.3 1:3,»: .- Lhasa rt:f::ul 4. .1 . E}! LIBRARIES ms it'llll‘llllllllllll\\\\\\\\\\\\\\l\\\\\ This is to certify that the dissertation entitled Genotypic Variation fin Response to Elevated Atmospheric Carbon Dioxide in Two Populations of Plantago lanceolata L. presented by Dawn Jenkins Klus has been accepted towards fulfillment of the requirements for Ph. D. degree in Botany flaw/4% Major professorfl DMCN0vember 17, 1995 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY E Michigan State} University i PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE "madam ” 3. MSU Is An Affirmative Action/Equal Opportunity Institution Warns-9.1 GENOTYPIC VARIATION IN RESPONSE TO ELEVATED ATMOSPHERIC CARBON DIOXIDE IN TWO POPULATIONS OF PLAN TAGO LANCEOLATA L. By Dawn Jenkins Klus A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1995 ABSTRACT GENOTYPIC VARIATION IN RESPONSE TO ELEVATED ATMOSPHERIC CARBON DIOXIDE IN TWO POPULATIONS OF PLAN TAGO LAN CEOLATA L. By Dawn Jenkins Klus The nature and the magnitude of possible adaptive evolutionary changes in response to elevated atmospheric CO2 will be determined by genetic variation in that response within populations and species. This study documented genotypic variation in the phenotypic response to elevated CO2 for maternal families of two populations of Plantago lanceolata grown in open topped chambers. Three groups of traits were measured over the course of one growing season: physiology, growth and biomass allocation. The overall effect of elevated CO2 did not appear to affect plant size traits as much as the allocation of photosynthate due to increased assimilation rates in the elevated CO2 environment. Root: shoot ratios and assimilation rates increased for the elevated CO2 grown plants; specific leaf area and stomatal conductance declined; nitrogen allocation to aboveground tissues increased. It appeared that Plantago lanceolata was flexible in terms of integrating its physiological processes with patterns of nutrient and biomass allocation within the plant. In ambient and elevated CO2 environments, there were highly significant population level differences for most variables. Significant CO2 x population interactions were detected for physiological and grth traits. Significant CO2 x family interactions were detected for nutrient allocation, physiological and growth traits. Individual families varied greatly in the magnitude and direction of response to elevated C02. Some families responded positively, some negatively, and some not at all to the elevated CO2 environment. Genotype x environment interactions resulted in families changing rankings with respect to one another across ambient and elevated CO2 environments, thus rendering it difficult to predict the performance of a family in elevated 002 from its performance in ambient C02. Because the response of the populations and the families of P. lanceolata to elevated CO2 was not uniform, some or all of the changes in physiology and allocation patterns may ultimately help certain families of Plantago lanceolata to survive, compete, and reproduce with greater success in an elevated CO2 world. Such variety in response to elevated CO2 at the family level has implications for both intra- and interspecific competitive ability in elevated CO2 and suggests that elevated atmospheric CO2 may have the potential to act as an agent of natural selection. Dedicated to my parents, Ruth and Louis Jenkins my husband, John Klus and to my son, Nicholas Klus ACKNOWLEDGEMENTS I would like to acknowledge Susan Kalisz, Steve Tonsor, Peter Curtis and James Teeri for their support: moral, material, and financial Without their collaboration, this project would not have been a success. Everyone at the W. K Kellogg Biological Station, past and present, has inspired me, encouraged me, and given me valuable assistance whenever I needed it. My thanks to all of you, but especially to: Denise Thiede Paco Moore Kim Thompson Jon Ervin Pam Woodrufl‘ Nina Consolatti Sandy Halstead Deb Walker Helen Garchow Carolyn Hammarsjkold Ann O'Neill Bryan Foster Andy Turner Pete Smith Char Adams Kay Gross Phil Robertson Hal Collins Carolyn Miller Alice Gillespie "On location" thanks go to Brenda Casper, Kel Wiederman, and Scott Starr for making a place for me in Philadelphia. My new fiiends among the faculty and staff of Sayre School in Lexington, Kentucky, have provided me with constant encouragement. Thank you. TABLE OF CONTENTS Page LIST OF TABLES .............................................................................................. vii LIST OF FIGURES ............................................................................................. xi LIST OF ABBREVIATIONS ............................................................................. xiv INTRODUCTION .................. ‘ .............................................................................. 1 CHAPTER ONE: IN TRASPECIFIC VARIABLE RESPONSES T O ELEVATED ATMOSPHERIC C02: RESOURCE PARTITIONING IN ABOVE- AND BELOWGROUND TISSUES IN PLAN TAGO LAN CEOLATA L .................... 7 CHAPTER TWO: PHYSIOLOGICAL VARIATION IN RESPONSE TO ELEVATED CARBON DIOXIDE IN PLANTAGO LAN CEOLATA ................................................................ 33 CHAPTER THREE: GROWTH RESPONSES TO ELEVATED CARBON DIOXIDE IN PLANT AGO LAN CEOLATA ..................... 72 APPENDIX ....................................................................................................... 1 1 1 LITERATURE CITED ...................................................................................... 123 I fl c—r'L-l‘. '«"" ’T‘" ': .. ‘_ _ . _ LIST OF TABLES Table Page 1. Analysis of variance for biomass variables. Overall model is presented .................................................................................................... 17 2. Analysis of variance for tissue composition components for three families (EL 15, KF 32, KF 35). Overall model for each variable is presented ................................................................................................. 23 3. Means (s.e.) for CO2 treatments for ambient and elevated CO2 measurements. Values for August and September are presented. N = 89 - 99 for each variable. CO2 treatment means were compared using univariate one-way AN OVAs. Letters represent sigmificant differences for each variable measured at the same CO2 level during same month. Lowercase letters, differences between means at ambient C02. Uppercase letters, differences between means at elevated C02. Differences were considered to be significant at the p < 0.025 level (Bonferroni correction for multiple comparisons of means. Table 3A August ...................................................................................... 46 Table 3.B. September ................................................................................. 46 4. Multivariate analysis of variance for the main effects of CO2 treatment, population, level, month, and family nested within population, and their two-way interactions. Part A. Overall MANOVA for four physiological variables: assimilation, transpiration, stomatal conductance, and intercellular CO2 concentration. Part B. Summary of univan'ate mixed model analysis of variance for each of the four physiological variables used in the MANOVA, plus instantaneous water use efficiency and specific leaf area. CO2 measurement level was not a main effect in the analysis of variance for specific leaf area. Table 4A. Overall multivariate analysis ...................................................... 47 Table 4B. Summary of probabilities from univariate mixed model AN OVAs. CO2 measurement level was not included in the model for SLA .......................................................................... 48 Table Page 5. Repeated measures analysis for variables measured at the ambient and elevated CO2 measurement level (identified as Level in the model). The analysis was conducted separately for each month, first in an overall multivariate analysis, then for each physiological variable separately. Significance levels: *1 p < 0.01; * p < 0.05; ** p < 0.01; *** p < 0.001. Table 5A August ...................................................................................... 51 Table 5B. September ................................................................................. 52 6. Repeated measures analysis for variables measured in August and September. The analysis was conducted separately for each CO2 measurement level, first in an overall multivariate analysis, then for each physiological variable separately. Significance levels: *1 p < 0.1; a: p < 0.05; ** p < 0.01; *** p < 0.001. Table 6A Ambient CO2 measurement level ................................................ 57 Table 6.B. Elevated CO2 measurement level ................................................ 58 7.A. 7.B. 8A 8B. Individual tests of treatment means for main experiment. Comparison of ambient CO2 chambers, elevated CO2 chambers, and unchambered controls. Shoot number and shoot diameter for Day 10 were not compared since each plant had only one shoot which was too small to be measured on that date. Means were compared using pairwise contrasts in an analysis of variance. Different letters indicate that means differ by p < 0.017 (B onferroni adjustment for multiple contrasts) ................................................................................................ 86 Individual tests of treatment means for growth experiment. Comparison of ambient CO2 chambers, elevated CO2 chambers, and unchambered controls. Shoot number and shoot diameter for Day 13 were not compared since each plant had only one shoot which was too small to be measured on that date. Means were compared using pairwise contrasts in an analysis of variance. Different letters indicate that means differ by p < 0.017 (B onferroni adjustment for multiple contrasts) ................................................................................................ 87 Repeated measures profile analysis of main experiment variables. Comparisons include ambient CO2 chambers, elevated CO2 chambers, and unchambered control sites. Between subjects tests detect differences among main effects and interactions, summed over the repeated measure. Within subjects tests detect main effect of the repeated variable and interaction of the repeated variable with the main experimental effects ................................................................................. 90 Comparisons of census dates for growth experiment. Tests for interactions between main effects and successive census dates .................. 91 1339:." ' ' Table Page 9.A. Repeated measures profile analysis of grth experiment variables. Comparisons of ambient CO2 chambers, elevated CO2 chambers, and unchambered control sites. Between subjects tests detect difl‘erences among main effects and interactions, summed over the repeated measure. Within subjects tests detect main effect of the repeated variable and interaction of the repeated variable with the main experimental efl‘ects ...................................................................................................... 93 9B. Comparisons of census dates for growth experiment. Tests for interactions between main effects and successive census dates ............ 94 10. Individual tests of treatment means for physiological variables in growth experiment. Comparison of ambient CO2 chambers, elevated CO2 chambers, and unchambered controls. Means were compared using pairwise contrasts in an analysis of variance. Letters indicate significant differences, p < 0.017 (Bonferroni adjustment for multiple contrasts) ............................................................... 99 l 1. Repeated measures profile analysis of physiological variables in growth experiment. Comparisons of ambient CO2 chambers, elevated CO2 chambers, and unchambered control sites. Between subjects tests detect differences among main effects and interactions, summed over the repeated measure. Within subjects tests detect main effect of the repeated variable and interaction of the repeated variable with the main experimental efl‘ects ............................ 102 12. Overall and population means (s.e.) for CO2 treatments for ambient and elevated CO2 measurements. CO2 treatment means were compared using univariate one-way AN OVAs. Letters represent significant differences for each variable between the two CO2 treatments. Differences were considered to be significant at the p < 0.025 level (Bonferroni adjustment for multiple comparisons of means) .................................................................................................. 1 l6 l3. Multivariate analysis of variance for the main effects of CO2 treatment, population, and family nested within population, and their one-way interactions. Part A. Overall MANOVA for three physiological variables: assimilation, transpiration, and stomatal conductance. Part B. Overall MANOVA for two biomass variables: aboveground biomass, belowground biomass. Table 13.A. Overall multivariate analysis for physiological variables ......... 117 Table 13.B. Overall multivariate analysis for biomass variables ................. 117 Table Page 14. Summary of probabilities from univariate mixed model AN OVAs. Part A Physiological variables: A=assimilation rate. E=transpiration rate. Gs=stomata1 conductance. Ca'ntercellular CO2 concentration. WUE=water use efficiency. Part B. Biomass variables: aboveground biomass, belowground biomass, total biomass, root: shoot ratio. Table 14.A. Physiological variables ............................................................. 1 18 Table 14.B. Biomass variables .................................................................... 118 LIST OF FIGURES Figure Page 1. Comparison of aboveground biomass, belowground biomass, and root: shoot ratio by CO2 treatment. For each category (overall, population, family) means for ambient CO2 appear to the left of means for elevated C02. Vertical bars indicate one standard error. Aboveground biomass, belowground biomass, and root: shoot ratios were considered separately for each category. Scale for root: shoot ratio appears to the right. Significance levels appear above bars for aboveground biomass, below bars for belowground biomass, and immediately adjacent symbol for root: shoot ratio. Whole plant biomass differed significantly across CO2 treatment for overall data (*), population KF (*), family EL 7 (*1), family EL 9 (*), and family KF 31 (*). Significance levels: *1 p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001 .............................. l9 2. Comparison of soluble sugar, % carbon content, % nitrogen content and carbon: nitrogen ratio by C02 treatment. For each category (overall and family) means for ambient CO2 appear to the left of means for elevated C02. Vertical bars indicate one s.e. Aboveground and belowground means were considered separately for each variable. Significance levels appear above bars for aboveground means and below bars for belowground means. Significance levels: *1p<0.1;*p<0.05;**p<0.01;***p<0.001 ......................................... 25 3. Comparison of overall means (:1: s.e.) for assimilation, transpiration, stomatal conductance, and intercellular CO2 concentration for August and September. In each month, plants grown in each CO2 treatment were measured at both CO2 levels, ambient CO2 and elevated C02. For intercellular CO2 concentration error bars may be obscured by symbol ............................................................................................................. 54 4. Comparison of individual family means for stomatal conductance in August and September. (top) Plants grown in ambient C02, measured in elevated C02. (bottom) Plants grown in elevated C02, measured in elevated CO2 ................................................................................................... 62 Figure Page 5. 10. 11. 12. 13. Assimilation versus intercellular leaf CO2 concentration at both CO2 measurement levels. At the ambient CO2 measurement level, inter- cellular CO2 concentration was close to 200 “bar. At the elevated CO2 measurement level, intercellular CO2 concentration was close to 500 ,ubar. Vertical bars indicate i: one s.e. for assimilation means. Horizontal bars indicate i one s.e. for intercellular CO2 means ..................................................................................................... 67 . Overall means for leaf number at each census date for main experiment and grth experiment (inset). Vertical bars indicate one s.e ......................... 83 . Overall means for basal shoot diameter at each census date for main experiment and growth experiment (inset). Vertical bars indicate one s.e. ......................................................................................................... 84 Overall means for vegetative shoot number at each census date for main experiment and growth experiment (inset). Vertical bars indicate one s.e. .......................................................................................................... 85 Comparison of overall means for aboveground biomass, belowground biomass, and root: shoot ratio for chambers and unchambered control sites. Vertical bars indicate one s.e. Scale for root: shoot ratio appears to the right. Means were tested by pairwise contrasts in an analysis of variance. Different letters indicate that means differ by p < 0.017 (Bonferroni adjustment for multiple contrasts) ............................................... 97 Comparison of overall growth experiment means (i: s.e.) for assimilation, transpiration, stomatal conductance, and water use efficiency at both ambient and elevated CO2 measurement levels ............................................. 101 Assimilation versus intercellular leaf CO2 concentration at both ambient and elevated CO2 measurement levels for growth experiment. At the ambient CO2 measurement level, intercellular CO2 concentration was close to 200 pbar for all treatments. At the elevated CO2 measurement level, intercellular CO2 concentration was close to 500 ,ubar for all treatments. Vertical bars indicate i s.e. for assimilation means. Horizontal bars indicate d: s. e. for intercellular CO2 means .......................... 104 Comparison of overall, population and family means for assimilation for ambient and elevated CO2 treatments for 1991 experiment ..................... 119 Comparison of overall, population and family means for transpiration for ambient and elevated CO2 treatments for 1991 experiment ..................... 120 Figure Page 14. Comparison of overall, population, and family means for aboveground biomass, belowground biomass, and root: shoot ratio by C02 treatment for 1991 experiment. For each category (overall, population, family) means for ambient CO2 treatment appear to the left of means for elevated CO2 treatment. Vertical bars indicate one s.e. Scale for root: shoot ratio appears to the right. Significance levels for individual tests of means appear above bars for aboveground biomass and below bars for belowground biomass. Significance levels: *lp<0.1;*p<0.05;**p<0.01;***p<0.001 ....................................... 122 LIST OF ABBREVIATIONS Assimilation (pmol CO2 m'2 s") ............................................................. A Transpiration (mmol H20 m’2 s") ........................................................... E Water Use Efliciency (mmol CO2 mol'1 H20) ...................................... WUE Stomatal Conductance (mol H20 m'2 s") ............................................... Gs Intercellular C02(,u1iter liter") .............................................................. Ci Specific Leaf Area (cm2 g‘1 dry weight) ............................................... SLA Relative Grth Rate (g biomass g'l day") ......................................... RGR xiv INTRODUCTION Evidence is now unequivocal that atmospheric CO2 levels are increasing worldwide (Keeling, 1986). Predictions indicate that the increase will continue well into the next century, with a doubling of current CO2 levels by 2050 being likely (Strain, 1987). Numerous studies have established what is now considered a picture of the "typical" whole plant response to elevated C02, at least in the short term. Exposure to elevated CO2 results in increased photosynthetic rates, decreased transpiration and stomatal conductance, and increased water use efficiency (see Bazzaz, 1990, for a review). Accompanying these physiological changes, alterations in tissue characteristics and biomass allocation often result in decreased specific leaf area (SLA), increased relative growth rate (RGR), and increased root: shoot ratios (Bazzaz, 1990). However, as the variety of species studied and length of experiments have increased, considerable interspecific variation in the magnitude and duration of this "typical" response to elevated CO2 has been documented. Assimilation rates may undergo negative acclimation, declining over time to a point equivalent to or even lower than that of plants photosynthesizing in ambient CO2 (DeLucia et al., 1985). Decreased stomatal conductance (Gs) may bring about increased leaf temperatures and vapor pressure deficits at the leaf surface, resulting in increases in transpiration (Bowes, 1993). Short-term increases in relative growth rate (RGR) may decline over time; early gains in biomass in l 2 plants exposed to elevated CO2 may not be maintained (Larigauderie et al., 1988; but see Poorter et al., 1988). Biomass allocation patterns in response to elevated CO2 appear to species-specific, with certain species tending to increase root: shoot ratios in an elevated CO2 environment, while other species do not (see Woodward et al., 1991, and Stulen and den Hartog, 1993, for reviews). Complicating the picture even firrther is the fact that environmental conditions also affect the magnitude and duration of response to elevated C02. For example, high nutrient environments may magnify and prolong CO2 - mediated changes in physiological and growth processes, while low nutrient conditions may have the opposite effect (Patterson and Flint, 1982; Larigauderie et al., 1988). Moreover, the nature of the response to elevated CO2 may be an intrinsic, genetically determined trait of certain species. For example, plant species from less productive communities such as serpentine grasslands or the arctic tundra do not appear to respond as strongly to elevated CO2 as plants from more productive environments (Williams et al., 1988; Oechel and Strain, 1985) The ability of a plant species to maintain a positive response to elevated CO2 over the long term may depend upon its life history and its ability to maintain appropriate internal source-sink relationships. Negative photosynthetic acclimation may not occur, or may not proceed as rapidly, when plants possess adequate sinks for carbohydrate in the form of roots, fruits, or other storage products (Woodward et al., 1991; Ziska and Teramura, 1992). Similarly, woody plants that can allocate carbohydrate to structural tissues or herbaceous plants that can export photosynthate to belowground storage in tubers or roots may have a greater capacity to maintain a positive long-term response to 3 elevated CO2 (O'Neill et al., 1987; Bhattacharya et al., 1985; Arp and Drake, 1991). Species-specific variation in ability to respond to elevated CO2 may affect community level interactions. Exposure to elevated CO2 may act as a release from carbon limitation for C3 species, putting them at a competitive advantage compared to C4 species which are not carbon-limited at current ambient CO2 levels (Patterson et al., 1984; Carter and Peterson, 1983). Intrinsic differences in species' abilities to allocate tissues to roots to increase the acquisition of nitrogen or water, or to reallocate nitrogen away from photosynthesis to other critical processes within the tissues, may increase the ability of some species to utilize these limiting nutrients at the expense of others (Arp, 1991). Since elevated CO2 can affect early growth parameters such as germination and seedling size (Wullf and Alexander, 1985), species which are able to take advantage of additional carbon early in the growmg season may alter the plant community in terms of plant density and species diversity (Goldberg and Miller, 1990). Thus, elevated CO2 may affect plant species interactions with the abiotic environment and with other members of the community (Tilman, 1993). The discovery of intraspecific variation in the ability to respond to elevated CO2 has provoked interest in determining whether plants may evolve in response to elevated CO2 (Geber and Dawson, 1993). Physiological traits display genetic variation and are heritable (Scheiner et al., 1984; Tonsor and Goodnight, 1996). Studies of agricultural crops have revealed that genetic variation in response to elevated CO2 affects traits such as seed number and seed size, aspects of yield that are comparable to evolutionary measures of fitness (E. g., Ziska and Teramura, 1992). Very few studies have been conducted that examine the extent of intraspecific variation in response to elevated CO2 4 for native plant populations, but the results of those studies indicate that heritable intraspecific variation in fitness traits in response to elevated CO2 exists (Garbutt and Bazzaz, 1984; Curtis et al., 1994). If elevated CO2 will act as an agent of natural selection, intraspecific variation in physiological response to elevated CO2 must translate into variation in such fitness components as survivorship, size, and fecundity (Geber and Dawson, 1993). Under elevated C02, increased rates of photosynthesis have been demonstrated to have a positive effect on growth and biomass accumulation, resulting in predictions of increased crop yield of 30—40 % (Cure and Acock, 1986). Yet the mechanistic connection between physiological traits and growth remains unclear (Pereira, 1994). Relative grth rate (RGR) is expressed as the product of net assimilation rate (NAR) and leaf area ratio (LAR): RGR = NAR x LAR NAR is usually positively correlated with photosynthetic rate, while LAR is usually positively correlated with specific leaf area (SLA) (Konings, 1989). Under current ambient CO2 conditions, high RGR appears to be more strongly associated with high SLA than high photosynthetic rates (Shipley, 1995). Because exposure to elevated CO2 often results in increased photosynthetic rates, but decreased SLA, the net effect of elevated CO2 on size and yield (surrogates for fitness) will depend upon how members of a species integrate these two potentially counteracting responses to elevated CO2 into growth. Because the interaction of a plant's genotype with its environment determines its phenotype, exposure of the genotype to a novel environment, such as twice-ambient C02, may result in the expression of a different phenotype. A novel environment may bring about new interactions among genes, potentially resulting in unexpected phenotypic outcomes. That is, the relative contribution of various genes to a given trait, such as 5 growth rate, may vary in different environments (Wright, 1969). Variation in gene interactions may result in shifts in the genetic correlations between traits, which in turn can change the constraints on the independent evolution of correlated traits (Via and Lande, 1983). Changes in genetic correlations as a response to elevated CO2 may result in a release of genetic variation currently masked by the ambient CO2 environment, and may make short-term evolutionary response to elevated CO2 possible (Bradshaw and McNeilly, 1991). Therefore, documenting the extent of phenotypic family level variation in response to elevated CO2 is a necessary first step in determining the extent of genetic variation upon which elevated CO2 may act as an agent of natural selection. A pilot experiment I conducted in 1991 at the Duke University Phytotron indicated that just such intraspecific variation in response to elevated CO2 existed for maternal families in two populations of Plantago lanceolata (see Appendix). Multivariate analysis of the physiological traits of assimilation, transpiration, and stomatal conductance revealed significant CO2 environment by family interactions in nutrient-rich, controlled greenhouse conditions. Family level variation was also detected for aboveground and belowground biomass, but not for root: shoot ratios. The results of the 1991 experiment led me to conduct another set of experiments in 1992 in the more natural conditions offered by open-topped environmental chambers set up in an old field at the W. K Kellogg Biological Station in Hickory Corners, Michigan. The goals of the 1992 experiment were three-fold: (l) to document the extent of genotypic variation in response to elevated CO2 for physiological, growth, and biomass allocation traits, (2) to determine the nature of the duration of response to elevated CO2 by comparing growth, allocation, and physiological traits after short-term exposure to elevated CO2 with the same traits after exposure to 6 elevated CO2 for an entire growing season, and (3) to explore how source-sink relations in an herbaceous perennial such as Plantago lanceolata might affect the ability of members of the species to integrate long-term response to elevated CO2 with respect to physiology, grth and biomass allocation. CHAPTER ONE INTRASPECIFIC VARIABLE RESPONSES TO ELEVATED ATMOSPHERIC C02: RESOURCE PARTITIONING IN ABOVE- AND BELOWGROUND TISSUES IN PLAN TAGO LAN CEOLATA L.* *This chapter was co-written with Susan Kalisz, Stephen J. Tonsor, Peter S. Curtis, and James A. Teeri for submission to Oecologia. INTRODUCTION Recent increases in atmospheric CO2 concentration have been clearly documented (Neftel et al., 1985; Keeling, 1986). Extrapolation from the current pattern indicates that this increase will continue, with a doubling of crurent CO2 levels being likely by the middle of the next century (Strain, 1987). Numerous studies have established what is now considered a picture of the "typical" whole plant response to elevated C02, at least in the short term In C3 species whole plant biomass and yield typically increase (Kimball et al., 1993; Poorter, 1993; Cure and Acock, 1986), and patterns of biomass and resource allocation within the plant are often altered. For example, root:shoot ratios generally increase under elevated C02. Despite increases in root biomass, nitrogen uptake may not keep up with carbon Supply, resulting in increased carbon: nitrogen ratios (see Bowes, 1993; Woodward et al., 1991; Bazzaz, 1990 for reviews ). However, as the variety of species studied and the length of experiments have increased, considerable interspecific variation in the magnitude and duration of this "typical" response to elevated atmospheric CO2 has been documented. While many reviews cite an increase in rootzshoot ratio under most conditions (Bowes, 1993 ; Woodward et al., 1991; Bazzaz, 1990), some studies have documented no increase in rootzshoot ratio (see Stulen and den Hartog, 1993, for a review). Some reports have shown that the initial increase in biomass of elevated CO2 grown plants compared to ambient grown plants was maintained throughout the experiment (Poorter et al., 1988; 9 Smith et al., 1987); in other studies early differences in biomass among CO2 environments disappeared by the end of the experiment (Larigauderie et al., 1988; Norby et al., 1987). There are several possible explanations for this variety of experimental results. Growth conditions may affect the magnitude and duration of response to elevated atmospheric C02. High nutrient environments may magnify and prolong COz-mediated increases in overall biomass, with no change in root:shoot ratio (Patterson and Flint, 1982). Low nutrient conditions may dampen the magnitude or curtail the duration of a positive biomass response to high CO2 as other resources besides carbon become limiting, while shifting the allocation of biomass to roots at the expense of shoots (Larigauderie et aL,l988) While environmental conditions play a role in governing the response of plants to elevated C02, some of the variation in the magnitude and duration of response may be related both to the nature of the environment to which the species is adapted, and to the life history and phenology of the species. That is, the ability to respond to elevated CO2 may be an intrinsic, genetically determined trait of certain species. For example, plant species from low resource communities such as serpentine grasslands or the arctic tundra do not appear to respond as strongly as plants from more productive environments (Williams et al., 1988; Oechel and Strain, 1985). The ability of a plant species to maintain a positive response to elevated CO2 over the long term may depend upon its life history and its ability to maintain appropriate source-sink relationships (Bowes, 1993; Arp, 1991). In nutrient-rich agricultural systems, crops with yield components primarily aboveground such as soybeans and cotton generally do not increase rootzshoot ratio in response to elevated CO2 (Idso et al., 1988). On the other hand, root crops such as radishes, carrots 10 and sweet potatoes generally allocate more to root growth and maintain a positive response to elevated C02, presumably because they have ample sink strength in underground organs for carbohydrate (Idso et al., 1988; Bhattacharya et al., 1985). In a salt marsh community, perennial herbaceous species that were able to maintain an adequate sink for carbohydrate belowground maintained a positive response to elevated CO2 for more than four years (Curtis et al., 1989; Arp and Drake, 1991). Thus, elevated atmospheric CO2 would seem to affect species which allocate biomass significantly to belowground tissues differently than species which do not. Because CO2 can have a direct effect on plant performance and since the response of plants to elevated atmospheric CO2 varies with species, there is growing interest in determining whether plant species will be capable of an evolutionary response to increased CO2 (Geber and Dawson, 1993). Ifa species' ability to respond to elevated atmospheric CO2 is genetically determined, intraspecific genetic variation in response to CO2 may also exist. Clearly, if CO2 is to be expected to act as an agent of natural selection, the species must exhibit heritable variation in the magnitude of response to elevated C02, and the response of a species to CO2 must be of lasting duration and affect plant fitness in some way. Studies of agricultural crops have revealed genotypic variation in elevated CO2 effects in characteristics such as seed number and seed size, aspects of yield that are comparable to evolutionary measures of fitness (e.g., Ziska and Teramura, 1992). In wild plant species, elevated CO2 may affect traits both directly and indirectly associated with fitness. For example, size at the end of one growing season may be associated with reproductive success in the next season in perennial species (Prirnack, 197 9). Very few studies have been conducted that examine the extent of intraspecific ll variation in fitness traits in response to CO2 explicitly (Garbutt and Bazzaz, 1984; Curtis et al., 1994), but the results of those studies indicate that heritable intraspecific variation in native plant populations in response to CO2 exists. The purpose of this experiment was two-fold: (1) to document the extent of intraspecific variation in whole plant traits associated with fitness in response to elevated atmospheric CO2 in a naturally occurring plant species; and (2) to explore the ways in which aboveground and belowground tissues are allocated in response to elevated CO2 in a perennial herbaceous species that has a significant belowground sink for carbohydrate. We selected two populations of the short-lived herbaceous perennial, Plantago lanceolata, and measured overall biomass of maternal families from those populations after exposure to ambient and elevated CO2 for one full growing season. Overall biomass is a trait that has been associated with lifetime fitness in this species (Primack and Antonovics, 1982). We also examined how that biomass was allocated between aboveground and belowground components (rootzshoot ratio). Finally, we measured tissue chemical composition in leaves and roots for parameters hypothesized to be associated with the maintenance of a positive response to elevated C02: soluble sugar and starch content, and percent carbon and percent nitrogen content. MATERIALS AND METHODS Plantago lanceolata has been used extensively in physiological, ecological and genetic studies (Tonsor and Goodnight, 1996; Kuiper and B05, 1992; Tonsor, 1985, 1990; Teramura, 1983). Populations of this species grow in a variety of habitats, and have been found to exhibit both local genetic variation (Tonsor, 1985, 1990; Tonsor et al., 1993; 12 Teramura 1983; Teramura and Strain, 1979) and phenotypic plasticity (Teramura and Strain, 1979; Antonovics and Primack, 1982). The species is known to be genetically variable for physiological and morphological traits (Tonsor and Goodnight, 1996; Teramura, 1983; Teramura and Strain, 1979), and is capable of undergoing rapid evolutionary change (Wolfi‘ and Van Delden, 1989; Wu and Antonovics, 1976). Moreover, the species maintains a significant belowground sink for carbohydrate in the form of a rhizome (Teramura, 1983) and is capable of shifting allocation between aboveground and belowground tissues under changing nutrient conditions (Van der Aart, 1985). Studies of the effects of elevated CO2 on P. lanceolata have revealed genetic variation in allelochemical content (Fajer et al., 1992) and early growth parameters (Wulff and Alexander, 1985). The study populations of Plantago lanceolata were chosen to represent two distinct habitats. The Ely Lake (EL) population (Allegan County, Michigan) grows on exposed sandy soil on a sunny lakeshore, experiencing high irradiance, low nutrient availability, and periodic water stress. The Kellogg Field (KF) population (Kalamazoo County, Michigan) grows in partial shade on the edge of a mown field. The Kellogg Field soil is higher in both organic matter and water availability. A total of 24 families, twelve randomly selected from each of the two populations, were used in the experiment. On June 5, 1992, all seeds from each family were divided into two equal groups, planted in separate flats, and placed in either ambient or twice- ambient (hereafter referred to as elevated) CO2 to germinate. Germination took place over 7 - 10 days. On June 17, six maternal siblings from each family in each CO2 environment were transplanted into separate 30-cm high pots made from 10-cm diameter -- 13 PVC pipe with mesh screen bottoms. The day of transplanting was considered Day 1 of the experiment. The pots were filled with a 50-50 mixture of low organic matter field soil (Kalamazoo loam) and sand. 144 pots (6 pots x 12 families x 2 populations) were distributed randomly among four replicate outdoor open-top chambers in each CO2 environment. Both ambient and elevated CO2 chambers had one-meter square internal dimensions, contained 36 pots each, and were constructed following the protocol of Curtis and Teeri (1992). To determine the effect of the chambers themselves, 72 additional seedlings from each population were planted in individual pots and distrrbuted - randomly among four 1 m2 unchambered control sites. The chambers and unchambered control sites were arrayed in four randomized blocks in a recently abandoned field adjacent to the Terrestrial Field Laboratory at Kellogg Biological Station, Hickory Comers, Michigan. Each block contained one ambient CO2 chamber, one elevated CO2 chamber, and one unchambered control site. Prior to the experiment, the site was cleared, herbicided, and disked to smooth out uneven patches. The experimental site experienced firll sun throughout the day. The plants were watered as needed, usually twice daily. There was no fertilizer supplementation. Pure C02, mixed with ambient air by ventilation fans, was supplied 24 hours per day to the elevated chambers. Ambient air was circulated within the ambient chambers by the same type of fan. CO2 levels were monitored continuously and levels were recorded on a computer at three-minute intervals (Curtis and Teeri, 1992). Mean daytime (0700- 1900 hours) CO2 partial pressure inside the elevated chambers was 72 i 6 Pa (:1: s.d.), and 36 :t 3 Pa inside the ambient chambers. Quantum sensors and shaded thermocouples attached to a L1- 1000 datalogger (LICOR Inc., Lincoln NB, USA) 14 recorded irradiance levels and temperature. Daytime temperatures were 1.7 i 0.6 °C higher inside the chambers than in the unchambered control sites, with no significant difference in temperature between ambient and elevated chambers. Three weeks into the experiment the young plants were exhibiting symptoms of light stress (prostrate growth and red pigmentation of the leaf bases) so all chambers and control sites were covered with neutral density shade cloth. The shade cloth reduced ambient light by 68%, and the plants recovered their normal phenotype. Leaf number was counted for each plant on Day 1, 30, 55, and 127 of the experiment. On October 21, after 127 days of growth, and following several frosts and a snowfall, the plants were harvested. The roots and leaves (shoots) were separated at the soil line, dried at 60°C for five days, and weighed. For a subset of three families, shoot and root tissue were analyzed separately for soluble sugar, starch, % carbon, and % nitrogen content. For this part of the analysis, we selected three families having a complete sample size of six individuals in each CO2 environment at the end of the experiment; we deliberately selected families which appeared to be thriving in both CO2 environments, but whose overall appearance was qualitatively distinct. Sample size for each of the two CO2 treatments for tissue content analysis was 18: n = 6 plants per family x (3 families). Plants fiom the unchambered control sites were not included in the tissue analysis. The shoot and root tissues of each plant were analyzed separately for % carbon and % nitrogen content with a Carlo Erba NA1500 Series II CHN analyzer (Fison's Instruments, Paramus NJ, USA). Starch content and the combined concentration (as glucose units) of sucrose, glucose and fi'uctose (mg soluble sugar/ g dry weight) were analyzed enzymatically in shoot and root tissue (Jones et al., 1977). 1 5 Statistical Analysis Data from the main experiment and the subset of three families used for tissue chemical composition analysis were analyzed by analysis of variance (PROC GLM) using SAS (SAS Institute, Inc., 1988). The model for the analysis of the main experiment was: Yum = u + at + 13,- + Yk + 5k] + (9er +0351” + ear- Block (a), C02 level (0), population (y), and family nested within population (6) were main effects in the model. CO2 x population (By) and CO2 x family nested within population ([35) were interaction terms. Sample size for each family in each CO2 environment was ranged from 2 - 6. Of the 24 families used in the main experiment, 6 families were excluded from analysis because 2 or fewer individuals germinated in one or both CO2 environments. The model for the analysis of the three families used for tissue chemical composition analysis was: Yij = u + a, + p, + (043)ij + eij. CO2 level (or) and family ([3) were main effects in the model; CO2 x family was the interaction term Because only one family from population EL was included in the tissue composition analysis, a population effect could not be determined. All variables in both analyses were normally distributed except rootzshoot ratio. A log transformation of root1shoot ratio to achieve normality did not change significance levels and the untransformed data were used in the overall AN OVA. Because a comparison of family means was intended initially, all main effects and interactions in the model were considered to be fixed (Gill, 197 8). A statistically significant population or family (population) value demonstrated that these populations or families could be distinguished fiom one another in their response to elevated C02. A l6 statistically significant interaction term indicated that the populations or families differed in the magnitude and/or direction of response to the elevated CO2 environment. To quantify further the nature of population and family level differences, each population and each family was tested individually for its response to CO2 environment, using one-way ANOVAs. Because of the small sample sizes (11 s 6 ) for each family we used both p < 0.05 and p < 0.1 level of statistical significance, as recommended by Gill (1978). RESULTS BIOMASS ALLOCATION -CO2 Environment Effects- Plants grown in both ambient and elevated chambers had larger biomass and smaller root:shoot ratios than plants in the unchambered control sites (data not shown). Because the chambers had similar effects on both ambient- and elevated-CO2 grown plants, the chamber effect was not incorporated into the analysis of experimental results and the unchambered control treatment will not be considered further. Plants grown in elevated CO2 had significantly greater belowground biomass, whole plant biomass and rootzshoot ratio than ambient CO2 grown plants (Table 1; Figure 1). Final aboveground biomass did not show a significant CO2 response, although leaf number was significantly greater (p<.05) for the plants grown in elevated CO2 earlier in the experiment (Days 27 and 59, data not shown). -Population Level Effects- Both populations EL and KP had significantly greater belowground biomass and rootzshoot ratios in elevated CO2 compared to the ambient CO2 environment (Figure 1). 17 - - N0 .0 - - 00v - - 0.0 - - ”3 NH Hobm #0050 5.0 H00 H0000 0.0 H .0m N0000 w .0 H. .w H 30.0 0.0 0 .0H 0H Amomvbfidam x N00 9000 M00 0000.0 3000 V0 HSH 0000 .0 m .0 0d 0m H00 m0 m0 H :oHuaHEHoAH x NOD 0m00 .0 HEN 00.0 vm00 .0 m .N 0 0H H H H000 0 .N N.mH N300 m .N 0.00 0H Anew—«Haemv bani Nw0w.0 m0 .0 0000 .0 N000 .0 m .3 w .moh 0000.0 0 . H H m .00 M000 .0 v .m H 0 .mmm H «80230.5 #000 .0 0 .mH M0 0300 H .0 m .00N N000 .0 0 .mH 0.0m mvH H .0 m .N 0 .N0 H N00 0300 m.m H0 :30 mH 5H0 M0000 m0 0H $00.0 0N 000 m VHconH & ”H m2 R "H mH>H oH "H m2 Q m mHZ .HHo 3:80 3% magma—lg 003380 E H358 H355 .moSeES $385 How confine/Ho mHmrmHaSs .H 203. 18 Figure 1. Comparison of aboveground biomass, belowground biomass, and root: shoot ratio by C02 treatment. For each category (overall, population, family) means for ambient CO2 appear to the left of means for elevated C02. Vertical bars indicate one standard error. Aboveground biomass, belowground biomass, and root: shoot ratios were considered separately for each category. Scale for root: shoot ratio appears to the right. Significance levels appear above bars for aboveground biomass, below bars for belowground biomass, and immediately adjacent symbol for root: shoot ratio. Whole plant biomass differed significantly across CO.2 treatment for overall data (*), population KF (*), family EL 7 (*1), family EL 9 (*), and family KF 31 (*). Significance levels: *1 p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. 19 Root: Shoot Ratio 1.5 5 I f ‘ \\\;¥\'\ ‘u. \\‘.\;\‘~'\:u‘.\; \:-:~’ "\\' 50" N ‘a ‘\2\\‘ §\\: \‘I.\. (V Sick-N sassaaa 5; ms. I--l « I3; azx‘sasa‘aaas w :m ‘- :aassmaaam 68 :l)l EILEKW £8 :l)i :sssaaaasaaaas I—r :aaamaas 99 :m l—I H W :l)i EV :l)i *1 2’? :lM + Root Shoot Ratlo 11 II I II amasaaaas 89 :lM = W 1—1—1 asssaaaasasaa as :m amass ;- LE :l>i l—i aaassmsaa :- 9 L13 l—l ‘— =l maamsss 0 l '13 l-l 1" as: 3... as H swam '3 613 I mamas * H Belowground Biomass mwa 9 '13 —: 3W "‘ =laasssma . 1 13 H 22:32 WE‘RE — rams 9 "13 E: =aasaaaas 17 "El —I saaas Z 13 —: 7:4: if H * Isaaaaaaaax ; :m dod _:: maasaaaaa a 13 docs r r, lammm E "BJSAQ i E] Aboveground Biomass IIIIIIIIlllllllllllllllllllll 0 lo 0 m o to O 6‘ (‘0 N N ‘— ‘_ v (10) (15) (5) sseuroga Population KF Figure 1 Population EL Summary Data 20 Aboveground biomass did not differ significantly for either population across CO2 treatments. Only population KF had significantly greater whole plant biomass (p < .05) in elevated C02. Within the ambient CO2 environment, there were significant differences between populations EL and KP for all biomass variables except rootzshoot ratio (Figure 1). Plants from Population KF were significantly larger than those from population EL for aboveground biomass, belowground biomass and total biomass (p < .05). Although there was no CO2 x population interaction (Table 1), population KF responded proportionally more to elevated CO2 for belowground and whole plant biomass than did population EL. Belowground biomass for population KF was 25% greater in the elevated than in the ambient CO2 environment; for population EL the increase in belowground biomass in the elevated CO2 environment was 22%. Whole plant biomass in the elevated CO2 environment was 16% greater for population KF and 11% greater for population EL relative to the ambient CO2 environment. -Family Level CO2 Effects- In contrast to the overall positive response to elevated CO2 at the population level, the family means showed significant variation in direction and magnitude of response (Figure 1). Three types of responses emerged when family means in the ambient and elevated CO2 environments were compared (Figure l). (1) Increase in biomass - Across both populations, nine families showed significant increases in one or more biomass variables in the elevated CO2 environment compared to the ambient environment. Family EL 9 exhibited increased aboveground, belowground and whole plant biomass in elevated C02. Family KF 31 had increased belowground and whole plant biomass under elevated C02. Two families (EL 15 and KP 42) only increased belowground biomass under 21 elevated C02. Five families (EL 4, EL 10, KF 39, KF 41, and KP 44) increased their root: shoot ratios under elevated C02. (2) No change in biomass - Eight families showed no significant response to the elevated CO2 environment for any of the biomass variables. (3) Decrease in biomass - One family, EL 7, was significantly smaller in aboveground and whole plant biomass in the elevated compared to the ambient CO2 environment. These three types of responses reveal how the families in these populations differed in which portions of their biomass, if any, responded to the elevated CO2 environment. It is clear that there is no consistent, generalized response to elevated CO2 at the family level for these biomass variables. When relative allocation to aboveground and belowground tissues (that is, root: shoot ratio) was compared among CO2 environments for the individual families, the general response was toward increased root: shoot ratios. However, the means by which the individual families achieved this response to elevated CO2 varied considerably. Of the five families with significantly greater rootzshoot ratios in elevated CO,, two routes to achieve this response were detected: (1) little change in aboveground biomass accompanied by a tendency to increase belowground biomass (families EL 4, EL 10, KF 39, and KP 44); and (2) aboveground biomass being smaller in elevated CO2 than in ambient CO2 with little change in belowground biomass (family KF 41). Thirteen of the eighteen families analyzed did not significantly increase root: shoot ratios in elevated C02, although the trend in nine of these families was toward greater root: shoot ratios in elevated C02. There was not always sufficient statistical power for resolving small differences in family mean responses among CO2 environments for small sample sizes (11 s 6 for each family with each CO2 environment). 