in.” a . . W} 2.
. u . .

.

2 3mm,
11. mm, .

1...
;.
.1.
i\

31...

 

.

41.2 .12. .2“... .. . . .. ..
111...; 2mm... “.23. .. ., a 2.9.2:
..12. $2... . . .221.
R9522 . . .
2.411%,
gm"
.5

.Bn: :

. .. 1..
. r2 .1 .

. ml

. . . . w...
. Jr; . 1.1 . . .

1. by... . . . . . $3.2 .
.2 . . . .
.. .

flux. ,
”had"

$12

:2 P

1. LL.
. l. 8.4 .l
3 I.

2 m1“...
. . :3... z .
9a .15x53mem1bh S

 

523‘ .§ Z.

:2139

. 2.V.n\.
.. 22.1,;

 

 

...Ԥr....n :....1

... 1.2.12.3

THESIS
Z.
Z 060

This is to certify that the
dissertation entitled

THE EVOLUTIONARY ECOLOGY OF SENESCENCE
IN DAPHNIA

presented by

JEFFRY L. DUDYCHA

has been accepted towards fulfillment
of the requirements for

Ph . D degree in ZOOLOGY

W

flajor professor

DateVX yécfh /?%

MSU is an Affirmative Action/Equal Opportunity Institution 0212771

 

 

LEBRARY
Michigan State
University

 

 

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m

c!ClRC/Dato0uo.p65-p.14

 

 

THE EVOLUTIONARY ECOLOGY OF SENESCENCE IN DAPHNIA

by
Jeffry L. Dudycha

A DISSERTATION

Submitted to
Michifgan State University
in partial ful llment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Program in Ewcolog}; ,Evolutionary Biolo and Behavior,
ellogg Biological Station

Department of Zoology
1999

ABSTRACT
THE EVOLUTIONARY ECOLOGY OF SENESCENCE IN DAPHNIA

by
Jeffry L. Dudycha

The evolutionary theory of senescence predicts that high extrinsic
mortality in natural populations should select for accelerated reproductive
investment and shortened lifespan. I examined the theory with natural
populations of the Daphnia pulex-pulicaria species complex, a group of common
freshwater crustaceans that spans an environmental gradient of habitat
permanence. I document substantial genetic variation in life history traits among
populations of this complex. Populations from temporary ponds have shorter
lifespans, earlier and faster increases of intrinsic mortality risk and earlier and
steeper declines in fecundity than populations from permanent lakes. I also
examine the age-specific contribution to fitness, which declines faster in
populations from ponds than those from lakes. Pond Daphnia also exhibit faster
juvenile growth and higher early fitness, measured as population growth rate (r).
I observed negative genetic correlations between r and indices of life history
timing, further suggesting tradeoffs between early- and late-life performance.

Understanding the evolution of senescence in nature requires knowledge
about genetic variation and ecological relevance of senescence plasticity because
senescence is responds to factors such as food and temperature. Therefore, I also
examined plasticity of senescence and variation of that plasticity within and
between two species complexes of Daphnia. I quantified senescence in four
environments (crossed high and low temperatures and food levels), to evaluate
the plasticity of mortality and fecundity with respect to the ecology of the

species’ habitats. Senescence was highly plastic, but species were plastic to

different degrees. Population growth rate (r) was most responsive to food, but
senescence was most responsive to temperature. Temperature strongly
influenced the degree to which genetic differentiation within complexes was
expressed, with little differentiation under temperature stress. In the D. pulex-
pulicaria complex, whose species use markedly different habitats, genetic
variation of senescence was strong and robust across environments. Species of
the D. mendotae-dentifera complex, which inhabit stratified lakes, had similar
senescence patterns. Most genetic variation of senescence occurred within,
rather than between, species complexes, indicating that divergence of senescence
in Daphnia is recent.

1 further tested for a relationship between extrinsic mortality risk and
ecological distribution of senescence variation by measuring natural
demographic rates (growth, birth and death) in twelve populations of D. pulex-
pulicaria. This work showed a negative relationship between ecological
mortality risk and investment in late-life fitness. By constructing a phylogeny of
these populations based on mtDNA, I was able to show that variation of
senescence is not easily attributed to simple genetic history. My results cannot be
explained by a tradeoff between survival and fecundity, nor by non-evolutionary
theories of senescence. Instead, the data support the evolutionary theory of
senescence because the genetic variation in life histories we observed is

congruent variation of death rates experienced in the wild.

Co ' tb
Jef erghDugycha
1999

to a fry-cook on Venus and
a young man who didn’t think he’d
seen anything good that day

ACKNOWLEDGEMENTS

I deeply appreciate the guidance and Alan Tessier has offered over the
years, along with a great variety of intellectual interations. My research would
have been much more incomplete and tentative had I not had the expert and
dedicated assistance of Pam Woodruff in a wide variety of capacities, ranging
from technical lab work and field work to watching after my dogs when needed.

I also appreciate that my breadth of conceptual understanding in ecology
and evolution was enhanced by interactions with people at KBS and in the BBB
program, in particular Tom Getty, Rich Lenski, Don Straney, Kevin Geedey, Liz
Smiley, Danny Rozen, Vaughn Cooper, Brian Black, Jeff Conner, Kris Desmarais,
Anita Davelos, Erin Boydston, Becky Fuller, Jeff Birdsley and Jim Smith. My
interest in ecology and evolution solidified in classes taught by Carol Folt, Dave
Peart, Dick Holmes, John Gilbert and Mark McPeek. I am in their debt for
showing me what it is to be an academic ecologist. In addition, Mark played a
pivotal role in my becoming a grad student at KBS.

Several others provided assitance at critical times or with important tasks.
Bryan Foster first showed me where Otislake HI was, when I had all but given up
on ever finding D. pulex. My father, Art Dudycha, was fun to have along when I
was doing fieldwork and isolating clones. Dieter Ebert gave me some useful
photographs and tips on finding parasites in cultures of Daphnia.

John Gorentz and Nina Consolatti went well beyond the call of duty in
simplifying the computational and logistic details of completing this work. Alice
Gillespie, Sally Shaw, Lisa Craft and Char Adams made a great deal of
administrative hassles vanish for me.

I had a great time testing some of my ideas out with people who visited
KBS, especially Marc Tatar, David Reznick, Bob Ricklefs, Mike Lynch, Steve

vi

Steams and Bob Holt. Critical comments on earlier drafts of some chapters were
kindly provided by Marc Tatar, Mark McPeek, Carla Caceres, Jeff Conner, Brian
Black, Danny Rozen, and Alan Tessier (several times over).

My ability to accomplish any molecular work was based on the patient
training by Mike Pfrender, with additional advice and support from Mike Lynch,
Niles Lehman, John Colbourne, Sue Stoltzfus and Les Dethlefsen. Niles and John
graciously granted access to and use of unpublished data.

The research I conducted was primarily supported by the U. S. National
Science Foundation and the Kellogg Biological Station. Additional funding came
from the program in Ecology and Evolutionary Biology, the Department of
Zoology, the College of Natural Science and the Graduate School at Michigan
State and the George H. Lauff Research Fund at the Kellogg Biological Station.

vii

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES .

CHAPTER ONE
INTRODUCTION: MICROCRUSTACEA AS A MODEL OF AGING

CHAPTER TWO
NATURAL GENETIC VARIATION OF LIFESPAN, REPRODUCTION
AND JUVENILE GROWTH IN DAPHNIA
Introduction. .
Objectives of this study .
Daphnia and the pond- -lake gradient
Methods .
Clonal isolation and lineage establishment
Life history trait measurement .
Quantifying as ects of senescence
Statistical Ana yses . .
Results
Mortality and fecundity
Combined fitness values
Juvenile size and growth
Discussion . . .
Tradeoffs
Literature Cited

CHAPTER THREE
A REACTION NORM APPROACH TO THE EVOLUTIONARY
ECOLOGY OF SENESCENCE .
Introduction.
Study system
Methods .
Life tables . .
Estimation of age-specific fitness .
Statistical analyses .
Results .
Discussion
Senescence m D. pulex-pulicaria
Senescence m D. mendotae-dentifera
Environmental effects on senescence
Literature Cited . .

