make it more difficult to achieve good levels of FHB resistance.
Our comparison of susceptibility to natural infection in the field has numerous
limitations. Most importantly, a number of environmental and morphological factors could
contribute to the higher resistance seen in intermediate wheatgrass and some perennial wheat
lines. To begin with, there were pronounced phenological differences between flowering dates
for annual wheat vs. the perennial species. Annual wheat flowered around June 6, approximately
7-14 days before first year perennial wheat, and 24 days before intermediate wheatgrass. June
2011 was characterized by generally warm and dry weather, and thus it is likely that spikes were
generally drier around the time of perennial wheat and intermediate wheatgrass flowering
compared to annual wheat. Although plant height was not measured, casual observations also
suggested that perennial wheat and wheatgrass were taller than annual wheat: as plant height is
associated with resistance to FHB, this could have contributed to lower infection in the perennial
species. Finally, the perennial species generally had tougher and thicker glumes (much tougher in
the case of intermediate wheatgrass, such that threshing was very difficult, and somewhat
tougher in the case of perennial wheat). Tougher, more recalcitrant glumes that wrapped more
tightly around the floret could also present a physical barrier to FHB infection. In addition, there
might be physiological mechanisms that suppress the spread of FHB infection. Thus it is
impossible at this point to disentangle phenological, morphological and physiological factors that
could all help explain differential susceptibility to FHB between annual wheat, perennial wheat
and intermediate wheatgrass.
Selection for high yield and PSCR in Thinopyrum / wheat amphiploids over several
generations does not appear, up to 2005, to have selected against FHB resistance. However, the
set of lines from WSU in 2009 were substantially less resistant, suggesting that resistance may

243

have been lost over the four cycles of breeding after 2005. It is unlikely that the perennial parents
had low resistance in the first place, as Thinopyrum spp. have been reliably used as a source of
resistance traits in the past, and observational evidence from wheatgrass breeders in Kansas
(DeHaan LR, personal communication) suggest that FHB disease pressure is low. The decline in
resistance hints that further selection for yield in FHB-free environments could result in the loss
of FHB resistance, and that further breeding efforts should screen for FHB resistance through
inoculation assays. The positive correlation between FHB resistance and PSCR suggests that
selection for increased perenniality in future could favour increased FHB resistance. The very
high levels of resistance found in some perennial cereal lines offer hopeful indications that
perennial wheat and wheatgrass could be grown successfully as perennial food grains even in
environments prone to infection by FHB.

Acknowledgements

We express our appreciation to Stephen Jones and Kevin Murphy at Washington State University.
Excellent technical and field assistance is also acknowledged from Steve Culman, Swasti
Mishra, Sue Hammar, and Randy Laurenz (Michigan State University).

244

APPENDIX

245

Tables and Figures for Chapter 5

246

Table 22. Genotypes involved in a 2010-2011 greenhouse study of Type II resistance to
Fusarium Head Blight (East Lansing, MI). Annual wheat checks and perennial cereal test lines
are included. ‘Amphiploid test lines’ include all lines with perennial grass ancestry.

Annual
Wheat
Checks
Line Name
Sumai
Wheaton
Truman
Ambassador

Life History
Annual spring
Annual spring
Annual winter
Annual winter

Resistance
Resistant
Susceptible
Resistant
Susceptible

Amphiploid
Test Lines
Line Name

Life Form

Perennial
Ancestry
100%

PSCR

Reference

1.00

DeHaan et al., 2010

50%
50%
50%
50%
37.5%

0.86
0.2*
0.1*
0.4*
0.00

TLI-C1
Rival Rye
LI-48
LI-47
LI-19
P3-2009

Perennial
wheatgrassa
Rye amphiploid
Wheat amphiploid
Wheat amphiploid
Wheat amphiploid
Wheat amphiploid

P5-2009

Wheat amphiploid

37.5%

P10-2009

Wheat amphiploid

37.5%

P15-2009

Wheat amphiploid

37.5%

P3-2005
P11-2005
P15-2005
P19-2005

Wheat amphiploid
Wheat amphiploid
Wheat amphiploid
Wheat amphiploid

25%
25%
25%
25%

*Unreplicated observation from 1 m row.
**Data obtained from separate pilot study in 2007-2008.
*** Obtained from Peters Seed Company, Oregon, USA.
.