22 Family performance in ambient CO2 was not a good predictor of family performance in elevated C02. Several families which were either the largest or smallest for a particular biomass variable in the ambient CO2 environment changed rankings with respect to the other families when in an elevated CO2 environment. Families EL 5 and KF 33 had comparatively small aboveground biomass in ambient CO2 but were close to the largest families in aboveground biomass in elevated C02, while families EL 7 and KP 41 decreased in aboveground biomass in elevated CO2 relative to the other families (Figure 1). Similarly, families EL 15 and KF 33 moved up in ranking for belowground biomass in the elevated CO2 environment while families EL 7, KF 35 and KF 37 moved down. By contrast, certain families did not change biomass ranking when exposed to elevated CO2 . For example, family KF 32 had the largest aboveground, belowground, and whole plant biomass and family KF 31 had the highest rootzshoot ratio of any family in population KF in ambient and elevated C02. TISSUE CHEMICAL COMPOSITION -CO2 Environment Effects- For families EL 15, KF 32 and KF 35, there were significant overall differences in soluble sugar content, carbon content (%DW) , nitrogen content (%DW), and carbonznitrogen ratios between ambient- and elevated-CO2 grown plants. Overall, whole plant soluble sugar was 15% higher in the elevated CO2 grown plants (59.3 mg/g dry weight vs. 51.8 mg/g, p<. 005) with this increase present in both aboveground and belowground tissues (Table 2, Figure 2). Starch content was less than 1% for all individuals in the three families in both environments. Whole plant % carbon was slightly, although significantly, higher in the elevated compared to the ambient CO2 environment - - n: - - es - - em - - was 8 3am ease to 3 «Sad os o.o ”ES 3 2 aged so an Haoesmaaa x N8 «was 3 3m 82 .o 3. S S: .o M: a: ”Sec 3 3% N assuage bass N 23 0800 5: NémH m0~00 0.0 N0 M000 .0 m .0 TN N080 m.m WSNN H NOD AN m mH>H K be venom - - 0.0m - - H0 - - Ne - - H.HmH 0m Bum 020.0 0.? w.h0H 0000.0 0.0 v0 0: m0 m0 0N H.00H .0 0H deN N Amomvrmmaam x ~00 m0000 H.HH 0.0mm H0000 m .NH 00 m00H .0 HMH NS 05¢N0 m .H N.NOH N 30320096 333.0 030.0 «av H.mmH 0030 m0 m0 0m000 0.0 mHv HmH00 0H. 0.NHO H NOU R ”H m2 oH «Ho 358m 03:8on mH 033.8,» Home How Hovoa HHSo>O .Amm ”UH “mm ”UH .mH Hm: 35:80 086 80 3800088 838088 25me new oooafigmo mHm>Hao< .N «Hana. 24 Figure 2. Comparison of soluble sugar, % carbon content, % nitrogen content and carbon: nitrogen ratio by C02 treatment. For each category (overall and family) means for ambient CO2 appear to the left of means for elevated C02. Vertical bars indicate one s.e. Aboveground and belowground means were considered separately for each variable. Significance levels appear above bars for aboveground means and below bars for belowground means. Significance levels: *1 p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. 25 N 2:9... mm my. mm ”2 3mm. =Em>o onv mm “2 mm ”2 3 ._m __Em>o A v k. k. «:1 m.—‘ Momw m E /o A? o w H o m ...... S m w T3 w w mm; m «a . D... , .I 9 v F n , ad 2390328 I 23235.? mm “2 mm “E 3 .m. =Eo>o A 0 mm ”2 mm “2 2 .m. .6330 l , - A80 F ea Amw % .. 39 my 0 AOVV m A0 5 m. A v m H o m cm mV wow 0 . 0 w n 0 J “won m. . 9. m 4. .1” 1 fl we. 8 m on « om 26 (38.9 % vs. 37.5 %, p < .05). The increase in whole plant % carbon was due entirely to an increase in carbon content in aboveground tissues in the elevated CO2 environment, as belowground tissues did not differ significantly in % carbon content (Figure 2). Whole plant nitrogen content (%DW) showed no CO2 response, although the distribution of nitrogen between aboveground and belowground tissues was affected by CO2 environment. In elevated C02, % nitrogen content was significantly higher in aboveground tissues, but lower in belowground tissues compared to plants in ambient CO2 (Figure 2). Carbon: nitrogen ratios were lower in aboveground and higher in belowground tissues under elevated C02 (Figure 2). -Family Level CO2 Effects- There were significant differences among the three families EL 15, KF 32, and KP 35 in response to elevated CO2 for soluble sugar content, % carbon, % nitrogen, and carbonznitrogen ratios (Figure 2). Families EL 15 and KP 35 had significantly more aboveground (and whole plant) soluble sugar under elevated CO2 , while the soluble sugar content in family KF 32 was not affected by C02 treatment (Figure 2). There were small but significant increases in aboveground % carbon content in families KF 32 and KF 35 in the elevated CO2 environment, but no increase in % carbon content in family EL 15 (Figure 2). Belowground soluble sugar and % carbon content did not change significantly with CO2 environment for any individual family. There were significant family level and CO2 x family interactions for aboveground % nitrogen content and aboveground carbon: nitrogen ratio (Table 2). Families KF 32 and KP 35 had higher aboveground % nitrogen content in the elevated CO2 environment than in the ambient, while family EL 15 showed no change. Belowground % nitrogen was 27 not afl‘ected by C02 in any of the three families, but families KF 32 and KF 35 both contained significantly less nitrogen belowground relative to aboveground under elevated C02. In family KF 32, the carbon: nitrogen ratio decreased aboveground, but increased belowground under elevated C02, while family EL 15 showed increased belowground carbon: nitrogen ratio only. Family KF 35 showed no response to C02 in above- or belowground carbon: nitrogen ratio. DISCUSSION Our results demonstrate significant within and among population variation in Plantago lanceolata for response to elevated C02. Belowground biomass, whole plant biomass, and rootzshoot ratios increased in the elevated CO2 environment, but we found considerable variation in the direction and magnitude of response to elevated CO2 at the level of the population and family. Populations EL and KP responded to elevated C02 similarly by increasing biomass. Belowground biomass was most responsive to elevated C02. This difl‘ered from the 1991 experiment in which aboveground biomass was also larger in the elevated CO2 environment under more productive greenhouse conditions (see Appendix). Increases in belowground biomass in the 1992 experiment resulted in increased rootzshoot ratio in both populations in the elevated CO2 environment. Yet the two populations were not identical in their responses to elevated 002 (Figure 1). Plants from Population KF were larger than those fi'om population EL for all biomass variables except rootzshoot ratio, and Population KF showed a significantly greater response to the elevated CO2 environment for belowground and whole plant biomass than Population EL. The family responses to elevated CO2 in biomass and biomass allocation were far 28 more complex than the population-level analyses suggest. Depending upon the family, either aboveground biomass, belowground biomass, both biomass components, or neither biomass component showed a response to elevated CO2 (Figure 1). In general, though, belowground biomass was more responsive to elevated C02. While only eight families maintained a greater than 10% increase in aboveground biomass for the entire growing season, a much larger number of families (thirteen out of eighteen) maintained a greater than 10% increase in belowground biomass in the elevated C02 environment. This resulted in the general response of increased root: shoot ratios in the families grown in elevated C02. However, different families may well have followed different physiological and deve10pmental patterns to achieve a similar response to elevated C02, illustrated by the different routes taken to achieve increased root: shoot ratio by the individual families. There are two possible explanations for the greater responsiveness of the belowground biomass component to the elevated CO2 environment. The first explanation draws upon the model of balanced carbon and nitrogen allocation between shoots and roots (Davidson, 1969; Thornley, 1972: Johnson, 1985). This model predicts that as aboveground tissues experience an increase in carbon supply, allocation to belowground tissues increases to balance the nutrients within the plant. The results of the tissue composition analysis were not consistent with this model. When the three families, KF 32, KF 35, and EL 15, were analyzed for carbon and nitrogen content, we found that families KF 32 and KP 35 increased nitrogen allocation to the shoots under elevated C02, without significantly increasing root: shoot ratio. Family EL 15, on the other hand, had a significantly greater root: shoot ratio in elevated COZ, but did not increase % nitrogen content aboveground. Thus, an increased root: shoot ratio did not result in increased 29 allocation of nitrogen aboveground. However, because such a small number of families was included in this portion of the experiment, it is clear that a much larger scale study is needed to explore the relationship between biomass, carbon and nitrogen allocation fully. A second explanation for an increase in allocation to belowground biomass under elevated CO2 relates to the life history of this particular species. Plantago lanceolata can increase carbon allocation to belowground biomass in order to maximize survival and reproductive potential in subsequent growing seasons (Van der Aart, 1985; Teramura, 1983). When all eighteen families were measured for changes in biomass allocation under elevated C02, our experiment revealed that a greater carbon supply (in the form of increased C02), coupled with a large belowground sink for carbon, makes the pattern of biomass allocation at high CO2 in Plantago lanceolata consistent with results from root crops, in which rootzshoot ratios increased in elevated CO2 environments regardless of the availability of other resources (Idso et al., 1988). Similarly, Bromus mollis, a naturally occurring perennial species, also accumulated significant belowground biomass in an elevated CO2 environment, even in soils that were not resource limited (Larigauderie et aL,1988) In an elevated CO2 world, it is possible that the population structure and community interactions of this species may change. In our experiment, it was not possible to predict how a family would respond to elevated CO2 simply by examining its performance in ambient C02. Several families changed rankings relative to one another in the elevated CO2 environment, becoming either relatively larger or relatively smaller in the elevated CO2 environment than members of the same family in the ambient CO2 environment (Figure 1). Under an elevated CO2 atmosphere in the natural community in 30 which Plantago lanceolata grows, changes in allocation like those documented in this study can be expected to alter the competitive interactions both among conspecifics and among species. Although most families did respond to elevated C02, it is also important to note that certain families did not respond to the elevated C02 environment by increasing biomass or rootzshoot ratios. The lack of response in these families to the increase in carbon supply may ultimately result in these families being at a competitive disadvantage compared to other families in a world of consistently increasing atmospheric C02. In fact, it is hard to imagine how the families that did not respond to elevated CO2 would be able to interact successfully in a community context in an elevated CO2 world. Ifcompetitive success depends on allocation properties, as has been widely suggested (for example, Grime, 1977, Goldberg, 1991), this study suggests that families that allocate more biomass belowground may be at a competitive advantage in an elevated CO2 world, but which families will respond in this way to elevated CO2 cannot be predicted by their performance under current ambient CO2 conditions. - Implications for Adaptive Response to CO2 Our results demonstrate suflicient variation in response to CO2 at the family level to indicate that CO2 by itself may act as an agent of natural selection in natural plant populations. Plantago lanceolata has been shown to respond evolutionarily to human- induced selective forces of recent occurrence and moderate strength. For example, Wu and Antonovics (1976) documented increased lead tolerance in roadside populations of Plantago lanceolata growing on higher lead soils near the source of automobile emissions, compared to populations away from the road. Rapid evolution in P. lanceolata was also demonstrated by Wolfl‘ and Van Delden (1989), who obtained a response to selection for 31 leaf angle in P. lanceolata in one generation. Other studies have found genotypic variation in this species in elevated CO2 environments for ecologically important traits such as germination rate and seedling size (Wulfi‘ and Alexander, 1985). In our experiment, families varied in biomass accumulation and allocation. Primack and Antonovics (1982) have shown that biomass accumulation and allocation patterns are associated with lifetime survival and reproductive potential (fitness) in Plantago lanceolata, and Tonsor and Goodnight (1996) have shown significant narrow sense heritability for plant size in this species. Thus, it is likely that Plantago lanceolata has the potential for an evolutionary response to increasing levels of atmospheric C02. In this experiment, the entire range of responses to elevated CO2 (positive response, negative response, no response) that has previously been documented at the species level ( e. g., Bazzaz, 1990) was seen within Plantago lanceolata at the family level. This range of responses occurred in a variety of traits, including biomass, rootzshoot ratios, % nitrogen content and ON ratios. Which biological traits are acted upon by natural selection will depend up on the ecological context in which families and populations are found. For example, in the presence of elevated C02, certain traits may experience strong selection in low nutrient or dry environments. One firrther implication of the great variation in response to elevated CO2 among the families in this study relates to recent developments in the measurement of selection (Arnold and Wade, 1984; Kalisz, 1986; Wade and Kalisz, 1990). When measuring the opportrmity for selection, a comparison of means is not necessarily the most illuminating procedure to use in situations in which phenotypic variation increases in one environment compared to another. In this study, the range of family mean values (largest family mean 32 minus smallest) for all of the biomass variables (aboveground, belowground and whole plant biomass, and root: shoot ratio) was greater in elevated CO2 than in ambient CO2 by an average of 33%. Because these measures of plant size are correlated with fitness in Plantago lanceolata (Primack and Antonovics, 1982; Tonsor and Goodnight, 1996), a broader range of mean values for biomass traits might translate into a greater range in relative fitness among the families in these populations in an elevated CO2 environment. Under these circumstances, those families exhibiting extreme responses (either very large positive responses or those in a direction opposite of expectation) may provide more information about the possible role natural selection may play in an elevated CO2 atmosphere than families which exhibit more "typical" responses. An increased variance in relative fitness among genotypes in an elevated CO2 environment would provide a greater opportunity for selection to occur, as defined by Arnold and Wade (1984). Whether shifis in species composition of a community or shifts in species function within a community are the predominant responses to elevated CO2 is largely a question of relative rates of response at the intraspecific genetic and the interspecific community level. It will be important to incorporate the complexity of genetic variation in response to elevated atmospheric CO2 into future studies directed toward understanding the impact of this aspect of environmental change on plant species and communities. CHAPTER TWO PHYSIOLOGICAL VARIATION IN RESPONSE TO ELEVATED CARBON DIOXIDE IN PLANTAGO LANCEOLATA 33 34 INTRODUCTION Carbon dioxide has been determined to be a limiting resource to plants which photosynthesize by means of the C3 pathway (Bowes, 1991). Atmospheric CO2 levels are currently increasing (Keeling, 1986); predictions indicate that the increase will continue well into the next century, with a doubling of current CO2 levels by 2050 likely (Strain, 1987). Elevated atmospheric CO2 has direct elfects on plant life (Strain and Cure, 1985). Exposure to elevated C02, at least in the short term, stimulates photosynthetic, or CO2 assimilation, rates (A, expressed as umol CO2 fixed in2 sec'l). Transpiration rates (E, mmol H20 m'2 s'l) often decline. The increase in assimilation, accompanied by a decrease in transpiration, results in increased instantaneous water use efficiency (calculated as A/E) (Bazzaz, 1990). Accompanying these changes in assimilation and transpiration are declines in stomatal conductance (Gs, expressed as mol H20 m'2 s") and decreases in specific leaf area (SLA, cm2 area g'1 dry weight) (Bazzaz, 1990). Often, relative growth rates increase and plants in elevated CO2 may accumulate more biomass than plants growing in ambient levels of C02, especially when other nutrients are not limiting (Klus, Ch. 1; Bazzaz, 1990; Kimball et al., 1993). However, these initial changes in physiology and growth in response to elevated CO2 may not be maintained. The capacity of a plant to maintain an overall positive response to elevated CO2 depends upon a complex interaction of many physiological traits. For example, the maintenance of increased water use efliciency results from the interaction of the internal processes of assimilation, transpiration, and stomatal 35 conductance. After prolonged exposure to elevated C02, decreases in stomatal conductance may result in increased leaf temperatures and vapor pressure deficits at the leaf surface. Under these conditions, transpiration rates may not decline, ofl‘setting potential gains in water balance due to lower Gs (Bowes, 1993: Schulze et al., 1987). Other physiological processes, such as assimilation and stomatal conductance, may undergo negative acclimation. Assimilation, in particular, may decline over time to a point equivalent to or even loWer than that of plants photosynthesizing in ambient CO2 (Wulfl‘ and Strain, 1982; DeLucia et al., 1985). Photosynthetic acclimation to elevated CO2 has been widely investigated. Most of the mechanisms proposed to explain negative photosynthetic acclimation recognize that an imbalance between carbon and other plant nutrients, such as nitrogen, may occur as supplies of carbon in the form of CO2 increase (Arp, 1991; Bowes, 1991). Much of a plant's nitrogen is tied up in the carbon-fixing enzyme Rubisco (ribulose bisphosphate carboxylase—oxygenase), the single most abundant enzyme in the world. Rubisco comprises 30- 50% of a plant's leaf protein content in C3 species (Bowes, 1991). Bowes (1991) has summarized three nonexclusive mechanisms by which negative photosynthetic acclimation to elevated CO2 may occur. First, plants may reallocate nitrogen from Rubisco by lowering either the quantity or the activation state of Rubisco, slowing down the accumulation of the products of photomthesis (Arp, 1991) . Second, the rate of regeneration of ribulose bisphosphate (RuBP) or inorganic phosphate (Pi) may lag behind the supply of carbon to Rubisco, again reducing the rate of carbon fixation (Sharkey, 1985). Third, starch may accumulate as an end product of photosynthesis and may physically interfere with the diffusion of CO2 into the thylakoids or otherwise disrupt the 36 function of the carbon fixation pathway (Wulff and Strain, 1982; DeLucia et al., 1985). All of these mechanisms proposed to explain the negative acclimation of photosynthesis under exposure to elevated CO2 may involve an imbalance in the supply of two critical nutrients, carbon and nitrogen, or an imbalance between the source of carbon compounds in the plant (photosynthesis) and the destination of those compounds (the sink) (Arp, 1991; Poorter, 1993). If nitrogen supply cannot keep up with carbon supply, or if photosynthate accumulates because the plant does not possess adequate sinks for carbohydrate, the rate of photosynthesis may decrease. Experiments conducted to explore the relationship between source and sink for carbohydrate under elevated CO2 suggest that negative acclimation may not occur, or may not proceed as rapidly, when plants possess adequate sinks for carbohydrate in the form of roots, fruits or other storage products (Arp, 1991; Woodward et al., 1991; Ziska and Teramura, 1992). The ability of a plant to modify physiological pathways and shift patterns of nutrient allocation and acquisition in response to elevated CO2 may be determined by the plant's life history and intrinsic grth properties. Poorter (1993) reviewed growth responses of 156 species to elevated CO2 and concluded that the plants which may be most successful in integrating physiological and allocational responses to elevated CO2 may be those with large source-sink strength and/or those with intrinsically high relative grth rates. Plants that export photosynthate to belowground storage products such as sweet potatoes (Bhattacharya et al., 1985), carrots, or radishes (Idso and Kimball, 1989) typically maintain long-term positive responses to elevated COZ. Woody plants (O'Neill et al., 1987) or herbaceous plants with large root systems may also have a greater capacity to integrate physiological and morphological processes to maintain a long-term positive 37 response to elevated CO2 (O'Neill et al., 1987; Larigauderie et al., 1988; Arp and Drake, 1991) Intrinsic differences in species' abilities to allocate tissues to roots to increase the acquisition of nitrogen or to reallocate nitrogen within tissues, may alter ecological interactions among species. Differences in the magnitude and duration of the CO2 response may increase the ability of some species to utilize limiting nutrients at the expense of others (Arp, 1991). Interspecific variation in the ability to respond in this way to elevated CO2 may alter how a plant interacts with its abiotic environment and with the other members of its community (Tilman, 1993). The discovery of intraspecific variation in the ability to respond to elevated CO2 has provoked interest in determining whether plants may evolve in response to elevated CO2 (Geber and Dawson, 1993). Physiological traits display genetic variation and are heritable (Scheiner et al., 1984; Geber and Dawson, 1990; Radin et al., 1994; Tonsor and Goodnight, 1996). To date, however, no experiments have explicitly documented the extent of genotypic variation in physiological traits in response to elevated C02. The nature of the intraspecific variation in physiological response to elevated CO2 may include variation in the instantaneous capacity to respond positively to elevated C02, and may also include variation in the ability to maintain a positive response to elevated CO2 over long time periods. Ifelevated CO2 will act as an agent of natural selection, intraspecific variation in physiological response to elevated CO2 must translate into variation in fitness components such as survivorship, size, and