CHAPTER FOUR
MORTALITY RISK AND THE ECOLOGICAL DISTRIBUTION OF
SENESCENCE VARIATION . .
Introduction.
Objective of this study
The model system
Genetic variation of senescence

viii

xi

106
106
110

112

Trait measurement .
Data analysis. .
Life history results .
Field Demography.
Methods
1 Reisuultia 'd h 1111111
Re ations etween emo a an 1 e sto
Local Ph lopgeny gr p y .ry
A sequencing
Gene tree construction
Discussion . . .
Literature Cited

LIST OF TABLES

Table 2-1. Values of fitness measures and the timing of fitness
contribution in D. pulex and D. pulicaria.

Table 3-1. Differences in mortality rate acceleration between .
nominal species of Daphnia.

Table 3-2. Repeated measures ANOVA of intrinsic value and
reproductive value in Daphnia.

Table 3-3. Tests of genetic variation in the plasticity of the age
when only 25% of intrinsic value remains.

52

96

97

98

Figure 2-1.
Figure 2-2.
Figure 2-3.
Figure 2-4.
Figure 2-5.
Figure 2-6.

Figure 2—7.

Figure 3-1.
Figure 3-2.
Figure 3-3.
Figure 3-4.
Figure 3-5.
Figure 3-6.
Figure 3-7.

LIST OF FIGURES

Survivorship of Daphnia pulex-pulicaria.

53

Mortality hazard function od D. pulex, D. pulicaria and hybrids 54

Per-capita fecundity of D. pulex, D. pulicaria and hybrids.
Reproductive value of Daphnia pulex-pulicaria
Intrinsic value of Daphnia pulex—pulicaria

Relationship between early fitness and life history timing
in Daphnia pulex-pulicaria.

Neonate and adult mass (in 11g) of clones 1n the Daphnia
pulex-pulicaria complex. . .

Survivorship of D. pulex-pulicaria in four environments. .
Intrinsic value of D. pulex-pulicaria. .

Reproductive value of D. pulex-pulicaria.

Survivorship of D. mendotae-dentifera in four environments.
Intrinsic value of D. mendotae-dentifera.

Reproductive value of D. mendotae—dentifera.

Estimated opulation growth rate and persistence of

fitness into late life 1n our laboratory environments.

Figure 4-1.
Figure 4-2.
Figure 4-3.
Figure 4-4.
Figure 4-5.
Figure 4-6.
Figure 4-7.
Figure 4-8.
Figure 4-9.

Survivorship of 12 populations of D. pulex—pulicaria.
Fecundity of 12 populations of D. pulex—pulicaria. .
Intrinsic value of 12 populations of D. pulex-pulicaria.

Population density of D. pulex-pulicaria over one year.

Comparison of growth rates in D. pulex—pulicaria over 1 year.
Instantaneous birth rates in D. pulex-pulicaria over one year.

Instantaneous death rates in D. pulex-pulicaria over one year.

55
56
57

58

59

99

100
101
102
103
104

105
137
138
139
140
141
142
143

Relationship between ecological mortality risk and senescence. 144

Gene tree of 25 clones of D. pulex-pulicaria.

xi

145

CHAPTER ONE: INTRODUCTION
MICROCRUSTACEA AS A MODEL OF AGING

The science of geriatric medicine is centrally concerned with the process of
human aging, with the goal of alleviating its detrimental effects. Crustaceans
may not seem to be an obvious choice as a medical model, but I will argue that
there are important fundamental questions for which they are ideal. In this
chapter, I discuss the evolutionary biology of aging and how certain crustaceans
could be used to integrate it with our understanding of the physiological effects
of aging. This is necessarily a ”looking forward” chapter, because currently little
work with crustaceans is related to the biology of aging, let alone geriatric
medicine.

Any sensible approach to science begins with clearly identifying questions
we wish to answer. Considering what we wish to know about aging and why
we wish to know this is the first stage of choosing a model organism. How does
the desired information mesh with the already substantial body of knowledge
which we have accumulated? These considerations lead to the potential for
crustaceans, particularly aquatic microcrustaceans, to be added to the relatively
narrow group of model organisms used to study the process of aging.

The first part of this chapter is devoted to defining some questions about
the evolutionary biology of aging which are either wholly unanswered or which
could profit from taxonomic expansion beyond current model systems. A good
general overview of the evolution of aging can be found in Rose (1991). The
second part addresses the relative merits and demerits of a few microcrustaceans
in the context of these questions. I pay particular attention to introducing aspects
of the basic biology of these organisms with which the reader may not be

familiar and the relevant aspects of basic evolutionary biology. The third part

briefly reviews the scant direct knowledge of microcrustaceans and aging. I
close with a discussion of fundamental tools which should be developed for

microcrustacea and the research potential for microcrustaceans in the field of

aging.

What is the relationship between nature and the evolution of aging?

This is a broad, overarching question that is largely unaddressed by
empirical information. Substantial theory exists to explain the evolution of aging
(Williams 1957, Hamilton 1966, Edney and Gill 1966, Charlesworth 1993), and
many key predictions and assumptions have withstood laboratory tests.
However, the careful control of the laboratory is a far cry from the complexities
of nature. Examining the evolution of aging in nature is important because it
will reveal whether or not there are critical details which have not been
considered in laboratory and theoretical work. Finding a discrepancy between
nature and current knowledge makes possible a search for the cause of that
discrepancy. Finding no discrepancy reassures us that we are on the right track,
a particularly important reassurance for geriatric scientists since their ultimate
subject is a naturally evolving species.

A broad component of the relationship between nature and the evolution
of aging is how aging evolves in nature, which itself has several subsidiary
questions. How much genetic variation is generally present for patterns of
aging? How closely is the natural selection on patterns of aging coupled to the
phenotypic expression of aging? What are the dominant forces (e. g., predation,
disease, food availability, habitat permanence, mate choice, parent-offspring
conflict) exerting selection on aging? How do these forces operate? What are the

relative impacts of fecundity selection and mortality selection? What are the

trade-offs associated with patterns of aging, and how does selection act on those
trade-offs (thus indirectly influencing aging)? What constraints are there on the
evolution of aging patterns? How does the aging pattern evolve if there is
temporal or spatial variation in the strength, direction, or type of selection?

It may be difficult to see the medical applications of answers to these
questions, and indeed they may be subtle and indirect. However, this is not to
say they are unreal or unimportant. In a human context, consider attempts
which may be made to mitigate the detrimental effects of aging. Understanding
the trade-offs which selection has created may suggest focused research on
potential unwanted side-effects, particularly those which appear only in
subsequent generations, either via genetic or maternal effect. Or, it could be that
humans have substantially altered the influence of selection on their own
patterns of aging, in which case it may be difficult to modify a phenomenon
which is in a state of evolutionary flux. Exploring how natural selection acts on
the integrated whole of an organism to shape the evolution of aging will afford
us a greater understanding of how the different aspects of bodily deterioration
are interrelated; a better understanding of the potential repercussions on
intervening in aging will be the result.

Natural selection is not the only component of the relationship between
nature and the evolution of aging: aging could also affect an organism’s
ecological interactions. Ecologists have largely (though not entirely, and it is
changing) assumed aging to be a minor force in ecological interactions and have
rarely looked for it, and have even more rarely explored its impacts. This is not
wholly unreasonable, given the extraordinary difficulty of observing aging in the
wild, and the observation that most populations are largely unsheltered from

extrinsic mortality — hence, few wild animals die of old age. Some aspects of

ecological interactions, however, such as mating behavior and thermotolerance,
have received some attention in a laboratory context.

Historical attitudes notwithstanding, it is still true that ecological aspects
of aging are worthy of study, and that their study in conjunction with
evolutionary biology is especially apropos. Good evolutionary biology is
intertwined with ecology, just as good ecology is intertwined with evolutionary
biology. The ecological context of aging is critical, though largely unknown from
an empirical perspective, because it is primarily responsible for the natural
selection exerted on aging. It is also important because humans are involved in
ecological interactions, even though this concept is rarely emphasized.
Understanding our own aging, therefore, requires an understanding of the
ecological impacts and the ecological context of aging. For example, as
organisms age, they may become more susceptible to disease, in turn having an
impact on their ability to provide for offspring. This is an extension of the
expression of aging beyond that which occurs in the body. The fact that we
humans are part of an ecological system is probably relevant to many other areas
of human biology. It is particularly relevant in terms of aging because we are the
one species most likely to experience substantial aging, thus we are also the
species in which aging is likely to have the most substantial impact on our
ecological interactions. However, a variety of reasons, particularly the ability of
researchers to control the accumulated experiences of an experimental subject,

force us to seek model organisms.