247

Reference
Bai and Suenaga, 2000
Busch et al., 1984
McKendry et al., 2005
Lewis et al., 2010

***
DeHaan et al., 2010
DeHaan et al., 2010
DeHaan et al., 2010
Murphy, K.M.
(unpublished)
0.20
Murphy, K. M.
(unpublished)
0.00
Murphy, K. M.
(unpublished)
0.76
Murphy K. M.
(unpublished)
0.68** Murphy et al., 2010
0.64** Murphy et al., 2010
0.90
Murphy et al., 2010
0.80
Murphy et al., 2010

Table 23. Results from comparing percent of Fusarium-infected spikelets in various
perennial grass and grass-cereal amphiploid lines to annual susceptible and resistant
controls. Data were taken from a 2010-2011 greenhouse study of susceptibility to Fusarium
head blight (FHB) in perennial cereals, conducted in East Lansing, MI.
* (H) indicates significantly higher than the indicated control
* (L) indicates significantly lower than the indicated control
n.s. indicates no significant difference from the indicated control (at α = 0.05).
“Resistant”, “Susceptible” and “Intermediate” classification are as defined in “Methods” section.

Genotype
TLI-1
P3-2005
P11-2005
P15-2005
P19-2005
P3-2009
P5-2009
P10-2009
P15-2009
LI-19
LI-47
LI-48
Rival Rye

vs.
Ambassador
* (L)
* (L)
* (L)
* (L)
* (L)
n.s.
* (L)
n.s.
* (L)
* (L)
* (L)
* (L)
* (L)

vs.
Wheaton
* (L)
* (L)
* (L)
* (L)
* (L)
n.s.
n.s.
n.s.
* (L)
* (L)
* (L)
* (L)
n.s

vs.
Truman
n.s.
n.s.
n.s.
n.s.
n.s.
* (H)
* (H)
* (H)
n.s.
n.s.
* (H)
* (H)
* (H)

248

vs.
Sumai 3
n.s.
n.s.
n.s.
n.s.
n.s.
* (H)
* (H)
* (H)
n.s.
n.s.
* (H)
* (H)
* (H)

Overall Rating
Resistant
Resistant
Resistant
Resistant
Resistant
Susceptible
Intermediate
Susceptible
Resistant
Resistant
Intermediate
Intermediate
Intermediate

Table 24. Results from comparing percent of Fusarium-infected rachis segments in various
perennial grass and grass-cereal amphiploid lines to annual susceptible and resistant
controls. Data were taken from a 2010-2011 greenhouse study of susceptibility to Fusarium
head blight (FHB) in perennial cereals, conducted in East Lansing, MI.
* (H) indicates significantly higher than the control
* (L) indicates significantly lower than the control
n.s. indicates no significant difference from the control (at α = 0.05).
“Resistant”, “Susceptible” and “Intermediate” classification are as defined in “Methods” section.

Genotype
TLI-1
P3-2005
P11-2005
P15-2005
P19-2005
P3-2009
P5-2009
P10-2009
P15-2009
LI-19
LI-47
LI-48
Rival Rye

vs. Ambassador
* (L)
* (L)
* (L)
* (L)
* (L)
n.s.
* (L)
n.s.
* (L)
* (L)
* (L)
* (L)
* (L)

vs. Wheaton
* (L)
* (L)
* (L)
* (L)
* (L)
n.s.
n.s.
n.s.
n.s.
* (L)
n.s.
n.s.
n.s

vs. Truman
n.s.
n.s.
n.s.
n.s.
n.s.
* (H)
* (H)
* (H)
n.s.
n.s.
* (H)
* (H)
n.s.

249

vs. Sumai 3
n.s.
n.s.
n.s.
n.s.
n.s.
* (H)
* (H)
* (H)
* (H)
n.s.
* (H)
* (H)
n.s.