fecundity (Geber and Dawson, 1993). Under elevated CO2 increased rates of photosynthesis have been demonstrated to have a positive effect on 38 growth and biomass accumulation, resulting in predictions of increased crop yield of 30 - 40% (Kimball, 1983; Cure and Acock, 1986). Yet the mechanistic connection between the physiological behavior of a plant and its growth remains unclear (McGraw and Wulff, 1983; Poorter, 1989; Pereira, 1994). Elevated CO2 often results in increased relative growth rates initially, but those early increases may decline under prolonged exposure (Bazzaz, 1990). Early gains in biomass, though, due to higher initial growth rates, may be maintained over the long term (Wulfi‘ and Strain, 1982; Poorter et al., 1988) and may be key in determining the outcome of competitive interactions. Under ambient CO2 conditions high relative growth rate appears to be more strongly associated with high SLA than high photosynthetic rates (Shipley, 1995 ). Because exposure to elevated CO2 often results in increased photosynthetic rates, but decreased SLA, the net effect of elevated CO2 on size and yield (surrogates for fitness) will depend on how members of a species integrate these two potentially counteracting responses. The goal of this experiment was to determine, for two populations and twenty-four families of Plantago lanceolata, the extent of genotypic variation in physiological response to elevated C02. We measured traits which are known to be involved with both the instantaneous capacity for response to elevated CO2 and with the long-term maintenance of positive response to elevated C02: assimilation, transpiration, stomatal conductance, and water use efficiency. We also measured specific leaf area, in an attempt to connect the physiological behavior of the plants with one aspect of their growth. We investigated variation in the capacity to respond physiologically to elevated CO2 by measuring each plant at both CO2 growth levels, ambient CO2 and twice ambient ("elevated") C02. This allowed us to compare the long-term physiological response of 39 plants grown in elevated CO2 to the instantaneous physiological response of plants grown in ambient CO2 but exposed temporarily to elevated CO2 and vice versa. We also measured physiological traits two times during the growing season, middle and late, to detect whether the physiological response to elevated CO2 changed as the growing season progressed. MATERIALS AND METHODS Plantago lanceolata has been used extensively in physiological, ecological and genetic studies (Tonsor and Goodnight, 1996; Kuiper and Bos, 1992; Tonsor, 1985, 1990; Teramura, 1983). The species is genetically variable for physiological and morphological traits (Teramura, 1983; Teramura and Strain, 1979); Tonsor and Goodnight (1996) found significant narrow sense heritabilities for transpiration rate, photosynthetic capacity, and specific leaf weight in a Michigan population of Plantago lanceolata under current ambient CO2 levels, indicating that the potential exists for evolutionary change in physiological traits for this species in the field. Populations of Plantago lanceolata grow in a variety of habitats, and have been found to exhibit local genetic variation (Tonsor, 1985, 1990; Tonsor et al., 1993; Teramura 1983; Teramura and Strain, 1979), phenotypic plasticity (Teramura and Strain, 197 9; Antonovics and Primack, 1982), and the capacity to undergo rapid evolutionary change (Wolff and Van Delden, 1989; Wu and Antonovics, 197 6). Studies of the effects of elevated CO2 on P. lanceolata have revealed genetic variation in allelochemical content (Fajer et al., 1992) and early growth parameters (Wullf and Alexander, 1985). The study populations of Plantago lanceolata were chosen to represent two 40 distinct habitats. The Ely Lake (EL) population (Allegan County, Michigan) grows on exposed sand on a sunny lakeshore, experiencing high irradiance and periodic water stress. The Kellogg Field (KF) population (Kalamazoo County, Michigan) grows in Kalamazoo loam in partial shade on the edge of a mown field. A total of 24 families, twelve randomly selected from each of the two populations, were used in the experiment. On June 5, 1992, all seeds from each family were divided into two equal groups, planted in separate flats, and placed in either ambient or twice- ambient (hereafter referred to as elevated) CO2 to germinate. Germination took place over 7 - 10 days. On June 17, six maternal siblings fiom each family in each CO2 environment were transplanted into separate 30-cm high pots made from 10-cm diameter PVC pipe with mesh screen bottoms. The day of transplanting was considered Day 1 of the experiment. The pots were filled with a 50-50 mixture of Kalamazoo loam and sand. 144 pots (6 pots x 12 families x 2 populations) were distributed randomly among four replicate outdoor open-top chambers in each CO2 environment. Both ambient and elevated CO2 chambers had one-meter square internal dimensions, contained 36 pots each, and were constructed following the protocol of Curtis and Teeri (1992). To determine the effect of the chambers themselves, seventy-two additional seedlings from each population were planted in individual pots and distributed randomly among four 1 m2 unchambered control sites. The chambers and unchambered control sites were arrayed in four randomized blocks in a recently abandoned field adjacent to the Terrestrial Field Laboratory at Kellogg Biological Station, Hickory Corners, Michigan. Each block contained one ambient CO2 chamber, one elevated CO2 chamber, and one unchambered control site. Prior to the experiment, the site was cleared, herbicided, and disked to 41 smooth out uneven patches. The experimental site experienced firll sun throughout the day. The plants were watered as needed, usually twice daily. There was no fertilizer supplementation. Pure C02, mixed with ambient air by ventilation fans, was supplied 24 hours per day to the elevated chambers. Ambient air was circulated within the ambient chambers by the same type of fan. CO2 levels were monitored continuously and levels were recorded on a computer at three-minute intervals (Curtis and Teeri, 1992). Mean daytime (0700-1900 hours) CO2 partial pressure inside the elevated chambers was 72 i 6 Pa (:t s.d.), with the mean daytime CO2 partial pressure inside the ambient chambers being 36 i 3 Pa. Quantum sensors and shaded thermocouples attached to a LI-1000 datalogger (LICOR Inc., Lincoln NB, USA) recorded irradiance levels and temperature. Daytime temperatures were 1.7 i 0.6 °C higher inside the chambers than in the unchambered control sites, with no significant difference in temperature between ambient and elevated chambers. Three weeks into the experiment the young plants were exhibiting symptoms of light stress (prostrate growth and red pigmentation of the leaf bases) and all chambers and control sites were covered with shade cloth. The shade cloth reduced ambient light by 68%, and the plants recovered their normal phenotype. Physiological measurements were made on all of the plants in the experiment in two separate periods: August 9 - 13 and September 4 - 10, 1992. Measurements were made in an open system using two portable infi'a-red gas analyzers (Models LCA-2 and LCA-3, Analytical Development Corporation, Hoddesdon, UK) and narrow Parkinson Leaf Chambers (Analytical Development Corp oration, Hoddesdon, UK) at both ambient and elevated C02. Elevated CO2 was provided from one of the elevated open-top 42 chambers. The CO2 source (ambient or elevated C02) was rotated to a different machine each day. The machines were set to have identical air flow rates (400 ml min") and saturating light levels (1300 ,uE) for all measurements. Mean leaf chamber temperature was 28.5 °C (s.d. 2.5) in August and 26.8 °C (s.d. 1.7) in September. Measurements of physiological traits were made on an intact, most recently fully expanded leaf of each plant. Each plant was measured at two CO2 levels, first at its own growth CO2 level (ambient or elevated), then at the other CO2 level. For each measurement, the leaf was allowed to equilibrate within the leaf chamber for three minutes before readings were taken. After measurement at both CO2 levels, the leaf was removed from the plant, the area of leaf inside the leaf chamber was excised (if the leaf did not entirely fill the leaf chamber), and all parts of the leaf were pressed, dried and weighed. Specific leaf area for each plant was calculated as the ratio of leaf area (cmz) to the dry weight (g) of the leaf used for the physiological measurements. Calculations of CO2 assimilation (,umol CO2 m'2 s"), transpiration (mmol H20 111’2 s'l), stomatal conductance (mol H20 m’2 s") and intercellular CO2 concentration (phat) for both months were made using equations from Analytical Development Corporation (1992) and von Caemmerer and Farquhar (1981). Statistical Analysis For each month's measurements, preliminary analysis indicated that the four gas exchange variables (assimilation, transpiration, stomatal conductance, and internal leaf CO2 concentration) differed according to experimental block, machine used for measurement, and date of measurement. The variables were adjusted for these differences 43 by subtracting the mean for each block, machine and date from each individual measurement and adding back the grand mean for each month summed over all blocks, machines, and dates (Tonsor and Goodnight, 1996). This procedure removed deviations due to experimental block, intrinsic differences between the two infra-red gas analyzers, and variation in environmental conditions on different days, while retaining the differences due to CO2 treatment and other main effects. The four gas exchange variables were initially analyzed using a multivariate analysis of variance (MAN OVA) to determine overall main and interaction effects. The multivariate analysis was followed by appropriate univariate mixed model analyses for the four primary physiological variables plus instantaneous water use efficiency (the ratio of assimilation to transpiration) and specific leaf area (cmz/ g dry weight of leaf tissue). The overall model for the analysis of the physiological variables included four fixed effects (CO2 treatment, month, CO2 measurement level , and population), one random effect (family nested within population), and two-way interactions for all main effects. Interaction terms containing the random effect were also treated as random effects. CO2 measurement level was omitted from the AN OVA for specific leaf area, since measurement level was not a treatment effect for that variable. Expected mean squares for the main effects in the mixed model AN OVA were divided by the appropriate interaction terms to produce the F values for significance tests. Significance levels for the main effects in the multivariate analysis were interpreted using Roy's Greatest Root, recommended because of its statistical power and the fact that it is applicable to post hoc statistical comparisons (Scheiner, 1993 ). CO2 treatment means for the chambers and unchambered control sites were individually tested for treatment differences, using 44 Bonferroni corrections for multiple tests of means. Sample size for each family in each CO2 environment was 11 s 6. Of the 24 families used in the main experiment, 6 families were excluded from the analyses because 2 or fewer individuals germinated in one or both CO2 environments. All variables in the analyses were normally distributed except transpiration (in August) and stomatal conductance. Log10 transformation of transpiration and square-root transformation of stomatal conductance improved normality. Since each plant in the experiment was measured at two levels of CO2 in each of two months, a repeated measures analysis was conducted. The repeated measures analysis is analogous to a split-plot analysis in which the repeated measure is considered to be a within- subject effect, and the experimental effects are considered between- subjects effects (Potvin et al., 1990; von Ende, 1993). In this experiment, there were two classifications of repeated measures: CO2 level (ambient and elevated), and month (August and September). Accordingly, we performed two separate repeated measures analyses: one to test the response of the physiological variables to the level of CO2 measurement, and the other to test the response of the physiological variables across the two months. The repeated measures analysis produces two kinds of output: a "between-subjects" analysis, and a "within-subjects" analysis. The between-subjects analysis tests the significance of each main effect and interaction in the overall experimental summed over the repeated measure. The first line of the within-subjects analysis shows the significance of the repeated measure, summed over all of the main effects. The following lines of the within- subjects analysis tests the significance of the interaction of the repeated measure and each main effect of the experimental model. All four physiological variables were initially tested together in a multivariate repeated measures analysis, then were considered 45 separately for their contribution to the multivariate results. RESULTS The CO2 treatment responses showed significant variation for each of the physiological traits, more so for the elevated CO2 measurement level than the ambient CO2 level, and more in September than in August (Table 3). CO2 Treatment Effects The ambient and elevated CO2 chambers were compared using a multivariate analysis of variance of the four variables assimilation, transpiration, stomatal conductance and intercellular CO2 concentration. Water use efficiency was not included in the overall multivariate analysis because it is a linear combination of assimilation and transpiration and would reduce the power of the multivariate analysis. The multivariate analysis revealed that all of the main effects were highly significant (Table 4. A). In addition, all of the two- way interaction terms involving CO2 as a treatment effect were also highly significant. Each of the four physiological variables used in the multivariate analysis was also tested using a univariate mixed model analysis of variance (Table 4. B); water use efficiency and specific leaf area were analyzed univariately as well. Assimilation rate and intercellular CO2 concentration contributed the most to the overall significance of the CO2 treatment effect in the multivariate analysis. Assimilation rates were higher for the elevated grown plants measured in elevated CO2 than for the ambient grown plants measured in ambient CO2 for both months. Intercellular CO2 was higher in plants at the elevated CO2 measurement level for all treatments. Water use efliciency and specific leaf area respond significantly to CO2 treatment. Water use efficiency was higher when ............... .698 8.2: .698 2.62 .698 8.82 82 080 208% .698 8.42 .698 6.86 .698 699% .69: 69mg .698 ~98~ .298 8.63 .00 888.35 023 03 do .60 .8 80 .698 93 ..6o .8 3.0 .698 2 .2 .698 ~9_ .698 E .2 a 3 >056 .698 86 .628 62 .698 09m .698 8.2 .698 _9_ .698 w: 9.03 .698 ~96 .698 2 6 .698 8.6 .6~.8 6.0 .6~ .8 8.6 .698 86 88885 .668 3.8 .698 8.2 96.8 8.8 .698 8.2 .698 8.: .698 092 8688...... o «:0 ~00 00096—0 ~00 00:30 0 8 ~00 00:30 0 ~00 80580 ~00 80380 ~00 80380 3 . > 8 0050008 .0020 0080008 A 00 8 00.8308 . 00 8 0050008 .0000 8 0080008 .~00 8 0080008 .~00 0000080000: 8 8205 000300 8 850.5 80580 8 8530 080080008: 8 8580 002320 8 830.5 80380 8 8580 80080800 .m .m 0300. ............... .698 2.02 .698 4902 .698 8.2~ 8.... 0.3 0888 .698 996mm .698 ~9~mm .698 8 .m~m .62 .8 2986 .698 2968 .66.: 2 .2 m~ .00 5885.: 00830: no .698 ~92 .698 9: .698 $0 .60 .8 2 .2 .698 9: .698 m3 3. 3 hoaofiowmxm ..698 2% .698 «.3 .698 Sm .698 m: .698 8.2 .698 6: a: .2 (A: .8 2.6 .698 696 .5 .8 8.6 .62 .8 86 .62 .8 296 .62 .8 696 80.835 .698 6.2 .698 9...: .698 2 .8 .298 6.2 .698 2.2 .698 692 8808...... 609%?» u .00 80.8 u .00 8.05 o .00 838. .00 .838. .00 .838. 033...; 8 00 008 . 0:0 8 0080008 . 00 8 0080008 . 00 8 0080008 .0000 8 00.82.08 .~00 8 0080008 .~00 000008.308 8 8330 000320 8 830.5 80380 8 8580 008008328 8 8580 000300 8 8380 80380 8 8580 lemma... 4 .m 2.3 A8008 0 80230800 00538 000 80000200 8020.809 00>“: 30.0 V 0 05 “a 800E008 on 8 008002800 303 0008000th ~00 002300 3 8008 800500 mooaouofimc 80:02 00080000 . 0 80380 “a 8008 800.509 000880.00 .8032 038033 .5808 0800 05 0880 $50100 0800 0.0 “0 0080008 0380.3 0000 200 0008080§0 839006 80000 08 80304 00.9072 @3080 0.2838: 080: 00000800 803 0508 80800.0 ~00 03283 0000 000 00 - 00 N 2 0080008 03 00080800 000 “00034. 000 00206» 3808080008 00 008300 080 80380 ~00 0808008 ~00 200 60.00 8002 .0 030.0. 47 3.880 68.90 880 .8 6. 886.0 .. .082 890 880 :80 .8 6. 6.88.0 1.50 0690 880 0:90 26 0. .082 183 280 6090 880 06 0. .082 .. 8.. 8.00 820 280 26 0. .260 .. 80 33.880 82 .m 880 .8 2 80.08... .. .00 33.880 28.. 880 26 0. .082 .. .00 5:880 82 .2 ~80 26 0 .250 ._. .00 .880 .896 6290 26 0. 8.. ._. .00 ._.._.._.880 8.6.. .02 0 .8 2 2888.8 600.388 8...... .3880 0. 8.2 880 26 0. 6.880 .082 33.880 88.0.8 869:. 26 0 0.5.0 0.3 8.820.300 .00 5.880 83.0 880 26 0. 08.0 80.080 .1880 0898 R2 0 26 0 0.000 8.8.5 .00 0 A 0 0:03 0 0000 880000 0.030 00 0000880000 00 00080802 0080m 0...... 3.0.28... 0.30 .< 0 .3... .0000 0000 030000 000 0088030 800080 0.0. 8 000.000 808 0 000 00.5 058 808088008 N00 .0000 008 030000 000 0088050 000 0803 0000808808 020 10502090 08 8 002. 003033 0008300830 800 08 00 3000 000 00003300 80280 00008 0088 88038. 00 ©0886 .m 800 8008800000 N00 88000088 000 6008000000 080808 80380808 80800808.. 003083 0008000830 800 000 «5073.2 0800 .0. 800 00030888 003-025 80.0. 000 80308000 83005 08000 00800 080 .3808 .0050 80008000 .8080008 N00 («0 800.00 808 08 000 008080.» 00 08.008. 808.3082 .0. 030,—. 48 83¢ 83% ENE “30:; $0.83 $85 2 aomvafl*£§2 ..... 3:3 $23 N32 33o $86 2 Seaman * ~23 ..... $83 33¢ £on $qu $33 H £82 * 354 xv; 853 $23 33.0 323 262 H 582 * mom ..... 3:2 833 335 $26 293 fi 354 * mom ”23 “"285 tax; 893 “.285 $386 2 Reagan * Now on; .o :28: mm: .o *386 223 1585 # €82 * Nco ..... **Looo.o its: "5de $de $886 fi 333 * Noo $36 82 .o 893 335 2:3 macho fl mom * Nov 235 ammo 82 .o 336 886 2:2 2 23:55: 233333 55m “3886 $183 its: $2 .o «22 $886 H 3:36 £82 ..... its: its: $183 :08: its: fl 22,2: :53 “anuaaz Nom $de $85 32 .o 333 38d 825 fi 3?: session $885 $183 1183 Sag 32 .o is: M 38 “88329 Noo <6 m5? No 6 m < by 85% .Sm 5 E56 2: E “.832: 8: 33 3%: 9:25:58:— NOO .m<>OZ< E58 3me 33.33:: 88m mowananoamo Exam .m .v 0359 49 measured in elevated CO2 for all CO2 treatments. Specific leaf area was greater for the ambient CO2 grown plants than for either the elevated CO2 grown plants or the unchambered control plants. At least one two-way interaction, including the only interaction that did not involve a repeated measure, CO2 treatment x family nested within population, was significant in the univariate analyses for assimilation, transpiration and intercellular CO2 concentration. This indicates that the families responded difl‘erentially to the CO2 treatment. Chamber Effects The open-topped chambers were compared with the unchambered control sites in order to determine the extent of a chamber efl‘ect. Plants in the ambient chambers did not differ from the plants in the unchambered control sites with respect to assimilation and stomatal conductance in August and September, at either CO2 measurement level (Table 3). At the elevated CO2 measurement level in August, plants in the ambient CO2 chambers had lower transpiration and higher water use efliciency than the unchambered control plants. Specific leaf area was higher for the ambient CO2 chamber plants than for the unchambered controls in August and September. Contrary to expectation, the elevated CO2 chambers did not differ from the unchambered control sites in the same way as the ambient CO2 chambers did. There were no significant physiological differences between the plants measured in the elevated CO2 chambers and the unchambered control sites in August, except for intercellular CO2 (higher for the elevated CO2 chamber) (Table 3). In September, plants in the unchambered control sites had greater assimilation rates, transpiration rates, and stomatal conductance at both CO2 measurement levels than did plants in the elevated CO2 chambers :1 50 (Table 3), but water use efiiciency and specific leaf area did not difler for the unchambered control plants and the plants in the elevated CO2 chambers. Repeated measures analysis of CO2 measurement levels The main efi‘ects and interactions involving the physiological measurements at each CO2 level were further examined using a repeated measures analysis of variance, which takes into account the correlation between repeated measurements on the same plant. The repeated measures analysis was used to indicate the presence of an interaction of the two CO2 levels with the treatment efl‘ect, to determine whether the variables responded difi‘erently at the two C02 measurement levels, and to determine whether there was an overall difference between the two CO2 measurement levels, pooled over all treatment effects. Effects of CO2 measurement level in August In August, the multivariate tests were significant for both the CO2 treatment elfect and the family nested within population effect (Table 5. A). The level of measurement had a significant effect on the physiological variables. For each of the variables tested individually, the interaction of CO2 measurement level and CO2 treatment was significant. Thus, the direction of response for each of the variables for the two CO2 growth environments differed between the ambient and elevated levels of measurement. Transpiration rate was similar for the two CO2 treatments at the ambient CO2 measurement level At the elevated C02 measurement level, however, the transpiration rate for the ambient CO2 grown plants was lower, while that of the elevated CO2 grown plants was higher (Figure 3). Stomatal conductance was similar for the elevated CO2 grown plants at both CO2 measurement levels in August; stomatal conductance for the 51 mz mz mz mz 2 mz m2 .2 28228230233 mz mz mz m2 2 mz 222 .2. 22.2....82052 mz mz mz mz 2 .. m2 .2 282282233 mz mz mz mz 2 mz 222 .2. 20.2.2032 ”33.. ”.33. .33» «x fl $3. on .v NOOLQVQH at; 5.3 m2 ”33.. fi ”223. omfi av mz mz mz mz 2 mz m2 .2 eoaaflmoo m2 m2 mz mz 2 mz 222 .2. achraoo mz mz a. .. 2 .3... m2 .2 22292282 m2 m2 mz mz 2 mz 222 .2. 28.2 .2. mz mz mz 2 .. 222 .v N00 20 any H 4 :3 22825 420%? "magma 88 Saw 5950“— 2m=w=< 4 .m 033. .2822 v 22 .2... M222.22 v 2 .3. 22.22 v 22 ._. m2.22 v 2 2... ”@2302 ooaomfla sages 02222.5, Rflmflofibfi some 8.2 22222 warm—28 0353:5222 22296 222w 222 62m .2288 22023 8.2 bun—22233 22803280 33 2322222 02:. .9258 222. 222 296A 22 woflpficu 2262 paoaogmmoa NOD 3:320 228 22203822 2: :2 25222238 83222223 How 23223 $2388 2.882232 .m 033. 52 mz mz mz mz 2 2... 22 .2 220.222.2222....8233 mz mz mz mz 2 mz 222 .2. 2o2*.oo*2o>o2 mz mz mz mz 2 .. 22 .2 22222228222233 mz mz 2... mz 2 mz 222 .2. 222222082 2.32322 m2 .22 “.3232. ~ 22332 m: “V. «0022—0254 2323.. 2:3. «is 23232. ~ “33.2 w: «v 3264 a 2222 a . 22m 222222 m2 mz m2 m2 2 mz 222 .2 222c22aan2*.oo mz mz m2 m2 2 mz 222 .2. 2222.60 mz mz m2 m2 2 .. 22 .2 2222228222 2... mz mz mz 2 mz 222 .2. 2222 .23; m2 m2 .2332 M 2233. w: 2? «CD 20 ”G H < .22 E2295 2%“? 288...:— 38 Saw 5950: .2onaoaaom .m .m 0322.2. 