How does the relationship between aging and the environment integrate into our

physiological understanding of aging?

Current well-developed models of aging have been focused on
understanding the genetic and developmental mechanisms involved in aging
(e.g., Drosophila, Caenorhabditis elegans) or an understanding of particular
physiological processes (e.g., mice). Due to limitations of these models, such as
an inability to evaluate their ecological context or a generation time which
inhibits thorough demographic work, it is unlikely that they will be useful for
integrating our understanding of the genetic, physiological and developmental
aspects of aging with the perspective of ecology. This integration is important,
because through it we will be able to understand the causes of the complex web
of interactions and trade-offs underlying physiological, developmental,
biochemical and other types of traits which degrade as an organism ages.
Therefore, if we are choosing a new model system for aging specifically to
integrate the ecological perspective, it is imperative that we are still able to
evaluate the traits of the organism that are the foundation of our interest in
aging.

Essentially, this question is about bridging the interests of the
gerontologist to the interests of the evolutionary biologist. To the gerontologist,
aging is a property of individuals, a deterioration of the functioning of the body.
To the evolutionary biologist aging is a property of populations, a demographic
phenomenon embodied by declining average fecundity and survival probability.
A clear understanding of how these two levels, the individual and the

population, are linked is important for a thorough understanding of either level.

The merits and demerits of microcrustaceans

Consider the general properties that make an organism amenable to

laboratory studies in evolutionary biology. First, the organism needs to be fairly

easy to maintain in the lab. This means that the organism should be small, so
that many individuals can be packed into limited lab space. It also means that
the environmental conditions required need to be well known and
straightforward. Furthermore, for evolutionary biology, it is desirable to have
high fecundity (enabling the researcher to generate adequate sample sizes).
Second, the organism should have a short generation time. This is because
many types of investigations will be improved by a multigenerational outlook.
Also, it is becoming more and more apparent that maternal effects (aspects of the
maternal environment/ genotype which have an influence on the phenotype of
offspring) must be controlled for, and preferably incorporated into studies.
Thus, several generations of lab acclimation (and perhaps also acclimation to
specific experimental conditions) may be important. A short generation time is
usually correlated with a short lifespan, a feature which will be particularly
useful to studies of aging. Third, it is always helpful to choose a study organism
for which there is a substantial amount of accumulated knowledge. This allows
one to proceed more directly to questions of interest, and allows one to support
reasonably the assumptions of an experiment and to define the conditions across
which results can be generalized.

Another property of model organisms that would be important for studies
of aging in particular is the ability to assay aging. This is incorporated in the
general properties if we want to use a demographic index of aging (which is the
main interest of the evolutionary biologist). However, particular physiological
features usually are of interest, such as oxygen consumption, stress tolerance,
disease resistance, or locomotive ability. This is especially true from the
gerontologist’s perspective. In developing a model system for aging, these
should be taken into account, because they allow relating the aspects of bodily

deterioration directly to evolution. In fact, we would ideally want an organism

that has an analog for as many of the characteristics of aging in humans that are
important as possible. A property which will facilitate assays of aging is
iteroparity (multiple reproductive events per lifetime). This is because the
portion of lifespan over which aging can be expected in animals which reproduce
once and then die is so compressed that making meaningful observations of the
process would be difficult.

Crustaceans represent an opportunity to more fully integrate ecology into
our knowledge of aging (Reznick 1993), something that would be very difficult
with currently widespread model systems. There are several candidate
crustaceans for research in the evolutionary biology of aging. The basic
considerations of laboratory amenability, short generations and high fecundity
lead to the microcrustaceans, aquatic (fresh and saltwater) crustaceans of the
groups Copepoda and Cladocera. I focus on the family Daphniidae (Cladocera)
because they are the family for which there is the most substantial body of
knowledge which is pertinent to investigations of the biology of aging.

Microcrustaceans are small (usually <4mm), poikilothermous arthropods
with well-developed circulatory, respiratory, neurological, digestive,
reproductive and muscular systems. In addition, there is tremendous diversity
in these physiological systems, necessary to cope with the diversity of habitats in
which they are found. For example, herbivores, omnivores and detritivores can
be found in the family Daphniidae; digestive systems must vary to assimilate the
different types of food. Variation in respiratory function is necessitated because
the oxygen concentration in water can vary widely. Daphnia pulex, a common
North American species, is known to have reproductive modes including
obligate asexuality, hermaphroditism, and sexuality in the wild; this presents an
unparalleled (and unexplored) opportunity for genetic control. Members of the
family Daphniidae are particularly well suited to physiological and

developmental biology because they are transparent and many intraorganismal
processes can be observed with nothing more than a light microscope. A large
body of physiological and developmental information has been generated (Peters
and de Bernardi 1987), although it has not been applied to aging.

Daphniidae have been the subject of research (ecological, developmental
and physiological) since the mid-1800’s. In that time, they have become as much
a model organism in ecology as there exists. In particular, they have been the
subject of work on population biology, community ecology and life history
studies. Culture methods are well-established (Goulden, et a1 1982, Peters 1987),
based on a good understanding of nutrition (Lampert 1987), and have been
defined with sufficient precision for Daphnia to become one of the standard
bioassay tools of ecotoxicology. They (and all other microcrustaceans) grow in a
series of instars, developmental stages punctuated by molting of the carapace.
The infrastructure required to support culture of daphniids is not significant, and
maintenance of cultures is fairly inexpensive. Source populations can be found
in most lakes and many ponds. They have a short generation time (first
reproduction can occur as early as 5 days) and their lifespan can probably range
from 20-200 days, depending on species, population and environmental
conditions. The importance of having a model organism for which there is a
great deal of variation in lifespan (and, presumably, rates of aging) cannot be
overemphasized: it is the examination of variation that permits scientific
explanation of causes and consequences.

In addition to their general acceptability, Daphniidae have a number of
particular qualities which affect their usefulness in the evolutionary biology of
aging. First, and vital in terms of the questions delineated earlier in this chapter,
is the tremendous background of ecological knowledge about Daphniidae and
the potential to link it to a substantial body of physiological knowledge. This is

absolutely critical for understanding how natural selection has shaped aging,
and is not a feature of animals which are commonly used as models of aging,
such as Drosophila or C. elegans. Importantly, the type of ecological knowledge
about daphniids we have is appropriate to selection on aging. The forces of
selection which shape aging are those which alter the relative fitness value of
particular age classes. This means that the natural risk of mortality must be
understood, and furthermore, that understanding the age-specificity of the forces
of mortality will enable a more refined approach to studying aging. For
daphniids, the factors which cause extrinsic mortality are well known (Threlkeld
1987): impacts of numerous predators (and their size-specificity) have been
quantified, the temporal changes of abiotic factors which kill daphniids (anoxia,
pond-drying) are easily measured, and there is currently a resurgence of interest
in the microparasites (i.e., diseases) of Daphnia which should prove important.
There are approximately 40 species of Daphnia in North America, and
perhaps 125 world-wide; this is the most well-studied genus of daphniids, and
one of the most well-studied microcrustaceans. Not all species are equally well
known, nor has one particular species become the leading research subject,
although there are a handful of species which receive the bulk of attention. This
taxonomic diversity should not be viewed as a drawback, though, because it is
accompanied by ecological diversity. This, combined with emerging knowledge
of their evolutionary relationships, makes the genus a good candidate for
comparative studies of aging. Different species of Daphnia exist along a range of
selection pressures on aging, and should exhibit different patterns of aging. This
would then be the raw material for comparing particular aspects of the
physiological and biochemical mechanisms of aging. These comparisons can be
supported by the growing knowledge of the evolutionary relationships among

the species, genera and families of Cladocera.

One of the most valuable ecological tools for understanding the evolution
of aging is not our knowledge of what might kill Daphnia, but our ability to
quantify mortality easily in a natural population (Threlkeld 1987). Slightly more
complicated, but also well-developed, are methods of determining size (as age)
specific mortality risk. This allows us to move beyond simple descriptions of
how selection should be influencing aging in a population, to measuring the
selection a population is currently experiencing, thus freeing one from the
nagging feeling that something else important is being overlooked. If we are
willing to make the assumption that current selection reflects past selection, we
can draw conclusions about how the current phenotype has evolved. This
critical assumption can be considered in light of our knowledge of the factors
which cause Daphnia death in order to evaluate it.

