Cost-Benefit Theory of Leaf Lifespan Using Seedlings of Tropical Tree Species

stefanescu_c ( .pdf )

stefanescu_cpage_Page_23.txt

stefanescu_cpage_Page_60.txt

stefanescu_cpage_Page_65.txt

stefanescu_cpage_Page_27.txt

stefanescu_cpage_Page_39.txt

stefanescu_cpage_Page_06.txt

stefanescu_cpage_Page_22.txt

stefanescu_cpage_Page_38.txt

stefanescu_cpage_Page_21.txt

stefanescu_cpage_Page_31.txt

stefanescu_cpage_Page_61.txt

stefanescu_cpage_Page_10.txt

stefanescu_cpage_Page_44.txt

stefanescu_cpage_Page_04.txt

stefanescu_cpage_Page_07.txt

stefanescu_cpage_Page_20.txt

stefanescu_cpage_Page_29.txt

stefanescu_cpage_Page_36.txt

stefanescu_cpage_Page_15.txt

stefanescu_cpage_Page_18.txt

stefanescu_cpage_Page_09.txt

stefanescu_cpage_Page_25.txt

stefanescu_cpage_Page_56.txt

stefanescu_cpage_Page_19.txt

stefanescu_cpage_Page_32.txt

stefanescu_cpage_Page_53.txt

stefanescu_cpage_Page_64.txt

stefanescu_cpage_Page_12.txt

stefanescu_cpage_Page_30.txt

stefanescu_cpage_Page_43.txt

stefanescu_cpage_Page_08.txt

stefanescu_cpage_Page_17.txt

stefanescu_cpage_Page_47.txt

stefanescu_cpage_Page_41.txt

stefanescu_cpage_Page_51.txt

stefanescu_cpage_Page_11.txt

stefanescu_cpage_Page_34.txt

stefanescu_cpage_Page_37.txt

stefanescu_cpage_Page_42.txt

stefanescu_cpage_Page_45.txt

stefanescu_cpage_Page_54.txt

stefanescu_cpage_Page_48.txt

stefanescu_cpage_Page_35.txt

stefanescu_cpage_Page_46.txt

stefanescu_c_pdf.txt

stefanescu_cpage_Page_01.txt

stefanescu_cpage_Page_59.txt

stefanescu_cpage_Page_40.txt

stefanescu_cpage_Page_33.txt

stefanescu_cpage_Page_28.txt

stefanescu_cpage_Page_62.txt

stefanescu_cpage_Page_14.txt

stefanescu_cpage_Page_58.txt

stefanescu_cpage_Page_24.txt

stefanescu_cpage_Page_63.txt

stefanescu_cpage_Page_16.txt

stefanescu_cpage_Page_52.txt

stefanescu_cpage_Page_02.txt

stefanescu_cpage_Page_05.txt

stefanescu_cpage_Page_50.txt

stefanescu_cpage_Page_13.txt

stefanescu_cpage_Page_49.txt

stefanescu_cpage_Page_57.txt

stefanescu_cpage_Page_26.txt

stefanescu_cpage_Page_03.txt

stefanescu_cpage_Page_55.txt

COST-BENEFIT THEORY OF LEAF LIFESPAN IN SEEDLINGS OF TROPICAL
TREE SPECIES














By

CARLA C. STEFANESCU


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006


































Copyright 2006

by

Carla C. Stefanescu















ACKNOWLEDGMENTS

I would like to thank my advisor, Kaoru Kitajima, for giving me the opportunity to

pursue a master's degree under her guidance, providing me with financial support during

my first field season and academic support throughout the completion of this degree. I

would also like to thank my committee members, Jack Putz and Tim Martin, who

provided invaluable comments and support throughout the thesis writing process. I

would also like to thank the Botany Department staff, in particular Kimberly Holloway,

who offered necessary logistic support throughout my time at the University of Florida.

I thank my parents, Drs. Doru and Lydia Stefanescu, whose integrity, courage, and

discipline have provided me with a lifelong example and reminder of what I one day

hope to become. I would also like to thank my sister, Alina Stefanescu, whose creativity

and free spirit have provided me with a better perspective on life. Last, but not least, I

would like to thank the smallest member of my family, Maxwell Stefanescu, whose sheer

presence serves as a reminder that some of life's miracles cannot be explained.

I would also like to thank my dear friends Andrea Crino, Christine Lucas, Silvia

Alvarez-Claire, Jenny Schafer, Leslie Boby, Heather Loring, Adrienne Frisbee, and

Hanna Lee, without which my time in Gainesville would have been much more

productive but much less worthwhile.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES ............................................... vi

LIST OF FIGURES .............................................. vii

ABSTRACT .................................................... ................. ix

CHAPTER

1 DECLINE OF PHOTOSYNTHETIC CAPACITY WITH LEAF AGE IN
SEEDLINGS OF ELEVEN TROPICAL TREE SPECIES....................................1...

In tro d u ctio n .................................................................................................. ............... 1
M eth o d s ...................................................................................................... ........ .. 7
S tu d y S ite ............................................................................. . ..................7
Plot Establishm ent ................................................... ............... .......... .......... .... .. 8
Measurement of Light and Photosynthesis in the Field ..................................8...
Laboratory Measurements and Statistical Analyses.......................................10
R e su lts ...................................................................................................... . .............. 1 1
D isc u ssio n ............................................................................................................... ... 1 2

2 LEAF AGE, PHOTOSYNTHETIC CAPACITY, AND LEAF LIGHT LEVEL
AS PREDICTORS OF DAILY CARBON GAIN USING SEEDLINGS OF FIVE
TR O PIC A L TR E E SPE C IE S .....................................................................................30

In tro d u ctio n ................................................................................................................ 3 0
M e th o d s ...................................................................................................................... 3 4
S tu d y S ite ............................................................................................................. 3 4
Plot E stablishm ent ........................................................................... ..... .............. .. 34
Measurement of Light and Photosynthesis in the Field .................................... 35
Calculation of D aily Carbon G ain.................................................. ................ 36
Statistical Analyses ..................................................... ............ 37
R e su lts....................................................................................................... ....... .. 3 8
D isc u ssio n ............................................................................................................... ... 3 9










L IST O F R EFER EN CE S ...... .................................................................... ................ 51

BIO GR APH ICAL SK ETCH ...................... .............................................................. 55


























































v















LIST OF TABLES


Table page

1-1 Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer
m ethod for each species studied.......................................................... ............... 18

1-2 Number of plants and leaves measured for Amax, Aarea, LMA, Narea, and PNUE
m easurem ents for each species ........................................................... ................ 18

1-3 Number of plants and leaves for which % PFD and Aarea were measured for a
su b set o f sp ecie s ....................................................................................................... 19

1-4 Mean leaf lifespan and regression statistics (x and y intercepts and slopes) for
area- and mass-based light-saturated net photosynthesis against leaf age (days
after full expansion) for seedlings of eleven tropical tree species. ........................20

1-5 Mean leaf lifespan and regression statistics (x and y intercepts and slopes) for
light-saturated net photosynthesis (Aarea in micromols of CO2 m-2 S-1) against
leaf age (days after full expansion) for eight tropical tree species........................20

2-1 Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer
m ethod for each species studied.......................................................... ................ 43

2-2 Number of seedlings and leaves for which % PFD and Aday were measured in
each of five species of tropical tree species ........................................ ................ 43

2-3 Multiple regression analysis of seedlings of five tropical tree species with Amax
and leaf PFD as predictor variables, and daily carbon gain as the response
v a ria b le .................................................................................................................. ... 4 3

2-4 Light curve parameters for five species of tropical tree seedlings used to
calculate daily carbon gain ....................................... ....................... ................ 44















LIST OF FIGURES


Figure page

1-1 Light saturated photosynthetic assimilation rates (Aarea) of leaves of different age
on seedlings of eleven tropical tree species ........................................ ................ 21

1-2 Mass-based photosynthesis (Amass) of leaves with contrasting leaf age in eleven
species of tropical trees .............. ............... ................................................ 22

1-3 Relationship between initial Aarea, Aarea-leaf age slope, Amass, Amass-leaf age
slope with median leaf lifespan for each species. ............................... ................ 23

1-4 The relationship between the actual median leaf lifespan and an estimate of the
time when photosynthetic capacity would reach zero (parameter b, the x
intercept of the leaf age-Aarea and leaf age-Amass regression)...............................24

1-5 Relationship between leaf mass per unit area (g m-2) and leaf age in seedlings of
eleven species of tropical trees............................................................ ................ 25

1-6 Relationship between leaf nitrogen (g m-2) and leaf age.....................................26

1-7 Relationship between photosynthetic nitrogen use efficiency (PNUE) and leaf
ag e .......................................................................................................... ........ .. 2 7

1-8 Decline in %PFD (the total daily PFD incident on the leaf relative to the total
daily PFD above the canopy) with leaf age in a subset of species and leaves .........28

1-9 Effect of % PFD (the total daily PFD incident on the leaf relative to the total
daily PFD above the canopy) on light-saturated photosynthesis (Aarea) and for a
sub set of species and leaves ....................................... ...................... ................ 29

2-1 A schematic representation of the expected relationship between leaf age, leaf
PFD Amax, and Aday ..... .. .......................... ........... ........................... .............. 45

2-2 The relationship between % PFD and mmol PFD m-2 day-1 for all leaves and
sp e cie s p o o led .......................................................................................................... 4 5

2-3 The percent of time leaves of different ages spent above and below their light
compensation point (LCP) relative to the total day length..................................46









2-4 The relationship between leaf age and three-day averages of daily carbon gain in
seedlings of five species of tropical trees............................................ ................ 47

2-5 The relationship between maximum photosynthetic capacity (Amax) and three
day averages of daily carbon gain for seedlings of five species of tropical trees. ...48

2-6 The relationship between leaf age and three day averages of light incident on
those leaves for seedlings of five species of tropical trees..................................49

2-7 The relationship between daily carbon gain and Amax or leaf PFD when all
m easurem ents and species w ere pooled ............................................. ................ 50















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

COST-BENEFIT THEORY OF LEAF LIFESPAN IN SEEDLINGS OF TROPICAL
TREE SPECIES

By

Carla C. Stefanescu

August 2006

Chair: Kaoru Kitajima
Major Department: Botany

Variation in leaf lifespan is highly correlated with variation in leaf physiological

and structural traits within and among plant communities. The cost-benefit theory of leaf

lifespan links interspecific variation in leaf lifespan with the decline of photosynthetic

capacity with leaf age, assuming that photosynthetic capacity (Amax) is linearly correlated

with daily net assimilation rate (Aday). In Chapter 1, I evaluated these alternative

predictions by comparing the decline rate of photosynthetic capacity with leaf age for

eleven tropical tree species from Barro Colorado Island, Panama. Leaves were sampled

from seedlings that had been grown for 2-3 years in common gardens established in

natural tree fall gaps. Leaves representing a range of leaf ages were sampled for

photosynthetic capacity (Amax), total daily photon flux density (PFD), leaf mass per unit

area (LMA), leaf nitrogen (leaf N), and photosynthetic nitrogen use efficiency (PNUE).

In all species, Amax, PFD, and PNUE decreased with leaf age. Leaf N decreased with leaf

age in seven out of eleven species. There was no consistent relationship between LMA









and leaf age. In most species, leaves were lost before their photosynthetic capacity

reached zero. Species with shorter leaf life spans had a high initial photosynthetic

capacity and a steep decline rate of photosynthetic capacity with leaf age. My results

favor the second prediction of the cost-benefit theory, suggesting that the total number of

leaves per seedling is constrained by factors other than decline of photosynthetic rates

with leaf age.

The above theoretical predictions assume that instantaneous photosynthetic

capacity (Amax) is linearly correlated with daily net carbon gain (Aday). In Chapter 2, I

evaluated how Aday may be correlated with Amax by estimating daily carbon gain for

seedling leaves of five tropical tree species that varied in median leaf lifespan. I

estimated daily net carbon gain averaged over three days, using photosynthetic light

response curves and PFD directly measured on individual leaves. While leaf age was

correlated with carbon gain in four out of the five species in this study, Amax was

correlated with carbon gain in three species. In contrast, leaf PFD was correlated with

carbon gain in all species. In a multiple regression using leaf PFD and Amax as predictive

variables for Aday, Amax was a predictor of Aday only in one species. These results can be

explained in terms of the light environment leaves experienced during the study.

Previous studies that documented a strong correlation between Amax and Aday measured

canopy leaves that received consistent light saturating conditions. In contrast, the leaves

in this study received light below their light compensation point for 12 to 52% of the total

day light. Based on the results of this study, it appears that leaf PFD is a better predictor

of daily carbon gain than Amax in conditions where light is not saturating.














CHAPTER 1
DECLINE OF PHOTOSYNTHETIC CAPACITY WITH LEAF AGE IN SEEDLINGS
OF ELEVEN TROPICAL TREE SPECIES

Introduction

Tropical rain forest species produce leaves with widely variable leaf lifespan. For

example, while leaves of pioneer species generally live less than one year, leaves of

shade-tolerant species can live five years or more (Coley 1988, Williams et al. 1989).

Variation in leaf lifespan among species reflects genotypic differences that evolved partly

in relation to life history strategies. Pioneers survive best in sites with high light and

nutrients where their inherent fast rates of growth and leaf turnover help them out-

compete other plants (Bazzaz 1979, Ellsworth and Reich 1996). Shade-tolerant species,

although rapidly out-competed in high light environments, persist in light-limited

environments where slower growth rates, lower photosynthetic capacity, and longer leaf

lifespans allow them to maintain a positive carbon budget by amortizing initial

construction costs (Bazzaz 1979, Koike 1988). When comparing species, it has been

shown that shade tolerant species have lower photosynthetic capacity, higher leaf mass

per unit area, and longer leaf lifespans than pioneer species under comparable light

conditions (Williams et al. 1989, Ellsworth and Reich 1996, Reich et al. 2004). These

patterns are thought to be the result of functional constraints that enhance either

productivity or nutrient conservation and defense. A plant can allocate resources to

leaves with a high photosynthetic capacity but short lifespan, or produce physically

resistant leaves that have a lower rate of carbon assimilation over a longer period of time









(Reich et al. 1992). For example, thicker cell walls are necessary to increase leaf

toughness which provides mechanical protection essential to keeping leaves for long

periods of time (Hikosaka 2004). However, thicker leaves require higher nitrogen

investments in cell walls rather than photosynthetic proteins and thus, have lower overall

carbon assimilation (Hikosaka 2004).

Variation in leaf lifespan is highly correlated with variation in leaf physiological

and structural traits within and among communities (Reich et al. 1991, Reich et al. 1992,

Reich et al. 1999). Mass-based photosynthetic capacity (Amass) and leaf nitrogen (Nmass)

are usually positively correlated, and both are negatively correlated with leaf mass per

unit area (LMA). Amass and Nmass decrease with increasing leaf lifespan, while LMA

increases with increasing leaf lifespan. The decline of Amass and Nmass with leaf age is

thought to be due to the retranslocation of nitrogen to younger leaves which maximizes

whole-shoot photosynthetic income rather than to uncontrolled physiological

deterioration (Field and Mooney 1983, Ackerly and Bazzaz 1995). Hikosaka et al.

(1994) showed that age-related changes due to retranslocation were more pronounced

when of light or nitrogen were limiting.

Photosynthetic nitrogen use efficiency (PNUE), the rate of photosynthesis

expressed per unit nitrogen, is an important leaf trait to characterize in relation to leaf

lifespan. Although a few studies did not find a change in PNUE with leaf age (Mooney

et al. 1981, Field and Mooney 1983), more recent studies have shown that PNUE

decreases with both leaf age and LMA (Reich et al. 1991, Kitajima et al. 1997, Escudero

and Mediavilla 2003, Mediavilla and Escudero 2003, Wright et al. 2004). The

retranslocation of nitrogen to younger leaves can result in the decline of both leaf









nitrogen and PNUE in older leaves which consequently leads to the decline of

photosynthetic capacity with leaf age. The decrease of PNUE with LMA can be

explained by the allocation of nitrogen to non-photosynthetic functions in leaves

(Escudero and Mediavilla 2003, Hikosaka 2004). Leaves with a high LMA allocate a

larger amount of nitrogen to cell wall proteins, and a lower amount of nitrogen to

photosynthetic proteins (Takashima et al. 2004). This trade-off in nitrogen allocation

between photosynthesis and structure results in a negative correlation between PNUE and

LMA (Hikosaka 2004). Thus, in studies attempting to explain age-related changes in

photosynthetic decline, the relationship between leaf age, leaf nitrogen, and PNUE must

be evaluated.

Cost-benefit analyses have been used to explain variation in leaf lifespan (Chabot

and Hicks 1982, Kikuzawa 1991, Kikuzawa and Ackerly 1999). Leaf lifespan is

typically considered to reflect a balance between lifetime leaf carbon gain and

construction and maintenance costs (Chabot and Hicks 1982). Thus, the affect of leaf age

on photosynthesis is important in cost-benefit theories of leaf lifespan. Maximum

photosynthetic capacity (Amax) typically declines monotonically with leaf age and can be

approximated by the following linear equation: Amax = a (1 t/b), where t is the leaf age

in days from the time of full leaf expansion (Kikuzawa 1991). The parameter a (y

intercept of the regression) represents the initial photosynthetic capacity after full leaf

expansion, and can be directly measured. The parameter b (x intercept of regression)

represents the time when photosynthetic capacity would reach zero. This parameter is a

statistical extrapolation determined as a function of the initial photosynthetic rate (a) and

the rate of its decline (parameter a/b, the slope of the regression).









Cost-benefit analyses of leaf lifespan have generated two alternate predictions. If

there are no external constraints on the maximum number of leaves produced and

photosynthesis declines with leaf age (i.e., self-shading causes photosynthetic decline and

limits the number of leaves per shoot; Case 1), a leaf is expected to maximize net carbon

gain over its entire lifetime (Kitajima et al. unpublished manuscript). In this case, the

optimum leaf lifespan should be close to the time when photosynthetic rates equal zero

(Parameter b). Since total daily net photosynthetic income should reach zero before net

photosynthetic rates equal zero, actual leaf lifespan is expected to be less than parameter

b (Kitajima et al. 2002). Parameter b approximates actual leaf longevity when the same

leaf is repeatedly measured for plants with very short lifespans (Ackerly and Bazzaz

1995). In contrast, in tropical tree species with leaf life-times ranging from 183-343

days, b was significantly greater than actual leaf longevity (Kitajima et al. 1997).

Alternatively, if the total number of leaves is limited by external resources (Case

2), a leaf is expected to maximize the rate of net carbon gain averaged over the lifetime of

the leaf (Kikuzawa 1991, Kikuzawa and Ackerly 1999, Kitajima et al. unpublished

manuscript). In this scenario, leaves are replaced when their net carbon gain per unit

time averaged over their entire lifespan has reached a maximum and optimal leaf lifespan

is predicted to be equal to (2bC/a)1/2 where C is the construction cost of the leaf

(Kikuzawa 1991). The parameters a and C can be considered on a mass or area basis.

Based on this theoretical model, it is expected that leaf lifespan will be short when the

initial photosynthetic rate of the leaf is high, and long when construction costs are high or

the rate of decline of photosynthetic rates with leaf age is slow (Kikuzawa 1991). Based

on a comparison of seven canopy tree species, Kitajima and others (unpublished









manuscript) found that Case 1 is most applicable only to short-lived leaves of pioneer

species, while Case 2 is equally applicable to both short- and long-lived leaves.

Numerous studies have analyzed the maximum photosynthetic rates of different

species (Reich et al. 1991, Reich et al. 1992, Reich et al. 1999). However, data on the

rate of photosynthetic decline with leaf age in species with different lifespans are

relatively scarce (Kikuzawa and Ackerly 1999). Such data are essential in estimating the

long-term carbon budget of individual leaves and entire plants. Here, we report the rate

of decline of photosynthetic capacity with leaf age for seedlings of eleven tropical tree

species with median leaf lifespans of from 140-797 days. Previous studies have focused

on comparing the rate of photosynthetic decline for canopy trees (Kitajima et al. 1997,

Kitajima et al. 2002) or have only evaluated changes in one species (Ackerly and Bazzaz

1995). Thus, my study appears to be the first to evaluate whether cost-benefit theory

explains variation in the leaf lifespan of seedlings. Why might leaf lifespan be different

in seedlings and adults? In both seedlings and adults, leaf lifetimes are expected to be

correlated with the cost and maintenance of leaves and supporting structures (Kikuzawa

and Ackerly 1999). The higher initial construction and maintenance costs, the longer a

leaf is expected to live in order to pay back those costs. As seedlings grow in height, the

costs of biomechanical support and the transport of water and nutrients to leaves are also

expected to increase (Kikuzawa and Ackerly 1999). To offset these increased costs, leaf

photosynthetic capacity should also increase with plant height. Indeed, in temperate

species, light-saturated photosynthetic rates have been documented to increase with

ontogeny, from seedlings to saplings and adult trees (Koike 1988). As leaves with higher

photosynthetic capacity and more supporting structures are expected to be more costly to









produce, they are also anticipated to require longer to pay back initial construction costs.

Thus, seedlings should have shorter leaf lifespans than adult trees of the same species

because construction and maintenance costs of supporting structures are expected to be

lower for seedlings than adults (Kikuzawa and Ackerly 1999). When leaf lifespan was

compared across species with different life forms, species with lower biomechanical and

transport costs, such as aquatic floating plants, had a shorter leaf lifespan than leaves with

higher construction and maintenance costs such as temperate deciduous trees (Kikuzawa

and Ackerly 1999). Alternatively, leaves in high light environments are expected to pay

back construction and maintenance costs more quickly than leaves in shaded

environments. When species are compared, plants grown in high light conditions often

have a faster leaf turnover than plants grown in the shaded understory (Bongers and

Pompa 1990). Therefore, it is likely that the average leaf lifespan of a seedling growing

in the forest understory will be longer than that of an adult tree with access to a higher

light environment.

Due to the effect of light environment on leaf lifespan, it is imperative that plants

be grown in similar light environments when comparing leaf lifespans across species or

life forms. Thus, all species in this study were planted in naturally occurring canopy

gaps. Canopy gaps are particularly important for tropical tree regeneration as species are

thought to be differentiated in part by their response to light conditions in gaps (Denslow

1980). This response is partially determined by differences in leaf dynamics among

species as species with shorter leaf lifespans are typically intolerant of shade, while

species with longer leaf lifespans are shade-tolerant (Bongers and Pompa 1990).

Detailed knowledge of species responses to similar gap micro-environments could lead to









an understanding of the degree of differentiation among species in the way they exploit

photosynthetically active radiation. This differentiation is most pronounced during early

stages of the tree life cycle when seedlings or saplings suffer high mortality.