53 Figure 3. Comparison of overall means (i s.e.) for assimilation, transpiration, stomatal conductance, and intercellular CO2 concentration for August and September. In each month, plants grown in each CO2 treatment were measured at both CO2 levels, ambient CO2 and elevated C02. For intercellular CO2 concentration error bars may be obscured by symbol. N 00 0| 0 Assimilation (umol m'zs'1) N o 10- l 4 3? \i A 01 1 1 l 1 l J R Q August September 54 Transpiration (mmol rn'2 s‘1 ) 2. E F” 0: August September _E-— Ambient CO 2 grown, Ambient Level + Ambient CO 2 grown, Elevated Level —e— Elevated CO 2 grown, Ambient Level + Elevated CO 2 grown. Elevated Level T7? Stomatal Conductance (mol m'2 s'1 ) P o: / Ki; ‘- : IF August September 600 9"”. C a 500 a 3.. ON 0 ; 400 E E 2% *5 300 57 ’ Q 200 August September Figure 3 55 ambient CO2 grown plants was higher than that of the elevated CO2 grown plants measured in ambient C02, but lower than the elevated CO2 grown plants measured in elevated CO2 (Figure 3). The between-subjects test for the overall CO2 treatment effect was only statistically significant for intercellular C02; the intercellular C02 concentration for the elevated CO2 grown plants was higher than that for the ambient CO2 grown plants. CO2 measurement level was a significant effect for each of the physiological variables (Table 5. A); the response of each variable in ambient CO2 differed from the response of each variable in elevated C02. For instance, both assimilation rate and intercellular CO2 concentration were higher at the elevated CO2 measurement level for both ambient and elevated CO2 grown plants. In all cases, the pattern of response for water use efficiency was similar to that for assimilation rate. Assimilation rate and water use efficiency were highly correlated (overall r2 = 0.75; p < .0001), while transpiration and water use efficiency were not as strongly correlated (r2 = -0.21; p < .0001). Effects of CO2 measurement level in September In September, both the CO2 treatment and family nested within population effects were significantly different in direction and magnitude in the multivariate analysis (Table 5. B). When the variables were examined separately, the assimilation and intercellular CO2 concentration results were similar. Both assimilation and intercellular C02 concentration were higher at the elevated CO2 measurement level than in ambient C02, but the plants grown in elevated CO2 had lower assimilation rates and higher intercellular CO2 concentrations in elevated C02 than did the ambient CO2 grown plants (Figure 3). For transpiration, there was a treatment x CO2 level interaction; the elevated CO2 grown plants maintained the same transpiration rate at both ambient and elevated CO2 measurement 56 levels, but the ambient grown plants transpired less in elevated CO2 than they did in ambient C02. The results of the individual tests for stomatal conductance were nonsignificant, except for the effect of CO2 level of measurement; stomatal conductance was lower for both ambient and elevated CO2 grown plants in elevated CO2 than in ambient C02. Repeated measures analysis for August and September A separate repeated measures analysis was conducted for the physiological measurements made during August and September, to take into account the correlation between measurements made at two separate times on the same plant. The repeated measures analysis was used to indicate the presence of an interaction of the two measurement times with the treatment effect, to determine whether the variable responded differently during August and September, and to determine whether there was an overall difference between the two months, pooled over all treatment effects. Differences between August and September at the ambient CO2 level When the months were compared at the ambient CO2 measurement level, the multivariate tests were significant for the CO2 treatment, family nested within population, and treatment x family (population) interaction effects (Table 6. A). Only assimilation rate showed significant results for both the direction and magnitude of response across the months when the variables were examined individually. The assimilation rate in September tended to be lower for the elevated CO2 grown plants in ambient CO2 than the ambient CO2 grown plants (Figure 3). Likewise, water use efliciency was lower for the elevated grown plants in September at the ambient CO2 measurement level than in August (Table 3; Figure 3). For transpiration and intercellular CO2 concentration, the 57 mz mz mz mz 2 .2. 2222 .2 2222222822....822282 mz mz mz mz 2 mz 222 .2. 220.2...oo*22282.,2 mz 222 mz mz 2 t. 222 2.2 22222282222252 mz .222 mz m2 2 mz 222 .2. 22222222282 mz mz mz .. 2 .. 22 .2. E 60.222322 2332 m2 m2 22 2 233" mfifi .2. mz mz mz mz 2 .3. 2222 .2 22282822....8 mz mz mz m2 2 .272 22 .2. 2292....8 .. mz mz .. 2 .3. 2222 .2 28.228222 2... mz 222 .. 2 mz 222 .2. .222 mz .. mz 2... 2 .. 22 .2. :2 .8 2U and mm d. Eu 212.206 :wnm—Dz Quechua avoid—pm £0050” .2962 22222822223222 NOU 2220222222. .2.. .o 22222.2. Aood V 22 “22232. m:26 V 22 .23.. Wed V 22 .2. 226 V 22 2* 22962 82222022222w2m 22822222228 022222.222; 22202wo2o2m>2222 220220 20.2 220222 2.2.2228 3832222222 22222026 2222 222 6222 .2962 22202220238222 N00 22023 2222 .22322222228 2208222228 92.5 29222222 02:. 2222228223 22222 mswfix 2 22322220222 322222.222; 2222 2.0522222 32.258222 2282202232 .0 0322.2. 58 mz ... mz mz 2 .. $2.2 28.228.239.228: mz mz mz mz 2 m2 £2 .2. 2222....822252 mz mz 2... mz 2 .. $2 .2 Eomvafigqoz mz ,. mz mz 2 m2 8.2 .2. 2282282 mz ’83, m2 ”.5. M ”33% omfi 2v PC. NOUtfihoE a" 2 m2 ”.33. fl «“3. ©N~ .v 5:02 822223 . 5:22: mz mz mz m2 2 m2 m2 .2 aomvafimoo mz mz .022 m2 2 m2 82 .2. h222....8 2... mz mz mz 2 .2. 82.2 222282 mz mz mz mz 2 m2 82 .2. 2222 .33.. m2 .2. 23.3.. ~ 233. ©N~ .v Fr «00 2U .G H < 52 22330 4mm? "38.5— 38 Sam :83qu— 2962 228832588 N00 @8565 .m .c 2an 59 performance of both ambient and elevated €02 grown plants measured in ambient CO2 was virtually identical in August and September. Stomatal conductance, however, was higher for the ambient C02 grown plants than the elevated CO2 grown plants in both August and September (Table 6.A). Differences between August and September at the elevated C02 level At the elevated CO2 measurement level, the overall multivariate tests were signifith for the CO2 treatment and the family nested within population eifects (Table 6. B); in addition, there was a significant CO2 treatment x family (population) interaction in the direction of response in the two months. When examined univariately, both assimilation and stomatal conductance responded across the months in a nonparallel direction. Stomatal conductance was similar for the ambient CO2 grown plants in both August and September, but it declined for the elevated CO2 grown plants in September compared to August. The ambient CO2 grown plants had higher assimilation rates, lower transpiration rates and lower intercellular CO2 concentrations at the elevated CO2 measurement level in both August and September than the elevated CO2 grown plants (Table 3; Figure 3). Assimilation rates and stomatal conductance tended to be lower in September than in August; intercellular CO2 concentration was higher in September than in August at the elevated CO2 measurement level (Figure 3). Population and Family Interactions with C02 Treatment The multivariate analysis revealed a significant CO2 treatment by population interaction for the physiological variables (Table 6. A), but the same interaction term was not significant for any of the physiological variables when examined individually (Table 6. B). Thus, overall the populations responded differently to CO2 treatment. The repeated 60 measures analysis indicated significant month by population interactions for assimilation rate at the ambient CO2 measurement level (Table 6. A), and for stomatal conductance at the elevated CO2 measurement level (Table 6. B). Ambient CO2 grown plants in P0pulation EL had lower stomatal conductance in September than in August (0.94 August v. 0.88 September), while ambient CO2 grown plants from Population KF had higher stomatal conductance in September than in August (0.82 August v. 0.93 September). The individual families showed significant interaction with CO2 treatment. This was revealed in the multivariate analysis (Table 6. A), and in the univariate mixed model analyses for assimilation, transpiration, intercellular CO2 concentration, and water use efficiency (Table 6. B). Figure 3 shows the individual family means for assimilation rate for both CO2 treatments at both CO2 measurement levels, in both months. Some families responded similarly to CO2 measurement level, whether they were grown in ambient CO2 or elevated CO2 (for example, families EL 7, EL 9, KF 33 and KP 44 in August). Some families showed very different assimilation patterns in September compared to August (e.g., families EL 5, EL 9, KF 33, KF 37, KF 39, KF 43, and KF 44). Therefore, the pattern of response of the families for assimilation and other physiological traits was variable. The repeated measures analysis revealed a significant interaction between month, CO2 treatment and family for stomatal conductance at the elevated CO2 measurement level (Table 6. B, Figure 4). As with assimilation, the types of response to CO2 treatment with respect to stomatal conductance were extremely variable. Regardless of CO2 grth environment, some families showed increased stomatal conductance in September compared to August, some families had decreased stomatal conductance in September, 61 Figure 4. Comparison of individual family means for stomatal conductance in August and September. (top) Plants grown in ambient C02, measured in elevated C02. (bottom) Plants grown in elevated C02, measured in elevated C02. 62 Ambient CO2 Treatment Elevated Measurement Level 0.50 Stomatal Conductance (mol m ‘2 s'1 ) 0.00 August September Elevated C02 Treatment Elevated Measurement Level 2.00 .A 01 O .a. o O | 0.50 Stomatal Conductance (mol m '2 s'1 ) 0.00 August September Figure 4 HMMMMMMHH EL 2 EL 4 EL 5 EL 7 EL 8 EL 9 EL 10 EL 15 KF 31 KF 32 KF 33 KF 35 KF 37 KF 39 KF 41 KF 42 KF 43 KF 44 63 and still other families showed little change in stomatal conductance in September compared to August. This variation contributed to the nonparallel responses between August and September when compared by C02 treatment (Figure 3). DISCUSSION Overall, plants grown and measured in elevated CO2 were significantly difi‘erent physiologically from plants grown and measured in ambient C02. Plants grown and measured in elevated CO2 had higher assimilation rates, higher water use efliciency, and lower stomatal conductance than plants grown and measured in ambient C02. Transpiration did not differ greatly between the ambient CO2 and elevated CO2 grown plants, so the increase in water use efficiency was directly related to the increase in assimilation rates in the elevated CO2 grown plants. These results differed from the 1991 experiment, in which ambient CO2 assimilation rates were more similar to elevated CO2 assimilation rates, and where transpiration rates were lower for the elevated CO2 grown plants (see Appendix). The failure of transpiration to decrease in 1992 as GS declined may have been due to increased leaf temperatures of vapor pressure deficits in the elevated C02 plants, neither of which variables was measured in this experiment. The differences between the plants in the two grth environments for assimilation, water use efficiency and stomatal conductance were not as marked in September as they were in August. Ambient CO2 grown plants maintained the same physiological potential to respond to elevated CO2 in September as in August, but assimilation and water use efficiency declined for the elevated CO2 grown plants in September. Even though the physiological performance of the elevated CO2 grown plants 64 declined in September, carbon fixation rates of the plants grown and measured in elevated CO2 remained higher than carbon fixation rates of the plants grown and measured in ambient C02. When compared to the unchambered control sites, some of the differences in the performance of plants in the ambient and elevated CO2 chambers could be attributed to the presence of the chambers, while other differences seemed to be due to the action of elevated C02. The plants in both types of chambers (ambient CO2 and elevated C02) had lower transpiration rates than the unchambered controls. This was probably due to reduced wind and increased relative humidity inside the chambers, a common occurrence in open topped chambers (Leadley and Drake, 1993). The leaves of plants inside the ambient CO2 chambers had greater specific leaf area than the unchambered controls, but this was not the case for the plants inside the elevated CO2 chambers. Since a reduction of specific leaf area is a well-known effect of increased CO2 (Bazzaz, 1990), the increase in specific leaf area in the ambient chambers was probably a chamber effect which was offset in the elevated CO2 chambers by the elevated C02. Overall, the physiological results suggest that negative acclimation to elevated CO2 was occurring in the elevated CO2 grown plants. While assimilation rates were similar for plants measured in both CO2 environments in August, the ambient CO2 grown plants had lower transpiration and intercellular CO2 concentration at the elevated CO2 measurement level than the elevated CO2 grown plants measured at the same level. The diiference in performance of the elevated CO2 grown plants compared to the ambient CO2 grown plants was even greater later in the experiment. In September, the plants grown in elevated CO2 had lower assimilation rates, lower stomatal conductance, and higher intercellular CO2 65 concentrations than the ambient CO2 grown plants at each measurement level. Both ambient CO2 and elevated CO2 grown plants declined in stomatal conductance and specific leaf area in September compared to August, occurrences common as the growing season progresses (Morison, 1987 ; den Hartog et al., 1993). The decline in assimilation rate between August and September only occurred in the elevated CO2 chambers, however, which indicates that this response was due more to long-term exposure to elevated CO2 than to the progression of the growmg season. The decline in photosynthetic rates for the elevated CO2 grown plants compared to the ambient CO2 grown plants in September occurred at both CO2 measurement levels, at equivalent intercellular CO2 concentrations (Figure 5). Data from other investigations (von Caemmerer and Farquhar, 1981; DeLucia et al., 1985; Sage et al., 1989) have revealed that declines in photosynthetic rates prior to the CO2 saturation point are generally due to a reduction either in the amount or in the activation state of Rubisco. Measurements made at ambient CO2 corresponded to this pre— saturation level. In ambient C02, the elevated CO2 grown plants measured in September were showing signs of Rubisco limitation (Figure 5). Measurements made in elevated CO2 corresponded to saturating intercellular CO2 concentrations. Again, the elevated CO2 grown plants measured in September also showed the greatest signs of RuBP regeneration limitation of assimilation rates (Figure 5). Nitrogen concentrations in ambient and elevated CO2 were measured for a subset of three families from the experiment. Per cent nitrogen was the same in the aboveground tissues of the elevated CO2 grown plants compared to the ambient CO2 grown plants (Klus, Ch. 1). Therefore, some of the nitrogen in the elevated CO2 grown plants may have been reallocated from photosynthetic machinery to other 66 Figure 5. Assimilation versus intercellular leaf CO2 concentration at both CO2 measurement levels. At the ambient CO2 measurement level, intercellular CO2 concentration was close to 200 ,ubar. At the elevated CO2 measurement level, intercellular CO2 concentration was close to 500 ubar. Vertical bars indicate 3: one s.e. for assimilation means. Horizontal bars indicate i: one s.e. for intercellular CO2 means. 00 O 4_l L 11 N 01 1 g I l l N O l l_ L l J —-\ 01 r 4 r r l Assimilation (umol rn'2 s'1 ) 3 i l 67 0 . ,. 100 l l I I l l T , , . . . 200 300 400 500 600 Intercellular CO 2 (ubar) I f r T I I I T T I I + Ambient CO 2 grown, August —E— Ambient CO 2 grown, September —9— Elevated CO 2 grown, September —.— Elevated CO 2 grown, August Figure 5 ,4 68 molecules in the aboveground tissues. Ecological and Evolutionary Implications It is important to note that while the photosynthetic performance of the elevated CO2 grown plants had declined from its maximum potential compared to the ambient CO2 grown plants, the elevated CO2 grown plants still assimilated at significantly higher rates in their own growth environment than the ambient CO2 grown plants in ambient C02. These results are similar to those of Hollinger (1987 ) and Yelle et al. (1989) in which plants grown in elevated CO2 maintained higher assimilation rates in their own environment. It is possible that as the growing season progressed, the plants in the elevated CO2 environment were integrating their physiological response to elevated CO2 by reallocating some nitrogen away from photosynthetic processes, possibly to other molecules in leaves or roots to increase their efliciency of nitrogen use and their capacity for nutrient uptake (Klus, Ch. 1). The significant CO2 by population interaction term in the overall multivariate analysis of the physiological variables (Table 6) indicated that the two populations responded difl‘erently to the CO2 treatment. A key variable contributing to this interaction was water use efliciency (Table 6. B). Water use efficiency was higher for Population EL than Population KF in ambient CO2 during both months (data not shown), but water use efficiency for Population KF increased to match that of Population EL in the elevated CO2 treatment. This change in population response in the two CO2 environments appeared to stem from the fact that, in several instances, assimilation rates were higher for Population EL than Population KF in ambient C02, but were similar for the two populations in elevated C02. The two populations may have been interpreting the ambient CO2 69 environment in slightly different ways due to their habitats of origin. Population EL grows in sandy soil near a lake, and experiences high irradiance and periodic drought stress. Population KF grows in the moister, shadier habitat of a mown field. Depending upon the nature of local changes in rainfall and air temperature that may accompany increasing atmospheric CO2 in the next 100 years, this difference in water use efficiency among the two populations in the elevated CO2 treatment may be critical in determining the success of these populations in a novel environment. Moreover, the intraspecific variation in instantaneous response to elevated CO2 and the variation in the capacity to maintain a positive response to elevated CO2 discovered in this experiment may mean that some families will be placed at an advantage _ or disadvantage relative to other families in an elevated CO2 world. Families that are able to maintain lower stomatal conductance and higher water use efficiency may be better able to survive in a world of increased drought. Some families may be better than others at allocating resources so as to maintain high assimilation rates while also maximizing the use and acquisition of nitrogen. Under such circumstances, families showing such variation in response to elevated CO2 may very well undergo changes in competitive rankings with respect to other members of their plant community. Depending upon the scale of interaction, this may result in changes in intraspecific and interspecific community dynamics. Significant CO2 x family interactions in physiological responses at both CO2 measurement levels and for both months suggests the existence of intraspecific genetic variation in response to elevated C02. It was not possible to predict an individual family's performance in elevated CO2 from its performance in ambient C02; therefore, families with 70 higher assimilation rates, stomatal conductance, or water use efliciency at one CO2 measurement level in August did not necessarily have equivalent rates at the other measurement level or during September. While the positive effects of CO2 declined for some families over time, other families maintained a positive physiological response to CO2 through the end of September. prhysiological variation somehow afl‘ects general plant performance (fitness) in the long term, then this physiological variation may result in variation in fitness up on which natural selection may act. The degree of variation in physiological response to elevated CO2 at the family level in this experiment were similar to that of biomass accumulation and allocation previously described (Klus, Ch. 1). Yet the patterns of physiological response and biomass allocation for each family in elevated CO2 did not correspond. Families grown in elevated CO2 that expressed higher assimilation and water use efliciency, or higher specific leaf area, were no bigger on average than families with lower rates of assimilation, lower water use efliciency and lower specific leaf area. It may be the case that exposure to elevated CO2 initiates responses in plants that have contradictory growth outcomes. For example, in current ambient CO2 conditions, high relative grth rate is often associated with high specific leaf area, but not necessarily with high photosynthetic rates (Lambers, 1987; Shipley, 1995). In an elevated CO2 environment, plants generally increase photosynthetic rates, but other changes in morphology may prevent higher assimilation rates from being translated into higher growth rates. Ifexcess carbohydrates formed by photosynthesis accumulate in the leaves, changes in leaf architecture may occur which might increase the diffusive resistance to CO2 inside the leaf and increase dry matter content on an area basis (decrease specific leaf 71 area) (Dijkstra and Lambers, 1989). Although starch did not accumulate in the leaves of the plants in this experiment, there was a small increase in % carbon in aboveground 1 tissues (Klus, Ch. 1), a decrease in specific leaf area, a decrease in stomatal conductance (Table 5), and a decrease in assimilation at equivalent intercellular CO2 concentrations in the elevated CO2 grown plants (Figure 5). An increase in the number of mesophyll cells devoted to photosynthesis (Shipley, 1995) or an increase in cell wall material, and an increase in diffusive resistance to internal CO2 in the leaves of the plants grown in elevated CO2 (Dijkstra and Lambers, 1989) might explain why the elevated grown plants showed some negative photosynthetic acclimation and why earlier increases in plant size were not maintained throughout the growing season. The mechanisms by which plants translate their physiological response to elevated to the level of the whole plant remain little explored. The chain of events in response to elevated CO2 by which plants integrate physiology with growth and biomass allocation must be more clearly elucidated before we will be able to predict the potential ecological and evolutionary impact of intraspecific variation in plant response to elevated atmospheric C02. CHAPTER THREE GROWTH RESPONSES TO ELEVATED CARBON DIOXIDE IN PLAN TAGO LAN CEOLATA 72 73 INTRODUCTION Increased atmospheric CO2 has direct effects on plant life (Strain and Cure, 1985). Physiological traits are altered in plants exposed to elevated C02, and grth often increases (reviewed in Bazzaz, 1990). Kimball (1983) and others (Cure and Acock, 1986) have predicted that increased growth in response to elevated CO2 in the next one hundred years may result in as much as a 30% greater yield in agricultural crops. Depending upon experimental conditions, though, plant grth responses to elevated CO2 have been found to be complicated by interactions with other environmental conditions (Bazzaz, 1990). While the positive effects of elevated CO2 on plant growth are most often seen when other nutrients are not limiting (Bazzaz, 1990), some studies have determined that CO2 may have positive effects on grth even in nutrient-poor conditions (Norby et al., 1986; Bowes, 1993). Often elevated CO2 positively afl‘ects early growth rates (Wullf and Alexander, 1985), but those efi‘ects may decline over time (Wulfl‘ and Strain, 1982). Yet, despite the decline in the effect of elevated CO2 on grth rate over time, early stimulation of growth may be suflicient to result in greater biomass in elevated CO2 grown plants compared to ambient CO2 grown plants (Wulfl and Strain, 1982; Poorter et al. , 1988) Important in determining the magnitude of growth response to elevated CO2 is a plant's physiological response to CO2 enrichment. It is generally agreed upon that 74 physiological characteristics influence grovvth characteristics, but the connection between physiology and growth is not clear (McGraw and Wulfi; 1983; Fichtner, 1994). The manufacture of plant tissue depends upon the supply of organic energy storage molecules and construction materials to the tissues manufactured by photosynthetic and biosynthetic pathways. Yet in agriculture, where plant growth and yield are critical issues, plants artificially selected for increased growth and yield often do not have increased photosynthetic rates (Evans, 1975). Accordingly, selection for increased photosynthetic rates does not always result in either increased growth or yield (Evans, 1975; McGraw and Wulfl‘, 1983). Photosynthesis and growth may be integrated through other plant traits such as specific leaf area (Poorter, 1993). Complicating attempts to understand the connection between photosynthesis and growth in response to elevated C02 is the fact that elevated CO2 may afiect individual plant traits in ways that may have contradictory efiects on growth. For example, while elevated CO2 is often associated with higher assimilation rates and greater biomass accumulation, elevated CO2 also generally results in decreased specific leaf area, a condition associated with low grth rates (Konings, 1989; Gamier, 1992). Elevated CO2 also afi‘ects other physiological processes such as stomatal conductance and transpiration which control water balance in the plant and interact with assimilation rates to determine overall water use efficiency (reviewed in Bazzaz, 1990). Therefore, the effect of elevated CO2 on growth is mediated through assimilation rates and the interaction of assimilation rates with stomatal characteristics, water use, and biomass allocation. It has been hypothesized that