The second major aspect of daphniid biology that is relevant to research in
the evolutionary biology of aging is their unusual mode of reproduction.
Daphniids are cyclic parthenogens, animals which typically reproduce by
ameiotic parthenogenesis (cloning of females) punctuated by generations of
sexual reproduction (Hebert 1987). This is the basis of the ease of working with
them in the lab, because large numbers of isogenetic females can be naturally
produced. Thus, it becomes easy to segregate environmental and genetic
influences on a particular trait. In the lab, it is usually simple to set conditions
under which parthenogenesis is the sole reproductive mode. Juveniles develop
directly into adults, and the timing of development can be adjusted with food
and temperature. Historically, breeding has been difficult due to a required
period of diapause for sexually produced eggs. Now, though, this presents only
a delay in hatching offspring, opening up the potential for elaborate breeding
designs involving cloning, selfing, inbreeding, outbreeding — even interspecific

hybridization is possible.

10

The unusual reproductive mode creates opportunities unavailable in other
organisms (How does maleness or femaleness influence aging? With daphniids,
genetically identical males and females can be produced to explore this), but it
also could be problematic for some types of questions. One unique advantage is
that the naturally occurring genetic diversity can be captured and maintained in
the lab, without resorting to methods which may induce inbreeding depression
or substantial change of variation. An equally important advantage is the ability
to explore the effects of a particular agent (a vitamin, a temperature, a hormone,
a predator) on a particular genotype. A potential difficulty is that some natural
lineages undergo large numbers of asexual generations (particularly obligately
asexual lineages) leading to generalized mutation accumulation, a complication
if one is interested in the genetic mechanisms by which aging evolves in Daphnia.

There is one glaring aspect of Daphnia biology that makes them less than
wonderful for aging research. This is the lack of a strong body of molecular
genetic and biochemical knowledge about this system. Although there is a wide
literature on the population genetics of Daphnia, based on allozyme characters,
and soon to be DNA characters, Daphnia simply have not been used in molecular
genetics. Historically, this probably was due to the inability of early geneticists
to find any phenotypic traits of Daphnia which have simple mendelian
inheritance (this was not for lack of trying) in combination with the difficulty of
breeding them. To my knowledge there are no known genes in Daphnia, they
have not been transformed, and are undeveloped as a system for genetic
manipulation. Even the number of chromosomes present in certain species is in
dispute. Furthermore, although there is a reasonable amount of whole-organism
physiological knowledge, this has not led to extensive exploration of molecular
biology and biochemistry. These have been the general directions aging research

is proceeding in other organisms, and Daphnia may not be able to catch up. This

11

is not to say they are not worth developing, though. Daphnia are more suitable
than current models for the major questions outlined above, and there is greater
potential for developing a robust molecular genetic knowledge of Daphnia than
there is of developing a robust ecological knowledge of, say, Drosophila.

One other, relatively minor, drawback to working with Daphnia is that as
ideal for lab studies as they are, sample sizes of 100,000 animals (which have
recently been used in Drosophila) are unrealistic for Daphnia. Not only would this
be very labor intensive, but space would be a problem. This is because as
Daphnia grow (from 0.5 to 3.5 mm in some species, but only to 1.2mm in others),
the volume of water in which they are kept must increase exponentially (due to
their feeding behavior) in order for the juvenile and the adult to experience the
same environment. Thus, not only are they slightly larger than other model
organisms, but the mechanism by which density affects them requires a greater-

than-linear increase in space relative to their growth.

Knowledge about aging in microcrustaceans

There have been rather few studies examining aging in daphniids, and
those that do exist typically report average lifespans of several clones under the
same conditions, or of a single clone under several conditions. The most
complete studies are from the 19305, when A. M. Banta and his associates made
several careful studies of aging in D. longispina. These studies are all the more
intriguing because they occurred before the theoretical explanation of aging was
developed, and before modern Daphnia culture methods became established
(they used Banta’s Manure Infusion as a food source).

In one study (Wood, et al 1939a), they were simply observing growth and

reproduction over the lifespan of a single clone, at 25°C and ”unlimited” food.

12

But their hourly observations over the =30 day lifespan illuminate life history
changes that occur during Daphnia senescence. While ”young” (until the ninth
instar), instars typically lasted 38 hours, and ended with the release of young
followed within minutes by molting. As animals aged, instar duration
lengthened and became irregular, and was decoupled from the release of
offspring (by up to 2 days). Clutch size also decreased as animals aged, and a
post-reproductive sterile period roughly equal to 1 instar was observed.
Furthermore, they report that animals ”gradually become sluggish in their
movements and may finally rest on the substratum,” a phenomenon I have
observed in four additional species.

Another study examined growth patterns of 6 different clones (inbred in
the lab), and is probably the first to report genetic variation of lifespan (Wood, et
al 1939b). In general, slow growth was associated with longer lifespan, but all 6
clones had shorter lifespans and were smaller (and usually slower growing) than
the standard, natural, clone used in the previous study. Other studies by the
same group showed that semi-starvation increased longevity, lowered heart rate
and reduced fecundity (Ingle 1933), but that semi-starvation followed by good
feeding increased longevity further while restoring lifetime fecundity to that of
individuals well-fed throughout their life (Ingle, et a1 1937).

A separate lab worked on the influence of temperature on lifespan in D.
magna, showing that a 10° increase approximately halved lifespan (MacArthur 8:
Baillie 1929a) and that metabolic rate (as indexed by heart rate) was inversely
proportional to lifespan (MacArthur and Baillie 1929b). Thus, by the end of the
19305, temperature and food level, two of the most important environmental
influences on lifespan had been demonstrated in Daphnia. These early studies
are the only microcrustacean studies focused on aging, and only rarely since has

lifespan even been reported in the large body of zooplankton life history

13

research. Korpelainen’s (1986) data confirms the effect of temperature in D.
magna, and that there is naturally occurring genetic variation in the response to
temperature among 5 clones. She also includes data on males of the same 5
clones, and the variation among their temperature response differs from that of
the females. Furthermore, she shows a similarly complex pattern of influences of
photoperiod on lifespan. None of these studies have explored the environmental
conditions which would have selected for the observed phenotypes. Very recent
experiments are starting to change that. Substantial genetic variation among
populations and species has been shown for four species, which, at least at the
species level, appears to be related to natural selection (Dudycha, unpublished

data).

Prospectus: Developing microcrustaceans as a model of aging

The current utility of microcrustaceans for the evolutionary biology of
aging is primarily in the comparative method. Comparative biology has not
been an important component of evolutionary aging research, but it presents the
best opportunity for understanding the process in nature. Daphniids appear to
be an ideal organism for this research, because there is a substantial range of
ecological variation within the family, including situations where evolutionary
acceleration and deceleration of aging is predicted. They also present a unique
opportunity where one can quantify naturally-experienced mortality risks
(selection) and rates of aging, as assayed by a wide variety of indices, under
specified conditions in a genetically controlled fashion. Relevant ecological
variation also occurs within species among populations, but comparisons may
not be straightforward due to factors such as gene flow or plasticity, both of

which tend to dampen evolutionary change. Overall, by applying the

14

comparative method to daphniids, we should be able to explore whether natural
patterns of aging are associated with particular environmental influences. More
specifically relevant to geriatrics, the subsequent ability to disentangle the suite
of traits which make up bodily deterioration will be enhanced through an
understanding of the components of natural patterns of aging and the
mechanisms by which they are influenced.

The future utility of Daphnia could include manipulative experiments if a
”genetics toolbox” is built as a foundation. The power of Daphnia as a model
organism would become truly impressive: it would encompass genetics,
molecular biology, development, physiology, demography, ecology and
evolution. The ability to identify particular loci that affect aging in Daphnia
would lead to the potential for describing natural allelic variation and its
relationship to the selective environment. The ability to manipulate particular
loci would enable experiments that investigate the mechanisms whereby aging
affects ecological interactions. These two distinct avenues of research could
greatly illuminate how aging of individuals in nature affects and is affected by
interactions with other organisms. Finally, we would have a model system
where aging could be explored throughout the biological hierarchy, guiding us

in our attempts to understand the aging of humans.

Literature Cited

Charlesworth, B. 1993. Evolutionary mechanisms of senescence. Genetica 91:
11-19.