Overall Rating
Resistant
Resistant
Resistant
Resistant
Resistant
Susceptible
Intermediate
Susceptible
Intermediate
Resistant
Intermediate
Intermediate
Intermediate

Table 25. Means (+ SE) and F-values (p-values in parentheses) from comparing incidence,
severity and total disease index of Fusarium graminearum infection on seed heads of
perennial rye (Rival Rye), intermediate wheatgrass (TLI-C1) and five wheat / Thinopyrum
amphiploids, growing in the field under natural and non-inoculated conditions. Data were
taken from a 2010-2011 greenhouse study of susceptibility to Fusarium head blight (FHB) of
perennial cereals in the field, conducted in Hickory Corners, MI. All lines were compared against
the annual wheat check, Triticum aestivum cv. Frankenmuth’.

Accession
Frankenmuth
TLI-C1
P19-2005
P3-2009
P5-2009
P10-2009
P15-2009
Rival Rye
Sources of Variation
Species
Block

Incidence
5.8 + 0.8
0.8 + 0.8
0.8 + 0.8
30.0 + 11.7
21.2 + 2.4
18.8 + 4.3
7.5 + 7.5
3.0 + 2.0

Severity
53.1 + 6.4
14.0 + 14.0
16.7 + 16.7
48.2 + 8.0
61.1 + 14.2
33.9 + 6.7
10.0 + 10.0
46.5 + 20.9

Disease Index
3.4 + 0.4
0.7 + 0.7
0.8 + 0.8
16.0 + 7.7
13.1 + 3.5
6.2 + 1.9
3.0 + 3.0
7.0 + 7.0

9.76 (0.0001)
1.04 (0.15)

2.38 (0.049)
0.79 (0.22)

5.28 (0.0007)
0.74 (0.23)

250

Figure 26. Percent of Fusarium-infected spikelets in 13 perennial grass and grass-cereal
amphiploid lines, compared against susceptible and resistant annual wheat controls. Data
were taken from a 2010-2011 greenhouse study of susceptibility to Fusarium head blight (FHB)
in perennial cereals, conducted in East Lansing, MI. Pink = susceptible annual wheat lines, red =
resistant annual wheat, blue = Thinopyrum intermedium, dark green = wheat x Th. junceiforme,
green = wheat x Th. intermedium, light green = wheat x Th. elongatum, brown = perennial rye.

251

Figure 27. Percent of Fusarium-infected rachis segments in 13 perennial grass and grasscereal amphiploid lines, compared against susceptible and resistant annual wheat controls.
Data were taken from a 2010-2011 greenhouse study of susceptibility to Fusarium head blight
(FHB) in perennial cereals, conducted in East Lansing, MI. Pink = susceptible annual wheat
lines, red = resistant annual wheat, blue = Thinopyrum intermedium, dark green = wheat x Th.
junceiforme, green = wheat x Th. intermedium, light green = wheat x Th. elongatum, brown =
perennial rye.