growth potential may be determined more by source- sink relations and the way a plant allocates photosynthate within and among its various J— 75 organs to utilize and acquire other nutrients efliciently than by intrinsic photosynthetic capacity (Evans, 1975; Fichtner, 1994). Grth responses to elevated CO2 are enhanced in plants which possess an adequate sink for the carbohydrate manufactured by photosynthesis (Poorter, 1993), for example, roots. Arp (1991) reviewed studies in which growth responses to elevated CO2 declined over time, and determined that a lack of growth response was most pronounced when the plants were limited in the amount of biomass they could allocate to belowground tissues (usually because of small pot sizes). Plants that manufacture other sinks for carbohydrate, such as fi'uits or woody tissue, often show a greater growth response to elevated CO2 than plants which do not possess such sinks (Poorter, 1993). Being able to allocate carbohydrate to sinks may help plants maintain an internal balance between the supply of carbon and the supply of nutrients, by moving the carbohydrate into storage organs, by using the carbohydrate to build more tissues to acquire nutrients, or by reallocating nutrients within the plants tissues to increase their effective use (Arp, 1991; Poorter, 1993). Therefore, plants which are flexible in their ability to allocate organic resources among various types of tissue, for example, between aboveground and belowground tissues, seem to have a greater capacity for successfirlly translating physiological potential into grth than plants for which the relationship between physiology and grth is more canalized (Poorter, 1993). Interspecific variation exists in the extent to which plants can reallocate tissues or nutrients in response to changes in their abiotic environment. For example, Plantago lanceolata can change allocation to roots and shoots in response to changes in nitrogen availability, but Plantago major does not seem to be as flexrble (Kuiper and Bos, 1992). Variation in grth responses to elevated CO2 that are 76 influenced by source-sink relationships have been documented for species possessing a variety of life histories (Poorter, 1993). Understanding the way elevated CO2 afl‘ects grth is critical in determining how plant life will respond to increasing levels of atmospheric CO2 on a large scale. Increasing atmospheric CO2 may have implications for plant-plant interactions (for example, competitive relationships) and evolutionary potential (Tilman, 1993; Geber and Dawson, 1993). Growth and biomass allocation patterns, not simple physiological potential, will ultimately determine whether plants will shift in competitive hierarchies within plants communities. Allocation of resources in response to elevated CO2 for survival and reproduction will afiect plant fitness, and therefore the capacity of plants to respond evolutionarily to elevated C02. Experiments conducted in 1992 explored the eflects of elevated CO2 on biomass allocation (Klus, Ch. 1), physiology (Klus, Ch. 2) and grth traits in two populations of Plantago lanceolata. Plantago lanceolata is a species especially well- suited for a study of the effects of elevated CO2 on grth traits. For herbaceous perennials such as P. lanceolata, size at the end of the first growing season is often an indicator of survival and fecundity in subsequent growing seasons (Primack, 1979; Solbrig, 1981). P. lanceolata also possesses an underground sink for carbon in the form of a corm, and is flexible in its allocation of biomass depending upon nutrient conditions (Kuiper and Bos, 1992). Another potential sink for carbohydrate are the vegetative side shoots which Plantago lanceolata produces under appropriate growth conditions (Teramura, 1983). To determine the response of early growth parameters to elevated CO2 on two populations of P. lanceolata, Iperformed censuses throughout one growing season for the aboveground 77 traits of leaf number, vegetative shoot number, and overall shoot diameter (the diameter of the base of the plant, including the side shoots, just above the soil surface). A subset of plants censused for growth traits was harvested midseason in order to determine biomass allocation patterns to above- and belowground tissues after exposure to elevated CO2 for approximately one-half of the growing season. The plants harvested midseason were also measured physiologically and in terms of final biomass allocation to match the treatment of the plants in the larger experiment. By examining all three aspects of response to elevated C02, I hoped to gain insight into how physiological traits and the allocation of resources to tissues interact in determining whole plant grth responses to an increased supply of carbon resources. MATERIALS AND METHODS The study populations of Plantago lanceolata were chosen to represent two distinct habitats. The Ely Lake (EL) population (Allegan County, Michigan) grows on exposed sand on a sunny lakeshore, experiencing high irradiance and periodic water stress. The Kellogg Field (KF) population (Kalamazoo County, Michigan) grows Kalamazoo loam in partial shade on the edge of a mown field. A total of 24 families, twelve randomly selected from each of the two populations, were used in the experiment. On June 5, 1992, all of the seeds from each family were divided into two equal groups, planted in separate flats, and placed in either ambient or twice-ambient (hereafter referred to as elevated) CO2 to germinate. Germination took place over 7 - 10 days. On June 17, six maternal siblings from each family in each CO2 environment were transplanted into separate 30-cm high pots made from 10-cm diameter 78 PVC pipe with mesh screen bottoms. The day of transplanting was considered Day 1 of the experiment. The pots were filled with a 50-50 mixture of relatively inorganic native field soil (Kalamazoo loam) and sand. 144 pots (6 pots x 12 families x 2 populations) were distributed randomly among four replicate outdoor open-top chambers in each CO2 environment. Both ambient and elevated CO2 chambers had one-meter square internal dimensions, contained 36 pots each, and were constructed following the protocol of Curtis and Teeri (1992). To determine the effect of the chambers themselves, seventy-two additional seedlings from each population were planted in individual pots and distributed randomly among four 1 m2 unchambered control sites. The chambers and unchambered control sites were arrayed in four randomized blocks in Bailey Field, an abandoned agricultural site at Kellogg Biological Station, Hickory Comers, Michigan. Each block contained one ambient CO2 chamber, one elevated CO2 chamber, and one unchambered control site. In preparation for the experiment, the site was cleared, herbicided, and disked to smooth out uneven patches. A smaller experiment (hereafter referred to as the growth experiment) was conducted in conjunction with the large- scale experiment (referred to as the main experiment) in order to follow early grth patterns and to determine biomass allocation and physiology in midseason without performing a destructive harvest within the main experiment. Seedlings from Populations EL and KP which had been germinated in ambient CO2 or elevated CO2 were transplanted into four-inch pots and placed on the ground between the larger main experiment pots. The grth experiment was set up in two ambient CO2 chambers, two elevated CO2 chambers, and two unchambered control sites, selected at random from those in the main experiment. Fifteen pots were placed in 79 each location, providing a sample size of 30 for each of the experimental treatments (ambient C02, elevated C02, and the unchambered controls). The experimental site at Bailey Field experienced full sun throughout the day. The plants were watered as needed, usually twice daily. There was no fertilizer supplementation. Pure C02, mixed with ambient air by ventilation fans, was supplied 24 hours per day to the elevated chambers. Ambient air was circulated within the ambient chambers by the same type of fan. CO2 levels were monitored continuously and levels were recorded on a computer at three-minute intervals (Curtis and Teeri, 1992). Mean daytime (0700-1900 hours) CO2 partial pressure inside the elevated chambers was 72 i 6 Pa (i s.d.), with the mean daytime CO2 partial pressure inside the ambient chambers being 36 i 3 Pa. Quantum sensors and shaded thermocouples attached to a LI-1000 datalogger (LICOR Inc., Lincoln NB, USA) recorded irradiance levels and temperature. Daytime temperatures were 1.7 :t 0.6 °C higher inside the chambers than in the unchambered control sites, with no significant difference in temperature between ambient and elevated chambers. Three weeks into the experiment the young plants were exhibiting symptoms of light stress (prostrate grth and red pigmentation of the leaf bases) and all chambers and control sites were covered with shade cloth. The shade cloth reduced ambient light by 68%, and the plants recovered their normal phenotype. Leaf number, the number of vegetative shoots (shoot number), and the basal diameter of the plant (shoot diameter) taken just above the soil surface, were measured during three censuses for the growth experiment (Day 13, Day 27, and Day 49) and during four censuses for the main experiment (Day 10, Day 59, Day 78, and Day 127). Physiological measurements were taken immediately prior to the final census for the plants 80 in the grth experiment (methodology for physiological measurement described in Chapter 2). The most recently fully expanded leaf of each plant in the grth experiment was measured physiologically, then removed from the plant and pressed. After drying fully, the area and mass of each leaf were measured and specific leaf area was calculated. The growth experiment plants were harvested on Day 49. The roots and shoots were separated at the soil surface, bagged, and dried at 60 °C. Aboveground biomass, belowground biomass, and root: shoot ratio were determined from the dried plant material. Plants in the main experiment were not harvested until the end of the growing season on Day 127. Statistical Analysis The census data for the main experiment and the growth experiment were analyzed using a repeated measures profile analysis, as recommended by von Ende (1993) and Potvin et al. (1990). The repeated measures analysis is analogous to a split-plot analysis in which the repeated measure is considered to be a within-subject effect, and the experimental effects are considered between-subjects efi‘ects (Potvin et al., 1990; von Ende, 1993). I performed two separate repeated measures analyses. The census data were analyzed for the repeated measure of time, and the physiological data gathered from the grth experiment were analyzed for the repeated measure of CO2 measurement level. The census data were first compared using a multivariate model; then each variable was analyzed individually in a separate repeated measures analysis to determine its influence on the multivariate analysis. In this experiment, time (i.e., the repeated census dates) was the repeated measure. Significance levels for the main effects in the 81 multivariate analysis were interpreted using Roy's Greatest Root, recommended because of its statistical power and the fact that it is applicable to post hoc statistical comparisons (Scheiner, 1993). Greenhouse-Geisser F statistics were used to determine the significance of the treatment effects for the individual variables. Greenhouse-Geisser F values are adjusted to account for inequalities of variance in the univariate analysis of repeated measures, and are considered to be conservative with respect to Type I errors (von Ende, 1993). The repeated measures analysis produces two kinds of output: a "between- subjects" analysis, and a "within-subjects" analysis. The between-subjects analysis tested the significance of each main effect and interaction in the overall experimental design summed over the repeated censuses. The first line of the within-subjects analysis showed the significance of time, summed over all of the main effects. The following lines of the within-subjects analysis showed the significance of the interaction of time and each main effect of the experimental model. For the main experiment, the main effects were Block, Treatment (including the ambient and elevated CO2 chambers and the unchambered control sites), Population, and Family nested within population. The growth experiment did not include the effect of individual families nested within the populations. Because the measurement of growth traits took place more than twice, follow-up contrasts were performed to determine the significance of each census interval and its interaction with the main effects. For the main experiment there were three follow-up contrasts: Day 10 with Day 59, Day 59 with Day 7 8, and Day 78 with Day 127. The grth experiment included two contrasts: Day 13 with Day 27, and Day 27 with Day 49. The physiological data for the main experiment were analyzed using a profile analysis similar to the census analysis, but with CO2 measurement level as the repeated 82 measure. Since there were only two repeated measures in the physiological analysis, ambient CO2 and elevated C02, follow-up contrasts were not required. The final biomass allocation data for the growth experiment were analyzed using analysis of variance to compare treatment means. Treatment means for the chambers and unchambered control sites were individually tested for treatment difierences, using Bonferroni corrections for multiple tests of means (Scheiner, 1993). Sample size for each treatment in the grth experiment was 17 - 29. Sample size for each treatment in the main experiment was 87 - 92. Sample size for each family in each CO2 environment for the main experiment was n s 6. Of the 24 families used in the main experiment, 6 families were excluded from analysis because 2 or fewer individuals germinated in one or both CO2 environments. RESULTS The mean values for leaf number, shoot number, and shoot diameter varied considerably over the course of the growing season for both the main experiment (Figures 6- 8, Table 7. A) and the grth experiment (Figures 6 - 8, insets; Table 7. B). Although the plants in the ambient CO2 chambers, the elevated CO2 chambers, and the unchambered control sites were indistinguishable in size for each of the variables at the beginning of the experiment, they assumed different grth trajectories early in the experiment. In some cases the early differences were maintained until the end of the experiment; in other cases, early differences among the main treatments in aboveground grth traits disappeared by the end of the experiment. Leaf number increased more for the plants in the elevated CO2 chambers during the first census interval than for the other two treatments (Figure 6). 83 .m.m mco 28:25 32 _mo_tm> .comcc 25:2:QO 5265 use Emetmaxm EmE Le Emu mzwcmo comm 5 23:5: 52 .8 game 5290 .o 2:9”. 292.:ch ~00 355m. Iml 282250 “223820;: + amnEmco NOOEmEE< IOI R2 >8 2 .28 8 28 L _ _ _ _ _ 2 _ _ _ _ _ me .28 _ _ R .28 2 .28 _ 2 _ 2 >8 .Iecj—urnN JBG1 Iomr CON 84 5.8 55 52865 28 50:55 .9850 E55598 5265 was 22525598 55E 5.. 528 8250 :08 E 55:56 2825 558 .28 255E =m5>0 N 559“. ESE—80 N00 555$ Iml £2250 5.822.883 L1| 559E280 «002255254 IOI 52 >8 3 >8 8 >8 2: >8 _ _ g _ _ 2 _ _ O K m .109 S 2 u. T: w 5 1. \ 2. >8 5. >8 .2 >8 now 0 _ _ _ _ o H DIV. IN HIme 2 m. -2. won 1 I0 mom 5 2. 85 .55 55 52865 38 28225> 2285:: 255598 5265 En E55598 :EE 52 828 2.550 :08 E 522:5: Eocm 5222285209 52 8582: =m5>0 .m 5:22am 552225220 ~00 um2m>52m2 Iml 522250 559E883 + 5522:5220 ~00 E5222E< IOI >~2 >8 E >8 8 >8 2: >8 2 2 2 _ 2 b 2 2 2 2 _ L E O Lt. m 2. >8 >~ >8 2>8 . mm 2 2 2 2 2 Or . - A: u. 19 - TN. 2 2 m- m2. requrnN rooqs 86 >288 >2 >28 8.22 >88 8.” >~2 >8 .888 822m >~m8 >3 >>~8 ~2 .22 .>w~.8 2.0 8 >8 .8852 28.2w >88 5.2 >88 2.: >88 92.: 8 >8 8.28:2 822m >88 2.222% >88 2.~.~m >38 82.3 >~2 >8 .5388 822m >288 8.8 >88 ~>.- >88 .238 8 >8 .2888 822m >98 2 >2 >38 8.2.2 >88 2 .2 2 >8 8808.8 288 >28 2.2.2 >228 8.82 >22 .8 8.22 >~2 >8 58872888 >2 .8 25.22 >58 2.2.2.2 >38 $.82 22> >8 58:57:82 >28 82.2 >88 ~38 >38 25.? 2 >8 5288.288 >88 mm.~ >88 wm.~ >88 wm.~ 2 >8 5028:7288 92 - 20 n a ~22 - 8. u a ~m - S u = "82.2353 5.520 3.8 nae—>2 3.8 :82 05.8 :8: 5:880 ~00 Bantam 8% 29850 85.282325 5.28.30 50 «53:3. .Ammabqoo 2.22228 5.2 258622.85 Eobomaomv Sod v >2 .3 5% mgofi «2:20 8822:: 85352 2.55me .8883 £328: a: 222 323280 823228 mafia 255888 553 882 .88 E2: 28 2232588 on 3 :28 03 83 22033 826 one as 2222 858 >208 822% 255888 E: 553 3 >82 Sm 558220 805 222222 5286 88m .2838 2022282228 225 888220 ~00 2088525 £59880 ~00 253:8 mo 2282.2288on 8552598 22282 5.2 2288 222582538 888 28222208 .< .9 538. 87 >33 8%: >85 88% >28 2.8..“ 9m ~83 aoEMaoA ofloomm >38 and >38 who >88 83 88 82m 88 >88 83 >88 ”no >88 8 .o 8 8.88 858828 >88 88 >88 88 >38 26 A8 8.88 858963.. >88 8: >8 .8 m2 >2 .8 8 .8 8. >8 .3852 82m >88 SA .88 3 >88 8 .8 R >8 838:2 825 >28 as >88 2 .m >38 8 .m 3.8 8. >8 .8088 828 >58 83. >8 .8 82 >38 «.3 8.8 R >8 8808.8 828 >88 >28 >88 8.: >8 8 8.8 8. >8 83888.8 >88 2.” >28 2 .o >28 >3 R >8 83888.8 >8 .8 83 >2 .8 «3 >38 >2 8 >8 82888.8 8.. - mm H a 3 - am u a 8 - mm H a "338.3 52.5 3.8 :32 A68 532 A68 532 5:880 N60 @3965— Smm 3.5.80 6888883 5:880 NOD 953:3. .Ammabnoo 28838 Sm «386%? Eobomdomv Sod v a .3 8% 8.88 :28 88:8 £032 «:08me .8588? $9823 3 a 32380 083:8 mam: 808888 0.83 832 .33 “8.. no @83on on 3 :85 2: 33 £033 80% 28 >88 88 :58 :23 88mm 38888 8: 803 2 .95 8m $388. 80% Ba >853: Scam 28:80 8888808 98 £39880 «00 88820 $88880 «00 3058s mo 888800 808585 6.5on Sm 838 25:588.? 383 33.3%.: .m .h 038. 88 This was true for both the main experiment and the grth experiment. For the growth experiment, all three treatments increased in leaf number in parallel afier Day 27. Leaf number was indistinguishable for the three treatments in the growth experiment at Day 49 (Table 7. B). In the main experiment, leaf number increased in the ambient CO2 chambers during the second census interval to run parallel to the elevated chambers. After Day 7 8, leaf number in the ambient CO2 and elevated CO2 chambers was indistinguishable (Table 7. A). Shoot diameter diverged very early on for the plants in the three treatments in the grth experiment (Figure 7, inset). The plants in the elevated CO2 treatment maintained a higher shoot diameter than either of the other two treatments through Day 49 (Table 7. B). In the grth experiment, shoot diameter did not diverge for the three treatments until after Day 59, when the increase in shoot diameter for the elevated CO2 grown plants was more rapid than for the other two treatments (Figure 7). After Day 7 8, the ambient CO2 chamber plants and the unchambered control plants continued to increase in shoot diameter to the same degree, while the increase in shoot diameter for the elevated CO2 grown plants tapered 01f . However, by the end of the experiment, the three treatments could not be distinguished from each other for shoot diameter (Table 7. A. ). The increase in shoot number for the plants in the experiment was quite difi‘erent than that for the other two size variables (Figure 8). Again, shoot number diverged very early on for the elevated CO2 plants in the growth experiment (Figure 8, inset), as it did for the plants in the main experiment. Shoot number declined for the plants in the main experiment between Day 59 and Day 78; thereafter, the ranking of the three treatments with respect to shoot number remained consistent until the end of the experiment. 89 Repeated Measures Analysis of Census Variables The overall multivariate profile analysis of the census variables revealed a significant main effect of block and an interaction of block with family efl‘ects (main experiment) (Table 8. A) or population effects (grth experiment) (Table 9. A). These block interactions were due to unequal representation of the families or the populations in the blocks, depending up on the experiment. The biologically relevant effects in the experimental model resulted fiom the treatments and the interactions of the treatments with the populations, families, or time of measurement. Overall, the treatment and time of measurement effects were significant for both the main experiment and the grth experiment (Table 8. A; Table 9. A). In addition, there were significant two- and three- way interactions involving treatments, populations, families, and time in the main experiment; only the two-way interaction of time with treatment was significant overall for the grth experiment. Repeated measures analysis of the main experiment The main effect of time was significant for each of the three comparisons of census dates for the main experiment, except for shoot number during the third census interval (Table 8. B). Leaf number (Figure 6) and shoot diameter (Figure 7) increased over time for all three census intervals, while shoot number (Figure 8) did not significantly increase after the third census. The significant time by treatment interaction terms during the first and second intervals indicated that the plants in the chambers and the unchambered control sites did not respond to the same degree for leaf number and shoot number (Figure 6, Figure 8). Both the ambient and elevated CO2 chambers had similar leaf numbers during the second and third census intervals; plants inside both types of chambers had greater leaf 90 ............... Sm A085 Sum ... mz mz R ...._. as .: aoaaaéezafi mz ... ...... 0 ... 02 .m 89.5888 in T." 1. CE “.3." as do Amomvawmt—oo—muwoamh mz mz m2 0 mz as .m 9838888: ...... mz mz M: ... as .0 53.829088 “In T, *** wn *** QNH 6N Amomvaamanogh mz mz mz m mz a .m 8&8“: * m2 m2 0 * wNH am threwogn. ii ** ..i% m *3. 02 .m xofimauogh *** *3. *** m i... NNH .m 08H 380.8 38.8.5 8E? uuuuuuuuuuuuuuu 0*: ..ii iii: HAVEN ... mz mz : ...... a2 .: aéaamrte mz ... ... N ... 02 .m 88¢ ... ... ... m0 ...... m2 .8 as: 88388 m2 m2 m2 m mz as .m 89288 mz mz mz 0 ... 02 .0 5.285 “In 7* "I, om *** mmfl 6N Amend 80% mz mz 02 H mz 5 .m mom mz mz m2 m ... 02 .m out 8083; “In." m2 1. m .51. as am xoofim hove-Haun— wociw hfln—gz “OOH—m hens—pl Maw..— .wfi 300M «magmas mmmovb =nuo>o H6 dug H6832 "maven—H 300 T—flm £003.50” 3% 385 .358: 8 .8250 38088088 808 05 HEB 05083 @8898 08 mo 880888 @80 038.88 @8890." 06 mo “8&0 808 8080@ 888 800.33 85;? 080008 @8398 05 ~26 @0883 .880888 @80 800MB 808 @880 m0080u0§@ 880@ 8m8 800 .33 800380m .085 3888 3838.328 @80 803880 N00 @8380 82—880 NOD 80380 08:08 8888800 .3308; 8088098 808 me 89:80 0598 0080008 @80090M 4 .w 030,—. 91 Table 8. B. Comparisons of census dates for growth experiment. Tests for interactions between main effects and successive census dates. m r’ m Growth Variables; Day 10—Day 59 df Leaf Number Shoot Number Shoot Diameter Time 1 *** *** *** Time*Block 3 *** NS *** Time*Trt 2 * * 1 NS Time*Pop 1 NS NS NS Time*Fam(Pop) 26 * NS * Time*Block*Trt 6 NS NS ** Time*Block*Pop 3 NS NS NS Time*Block*Fam(Pop) 60 * l * l * Time*Trt*Pop 2 NS NS NS Time*Trt*Fam(Pop) 17 NS NS * 1 Error 129 --------------- Day 59-Day 78 (if Leaf Number Shoot Number Shoot Diameter Time 1 ”a: an“: anal: Time*Block 3 ** ** *** Time*Trt 2 * NS *** Time*Pop 1 NS NS NS Time*Fam(Pop) 26 NS NS NS Time*Block*Trt 6 * NS *** Time*Block*Pop 3 NS NS NS Time*Block*Fam(Pop) 62 NS NS * Time*Trt*Pop 2 NS NS NS Time*Trt*Fam(Pop) 17 *1 NS *1 Error 149 ............... Table 8. B. (cont) 92 Day 78-Day 127 (If Leaf Number Shoot Number Shoot Diameter Time 1 *** NS *** Time*Block 3 * NS * Tirne*Trt 2 NS NS NS Time*Pop 1 NS NS NS Time*Fam(Pop) 26 ** NS NS T ime*Block*Trt 6 NS NS NS Time*Block*Pop 3 NS NS NS T ime*Block*Fam(Pop) 63 NS NS * 1 Time*Trt*Pop 2 * 1 NS NS Time*Trt*Fam(Pop) 17 NS NS NS Error 149 --------------- 93 Table 9. A Repeated measures profile analysis of growth experiment variables. Comparisons of ambient CO2 chambers, elevated CO2 chambers, and unchambered control sites. Between subjects tests detect differences among main effects and interactions, summed over the repeated measure. Within subjects tests detect main effect of the repeated variable and interaction of the repeated variable with the main experimental effects. Experimental Effects Growth Variables Between Subjects Num df, Overall (Roy's Leaf Shoot Shoot m Den df Greatest Root) df Number Number Diameter Block 3, 69 NS 2 NS NS NS Treatment (T rt) 3, 69 *** 2 NS NS ** Pop 3, 68 NS 1 NS NS NS Block*Trt 3, 68 NS 1 NS NS NS Block*Pop 3, 69 * 1 2 *1 NS NS Trt*Pop 3, 69 NS 2 NS NS NS Error 70 ----- 70 ----- ---—- ----- Within Subjects Effects: Time 3, 68 *2“: 2 an“: *** *** Time*Block 3, 69 NS 4 NS NS NS Time*Trt 3, 69 ** 4 NS NS ** Time*Pop 3, 68 NS 2 NS NS NS Time*Block*Trt 3, 68 NS 2 NS NS NS Time*Block*Pop 3, 69 NS 4 NS NS NS Time*Trt*Pop 3, 69 NS 4 NS NS NS Error (Time) 70 ..... 14o ............... 