Edney, E. G. and R. W. Gill. 1968. Evolution of senescence and specific
longevity. Nature 220: 281-82.

15

Goulden, C.E., R. M. Comotto, J. A. Hendrickson In, L. L. Henry and K. L.
Johnson. 1982. Procedures and recommendations for the culture and use of
Daphnia in bioassay studies. In Aquatic Toxicology and Hazard Assessment, 5th
Conf. Am. Soc. Testing Mater., ASTM STP. 148-169.

Hamilton, W. D. 1966. The moulding of senescence by natural selection. Journal
of Theoretical Biology 12: 12-45.

Hebert, P. D. N. 1987. Genetics of Daphnia. In Daphnia, Peters and de Bernardi,
eds. Istituto Italiano di Idrobiologia, Verbania Pallanza.

Ingle, L. 1933. Effects of environmental conditions on longevity. Science 78:
511.

Ingle, L., T. R. Wood, and A. M. Banta. 1937. A study of longevity, growth,
reproduction and heart rate in Daphnia longispina as influenced by limitations
in quantity of food. Journal of Experimental Zoology 76: 325-352.

Jurine, L. 1820. Histoire des Monocles qui se trovent aux environs de Geneve. J.J.
Paschoud, Geneve.

Korpelainen, H. 1986. The effects of temperature and photoperiod on life history
parameters of Daphnia magna (Crustacea: Cladocera). Freshwater Biology 16:
615-620.

Lampert, W. 1987. Feeding and nutrition in Daphnia. In Daphnia, Peters and de
Bernardi, eds. Istituto Italiano di Idrobiologia, Verbania Pallanza.

MacArthur, J. W. and W. H. T. Baillie. 1929a. Metabolic activity and duration of
life. 1. Influence of temperature on longevity in Daphnia magna. Journal of
Experimental Zoology 53: 221-242.

MacArthur, J. W. and W. H. T. Baillie. 1929b. Metabolic activity and duration of
life. H. Metabolic rates and their relation to longevity in Daphnia magna.

Journal of Experimental Zoology 53: 243-268.

16

Peters, R. H. 1987. Daphnia culture. In Daphnia, Peters and de Bernardi, eds.
Istituto Italiano di Idrobiologia, Verbania Pallanza.

Peters, R. H. and R. de Bernardi, eds. 1987. Daphnia. Istituto Italiano di
Idrobiologia, Verbania Pallanza.

Reznick, D. 1993. New model systems for studying the evolutionary biology of
aging: Crustacea. Genetica 91: 79-88.

Rose, M. R. 1991. Evolutionary Biology of Aging. Oxford University Press

Threlkeld, S. T. 1987 . Daphnia population fluctuations: Patterns and
mechanisms. In Daphnia, Peters and de Bernardi, eds. Istituto Italiano di
Idrobiologia, Verbania Pallanza.

Williams, G. C. 1957. Pleiotropy, natural selection and the evolution of
senescence. Evolution 11: 398-411.

Wood, T. R, L. Ingle and A. M. Banta. 1939a. Growth and reproductive
characteristics of Daphnia longispina. In Studies on the Physiology, Genetics and
Evolution of Some Cladocera. A. M. Banta, ed. Carnegie Institute of
Washington.

Wood, T. R., A. M. Banta and L. Ingle. 1939b. Growth of genetically different
clones. In Studies on the Physiology, Genetics and Evolution of Some Cladocera. A.
M. Banta, ed. Carnegie Institute of Washington.

17

CHAPTER TWO: NATURAL GENETIC VARIATION OF LIFESPAN,
REPRODUCTION AND JUVENILE GROWTH IN DAPHNIA

with A. J. Tessier

Senescence is a post-maturation decline in the physiological state of
organisms as they grow older. From an evolutionary perspective, this decline is
most relevant as a decrease in age-specific rates of survivorship and fecundity.
The evolutionary theory of senescence (ETS) argues that the ultimate cause of
senescence is a decline in the force of selection as organisms age (Medawar 1952;
Williams 1957; Hamilton 1966; Charlesworth 1980). The force of selection
declines because phenotypic effects at later ages influence smaller and later
portions of reproduction relative to early phenotypic effects. Furthermore, these
later effects are not exposed to selection in individuals who die before
expression. Populations facing a greater risk of extrinsic mortality have a
stronger decline in the force of selection and greater senescence should evolve in
them. ETS has substantial support from laboratory experiments that have tested
predictions about the response to selection on late-life performance (e. g., Rose &
Charlesworth 1980; Mueller 1987; Partridge 8: Fowler 1992; Zwaan, et al. 1995a).
Tests of critical assumptions, including the existence of mutations with age-
specific effects (Pletcher, et a1. 1998) and age-specific patterns of genetic variance
in mortality (Promislow, et a1. 1996) and fecundity (Tatar, et a1. 1996) also
support ETS. Research has focused on understanding and discriminating among
the genetic mechanisms that permit the evolution of senescence (e. g., Rose and
Charlesworth 1981; Rose 1984; Partridge and Fowler 1992). However,
ascertaining the relative contributions of the non-mutually exclusive genetic

mechanisms will not illuminate whether the general theory adequately explains

18

the wide diversity of senescence patterns found in nature. Relatively little work
has examined the ecological context of senescence evolution (Reznick 1993; Tatar,

et a1. 1997; Dudycha, in press).

Tests of ETS by artificial selection on stock laboratory populations have
dealt with situations where selection is direct and constant. Natural systems
may be more complicated due to indirect and changing selection pressures
(Abrams 1991, 1993), but in general, we expect higher extrinsic mortality to select
for greater senescence (Williams 1957; Hamilton 1966; Edney and Gill 1968;
Charlesworth 1980, 1993). However, if the mortality falls predominantly on
juveniles, reduced senescence is expected because adults are evolutionarily more
valuable (Hamilton 1966; Charlesworth 1980; Abrams 1993). Although there is
no general reason to expect that the relative magnitude of juvenile versus adult
mortality is correlated with the absolute magnitude of extrinsic mortality in
nature, they are often confounded in selection experiments (Clark 1987). Spatial
or temporal changes in factors exerting mortality in nature may shift the net
direction of selection on senescence through changes in the age-specificity of
mortality or changes in the total level of mortality. It is not clear that the success
of ETS in explaining senescence evolution in relatively straightforward
laboratory settings will translate into success in explaining senescence as it has

evolved in the wild.

Comparisons among taxa have suggested that maximum or mean
lifespan, which represent only one aspect of senescence, are consistent with ETS.
The ability to fly, a trait that presumably reduces extrinsic mortality risk, is
associated with longer maximum lifespan in mammals (Prothero and Jurgens

1987) and birds (Holmes and Austad 1995). Keller 8: Genoud (1997)

19

demonstrated a cross-taxon association between social system and estimated
mean lifespan of female ants. They argued that in captivity, queens live longer
than reproductive females from solitary taxa because queens’ lifespans have
evolved in a relatively protected colonial environment. A recent analysis of
avian and mammalian mortality schedules (Ricklefs 1998) moves beyond
lifespan and shows, under reasonable extrinsic mortality assumptions, that
acceleration of adult mortality rate is positively related to the extrinsic mortality
risk faced by young adults. These studies used organismal traits to approximate
vuh1erability to extrinsic mortality and estimate risk based on those
approximations. One study (Tatar, et al. 1997) has looked at an environmental
axis to make predictions based on exposure to mortality risk. Tatar, et a1. (1997)
showed that, when raised in a common environment, male grasshoppers from
low elevations (i.e., a long season) live longer and have slower increases in age-

specific intrinsic mortality than those from high elevations.

Complete life table data are difficult to acquire, particularly when
quantifying senescence, because one normally needs to shield individuals from
extrinsic mortality (but see Promislow 1991; Ricklefs 1998). This difficulty places
limits on the interpretation of comparative studies, which have usually relied on
summary statistics such as the maximum or mean lifespan. First, senescence is a
non-linear change over time whose trajectory can vary in shape and magnitude.
Simple summary statistics may not capture this variation. Second, mortality is
only part of the fitness equation. Testing predictions of ETS has often yielded
supportive data, but many empirical reports (e. g., Curtsinger, et al. 1992; Carey,
et a1. 1992; Tatar, et a1. 1993) have revealed patterns of mortality not completely
predicted by ETS. The unexpected patterns may be due in whole or in part to
shortcomings of the theory (Blarer, et a1. 1995; McNamara and Houston 1996;

20

Pletcher and Curtsinger 1998), or it may simply be that the expected senescence
pattern was expressed in unmeasured traits. Reproduction must be included for
an accounting of fitness reduction due to senescence because age-dependent
changes in a tradeoff between survival ability and reproduction may cause
declines in either ability (Partridge and Barton 1993, 1996). We are unaware of
comparative work specifically incorporating reproductive declines that evolved

in nature.