252

REFERENCES

253

REFERENCES

Acharya SN, Mir Z, Moyer JR, Orshinsky BR, Thomas JE. 2003. Effect of row spacing and
seeding date on forage yield and quality of perennial cereal rye (Secale cereale L.). Canadian
Journal of Plant Science 83: 363-369.
Acharya SN, Mir Z, Moyer JR. 2004. ACE-1 perennial rye. Canadian Journal of Plant Science
84: 819-821.
Bai G, Shaner G. 2004. Management and resistance in wheat and barley to Fusarium head blight.
Annual Review of Phytopathology 42: 135-161.
Ban T, Suenaga K. 2000. Genetic analysis of resistance to Fusarium head blight caused by
Fusarium graminearum in Chinese wheat cultivar ‘Sumai 3’ and the Japanese cultivar ‘Saikai
16S’. Euphytica 113: 87-99.
Busch R, McVey D, Rauch T, Baumer J, Elsayed F. 1984. Registration of ‘Wheaton’ wheat. Crop
Science 24: 622.
Buerstmayr H, Ban T, Anderson, JA. 2009. QTL mapping and marker-assisted selection for
Fusarium head blight resistance in wheat: a review. Plant Breeding 128: 1-26.
Cai X, Chen PD, Xu SS, Oliver RE, Chen X. 2005. Utilization of alien genes to enhance
Fusarium head blight resistance in wheat- A review. Euphytica 142: 309-318.
Capelli RA, Peterson JL. 1965. Macroconidium formation in submerged cultures by a nonsporulating strain of Gibberella zeae. Mycologia 57: 962-966.
Cox TS, Bender M, Picone C, Van Tassel DL, Holland JB, Brummer EC, Zoeller BE, Paterson,
AH, Jackson W. 2002. Breeding perennial grain crops. Critical Reviews in Plant Science 21: 5991.
Cox CM, Murray TD, Jones SS. 2002. Perennial wheat germ plasm lines resistant to eyespot,
Cephalosporium stripe, and wheat streak mosaic. Plant Disease 86: 1043-1048.
Cox CM, Garrett KA, Cox TS, Bockus WW, Peters T. 2005. Reactions of perennial grain
accessions to four major cereal pathogens of the Great Plains. Plant Disease 89: 1235-1240.
Cox TS., Glover JD., Van Tassel DL., Cox CM., DeHaan, LR. 2006. Prospects for developing
perennial grain crops. BioScience 56: 649-659.
DeHaan LR, Van Tassel DL, Cox TS. 2005. Perennial grain crops: a synthesis of ecology and
plant breeding. Renewable Agriculture and Food Systems 20: 5-14.

254

Fedak G, Han F. 2005. Characterization of derivatives from wheat-Thinopyrum wide crosses.
Cytogenetic and Genome Research 109: 360-367.
Fu S, Lu Z, Qi B, Guo X, Li J, Liu B, Han F. 2011. Molecular cytogenetic characterization of
wheat-Thinopyrum elongatum addition, substitution and translocation lines with a novel source
of resistance to wheat Fusarium Head Blight (FHB). Journal of Genetics and Genomics 39: 103110.
Gilbert J, Fedak G, Procunier JD, Aung T, Tekauz A. 1996. “Strategies for Breeding for
Resistanceto FHB in Canadian Spring Wheat.” In “Fusarium Head Scab: Global Status and
Future Prospects.” Proceedings of a Workshop Held at CIMMYT, El Batan, Mexico, 13-17 Oct
1996.
Han FP, Fedak G, Benabdelmouna A, Armstrong K, Oullet T. 2003. Characterization of six wheat
x Thinopyrum intermedium derivatives by GISH, RFLP, and multicolor GISH. Genome 46: 490495.
Khavkin EE, Zabrodina MV, and Sillis DY. 1996. Isozymes of aspartate aminotransferase in
perennial and annual rye and their hybrids. Genome 39: 513-519.
Lammer D, Cai X, Arterburn M, Chatelain J, Murray T and Jones S. 2004. A single chromosome
addition from Thinopyrum elongatum confers a polycarpic, perennial habit to annual wheat.
Journal of Experimental Botany 55: 1715-1720.
Lewis JM, Siler L, Souza E, Ng, PKW, Dong Y, Brown-Guedira G, Jiang G-L, Ward RW. 2010.
Registration of ‘Ambassador’ wheat. Journal of Plant Registrations 4: 195-204.
Li H, Wang X. 2009. Thinopyrum ponticum and Th. intermedium: promising sources of
resistance to fungal and viral diseases of wheat. Journal of Genetics and Genomics 36: 557-565.
Liu S, Hall M.D., Griffey, C. A., and McKendry, A. L. 2009. Meta-analysis of QTL associated
with Fusarium head blight resistance in wheat. Crop Science 49: 1955-1968.
Mackintosh CA, Lewis J, Radmer LE, Shin S, Heinen SJ, Smith LA, Wycoff MN, Dill-Macky R,
Evans CK, Kravchenko S, Baldridge GD, Zeyen RJ, Muehlbauer GJ. 2007. Overexpression of
defense response genes in transgenic wheat enhances resistance to Fusarium head blight. Plant
Cell Reports 26: 479-488.
McKendry AL, Tague DN, Wright RL, Tremain JA, Conley SP. 2005. Registration of ‘Truman’
wheat. Crop Science 45: 421-423.
Mesterházy A. 1995. Types and components of resistance to Fusarium head blight of wheat.
Plant Breeding 114: 377-386.
Mesterházy A, Bartók T, Lamper, C. 2003. Influence of wheat cultivar, species of Fusarium, and
isolate aggressiveness on the efficacy of fungicides for control of Fusarium head blight. Plant
255