The objectives of this study were: (1) to fill a gap in empirical data on declines of

photosynthetic capacity with leaf age for seedlings of tropical tree species with a range of

leaf lifespans; (2) to evaluate two alternative predictions of cost-benefit theory; and, (3)

to explore the functional basis for relationships among leaf age, leaf position, light

microenvironment, leaf mass per area, nitrogen content, and photosynthetic capacity. We

compared the x intercept of the regression of leaf age with photosynthetic capacity

(parameter b) with the observed leaf lifespan of each species. A relationship between the

parameter b, the x intercept of the regression of leaf age with photosynthetic capacity,

and species leaf lifespan would provide qualitative support for Case 1 in which external

resources do not limit the total number of leaves produced. Alternatively, if plants are

limited by external resources (Case 2), I expected that the rate of photosynthetic decline

would be faster (i.e. a/b would be larger) in species with shorter leaf lifespans according

to Kikuzawa's (1991) predictions. I also predicted that photosynthetic rate, nitrogen per

unit area, and photon flux density (PFD) measured at the leaf surface would decrease

with leaf age.

Methods

Study Site

The study was conducted on Buena Vista Peninsula, a 60-year-old secondary

lowland tropical forest area in the Barro Colorado National Monument (BCNM),

Republic of Panama (909' N, 7951'W). The species composition, climate, and ecology

of the Buena Vista Peninsula is similar to that of the young forests on Barro Colorado









Island (Croat 1978, Leigh 1982). The BCNM forest is semi-deciduous during the

pronounced dry season that usually lasts from mid-December to April, and receives

annual rainfall ca. 2,600 mm (Daws et al. 2002).

Plot Establishment

Three gap plots were established in natural tree fall gaps of approximately 100 m2

in May 2002. Seeds were collected on Barro Colorado Island from at least two parent

trees and germinated in a greenhouse. At radicle emergence, seedlings were transplanted

into three common gardens located natural tree fall gaps on Buena Vista Peninsula. Each

garden was 7 x 7 m2 and enclosed in a hardwire fence to exclude vertebrate herbivores.

Based upon the mean percentage of total daily PFD in the center of the plot relative to

total daily PFD above the canopy, the light environment in each plot was determined to

be 50.7%, 37.7%, 23.4% of full sun in plots 1, 2 and 3, respectively (Kitajima

unpublished data). Newly expanded leaves were marked with unique identification

numbers and monitored for survival during monthly census. Median leaf lifespan was

estimated for each species at the time from full leaf expansion to abscission with the

Kaplan-Meyer method (Donovan and Weldin 2002), which accounts for censored leaves

(leaves removed from the study before their death or alive at the last survival census).

Measurement of Light and Photosynthesis in the Field

I selected eleven species of tropical trees common to BCNM that varied in median

leaf lifespan and shade tolerance for sampling (Table 1-1). Species are hereafter referred

to by their generic names. Median leaf lifespan varied between 140 days in Ceiba and

792 days in Aspidosperma. All species are canopy trees, except for Pentagonia which is

a small tree commonly found in gaps but also encountered in the shaded understory.

Seeds of Aspidosperma, Ceiba, and Vochysia are wind-dispersed, while those of the other









eight species are dispersed by birds and mammals. Ceiba, Pentagonia, and Vochysia are

early successional species while other species are intermediately shade tolerant

(Tabernaemontana, Genipa) or very shade tolerant (Aspidosperma, Calophyllum,

Gustavia, Tetragastris, Trichilia and Virola) based on the abundance and survival of

juveniles in the shaded understory of BCNM (Kitajima personal observation, CTFS data

set available online, S. J. Wright unpublished data).

To include a wide-range of leaf ages for each species, I selected four marked leaves

of contrasting ages on each individual seedling for measurements using the leaf census

data (Table 1-2). For each selected leaf, I measured the rates of light-saturated net CO2

assimilation (=photosynthetic capacity per unit area, Aarea), stomatal conductance (Gs) at

photon flux density (PFD) of 1000 itmol m-2 s-1, and then measured dark respiration with

a portable infra-red gas analyzer (LI-6400, LICOR, Lincoln, Nebraska, USA). Light was

supplied with blue-red light emitting diodes. The CO2 concentration of the reference air

entering the leaf chamber was adjusted with a CO2 mixer control unit so that the

"reference" air entering the chamber had [CO2] = 38 Pa. Sample [CO2] ranged from 36.4

to 38.2 Pa. The chamber temperature was stabilized by maintaining the Peltier block

temperature at 280C. The relative humidity of the sample was kept as close to ambient as

possible (typically between 70-85%) and air flow rate was set at 400 .imol s-1. Gas-

exchange measurements were made between 0800-1200 h between May and August of

2004. Additional leaves were measured between July and September of 2005 to

supplement the age range for species with long-lived leaves.

Total daily PFD at the leaf surface was measured for 3 days for a subset of leaves

representing the full range of leaf age of five species after gas exchange measurements









(Table 1-3). A calibrated GaAsP sensor (Hamamatsu, Japan) was attached to the adaxial

surface of each sampled leaf. Campbell Dataloggers (Models 21X and 10X) sampled

PFD sampled every 2 s and recorded a mean for each minute. The results were expressed

as %PFD (the mean percentage of the total daily PFD above the canopy) for each leaf.

All light measurements were made between May and August of 2004.

Laboratory Measurements and Statistical Analyses

Leaf mass per area (LMA) for each leaf was determined from two 19.6 cm2 leaf

disks dried at 60C until a stable dry mass was reached. Nitrogen content per unit mass

(Nmass) and per unit area (Narea) were determined for these leaf disks with an elemental

analyzer (Costech Analytical Model 4010, Valencia, California, USA). Instantaneous

photosynthetic nitrogen use efficiency (PNUE) was determined by dividing Aarea by Narea.

Statistical analyses, including leaf survival analysis and regression and correlation

analyses among leaf traits, were completed using JMP V.5.1 (SAS, 2003). The

relationships of Aarea, Amass, LMA, Narea, and PNUE with leaf age were modeled with

ANCOVA for each species with leaf age as the covariate and plant as a random effect.

The rate of photosynthetic decline (parameter a/b) was estimated as the slope of the

regression line of leaf age and Aarea or Amass (hereafter leaf age-Aarea or leaf age-Amass).

They and x intercept of the leaf age-Aarea or leaf age-Amass regression was used as an

estimate initial photosynthetic capacity (parameter a) and the leaf age when

photosynthesis would equal zero (parameter b) respectively. Significance of PFD

relationships with leaf age and Aarea were tested with a simple linear regression with

leaves, plants, and plots pooled for each species.









Results

Aarea declined significantly with leaf age in all species (Figure 1-1). The rate of

decline varied among species as indicated by regression slopes of Aarea against leaf age

that ranged from -0.008 to -0.039 [tmol CO2 m-2 s-1 d-1. Amass also declined significantly

with leaf age in all species (Figure 1-2). Regression slopes of Amass against leaf age

ranged from -0.0007 to -0.0001 [tmol CO2 g-2 S-1 d-1. While there was no relationship

between initial Aarea and median leaf lifespan among species, the regression slope of Aarea

against leaf age increased significantly with increasing median leaf lifespan among

species (Figure 1-3a and 1-3c). Although it was still included in statistical analyses, the

slope value for Vochysia was outside of the 95th percentile, and was thus considered an

outlier. It had the steepest slope on both a mass and area basis, but only an intermediate

median leaf lifespan (Figure 1-3a and 1-3c). Initial Amass decreased with median leaf

lifespan, whereas the slope of Amass against leaf age increased with median leaf lifespan

(Figure 1-3b and 1-3d).

In all species, x intercepts were greater than median leaf lifespan by 89 (Trichilia)

to 403 days (Ceiba) when photosynthetic capacity was measured on an area basis (Table

1-4). When photosynthetic capacity was measured on a mass basis, x intercepts

overestimated median leaf lifespan by 4 (Gustavia) to 403 days (Aspidosperma) and

underestimated median leaf lifespan by 30 days in Trichilia (Table 1-4). There was a

positive relationship between median leaf lifespan and the x intercept of species' leaf age-

Aarea and leaf age-Amass regression (Figure 1-4).

There was no consistent relationship within species between LMA and leaf age

(Figure 1-5). In some species, LMA increased with leaf age (Ceiba, Genipa, Gustavia,

Tetragastris, and Trichilia), in others, it decreased (Calophyllum and Pentagonia), while









in Aspidosperma, Tabernaemontana, Virola, and Vochysia there was no relationship.

The percent change in LMA from a 30-day old leaf to a 300-day old leaf ranged between

-5% in Calophyllum and 35% in Genipa. Seven out of the eleven species showed a

decline in Narea with leaf age (Figure 1-6), but PNUE decreased with leaf age in all

species (Figure 1-7).

Light availability declined significantly with leaf age in the five species for light

data were available (Figure 1-8). Aarea was positively correlated with light availability in

three out of five species (Figure 1-9) which also showed a decline in Narea with leaf age

(Figure 1-5). In contrast, Aspidosperma and Genipa did not increase in Aarea with leaf age

although they did decline in Narea with leaf age.

Discussion

This paper provides further support for a negative relationship between leaf age and

photosynthetic capacity across species (Reich et al. 1992, Reich et al. 1999, Wright et al.

2004) while also supplying rare data on the rate of photosynthetic decline with leaf age

which allows for the evaluation of cost-benefit theories of leaf lifespan (Kikuzawa 1991,

Kikuzawa and Ackerly 1999, Kitajima et al. unpublished manuscript). Cost benefit

theories have generated two alternative optimization criteria for predicting leaf lifespan

(Kikuzawa 1991, Kitajima et al. unpublished manuscript). Depending upon whether or

not external resources are limiting, a leaf should either maximize total lifetime carbon

gain (Case 1) or average daily carbon gain (Case 2). In Case 1, leaf lifespan should be

similar to the x intercept of the Aarea-leaf age relationship (parameter b in Kikuzawa's

model). Because total daily net photosynthetic income should reach zero before

photosynthetic capacity reaches zero, the x intercept is expected to be greater than the

actual leaf life time of a given species. In a previous study, estimated x-intercepts









approximated the mean leaf lifespan for seedlings of a tropical pioneer tree with very

short mean leaf lifetimes (Ackerly and Bazzaz 1995; Table 1-5). In contrast, the x-

intercept of the Aarea-leaf age relationship in this study consistently overestimated median

leaf lifespan by 89-403 days (Table 1-4). Kitajima and others (1997) also reported x-

intercepts that were greater than the mean leaf life times of five tropical tree species

(Table 5). Contrary to an earlier study by Kitajima and others (2002), I found no

relationship between the discrepancy in the x intercept and median leaf lifespan with

species-specific median leaf lifespans. Instead, there was a significant positive

relationship between the x intercepts and median leaf lifespans which provides qualitative

support for Kikuzawa's model. However, rates of photosynthetic capacity declined

sharply before reaching the potential leaf lifespan b suggesting that actual leaf lifespan is

much shorter than the potential. Thus, it appears that Case 1 applies only to very short-

lived leaves.

The results of this study supported predictions that leaves will maximize daily

average carbon gain (Case 2) as suggested by Kikuzawa's cost-benefit model (1991) and

two earlier empirical studies (Kitajima et al. 1997, Kitajima et al. 2002). In general,

species with shorter leaf lifespans had (1) a higher initial photosynthetic capacity and (2)

a steeper rate of decline of photosynthetic capacity with leaf age (parameter a/b in

Kikuzawa's model) than species with longer leaf lifespans. Aarea-leaf age slopes were

steeper for the five species with short median leaf lifespans (140-280 days) than the six

shade-tolerant species with long median leaf lifespans (504-797 days; Table 1-4). These

slope values are similar to Aarea-leaf age slopes reported in previous studies (Table 5).

Kitajima et al. (1997) found Aarea-leaf age slopes ranging from -0.032 to -0.0018 [tmol









CO2 m-2 d-1 for adults of five tropical tree species with mean leaf lifespans of 174-315

days. Adults of two tropical pioneer species with very short leaf lifespans (74-93 days)

had much steeper slopes (-0.2 to -0.25 tmol CO2 m-2 d-1; (Kitajima et al. 2002). Even

steeper slopes between -0.57 and -1.32 [tmol CO2 m-2 d1 have been documented for

seedlings of Heliocarpus appendiculatus, a tropical pioneer tree with mean leaf life times

of only 28-37 days (Ackerly and Bazzaz 1995).

Although a general dichotomy between Aarea-leaf age slopes of species with short

and long-lived leaves is evident, some species do not fit the predicted pattern. For

instance, Vochysia had both the steepest rate of photosynthetic decline with leaf age and

the highest initial Aarea, yet had the longest median leaf lifespan (280 days) of the pioneer

species studied. The species with the longest median leaf lifespan (797 days)

Aspidosperma, had the shallowest slope, but not the lowest initial Aarea as Kikuzawa's

cost-benefit analysis would predict. Nevertheless, this study still provides strong support

for Case 2 which provides leaf lifetime optimization criteria for plants of various leaf

lifespans that are primarily limited by external resources.

Species differences weakened the among species relationship between Amass and

leaf lifespan (Figure 1-3b, r2= 0.56, p = 0.009). Previous studies have documented

stronger correlations (r2 values typically between 0.70-0.91, with p-values < 0.001; Reich

et al. 1992, Reich et al. 2004, Wright et al. 2004). I believe this can be explained by the

greater range in leaf lifespans as well as the greater number of species measured in

previous studies. Reich et al. (1992) reported leaf lifespans up to 4200 days, while the

maximum leaf lifespan in my study was 792 days. This five-fold greater range in leaf

lifespans could result in stronger correlations, particularly when coupled with a greater









number of species for comparison. The absence of a significant relationship between

initial Aarea and leaf lifespan among species has been frequently cited in earlier studies

(Reich et al. 1992, Reich et al. 2004, Wright et al. 2004) and is due to the effect of LMA

on photosynthetic capacity (Peterson 1999). When LMA is controlled statistically, the

slope of the relationship between photosynthesis and leaf nitrogen is the same for both

area and mass-based expressions (Peterson 1999).

Aarea, Amass, Narea, and PFD decreased with leaf age as predicted for all species

measured while LMA did not vary consistently with leaf age across species. In most

species, the changes in LMA with leaf age were so small that they could not have

contributed much to the decrease of Aarea, Amass, and Narea. In contrast, the observed

decline in Narea with leaf age did contribute to the decline in photosynthetic capacity in

seven out of the eleven species and was likely caused by the reallocation of nitrogen to

newer leaves where PFD was greater. Indeed, due to shading by younger leaves, leaf

light levels decreased with leaf age for the five species for which leaf PFD was measured

in this study. This pattern is consistent with theories of optimal resource allocation in

which nitrogen from older shaded leaves is reallocated to younger unshaded leaves,

resulting in a more favorable carbon gain for the entire plant (Field and Mooney 1983,

Hirose and Werger 1987, Hikosaka et al. 1994). In all species, the decrease in PNUE

with leaf age also contributed to the decline of photosynthetic capacity. The retention of

leaves despite their low photosynthetic capacity and PNUE has been explained as a

means of increasing total production due to the longer retention of plant nitrogen

(Escudero and Mediavilla 2003).









This study is apparently the first to evaluate whether cost-benefit theory explains

variation in leaf lifespan in several species of seedlings ranging in shade tolerance and

median leaf lifespan. I have no data on the leaf lifespan of adult life forms. However,

sapling leaf lifespans were measured for three of the eleven species I studied (Coley

1988, King 2003). Gap-grown saplings of Tetragastris and Calophyllum 1-2 years of age

had mean leaf lifespans of 690 and 1050 days respectively (Coley 1988). This value is

only a little greater than the median leaf lifespan of 671 days for Tetragastris reported in

my study. For Calophyllum, 1050 days is much greater than the median leaf lifespan of

700 days reported in this study. King (1993) evaluated the mean leaf lifespans of

Trichilia and Calophyllum saplings grown over a range of light environments and found

mean leaf lifespans of 1350 and 2640 days, respectively, which is much greater than the

values of 630 and 700 days reported in this study. However, since King (1993) evaluated

mean leaf lifespan in saplings across gap, understory and intermediate light conditions, it

is difficult to account for the effect of light on leaf lifespan. Leaf lifespan is expected to

be greater in shaded than in gap conditions (Bongers and Pompa 1990). For all three

species in which sapling leaf lifespan was measured, leaf lifespan was greater in saplings

than in seedlings, providing support for the idea that leaf lifespan should increase with

increasing construction and maintenance costs of supporting structures (Kikuzawa and

Ackerly 1999). In order to evaluate this hypothesis further, sapling and adult leaf

lifespans should be measured for the eleven species in this study under similar light

conditions.

In summary, general predictions made by Kikuzawa's cost-benefit model (1991)

were supported by the results of this study. Leaf lifespan was shorter for species with









higher initial photosynthetic rates and longer when the rate of photosynthetic decline with

leaf age was slow. The x-intercept of the Aarea-leaf age relationship consistently

overestimated median leaf lifespan, indicating that the observed leaf lifespan was much

shorter than the potential. Thus, it appears the optimal strategy for the seedlings of

tropical tree species in this study is to maximize daily carbon gain as external resources

are limiting (Case 2). Leaf age was a strong determinant of photosynthetic capacity,

supporting the idea that knowledge of the distribution of leaf age in a canopy will allow

for estimates of canopy photosynthetic performance.










Table 1-1. Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer
method for each species studied. Species are ranked from shortest median
leaf lifespan to longest median leaf lifespan. Nomenclature follows Croat
(1978).


Soecies


Ceiba pentandra
Genipa americana
Pentagonia macrophylla
Tabernaemontana arborea
Vochysia ferruginea
Virola surinamensis
Gustavia superba
Trichilia tuberculata
Tetragastris panamensis
Calophyllum longifolium
Aspidosperma cruenta


Family


Total
Leaves


+ + F


Bombacaceae
Rubiaceae
Rubiaceae
Apocynaceae
Vochysiaceae
Myristicaceae
Lecythidacaceae
Meliaceae
Burseraceae
Clusiaceae
Apocynaceae


Table 1-2. Number of plants and leaves measured for Amax, Aarea, LMA, Narea, and PNUE
measurements for each species. Species are ranked from shortest median leaf
lifespan to longest median leaf lifespan.
# of Leaves

Species P A area Amass LMA Narea PNUE
Plants


Ceiba
Genipa
Pentagonia
Tabernaemontana
Vochysia
Virola
Gustavia
Trichilia
Tetragasteris
Calophyllum
Aspidosperma


28
43
31
32
42
105
44
27
46
42
69


28
43
31
32
42
51
32
27
30
24
38


28
43
31
32
42
78
39
36
38
41
49


Median


Mean


156.1
234.8
274.2
277.6
280.3
494.3
591.2
626.1
648.3
679.9
745.3


Std
Dev


3.10
4.96
6.03
3.94
3.41
6.56
17.17
20.12
13.87
12.79
12.51







19



Table 1-3. Number of plants and leaves for which % PFD and Aarea were measured for a
subset of species. Species are ranked from shortest median leaf lifespan to
longest median leaf lifespan.
# of Leaves
# of % PFD
Species Plats %PFD ad ..
Plants and Aarea
Genipa 5 19 9
Tabernaemontana 2 10 8
Virola 4 17 13
Tetragasteris 4 17 15
Aspidosperma 2 8 8









Table 1-4. Mean leaf lifespan and regression statistics (x and y intercepts and slopes) for area- and mass-based light-saturated net
photosynthesis against leaf age (days after full expansion) for seedlings of eleven tropical tree species.
Median Aarea (jmol m-2 S-1) Amass (jmol g-1 s-1)
Species Leaf x intercept y intercept Slope x intercept Slope
intercept
Lifespan (b) (a) ( ') (b) (a)nterce ( ')
Ceiba 140 543 11.4 -0.021 443 0.31 -0.0007
Genipa 224 512 8.7 -0.017 400 0.20 -0.0005
Pentagonia 252 525 8.4 -0.016 700 0.14 -0.0002
Tabernaemontana 279 489 9.3 -0.019 433 0.26 -0.0006
Vochysia 280 369 14.4 -0.039 400 0.28 -0.0007
Virola 504 730 7.3 -0.010 750 0.15 -0.0002
Gustavia 629 882 9.7 -0.011 633 0.19 -0.0003
Trichilia 630 800 8.0 -0.010 600 0.12 -0.0002
Tetragastris 671 888 7.1 -0.008 700 0.14 -0.0002
Calophyllum 700 887 13.3 -0.015 750 0.15 -0.0002
Aspidosperma 797 1125 9.0 -0.008 1200 0.12 -0.0001

Table 1-5. Mean leaf lifespan and regression statistics (x and y intercepts and slopes) for light-saturated net photosynthesis (Aarea in
micromols of CO2 m-2 S-1) against leaf age (days after full expansion) for eight tropical tree species measured in previous
studies (Ackerly and Bazazz 1995, Kitajima et al 1997, 2002). Since measurements were made for a range of nutrient and
light conditions for Heliocarpus appendiculatus, average values for each parameter are reported. Species are arranged from
shortest to longest mean leaf lifespan.
Species Mean Leaf x intercept y intercept Slope Source
S Lifespan (b) (a) (, ',)
Heliocarpus appendiculatus 33 36 14.6 -0.915 Ackerly and Bazzaz 1995
Cecropia longipes 74 133 23.7 -0.179 Kitajima et al. 2002
Urera caracasana 93 98 21.2 -0.255 Kitajima et al. 2002
Antirrhoea trichantha 174 550 12.1 -0.022 Kitajima et al. 1997
Luehea seemannii 201 1200 10.8 -0.009 Kitajima et al. 1997
Castilla elastica 206 365 11.7 -0.032 Kitajima et al. 1997
Annona spraguei 221 482 10.6 -0.022 Kitajima et al. 1997
Anacardium excelsum 315 985 6.9 -0.007 Kitajima et al. 1997


















Ceiba
y= -0.021x + 11.43
15 r = 0.56, P = 0.0012





5-


20 -
Genipa
y = -0.017x + 8.66
15 r= 0.63, P < 0.0001









Pentagonia
y =-0.016x + 8.40
15 = 0.63, P < 0.0001


10- %


















y = -0.039x + 14.35
S.



















15- r2 0.76, P < 0.0001"




5- go
y=-0.019x+9.31
15 = 0.78, P< 0.0001


















5 ----------------------------


0 200 400 600

Leaf Age (days)


800 1000


Virola
y= -0.01 Ox + 7.33
15 r2 = 0.77, P < 0.0001


10






Gustavia
y = -0.01 lx + 9.71
15 = 0.70, P < 0.0001
*

10 A


5


20
Trichilia
y = -0.01 Ox + 7.99
15 r2 = 0.38, P = 0.0009


10 -


5 C h *aopy

20 -
Tetragastris
y = -0.008x +7.08
15 r2 = 0.64, P < 0.0001


10 *











5- A A



20 --- A
Aspidosperma
y = -0.0108x +13 8.9833
15 r2= 0.8571, P < 0.0001
10 A' A A



5

0


Leaf Age (days)

Figure 1-1. Light saturated photosynthetic assimilation rates (Aarea) of leaves of different

age on seedlings of eleven tropical tree species. Each point represents a leaf.