Objectives of this Study

We compare senescence patterns in closely related taxa living across a
known gradient of mortality risk. First, we quantify the magnitude and
distribution of natural genetic variation in senescence using full mortality and
fecundity data collected under standard conditions that shield the animals from
extrinsic mortality. Second, we look for tradeoffs (negative genetic correlations)
between composite indices of senescence and early-life performance. Such
tradeoffs are a mechanism by which rapidly senescing genotypes could persist in
the face of invasion by long-lived genotypes and provide a framework for
understanding how senescence is integrated into the total life history. Third, we
consider the potential for further evolution of senescence in nature by examining
differentiation of senescence among populations within species. We focus on
multiple populations of the species complex Daphnia pulex-pulicaria, a group of
freshwater microcrustaceans well-suited to studies of senescence and ecology

(Bell 1984; Reznick 1993; Dudycha, in press).

21

Daphnia and the Pond-lake Gradient

The taxonomic status of Daphnia pulex and D. pulicaria is controversial.
Two recent mtDNA phylogenies conclude that D. pulex and D. pulicaria are not
distinct clades (Lehman, et a1. 1995; Crease, et a1. 1997). However, the most
recent taxonomic revision (Hebert 1995) considered them separate species, based
on differences in habitat occupied (pond or lake) and fixed allelic variation at the
LDH allozyme locus. Because we are mainly interested in the habitat differences
among taxa, here we discuss them as distinct species. Our view is that the
nominal species are ecologically distinct, but extremely close genetic relatives.
Naturally occurring hybrid (LDH heterozygotic) populations are also fairly

common and so are included in our study.

Daphnia pulex and D. pulicaria are small (< 3.5mm) crustaceans, which
appear to segregate along an environmental gradient of habitat permanence
(Deng 1997). Daphnia pulex is typically found in temporary ponds, while D.
pulicaria is found in deep lakes. Populations of D. pulex must recruit from
diapausing eggs in spring and undergo a variable period of reproduction via
immediately developing eggs before producing diapausing eggs. Anoxia or
desiccation kills individuals that survive to early summer. Variation in pond
size, local hydrology and annual weather patterns create variation in the length
of time any one pond is habitable by D. pulex, but all of our study ponds dry by
summer. Hence, maximum lifespan of D. pulex is constrained to a few months
by the temporary nature of their habitat. In contrast, D. pulicaria lives in a
permanent and environmentally more stable habitat, and can persist year-round
in deep lakes (Geedey, et a1. 1996; Geedey 1997). In the absence of any strong

abiotic constraints, resources and predators regulate D. pulicaria population

22

dynamics, generating predictable peaks of habitat quality in the spring and fall
(Hutchinson 1967). Planktivorous fish constitute the most important predator,
but D. pulicaria can effectively avoid them in deep water (Tessier and Welser
1991). Populations of D. pulicaria display low birth and death rates even at the
time of peak planktivory (Leibold and Tessier 1998). Perennial populations of D.
pulicaria also produce both immediately developing and diapausing eggs, but we
have observed that diapause investment is much less than in D. pulex (Geedey

1997; Dudycha, unpubl. data).

Hybrids of D. pulex and D. pulicaria occur in a broader range of waterbody
type than either parent, including both shallow ponds and lakes. In some cases
they co-occur with one of the parents, but they typically occupy habitats poorly
exploited by either parent, such as permanent fishless ponds. Little is known
about hybrid population dynamics, however a few local populations become

undetectable in late summer. (C. Steiner, pers. com., 1998)

Gradients of habitat permanence cause strong differences in extrinsic
mortality faced by different populations if individuals have no mechanism to
escape habitat loss (e. g., migration). As habitat duration lengthens, the time until
certain death is extended, and therefore the duration of potentially useful adult
lifespan is also extended. Based on an expected lifespan driven by abiotic and
biotic factors regulating population dynamics, we predicted the evolution of
faster senescence in pond D. pulex than in lake D. pulicaria. We made no a priori
predictions regarding the senescence of hybrids, due to uncertainty regarding
how the parents’ genomes would combine and to the lack of any reasonable

expectation for adaptation to their place on the habitat duration gradient.

23

METHODS

Clonal Isolation and Lineage Establishment

Daphnia will reproduce in the lab by strictly ameiotic parthenogenesis,
allowing us to capture the genetic variation in nature by randomly sampling a
population and establishing isogenetic lines. We isolated five randomly chosen
clones from each of three populations of D. pulex, three populations of D.
pulicaria, and two populations of naturally occurring hybrids. We did not choose
populations randomly; rather, we chose populations of the parent species to
represent the extremes of the habitat permanence gradient. This consideration
did not have much effect on the choice of D. pulex populations, but we chose D.
pulicaria populations based on data indicating they were perennial populations
(Geedey 1997). We chose hybrid populations from permanent waters free of fish
planktivory. All populations are in SW Michigan, except one of D. pulex, which
is in Michigan’s Upper Peninsula. Clones were isolated in April and May 1996,
except for a few clones isolated in spring 1995. Collecting clones at this time of
year maximizes the genetic variation we sample due to recent hatching of
sexually produced diapausing eggs (Lynch, 1984; Tessier, et a1. 1992; Geedey, et
a1. 1996). We established lineages from all clones and acclimated them to lab
conditions (20-22°, satiating food) for 2 3 generations prior to life history trait
measurement (Tessier and Consolatti 1989, 1991). Routine monitoring and

microscopic examination (at 460X) verified that lineages were free of parasites.

24

Life History Trait Measurement

We started life tables with neonates whose mothers were all raised
simultaneously from birth for several weeks at low density (1 / 50 mL) to
minimize variable maternal effects. Life tables began concurrently with neonates
(~12 hr old) from the third or later clutches of offspring. For each population, 3
cohorts consisting of 2 neonates from each of the 5 clones (30
neonates / population) were set up. Some minor deviations from this ideal
occurred because a few clones failed to produce an adequate number of female
neonates on the day the experiment started. Animals were changed to fresh
water every 2 d, fed a satiating food level (20,000 cells / mL Ankistrodesmus

falcatus daily), and incubated at 20° on a 16:8 L:D cycle. Water volume was
adjusted as animals grew and died such that a cohort could not filter more than
50% of their water between feedings (increased from 10 mL / neonate to 300

mL / animal at the largest size; Knoechel and Holtby 1986). We recorded
mortality daily and fecundity every 2 d until all animals died.

Daphnia have two modes of reproduction, parthenogenetic reproduction
that produces immediately developing offspring, and (usually) sexual
reproduction that produces diapausing eggs. Only immediately developing
offspring were included in our estimates of age-specific daily fecundity (mx).
When a female produced an ephippium, the easily-identified structure into
Which diapausing eggs are deposited, we adjusted fecundity values by
considering that female to be unavailable for reproduction until the next instar
(Lynch 1989). Thus, mx was calculated by dividing the number of directly
developing offspring produced by the number of live females that were not

producing an ephippium. Normally, no eggs will be deposited in the ephippium

25

when males are not present. We assumed that individual females did not engage

in continuous ephippial production.

Juvenile survival was nearly perfect for most populations. As an
additional measure of juvenile performance, we estimated juvenile growth rates
of 50 clones of D. pulicaria, 11 clones of D. pulex and 16 hybrid clones. These
clones were isolated as part of another project from a broader array of Michigan
populations than those for the life tables were. Mothers were raised as in the life
tables. For each clone 18-24 randomly chosen neonates were immediately
harvested; an additional 10 neonates were raised to maturity (here, defined as
the day eggs were released into the brood chamber) under the same conditions
as in the life tables, except they were fed a higher food level (40,000 cells/ day).
This food level is functionally equivalent (satiating) to the food level used in the
life tables (Lampert 1987). Neonates and mature animals were dried overnight at
55°, then weighed (to within 0.1 11g) on a Cahn microbalance. We used time to

maturity in conjunction with neonatal and mature mass to calculate juvenile

specific growth rate (pg ugl d'1;Tessier and Goulden 1987).