94 Table 9. B. Comparisons of census dates for growth experiment. Tests for interactions between main effects and successive census dates. Comparison; Growth Yariablgs; Day 13-Day 27 df Leaf Number Shoot Number Shoot Diameter Time 1 *** an: aura: Time*Block 2 NS NS NS Time*Trt 2 * 1 * 1 ** Time*Pop 1 NS NS NS Time*Block*Trt 1 NS NS NS Time*Block*Pop 2 * 1 * 1 *1 Time*Trt*Pop 2 NS NS NS Error 70 --------------- Day 27-Day 49 df Leaf Number Shoot Number Shoot Diameter Time 1 ma: a: *** Time*Block 2 NS NS * 1 Time*Trt 2 NS NS NS Time*Pop 1 NS NS NS Time*Block*Trt 1 NS NS NS Time*Block*Pop 2 NS NS NS Time*Trt*Pop 2 NS NS NS Error 70 --------------- 95 numbers than the unchambered control sites (Table 7. A). Plants in the elevated CO2 chambers had greater shoot numbers than the unchambered control sites also, but the ambient CO2 chambers could not be distinguished from the either the elevated C02 chambers or the unchambered control sites (Table 7. A). Repeated measures analysis of the growth experiment In the growth experiment, the time effect was significant for both census intervals for all three census variables (Table 9. B). All three variables showed increases in size throughout the experiment (Figures 6 - 8, insets). The interaction of time with treatment, however, was only significant for the first census interval. The chambers and unchambered control sites increased in leaf number, shoot number and shoot diameter at different rates during the first census interval, but after that point the responses increased in parallel By the time the plants in the grth experiment were harvested, the plants in the chambers and the unchambered control sites could not be distinguished fiom each other for leaf number or shoot number, but the plants in the elevated CO2 chambers had a greater shoot diameter than the plants in either the ambient CO2 chambers or the unchambered control sites (Table 7. B). Biomass Allocation in the Growth Experiment After the grth experiment was harvested, biomass allocation patterns were compared for the three treatments. Plants in the ambient CO2 chambers were smaller in terms of overall biomass and root: shoot ratio than in either of the other two treatments (Figure 9), and they had greater specific leaf area than the elevated CO2 plants or the unchambered control plants (Table 7. B). Plants in the elevated CO2 chambers and the unchambered control sites could not be distinguished from one another for aboveground 96 Figure 9. Comparison of overall means for aboveground biomass, belowground biomass, and root: shoot ratio for chambers and unchambered control sites. Vertical bars indicate one s. e. Scale for root: shoot ratio appears to the right. Means were tested by pairwise contrasts in an analysis of variance. Different letters indicate that means difl‘er p < 0.017 (B onferroni adjustment for multiple contrasts). Biomass (g) 97 1.2 H2.00 0.8 1.50 , T T / , ’ ’ —1.00 0.4 "l" — /, "7"" —O_50 00* —0.00 —-0.50 (0.4) —-1.00 (0'8) Ambient coz Unchambered Elevated 002 ‘1'50 Chambers Control Sites Chambers I Belowground Biomass @ Aboveground Biomass + Root:Shoot Ratio Figure 9 0923 roous noon 98 biomass, belowground biomass, root: shoot ratio, or specific leaf area (Figure 9; Table 7. B). Therefore, the plants in the ambient CO2 chambers had less biomass overall and produced leaves of smaller mass per unit area than the plants in either of the other treatments. Because the plants in the ambient CO2 chambers showed increases in leaf number and shoot number similar to the plants in the elevated CO2 chambers, but were smaller in terms of overall biomass, the plants in the ambient CO2 chambers were increasing in size without similarly increasing in biomass compared to elevated CO2 treatment. Therefore, at the time of harvest, the plants grown in both types of chambers were the same size, but the plants in the elevated CO2 chambers were heavier. Conversely, the unchambered control plants were smaller than the elevated CO2 chamber plants in terms of leaf number and shoot number at the final harvest, but were the same as the elevated CO2 plants in terms of final biomass and biomass allocation. Physiology in the Growth Experiment A repeated measures analysis was used to determine the response of the plants in the growth experiment to CO2 measurement level. Plants from the three treatments could not be distinguished from each other for assimilation rates and water use efficiency at the ambient CO2 measurement level (Table 10); however, the plants responded in different degrees to the elevated CO2 measurement level (Figure 10). While the plants from all three treatments had higher assimilation rates and instantaneous water use efficiency in the elevated CO2 measurement level than in ambient CO2 , both the ambient CO2 grown plants and the unchambered control plants had greater photosynthetic rates and water use efficiency in elevated CO2 than the elevated CO2 grown plants (Figure 10; Table 11). A 99 Table 10. Individual tests of treatment means for physiological variables in growth experiment. Comparison of ambient CO2 chambers, elevated CO2 chambers, and unchambered controls. Means were compared using pairwise contrasts in an analysis of variance. Letters indicate significant differences, p < 0.017 (Bonferroni adjustment for multiple contrasts). Ambient CO2 Unchambered Elevated C02 Chamber Mean Control Site Mean Chamber Mean (s.e.) (s.e.) (s.e.) Variable n = 17 - 18 n = 29 n = 29 Assimilation Rate, 10.94 (0.45)“ 10.33 (0.55)“ 9.41 (0.39)“ Ambient CO2 Level Assimilation Rate, 22.22 (0.68)“ 20.43 (0.88)“ 16.59 (0.73)b Elevated CO2 Level Transpiration Rate, 5.99 (0.28)“ 5.78 (0.26)“b 5.05 (0.21)b Ambient CO2 Level Transpiration Rate, 5.96 (0.21)“ 5.85 (0.22)“ 5.54 (0.24)“ Elevated CO2 Level Water Use Efficiency, 1.86 (0.07)“ 1.79 (0.06)“ 1.90 (0.08)“ Ambient CO2 Level Water Use Efficiency, 3.77 (0.13)a 3.54 (0.13)“ 3.04 (0.11)b Elevated CO2 Level Stomatal Conductance, 0.34 (0.03)b 0.57 (0.06)“ 0.27 (0.02)b Ambient CO2 Level Stomatal Conductance, 0.49 (0.04)“ 0.37 (0.03)“ 0.39 (0.04)“ Elevated CO2 Level Intercellular C02, 218.06 (3.59)“ 245.97 (3.18)“ 236.70 (3.92)“ Ambient CO2 Level Intercellular C02, 520.63 (6.65)“ 519.78 (8.38)“ 541.05 (6.00)“ Elevated CO2 Level 100 Figure 10. Comparison of overall growth experiment means (d: s.e.) for assimilation, transpiration, stomatal conductance, and water use efficiency at both ambient and elevated CO2 measurement levels. N 01 \ \B\ o Transpiration (mmol m '2 s'1 ) O O Stomatal Conductance (mol m '2 s'1 ) Water Use Efficiency 9 101 L1 1 Assnmllatlon (umol m 25 1) A o 5 8 6 e— g , B” 4 2 0 .8 4 I><§ nag/i .2 0 5 4 / 3 Va 2 / 1 0 Ambient 002 Elevated C02 Measurement Measurement Level Level Ambient CO 2 Chambers Unchambered Control Sites Elevated CO 2 Chambers Figure 10 102 .................... on ---: ---: 20865 888m mz mz mz mz N mz 880 .N 8898.85 N88.8958 mz mz mz mz 8 02 S .N 8888058 ....r. .2.... 8* ...... N ....r. we .N 88.8. N88.88.88 ....2. m2 ... .2.... 8 .2.... S .m 8058 N00 uuuuuuuuuuuuuuu ..--u 00 ii. ......i Roam m2 8... mz mz N mz we .N 8.8.85 Now mz mz 8... mz 8 8* S .m 88cm ...... ... mz I... N .2.... 8 .N 5 N8 .0 .c 88 < 888 8838 888880 .888 an 8.8388 88805 ..888 8888878 .800b0 888088.80 9.80 888 08: 883 038.808» @8888 08.8 88880888 @880 058883 @88808808 08.80 808.80 8888 8.080@ 808 800.38 88.88;? 088808808 @8888 0888 8028 @0888 .888088888 @8888 800b0 8888 88888 0008808083 8080@ 888 800.3880 88003838 .0880 80.888 3808888888088 @8888 80888880 N00 @8880 808880880 NOD 803888 mo 888888800 808.8090 53on 8 00508.88 888082888830 89880 0508 00885008 @8008838 .88 0388.8. 103 Figure 11. Assimilation versus intercellular leaf CO2 concentration at both ambient and elevated C02 measurement levels for growth experiment. At the ambient CO2 measurement level, intercellular CO2 concentration was close to 200 ,ubar for all treatments. At the elevated CO2 measurement level, intercellular CO2 concentration was close to 500 ubar for all treatments. Vertical bars indicate 2t s.e. for intercellular CO2 means. 104 25‘ w 2? A20 //+ We» : // €15- //Z //% g : // / §§1o:i i¥::://///// .g _ IE 3 _ < 5 Ofi mm 0 100 200 300 400 500 Intercellular CO 2 (ubar) _._ Ambient C0 2 Chambers + Unchambered Control Sites _9_ Elevated CO 2 Chambers Figure 11 600 105 graph of the relationship between assimilation and intercellular CO2 showed that the elevated €02 grown plants had lower assimilation rates at elevated CO2 than the other two groups of plants for an equivalent intercellular CO2 concentration (Figure 11). The relationship between assimilation rate and water use efficiency at the two C02 measurement levels in the growth experiment was similar to that in the main experiment (Klus, Ch. 2). Surprisingly, both the ambient CO2 grown plants and the elevated CO2 grown plants had higher stomatal conductance in elevated CO2 than they did in ambient CO2 (Figure 10), opposite the response of the plants to CO2 level in the main experiment (Ch. 2). Only the unchambered control plants had lower stomatal conductance in elevated CO2 than in ambient C02. Therefore, stomatal conductance had a significant treatment by level interaction term in the analysis (Table 11). Transpiration was the least variable of the physiological traits measured at both ambient and elevated CO2 (Table 10; Figure 11). Transpiration rates were similar for all three groups of plants at both measurement levels. DISCUSSION The effects of elevated CO2 on the aboveground size components of leaf number, shoot number and shoot diameter appeared to be transitory in this experiment. The stimulatory effects of the elevated CO2 treatment occurred within the first census intervals, after which the plants growing in the ambient CO2 chambers grew to match the plants in the elevated C02 with respect to the size traits. These results are consistent with previous experiments investigating grth responses to elevated C02, in which the positive grth influence of elevated CO2 was short-lived (Wulif and Strain, 1982; Poorter, 1993). In 106 fact, in this experiment, the effect of the open-topped chambers seemed to be at least as important an influence on the aboveground size traits as was the elevated C02. The plants in both chambers, with or without added C02, were more similar to each other in terms of aboveground size characteristics than they were to the unchambered control plants. Even though the aboveground size traits often ended up to similar for the treatments, some differences between the elevated CO2 grown plants and the ambient CO2 grown plants were maintained over the extent of the experiment. The allocation of biomass was affected by the presence of elevated C02. For both the main experiment and the growth experiment, specific leaf area was larger for the plants in the ambient CO2 chambers (probably a chamber effect), but the plants in the elevated CO2 chambers had decreased specific leaf area, probably due to the presence of elevated CO2 counteracting the chamber efi‘ect (Klus, Ch. 2). The plants in the main experiment difi‘ered with respect to final biomass allocation after 127 days of growth, primarily by increasing root: shoot ratios in the elevated CO2 environment (Klus, Ch. 1). The biomass allocation patterns of the plants harvested in the grth experiment after 49 days followed this general pattern. That the pattern of biomass allocation response was similar in the growth experiment compared to the main experiment occurred despite difi‘erences in the grth conditions of the two sets of plants. The plants in the grth experiment were planted in four-inch square pots; to maintain air flow within the chambers, the grth experiment pots were placed on the ground between the taller cylindrical pots used for the main experiment. For part of the day the smaller grth experiment pots were shaded by the pots in the main experiment. This condition worsened as the main experiment plants grew larger. Perhaps as a consequence of the shading, the grth experiment plants were much 107 smaller at the time of their harvest (Day 49) than the main experiment plants were only ten days later at their second census, but an examination of the growth response variables indicated that the growth experiment plants had not plateaued in response to elevated CO2 under the shaded conditions (Figures 6 - 8, insets). Despite the difference in grth conditions between the plants in the growth experiment and the plants in the main experiment, the plants in both experiments responded in similar ways to CO2 treatment. The allocation of biomass to belowground tissues in the grth experiment was greater in the elevated CO2 grown plants than it was in the plants in the ambient CO2 chambers; these results matched the results of the main experiment (Klus, Ch. 1) Physiologically, the plants grown and measured in elevated C02 in the growth experiment were photosynthesizing at approximately twice the rate of the plants grown and measured in ambient CO2 (Figure 10), similar to the plants measured in the main experiment (Klus, Ch. 2). Previous studies have determined that the physiological response to elevated CO2 is not hindered by a shade treatment (Ehret and J olifl‘e, 1985). Yet the decline in maximum photosynthetic potential described for the plants in the main experiment (Klus, Ch. 2) had also occurred in the growth experiment plants by Day 49 of the experiment. The plants grown and measured in elevated C02 had lower assimilation rates than the plants grown in ambient CO2 and measured in elevated C02, both those in the ambient CO2 chambers and in the unchambered control sites (Table 10; Figure 10). The relationship between assimilation rates and intercellular CO2 concentration (Figure 11) indicates that either the quantity or the activation state of Rubisco had declined during the period of exposure to elevated CO2 for the plants in the elevated CO2 chambers (von Caemmerer and Farquhar, 1981; Klus, Ch. 2). 108 The response of stomatal conductance in the plants at the elevated CO2 measurement level was atypical Only one other study, conducted on an arctic alpine tundra ecosystem, has found increased stomatal conductance in plants exposed to elevated CO2 (Oberbauer et al., 1986). In the tundra study, the increase in stomatal conductance was associated with increased temperatures, up to 4 °C. The chambers in this study were only 1.7 °C warmer than the unchambered control sites; moreover, the stomatal conductance results from the growth experiment did not match the results from the main experiment (Klus, Ch. 2), even though temperature conditions were the same for both experiments. Therefore, some other environmental condition must have been interacting with the increased intercellular CO2 to increase the stomatal conductance. The mechanics of stomatal dynamics are complicated and not well understood (Zeiger et al., 1987 ). Other than C02, water relations play a role in controlling stomatal aperture (Morison, 1987 ; Mott, 1990); perhaps the water relations of the plants in the grth experiment which were grong in at least partial shade affected the relationship between stomatal aperture and intercellular C02. Given the positive relationship between stomatal conductance and assimilation rate (Pereira, 1994), any environmental condition which would have enabled the growth experiment plants to keep their stomates open despite the increase in intercellular CO2 would have helped them to maximize their photosynthetic response to elevated C02. In this experiment, the overall effect of elevated CO2 did not appear to affect plant size traits as much as the allocation of photosynthate due to increased assimilation rates in the elevated CO2 environment. Root: shoot ratios increased for the elevated CO2 grown plants, while specific leaf areas declined. These results were the same for both the growth 109 experiment and the main experiment. Since assimilation rates did decrease somewhat from their mardmum potential in the elevated CO2 plants, the question arises concerning the fate of nitrogen that may have been reallocated from Rubisco. For a small subset of plants from the main experiment that were analyzed for nitrogen concentration, nitrogen concentration was the same aboveground for the plants grown in ambient CO2 and elevated C02. The plants appeared to mobilize nitrogen from their belowground tissues, because carbon: nitrogen ratios increased belowground and decreased aboveground, resulting in no net change in overall C:N ratio (Klus, Ch. 1). Since the types of nitrogen compounds the plants manufactured was not analyzed, it was not possible to determine the nature of the reallocation of nitrogen. The additional carbon compounds that resulted in decreased specific leaf area in elevated CO2 were not being stored either as starch or soluble sugars (Klus, Ch. 1); it is possible that the plants were manufacturing additional quantities of cellulose or secondary metabolites (Dijkstra and Lambers, 1989; but see Fajer et al., 1992). It was not possible to determine the specific nature of the carbon compounds made in response to elevated CO2 in this experiment. Overall, however, it appeared that Plantago lanceolata was flexible in terms of integrating its physiological processes with patterns of nutrient and biomass allocation within the plant. Even though I examined physiology, grth and allocation patterns in response to elevated CO2 in this experiment, I was not able to determine how elevated CO2 would ultimately affect this species' ability to interact in its community. Because the response of the populations and the families of P. lanceolata to elevated CO2 was not uniform, some or all of the changes in physiology and allocation patterns may ultimately help certain 110 families of Plantago lanceolata to survive, compete, and reproduce with greater success in an elevated CO2 world. Thus, the genotypic variation in response to elevated CO2 that was documented in this study suggests that elevated CO2 may act as an agent of natural selection. The potential for ecological and evolutionary change in response to elevated CO2 should be addressed by a longer-term study that follows Plantago lanceolata across several growing seasons and several generations in an elevated CO2 environment. Only in this way will it be possible to explore the ultimate effects of changes in physiology and allocation patterns in response to elevated CO2 on the fate of Plantago lanceolata in its community. APPENDIX APPENDIX METHODS In the summer of 1991, I conducted a pilot experiment to investigate the nature of genetic variation in response to elevated CO2 for physiological and biomass allocation traits in two populations of Plantago lanceolata. The Kellogg Field (KF) population (Kalamazoo County, Michigan) was located in partial shade on the edge of a mown field. The Ray Boom (RB) population was located in suburban Chicago, Illinois. A total of 50 randomly selected families, 25 from each population, were used in the experiment. In early July, 1991, all seeds fi'om each family were sown in flats and placed in a greenhouse at W. K Kellogg Biological Station, Hickory Comers, Michigan. Germination took place over 7 - 10 days. After germination, the seedlings were transported to the Duke University Phytotron in Durham, North Carolina. Ten seedlings from each family were transplanted into separate 30-cm high pots made from 10-cm diameter PVC pipe with mesh screen bottoms. The potswere filled with a soil-less composition of hardened clay particles to provide a uniform substrate. Five seedlings fiom each family were placed into either an ambient (35.5 Pa) or twice-ambient (71.0 Pa, hereafter referred to as ”elevated") greenhouse in the Phytotron. The plants were watered ad libitum in the morning with half-strength Hoagland's solution. Ifthe plants showed signs of wilting in the afternoon, they were watered again 111 112 with deionized water. Daytime temperature was 28 °C; nighttime temperature was 22 °C. 5 5 days into the experiment, physiological measurements of assimilation and transpiration were made using three Model LCA-2 and one Model LCA-3 portable infra-red gas analyzers (Analytical Development Corp oration, Hoddesdon, UK) and narrow Parkinson Leaf Chambers (Analytical Development Corp oration, Ho ddesdon, UK) at both ambient and elevated C02. Elevated CO2 was provided from the elevated CO2 greenhouse. The CO2 source (ambient or elevated C02) was rotated to a different machine each day. Measurements of physiological traits were made on an intact, most recently fully expanded leaf of each plant. Calculations of CO2 assimilation (umol CO2 m'2 s"), transpiration (mmol H20 m‘2 s"), stomatal conductance (mol H20 m'2 s'l) and intercellular CO2 concentration (ubar) were made using equations from Analytical Development Corporation (1992) and von Caemmerer and Farquhar (1981). At the end of 60 days the plants were harvested. The roots and leaves (shoots) were separated at the soil line, dried at 60°C for at least five days, and weighed. Sgtistical Analysis Preliminary analysis indicated that the four gas exchange variables (assimilation, transpiration, stomatal conductance, and internal leaf CO2 concentration) difl‘ered according to machine used for measurement and date of measurement. The variables were adjusted for these difl‘erences by subtracting the mean for each machine and date from each individual measurement and adding back the grand mean summed over all machines and dates (Tonsor and Goodnight, 1996). This procedure removed deviations due to intrinsic differences between the infra-red gas analyzers, and variation in environmental conditions on different days, while retaining the difl'erences due to CO2 treatment and 113 other main effects. The gas exchange variables assimilation, transpiration, and stomatal conductance were initially analyzed using a multivariate analysis of variance (MAN OVA) to determine overall main and interaction effects. The multivariate analysis was followed by appropriate univariate mixed model analyses for these physiological variables plus instantaneous water use efficiency (the ratio of assimilation to transpiration) and intercellular CO2 concentration. The overall model for the analysis of the physiological variables included the fixed effects of CO2 treatment, and population, one random effect (family nested within population), and interactions for all main effects. Interaction terms containing the random effect were also treated as random effects. Expected mean squares for the main effects in the mixed model AN OVA were divided by the appropriate interaction terms to produce the F values for significance tests. Significance levels for the main effects in the multivariate analysis were interpreted using Roy's Greatest Root, recommended because of its statistical power and the fact that it is applicable to post hoc statistical comparisons (Scheiner, 1993). Sample size for each family in each CO2 environment was 11 = 3-5. Of the 50 families used in the main experiment, 6 families were excluded from the analyses because 2 or fewer individuals germinated. A separate multivariate analysis was conducted for aboveground and belowground biomass, also using Roy's Greatest Root. Mixed model univariate analyses were subsequently conducted for aboveground biomass, belowground biomass, whole plant (total) biomass, and root: shoot ratio. 114 RESULTS CO2 treatment means differed significantly for all of the physiological and biomass allocation variables (Table 12). Multivariate analysis of the physiological variables revealed that the main effects of CO2 treatment and family were highly significant (Table 13. A). In addition, the interaction of CO2 treatment with family was also highly significant, indicating that the families responded difl‘erentially to the elevated CO2 treatment (Figures 12 and 13). The multivariate analysis of the biomass variables showed that all of the main efl‘ects were highly significant (Table 14. B). No interaction terms were statistically significant for the biomass variables. Each of the physiological variables used in the multivariate analysis was also tested using a univariate mixed model analysis of variance (Table 14. A); water use efficiency and intercellular CO2 concentration were analyzed univariately as well. Each of the physiological variables showed a significant response to CO2 treatment. Assimilation rates, water use efliciency, and intercellular CO2 concentration were higher in the elevated CO2 treatment; transpiration and stomatal conductance were lower in the elevated CO2 treatment than in the ambient CO2 treatment. Univariate mixed model analysis of the biomass variables showed significant main effects of CO2 treatment, population, and family for aboveground, belowground and total biomass, and a significant CO2 treatment effect for root: shoot ratio (Table 14. B). Aboveground, belowground, whole plant biomass, and root: shoot ratio were all greater in the elevated CO2 treatment than in ambient C02. Population KF was significantly larger than Population RB for each biomass variable (p < .05), yet the two populations were similar physiologically. The families showed considerable variation in magnitude of response to elevated CO2 for each biomass variable 1 15 (Figure 14). The CO2 treatment x family interactions for the physiological traits and the considerable variation in magnitude and degree of response for the biomass traits for the different families clearly indicated that intraspecific genotypic variation in response to elevated CO2 exists for these two populations of Plantago lanceolata. Genotypic variation in response to elevated CO2 occurred for both instantaneous physiological traits and biomass allocation traits which might influence long-term survival and reproduction in this species. Moreover, the significant differences in response at the population level suggested that a population's habitat of origin might affect its capacity to respond to elevated C02. Yet, the design of the 1991 experiment did not allow me to determine unequivocally either the time of onset of the CO2 response or its duration past 60 days. Additionally, pseudoreplication was a problem, because cost restricted the experiment to only one elevated and one ambient CO2 greenhouse. Therefore, I decided to conduct follow-up experiments in 1992 at the W. K Kellogg Biological Station to explore more My the nature of genotypic variation in response to elevated CO2 under more natural conditions. Setting up the experiment in op en-topped chambers in an old field allowed me to investigate whether the variation uncovered in controlled, high nutrient greenhouse conditions would manifest itself over the course of an entire grong season under less productive conditions. The design of the 1992 experiment also allowed me to measure early growth traits to determine the time of onset of CO2 response, and to investigate possible changes in patterns of physiological and biomass allocation response to elevated CO2 over the course of an entire growing season. 