Circles were leaves measured in 2004; triangles were leaves measured in

2005.



















Ceiba
0 4 y = -0.0007x + 0.31
r = 0.46, P= 0.0005
03 -<

02-

01


Genipa

04 Y = -0.0005x + 0.20
r2 = 0.72, P < 0.0001

03 -

0.3
02 S



01 0
05 --
Pentagonia
04 y = -0.0002x + 0.14
r2 = 0.70, P < 0.0001
03 -

02

v01 0- S:----.- *


05 0
STabernaemontana

04 y = -0.0006x + 0.26

03 -

02 -o



05 .
Vochysia
04- y=-0.0007x + 0.28
r2= 0.76, P < 0.0001
03-




01


0 200 400o 6o00 800

Leaf Age (days)


Virola
4 y = -0.0002x + 0.15
r2 = 0.70, P < 0.0001
0 3

02

0 1 -

05
Gustavia
04 y = -0.0003x + 0.19
S= 0.82, P < 0.0001
03

02 S

01 --



Trichilia
04 y = -0.0002x + 0.12
= 0.48, P= 0.0002
0 3 -

0 2 -

01 -


Tetragastris
04 y = -0.0002x + 0.14
r2 = 0.88, P < 0.0001
03

0 2




05
Calophyllum
04 y= -0.0002x + 0.15
r2 = 0.87, P < 0.0001
03

02 *

01 O T

05
Aspidosperma
04 y= -0.0001x + 0.12
= 0.57, P = 0.0001
0 3 -

0 2 -


6
0 1 W0---- i -

0 200 400 600


Leaf Age (days)

Figure 1-2. Mass-based photosynthesis (Amass) of leaves with contrasting leaf age in

eleven species of tropical trees. Each point represents a leaf













16 04

14 a b
14

0 12 03


E 04 Vochysa o
S05
<{ 6 < E

E 4- 01

2

p 4 00 60r 0 04 p 200 400 60088 8r2 0 55


r--001 -_6

'E 00
o-0020 0


-003


o 004- 0 V 0 0 -Vochysa
p = 0 0334, r2 0 41 p 00093 r2 055
-0051 -p 09 r 055
0 200 400 600 800 _3? 0 200 400 600 800

Median leaf lifespan (days)
Figure 1-3. Relationship between initial Aarea, A,,ea-leaf age slope, Amass, Am.s,-leaf age
slope with median leaf lifespan for each species. (a) Relationship between
initial Aarea and median leaf lifespan for each species. (b) Relationship
between the Aarea-leaf age slope and the median leaf lifespan for each species.
(c) Relationship initial Amass and median leaf lifespan for each species. (d)
Relationship between the Amass-leaf age slope and median leaf lifespan for
each species.












1400


1200 -


1000 -


800 -


600 -


400 -


100 200 300 400 500 600 700 800


Median Leaf Lifespan (days)
Figure 1-4. The relationship between the actual median leaf lifespan and an estimate of
the time when photosynthetic capacity would reach zero (parameter b, the x
intercept of the leaf age-Aarea and leaf age-Amass regression). Solid circles and
the solid line represent the relationship when photosynthetic capacity is
presented on an area basis (r2 = 0.87, P < 0.0001), while empty circles and the
dashed line represent the relationship when photosynthetic capacity is
presented on a mass basis (r2 = 0.58, P < 0.01).


y = 0.9x + 271.2
y=0.8x+283.4


















140 -
Ceiba
120- y = 0.029x + 38.00

100 2 = 0.86, P = 0.0454

80 -

60 -

40

20 -

140 -
Genipa
120 y = 0.056x + 40.8

100 r= 0.84, P < 0.0001

80 -


40 -

20 -

140 -
Pentagonia
120 y = -0.036x + 60.34

100- 2 = 0.40, P = 0.0031

80 -


40 0

20 -

140 -
Tabernaemontana
120 y = 0.005x + 36.65

100- r2= 0.19, P= 0.5243

80 -

60 -


20 -

140 -
Vochysia
120 y = -0.020x + 51.52

100 -2 = 0.57, P = 0.0577

80 -

60 -

40 :

20 -
0 ,-,


200 400

Leaf Age (days)


600 800


140 -
Virola
120 y = -0.0008x + 53.34

100 = 0.42, P = 0.7944

80 0







Gustavia
120 y = 0.020x + 51.99

100 r2 = 0.34 P = 0.0052

80-


40-

20
140 -
Trichilia
120 y = 0.013x + 68.69

100 = 0.67, P = 0.0147


60 -

40-
20
140 -

Tetragastris
120 y = 0.010x + 57.12
100 = 0.60, P = 0.0246

80 *


40-

20

140

120

100 L AO A
8 AA

60 -

40 Calophyllum

20 y = -0.016x + 98.46
r2= 0.39, P= 0.0105
140

120

** A. A
80 ... **..u .W 8 I .. A.... .,-s. .
60- *

40 Aspidosperma

20 y = 0.007x + 78.55
r2 = 0.54, P= 0.1935


0 200 400 600

Leaf Age (days)


Figure 1-5. Relationship between leaf mass per unit area (g m-2) and leaf age in seedlings

of eleven species of tropical trees. Each point represents a leaf. Circles were

leaves measured in 2004; triangles were leaves measured in 2005. Significant

and nonsignificant regression lines are indicated by solid and broken lines,

respectively.


1000


















Ceiba
20 y = -0.0002x + 1.18
r2 = 0.66, P = 0.6868

15

10 -

05-

25-
Genipa
20 = 0.0001x + 0.82
S= 0.69, P = 0.6559
15 -

10 I se

05-

25-
Pentagonia
20 y = -0.001x + 1.02
r = 0.43, P = 0.0027
15 -




05 *

25- Vochysia
Tabernaemontana
20 Y = -0.001x + 1.12
r2 = 0.47, P <= 0.0005
15

10 lim

05-


25 -Vochysia

20y = -0.002x + 1.56
r= 0.63, P < 0.0001





05 -

00 -


0 200 400 600
Leaf Age (days)


800 1000


Virola
20 y = -0.0005x + 1.01
r = 0.44, P < 0.0001
15 *
10 A




25



15 *
20 4P A A

A A
15 ... .. .....*.. .......................a



Gustavia
0o5y = -0.0001x + 1.57
25 -r2 = 0.40 P = 0.5379
Trichilia
20 y = -0.0006x + 1.49
r2 = 0.54, P = 0.0004





05

25 -
Tetragastris
20 y = -0.0003x + 1.01
r = 0.24, P = 0.0604
15- 0

10 AAA A
..........

05 o-

25
Calophyllum
20 y = -0.0007x + 1.35
r= 0.62, P < 0.0001
15 A




05 -


Aspidosperma
y = -0.0002x + 1.32
*. r2 = 0.72, P= 0.0461


-- -t ^ -- -
10
10 ~ A *S A A A

05

0 0 80 1 0
0 200 400 600 800 1000


Leaf Age (days)

Figure 1-6. Relationship between leaf nitrogen (g m-2) and leaf age. Each point

represents a leaf. Circles were leaves measured in 2004; triangles were leaves

measured in 2005. Significant and nonsignificant regression lines are

indicated by solid and broken lines, respectively.

















14 14 -
Ceiba
12- y = -0.019x + 10.03 12-

10o *| ? = 0.42, P= 0.0005 10 -

8 8-






14- 14
Genipa
12 y= -0.020x + 10.37 12-
10 r2 = 0.65, P < 0.0001 0 -



3 6 0
4 4-
2 -
14- 14
Pentaqonia
12 y =-0.012x + 8.50 12-

o10 = 0.72, P < 0.0001 10 -



*2


10- = 0. P < 0.0001 10 -








4- 4 600 4-

2 --
14- 14-
Tabernaemontana
12 y =-0.015x + 9.04 12-
o10- r = 0.66, P < 0.0001 10-

8- 8-
6 6

4- -f < 4 -
2 S 2
2 12-
14L a 14
Vochysia
12. y = -0.020x + 9.69 12-
10 r2 = 0.54, P < 0.0001 1o-




S



0 14
0 200 400 600 800 1000
12-
Leaf Age (days)


Virola
y = -0.010x + 7.73
r = 0.62, P < 0.0001


* y





Gustavia
y = -0.008x + 6.60
= 0.68, P < 0.0001









Trichilia
y = -0.007x + 5.64
r = 0.33, P = 0.0022









Tetragastris
y =-0.012x + 7.90
*= 0.81, P < 0.0001









*. *, Calophyllum
y = -0.010x + 10.75
r = 0.86, P < 0.0001


II


S
0g.. .1


S gO S
6
S


0 200 400 600 800 1000


Leaf Age (days)

Figure 1-7. Relationship between photosynthetic nitrogen use efficiency (PNUE) and leaf

age. Each point represents a leaf


Aspidosperma
y =-0.005x + 7.10
r2 = 0.74, P = 0.0011














50


40 -


30-


20 -


10 -


50


40 -


30 -


C 20-
0-
U0

10 -


0


600 800


Genipa
y = -0.122x + 29.93
* r2 = 0.380, P= 0.005C









o




Tabernaemontana
y = -0.024x + 8.58
r2 = 0.549, P = 0.0355











.3


Leaf Age (Days)
Figure 1-8. Decline in %PFD (the total daily PFD incident on the leaf relative to the total
daily PFD above the canopy) with leaf age in a subset of species and leaves.
Each point represents a leaf


400

Leaf Age (Days)


Virola
y= -0.016x + 10.27
r2 = 0.505, P = 0.0065














Tetragasteris
y = -0.009x + 6.49
r2 = 0.396, P= 0.0159














Aspidosperma
y = -0.072x + 31.24
r2 = 0.810, P= 0.0023



*











0 200 400 600 800


0 200














14

12

10

8

6

4

2

14

12




E 8
0

S6

6 4


0 10 20

% PFD


0 10 20 30 40

% PFD
Figure 1-9. Effect of % PFD (the total daily PFD incident on the leaf relative to the total
daily PFD above the canopy) on light-saturated photosynthesis (Aarea) and for
a subset of species and leaves. Each point represents a leaf. Significant and
nonsignificant regression lines are indicated by solid and broken lines,
respectively.


Genipa
r2= 0.0, P= 0.E
*
*














Tabernaemontana
y = 0.490x + 4.93
r2 = 0.668, P= 0.0145




* *.


30 40














CHAPTER 2
LEAF AGE, PHOTOSYNTHETIC CAPACITY, AND LEAF LIGHT LEVEL AS
PREDICTORS OF DAILY CARBON GAIN IN SEEDLINGS OF FIVE TROPICAL
TREE SPECIES

Introduction

Cost-benefit theories of leaf lifespan are a powerful approach that explains global

patterns of leaf lifespan across latitudes (Kikuzawa 1991, Kikuzawa and Ackerly 1999)

and the negative correlation between leaf lifespan and photosynthetic capacity (Reich et

al. 1997, Reich et al. 1999, Wright et al. 2004). A basic assumption central to cost-

benefit theory is that leaves, as the primary carbon gaining organs of plants, must

optimize photosynthetic gain (Aday) relative to construction and maintenance costs

(Chabot and Hicks 1982). Because photosynthetic characteristics change with leaf age,

leaf-age effects are an important consideration in cost-benefit theories of leaf lifespan.

Understanding the relationship between photosynthetic gain and leaf age can facilitate the

integration of photosynthetic carbon gain from individual leaves to individual crowns to

the entire forest canopy (Kitajima et al. 1997, Kitajima et al. 2002). Because Aday is

difficult to measure directly for many leaves, previous studies have focused on the

relationship between photosynthetic capacity and leaf age, assuming that Aday can be

approximated by maximum photosynthetic capacity, or Amax (Kitajima et al. 1997,

Kitajima et al. 2002, Kitajima et al. unpublished manuscript, Chapter 1). This

assumption is based upon a strong correlation between Amax and Aday reported across

several tropical canopy tree and epiphyte species (Zotz and Winter 1993, 1994).

However, the degree to which Aday, not Amax, varies with leaf environmental and









physiological factors has not been adequately explored. It is particularly important to

elucidate the relationship between Aday and Amax as predictions made using cost-benefit

theory are based on Aday, not Amax (Kikuzawa 1991).

Photosynthetic capacity (Amax) and daily (24 hour) photosynthetic gain (Aday)

typically decline monotonically with leaf age and can be approximated by two similar

equations that differ only by the notation which indicates daily rather than instantaneous

parameters. The equations are: Amax = a (1 t/b) or Aday = a' (1 t/b '), where t is the

leaf age in days from the time of full leaf expansion (Kikuzawa 1991). The parameters a

and a' represent the photosynthetic rate at the time of full leaf expansion. The parameters

b and b' is a statistical extrapolation determined as a function of the initial photosynthetic

rate (a or a') and the rate of its decline (parameter a/b or a'/b'). Thus, parameter b is an

estimate of the time when photosynthetic capacity would reach zero. Studies exploring

the relationship between leaf age and photosynthetic capacity have reported a tight

relationship between leaf lifespan and the parameter a/b indicating that it is possible to

estimate the effect of leaf age on photosynthetic gain if the mean leaf lifespan of the

species is known (Kitajima et al. 1997, Kitajima et al. 2002, Kitajima et al. unpublished

manuscript, Chapter 1).

Based upon the approximately linear decline of photosynthesis with leaf age

(Kitajima et al. 1997, Kitajima et al. 2002, Chapter 1), cost-benefit analyses of leaf

lifespan have generated two alternate predictions. If there are no external constraints on

the maximum number of leaves produced and maintained (Case 1), a leaf is expected to

maximize net carbon gain over its entire lifetime (Kikuzawa 1991, Kitajima et al.

unpublished manuscript). In this case, the optimum leaf lifespan should be close to the









time when photosynthetic rates equal zero (parameter b). Alternatively, if the total

number of leaves is limited by external resources (Case 2), a leaf is expected to maximize

the rate of net carbon gain averaged over the lifetime of the leaf (Kikuzawa 1991,

Kikuzawa and Ackerly 1999). In this scenario, leaves are replaced when their net carbon

gain per unit time over their entire lifespan has reached a maximum and optimal leaf

lifespan is predicted to be equal to (2bC/a)1/2 where C is the construction cost of the leaf

(Kikuzawa 1991). Based on this theoretical model, it is expected that leaf lifespan will

be short when the initial photosynthetic rate of the leaf is high, and long when

construction costs are high or the rate of decline of photosynthetic rates with leaf age is

slow (Kikuzawa 1991).

Daily photosynthetic gain can be affected by both environmental and physiological

factors, some of which are directly correlated with leaf ageing. Environmental factors

that affect daily photosynthetic gain can vary unpredictably over short time periods,

while other factors change predictably over the lifetime of a leaf. For instance,

photosynthetic gain is expected to vary in response to daily changes in weather, with

decreases in carbon gain during cloudy conditions. Alternatively, over the course of a

leaf's lifetime, a leaf can become increasingly shaded as new leaves are produced at

higher positions on the plant (Field 1983, Ackerly and Bazzaz 1995, Ackerly 1999). A

reduction in light levels experienced by the leaf, whether due to self-shading, shading by

surrounding vegetation, or weather, contributes to a decrease in photosynthetic gain as

lower light levels directly reduce photosynthetic rates (Field 1983).

Leaf physiological factors such as photosynthetic capacity and leaf nitrogen also

have a substantial impact on daily leaf carbon gain and generally decrease with leaf life









span across plant species and communities (Reich et al. 1991, Reich et al. 1992, Reich et

al. 1999). Decreases in Narea directly contribute to decreases in Amass and subsequent

decreases in daily carbon gain as leaves age. The decline of Amass and Nmass with leaf age

is thought to be due to the retranslocation of nitrogen to younger leaves which maximizes

whole-shoot photosynthetic gain rather than uncontrolled physiological deterioration

(Field and Mooney 1983, Ackerly and Bazzaz 1995). Hikosaka et al. (1994) showed that

age-related changes due to nitrogen retranslocation occurred primarily when light or

nitrogen availability is limiting.

The primary objective of this study was to examine how physiological and

environmental factors may predict daily carbon gain in seedlings of five tropical tree

species that differ in median leaf lifespan and shade tolerance. Since mortality rates are

typically highest in the seed and seedling life stages, understanding how seedling leaf

carbon gain varies with physiological and environmental factors can be helpful in

understanding adult abundance and distribution. In particular, I was interested in

determining whether changes in daily carbon gain were more highly correlated with Amax

or photon flux density (PFD) at the leaf level. While some studies reported a high

correlation between Amax and diel carbon gain (Zotz and Winter 1993, 1994), others have

cited leaf light levels as the determinant of photosynthetic capacity and resulting carbon

gain (Chazdon 1986, Chazdon and Pearcy 1986) Previous studies have documented a

strong correlation between leaf age, leaf PFD, and Amax (Kitajima et al. 1997, Kitajima et

al. 2002, Chapter 1). Thus, I hypothesized that leaf age would be highly correlated with

daily carbon gain since it directly affects both photosynthetic capacity and leaf PFD

(Chapter 1), which in turn directly affect carbon gain (Figure 2-1). Given that leaf PFD









has a strong and documented affect on Amax (Field 1983), and both variables affect

carbon gain, I hypothesized that leaf PFD would be the best predictor of daily carbon

gain.

Methods

Study Site

The study was conducted on Buena Vista Peninsula, a 60-year-old secondary

lowland tropical forest area in the Barro Colorado National Monument (BCNM),

Republic of Panama (9'09' N, 7951'W). The species composition, climate, and ecology

of the Buena Vista Peninsula are similar to that of the young forests on Barro Colorado

Island (Croat 1978, Leigh 1982). The BCNM forest is semi-deciduous during the

pronounced dry season that usually lasts from mid-December to April, and receives

annual rainfall ca. 2,600 mm (Daws et al. 2002).

Plot Establishment

Common gardens were established in three recent -100 m2 tree fall gaps in May

2002 on Buena Vista Peninsula. Each garden was 7 x 7 m2 and enclosed in a hardwire

fence with 1 cm mesh to exclude vertebrate herbivores. Seeds were collected on Barro

Colorado Island from at least two parent trees and germinated in a greenhouse. Seedlings

were transplanted at least 20cm apart at radicle emergence. Seedlings received water

only from rainfall. Average daily rainfall from May to August of 2004 was 11.9 mm day-

1 (May = 15.2 mm day-1, June= 10.7 mm day-1, July= 8.8 mm day-1, August = 13.4 mm

day-1; S. Patton, data available online). The total daily PFD in the center of plots 1, 2,

and 3, were 50.7%, 37.7%, and 23.4% of the light above the canopy, respectively

(Kitajima unpublished data). Difference in light environment reflected differences in the

size of the gaps. Understory vegetation and small stems were removed prior to









transplanting seedlings. Newly expanded leaves were marked with unique identification

numbers with a permanent marker and monitored for survival during monthly census.

Median leaf lifespan was estimated for each species at the time from full leaf expansion

to abscission with the Kaplan-Meyer method (Donovan and Weldin 2002), which

accounts for censored leaves (leaves removed from the study before their death or alive at

the last survival census).

Measurement of Light and Photosynthesis in the Field

For this study, I selected five species of canopy trees common to BCNM that varied

in median leaf lifespan and shade tolerance (Table 2-1; species are hereafter referred to

by their generic names). Median leaf lifespan varied from 478 days in Genipa to 792

days in Aspidosperma. Seeds of Aspidosperma are wind-dispersed, while those of the

other four species are dispersed by birds and mammals. Tabernaemontana and Genipa

are intermediately shade tolerant while Aspidosperma, Calophyllum, Gustavia,

Tetragastris and Virola are very shade tolerant based on the abundance and survival of

juveniles in the shaded understory of BCNM (Kitajima, personal observation, CTFS data

set available online, S. J. Wright unpublished data).

To include a wide-range of leaf ages for each species, I selected four marked leaves

of contrasting ages on each individual seedling using the leaf census data. For each

selected leaf, I measured the rates of net CO2 assimilation (=photosynthetic capacity per

unit area, Aarea), and stomatal conductance (Gs) at photon flux density (PFD) of 1000

mrnol m-2 s-1 supplied with blue-red emitting diodes, and then measured dark respiration

with a portable infra-red gas analyzer (LI-6400, LICOR, Lincoln, Nebraska, USA). The

CO2 concentration of the reference air entering the leaf chamber was adjusted with a CO2

mixer control unit so that the "reference" air entering the chamber had [CO2] = 38 Pa.









Sample [CO2] ranged from 36.4 to 38.2 Pa. The chamber temperature was controlled by

maintaining the Peltier block temperature at 280C. The relative humidity of the sample

was kept as close to ambient as possible (typically between 70-85%) and air flow rate

was set to 400 imol s-1. Gas-exchange measurements were made between 0800-1200 h

between May and August of 2004. Additional leaves were measured between July and

September of 2005 to supplement the age range for species with long-lived leaves.

Photosynthetic light response curves were measured for two of the four leaves

selected for gas exchange measurements using a portable infra-red gas analyzer (LI-6400,

LICOR, Lincoln, Nebraska, USA). Leaves were selected to represent a range of leaf

ages. Light supplied with blue-red emitting diodes was decreased in steps from 1000 to 0

.mol photons m-2 s1 after CO2 uptake rates reached a steady-state at each light level.

After gas exchange measurements, total daily PFD at the leaf surface was measured

for 3 days between May and August of 2004 for a subset of leaves representing the full

range of leaf ages (Table 2-2). A calibrated GaAsP sensor (Hamamatsu, Japan) was

attached to the adaxial surface of each sampled leaf. Campbell Dataloggers (Models 21X

and 10X) sampled PFD every 2 s and recorded a mean for each minute. Results were

expressed as mmol PFD m-2 day The % PFD (mean percentage of the total daily PFD

above the canopy) for each leaf was also presented (Figure 2-2).

Calculation of Daily Carbon Gain

Photosynthetic light curves were fitted to non-rectangular hyperbola (Johnson and

Thornley 1984) using Photosyn Assistant (Version 1.1.2, Dundee Scientific, Dundee,

UK). Best-fit estimates were made of the maximum rate of light-saturated net CO2

assimilation (= maximum photosynthetic capacity per unit area, Amax) as well as quantum

yield (P), curvature (0), and dark respiration (R).