Disease 87: 1107-1115.
Okubara PA, Jones SS. 2011. Seedling resistance to Rhizoctonia and Pythium spp. in wheat
chromosome group 4 addition lines from Thinopyrum spp. Canadian Journal of Plant Pathology
33: 416-423.
Oliver RE, Cai X, Xu SS, Chen X, Stack R. 2005. Wheat-alien species derivatives: a novel
source of resistance to Fusarium head blight in wheat. Crop Science 45: 1353-1360.
Oliver RE, Xu SS, Stack RW, Friesen TL, Jin Y and Cai X. 2006. Molecular cytogenetic
characterization of four partial wheat- Thinopyrum ponticum amphiploids and their reactions to
Fusarium head blight, tan spot, and Stagonospora nodorum blotch. Theoretical and Applied
Genetics 112: 1473-1479.
Oliver RE, Cai X, Wang R-C, Xu SS, Friesen TL. 2008. Resistance to tan spot and
Stagonospora nodorum blotch in wheat-alien species derivatives. Plant Disease 92:150-157.
Pirgozliev SR, Ray RV, Edwards SG, Hare MC, and Jenkinson P. (2010). Effect of timing of
fungicide application on the development of Fusarium head blight and the accumulation of
deoxynivalenol (DON) in winter wheat grain. Cereal Research Communications 36: 289-299
Pritsch C, Muehlbauer GJ, Bushnell WR, Somers DA, Vance CP. 2000. Fungal development and
induction of defense response genes during early infection of wheat spikes by Fusarium
graminearum. Molecular Plant-Microbe Interactions 13: 159-169.
Reimann-Philipp, R. 1986. Perennial spring rye as a crop alternative. Journal of Agronomy and
Crop Science 157: 281-285.
Scheinost PL, Lammer DL, Cai X, Murray TD, Jones, SS. 2001. Perennial wheat: the
development of a sustainable cropping system for the U.S., Pacific Northwest. American Journal
of Alternative Agriculture 16: 147-151.
Schroeder HW, Christensen JJ. 1963. Factors affecting resistance of wheat to scab caused by
Gibberella zeae. Phytopathology 53: 831-838.
Shen X, Kong L, Ohm H. 2004. Fusarium head blight resistance in hexaploid wheat (Triticum
aestivum)-Lophopyrum genetic lines and tagging of the alien chromatin by PCR markers.
Theoretical and Applied Genetics 108: 808-813.
Shen X, Ohm H. 2007. Molecular mapping of Thinopyrum-derived Fusarium head blight
resistance in common wheat. Molecular Breeding 20: 131-140.
Van Tassel DL, DeHaan LR, Cox TS. 2010. Missing domesticated plant forms: can artificial
selection fill the gap? Evolutionary Applications 3: 434-452.
Wan Y-F, Yen C, Yang J-L. 1997. The diversity of head scab resistance in Triticeae and their
256