116 Table 12. Overall and population means (s. e.) for CO2 treatments for ambient and elevated CO2 measurements. CO2 treatment means were compared using univariate one-way AN OVAs. Letters represent significant differences for each variable between the two CO2 treatments. Difi‘erences were considered to be significant at the p < 0.025 level (Bonferroni adjustment for multiple comparisons of means). Physiological Variables: Overall Population KF Population RB Assimilation: Ambient CO2 29.9 (0.5)A 29.4 (0.7)A 30.6 (0. 8) Elevated co2 32.8 (0.9)B 32.9 (1.2)B 32.6 (1.2) Transpiration: Ambient CO2 10.0 (0.2)A 9.9 (0.2)“ 10.2 (0.2)A Elevated co2 8.5 (0.2)B 8.2 (0.3)B 8.8 (0.3)B Water Use Efficiency: Ambient CO2 3.0 (0.04)A 3.0 (0.1)” 3.0 (0.1)A Elevated CO2 4.3 (0.3)B 4.6 (0.5)B 3.9 (0.2)B Stomatal Conductance: Ambient C02 1.41 (0.08)A 1.35 (0.10)A 1.52 (0.14)A Elevated C02 0.89 (0.06)B 0.94 (0.09)B 0.82 (0.08)B Intercellular 00,: Ambient (:02 345.0 (2.0)A 343.9 (3.0)A 345.6 (3.0)A Elevated CO2 523.5 (5.5)B 523.0 (8.2)B 524.1 (7.4)B Biomass Variables: Aboveground Biomass: Ambient CO2 10.9 (0.4)A 13.7 (0.5)A 7.6 (0.5)A Elevated co2 14.9 (0.4)B 17.6 (0.6)B 12.0 (0.5)B Belowground Biomass: Ambient CO2 5.9 (0.3)0 7.6 (0.5)A 3.8 (0.4)A Elevated co2 10.2 (0.6)B 12.1 (0.9)B 8.0 (0.7)B Whole Plant Biomass: Ambient CO2 16.8 (0.7)0 21.3 (0.8)A 11.4 (0.8)A Elevated co2 25.1 (0.9)B 29.7 (1.3)B 20.0 (1.1)B Root: Shoot Ratio: Ambient co2 0.53 (0.10)A 0.56 (0.11)A 0.48 (0.09)A Elevated co2 0.66 (0.16)B 0.66 (0.04)B 0.66 (0.04)B 88.0 888.88 8.88 .0 88 N8. 368088888 .. Noo 803.88 N83 eNood 08 N 8A8 .. Noo 4.88.883 8.». E828 83 Ne Eeaaemv 268888888680 88888.88 4.48080 8838. 89.88 02 N 36.: 858888888868 48.480888 8.88 eNNNe CNN N 38 888880; N00 m A a 030.» h 83— «m0uaoumv 9.25— :e 8382:8280: :0 8330852 "00.55 0.050988» 000805 80m $038280 083803038 =80>O .m .2 033. m ....»Neoe 888.8 NENe N8N oe 8886808888848 ... Nov 8828 08.88 8380.88 88N N 8,8 ... aoo $868.0 8.8 eNNNe N8N N8. 2.863581886888868188858 gene 30 0838 88N m 388.: 8868888888 83.48080 80.88 mete 88N N 38 8888888; Noe h A a 030.» m «00¢ 3800.5 {H3— 3 8303:8280: .6 8382058878 "00.58 0052.8g Fofiofiommham Sm flmbafi 083802038 =080>O .< .2 050,—. $0805 “858083820: .8082: “Sesame/one 80038858» 000803 95 80m «50722 8206 .m 880m 00888086880 #888830 28 .aeugmmfiwb douemfifime M05058» fiaomwofimmefim 00:: How <>OZmo membfia 083030—32 .mH 033. 118 80088 Neg N023 0888.0 N8. Reagan ... Noo mNeNe 28.0 880.0 88.0 8 8.8 ... aoo 88.0 1488.0 4580.0 ......4883 N8. 28835888885 968888886880 888858 8328 $488.88 448.8080 444808.88 8 3...: 8588888888898 44488.0 4.848883 teased 8.44883 8 308 88868888; N00 060% 002—0325— 208880:— ESH caucuwkefim cgeuwgefie :8 "00.55 88.88.88 88888580 .m 8.8 088$ 088N088 688.0 28.3 8.03.88 88.88 888. 36808888 ... Noo 8888.0 086.0 N882 688.0 N828 8 .898 .. Nov 3828 eNmS RmNe 86E; NmNee N8. 28.58.86 Aee88e8=eemv 88888888 wwNNe SNMS 2.8828 88.0 83.0.0 8 :89: 8888888868 48.48838 $488.0 $488880 444N080 8.9.80.0 8 38888888888; N00 853 .0 .e m a. 8I88 . 6.885 888.888., 88888668888 .< .88 e888? .0308 80:088.: .3082: :33 $3803 8888089680: .3083: w§08w0>on< 8003093 0088805 .m tam 06880650 00: 80:83 n E dowebfioaoo NOD 8883:0803 H 80 00883038880 830806 N m0 080.8 8830:9058: Mm .022 noueEEmma u < 80058893 Rummaxofiba 4 880m .m<>OZ< #08008 000888 08983:: 89a 00358830830 beam .3 030,—. 119 53.00 48.00 ' /+ w‘” 43.00 I; E / N 8 38.00 9» I 6 E 3 c 33.00 .9. E E '7, 28.00 on < 23.00 g 18.00 Ambient C02 Elevated CO2 —l— Overall —*—- KFB + KF20 —E— R3202 —0—- R3412 + PopKF —U— KF9 —I— KF21 -%— R3222 + R3420 + PopRB —$¢— KF10 + KF22 —G— R3253 46— R3432 + KF1 ——X— KF11 + KF23 —I— R3263 + R3442 -E— KF2 —B— KF12 + KF24 + R3272 —I— R3451 —9— KFS —-fl-— KF14 —B— KF25 + R3291 —.— R8482 —A— KF4 + KF15 —9— R3113 —-l— R3352 + R8493 —9— KF5 —>— KF16 —A— R3133 —-B-— R3373 + R3530 —E— KF6 —-+— Kim; —9— R3193 + R3392 —E— R3610 -E— KF7 —$¢— KF19 Figure 12. Comparison of overall, population and family means for assimilation for ambient and elevated 002 treatments for 1991 experiment. 120 14.00 13.00 A 12.00 'm N E 11.00 0 N 10.00 I E E 9.00 E, E 8.00 '1'? 8 7.00 C S '— 6.00 5.00 4.00 Ambient CO2 Elevated CO2 —I— Overall + KF8 + 1(on —E— R3202 —0— R3412 + PopKF —B— KF9 —I— KF21 —E— R3222 —+— R3420 + PopRB 4G— KF10 + KF22 —G— R3253 —-36— R3432 + KF1 —><— KF11 + KF23 —I— R3233 + R3442 —B— KF2 —E— KF12 + KF24 —I— R3272 —I— R3451 —-G— KF3 —-€-— KF14 —E— KF25 + R3291 —.— R3482 —A— KF4 + KF15 —G— R3113 + 33352 + R3493 —G— KF5 —>— KF16 —A— R3133 + R3373 + R3530 —E— KF6 —+— KF18 —€— R3193 + R3392 —E— R3610 —E— KF7 + KF19 Table 13. Comparison of overall, population and family means for transpiration for ambient and elevated COZ treatments for 1991 experiment. 121 Figure 14. Comparison of overall, population, and family means for aboveground biomass, belowground biomass, and root: shoot ratio by C02 treatment for 1991 experiment. For each category (overall, population, family) means for ambient CO2 treatment appear to the left of means for elevated CO2 treatment. Vertical bars indicate one s.e. Scale for root: shoot ratio appears to the right. Significance levels for individual tests of means appear above bars for aboveground biomass and below bars for belowground biomass. Significance levels: *1 p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. 122 Root: Shoot Ratio io shoot rat + Root % Below ground biomass D Aboveground biomass (6) 9831110153 A A A A N ‘- O I E J t 3 i V v v v 1 I l l I l l l 1 I l 1 l l - — A\\\'.\\‘.\\\\'.\\‘.\\\xxxxsxx'.‘.‘.v.'.\v.' I V— __f T-.\.\\\\\-- ‘ olgaa _ _ \3. .~.\-. :xxxsxxx : — ... 089 8 El I - :8 8: _ MN\\\\N\\\\“.\\‘.\~ ‘ ‘ ‘— —m.\' 967 88 —_: ** _n. u \ “re-8......“ -_ — \\'\N\‘ zg v a a - — \xvuvu ' ‘—c\1\:~' L9? 88 : _ ‘_ mxvua- .u- .x -.. , _ .......... 21?? 8E! -: : — \‘s\\\\\\\\'—\‘ ' _ “-0 ' zev 8 a : : _ M\N:\Y\\\‘u'\\ ' -—~ OZ? 8&1 _ — :III'U \ * — - =— - Z W 88 — — cunt“ - — — (\N‘\\\\\ 36 e 8 a _ _ Mxvuxx- ._u ‘— " - 8 E18 . \i)\_\\) n,-\ u '- n-fh ----- . CL H—' __ snmuxsxxasx .' — 29$ 88 F _— — n‘n'x‘uxx‘vA-vn ‘ 1' —‘— — w L63 8 H _ _ “NV: -- =--- 212 88 _ t — — u...“ 1.x u . .,. . ‘ —-,— _—_-t\~~n-- sea 88 — . . — ~\\\\\\\u.\ — ‘ _t—uuxfl' 892 88 _ — \M“N\N\\\‘.\NV ‘— — t : — Minx-«303.c- ZZZ a 8 ¥ _ ; _'- _ M\\\“\\N\N%\\%‘ "I_ :_)\.~-.\.\— zoz 88 — Z : _ \wumxm-uvvrm's -— —mm~~\w_ set 88 — : 2 — Mixx‘uxijxi'nltnru . . . g - — ...... - 89 L 8 El - ; : _ .vvwxxsxxu-exv — _ \mmnxxvsx 8 L L a a — ' — \uxwxxvnxxu -—-_ _‘_O<\x\\.\x\:\xn SZ :1)! — : : _ AWM\N\\N\\SN\_NN\\\\NN\V ' u — ‘ _— nxvnxvs-Axsxx 5._ vz dM v — : : — \‘n‘YWVIm'l-fi‘amnnfixfix'xT-Tsx"'rn‘» _ — mmuvnvsuu '— e Z d M ‘— — : : — m‘nxmmnxur _ = :- _ M\“\“\h\\\fi\\\_ 82 :l M a, — _ numxxucoocu L- _—. f_vw\“\w\\w Lad» 3 L‘: : — “m“\“\“\vxsx\kx\l.xk'.\\'.\.\-.\\-.\\.-.\.\.-.\.\I.\.\ - _ _: ‘T — mnnxxt-xxxxn \‘(Ifx‘h - — 0 Z d» m _ A5mm~m~v~\~s\\\\\ - - _ _ A“\\W“\\\ ' 6 L :1» I I — : : _ “W1." "_‘ —v.-vw' 9L3» mwwmxnxaxxmxxxxx-nx-sn _ — : : — mwmxsxs-e-nsu- — 9 L d)‘ — : : _ xvmtmx‘. ’ — : : — mm 9 L d)‘ — : : — WMWNKAV“\\N\\\\IJ. —— : :—mn1tm-u\xnu-— v L 3M — — hmwmxsxxxxxx _ “W'- ‘ ' z L fl)‘ “x“mswwsl-Lxxxuv ' — — — “xv“.xsvsu. * L L :lM — _ : — mmwvmxxxx- — — : . — .mxmmmxu-n O 1. 5M — :r’ _. ;. : _ xt-ll‘in‘lxxx- ‘11-‘17.)3‘ . . . . . . . . .. — sxxmmmsmv. — 6 3M F i _: :— M\\\\m\“\\\v\\\\\\\v \ — — _ mw' . 8 d M _ A\.\AM“\\A\\' - ’ M“\\\S“Y\\\\\\\N\S$\\v — - :. — — \\~.\A-.\AI.\A\\\ -— 9 d)! _— 8_ :: — mmvhwswmv " — _ .smsxwxxxsmsx —' 9 AM _ _ : — KIOLIQIIIXIII 111- in. C _ _' — “ms- t d)‘ ‘ Lnxxxxxx; \\ Ix‘lx .7. ".11 . I — — _ Nam“ 8 :I)‘ ' —_‘. _'_ .“xxxxxtxxkzsxxxmxx- .w .v I -— — XI“‘\\I Iii z a» I _ — wwsmsxxxv l _ — muxxxxxxu- L d» * _ — mmxsav ‘k d ‘3 ' _ ' — mvw " a a 0 d ’ \\\\m““\“‘e\\\\‘u\\‘ * E ‘ _ _ \“xvs-ns-‘smx ' 1K d M 0 d ‘K _ ‘8 _ ur- (tuuuax. u . . '0 a _........... 1: "BJGAO I I I I I I I I I I I I I I I I I I I I I I I I O O O A A A A N .8 o o o o ‘— N (‘0 V‘ | LITERATURE CITED LITERATURE CITED Analytical Development Corp oration (1992) LCA-3 Operations Manual. Hoddesdon, UK Antonovics J, Primack RB (1982) Experimental ecological genetics in Plantago. VI. The demography of seedling transplants of P. lanceolata. J Ecol 7 0:55-75 Arnold SJ, Wade MJ (1984) On the measurement of natural and sexual selection: theory. Evolution 382709-719 Arp WJ (1991) Effects of source-sink relations on photosynthetic acclimation to elevated C02. Plant Cell Environ 14:869-875 Arp WJ, Drake BG (1991) Increased photosynthetic capacity of Scirpus olneyi after 4 years of exposure to elevated C02. Plant Cell Environ 14: 1003-1006 Bazzaz FA (1990) The response of natural ecosystems to the rising global CO2 level. Annu Rev Ecol Syst 21: 167-196 Bhattacharya NC, Biswas PK, Bhattacharya Ch , Sionit N, Strain BR (1985) Grth and yield response of sweet potato to atmospheric CO2 enrichment. Crop Sci 25:97 5- 981 Bowes G (1991) Growth at elevated C02: photosynthetic responses mediated through Rubisco. Plant Cell Environ 14:795-806 Bowes G (1993) Facing the inevitable: plants and increasing atmospheric C02. Annu Rev Plant Physiol Plant Mol Biol 44:309-332 Bradshaw AD, McNeilly T (1991) Evolutionary response to global climatic change. Ann Bot 67 (Suppl 1):5-l4 Carter DR, Peterson KM (1983) Effects of a carbon dioxide-enriched atmosphere on the grth and competitive interaction of a C3 and a C4 grass. Oecologia 58: 188-193 Cure JD, Acock B (1986) Crop responses to carbon dioxide doubling: a literature survey. Agric For Meteor 38: 127-145 123 124 Curtis PS, Drake BG, Whigham DF (1989) Nitrogen and carbon dynamics in C3 and C4 estuarine marsh plants grown under elevated CO2 in situ. Oecologia 7 8:297-301 Curtis PS, Snow AA, Miller AS (1994) Genotype-specific effects of elevated CO2 on fecundity in wild radish (Raphanus raphanistrum). Oecologia 97: 100- 105 Curtis PS, Teeri JA (1992) Seasonal responses of leaf gas exchange to elevated carbon dioxide in Populus grandidentata. Can J For Res 22: 1320-1325 Davidson RL (1969) Effect of root/leaf temp erature difierentials on root/ shoot ratios in some pasture grasses and clover. Ann Bot 332561-569 DeLucia EH, Sasek TW, Strain BR (1985) Photosynthetic inhibition after long-term exposure to elevated levels of atmospheric carbon dioxide. Photosyn Res 7: 175- 184. den Hertog J, Stulen I, Lambers H (1993) Assimilation, respiration and allocation of carbon in Plantago major as afl‘ected by atmospheric CO2 levels. Vegetatio 104/105:369-378 Dijkstra P, Lambers H (1989) A physiological analysis of genetic variation in relative growth rate within Plantago major L. Func Ecol 3:57 7-5 87 Ehret DL, Jolifl‘e PA (1985) Photosynthetic carbon dioxide exchange of bean plants grown at elevated carbon dioxide concentrations. Can J Bot 63:2026-2030 Evans LT (1975) The physiological basis of crop yield. In: Evans LT (ed) Crop physiology: some case histories. Cambridge Univ Press, Cambridge UK, pp 327- 355 F ajer ED, Bowers MD, Bazzaz FA (1992) The effect of nutrients and enriched CO2 environments on production of carbon-based allelochemicals in Plantago: a test of the carbon/nutrient balance hypothesis. Am Nat 140:707-723 Fichtner K, Koch GW, Mooney HA (1994) Photosynthesis, storage, and allocation. In: Schulze E-D, Caldwell MM (eds) Ecophysiology of photosynthesis. Springer- Verlag, Berlin, pp 133-146 Garbutt K, Bazzaz FA (1984) The efl‘ects of elevated CO2 on plants. III. Flower, fruit and seed production and abortion. New Phytol 982433-446 Garnier E (1992) Growth analysis of congeneric annual and perennial grass species. J Ecol 80:665-675 125 Geber MA, Dawson TE (1990) Genetic variation in and covariation between leaf gas exchange, morphology, and development in Polygonum arenastrum, an annual plant. Oecologia 85:153-158 Geber MA, Dawson TE (1993) Evolutionary responses of plants to global change. In: Kareiva PM, Kingsolver JG, Huey RB (eds) Biotic interactions and global change. Sinauer Associates Inc, Sunderland MA USA, pp 179-197 Gill JL (1978) Design and analysis of experiments in the animal and medical sciences. Iowa State University Press, Ames IA USA, 409 pp Goldberg DE, Landa K (1991) Competitive efl‘ect and response: hierarchies and correlated traits in the early stages of competition. J Ecol 79: 1013-1030 Goldberg DE, Miller TE (1990) Effects of different resource additions on species diversity in an annual plant community. Ecology 71:213-225 Grime JP (197 7) Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am Nat 111: 1 169-1 194 Hollinger DY (1987) Gas exchange and dry matter allocation responses to elevation of atmospheric CO2 concentration in seedlings of three tree species. Tree Physiol 3: 193-202 Idso SB, Kimball BA (1989) Growth response of carrot and radish to atmospheric CO2 enrichment. Environ Exper Bot 29: 135-139 Idso SB, Kimball BA, Mauney JR (1988) Effects of atmospheric CO2 enrichment on rootzshoot ratios of carrot, radish, cotton and soybean. Agric Ecosys Envt 21:293- 299 Johnson IR ( 1985) A model of the partitioning of growth between the shoots and roots of vegetative plants. Ann Bot 55:421-431 Jones MGK, Outlaw WH Jr, Lowry OH (197 7) Procedure for the assay of sucrose in the range of 10'7 - 10'14 moles. Plant Phys 602379-383 Kalisz, S (1986) Variable selection on the timing of germination in Collinsz'a verna (Scrophulariaceae) Evolution 40:479-491 Keeling CD (1986) Atmospheric CO2 concentrations. Mauna Loa Observatory, Hawaii 1958-1986. NDP-OOl/Rl. Carbon dioxide information analysis center. Oak Ridge Natl Lab USA, Oak Ridge TN USA 126 Kimball BA (1983) Carbon Dioxide and agricultural yield: An assemblage and analysis of 430 prior observations. Agron J 7 5:779-788 Kimball BA, Mauney JR, Nakayama FS, Idso SB (1993) Efl‘ects of increasing atmospheric CO2 on vegetation. Vegetatio 104/105:65-75 Konings H (1989) Physiological and morphological difierences between plants with a high NAR or a high LAR as related to environmental conditions. In: Lambers H, Cambridge NIL, Konings H, Pons TL (eds) Causes and consequences of variation in growth rate and productivity of higher plants. SPB Academic Publishing, The Hague, pp 101-123 Kuiper PJC, Bos M (eds) (1992)P1antago: a multidisciplinary study. Springer-Verlag, Berlin Lambers H (1987 ) Does variation in photosynthetic rate explain variation in growth rate and yield? Neth J Agric Sci 35:505-519 Larigauderie A, Hilbert DW, Oechel WC (1988) Efl‘ect of CO2 enrichment and nitrogen availability on resource acquisition and resource allocation in a grass, Bromus mollis. Oecologia 7 7 :544-549 Leadley PW, Drake BG (1993) Open top chambers for exposing plant canopies to elevated CO2 concentrations and for measuring net gas exchange. Vegetatio 104/105:3-15 McGraw J B, Wulfl‘ RD (1983) The study of plant growth: a link between the physiological ecology and population biology of plants. J Theor Biol 103:21-28 Morison JIL (1987) Intercellular CO2 concentration and stomatal response to C02. In: Zeiger E, Farquhar GD, Cowan IR (eds) Stomatal fimction. Stanford Univ Press, Stanford CA USA, pp 229-251 Mott KA (1990) Sensing of atmospheric CO2 by plants. Plant Cell Environ 13:731-737 Nefiel A, Moor E, Oeschger H, Staufl‘er B (1985) The increase of atmospheric CO2 in the last two centuries. Nature 315245-47 Norby RJ, O'Neill EG, Luxmoore RJ (1986) Effects of atmospheric CO2 enrichment on the growth and mineral nutrition of Quercus alba seedlings in nutrient-poor soil. Plant Physiol 82:83-89 Norby RJ, O'Neil] EG, Hood WG, Luxmoore RJ (1987 ) Carbon allocation, root exudation and mycorrhizal colonization of Pinus echinata seedlings grown under CO2 enrichment. Tree Physio] 3:203-210 127 Oberbauer SF, Oechel WC, Riechers GH (1986) Soil respiration of Alaskan tundra at elevated atmospheric carbon dioxide concentrations. Plant Soil 96: 145-148 Oechel WC, Strain BR (1985) Native species responses to increased carbon dioxide concentration. In: Strain BR, Cure JD (eds) Direct efiects of increasing carbon dioxide on vegetation, DOE/ER-023 8, US Dept of Energy, Washington DC USA, pp 1 17- 154 O'Neil] EG, Luxmoore RJ, Norby RJ (1987 ) Increases in mycorrhizal colonization and seedling growth in Pinus echinata and Quercus alba in an enriched CO2 atmosphere. Can J For Res 172878-883 Patterson DT, Flint EP (1982) Interacting effects of CO2 and nutrient concentration. Weed Sci 30:389-394 Patterson DT, Flint EP, Beyers JL (1984) Efl‘ects of carbon dioxide enrichment on competition between a C4'weed and C3 crop. Weed Sci 32: 101-105 Pereira J S (1994) Gas exchange and growth. In: Schulze E-D, Caldwell MM (eds) Ecophysiology of photosynthesis. Springer-Verlag, Berlin, pp 147—181 Poorter H (1989) Interspecific variation in relative growth rate: on ecological causes and physiological consequences. In: Lambers H, Cambridge ML, Konings H, Pons TL (eds) Causes and consequences of variation in growth rate and productivity of higher plants. SPB Academic Publishing, The Hague, pp 45-68 Poorter H (1993) Interspecific variation in the growth response of plants to an elevated ambient CO2 concentration. Vegetatio 104/105:77-97 Poorter H, Pot S, Lambers H (1988) The elfect of an elevated atmospheric CO2 concentration on growth, photosynthesis and respiration of Plantago major. Phys Plant 73: 553-559 Potvin C, Lechowicz MJ, Tardif S (1990) The statistical analysis of ecophysiological response curves obtained from experiments involving repeated measures. Ecology 71: 1389-1400 Primack RB ( 197 9) Reproductive effort in annual and perennial species of Plantago (Plantaginaceae). Am Nat l 14:51-62 Primack RB, Antonovics J (1982) Experimental ecological genetics in Plantago. VII. Reproductive effort in populations of P. lanceolata L. Evolution 362742-752 128 Radin JW, Lu Z, Percy RG, Zeiger E (1994) Genetic variability for stomatal conductance in Pima cotton and its relation to improvements of heat adaptation. Proc Natl Acad Sci 91:7217-7221 Sage RF, Sharkey TD, Seemann JR (1989) Acclimation of photosynthesis to elevated CO2 in five C3 species. Plant Physiol 89:590-596 SAS Institute (1988) SAS/ STAT User's Guide, Release 6.03 Edition. SAS Institute Inc, Cary NC USA, 1028 pp Scheiner SM (1993) MANOVA: Multiple response variables and multispecies interactions. In: Scheiner SM, Gurevitch J (eds) Design and analysis of ecological experiments. Chapman and Hall, New York NY USA, pp 94- 1 12 Scheiner SM, Gurevitch J (eds) (1993) Design and analysis of ecological experiments. Chapman and Hall, New York NY USA, 445 pp Scheiner SM, Gurevitch J, Teeri JA (1984) A genetic analysis of the photosynthetic properties of populations of Danthonia spicata that have difi‘erent growth responses to light level. Oecologia 64: 74-77 Sharkey TD (1985) Photosynthesis in intact leaves of C3 plants: physics, physiology and rate limitations. Bot Rev 51:53-105 Shipley B (1995) Structured interspecific determinants of specific leaf area in 34 species of herbaceous angiosperms. Funct Ecol 91312-3 19 Smith SD, Strain BR, Sharkey TD (1987) Effects of CO2 enrichment on four Great Basin grasses. Funct Eco] 12139-143 Solbrig OT (1981) Studies on the population biology of the genus Viola. H. The efiect of plant size on fitness in Viola sororia. Evolution 35: 1080- 1093 Strain BR (1987) Direct effects of increasing atmospheric CO2 on plants and ecosystems. TREE 2: 18-21 Strain BR, Cure JD (eds) (1985) Direct effects of increasing carbon dioxide on vegetation. DOE/ER-023 8, US Dept of Energy, Washington DC USA, 286 pp Stulen I, den Hartog J (1993) Root growth and fimctioning under atmospheric CO2 enrichment. Vegetatio 104/105 99-1 15 Teramura AH (1983) Experimental ecological genetics in Plantago. IX. Differences in grth and vegetative reproduction in Plantago lanceolata L. (Plantaginaceae) fiom adjacent habitats. Amer J Bot 70:53-58 129 Teramura AH, Strain BR (1979) Localized populational differences in the photosynthetic response to temperature and irradiance in Plantago lanceolata. Can J Bot 57:2559-2563 Thornley JHM (1972) A balanced quantitative model for root: shoot ratios in vegetative plants. Ann Bot 36: 431-441 Tilman D (1993) Carbon dioxide limitation and potential direct effects of its accumulation on plant communities. In: Kareiva PM, Kingsolver JG, Huey RB (eds) Biotic interactions and global change. Sinauer Associates Inc, Sunderland MA USA, pp 333-346 Tonsor SJ (1985) Intrapopulational variation in pollen-mediated gene flow in Plantago lanceolata L. Evolution 39:7 7 5-782 Tonsor SJ (1990) Spatial patterns of difl‘erentiation for gene flow in Plantago lanceolata. Evolution 44: 1373- 1378 Tonsor SJ, Kalisz S, Fisher J, Holtsford TP (1993) A life-history based study of population genetic structure: seed bank to adults in Plantago lanceolata. Evolution 47 :833-843 Tonsor SJ, Goodnight CJ (1996) Evolutionary predictability in natural populations: Do mating wstem and epistasis interact to afi‘ect heritabilities in Plantago lanceolata L.? Evolution: in press Van der Aart PJM (1985) Demographic, genetic and ecophysiological variation in Plantago major and P. lanceolata in relation to vegetation type. In: White J (ed) The population structure of vegetation. Handbook of vegetation science, Part 111. W Junk, Boston MA USA Via S, Lande R (1983) Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39:505-522 von Caemmerer S, Farquhar GD (1981) Some relationships between the biochemistry of photosynthesis and the gas exchange of leaves. Planta 153: 376-3 87 von Ende CN (1993) Repeated-measures analysis: growth and other time-dependent measures. In: Scheiner SM and Gurevitch J (eds) Design and analysis of ecological experiments. Chapman and Hall, New York NY USA, pp 113-137 Wade MJ, Kalisz S (1990) The causes of natural selection. Evolution 44: 1947-1955 130 Williams WE, Garbutt K, Bazzaz FA (1988) The response of plants to elevated C02. V. Performance of an assemblage of serpentine grassland herbs. Envt Expt Bot 28: 123- 13 0 WolifK, van Delden W (1989) Genetic analysis of ecological relevant morphological variability in Plantago lanceolata L. IV. Response and correlated response to bidirectional selection for leaf angle. Heredity 62: 153- 160 Woodward FI, Thompson GB, McKee [F (1991) The effects of elevated concentrations of carbon dioxide on individual plants, populations, communities and ecosystems. Ann Bot 67 (Sl):23-38 Wright S (1969) Evolution and the Genetics of Populations. . Univ Chicago Press, Chicago IL USA Wu L, Antonovics J (197 6) Experimental ecological genetics in Plantago. 11. Lead tolerance in Plantago lanceolata and Cynodon daczylon from a roadside. Ecology 571205-208 Wulfl‘ RD, Alexander HM (1985) Intraspecific variation in the response to CO2 enrichment in seeds and seedlings of Plantago lanceolata L. Oecologia 66:45 8- 460 Wulff RD, Strain BR ( 1982) Efl‘ects of CO2 enrichment on grth and photosynthesis in Desmodium paniculatum. Can J Bot 60: 1084-1091 Yelle S, Beeson RC Jr, Trude] MJ, Gosselin A (1989) Acclimation of two tomato species to high atmospheric C02. Plant Physiol 90:1465-1472 Zeiger E, Farquhar GD, Cowan IR (eds) Stomatal fimction. Stanford Univ Press, Stanford CA USA, 503 pp Ziska LH, Teramura AH (1992) Intraspecific variation in the response of rice (Oryza sativa) to increased CO2 - photosynthetic, biomass and reproductive characteristics. Physiol Plant 84:269-276 I 0 W i . .3... < ....»Hri 3.1... )1 ill 2 4 .3 0 2 4 O 3 9 2