I estimated net photosynthetic carbon gain for each minute during the daytime

(0630-1800 h) using the following equation:

P(I) = {(Q/+ Amax {(QDI+ Amax)2 40(Amaxl}5 / 20} R

where P(I) is the instantaneous photosynthetic rate as a function of the apparent quantum

yield (0), I is the photon flux density incident on the leaf averaged over one minute, Amax

is the light-saturated photosynthetic capacity per unit area, 0 is the curvature of the non-

rectangular hyperbola, and R is area-based daytime respiration (Johnson and Thomley

1984). Carbon gain was estimated for each minute for which PFD was measured, then

averaged over three days to calculate average diurnal carbon assimilation (Table 2-4).

Nocturnal respiration was estimated to be 10% of diurnal carbon gain (Zotz and Winter

1993). The estimated Aday does not include changes in respiration rates or stomatal

closure.

For a subset of leaves that lacked light curve measurements, species averages of

curvature and apparent quantum yield and leaf-specific measurements of Amax and dark

respiration were fitted to non-rectangular hyperbola using Photosyn Assistant. This

provided a best-fit estimate of Amax for each leaf. Due to the low [CO2] differential, large

measurements errors were associated with individual dark respiration rates. Thus, dark

respiration rates were calculated from the regression between leaf age and dark

respiration in all species except Genipa. The species-specific average of dark respiration

was used for Genipa because it showed no relationship between leaf age and dark

respiration.

Statistical Analyses

Statistical analyses were carried out using JMP V.5.1 (SAS, 2003). Significance of

the relationship of average daily carbon gain with leaf age, Amax, and leaf PFD was tested









using a linear regression with leaves, plants, and plots pooled for each species. A

multiple regression was used to determine whether Amax or leaf PFD better predicted

daily carbon gain with leaves, plants, and plots pooled for each species. A simple linear

regression was used to determine the relationship between daily carbon gain, Amax, and

leaf PFD with all measurements and species pooled.

Results

Plants experienced light-limiting conditions primarily due to self-shading and

weather, although shading by surrounding plants might have contributed to light-

limitation to an unknown extent (C. Stefanescu, personal observation). Depending on

their age, leaves spent 12.9 to 51.6% of the total time they were exposed to day light

below their LCP (Figure 2-3).

Average daily carbon gain of individual leaves varied within and among species.

The lowest value observed was -1.8 mmol CO2 day-1 for a very old leaf of Tetragastris,

whereas the highest was 216.4 mmol CO2 day' for a young leaf of Genipa. The

relationship between average daily carbon gain and leaf age was significant in

Tabernaemontana (r2 = 0.5, P = 0.05), Virola (r2 = 0.5, P = 0.009), Tetragastris (r2 = 0.5,

P = 0.003), and Aspidosperma (r2 = 0.8, P = 0.002), but not in Genipa (r2 = 0.2, P = 0.2;

Figure 2-4). Average daily carbon gain was also positively correlated with Amax in

Tabernaemontana (r2 = 0.5, P = 0.05), Virola (r2 = 0.7, P = 0.0006), and Tetragastris (r2

= 0.6, P = 0.001), but not in Genipa (r2 = 0.1, P = 0.4) or Aspidosperma (r2 = 0.1, P = 0.5;

Figure 2-5). Average daily carbon gain was positively correlated with leaf PFD in all

species; Genipa (r2 = 0.8, P = 0.0005), Tabernaemontana (r2 = 0.9, P = 0.0004), Virola

(r2 = 0.9, P < 0.0001), Tetragastris (r2 = 0.9, P < 0.0001), and Aspidosperma (r2 = 0.9, P

= 0.002; Figure 2-6). Amax was a poor predictor of daily carbon gain in all species except









for Virola (P < 0.005; Table 2-3). Leaf PFD was a better predictor of daily carbon gain

that Amax in all species (Table 2-3). When all leaves of all species were analyzed

together, both Amax and leaf PFD were highly correlated with daily carbon gain (Figure 2-

7).

Discussion

Due to the documented affect of leaf age on both photosynthetic capacity and leaf

PFD for these species (Chapter 1), it is not surprising that leaf age was correlated with

average daily carbon gain in four out of the five species in this study. In Aspidosperma,

the lack of relationship between leaf age and daily carbon gain may reflect the limited

range of leaf age sampled relative to the actual leaf lifespan of this species; the median

leaf lifespan for Aspidosperma is 797 days, but only leaves between 37 and 365 days old

were measured. Leaf PFD was strongly correlated with daily carbon gain in all species.

In contrast, Amax was correlated with daily carbon gain in only three out of five species.

Furthermore, when evaluated in combination with leaf PFD, Amax was only correlated

with daily carbon gain in Virola.

When results were pooled for the five species in this study, there was a strong

relationship between Amax and daily carbon gain (r2 = 0.48, P < 0.0001, Figure 2-7) and

leaf PFD and daily carbon gain (r2 = 0.89, P < 0.0001, Figure 2-7). A previous study by

Zotz and Winter (1993) documented a higher correlation coefficient for the relationship

between Amax and Aday for 64 diel courses of net CO2 exchange for eight tropical species

(r2 = 0.92, P < 0.0001). Zotz and Winter (1993) did not encounter a strong relationship

between Aday and leaf PFD (r2 = 0.13, P < 0.01). Indeed, there was no relationship

between leaf PFD and Aday for a Ceibapentandra, a tropical canopy tree (r2 = 0.48, P <

0.1; (Zotz and Winter 1994). I believe the contrasting results between these two studies









can be explained by variation in leaf light levels. Even on clouded days, leaf PFD was

relatively high for leaves of the eight species measured by Zotz and Winter (1993), which

resulted in light-saturated rates of photosynthesis (Zotz and Winter 1993). The same

cannot be said for leaves of the five species in this study. In all species, all leaves

experienced light levels below their light compensation point at some time during the

three days during which light was measured. Even young leaves at the top of the plant

experienced light-limiting conditions for 12-19% of day light due to plant positioning

inside the gap and possible shading by surrounding plants. Thus, it appears that when

leaves are exposed to light conditions that allow for performance at maximum

photosynthetic capacity, daily carbon gain is highly correlated with maximum

photosynthetic capacity. However, when leaves are not constantly exposed to light-

saturating conditions, leaf light levels are more highly correlated with daily carbon gain.

This study documented a negative carbon gain for two out of the fourteen measured

leaves of Tetragastris. Both of these leaves were old (512 and 596 days) at the time of

measurement. At first, this result was surprising as all plants were grown in canopy gaps

in which the light environment above the seedlings ranged from 23-52% of full sun

(Kitajima unpublished data). However, the light conditions within each gap were

variable due to plant positioning inside the gap, and the height of surrounding plants. For

instance, above plant light measurements for Tetragastris were 8.4%, while leaves of

Aspidosperma and Genipa experienced much higher light levels with some leaves

exposed to as much as 30-38% of full sun. When taking into account that the above plant

light environment for Tetragastris was only 8.4% of full sun, negative carbon balances

for older, shaded leaves at the bottom of the plant are reasonable. Furthermore, the daily









carbon gains presented are means of three days of measurements for which the standard

deviation of carbon gain varied from 2-3 mmol CO2 S-2 day-1 depending on the leaf. Even

leaves with average negative daily carbon gain generally experienced some days with a

low positive carbon gain.

This study used photosynthetic light-response curves to calculate carbon gain from

daily courses of photon flux density (PFD). While this approach has been commonly

used in the past (Chazdon 1986, Williams et al. 1989, Poorter et al. 2006), the equation

used to calculate carbon gain (P(I) = {I + Amax {(I + Amax)2 40(IAmaxl}05 / 20} -

R, Johnson and Thornely 1984) does not take induction time into account (Kikuzawa et

al. 2004). Carbon gain can also be limited by the mid-day depression of photosynthetic

rates caused by stomatal closure in response to short-term drought during the hottest and

most humid times of day (Zotz and Winter 1993). Thus, it is likely that the carbon gain

calculated for leaves in this experiment are slightly overestimated. The most direct

method of calculating carbon gain is by measuring daily courses of photosynthetic rates

(Zotz and Winter 1993, 1994), but this approach was not used in this study.

Leaf PFD was the best predictor of carbon gain in this study regardless of whether

analyses were done separately for each species or together by pooling samples across all

species. It is well known that the photosynthetic capacity of leaves is correlated with the

light conditions under which they develop, resulting in a relationship between Amax and

leaf PFD (Field 1988). Particularly in shaded conditions, leaf PFD is more highly

correlated with leaf carbon gain than photosynthetic capacity (Chazdon and Pearcy

1986). In this experiment, leaf light ranged from 0.3 to 44% of daily PFD above the

canopy. While this range is much greater than the 1-2% of light typically experienced by






42


understory plants, the average daily leaf PFD across leaves and species was only 10% of

full sun, suggesting that leaves were not able to perform at maximum photosynthetic

capacity throughout the day. These results indicate that in conditions where light is not

saturating, leaf PFD is a better predictor of daily carbon gain than Amax.










Table 2-1. Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer
method for each species studied. Species are ranked from shortest median
leaf lifespan to longest median leaf lifespan. Nomenclature follows Croat
(1978).
Total Std
Species Family Leaves Median Mean Dev
Genipa americana Rubiaceae 478 224 234.8 4.96
Tabernaemontana arborea Apocynaceae 542 279 277.6 3.94
Virola surinamensis Myristicaceae 709 504 494.3 6.56
Tetragastris panamensis Burseraceae 89 671 648.3 13.87
Aspidosperma cruenta Apocynaceae 31 797 745.3 12.51

Table 2-2. Number of seedlings and leaves for which % PFD and Aday were measured in
each of five species of tropical tree species. Species are ranked from shortest
median leaf lifespan to longest median leaf lifespan.
# of Leaves
# of
Species Plts %PFD Aday

Genipa 5 9 9
Tabernaemontana 2 8 8
Virola 4 12 12
Tetragasteris 4 14 14
Aspidosperma 2 7 7

Table 2-3. Multiple regression analysis of seedlings of five tropical tree species with Amax
and leaf PFD as predictor variables, and daily carbon gain as the response
variable. The coefficient of multiple determination (r2) represents the
proportion of the variance explained by the two variables together, while P-
values represent the significance of each variable as it contributed to the
model.
Amax Leaf PFD
Species r2 P P
Genipa 0.90 0.12 0.0005
Tabernaemontana 0.91 0.43 0.006
Virola 0.96 0.006 < 0.0001
Tetragastris 0.87 0.17 0.0005
Aspidosperma 0.88 0.52 0.007









Table 2-4. Light curve parameters for five
clac ulate daily carbon gain


species of tropical tree seedlings used to


Amax (mnol m-2 s-) R (ftmol m-2 s-) Quantum Curvature
30 day 200 day 30 day 200 day Yield (0D) (0)
Genipa 13.5 6.61 -0.494 -0.494 0.058 0.697
Tabernaemontana 9.36 8.99 -0.486 -0.333 0.049 0.674
Virola 6.05 4.91 -0.349 -0.298 0.047 0.701
Tetragastris 8.61 6.58 -0.416 -0.331 0.049 0.673
Aspidosperma 8.22 10.58 -0.447 -0.396 0.056 0.734













Leaf PFD


Leaf Age


Aday


Amax


Figure 2-1. A schematic representation of the expected relationship between leaf age, leaf
PFD, Amax, and Aday. The relationships between leaf age and leaf PFD and
Amax were explored in Chapter 1.


40




30




UL
.. 20




10




0


0 2000 4000 6000 8000 10000


12000


PFD mmol m-2 day-'
Figure 2-2. The relationship between % PFD and mmol PFD m-2 day1 for all leaves and
species pooled. Each point is a three day average.
































<100 <200 <300 <400 <500 <600 <700


Leaf Age (days)
. The percent of time leaves of different ages spent above and below their light
compensation point (LCP) relative to the total day length. Depending on their
age, leaves spent 12.9 to 51.6% of the total time they were exposed to day
light below their LCP.


0 -



Figure 2-3





























50


E o0
" 250
0





-6
E 200
E

S150

0
S100


0 200 400 600

Leaf Age (Days)


800

200


Genipa
r2= 0.20, P = 0.22
*
















Tabernaemontana
y = -0.206x + 69.01
r2 = 0.50, P = 0.0501













49


Og
0


0 0


0 200 400 600 800

Leaf Age (Days)
Figure 2-4. The relationship between leaf age and three-day averages of daily carbon gain
in seedlings of five species of tropical trees.


Virola
y =-0.141x + 82.98
r2 = 0.51, P = 0.0091
















Tetragasteris
y = -0.070x + 41.60
r2 = 0.54, P = 0.0028















Aspidosperma
y = -0.267x + 157.70
r2 = 0.85, P = 0.0029
























































200 -


Amax (.mol m-2 s1)


Virola
y = 9.77x 7.57
r2 = 0.71. P = 0.0006


Genipa
r2= 0.1, P= 0.4







00








Tabernaemontana
y = 8.63x 39.29
r2 = 0.51, P= 0.0461










--

09
0M


Tetragasteris
y = 5.10x 4.47
r2 = 0.60, P = 0.0011


Aspidosperma
r2 = 0.1, P = 0.5


*0 0


Amax (pmol m-2 s1)

Figure 2-5. The relationship between maximum photosynthetic capacity (Amax) and three
day averages of daily carbon gain for seedlings of five species of tropical
trees.


S

S
0

0
0


50


E 0
" 250
0
-6

E 200
E


S150


r 100


a 50





























50


E 0
" 250
0
-6

E 200


3 150


r 100


S s50


Leaf PFD (mmol m-2 day1)


200


Virola
y = 0.026x 7.49
r2 = 0.91. P < 0.0001


S


0
0


Tetragasteris
y = 0.015x + 0.62
r2 = 0.85, P < 0.0001


Aspidosperma
y = 0.013x + 37.79
r2= 0.87, P = 0.0022




loo


Leaf PFD (mmol m2 day1)
Figure 2-6. The relationship between leaf age and three day averages of light incident on
those leaves for seedlings of five species of tropical trees. Leaf PFD increased
significantly with carbon gain in all species.


Genipa
y = 0.014x + 23.81
r2 = 0.84, P= 0.0005


Tabernaemontana
y = 0.019x + 2.58
r2 = 0.89, P = 0.0004


S S







50




250 250-
y = 9.14x 11.17 y = 0.02x + 6.17
r2 =0.48, P < 0.0001 r = 0.89, P < 0.0001
'C 200 200 -
E




- 100- 100 -


o 50 ** **

5 C* 0*


SI N (0 Ib I I

Amax (tmo1 m-2 S-1) Leaf PFD (mmol m-2 day1)

Figure 2-7. The relationship between daily carbon gain and Amax or leaf PFD when all
measurements and species were pooled.















LIST OF REFERENCES

Ackerly, D. 1999. Self-shading, carbon gain and leaf dynamics: a test of alternative
optimality models. Oecologia 119:300-310.

Ackerly, D. D., and F. A. Bazzaz. 1995. Leaf dynamics, self-shading and carbon gain in
seedlings of a tropical pioneer tree. Oecologia 101:289-298.

Bazzaz, F. A. 1979. Physiological ecology of plant succession. Annual Review of
Ecology and Systematics 10:351-371.

Bongers, F., and J. Pompa. 1990. Leaf dynamics of seedlings of rain-forest species in
relation to canopy gaps. Oecologia 82:122-127.

Chabot, B. F., and D. J. Hicks. 1982. The ecology of leaf life spans. Annual Review of
Ecology and Systematics 13:229-259.

Chazdon, R. L. 1986. Light variation and carbon gain in rain-forest understorey palms.
Journal of Ecology 74:995-1012.

Chazdon, R. L., and R. W. Pearcy. 1986. Carbon gain and photosynthetic efficiency
during lightflecks. Oecologia 69:524-531.

Coley, P. D. 1988. Effects of plant-growth rate and leaf lifetime on the amount and type
of anti-herbivore defense. Oecologia 74:531-536.

Croat, T. B. 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford.

Daws, M. I., C. E. Mullins, D. Burslem, S. R. Paton, and J. W. Dalling. 2002.
Topographic position affects the water regime in a semi-deciduous tropical forest in
Panama. Plant and Soil 238:79-90.

Denslow, J. S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12:47-
55.

Donovan, T. M., and C. W. Weldin. 2002. Spreadsheet Exercises in Ecology and
Evolution. Sinauer Associates, Sunderland, MA.

Ellsworth, D. S., and P. B. Reich. 1996. Photosynthesis and leaf nitrogen in five
Amazonian tree species during early secondary succession. Ecology 77:581-594.









Escudero, A., and S. Mediavilla. 2003. Decline in photosynthetic nitrogen use efficiency
with leaf age and nitrogen resorption as determinants of leaf life span. Journal of
Ecology 91:880-889.

Field, C. 1983. Allocating leaf nitrogen for the maximization of carbon gain-leaf age as a
control on the allocation program. Oecologia 56:341-347.

Field, C., and H. A. Mooney. 1983. Leaf age and seasonal effects on light, water, and
nitrogen use efficiency in a California shrub. Oecologia 56:348-355.

Hikosaka, K. 2004. Interspecific difference in the photosynthesis-nitrogen relationship:
patterns, physiological causes, and ecological importance. Journal of Plant
Research 117:481-494.

Hikosaka, K., I. Terashima, and S. Katoh. 1994. Effects of leaf age, nitrogen nutrition
and photon flux-density on the distribution of nitrogen among leaves of a vine
(Ipomoea tricolor) grown horizontally to avoid mutual shading of leaves.
Oecologia 97:451-457.

Hirose, T., and M. J. A. Werger. 1987. Maximizing daily canopy photosynthesis with
respect to the leaf nitrogen allocation pattern in the canopy. Oecologia 72:520-526.

Johnson, I. R., and J. H. M. Thornley. 1984. A model of instantaneous and daily canopy
photosynthesis. Journal of Theoretical Biology 107:531-545.

Kikuzawa, K. 1991. A cost-benefit-analysis of leaf habit and leaf longevity of trees and
their geographical pattern. American Naturalist 138:1250-1263.

Kikuzawa, K., and D. Ackerly. 1999. Significance of leaf longevity in plants. Plant
Species Biology 14:39-45.

Kikuzawa, K., H. Shirakawa, M. Suzuki, and K. Umeki. 2004. Mean labor time of a leaf.
Ecological Research 19:365-374.

King, D. A. 2003. Allocation of above-ground growth is related to light in temperate
deciduous saplings. Functional Ecology 17:482-488.

Kitajima, K., K. Kikuzawa, D. Ackerly, S. S. Mulkey, and S. J. Wright. 2003. Leaf-age
effects on photosynthetic rates as a key parameter in the theory of leaf longevity.
Unpublished manuscript.

Kitajima, K., S. S. Mulkey, M. Samaniego, and S. J. Wright. 2002. Decline of
photosynthetic capacity with leaf age and position in two tropical pioneer tree
species. American Journal of Botany 89:1925-1932.

Kitajima, K., S. S. Mulkey, and S. J. Wright. 1997. Decline of photosynthetic capacity
with leaf age in relation to leaf longevities for five tropical canopy tree species.
American Journal of Botany 84:702-708.









Koike, T. 1988. Leaf structure and photosynthetic performance as related to the forest
succession of deciduous broad-leaved trees. Plant Species Biology 3:77-87.

Leigh, E. G., Jr., Windsor, D.M., Rand, S.A. 1982. The Ecology of a Tropical Forest:
Seasonal Rhythms and Long-Term Changes, Second Edition. Smithsonian
Institution Press, Washington, D.C.

Mediavilla, S., and A. Escudero. 2003. Leaf life span differs from retention time of
biomass and nutrients in the crowns of evergreen species. Functional Ecology
17:541-548.

Mooney, H. A., C. Field, S. L. Gulmon, and F. A. Bazzaz. 1981. Photosynthetic capacity
in relation to leaf position in desert versus old-field annuals. Oecologia 50:109-112.

Peterson, A. G. 1999. Reconciling the apparent difference between mass- and area-based
expressions of the photosynthesis-nitrogen relationship. Oecologia 118:144-150.

Poorter, H., S. Pepin, T. Rijkers, Y. de Jong, J. R. Evans, and C. Korner. 2006.
Construction costs, chemical composition and payback time of high- and low-
irradiance leaves. Journal of Experimental Botany 57:355-371.

Reich, P. B., D. S. Ellsworth, M. B. Walters, J. M. Vose, C. Gresham, J. C. Volin, and W.
D. Bowman. 1999. Generality of leaf trait relationships: a test across six biomes.
Ecology 80:1955-1969.

Reich, P. B., C. Uhl, M. B. Walters, and D. S. Ellsworth. 1991. Leaf life-span as a
determinant of leaf structure and function among 23 Amazonian tree species.
Oecologia 86:16-24.

Reich, P. B., C. Uhl, M. B. Walters, L. Prugh, and D. S. Ellsworth. 2004. Leaf
demography and phenology in Amazonian rain forest: A census of 40 000 leaves of
23 tree species. Ecological Monographs 74:3-23.

Reich, P. B., M. B. Walters, and D. S. Ellsworth. 1992. Leaf life-span in relation to leaf,
plant, and stand characteristics among diverse ecosystems. Ecological Monographs
62:365-392.

Reich, P. B., M. B. Walters, and D. S. Ellsworth. 1997. From tropics to tundra: global
convergence in plant functioning. Proceedings of the National Academy of
Sciences of the United States of America 94:13730-13734.

Takashima, T., K. Hikosaka, and T. Hirose. 2004. Photosynthesis or persistence: nitrogen
allocation in leaves of evergreen and deciduous Quercus species. Plant Cell and
Environment 27:1047-1054.

Williams, K., C. B. Field, and H. A. Mooney. 1989. Relationships among leaf
construction cost, leaf longevity, and light environment in rain-forest plants of the
Genus Piper. American Naturalist 133:198-211.






54


Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J.
Cavender-Bares, T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Gamier,
P. K. Groom, J. Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J.
Midgley, M. L. Navas, U. Niinemets, J. Oleksyn, N. Osada, H. Poorter, P. Poot, L.
Prior, V. I. Pyankov, C. Roumet, S. C. Thomas, M. G. Tjoelker, E. J. Veneklaas,
and R. Villar. 2004. The worldwide leaf economics spectrum. Nature 428:821-827.

Zotz, G., and K. Winter. 1993. Short-term photosynthesis measurements predict leaf
carbon balance in tropical rain-forest canopy plants. Planta 191:409-412.

Zotz, G., and K. Winter. 1994. Photosynthesis of a tropical canopy tree, Ceibapentandra,
in a lowland forest in Panama. Tree Physiology 14:1291-1301.