Quantifying Aspects of Senescence

Mortality data reflects the capability of the average individual to maintain
its body under the daily rigors of life. Any degradation in that capability
appears as an increase in the age-specific conditional probability of mortality
(given survival to the beginning of the age class), commonly called the hazard
function. Even if maintenance capability is constant, survivorship will still
decline if it is imperfect. Hence, it is an increasing hazard that signifies

decreasing maintenance capability, and is considered evidence of senescence.

26

Fecundity data reflects the capability of the average individual to produce
progeny. Any degradation of this capability is manifest simply as a decline in

the age-specific fecundity rate, and is considered evidence of senescence.

Observing a decline in either survival ability or fecundity alone may be
evidence of senescence, but it may instead reflect a reallocation of effort between
these two fitness components. As an organism ages, it may trade off mortality
risk (i.e., somatic maintenance) with fecundity. Therefore, when comparing rates
of senescence among taxa it is important to examine age-specific patterns of
survival and fecundity jointly. We use two general approaches in our

comparison of populations of the Daphnia pulex-pulicaria complex.

We first analyze age-specific patterns of survivorship and fecundity as
separate fitness traits. Separately comparing the degradation of each trait across
taxa is straightforward since each reflects a component of age-specific fitness
(mortality hazard or per-capita fecundity) for the average individual. However,
only when both traits lead to the same conclusion will this approach address
senescence satisfactorily. Therefore, we also compare the taxa by combining

survival and fecundity into single measures of age-specific fitness.

One of the most widely used summary measures of age-specific fitness is
reproductive value, which specifies the contribution of each age class to r, the
rate of population increase. Reproductive value (v,) has been suggested as a
useful measure for senescence studies (Partridge and Barton 1996). However,
interpretation of v, with respect to senescence as a process of physiological
degradation is not at all straightforward. Because v, weights each age class’s

expected fecundity by its contribution to r, early reproduction has greater

27

”value” than equal reproduction at a later age. Therefore, as an alternative
measure of age-specific fitness, we also examine the unweighted contribution of
each age class to R0, the total lifetime reproduction. We refer to this age-specific

measure as the intrinsic value (ix).

Because our newborns enter age-class 1, we calculate reproductive value,

vx, with this equation (Goodman 1982):

 

er(x—l) oo .
__ -r1
vx — l 28 ljmj
x j=x

where Ix (cumulative survivorship) and 111, (per-capita fecundity) are taken
directly from the life table data and r is iteratively calculated by Newton-
Raphson approximation. Intrinsic value, i x, is calculated as the ratio of current
and expected future reproduction to the expected lifetime reproduction of an

individual:
I=x 1=0

We scale by expected lifetime reproduction (R0) to simplify comparison across
taxa that differ in overall capability. The scaling by R, causes intrinsic value to
reflect relative changes in age-specific fitness rather than absolute changes. In
some cases, comparison of absolute changes may be appropriate, however some
tradeoffs (e.g., between offspring size and number) may bias an absolute

comparison.

28

Statistical Analyses

To evaluate whether there was potential evidence for senescence in the

mortality data alone, we pooled the data for each taxon and fit it to a Weibull
model. Weibull models normally express age-specific mortality (11) at age x as llx

= 117(k)“ (Lee 1980). When 7, the dimensionless ”shape parameter,” equals

one, the hazard function is constant and there is no evidence of senescence. If y >

1, the hazard function increases with age; since this indicates that the conditional

per-capita mortality rate increases with age, it is potentially (pending analysis of

fecundity) evidence of senescence. The ”scale parameter,” A, scales the model to

a baseline rate of mortality, but is not indicative of senescence. We estimated 7

with the SAS v. 6.09 Lifereg procedure.

We applied two types of regression model to the mortality data to test for
differences among taxa in the pattern of mortality hazard. First, we used an
accelerated failure-time (AFT) model to test for differences in the tinting of high
mortality hazard. AFT models assume that a factor (in this case, taxon) affects
failure time (lifespan) multiplicatively, shifting hazardous periods along the
timeline (Kalbfleisch and Prentice 1980; Fox 1993). Analyses of failure-time data
in other fields normally address shifts in the time-period when failures occur
(Kalbfleisch and Prentice 1980; Lawless 1982; Fox 1993), but this has rarely been
done in investigations of senescence. Instead, lifespan is commonly examined
due to its inherent interest, but it serves to identify only the endpoint of
senescence. To apply the AFT model, we chose an underlying gamma
distribution because its flexibility allows, but does not require, the theoretical
expectation of monotonically increasing age-specific mortality (see Kalbfleisch

and Prentice 1980 or Fox 1993 for a general discussion of survival distributions).

29

We also ran comparisons using alternative biologically plausible underlying
distributions, including the Weibull distribution, and found our results were
robust regardless of distribution. Second, we used a proportional-hazards (PH)
model to test for differences among taxa in the magnitude of mortality hazard
increase. PH models assume that a factor affects hazard functions multiplicatively,
changing the magnitude of hazard in a given period (Kalbfleisch and Prentice
1980; Fox 1993). Juvenile mortality was excluded from analyses. All
survivorship analyses were run in SAS v. 6.09 with the Lifereg and PHreg

procedures.

We used repeated-measures AN OVA to test for differences among taxa in
the pattern of age-specific fecundity. A significant time X taxon interaction
indicates differences in age-specific changes of fecundity profiles and may be
evidence of differences in senescence if horizontal (age) compression of one
taxon’s profile relative to the other’ 8 is evident. If there is a significant
interaction and mortality analyses show the same relative contrast, then we can
conclude there is evidence of differences in overall senescence. We also used
repeated-measures AN OVA to test for differences among taxa in the shapes of
their reproductive value and intrinsic value curves. For the combined measures,

a significant time x taxon interaction indicates differences in the rate of change of

age-specific fitness with respect to r (reproductive value) or R0 (intrinsic value).

We tested for correlations between r and two indices that summarize the

major differences among populations in temporal distribution of age-specific

fitness. The first index was the age when vx peaks, which we determined by
fitting cubic polynomials, the simplest equations allowing asymmetry, to each

reproductive value curve. The second index was the age when only 25% of

30

intrinsic value remained. We chose 25% to allow most of the divergence between

taxa to occur, while minimizing any influence of ”tails” in the data.

RESULTS

Mortality and Fecundity

Daphnia pulex had a substantially shorter intrinsic lifespan (median = 38d,
max = 54d) than D. pulicaria (median = 62d, max = 99d; Fig. 1A). Both species
showed decreasing ability to maintain their bodies as they aged. This was
evident from both the increasing mortality hazard functions (Fig. 2) and Weibull
shape parameters > 1 (D. pulex y = 7.7, 95% CI = 5.1, 15.5; D. pulicaria 'y = 2.9, 95%

CI = 2.4, 3.6). The timing of high mortality hazard was significantly earlier in D.
pulex than in D. pulicaria (AFT model; p < 0.0001, 12 = 566.2, df = 1). The
magnitude of mortality hazard increase was also significantly greater in D. pulex
than in D. pulicaria (PH model; p < 0.0001, x2 = 57.6, df = 1). Hence, intrinsic
adult mortality increased both more and earlier in D. pulex than in D. pulicaria.
The survivorship curve of hybrids (median lifespan = 58 d, max = 91 d; Fig. 1B)
had a roughly similar shape to D. pulex (hybrid y = 6.1, 95% CI = 5.1, 7.7), yet it
was delayed, similar to D. pulicaria (Fig. 2). Within both parent species, there
was also significant among-population variation in the timing of high mortality
hazard (D. pulex: p < 0.0003, x2 = 16.3, df = 2; D. pulicaria: p < 0.0001, 12 = 33,1251,
df = 2). Similarly, variation in the magnitude of mortality hazard increase was
significant (D. pulicaria: p < 0.0001, x2 = 18.7, df = 1) or marginally so (D. pulex: p
= 0.066, x2 = 3.4, df = 1; Fig. 1A).