relation to ecological conditions. Euphytica 97: 277-281
Wang JI-R, Wang L, Gulden S, Rocheleau H, Balcerzak M, Hattori J, Cao W, Han F, Zheng Y-L,
Fedak G, Ouellet T. 2010. RNA profiling of fusarium head blight resistant wheat addition lines
containing the Thinopyrum elongatum chromosome 7E. Canadian Journal of Plant Pathology
32: 188-214.
Weik L, Kaul H-P, Kübler E, and Aufhammer W. 2002. Grain yields of perennial grain crops in
pure and mixed stands. Journal of Agronomy and Crop Science 188: 342-349.
Xue AG, Voldeng HD, Savard ME, Fedak G. 2009. Biological management of Fusarium head
blight and mycotoxin contamination in wheat. World Mycotoxin Journal 2: 193-201.
Zhang L, Luo P, Ren Z, Zhang H. 2011. Controlling Fusarium head blight of wheat (Triticum
aestivum L.) with genetics. Advances in Bioscience and Biotechnology 2: 263-270.
Zhang X, Shen X, Hao Y, Cai J, Ohm H, Kong L. 2011. A genetic map of Lophopyrum ponticum
chromosome 7E, harboring resistance genes to Fusarium head blight and leaf rust. Theoretical
and Applied Genetics 122: 263-270.
Zhou MP, Hayden MJ, Zhang ZY, Lu WZ and Ma, X-H. 2010. Saturation and mapping of a
major Fusarium head blight resistance QTL on chromosome 3BS of Sumai 3 wheat. Journal of
Applied Genetics 51: 19-25.

257

CONCLUSION. BROADER IMPACTS OF PERENNIAL CEREALS

258

In Chapter 1 of this dissertation, we considered environmental implications of perennial cereals
within agricultural landscapes. Some of the specific ecosystem services which perennial cereals
could be expected to contribute, include the following:
Edible grain yield (as this comprises an ecosystem service if humans and domesticated
animals are considered part of the ecosystem)
Reduced nutrient leaching
Increased carbon sequestration
Reduced soil erosion
Increased tolerance/resistance to insect herbivores and pathogens, due to structural,
physiological and biochemical traits inherited from perennial ancestors
Increased populations of pollinator and predatory insects.

In contrast, perennial cereals might also contribute certain ecosystem disservices. For example,
perennial plants which persisted year round in the field could serve as hosts for pathogens and
insect herbivores during periods of the year when annual stands were not present or were very
sparse (e.g. winter for the fungal pathogen Fusarium graminearum, late summer for the Hessian
fly Mayetiola destructor). These herbivores or pathogens might be able to maintain or increase
their populations during the off-season, due to the presence of perennial cereal hosts, and could
thus spread to adjacent agricultural fields during the late fall or spring. By serving as an alternate
host and reservoir for pathogens and insects, perennial cereals could contribute ecosystem
disservices in addition to the hypothesized ecosystem services.
Our research allows several of these questions to be addressed. In Chapter 1 of this

259

dissertation, we found that perennial wheat and perennial rye achieved approximately 50% and
70% of annual wheat and annual rye grain yields, respectively, and that these yields were
maintained over at least two years. No additional benefits in terms of extra forage production
were achieved. Thus perennial wheat and perennial rye are at this point still greatly inferior to
their annual counterparts in terms of one important ecosystem service (grain yield) and further
breeding efforts must strive to increase grain and forage yields.
How could breeding efforts most effectively increase seed production? One approach
might be to increase capacity to acquire energy and nutrient resources, while another approach
would focus on breeding directly for increased reproductive sinks., Currently, as demonstrated in
Chapter 4, perennial cereal species do not appear to be source limited, and have excess capacity
to photosynthesize and mobilize stored carbohydrates. Changes in sink / source ratio, especially
in perennial wheat, do not appear to strongly affect seed size. Thus breeding efforts should most
likely focus on breeding for greater spikelet number, lower sterility, and increased germination
percentage. Preliminary observations (data not published) suggest that perennial wheat currently
has longer spikes and more spikelets than annual wheat, but is markedly inferior in terms of
germination ability and sterility, with many spikelets never being fertilized or filled. Thus
decreased sterility and increased ability to germinate seem to be important pathways to achieving
higher grain yield in perennial cereals. Preliminary research (Culman S. et al., 2012) indicates
that intermediate wheatgrass yields far below annual wheat, although yields improve in the
second year. Thus further wheatgrass breeding efforts must also focus intensively on improving
sink strength, reproductive allocation and seed size.
Our research did not directly address the question of whether perennial cereals do in fact
show less nutrient leaching than annual relatives. However, ongoing research at Kellogg