BIOGRAPHICAL SKETCH

Carla C. Stefanescu was born in Madison, WI, in May 1980. She attended Central

High School in Tuscaloosa, AL, where she became fascinated by biology. Carla earned

her Bachelor of Science degree in biology with a concentration in environmental studies

at the University of the South in May 2002. She became interested in tropical ecology

during the pursuit of her bachelor's degree when she studied with the Organization of

Tropical Studies in Costa Rica for the Spring semester of 2001. During her

undergraduate studies, she also worked as a research assistant in Sewanee, TN, and

Tuscaloosa, AL. In August 2006, she obtained her Master of Science degree in botany at

the University of Florida.

PAGE 1

COST-BENEFIT THEORY OF LEAF LIFE SPAN IN SEEDLINGS OF TROPICAL TREE SPECIES By CARLA C. STEFANESCU A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Carla C. Stefanescu

PAGE 3

iii ACKNOWLEDGMENTS I would like to thank my advisor, Kaoru Kitajima, for giving me the opportunity to pursue a masters degree under her guidance, providing me with financial support during my first field season and academic support throughout the completion of this degree. I would also like to thank my committee members, Jack Putz and Tim Martin, who provided invaluable comments and support th roughout the thesis wr iting process. I would also like to thank the Botany Department staff, in particular Kimberly Holloway, who offered necessary logistic support throughout my time at the University of Florida. I thank my parents, Drs. Doru and Lydia Stefanescu, whose integrity, courage, and discipline have provided me with a lifelong example and reminder of what I one day hope to become. I would also like to thank my sister, Alina Stefanescu, whose creativity and free spirit have provided me with a better perspective on life. Last, but not least, I would like to thank the smallest member of my family, Maxwell Stefanescu, whose sheer presence serves as a reminder that some of lifes miracles cannot be explained. I would also like to thank my dear friends Andrea Crino, Christine Lucas, Silvia Alvarez-Claire, Jenny Schafer, Leslie B oby, Heather Loring, Adrienne Frisbee, and Hanna Lee, without which my time in Ga inesville would have been much more productive but much less worthwhile.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT....................................................................................................................... ix CHAPTER 1 DECLINE OF PHOTOSYNTHETIC CAPACITY WITH LEAF AGE IN SEEDLINGS OF ELEVEN TROPICAL TREE SPECIES..........................................1 Introduction................................................................................................................... 1 Methods........................................................................................................................ 7 Study Site...............................................................................................................7 Plot Establishment.................................................................................................8 Measurement of Light and P hotosynthesis in the Field........................................8 Laboratory Measurements a nd Statistical Analyses............................................10 Results........................................................................................................................ .11 Discussion...................................................................................................................12 2 LEAF AGE, PHOTOSYNTHETIC CAPACITY, AND LEAF LIGHT LEVEL AS PREDICTORS OF DAILY CARBON GAIN USING SEEDLINGS OF FIVE TROPICAL TREE SPECIES.....................................................................................30 Introduction.................................................................................................................30 Methods......................................................................................................................34 Study Site.............................................................................................................34 Plot Establishment...............................................................................................34 Measurement of Light and P hotosynthesis in the Field......................................35 Calculation of Daily Carbon Gain.......................................................................36 Statistical Analyses..............................................................................................37 Results........................................................................................................................ .38 Discussion...................................................................................................................39

PAGE 5

v LIST OF REFERENCES...................................................................................................51 BIOGRAPHICAL SKETCH.............................................................................................55

PAGE 6

vi LIST OF TABLES Table page 1-1 Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer method for each species studied...............................................................................18 1-2 Number of plants and leaves measured for Amax, Aarea, LMA, Narea, and PNUE measurements for each species................................................................................18 1-3 Number of plants and leaves for which % PFD and Aarea were measured for a subset of species.......................................................................................................19 1-4 Mean leaf lifespan and regression statistics ( x and y intercepts and slopes) for areaand mass-based light-saturated ne t photosynthesis against leaf age (days after full expansion) for seedlings of eleven tropical tree species...........................20 1-5 Mean leaf lifespan and regression statistics ( x and y intercepts and slopes) for light-saturated net photosynthesis (Aarea in micromols of CO2 m-2 s-1) against leaf age (days after full expansion) for eight tropical tree species...........................20 2-1 Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer method for each species studied...............................................................................43 2-2 Number of seedlings and leaves for which % PFD and Aday were measured in each of five species of tropical tree species.............................................................43 2-3 Multiple regression analysis of seedli ngs of five tropical tree species with Amax and leaf PFD as predictor variables, and daily car bon gain as the response variable.....................................................................................................................43 2-4 Light curve parameters for five spec ies of tropical tree seedlings used to calculate daily carbon gain.......................................................................................44

PAGE 7

vii LIST OF FIGURES Figure page 1-1 Light saturated photosynthetic assimilation rates (Aarea) of leaves of different age on seedlings of eleven tropical tree species.............................................................21 1-2 Mass-based photosynthesis (Amass) of leaves with contras ting leaf age in eleven species of tropical trees............................................................................................22 1-3 Relationship between initial Aarea, Aarea-leaf age slope, Amass, Amass-leaf age slope with median leaf lifespan for each species.....................................................23 1-4 The relationship between the actual medi an leaf lifespan and an estimate of the time when photosynthetic capacity would reach zero (parameter b the x intercept of the leaf age-Aarea and leaf age-Amass regression)...................................24 1-5 Relationship between leaf mass per unit area (g m-2) and leaf age in seedlings of eleven species of tropical trees.................................................................................25 1-6 Relationship between leaf nitrogen (g m-2) and leaf age..........................................26 1-7 Relationship between photosynthetic n itrogen use efficiency (PNUE) and leaf age............................................................................................................................ 27 1-8 Decline in %PFD (the total daily PFD in cident on the leaf relative to the total daily PFD above the canopy) with leaf ag e in a subset of species and leaves.........28 1-9 Effect of % PFD (the total daily PFD in cident on the leaf relative to the total daily PFD above the canopy) on light-saturated photosynthesis (Aarea) and for a subset of species and leaves.....................................................................................29 2-1 A schematic representation of the expect ed relationship between leaf age, leaf PFD, Amax, and Aday..................................................................................................45 2-2 The relationship betwee n % PFD and mmol PFD m-2 day-1 for all leaves and species pooled..........................................................................................................45 2-3 The percent of time leaves of different ages spent above and below their light compensation point (LCP) relative to the total day length.......................................46

PAGE 8

viii 2-4 The relationship between l eaf age and three-day averages of daily carbon gain in seedlings of five species of tropical trees.................................................................47 2-5 The relationship between ma ximum photosynthetic capacity (Amax) and three day averages of daily carbon gain for seed lings of five species of tropical trees....48 2-6 The relationship between leaf age and three day averages of light incident on those leaves for seedlings of fi ve species of tropical trees.......................................49 2-7 The relationship betwee n daily carbon gain and Amax or leaf PFD when all measurements and species were pooled...................................................................50

PAGE 9

ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science COST-BENEFIT THEORY OF LEAF LIFE SPAN IN SEEDLINGS OF TROPICAL TREE SPECIES By Carla C. Stefanescu August 2006 Chair: Kaoru Kitajima Major Department: Botany Variation in leaf lifespan is highly correlated with vari ation in leaf physiological and structural traits within and among plant communities. The cost-benefit theory of leaf lifespan links interspecific variation in leaf lifespan with the d ecline of photosynthetic capacity with leaf age, assumi ng that photosynthetic capacity (Amax) is linearly correlated with daily net assimilation rate (Aday). In Chapter 1, I evaluated these alternative predictions by comparing the decline rate of photosynthetic capacity with leaf age for eleven tropical tree species from Barro Colo rado Island, Panama. Leaves were sampled from seedlings that had been grown for 23 years in common gardens established in natural tree fall gaps. Leaves representing a range of leaf ages were sampled for photosynthetic capacity (Amax), total daily photon flux dens ity (PFD), leaf mass per unit area (LMA), leaf nitrogen (leaf N), and photosynthetic nitrog en use efficiency (PNUE). In all species, Amax, PFD, and PNUE decreased with leaf age. Leaf N decreased with leaf age in seven out of eleven species. There was no consistent rela tionship between LMA

PAGE 10

x and leaf age. In most species, leaves we re lost before their photosynthetic capacity reached zero. Species with shorter leaf life spans had a high initial photosynthetic capacity and a steep decline rate of photosynthe tic capacity with leaf age. My results favor the second prediction of the cost-benefit theory, suggesting that the total number of leaves per seedling is constrained by factor s other than decline of photosynthetic rates with leaf age. The above theoretical predictions assu me that instantaneous photosynthetic capacity (Amax) is linearly correlated w ith daily net carbon gain (Aday). In Chapter 2, I evaluated how Aday may be correlated with Amax by estimating daily carbon gain for seedling leaves of five tropical tree species that varied in medi an leaf lifespan. I estimated daily net carbon gain averaged over three days, using photosynthetic light response curves and PFD directly measured on individual leaves. While leaf age was correlated with carbon gain in four ou t of the five species in this study, Amax was correlated with carbon gain in three species. In contrast, leaf PFD was correlated with carbon gain in all species. In a multiple regression using leaf PFD and Amax as predictive variables for Aday, Amax was a predictor of Aday only in one species. These results can be explained in terms of the light environm ent leaves experienced during the study. Previous studies that documente d a strong correlation between Amax and Aday measured canopy leaves that received consistent light sa turating conditions. In contrast, the leaves in this study received light below their light compensation point for 12 to 52% of the total day light. Based on the results of this study, it appears that leaf PFD is a better predictor of daily carbon gain than Amax in conditions where light is not saturating.

PAGE 11

1 CHAPTER 1 DECLINE OF PHOTOSYNTHETIC CAPACITY WITH LEAF AGE IN SEEDLINGS OF ELEVEN TROPICAL TREE SPECIES Introduction Tropical rain forest species produce leaves with widely variable leaf lifespan. For example, while leaves of pioneer species ge nerally live less than one year, leaves of shade-tolerant species can live five years or more (Coley 1988, Williams et al. 1989). Variation in leaf lifespan among species reflect s genotypic differences that evolved partly in relation to life history stra tegies. Pioneers survive best in sites with high light and nutrients where their inherent fast rates of growth and leaf turnover help them outcompete other plants (Bazzaz 1979, Ellsworth an d Reich 1996). Shade-tolerant species, although rapidly out-competed in high light environments, persist in light-limited environments where slower growth rates, lower photosynthetic cap acity, and longer leaf lifespans allow them to maintain a positive carbon budget by amortizing initial construction costs (Bazzaz 1979, Koike 1988). When comparing species, it has been shown that shade tolerant species have lo wer photosynthetic capacity, higher leaf mass per unit area, and longer leaf lifespans than pioneer species under comparable light conditions (Williams et al. 1989, Ellsworth and Reich 1996, Reich et al. 2004). These patterns are thought to be the result of f unctional constraints that enhance either productivity or nutrient conser vation and defense. A plant can allocate resources to leaves with a high photosynthetic capacity but short lifespan, or produce physically resistant leaves that have a lower rate of carbon assimilation over a longer period of time

PAGE 12

2 2 (Reich et al. 1992). For example, thicker cell walls are necessary to increase leaf toughness which provides mechanical protecti on essential to keeping leaves for long periods of time (Hikosaka 2004). However, thicker leaves requi re higher nitrogen investments in cell walls rather than photosynthetic proteins and thus, have lower overall carbon assimilation (Hikosaka 2004). Variation in leaf lifespan is highly correlated with vari ation in leaf physiological and structural traits within and among co mmunities (Reich et al. 1991, Reich et al. 1992, Reich et al. 1999). Mass-based photosynthetic capacity (Amass) and leaf nitrogen (Nmass) are usually positively correlated, and both are negatively correlated with leaf mass per unit area (LMA). Amass and Nmass decrease with increasing leaf lifespan, while LMA increases with increasing leaf lifespan. The decline of Amass and Nmass with leaf age is thought to be due to the retr anslocation of nitrogen to younger leaves which maximizes whole-shoot photosynthetic income rather than to uncontrolled physiological deterioration (Field and Mooney 1983, Acke rly and Bazzaz 1995). Hikosaka et al. (1994) showed that age-related changes due to retranslocation were more pronounced when of light or nitrogen were limiting. Photosynthetic nitrogen use efficiency (PNUE), the rate of photosynthesis expressed per unit nitrogen, is an important leaf trait to ch aracterize in relation to leaf lifespan. Although a few studies did not find a change in PNUE with leaf age (Mooney et al. 1981, Field and Mooney 1983), more r ecent studies have shown that PNUE decreases with both leaf age and LMA (Rei ch et al. 1991, Kitajima et al. 1997, Escudero and Mediavilla 2003, Mediavilla and Esc udero 2003, Wright et al. 2004). The retranslocation of nitrogen to younger leaves can result in the decline of both leaf

PAGE 13

3 3 nitrogen and PNUE in older leaves whic h consequently leads to the decline of photosynthetic capacity with leaf age. The decrease of PNUE with LMA can be explained by the allocation of nitrogen to non-photosynthetic f unctions in leaves (Escudero and Mediavilla 2003, Hikosaka 2004) Leaves with a high LMA allocate a larger amount of nitrogen to cell wall pr oteins, and a lower amount of nitrogen to photosynthetic proteins (Takashima et al. 2004 ). This trade-off in nitrogen allocation between photosynthesis and structure results in a negative correlation between PNUE and LMA (Hikosaka 2004). Thus, in studies atte mpting to explain age-related changes in photosynthetic decline, the rela tionship between leaf age, l eaf nitrogen, and PNUE must be evaluated. Cost-benefit analyses have been used to explain variation in leaf lifespan (Chabot and Hicks 1982, Kikuzawa 1991, Kikuzawa and Ackerly 1999). Leaf lifespan is typically considered to reflect a bala nce between lifetime leaf carbon gain and construction and maintenance costs (Chabot and Hicks 1982). Thus, the affect of leaf age on photosynthesis is important in cost-benef it theories of leaf lifespan. Maximum photosynthetic capacity (Amax) typically declines monotonically with leaf age and can be approximated by the following linear equation: Amax = a (1 – t/b), where t is the leaf age in days from the time of full leaf ex pansion (Kikuzawa 1991). The parameter a ( y intercept of the regression) re presents the initia l photosynthetic capacity after full leaf expansion, and can be directly measured. The parameter b ( x intercept of regression) represents the time when photos ynthetic capacity would reach zero. This parameter is a statistical extrapolation dete rmined as a function of the initial photosynthetic rate ( a ) and the rate of its decline (parameter a/b the slope of the regression).

PAGE 14

4 4 Cost-benefit analyses of l eaf lifespan have generated tw o alternate predictions. If there are no external constraints on the ma ximum number of leaves produced and photosynthesis declines with l eaf age (i.e., self-shading cau ses photosynthetic decline and limits the number of leaves per shoot; Case 1) a leaf is expected to maximize net carbon gain over its entire lifetime (Kitajima et al. unpublished manus cript). In this case, the optimum leaf lifespan should be close to th e time when photosynthetic rates equal zero (Parameter b ). Since total daily net photosynthetic income should reach zero before net photosynthetic rates equal zero, act ual leaf lifespan is expected to be less than parameter b (Kitajima et al. 2002). Parameter b approximates actual leaf longevity when the same leaf is repeatedly measured for plants with very short lifespans (Ackerly and Bazzaz 1995). In contrast, in tropi cal tree species with leaf life-times ranging from 183-343 days, b was significantly greater th an actual leaf longevity (Kitajima et al. 1997). Alternatively, if the total number of leaves is limited by external resources (Case 2), a leaf is expected to maximize the rate of net carbon gain averaged over the lifetime of the leaf (Kikuzawa 1991, Kikuzawa and Ackerly 1999, Kitajima et al. unpublished manuscript). In this scenari o, leaves are replaced when their net carbon gain per unit time averaged over their entire lifespan has reached a maximum and optimal leaf lifespan is predicted to be equal to (2 b C/ a )1/2 where C is the construction cost of the leaf (Kikuzawa 1991). The parameters a and C can be considered on a mass or area basis. Based on this theoretical model, it is expected that leaf lifespan will be short when the initial photosynthetic rate of the leaf is high, and long when construction costs are high or the rate of decline of photosynt hetic rates with leaf age is slow (Kikuzawa 1991). Based on a comparison of seven canopy tree speci es, Kitajima and others (unpublished

PAGE 15

5 5 manuscript) found that Case 1 is most appli cable only to short-lived leaves of pioneer species, while Case 2 is equally applicab le to both shorta nd long-lived leaves. Numerous studies have analyzed the ma ximum photosynthetic rates of different species (Reich et al. 1991, Reich et al. 1992, Reich et al. 1999). However, data on the rate of photosynthetic decline with leaf ag e in species with different lifespans are relatively scarce (Kikuz awa and Ackerly 1999). Such data are essential in estimating the long-term carbon budget of individual leaves and entire plants. Here we report the rate of decline of photosynthetic cap acity with leaf ag e for seedlings of eleven tropical tree species with median leaf lifes pans of from 140-797 days. Previous studies have focused on comparing the rate of photosynthetic dec line for canopy trees (K itajima et al. 1997, Kitajima et al. 2002) or have only evaluated changes in one species (Ackerly and Bazzaz 1995). Thus, my study appears to be the firs t to evaluate whethe r cost-benefit theory explains variation in the leaf lifespan of seedlings. Why migh t leaf lifespan be different in seedlings and adults? In both seedlings and adu lts, leaf lifetimes ar e expected to be correlated with the cost and maintenance of leaves and supporting structures (Kikuzawa and Ackerly 1999). The higher initial construc tion and maintenance costs, the longer a leaf is expected to live in order to pay back those costs. As seedli ngs grow in height, the costs of biomechanical support and the transport of water and nutrients to leaves are also expected to increase (Kikuzawa and Ackerly 1999) To offset these increased costs, leaf photosynthetic capacity should al so increase with plant hei ght. Indeed, in temperate species, light-saturated photosynthetic rate s have been documented to increase with ontogeny, from seedlings to saplings and adult tr ees (Koike 1988). As leaves with higher photosynthetic capacity and more supporting structures are expect ed to be more costly to

PAGE 16

6 6 produce, they are also anticipated to require longer to pay back initial construction costs. Thus, seedlings should have shorter leaf lifes pans than adult trees of the same species because construction and maintenance costs of supporting structures are expected to be lower for seedlings than adu lts (Kikuzawa and Ackerly 1999). When leaf lifespan was compared across species with different life fo rms, species with lower biomechanical and transport costs, such as aquatic floating plants had a shorter leaf lifespan than leaves with higher construction and maintenance costs su ch as temperate deciduous trees (Kikuzawa and Ackerly 1999). Alternatively, leaves in hi gh light environments are expected to pay back construction and maintenance costs more quickly than leaves in shaded environments. When species are compared, pl ants grown in high light conditions often have a faster leaf turnover than plants gr own in the shaded unde rstory (Bongers and Pompa 1990). Therefore, it is likely that th e average leaf lifespan of a seedling growing in the forest understory will be longer than th at of an adult tree with access to a higher light environment. Due to the effect of light environment on l eaf lifespan, it is imperative that plants be grown in similar light environments when comparing leaf lifespans across species or life forms. Thus, all species in this st udy were planted in naturally occurring canopy gaps. Canopy gaps are particular ly important for tropical tree regeneration as species are thought to be differentiated in part by their re sponse to light conditi ons in gaps (Denslow 1980). This response is partially determin ed by differences in leaf dynamics among species as species with shorter leaf lifespans are typically intolerant of shade, while species with longer leaf lifespans are shade-tolerant (Bongers and Pompa 1990). Detailed knowledge of species responses to si milar gap micro-environments could lead to

PAGE 17

7 7 an understanding of the degree of differentia tion among species in the way they exploit photosynthetically active radiation. This di fferentiation is most pronounced during early stages of the tree life cycle when seedli ngs or saplings suffe r high mortality. The objectives of this study were: (1) to fill a gap in empirical data on declines of photosynthetic capacity with leaf age for seedlings of tropical tree species with a range of leaf lifespans; (2) to evaluate two alternative predictions of cost-benefit theory; and, (3) to explore the functional basis for relations hips among leaf age, leaf position, light microenvironment, leaf mass per area, nitroge n content, and photosynt hetic capacity. We compared the x intercept of the regression of leaf age with ph otosynthetic capacity (parameter b) with the observed leaf lifesp an of each species. A relationship between the parameter b, the x intercept of the regression of leaf ag e with photosynthetic capacity, and species leaf lifespan woul d provide qualitative support for Case 1 in which external resources do not limit the tota l number of leaves produced. Alternatively, if plants are limited by external resources (C ase 2), I expected that the rate of photosynthetic decline would be faster (i.e. a/b would be larger) in species with shorter leaf lifespans according to KikuzawaÂ’s (1991) predictions. I also pred icted that photosynthetic rate, nitrogen per unit area, and photon flux density (PFD) meas ured at the leaf surface would decrease with leaf age. Methods Study Site The study was conducted on Buena Vista Pe ninsula, a 60-year-old secondary lowland tropical forest area in the Ba rro Colorado National Monument (BCNM), Republic of Panama (9 09Â’ N, 79 51Â’W). The species compos ition, climate, and ecology of the Buena Vista Peninsula is similar to that of the young forests on Barro Colorado

PAGE 18

8 8 Island (Croat 1978, Leigh 1982). The BCNM forest is semi-deciduous during the pronounced dry season that usually lasts from mid-December to April, and receives annual rainfall ca. 2,600 mm (Daws et al. 2002). Plot Establishment Three gap plots were established in natu ral tree fall gaps of approximately 100 m2 in May 2002. Seeds were collected on Barro Colorado Island from at least two parent trees and germinated in a greenhouse. At radi cle emergence, seedlings were transplanted into three common gardens located natural tree fall gaps on Buena Vista Peninsula. Each garden was 7 x 7 m2 and enclosed in a hardwire fence to exclude vertebrate herbivores. Based upon the mean percentage of total daily PFD in the center of the plot relative to total daily PFD above the canopy, the light envi ronment in each plot was determined to be 50.7%, 37.7%, 23.4% of full sun in pl ots 1, 2 and 3, respectively (Kitajima unpublished data). Newly expanded leaves were marked with unique identification numbers and monitored for survival during m onthly census. Median leaf lifespan was estimated for each species at the time from full leaf expansion to abscission with the Kaplan-Meyer method (Donovan and Weldin 2002), which accounts for censored leaves (leaves removed from the study before their deat h or alive at the last survival census). Measurement of Light and Phot osynthesis in the Field I selected eleven species of tropical trees common to BCN M that varied in median leaf lifespan and shade tolerance for sampling (Table 1-1). Species are hereafter referred to by their generic names. Median l eaf lifespan varied between 140 days in Ceiba and 792 days in Aspidosperma All species are canopy trees, except for Pentagonia which is a small tree commonly found in gaps but also encountered in the shaded understory. Seeds of Aspidosperma, Ceiba, and Vochysia are wind-dispersed, whil e those of the other

PAGE 19

9 9 eight species are dispersed by birds and mammals. Ceiba, Pentagonia, and Vochysia are early successional species while other spec ies are intermediately shade tolerant ( Tabernaemontana, Genipa ) or very shade tolerant ( Aspidosperma Calophyllum, Gustavia, Tetragastris, Trichilia and Virola ) based on the abundance and survival of juveniles in the shaded u nderstory of BCNM (Kitajima personal observation, CTFS data set available online, S. J. Wright unpublished data). To include a wide-range of leaf ages for each species, I selected four marked leaves of contrasting ages on each individual seed ling for measurements using the leaf census data (Table 1-2). For each selected leaf, I measured the rates of light-saturated net CO2 assimilation (=photosynthetic capacity per unit area, Aarea), stomatal conductance (Gs) at photon flux density (PFD) of 1000 mol m-2 s-1, and then measured dark respiration with a portable infra-red gas analyzer (LI-6400, LI COR, Lincoln, Nebraska USA). Light was supplied with blue-red light emitting diodes. The CO2 concentration of the reference air entering the leaf chamber was adjusted with a CO2 mixer control unit so that the “reference” air entering the chamber had [CO2] = 38 Pa. Sample [CO2] ranged from 36.4 to 38.2 Pa. The chamber temperature was stabilized by maintaining the Peltier block temperature at 28C. The relativ e humidity of the sample was kept as close to ambient as possible (typically between 70-85%) and air flow rate was set at 400 mol s-1. Gasexchange measurements were made betw een 0800-1200 h between May and August of 2004. Additional leaves were measured be tween July and September of 2005 to supplement the age range for sp ecies with long-lived leaves. Total daily PFD at the leaf surface was measured for 3 days for a subset of leaves representing the full range of leaf age of fi ve species after gas ex change measurements

PAGE 20

10 10 (Table 1-3). A calibrated GaAsP sensor (Ham amatsu, Japan) was attached to the adaxial surface of each sampled leaf. Campbell Dataloggers (Models 21X and 10X) sampled PFD sampled every 2 s and recorded a mean fo r each minute. The results were expressed as %PFD (the mean percentage of the tota l daily PFD above the canopy) for each leaf. All light measurements were made between May and August of 2004. Laboratory Measurements and Statistical Analyses Leaf mass per area (LMA) for each l eaf was determined from two 19.6 cm2 leaf disks dried at 60C until a stable dry mass was reached. Nitrogen content per unit mass (Nmass) and per unit area (Narea) were determined for these leaf disks with an elemental analyzer (Costech Analytical Model 4010, Va lencia, California, USA). Instantaneous photosynthetic nitrogen use efficiency (PNUE) was determined by dividing Aarea by Narea. Statistical analyses, including leaf surviv al analysis and regression and correlation analyses among leaf traits, were comp leted using JMP V.5.1 (SAS, 2003). The relationships of Aarea, Amass, LMA, Narea, and PNUE with leaf age were modeled with ANCOVA for each species with leaf age as th e covariate and plant as a random effect. The rate of photosynthetic dec line (parameter a/b) was estimated as the slope of the regression line of leaf age and Aarea or Amass (hereafter leaf age-Aarea or leaf age-Amass). The y and x intercept of the leaf age-Aarea or leaf age-Amass regression was used as an estimate initial photosynthetic capacity (p arameter a) and the leaf age when photosynthesis would equal zero (parameter b) respectively. Significance of PFD relationships with leaf age and Aarea were tested with a simple linear regression with leaves, plants, and plots pool ed for each species.

PAGE 21

11 11 Results Aarea declined significantly with leaf age in all species (Figure 11). The rate of decline varied among species as in dicated by regression slopes of Aarea against leaf age that ranged from -0.008 to -0.039 mol CO2 m-2 s-1 d-1. Amass also declined significantly with leaf age in all species (Fi gure 1-2). Regression slopes of Amass against leaf age ranged from -0.0007 to -0.0001 mol CO2 g-2 s-1 d-1. While there was no relationship between initial Aarea and median leaf lifespan among species, the regression slope of Aarea against leaf age increased significantly w ith increasing median leaf lifespan among species (Figure 1-3a and 1-3c). Although it was stil l included in statis tical analyses, the slope value for Vochysia was outside of the 95th percentile, and was thus considered an outlier. It had the steepest slope on both a ma ss and area basis, but only an intermediate median leaf lifespan (Figure 1-3a and 1-3c). Initial Amass decreased with median leaf lifespan, whereas the slope of Amass against leaf age increased with median leaf lifespan (Figure 1-3b and 1-3d). In all species, x intercepts were greater than median leaf lifespan by 89 ( Trichilia ) to 403 days ( Ceiba ) when photosynthetic capacity was m easured on an area basis (Table 1-4). When photosynthetic capacity was measured on a mass basis, x intercepts overestimated median leaf lifespan by 4 ( Gustavia ) to 403 days ( Aspidosperma ) and underestimated median leaf lifespan by 30 days in Trichilia (Table 1-4). There was a positive relationship between median leaf lifespan and the x intercept of speciesÂ’ leaf ageAarea and leaf age-Amass regression (Figure 1-4). There was no consistent re lationship within species between LMA and leaf age (Figure 1-5). In some species, LMA increased with leaf age ( Ceiba, Genipa Gustavia Tetragastris and Trichilia ) in others, it decreased ( Calophyllum and Pentagonia ), while

PAGE 22

12 12 in Aspidosperma, Tabernaemontana, Virola, and Vochysia there was no relationship. The percent change in LMA from a 30-day ol d leaf to a 300-day ol d leaf ranged between -5% in Calophyllum and 35% in Genipa Seven out of the eleven species showed a decline in Narea with leaf age (Figure 1-6), but PN UE decreased with leaf age in all species (Figure 1-7). Light availability declined significantly with leaf age in the five species for light data were available (Figure 1-8). Aarea was positively correlated with light availability in three out of five species (Figure 1-9) which also showed a decline in Narea with leaf age (Figure 1-5). In contrast, Aspidosperma and Genipa did not increase in Aarea with leaf age although they did decline in Narea with leaf age. Discussion This paper provides further support for a ne gative relationship be tween leaf age and photosynthetic capacity across sp ecies (Reich et al. 1992, Reich et al. 1999, Wright et al. 2004) while also supplying rare data on the ra te of photosynthetic d ecline with leaf age which allows for the evaluation of cost-benef it theories of leaf lifespan (Kikuzawa 1991, Kikuzawa and Ackerly 1999, Kitajima et al. unpublished manuscript). Cost benefit theories have generated two alternative optim ization criteria for predicting leaf lifespan (Kikuzawa 1991, Kitajima et al. unpublished manuscript). Depending upon whether or not external resources are limiting, a leaf should either maximize total lifetime carbon gain (Case 1) or average daily carbon gain (Case 2). In Case 1, leaf lifespan should be similar to the x intercept of the Aarea-leaf age relationship (parameter b in KikuzawaÂ’s model). Because total daily net photosynthetic income should reach zero before photosynthetic capacity reaches zero, the x intercept is expected to be greater than the actual leaf life time of a give n species. In a previous study, estimated x-intercepts

PAGE 23

13 13 approximated the mean leaf lifespan for seedlin gs of a tropical pione er tree with very short mean leaf lifetimes (Ackerly and B azzaz 1995; Table 1-5). In contrast, the xintercept of the Aarea-leaf age relationship in this study consistently overestimated median leaf lifespan by 89-403 days (Table 1-4). Kitajima and others (1997) also reported xintercepts that were greater than the mean leaf life times of five tropical tree species (Table 5). Contrary to an earlier study by Kitajima and others (2002), I found no relationship between the discrepancy in the x intercept and median leaf lifespan with species-specific median leaf lifespans. Instead, there was a significant positive relationship between the x intercepts and median leaf lifespans which provides qualitative support for KikuzawaÂ’s model. However, rates of photosynthetic capacity declined sharply before reaching the poten tial leaf lifespan b suggesting that actual leaf lifespan is much shorter than the potential. Thus, it ap pears that Case 1 app lies only to very shortlived leaves. The results of this study supported predic tions that leaves will maximize daily average carbon gain (Case 2) as suggested by KikuzawaÂ’s co st-benefit model (1991) and two earlier empirical studies (Kitajima et al. 1997, Kitajima et al. 2002). In general, species with shorter le af lifespans had (1) a higher ini tial photosynthetic capacity and (2) a steeper rate of decline of photosynthetic capacity with leaf age (parameter a/b in KikuzawaÂ’s model) than species wi th longer leaf lifespans. Aarea-leaf age slopes were steeper for the five species with short medi an leaf lifespans (140280 days) than the six shade-tolerant species with long median leaf lifespans (504-797 days; Table 1-4). These slope values are similar to Aarea-leaf age slopes reported in pr evious studies (Table 5). Kitajima et al (1997) found Aarea-leaf age slopes ranging from -0.032 to -0.0018 mol

PAGE 24

14 14 CO2 m-2 d-1 for adults of five tropical tree speci es with mean leaf lifespans of 174-315 days. Adults of two tropical pioneer species with very shor t leaf lifespans (74-93 days) had much steeper slopes (-0.2 to -0.25 mol CO2 m-2 d-1; (Kitajima et al. 2002). Even steeper slopes between -0.57 and -1.32 mol CO2 m-2 d-1 have been documented for seedlings of Heliocarpus appendiculatus a tropical pioneer tree with mean leaf life times of only 28-37 days (Ackerly and Bazzaz 1995). Although a general dichotomy between Aarea-leaf age slopes of species with short and long-lived leaves is evident, some sp ecies do not fit the predicted pattern. For instance, Vochysia had both the steepest rate of phot osynthetic decline with leaf age and the highest initial Aarea, yet had the longest median leaf lifespan (280 days) of the pioneer species studied. The species with the longest median leaf lifespan (797 days) Aspidosperma had the shallowest slope, but not the lowest initial Aarea as KikuzawaÂ’s cost-benefit analysis w ould predict. Nevertheless, this study still provides strong support for Case 2 which provides leaf lifetime optimi zation criteria for plan ts of various leaf lifespans that are primarily lim ited by external resources. Species differences weakened the among species relationship between Amass and leaf lifespan (Figure 1-3b, r2 = 0.56, p = 0.009). Previous st udies have documented stronger correlations (r2 values typically between 0.700.91, with p-values < 0.001; Reich et al. 1992, Reich et al. 2004, Wri ght et al. 2004). I believe th is can be explained by the greater range in leaf lifespans as well as the greater number of species measured in previous studies. Reich et al. (1992) reported leaf lifespa ns up to 4200 days, while the maximum leaf lifespan in my study was 792 days This five-fold greater range in leaf lifespans could result in stronge r correlations, particularly when coupled with a greater

PAGE 25

15 15 number of species for comparison. The abse nce of a significant relationship between initial Aarea and leaf lifespan among species has been frequently cited in earlier studies (Reich et al. 1992, Reich et al 2004, Wright et al. 2004) and is due to the effect of LMA on photosynthetic capacity (Peterson 1999). Wh en LMA is controlled statistically, the slope of the relationship betw een photosynthesis and leaf ni trogen is the same for both area and mass-based expressions (Peterson 1999). Aarea, Amass, Narea, and PFD decreased with leaf age as predicted for all species measured while LMA did not vary consistently with leaf age across species. In most species, the changes in LMA with leaf age were so small that they could not have contributed much to the decrease of Aarea, Amass, and Narea. In contrast, the observed decline in Narea with leaf age did contribute to th e decline in photos ynthetic capacity in seven out of the eleven species and was likel y caused by the reallocation of nitrogen to newer leaves where PFD was greater. Inde ed, due to shading by younger leaves, leaf light levels decreased with l eaf age for the five species for which leaf PFD was measured in this study. This pattern is consistent w ith theories of optimal resource allocation in which nitrogen from older shaded leaves is reallocated to younger unshaded leaves, resulting in a more favorable carbon gain for the entire plant (Field and Mooney 1983, Hirose and Werger 1987, Hikosak a et al. 1994). In all spec ies, the decrease in PNUE with leaf age also contributed to the decline of photosyntheti c capacity. The retention of leaves despite their low photosynthetic cap acity and PNUE has been explained as a means of increasing total production due to the longer retentio n of plant nitrogen (Escudero and Mediavilla 2003).

PAGE 26

16 16 This study is apparently the first to evaluate whether co st-benefit theory explains variation in leaf lifespan in several species of seedlings ra nging in shade tolerance and median leaf lifespan. I have no data on the leaf lifespan of adult life forms. However, sapling leaf lifespans were measured for th ree of the eleven species I studied (Coley 1988, King 2003). Gap-grown saplings of Tetragastris and Calophyllum 1-2 years of age had mean leaf lifespans of 690 and 1050 days respectively (Coley 1988). This value is only a little greater th an the median leaf lifespan of 671 days for Tetragastris reported in my study. For Calophyllum 1050 days is much greater than the median leaf lifespan of 700 days reported in this study. King (1993) evaluated the mean leaf lifespans of Trichilia and Calophyllum saplings grown over a range of light environments and found mean leaf lifespans of 1350 and 2640 days, respectively, which is much greater than the values of 630 and 700 days reported in this study. However, since King (1993) evaluated mean leaf lifespan in saplings across gap, unde rstory and intermediate light conditions, it is difficult to account for the effect of light on leaf lifespan. Leaf lifespan is expected to be greater in shaded than in gap conditions (Bongers a nd Pompa 1990). For all three species in which sapling leaf lifespan was meas ured, leaf lifespan was greater in saplings than in seedlings, providing support for the id ea that leaf lifespan should increase with increasing construction and maintenance cost s of supporting struct ures (Kikuzawa and Ackerly 1999). In order to evaluate this hypothesis further, sapling and adult leaf lifespans should be measured for the eleven species in this st udy under similar light conditions. In summary, general predictions made by KikuzawaÂ’s cost-benefit model (1991) were supported by the results of this study. Leaf lifespan wa s shorter for species with

PAGE 27

17 17 higher initial photos ynthetic rates and longe r when the rate of phot osynthetic decline with leaf age was slow. The x-intercept of the Aarea-leaf age relationship consistently overestimated median leaf lif espan, indicating that the obser ved leaf lifespan was much shorter than the potential. Thus, it appears the optimal st rategy for the seedlings of tropical tree species in this st udy is to maximize daily carbon gain as external resources are limiting (Case 2). Leaf age was a st rong determinant of p hotosynthetic capacity, supporting the idea that knowledge of the distribution of leaf age in a canopy will allow for estimates of canopy photosynthetic performance.

PAGE 28

18 18 Table 1-1. Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer method for each species studied. Species are ranked from shortest median leaf lifespan to longest median leaf lifespan. Nomenclature follows Croat (1978). Species Family Total Leaves Median Mean Std Dev Ceiba pentandra Bombacaceae 296 140 156.1 3.10 Genipa americana Rubiaceae 478 224 234.8 4.96 Pentagonia macrophylla Rubiaceae 304 252 274.2 6.03 Tabernaemontana arborea Apocynaceae 542 279 277.6 3.94 Vochysia ferruginea Vochysiaceae 549 280 280.3 3.41 Virola surinamensis Myristicaceae 709 504 494.3 6.56 Gustavia superba Lecythidacaceae111 629 591.2 17.17 Trichilia tuberculata Meliaceae 59 630 626.1 20.12 Tetragastris panamensis Burseraceae 89 671 648.3 13.87 Calophyllum longifolium Clusiaceae 108 700 679.9 12.79 Aspidosperma cruenta Apocynaceae 31 797 745.3 12.51 Table 1-2. Number of plants and leaves measured for Amax, Aarea, LMA, Narea, and PNUE measurements for each species. Species are ranked from shortest median leaf lifespan to longest me dian leaf lifespan. # of Leaves Species # of Plants AareaAmassLMANarea PNUE Ceiba 7 28 28 28 28 28 Genipa 11 43 43 43 43 43 Pentagonia 8 31 31 31 31 31 Tabernaemontana 9 32 32 32 32 32 Vochysia 13 42 42 42 42 42 Virola 10 105 51 78 83 78 Gustavia 8 44 32 39 39 39 Trichilia 5 27 27 36 36 36 Tetragasteris 7 46 30 38 38 38 Calophyllum 6 42 24 41 41 41 Aspidosperma 9 69 38 49 50 49

PAGE 29

19 19 Table 1-3. Number of plants and leaves for which % PFD and Aarea were measured for a subset of species. Species are ranked fr om shortest median leaf lifespan to longest median leaf lifespan. # of Leaves Species # of Plants %PFD % PFD and Aarea Genipa 5 19 9 Tabernaemontana 2 10 8 Virola 4 17 13 Tetragasteris 4 17 15 Aspidosperma 2 8 8

PAGE 30

20Table 1-4. Mean leaf lifespan and regression statistics ( x and y intercepts and slopes) for areaand mass-based light-saturated net photosynthesis against leaf age (days after full expansion) for seedlings of eleven tropical tree species. Aarea (mol m-2 s-1) Amass (mol g-1 s-1) Species Median Leaf Lifespan x intercept ( b ) y intercept ( a ) Slope ( a/b ) x intercept ( b ) y intercept ( a ) Slope ( a/b ) Ceiba 140 543 11.4 -0.021 443 0.31 -0.0007 Genipa 224 512 8.7 -0.017 400 0.20 -0.0005 Pentagonia 252 525 8.4 -0.016 700 0.14 -0.0002 Tabernaemontana 279 489 9.3 -0.019 433 0.26 -0.0006 Vochysia 280 369 14.4 -0.039 400 0.28 -0.0007 Virola 504 730 7.3 -0.010 750 0.15 -0.0002 Gustavia 629 882 9.7 -0.011 633 0.19 -0.0003 Trichilia 630 800 8.0 -0.010 600 0.12 -0.0002 Tetragastris 671 888 7.1 -0.008 700 0.14 -0.0002 Calophyllum 700 887 13.3 -0.015 750 0.15 -0.0002 A s p idos p erma 79711259.0-0.0081200 0.12-0.0001 Table 1-5. Mean leaf lifesp an and regression statistics ( x and y intercepts and slopes) for light -saturated net photosynthesis (Aarea in micromols of CO2 m-2 s-1) against leaf age (days after full expansion) for eight tropical tree species measured in previous studies (Ackerly and Bazazz 1995, Kitajima et al 1997, 2002). Since measurements were made for a range of nutrient and light conditions for Heliocarpus appendiculatus average values for each parameter ar e reported. Species are arranged from shortest to longest mean leaf lifespan. Species Mean Leaf Lifespan x intercept ( b ) y intercept ( a ) Slope ( a/b ) Source Heliocarpus appendiculatus 33 36 14.6 -0.915 Ackerly and Bazzaz 1995 Cecropia longipes 74 133 23.7 -0.179 Kitajima et al. 2002 Urera caracasana 93 98 21.2 -0.255 Kitajima et al. 2002 Antirrhoea trichantha 174 550 12.1 -0.022 Kitajima et al. 1997 Luehea seemannii 201 1200 10.8 -0.009 Kitajima et al. 1997 Castilla elastica 206 365 11.7 -0.032 Kitajima et al. 1997 Annona spraguei 221 482 10.6 -0.022 Kitajima et al. 1997 Anacardium excelsum 315 985 6.9 -0.007 Kitajima et al. 1997

PAGE 31

21 5 10 15 20 0 5 10 15 20 5 10 15 20 Ceiba y = -0.021x + 11.43 r2 = 0.56, P = 0.0012 5 10 15 20 Tabernaemontana y = -0.019x + 9.31 r2 = 0.78, P < 0.0001 Pentagonia y = -0.016x + 8.40 r2 = 0.63, P < 0.0001 02004006008001000 0 5 10 15 20 Vochysia y = -0.039x + 14.35 r2 = 0.76, P < 0.0001 5 10 15 20 Genipa y = -0.017x + 8.66 r2 = 0.63, P < 0.0001 02004006008001000 0 5 10 15 20 Aspidosperma y = -0.008x + 8.98 r2 = 0.71, P < 0.0001 5 10 15 20 Caloyphyllum y = -0.015x +13.33 r2 = 0.85, P < 0.0001 Virola y = -0.010x + 7.33 r2 = 0.77, P < 0.0001 0 5 10 15 20 Tetragastris y = -0.008x +7.08 r2 = 0.64, P < 0.0001 5 10 15 20 Trichilia y = -0.010x + 7.99 r2 = 0.38, P = 0.0009 0 5 10 15 20 Aarea ( mol m-2 s-1)Leaf Age (days) Leaf Age (days) Gustavia y = -0.011x + 9.71 r2 = 0.70, P < 0.0001 Figure 1-1. Light saturated photos ynthetic assimilation rates (Aarea) of leaves of different age on seedlings of eleven tropical tree species. Each point represents a leaf. Circles were leaves measured in 2004; triangles were leaves measured in 2005.

PAGE 32

22 22 0.1 0.2 0.3 0.4 0.5 Ceiba y = -0.0007x + 0.31 r2 = 0.46, P = 0.0005 0.1 0.2 0.3 0.4 0.5 Tabernaemontana y = -0.0006x + 0.26 r2 = 0.83, P < 0.0001 0.1 0.2 0.3 0.4 0.5 Pentagonia y = -0.0002x + 0.14 r2 = 0.70, P < 0.0001 0200400600800 0.0 0.1 0.2 0.3 0.4 0.5 Vochysia y = -0.0007x + 0.28 r2 = 0.76, P < 0.0001 0.1 0.2 0.3 0.4 0.5 Genipa y = -0.0005x + 0.20 r2 = 0.72, P < 0.0001 0200400600800 0.0 0.1 0.2 0.3 0.4 0.5 Aspidosperma y = -0.0001x + 0.12 r2 = 0.57, P = 0.0001 0.1 0.2 0.3 0.4 0.5 Calophyllum y = -0.0002x + 0.15 r2 = 0.87, P < 0.0001 0.1 0.2 0.3 0.4 0.5 Virola y = -0.0002x + 0.15 r2 = 0.70, P < 0.0001 0.1 0.2 0.3 0.4 0.5 Tetragastris y = -0.0002x + 0.14 r2 = 0.88, P < 0.0001 0.1 0.2 0.3 0.4 0.5 Trichilia y = -0.0002x + 0.12 r2 = 0.48, P = 0.0002 0.1 0.2 0.3 0.4 0.5 Gustavia y = -0.0003x + 0.19 r2 = 0.82, P < 0.0001Amass ( mol g-1 s-1)Leaf Age (days) Leaf Age (days) Figure 1-2. Mass-based photosynthesis (Amass) of leaves with cont rasting leaf age in eleven species of tropical trees. E ach point represents a leaf.