260

Biological Station (Sprunger et al., unpublished data) does indicate that intermediate wheatgrass
shows much lower nitrogen leaching, and much lower concentrations of nitrogen in groundwater,
than annual wheat. This could be related to higher rates of nutrient uptake by larger and more
persistent root networks. Further research should investigate whether other perennial cereal
species show lower nutrient leaching rates than annual relatives as well.
Our study did not directly investigate insect pressure or composition of insect
communities in perennial cereal plots. However, with respect to pathogen pressure, we did find
that perennial rye was generally fairly susceptible to Fusarium graminearum, while intermediate
wheatgrass was extremely resistant, and perennial wheat lines varied in resistance.
With respect to this particular pathogen, therefore, the evidence suggests that perennial
rye does present risks of contaminating adjacent annual rye fields with a ready source of
inoculum in the spring. This may be less of a risk during the first year of a perennial rye crop, as
this species matures later than annual rye, so that by the time F. graminearum has infected
perennial rye spikes and been able to multiply, annual rye would already be past the point of
maximal susceptibility. In later years, however, as perennial rye and annual rye appear to mature
at the same time, the danger would be greater. Another pathogen to which perennial rye appears
to be particularly vulnerable is the fungal pathogen ergot (Claviceps purpurea): this pathogen
may be particularly prevalent in small plots, as it preferentially attacks unfertilized heads, and
perennial rye is a cross pollinated species which depends on nearby conspecific plants to be
pollinated. Observations in Part 2 of this dissertation suggest that perennial rye and annual rye
were both highly susceptible to C. purpurea, and thus perennial rye could serve as a year round
host and source of inoculums for ergot infection of annual rye fields. Future selection and
cultivation practices, therefore, should focus on breeding more disease resistant perennial rye

261

lines and on optimizing cultivation practices to minimize ergot infection (e.g. larger fields,
deeper planting).
Intermediate wheatgrass, by contrast, was highly resistant to F. graminearum, as were
some of our perennial wheat lines. Our study suggests that the likelihood of this particular
ecosystem disservice (serving as a reservoir for Fusarium infection) is low in intermediate
wheatgrass. Casual observation of the intermediate wheatgrass lines growing in the field in Part
3 of this dissertation suggest that they had very low susceptibility to other common diseases in
the region as well, as very few disease symptoms were seen. This corroborates reports of very
low disease pressure on intermediate wheatgrass lines in Kansas (DeHaan LR, personal
communication). Intermediate wheatgrass appears, in general, to be a species that would be
highly disease resistant on its own, and would probably not serve as a source of inoculum for
nearby annual crops. It might thus be a net contributor of ecosystem services with respect to
disease prevalence, inasmuch as intermediate wheatgrass could potentially be grown with less
reliance on fungicides. Perennial wheat, by contrast, showed more Fusarium head blight
susceptibility, and field observations of other disease symptoms indicated that some lines showed
generally less foliar and spike disease symptoms than annual wheat, but more than intermediate
wheatgrass. Other lines, by contrast, showed as much or more disease pressure (from pathogens
like barley yellow dwarf virus, leaf rust, and Septoria leaf blotch) as annual wheat. Further
breeding efforts in perennial wheat should be based on the lines identified as showing greatest
disease resistance, and should focus on selection for adequate levels of disease resistance
simultaneous with strong regrowth.
The one ecosystem service which our research addresses most directly, is carbon
sequestration. While we did not directly measure carbon sequestration in terms of inputs of fixed