PAGE 33

23 23 Initial Aarea ( mol m-2 s-1) 2 4 6 8 10 12 14 16 0200400600800 Slope of Aarea ( mol m-2 s-1) -0.05 -0.04 -0.03 -0.02 -0.01 0.00 Initial Amass ( mol g-1 s-1) 0.1 0.2 0.3 0.4 0200400600800 Slope of Amass ( mol g-1 s-1) -0.0008 -0.0007 -0.0006 0.0005 0.0004 0. 00 03 -0.0002 -0.0001 0.0000 a c b d Median leaf lifespan (days)p = 0.0093, r2 = 0.55 p = 0.5, r2 = 0.04 p = 0.0088, r2 = 0.55 p = 0.0334, r2 = 0.41 Vochysia Vochysia Figure 1-3. Relationship between initial Aarea, Aarea-leaf age slope, Amass, Amass-leaf age slope with median leaf lifespan for each species. (a) Relationship between initial Aarea and median leaf lifespan for each species. (b) Relationship between the Aarea-leaf age slope and the median leaf lifespan for each species. (c) Relationship initial Amass and median leaf lifespa n for each species. (d) Relationship between the Amass-leaf age slope and median leaf lifespan for each species.

PAGE 34

24 24 Median Leaf Lifespan (days) 100200300400500600700800900 Parameter b (days) 200 400 600 800 1000 1200 1400 y = 0.9x + 271.2 y = 0.8x + 283.4 Figure 1-4. The relationship betw een the actual median leaf lifespan and an estimate of the time when photosynthetic capac ity would reach zero (parameter b the x intercept of the leaf age-Aarea and leaf age-Amass regression). Solid circles and the solid line represent the relations hip when photosynthetic capacity is presented on an area basis (r2 = 0.87, P < 0.0001), while empty circles and the dashed line represent the relations hip when photosynthetic capacity is presented on a mass basis (r2 = 0.58, P < 0.01).

PAGE 35

25 25 20 40 60 80 100 120 140 Ceiba y = 0.029x + 38.00 r2 = 0.86, P = 0.0454 20 40 60 80 100 120 140 Tabernaemontana y = 0.005x + 36.65 r2 = 0.19, P = 0.5243 20 40 60 80 100 120 140 Pentagonia y = -0.036x + 60.34 r2 = 0.40, P = 0.0031 0200400600800 0 20 40 60 80 100 120 140 Vochysia y = -0.020x + 51.52 r2 = 0.57, P = 0.0577 20 40 60 80 100 120 140 Genipa y = 0.056x + 40.8 r2 = 0.84, P < 0.0001 02004006008001000 0 20 40 60 80 100 120 140 Aspidosperma y = 0.007x + 78.55 r2 = 0.54, P = 0.1935 0 20 40 60 80 100 120 140 Calophyllum y = -0.016x + 98.46 r2 = 0.39, P = 0.0105 20 40 60 80 100 120 140 Virola y = -0.0008x + 53.34 r2 = 0.42, P = 0.7944 20 40 60 80 100 120 140 Tetragastris y = 0.010x + 57.12 r2 = 0.60, P = 0.0246 20 40 60 80 100 120 140 Trichilia y = 0.013x + 68.69 r2 = 0.67, P = 0.0147 20 40 60 80 100 120 140 Leaf Age (days) Leaf Age (days)LMA (g m-2)Gustavia y = 0.020x + 51.99 r2 = 0.34 P = 0.0052 Figure 1-5. Relationship between leaf mass per unit area (g m-2) and leaf age in seedlings of eleven species of tropical trees. E ach point represents a leaf. Circles were leaves measured in 2004; triangles were leaves measured in 2005. Significant and nonsignificant regression lines are indicated by solid and broken lines, respectively.

PAGE 36

26 26 0.5 1.0 1.5 2.0 2.5 Ceiba y = -0.0002x + 1.18 r2 = 0.66, P = 0.6868 0.5 1.0 1.5 2.0 2.5 Tabernaemontana y = -0.001x + 1.12 r2 = 0.47, P = 0.0005 0.5 1.0 1.5 2.0 2.5 Pentagonia y = -0.001x + 1.02 r2 = 0.43, P = 0.0027 02004006008001000 0.0 0.5 1.0 1.5 2.0 2.5 Vochysia y = -0.002x + 1.56 r2 = 0.63, P < 0.0001 0.5 1.0 1.5 2.0 2.5 02004006008001000 0.0 0.5 1.0 1.5 2.0 2.5 Aspidosperma y = -0.0002x + 1.32 r2 = 0.72, P = 0.0461 0.5 1.0 1.5 2.0 2.5 Calophyllum y = -0.0007x + 1.35 r2 = 0.62, P < 0.0001 0.5 1.0 1.5 2.0 2.5 Virola y = -0.0005x + 1.01 r2 = 0.44, P < 0.0001 0.5 1.0 1.5 2.0 2.5 Tetragastris y = -0.0003x + 1.01 r2 = 0.24, P = 0.0604 0.5 1.0 1.5 2.0 2.5 Trichilia y = -0.0006x + 1.49 r2 = 0.54, P = 0.0004 0.5 1.0 1.5 2.0 2.5 Gustavia y = -0.0001x + 1.57 r2 = 0.40 P = 0.5379 Leaf Age (days) Leaf Age (days)Leaf N (g/m2)Genipa y = 0.0001x + 0.82 r2 = 0.69, P = 0.6559 Figure 1-6. Relationship betw een leaf nitrogen (g m-2) and leaf age. Each point represents a leaf. Circles were leaves measured in 2004; triangles were leaves measured in 2005. Significant and n onsignificant regression lines are indicated by solid and broken lines, respectively.

PAGE 37

27 27 2 4 6 8 10 12 14 Ceiba y = -0.019x + 10.03 r2 = 0.42, P = 0.0005 2 4 6 8 10 12 14 Tabernaemontana y = -0.015x + 9.04 r2 = 0.66, P < 0.0001 2 4 6 8 10 12 14 Pentaqonia y = -0.012x + 8.50 r2 = 0.72, P < 0.0001 02004006008001000 0 2 4 6 8 10 12 14 Vochysia y = -0.020x + 9.69 r2 = 0.54, P < 0.0001 2 4 6 8 10 12 14 Genipa y = -0.020x + 10.37 r2 = 0.65, P < 0.0001 02004006008001000 0 2 4 6 8 10 12 14 Aspidosperma y = -0.005x + 7.10 r2 = 0.74, P = 0.0011 2 4 6 8 10 12 14 Calophyllum y = -0.010x + 10.75 r2 = 0.86, P < 0.0001 2 4 6 8 10 12 14 Virola y = -0.010x + 7.73 r2 = 0.62, P < 0.0001 2 4 6 8 10 12 14 Tetragastris y = -0.012x + 7.90 r2 = 0.81, P < 0.0001 2 4 6 8 10 12 14 Trichilia y = -0.007x + 5.64 r2 = 0.33, P = 0.0022 2 4 6 8 10 12 14 Leaf Age (days) Leaf Age (days)PNUE ( mol g-1N s-1)Gustavia y = -0.008x + 6.60 r2 = 0.68, P < 0.0001 Figure 1-7. Relationship between photosynthetic nitrog en use efficiency (PNUE) and leaf age. Each point represents a leaf.

PAGE 38

28 28 0200400600800 0 10 20 30 40 50 Tabernaemontana y = -0.024x + 8.58 r2 = 0.549, P = 0.0355 10 20 30 40 50 Genipa y = -0.122x + 29.93 r2 = 0.380, P = 0.0050 10 20 30 40 50 Virola y = -0.016x + 10.27 r2 = 0.505, P = 0.0065 10 20 30 40 50 Tetragasteris y = -0.009x + 6.49 r2 = 0.396, P = 0.0159 0200400600800 0 10 20 30 40 50 Aspidosperma y = -0.072x + 31.24 r2 = 0.810, P = 0.0023Leaf Age (Days)% PFDLeaf Age (Days) Figure 1-8. Decline in %PFD (the total daily PFD incident on the leaf relative to the total daily PFD above the canopy) with leaf age in a subset of species and leaves. Each point represents a leaf.

PAGE 39

29 29 010203040 2 4 6 8 10 12 14 Tabernaemontana y = 0.490x + 4.93 r2 = 0.668, P = 0.0145 2 4 6 8 10 12 14 Genipa r2 = 0.0, P = 0.8 2 4 6 8 10 12 14 Virola y = 0.489x + 1.79 r2 = 0.627, P = 0.0021 2 4 6 8 10 12 14 Tetragasteris y = 0.655x + 1.15 r2 = 0.542, P = 0.0027 010203040 2 4 6 8 10 12 14 Aspidosperma r2 = 0.14, P = 0.4Aarea ( mol m-2 s-1)% PFD % PFD Figure 1-9. Effect of % PFD (the total daily PFD incident on the leaf relative to the total daily PFD above the canopy) on light-saturated photosynthesis (Aarea) and for a subset of species and leaves. Each point represents a leaf. Significant and nonsignificant regression lines are indicated by solid and broken lines, respectively.

PAGE 40

30 CHAPTER 2 LEAF AGE, PHOTOSYNTHETIC CAPACITY, AND LEAF LIGHT LEVEL AS PREDICTORS OF DAILY CARBON GAIN IN SEEDLINGS OF FIVE TROPICAL TREE SPECIES Introduction Cost-benefit theories of l eaf lifespan are a powerful appr oach that explains global patterns of leaf lif espan across latitudes (Kikuzaw a 1991, Kikuzawa and Ackerly 1999) and the negative correlation betw een leaf lifespan and photosynth etic capacity (Reich et al. 1997, Reich et al. 1999, Wri ght et al. 2004). A basic as sumption central to costbenefit theory is that leaves, as the pr imary carbon gaining organs of plants, must optimize photosynthetic gain (Aday) relative to construction and maintenance costs (Chabot and Hicks 1982). Because photosynthetic characteristics change with leaf age, leaf-age effects are an important consideration in cost-benefit theories of leaf lifespan. Understanding the relationship between photosynthet ic gain and leaf age can facilitate the integration of photosynthetic car bon gain from individual leaves to individual crowns to the entire forest canopy (Kitajima et al 1997, Kitajima et al. 2002). Because Aday is difficult to measure directly for many leav es, previous studies have focused on the relationship between photosynthetic cap acity and leaf age, assuming that Aday can be approximated by maximum phot osynthetic capacity, or Amax (Kitajima et al. 1997, Kitajima et al. 2002, Kitajima et al. unpub lished manuscript, Chapter 1). This assumption is based upon a strong correlation between Amax and Aday reported across several tropical canopy tree and epiphyte species (Zot z and Winter 1993, 1994). However, the degree to which Aday, not Amax, varies with leaf environmental and

PAGE 41

31 physiological factors has not been adequately explored. It is particularly important to elucidate the relationship between Aday and Amax as predictions made using cost-benefit theory are based on Aday, not Amax (Kikuzawa 1991). Photosynthetic capacity (Amax) and daily (24 hour) photosynthetic gain (Aday) typically decline monotonically with leaf age and can be approximated by two similar equations that differ only by the notation ’ whic h indicates daily rather than instantaneous parameters. The equations are: Amax = a (1 – t/ b ) or Aday = a’ (1 – t/ b’ ), where t is the leaf age in days from the time of full leaf expansion (Kikuzawa 1991). The parameters a and a’ represent the photosynthetic rate at the time of full leaf expansion. The parameters b and b’ is a statistical extrapolation determined as a function of the initial photosynthetic rate ( a or a’ ) and the rate of its decline (parameter a/b or a’/b’ ). Thus, parameter b is an estimate of the time when photosynthetic capac ity would reach zero. Studies exploring the relationship between leaf age and photosynthetic capaci ty have reported a tight relationship between leaf lifespan and the parameter a/b indicating that it is possible to estimate the effect of leaf age on photosyntheti c gain if the mean leaf lifespan of the species is known (Kitajima et al. 1997, Kitaji ma et al. 2002, Kitajima et al. unpublished manuscript, Chapter 1). Based upon the approximately linear decl ine of photosynthesis with leaf age (Kitajima et al. 1997, Kitajima et al. 2002, Chap ter 1), cost-benefit analyses of leaf lifespan have generated two alternate predictions If there are no exte rnal constraints on the maximum number of leaves produced and ma intained (Case 1), a leaf is expected to maximize net carbon gain over its entire lifetime (Kikuzawa 1991, Kitajima et al. unpublished manuscript). In this case, the opt imum leaf lifespan s hould be close to the

PAGE 42

32 time when photosynthetic rates equal zero (parameter b ). Alternatively, if the total number of leaves is limited by external resource s (Case 2), a leaf is expected to maximize the rate of net carbon gain averaged over the lifetime of the leaf (Kikuzawa 1991, Kikuzawa and Ackerly 1999). In this scenari o, leaves are replaced when their net carbon gain per unit time over their entire lifespan has reached a maximum and optimal leaf lifespan is predicted to be equal to (2 bC / a )1/2 where C is the construction cost of the leaf (Kikuzawa 1991). Based on this theoretical mode l, it is expected that leaf lifespan will be short when the initial photosynthetic ra te of the leaf is high, and long when construction costs are high or the rate of dec line of photosynthetic rates with leaf age is slow (Kikuzawa 1991). Daily photosynthetic gain can be affected by both environmental and physiological factors, some of which are directly correlate d with leaf ageing. Environmental factors that affect daily photosynthetic gain can vary unpredictably over short time periods, while other factors change predictably ove r the lifetime of a l eaf. For instance, photosynthetic gain is expected to vary in response to daily changes in weather, with decreases in carbon gain duri ng cloudy conditions. Alterna tively, over the course of a leafÂ’s lifetime, a leaf can become increasingly shaded as new leaves are produced at higher positions on the plant (Field 1 983, Ackerly and Bazzaz 1995, Ackerly 1999). A reduction in light levels expe rienced by the leaf, whether due to self-shading, shading by surrounding vegetation, or weather, contributes to a decrease in photosynthetic gain as lower light levels directly redu ce photosynthetic ra tes (Field 1983). Leaf physiological factors such as photos ynthetic capacity and le af nitrogen also have a substantial impact on daily leaf carbon ga in and generally decrea se with leaf life

PAGE 43

33 span across plant species and communities (Reich et al. 1991, Reich et al. 1992, Reich et al. 1999). Decreases in Narea directly contribute to decreases in Amass and subsequent decreases in daily carbon gain as leaves age. The decline of Amass and Nmass with leaf age is thought to be due to the retranslocation of nitrogen to younger leaves which maximizes whole-shoot photosynthetic gain rather than uncontrolled physiol ogical deterioration (Field and Mooney 1983, Ackerly a nd Bazzaz 1995). Hikosaka et al. (1994) showed that age-related changes due to nitrogen retransl ocation occurred primarily when light or nitrogen availabili ty is limiting. The primary objective of this study was to examine how physiological and environmental factors may predict daily car bon gain in seedlings of five tropical tree species that differ in median leaf lifespan a nd shade tolerance. Since mortality rates are typically highest in the seed and seedling life stages, understandi ng how seedling leaf carbon gain varies with physiological and e nvironmental factors can be helpful in understanding adult abundance and distribution. In partic ular, I was interested in determining whether changes in daily carbon gain were mo re highly correlated with Amax or photon flux density (PFD) at the leaf leve l. While some studies reported a high correlation between Amax and diel carbon gain (Zotz and Winter 1993, 1994), others have cited leaf light levels as th e determinant of photosynthetic capacity and resulting carbon gain (Chazdon 1986, Chazdon and Pearcy 1986) Previous studies have documented a strong correlation between l eaf age, leaf PFD, and Amax (Kitajima et al. 1997, Kitajima et al. 2002, Chapter 1). Thus, I hypothesized that leaf age would be hi ghly correlated with daily carbon gain since it directly affect s both photosynthetic ca pacity and leaf PFD (Chapter 1), which in turn di rectly affect carbon gain (Figur e 2-1). Given that leaf PFD

PAGE 44

34 has a strong and documented affect on Amax (Field 1983), and both variables affect carbon gain, I hypothesized that leaf PFD woul d be the best predictor of daily carbon gain. Methods Study Site The study was conducted on Buena Vista Pe ninsula, a 60-year-old secondary lowland tropical forest area in the Ba rro Colorado National Monument (BCNM), Republic of Panama (9 09Â’ N, 79 51Â’W). The species compos ition, climate, and ecology of the Buena Vista Peninsula are similar to that of the young forests on Barro Colorado Island (Croat 1978, Leigh 1982). The BCNM forest is semi-deciduous during the pronounced dry season that usually lasts from mid-December to April, and receives annual rainfall ca. 2,600 mm (Daws et al. 2002). Plot Establishment Common gardens were establis hed in three recent ~100 m2 tree fall gaps in May 2002 on Buena Vista Peninsula. Each garden was 7 x 7 m2 and enclosed in a hardwire fence with 1 cm mesh to exclude vertebrate herbivores. Seeds were collected on Barro Colorado Island from at least two parent trees and germinated in a greenhouse. Seedlings were transplanted at least 20cm apart at ra dicle emergence. Seedlings received water only from rainfall. Average daily rainfa ll from May to August of 2004 was 11.9 mm day1 (May = 15.2 mm day-1, June= 10.7 mm day-1, July= 8.8 mm day-1, August = 13.4 mm day-1; S. Patton, data available online). The to tal daily PFD in the center of plots 1, 2, and 3, were 50.7%, 37.7%, and 23.4% of th e light above the canopy, respectively (Kitajima unpublished data). Difference in light environment reflected differences in the size of the gaps. Understory vegetation and small stems were removed prior to

PAGE 45

35 transplanting seedlings. Newly expanded leav es were marked with unique identification numbers with a permanent marker and monito red for survival during monthly census. Median leaf lifespan was estimated for each sp ecies at the time from full leaf expansion to abscission with the Kaplan-Meyer method (Donovan and Weldin 2002), which accounts for censored leaves (leaves removed from the study before thei r death or alive at the last survival census). Measurement of Light and Phot osynthesis in the Field For this study, I selected five species of canopy trees common to BCNM that varied in median leaf lifespan and shade tolerance (T able 2-1; species are hereafter referred to by their generic names). Median leaf lifespan varied from 478 days in Genipa to 792 days in Aspidosperma Seeds of Aspidosperma are wind-dispersed, while those of the other four species are disp ersed by birds and mammals. Tabernaemontana and Genipa are intermediately shade tolerant while Aspidosperma Calophyllum, Gustavia, Tetragastris and Virola are very shade tolerant based on the abundance and survival of juveniles in the shaded understory of BCN M (Kitajima, personal observation, CTFS data set available online, S. J. Wright unpublished data). To include a wide-range of leaf ages for each species, I selected four marked leaves of contrasting ages on each individual seed ling using the leaf census data. For each selected leaf, I measured the rates of net CO2 assimilation (=photos ynthetic capacity per unit area, Aarea), and stomatal conductance (Gs) at photon flux dens ity (PFD) of 1000 mol m-2 s-1 supplied with blue-red emitting diodes, and then measured dark respiration with a portable infra-red gas analyzer (L I-6400, LICOR, Lincoln, Nebraska, USA). The CO2 concentration of the reference air enteri ng the leaf chamber was adjusted with a CO2 mixer control unit so that the “refer ence” air entering the chamber had [CO2] = 38 Pa.

PAGE 46

36 Sample [CO2] ranged from 36.4 to 38.2 Pa. The chamber temperature was controlled by maintaining the Peltier block temperature at 28C. The relative humidity of the sample was kept as close to ambient as possible (t ypically between 70-85 %) and air flow rate was set to 400 mol s-1. Gas-exchange measurements were made between 0800-1200 h between May and August of 2004. Additional le aves were measured between July and September of 2005 to supplement the age rang e for species with long-lived leaves. Photosynthetic light response curves were measured for two of the four leaves selected for gas exchange measurements us ing a portable infra-red gas analyzer (LI-6400, LICOR, Lincoln, Nebraska, USA). Leaves were selected to represent a range of leaf ages. Light supplied with blue-red emitting di odes was decreased in steps from 1000 to 0 mol photons m-2 s-1 after CO2 uptake rates reached a steady-s tate at each li ght level. After gas exchange measurements, total da ily PFD at the leaf surface was measured for 3 days between May and August of 2004 for a subset of leaves representing the full range of leaf ages (Table 2-2). A calib rated GaAsP sensor (Hamamatsu, Japan) was attached to the adaxial surf ace of each sampled leaf. Campbell Dataloggers (Models 21X and 10X) sampled PFD every 2 s and recorded a mean for each minute. Results were expressed as mmol PFD m-2 day-1. The % PFD (mean percentage of the total daily PFD above the canopy) for each leaf wa s also presented (Figure 2-2). Calculation of Daily Carbon Gain Photosynthetic light curves were fitted to non-recta ngular hyperbola (Johnson and Thornley 1984) using Photosyn Assistant (Version 1.1.2, Dundee Scientific, Dundee, UK). Best-fit estimates were made of th e maximum rate of light-saturated net CO2 assimilation (= maximum photosynt hetic capacity per unit area, Amax) as well as quantum yield ( ), curvature ( ), and dark respiration (R).