262

carbon into the soil, we do have indirect evidence from Chapter 3 of this dissertation that
suggests perennial cereals may fix more carbon from the atmosphere through photosynthesis.
Photosynthetic rates during mid-season were found to be approximately 29% higher in perennial
wheat and 45% higher in intermediate wheatgrass compared to annual wheat, and 29% higher in
perennial rye compared to annual rye (averaging across measurement dates). These
measurements are based on snapshots at only two or three time points during the season, and do
not represent an integrated measure of photosynthesis over the season: nevertheless they are at
least indicative that at a key time point during the season, photosynthetic carbon fixation in the
perennial cereals studied is greater than in their annual relatives. As carbon export belowground,
and subsequent carbon sequestration is dependent on the amount of carbon fixed from the
atmosphere, higher photosynthetic rates could in the long term contribute to greater carbon
sequestration within perennial cereal agroecosystems.
As yet, our measures of photosynthesis are taken on a per leaf area basis, and do not
represent measurements of carbon fixation for a whole stand. Nevertheless, by making some
broad assumptions we can at least make a rough estimate of carbon fixation on a per-plot level.
To do this, we assume that photosynthetic rates are constant over the growing season. While this
is certainly not true, we can use this broad assumption to help formulate some initial estimates,
which can later be adjusted on the basis of more accurate models of how photosynthetic rate
varies seasonally. We further assume, on the basis of observations over four summers between
2009 and 2012:
that annual wheat, annual rye and first year perennial cereals are planted around
Oct 15,
that annual rye senesces around June 15 and annual wheat around June 25,

263

that perennial rye senesces 14 days after annual rye (based on observations of
flowering date in Chapter 2, plus the observation that perennial cereals senesce
somewhat later after flowering than annuals)
that perennial wheat senesces 14 days after annual wheat (based, again, on data
from Chapter 2 plus the same assumption of later senescence after flowering)
that intermediate wheatgrass flowers one month after annual wheat and takes one
week longer for leaves to senesce after flowering (based on unreported
observational data in 2011 and 2012),
that all plants are inactive between November 15 and March 15, due to cold and
snow cover
that perennial cereals begin regrowing again on August 15th.
We then use the data collected in Chapter 3 of this indication, that find that leaf mass ratio does
not vary between annual and perennial cereal species, and that leaf mass per area (LMA) is 28%
and 26% higher in perennial wheat and intermediate wheatgrass, respectively, compared to
annual wheat. Finally, we use data on plant biomass per plot from first-year perennial and annual
cereal species in 2011 (data not reported here) that indicate perennial wheat and intermediate
wheatgrass do not produce significantly different quantities of biomass on a per-plot basis than
annual wheat, while perennial rye produces 49% less than annual rye (t = 3.55, p = 0.016, n = 5).
Our estimate of total net primary production per plot is thus:

NPP = Plant Biomass / Area x (LMR) / (LMA) x (Amax) x (Length of Growing Season),

264

with all parameters standardized such that the annual species (annual rye or annual wheat,
respectively) has a value of 1 for each.

This gives us as an initial estimate that intermediate wheatgrass, in its first year, has a net
fixation of 102% more CO2 than annual wheat, while perennial wheat fixes 58% more CO2
than annual wheat. Perennial rye fixes approximately 101% as much CO2 as annual rye. These
estimates are due to the existence of a substantial time period in the fall for additional carbon
accumulation, later flowering, and higher photosynthetic rates per unit leaf area, and are subject
to the limitations mentioned above. Nevertheless, if these higher rates of carbon accumulation
are reflected in higher rates of carbon export and sequestration belowground, perennial cereals
could represent a substantially superior sink for atmospheric CO2, compared with annual wheat.
Preliminary work at Kellogg Biological Station (Sprunger C., unpublished data) does indicate
that intermediate wheatgrass does, even over a short time period of three years, show greater
carbon sequestration than annual wheat. As dealing with elevated atmospheric carbon dioxide is
one of the great environmental challenges of the near future, the suggestive indications that
perennial cereals could be a a carbon sink, point to what might be their most important and
reliable ecosystem service.
In conclusion, our research has found that these three perennial cereals, in spite of their as
yet low grain yields compared to annual equivalents, do have some interesting physiological
traits (disease resistance and high photosynthetic rates) that could allow them, in future, to form
part of more environmentally sustainable agroecosystems. Two species in particular, perennial
wheat and perennial rye, are already approaching the yield thresholds at which they may be

265

economically feasible to introduce as crops to marginal areas, and their greater degree of sink
limitation compared to annual equivalents suggests that they could be successfully bred for
higher grain yields in future without compromising perennial growth habit. We thus conclude our
study of these novel, high yielding herbaceous perennials with a suggestion that future breeding
efforts could develop these species into agronomically important and economically productive
crops, as well as physiologically interesting representatives of an unusual growth form.

266