PAGE 47

37 I estimated net photosynthetic carbon ga in for each minute during the daytime (0630-1800 h) using the following equation: P(I) = { I + Amax – {( I + Amax)2 – 4 AmaxI }0.5 / 2 } – R where P(I) is the instantaneous photosynthetic ra te as a function of the apparent quantum yield ( ), I is the photon flux density incident on the leaf averaged over one minute, Amax is the light-saturated photosynt hetic capacity per unit area, is the curvature of the nonrectangular hyperbola, and R is area-based daytime respiration (Johnson and Thornley 1984). Carbon gain was estimated for each minute for which PFD was measured, then averaged over three days to calculate average diurnal car bon assimilation (Table 2-4). Nocturnal respiration was estimated to be 10% of diurnal carbon gain (Zotz and Winter 1993). The estimated Aday does not include changes in re spiration rates or stomatal closure. For a subset of leaves that lacked light curve measurements, species averages of curvature and apparent quantum yield and leaf-specific measurements of Amax and dark respiration were fitted to non-rectangular hyperbola using Photosyn Assistant. This provided a best-fit estimate of Amax for each leaf. Due to the low [CO2] differential, large measurements errors were associated with i ndividual dark respira tion rates. Thus, dark respiration rates were calculated from the regression between leaf age and dark respiration in all species except Genipa The species-specific average of dark respiration was used for Genipa because it showed no relationship between leaf age and dark respiration. Statistical Analyses Statistical analyses were carried out us ing JMP V.5.1 (SAS, 2003). Significance of the relationship of average daily carbon gain with leaf age, Amax, and leaf PFD was tested

PAGE 48

38 using a linear regression with leaves, plan ts, and plots pooled for each species. A multiple regression was used to determine whether Amax or leaf PFD better predicted daily carbon gain with leaves, plants, and pl ots pooled for each species. A simple linear regression was used to determine the relationship between daily carbon gain, Amax, and leaf PFD with all measurements and species pooled. Results Plants experienced light-limiting conditions primarily due to self-shading and weather, although shading by surrounding plan ts might have contributed to lightlimitation to an unknown extent (C. Stefanescu, personal ob servation). Depending on their age, leaves spent 12.9 to 51.6% of the total time they were exposed to day light below their LCP (Figure 2-3). Average daily carbon gain of individual leaves varied within and among species. The lowest value observed was -1.8 mmol CO2 day-1 for a very old leaf of Tetragastris whereas the highest was 216.4 mmol CO2 day-1 for a young leaf of Genipa The relationship between average daily carbon gain and leaf age was significant in Tabernaemontana ( r2 = 0.5, P = 0.05), Virola ( r2 = 0.5, P = 0.009), Tetragastris ( r2 = 0.5, P = 0.003), and Aspidosperma ( r2 = 0.8, P = 0.002), but not in Genipa ( r2 = 0.2, P = 0.2; Figure 2-4). Average daily carbon gain was also positively correlated with Amax in Tabernaemontana ( r2 = 0.5, P = 0.05), Virola ( r2 = 0.7, P = 0.0006) and Tetragastris ( r2 = 0.6, P = 0.001), but not in Genipa ( r2 = 0.1, P = 0.4) or Aspidosperma ( r2 = 0.1, P = 0.5; Figure 2-5). Average daily carbon gain was pos itively correlated with leaf PFD in all species; Genipa ( r2 = 0.8, P = 0.0005), Tabernaemontana ( r2 = 0.9, P = 0.0004) Virola ( r2 = 0.9, P < 0.0001) Tetragastris ( r2 = 0.9, P < 0.0001), and Aspidosperma ( r2 = 0.9, P = 0.002; Figure 2-6). Amax was a poor predictor of daily car bon gain in all species except

PAGE 49

39 for Virola (P < 0.005; Table 2-3). Leaf PFD was a better predictor of daily carbon gain that Amax in all species (Table 2-3). When all leaves of all species were analyzed together, both Amax and leaf PFD were highly correlated with daily carbon gain (Figure 27). Discussion Due to the documented affect of leaf age on both photosynthetic capacity and leaf PFD for these species (Chapter 1), it is not surprising that leaf age was correlated with average daily carbon gain in four out of the five species in this study. In Aspidosperma the lack of relationship betw een leaf age and daily carbon gain may reflect the limited range of leaf age sampled relativ e to the actual leaf lifespan of this species; the median leaf lifespan for Aspidosperma is 797 days, but only leaves between 37 and 365 days old were measured. Leaf PFD was strongly correl ated with daily carbon ga in in all species. In contrast, Amax was correlated with daily carbon gain in only three out of five species. Furthermore, when evaluated in combination with leaf PFD, Amax was only correlated with daily carbon gain in Virola When results were pooled for the five species in this study, there was a strong relationship between Amax and daily carbon gain (r2 = 0.48, P < 0.0001, Figure 2-7) and leaf PFD and daily carbon gain (r2 = 0.89, P < 0.0001, Figure 2-7). A previous study by Zotz and Winter (1993) documented a higher co rrelation coefficient for the relationship between Amax and Aday for 64 diel courses of net CO2 exchange for eight tropical species (r2 = 0.92, P < 0.0001). Zotz and Winter (1993) did not encounter a strong relationship between Aday and leaf PFD (r2 = 0.13, P < 0.01). Indeed, there was no relationship between leaf PFD and Aday for a Ceiba pentandra a tropical canopy tree (r2 = 0.48, P < 0.1; (Zotz and Winter 1994). I believe the co ntrasting results betw een these two studies

PAGE 50

40 can be explained by variation in leaf light levels. Even on clouded days, leaf PFD was relatively high for leaves of th e eight species measured by Zo tz and Winter (1993), which resulted in light-saturated rates of phot osynthesis (Zotz and Winter 1993). The same cannot be said for leaves of the five species in this study. In a ll species, all leaves experienced light levels below their light compensation point at some time during the three days during which light was measured. Even young leaves at the top of the plant experienced light-limiting conditions for 1219% of day light due to plant positioning inside the gap and possible shading by surr ounding plants. Thus, it appears that when leaves are exposed to light conditions that allow for performance at maximum photosynthetic capacity, daily carbon gain is highly correlated with maximum photosynthetic capacity. However, when leav es are not constantly exposed to lightsaturating conditions, leaf light levels are mo re highly correlated with daily carbon gain. This study documented a negative carbon gain for two out of the fourteen measured leaves of Tetragastris Both of these leaves were ol d (512 and 596 days) at the time of measurement. At first, this result was surp rising as all plants we re grown in canopy gaps in which the light environment above th e seedlings ranged from 23-52% of full sun (Kitajima unpublished data). However, the light conditions within each gap were variable due to plant positioning inside the gap, and the height of surrounding plants. For instance, above plant light measurements for Tetragastris were 8.4%, while leaves of Aspidosperma and Genipa experienced much higher light levels with some leaves exposed to as much as 30-38% of full sun. When taking into account that the above plant light environment for Tetragastris was only 8.4% of full sun, negative carbon balances for older, shaded leaves at the bottom of the plant are reasonable. Furthermore, the daily

PAGE 51

41 carbon gains presented are means of three days of measurements for which the standard deviation of carbon gain varied from 2-3 mmol CO2 s-2 day-1 depending on the leaf. Even leaves with average negative daily carbon gain generally experienced some days with a low positive carbon gain. This study used photosynthetic light-respons e curves to calculate carbon gain from daily courses of photon flux density (PFD). While this approach has been commonly used in the past (Chazdon 1986, Williams et al. 1989, Poorter et al. 2006), the equation used to calculate car bon gain (P(I) = { I + Amax – {( I + Amax)2 – 4 AmaxI }0.5 / 2 } – R, Johnson and Thornely 1984) does not take induction time into ac count (Kikuzawa et al. 2004). Carbon gain can also be limite d by the mid-day depression of photosynthetic rates caused by stomatal closure in response to short-term drought during the hottest and most humid times of day (Zotz and Winter 1993) Thus, it is likely that the carbon gain calculated for leaves in this experiment ar e slightly overestimated. The most direct method of calculating carbon ga in is by measuring daily co urses of photosynthetic rates (Zotz and Winter 1993, 1994), but this a pproach was not used in this study. Leaf PFD was the best predictor of carbon gain in this study regardless of whether analyses were done separately for each speci es or together by pooling samples across all species. It is well known that the photosynthetic capacity of l eaves is correlated with the light conditions under which they develop, resulting in a rela tionship between Amax and leaf PFD (Field 1988). Particularly in sh aded conditions, leaf PFD is more highly correlated with leaf carbon gain than photosynthetic cap acity (Chazdon and Pearcy 1986). In this experiment, leaf light range d from 0.3 to 44% of daily PFD above the canopy. While this range is much greater than the 1-2% of light t ypically experienced by

PAGE 52

42 understory plants, the average daily leaf PFD across leaves and species was only 10% of full sun, suggesting that leaves were not ab le to perform at maximum photosynthetic capacity throughout the day. These results indica te that in conditions where light is not saturating, leaf PFD is a better predictor of da ily carbon gain than Amax.

PAGE 53

43 Table 2-1. Median and mean leaf lifespan of marked leaves using the Kaplan-Meyer method for each species studied. Species are ranked from shortest median leaf lifespan to longest median leaf lifespan. Nomenclature follows Croat (1978) Species Family Total Leaves Median Mean Std Dev Genipa americana Rubiaceae 478 224 234.8 4.96 Tabernaemontana arborea Apocynaceae 542 279 277.6 3.94 Virola surinamensis Myristicaceae 709 504 494.3 6.56 Tetragastris panamensis Burseraceae 89 671 648.3 13.87 Aspidosperma cruenta Apocynaceae 31 797 745.3 12.51 Table 2-2. Number of seedlings and leaves for which % PFD and Aday were measured in each of five species of tropical tree sp ecies. Species are ranked from shortest median leaf lifespan to l ongest median leaf lifespan. # of Leaves Species # of Plants %PFD Aday Genipa 5 9 9 Tabernaemontana 2 8 8 Virola 4 12 12 Tetragasteris 4 14 14 Aspidosperma 2 7 7 Table 2-3. Multiple regression analysis of seed lings of five tropical tree species with Amax and leaf PFD as predictor variables, and daily car bon gain as the response variable. The coefficient of multiple determination ( r2 ) represents the proportion of the variance explained by the two vari ables together, while Pvalues represent the significance of e ach variable as it contributed to the model. Amax Leaf PFD Species r2 P P Genipa 0.90 0.12 0.0005 Tabernaemontana 0.91 0.43 0.006 Virola 0.96 0.006 < 0.0001 Tetragastris 0.87 0.17 0.0005 Aspidosperma 0.88 0.52 0.007

PAGE 54

44 Table 2-4. Light curve parame ters for five species of tr opical tree seedlings used to calculate daily carbon gain. Amax (mol m-2 s-1) R (mol m-2 s-1) Species 30 day 200 day 30 day 200 day Quantum Yield ( ) Curvature ( ) Genipa 13.5 6.61 -0.494 -0.494 0.058 0.697 Tabernaemontana 9.36 8.99 -0.486 -0.333 0.049 0.674 Virola 6.05 4.91 -0.349 -0.298 0.047 0.701 Tetragastris 8.61 6.58 -0.416 -0.331 0.049 0.673 Aspidosperma 8.22 10.58 -0.447 -0.396 0.056 0.734

PAGE 55

45 Figure 2-1. A schematic representa tion of the expected relations hip between leaf age, leaf PFD, Amax, and Aday. The relationships between leaf age and leaf PFD and Amax were explored in Chapter 1. PFD mmol m-2 day-1 020004000600080001000012000 % PFD 0 10 20 30 40 r2 = 0.85, P < 0.0001 y = 0.003x + 0.82 Figure 2-2. The relationship be tween % PFD and mmol PFD m-2 day-1 for all leaves and species pooled. Each point is a three day average. Leaf Age Leaf PFD Amax Aday

PAGE 56

46 Leaf Age (days) <100<200<300<400<500<600<700 Percent of Time Above or Below LCP 0 20 40 60 80 100 Below LCP Above LCP Figure 2-3. The percent of time leaves of different ages spent above and below their light compensation point (LCP) relative to th e total day length. Depending on their age, leaves spent 12.9 to 51.6% of the to tal time they were exposed to day light below their LCP.

PAGE 57

47 0200400600800 0 50 100 150 200 250 Tabernaemontana y = -0.206x + 69.01 r2 = 0.50, P = 0.0501 0 50 100 150 200 250 Genipa r2 = 0.20, P = 0.22 0 50 100 150 200 250 Virola y = -0.141x + 82.98 r2 = 0.51, P = 0.0091 0 50 100 150 200 250 Tetragasteris y = -0.070x + 41.60 r2 = 0.54, P = 0.0028 0200400600800 0 50 100 150 200 250 Aspidosperma y = -0.267x + 157.70 r2 = 0.85, P = 0.0029Leaf Age (Days)Daily Carbon Gain (mmol CO2 m-2 day-1)Leaf Age (Days) Figure 2-4. The relationship between leaf age an d three-day averages of daily carbon gain in seedlings of five sp ecies of tropical trees.

PAGE 58

48 0 5 10 15 2 0 0 50 100 150 200 250 Tabernaemontana y = 8.63x 39.29 r2 = 0.51, P = 0.0461 0 50 100 150 200 250 Genipa r2 = 0.1, P = 0.4 0 50 100 150 200 250 Virola y = 9.77x 7.57 r2 = 0.71, P = 0.0006 0 50 100 150 200 250 Tetragasteris y = 5.10x 4.47 r2 = 0.60, P = 0.0011 0 5 1 0 15 20 0 50 100 150 200 250 Aspidosperma r2 = 0.1, P = 0.5Amax ( mol m-2 s-1)Daily Carbon Gain (mmol CO2 m-2 day-1)Amax ( mol m-2 s-1) Figure 2-5. The relationship between maximum photosynthetic capacity (Amax) and three day averages of daily carbon gain for seedlings of five species of tropical trees.

PAGE 59

49 0 20 00 4000 60 00 8000 1000 0 12 00 0 0 50 100 150 200 250 Tabernaemontana y = 0.019x + 2.58 r2 = 0.89, P = 0.0004 0 50 100 150 200 250 Genipa y = 0.014x + 23.81 r2 = 0.84, P = 0.0005 0 50 100 150 200 250 Virola y = 0.026x 7.49 r2 = 0.91, P < 0.0001 0 50 100 150 200 250 Tetragasteris y = 0.015x + 0.62 r2 = 0.85, P < 0.0001 0 2000 4 000 6000 8000 1 000 0 12000 0 50 100 150 200 250 Aspidosperma y = 0.013x + 37.79 r2 = 0.87, P = 0.0022Leaf PFD (mmol m-2 day-1)Daily Carbon Gain (mmol CO2 m-2 day-1)Leaf PFD (mmol m-2 day-1) Figure 2-6. The relationship between leaf age an d three day averages of light incident on those leaves for seedlings of five species of tropical trees. Leaf PFD increased significantly with carbon gain in all species.

PAGE 60

50 0 2 4 6 8 1 0 12 14 0 50 100 150 200 250 0 20 00 4 0 00 60 00 80 00 10000 1 200 0 0 50 100 150 200 250 Daily Carbon Gain (mmol CO2 m-2 day-1)Amax ( mol m-2 s-1) Leaf PFD (mmol m-2 day-1)y = 9.14x 11.17 r2 = 0.48, P < 0.0001 y = 0.02x + 6.17 r2 = 0.89, P < 0.0001 Figure 2-7. The relationship betw een daily carbon gain and Amax or leaf PFD when all measurements and species were pooled.

PAGE 61

51 LIST OF REFERENCES Ackerly, D. 1999. Self-shading, carbon gain a nd leaf dynamics: a te st of alternative optimality models. Oecologia 119:300-310. Ackerly, D. D., and F. A. Bazzaz. 1995. Leaf dynamics, self-shading and carbon gain in seedlings of a tropical pion eer tree. Oecologia 101:289-298. Bazzaz, F. A. 1979. Physiological ecology of plant succession. Annual Review of Ecology and Systematics 10:351-371. Bongers, F., and J. Pompa. 1990. Leaf dynamics of seedlings of rain -forest species in relation to canopy gaps. Oecologia 82:122-127. Chabot, B. F., and D. J. Hicks. 1982. The ecolo gy of leaf life span s. Annual Review of Ecology and Systematics 13:229-259. Chazdon, R. L. 1986. Light variation and carbon gain in rain-forest understorey palms. Journal of Ecology 74:995-1012. Chazdon, R. L., and R. W. Pearcy. 1986. Car bon gain and photosynthetic efficiency during lightflecks. Oecologia 69:524-531. Coley, P. D. 1988. Effects of plant-growth ra te and leaf lifetime on the amount and type of anti-herbivore defens e. Oecologia 74:531-536. Croat, T. B. 1978. Flora of Barro Colorado Isla nd. Stanford University Press, Stanford. Daws, M. I., C. E. Mullins, D. Burslem, S. R. Paton, and J. W. Dalling. 2002. Topographic position affects the water regime in a semi-deciduous tropical forest in Panama. Plant and Soil 238:79-90. Denslow, J. S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12:4755. Donovan, T. M., and C. W. Weldin. 2002. Spreadsheet Exercises in Ecology and Evolution. Sinauer Associates, Sunderland, MA. Ellsworth, D. S., and P. B. Reich. 1996. P hotosynthesis and leaf nitrogen in five Amazonian tree species during early secondary succession. Ecology 77:581-594.

PAGE 62

52 Escudero, A., and S. Mediavilla. 2003. Declin e in photosynthetic nitrogen use efficiency with leaf age and nitrogen resorption as determinants of leaf life span. Journal of Ecology 91:880-889. Field, C. 1983. Allocating leaf nitrogen for the maximization of carbon gain-leaf age as a control on the allo cation program. Oecologia 56:341-347. Field, C., and H. A. Mooney. 1983. Leaf age and seasonal effects on light, water, and nitrogen use efficiency in a Ca lifornia shrub. Oecologia 56:348-355. Hikosaka, K. 2004. Interspecific difference in the photosynthesis-ni trogen relationship: patterns, physiological causes, and ec ological importance. Journal of Plant Research 117:481-494. Hikosaka, K., I. Terashima, and S. Katoh. 1994. Effects of leaf age, nitrogen nutrition and photon flux-density on the distribution of nitrogen among leaves of a vine ( Ipomoea tricolor ) grown horizontally to avoid mutual shading of leaves. Oecologia 97:451-457. Hirose, T., and M. J. A. Werger. 1987. Maximizing daily canopy photosynthesis with respect to the leaf nitrog en allocation pattern in the canopy. Oecologia 72:520-526. Johnson, I. R., and J. H. M. Thornley. 1984. A model of instantaneous and daily canopy photosynthesis. Journal of Th eoretical Biology 107:531-545. Kikuzawa, K. 1991. A cost-benefitanalysis of leaf habit and leaf longevity of trees and their geographical pattern. Amer ican Naturalist 138:1250-1263. Kikuzawa, K., and D. Ackerly. 1999. Significan ce of leaf longevity in plants. Plant Species Biology 14:39-45. Kikuzawa, K., H. Shirakawa, M. Suzuki, and K. Umeki. 2004. Mean labor time of a leaf. Ecological Research 19:365-374. King, D. A. 2003. Allocation of above-ground grow th is related to light in temperate deciduous saplings. Functional Ecology 17:482-488. Kitajima, K., K. Kikuzawa, D. Ackerly, S. S. Mulkey, and S. J. Wright. 2003. Leaf-age effects on photosynthetic rates as a key para meter in the theory of leaf longevity. Unpublished manuscript. Kitajima, K., S. S. Mulkey, M. Samani ego, and S. J. Wright. 2002. Decline of photosynthetic capacity with leaf age and position in two tropical pioneer tree species. American Journal of Botany 89:1925-1932. Kitajima, K., S. S. Mulkey, and S. J. Wr ight. 1997. Decline of photosynthetic capacity with leaf age in re lation to leaf longevities for five tropical canopy tree species. American Journal of Botany 84:702-708.

PAGE 63

53 Koike, T. 1988. Leaf structure and photosynthe tic performance as related to the forest succession of deciduous broad-leaved trees. Plant Species Biology 3:77-87. Leigh, E. G., Jr., Windsor, D.M., Rand, S.A. 1982. The Ecology of a Tropical Forest: Seasonal Rhythms and Long-Term Changes, Second Edition. Smithsonian Institution Press, Washington, D.C. Mediavilla, S., and A. Escude ro. 2003. Leaf life span diffe rs from retention time of biomass and nutrients in the crowns of evergreen species. Functional Ecology 17:541-548. Mooney, H. A., C. Field, S. L. Gulmon, and F. A. Bazzaz. 1981. Photosynthetic capacity in relation to leaf position in desert ve rsus old-field annuals Oecologia 50:109-112. Peterson, A. G. 1999. Reconciling the apparent difference between massand area-based expressions of the photosynthesis-nitr ogen relationship. Oecologia 118:144-150. Poorter, H., S. Pepin, T. Rijkers, Y. de Jong, J. R. Evans, and C. Korner. 2006. Construction costs, chemical compositi on and payback time of highand lowirradiance leaves. Journal of Experimental Botany 57:355-371. Reich, P. B., D. S. Ellsworth, M. B. Walters, J. M. Vose, C. Gresham, J. C. Volin, and W. D. Bowman. 1999. Generality of leaf trait relationships: a test across six biomes. Ecology 80:1955-1969. Reich, P. B., C. Uhl, M. B. Walters, and D. S. Ellsworth. 1991. Leaf life-span as a determinant of leaf structure and f unction among 23 Amazonian tree species. Oecologia 86:16-24. Reich, P. B., C. Uhl, M. B. Walters, L. Prugh, and D. S. Ellsworth. 2004. Leaf demography and phenology in Amazonian rain forest: A census of 40 000 leaves of 23 tree species. Ecological Monographs 74:3-23. Reich, P. B., M. B. Walters, and D. S. Ellswo rth. 1992. Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosystems. Ecological Monographs 62:365-392. Reich, P. B., M. B. Walters, and D. S. Ells worth. 1997. From tropics to tundra: global convergence in plant functioning. Proceedings of the National Academy of Sciences of the United Stat es of America 94:13730-13734. Takashima, T., K. Hikosaka, and T. Hirose. 2004. Photosynthesis or persistence: nitrogen allocation in leaves of evergreen and deciduous Quercus species. Plant Cell and Environment 27:1047-1054. Williams, K., C. B. Field, and H. A. Mooney. 1989. Relationships among leaf construction cost, leaf longevity, and light environment in rain-f orest plants of the Genus Piper American Naturalist 133:198-211.

PAGE 64

54 Wright, I. J., P. B. Reich, M. Westoby, D. D. Ackerly, Z. Baruch, F. Bongers, J. Cavender-Bares, T. Chapin, J. H. C. Cornelissen, M. Diemer, J. Flexas, E. Garnier, P. K. Groom, J. Gulias, K. Hikosaka, B. B. Lamont, T. Lee, W. Lee, C. Lusk, J. J. Midgley, M. L. Navas, U. Niinemets, J. Ol eksyn, N. Osada, H. Poorter, P. Poot, L. Prior, V. I. Pyankov, C. Roumet, S. C. Thomas, M. G. Tjoelker, E. J. Veneklaas, and R. Villar. 2004. The worldwide leaf economics spectrum. Nature 428:821-827. Zotz, G., and K. Winter. 1993. Short-term photosynthesis measurements predict leaf carbon balance in tropical rain-fores t canopy plants. Planta 191:409-412. Zotz, G., and K. Winter. 1994. Photos ynthesis of a tropical canopy tree, Ceiba pentandra in a lowland forest in Panama. Tree Physiology 14:1291-1301.

PAGE 65

55 BIOGRAPHICAL SKETCH Carla C. Stefanescu was born in Madison, WI, in May 1980. She attended Central High School in Tuscaloosa, AL, where she beca me fascinated by biology. Carla earned her Bachelor of Science degr ee in biology with a concentra tion in environmental studies at the University of the South in May 2002. She became interested in tropical ecology during the pursuit of her bachelorÂ’s degree when she studied with the Organization of Tropical Studies in Costa Rica for the Spring semester of 2001. During her undergraduate studies, she also worked as a research assistant in Sewanee, TN, and Tuscaloosa, AL. In August 2006, she obtained he r Master of Science degree in botany at the University of Florida.