31

The shape of the fecundity curve for D. pulex was significantly
compressed horizontally (changes occurred faster) relative to D. pulicaria (time x
taxon interaction; F = 4.12, p < 0.0001, df = 1, 46; Fig. 3A). The difference between
the species in their degradation of reproductive output was striking, because
over the entire time that D. pulex was sharply declining (ages 35-55 d), D.
pulicaria continued to improve its daily fecundity, declining moderately only
after reaching 65 days old. Hybrids had a greater overall fecundity rate than
either parent (Fig. 3B). The shape of the hybrid curve was strongly peaked, like
that of D. pulex, but extended to much later ages like that of D. pulicaria. There
was also significant variation in the shapes of the fecundity curves among
populations within D. pulicaria and marginally significant variation within D.
pulex (D. pulicaria: F = 1.94, p < 0.0001, df = 2, 94; D. pulex: F = 1.38, p < 0.069, df
= 2, 54).

Combined Fitness Values

Despite exhibiting an earlier and steeper increase of mortality hazard and
a more rapid decline of fecundity, D. pulex had a higher r than D. pulicaria (t =
2.654, df = 2, p = 0.057). This was a consequence of D. pulex’s more rapid
increase in fecundity early in life (Fig. 3). There was no obvious pattern of

difference in expected lifetime reproduction (Table 1).

The shape of the reproductive value profile for D. pulex was significantly

compressed horizontally relative to D. pulicaria (time x taxon interaction; F =

7.43, p < 0.0001, df = 1, 46; Fig. 4A). Note that by the age when D. pulicaria peaks
in its ability to contribute to population growth, D. pulex has lost essentially all of

its ability, even though both species achieved similar maximal age-specific

32

contributions. Therefore, D. pulex lost its reproductive value faster. The
reproductive value curve of hybrids (Fig. 4B) rose rapidly like D. pulex, peaked
midway between the parents, and declined at late ages similar to D. pulicaria.
There was also significant variation within each parent species in the shapes of
the reproductive value profiles (D. pulex: F = 4.87, p < 0.0001, df = 2, 54; D.
pulicaria: F = 1.45, p = 0.010, df = 2, 94).

Intrinsic value is defined to be 1 at age = 0 and it remained 1 until animals
matured, then declined roughly linearly until about 95% of intrinsic value had
been lost. However, D. pulex lost its ability to contribute to R0 substantially faster
than D. pulicaria (time X taxon interaction; F = 23.24, p < 0.0001, df = 1, 46; Fig.
5A), creating a difference > 30 days between the ages when 25% of intrinsic value
remained. The hybrid populations, again, were intermediate to the parental taxa

(Fig. 5B).

There was a strong, significant negative correlation between r and the age
when v, peaks (r = - 0.923, p = 0.01) and a weaker correlation between r and the
age when 25% of intrinsic value remained (r = - 0.679, p = 0.064; Table 1 and Fig.
6). These correlations both suggest a genetic tradeoff between early life fitness
(which maximizes r) and late life fitness (which extends the age when vJr peaks
and when 25% i, remains). Since r is used in the calculation of v,, we looked for
any mathematical dependence between r and the age when U, peaks by running
simulations with hypothetical life history data (combinations of 4 Ix and 28 mx
profiles). In all life histories, reproduction began at age 11 and continued until
age 50. With these hypothetical life histories, we generated points that evenly
filled the range of parameter space (r: 005; age when v, peaks: 10-50 d) defined

33

by our empirical data (Dudycha, unpublished data). This was true for both non-
senescent (constant or decreasing age-specific mortality rate and constant or
increasing fecundity) and senescent (increasing age-specific mortality and/ or
decreasing fecundity) life history profiles. These simulations also confirmed that
the age when 0,, peaks is substantially affected by late-life traits, whereas r is not.
Thus we are certain that the tradeoff we inferred is biological, not mathematical,

in nature.

Juvenile Size and Growth

In our comparison of juvenile performance, we observed that D. pulicaria
matured in 6.4 :i: 0.1 SE, D. pulex in 5.4 i 0.2 SE and hybrids in 6.0 :l: 0.2 SE days.
These differences are small relative to lifespan, but are a substantial portion of
the juvenile period. One-way ANOVA revealed that per-offspring investment,
measured as neonate mass, varied among the taxa (p <0.0001, F = 11.8, df=2, 77).
D. pulex had the smallest neonates (2.4 i 0.2 SE 11g), followed by hybrids (2.9 i 0.2
11g) and D. pulicaria (3.5 :i: 0.1 ug). Post-hoc pairwise comparison showed that D.
pulicaria was significantly different from D. pulex (p < 0.0001) and from hybrids
(p = 0.0243). Differences in per-offspring investment between the parents are
even more obvious when comparing clones with similar mass at maturity (Fig.
7). Daphnia pulicaria had a significant correlation between neonatal and adult
mass (r: 0.83, p < 0.001; Fig. 7). The relationship was isometric (reduced major
axis regression slope = 1.0 i 0.08), showing that D. pulicaria clones had the same
juvenile specific growth rate (0.378 d: 0.003 SE pg ugl d'l), regardless of their
mass at maturity. Neonatal and adult mass of hybrids were also correlated (r =
0.68, p = 0.004) and isometric (r. m. a. regression slope = 0.98 i 0.19). However,
the elevation of the hybrids’ slope was distinctly below that of D. pulicaria

34

(ANCOVA: F = 88.2; df = l, 63; p < 0.001) and they matured slightly earlier,
contributing to an overall faster juvenile specific growth rate in hybrids (0.514 :1:
0.009 SE 11g ug'l d'l). Our sample of D. pulex clones spanned a limited range of
adult sizes, and there was no significant correlation between neonate and adult
mass (r = 0.3, p = 0.37). These D. pulex clones largely fell within the 95%
probability ellipse for the hybrids (Fig. 7), and because they had an even shorter

maturation time, their juvenile specific growth rates were faster (0.613 21:0.015 SE
ug ug'l d'l) than D. pulicaria (F = 611.9, p < 0.0001, df = 1, 59).

DISCUSSION

We conclude that senescence occurs in all of our study populations
because both the hazard functions increased and fecundity rates declined over
later life. Our analyses of mortality and fecundity combined also showed a
degradation of age-specific fitness in later life, excluding the possibility that the
separate degradations of survival and reproductive capabilities create an illusion
of senescence via a changing tradeoff between them. Furthermore, our data
show that D. pulex, a resident of ephemeral habitats, experiences greater
senescence than D. pulicaria, a resident of permanent habitats, as expected by the
evolutionary theory of senescence. Specifically, D. pulex has a shorter lifespan,
an earlier and steeper rise in mortality rate, an earlier and sharper decline of
fecundity, a horizontally compressed reproductive value and a more rapidly

declining intrinsic value than does D. pulicaria.

Our approaches to analyzing life history data with respect to the concept
of senescence are complementary. The separate mortality and fecundity analyses

document that both survival and reproductive abilities degrade in concert,

35

supporting a common assumption (Partridge and Barton 1993). Our analysis of
reproductive value specifies how much the combined degradation of fitness
components affects age-dependent loss of the ability to contribute to population
growth. This view of senescence highlights how the contribution of alleles to
phenotypic evolution changes with age, and may be especially important in
studies that seek to measure selection on senescence. Finally, our analysis of
intrinsic value gives a summary picture of biological abilities, showing the
overall degradation with age of an average individual. Under laboratory
conditions, where survivorship declines primarily due to intrinsic breakdown,
this measure is closer to a physiological view of senescence than reproductive

value.

Studies that examine traits under a single set of laboratory conditions, but
deduce predictions from field conditions, are potentially biased by genotype-by-
environment interactions. We conducted our life tables under abiotic conditions
that are benign and at a satiating food level. These conditions are not unusual
for any of our populations in the field. However, the average temperature D.
pulex experiences in ponds is warmer and closer to 20° than the temperature D.
pulicaria experiences in lakes, because D. pulicaria spends summer in cold deep
water (Leibold and Tessier 1998). Resource levels can be high in lakes or ponds,
but are usually lower in lakes than in ponds (Leibold and Tessier 1998; Tessier,
unpubl. data). Therefore, the lab environment resembles the pond habitat of D.
pulex more closely than the lake habitat of D. pulicaria. It is unlikely that an even
closer match to the pond environment would improve late-life fitness in D. pulex
enough to reverse the fundamental differences between the species’ senescence

patterns through a genotype-by-environment interaction.

36

Assuming that the lab environment is sufficiently similar to natural
conditions to make an unbiased extrapolation to the field, we can ask to what
degree the observed life history differences are functionally significant in the
field. The lifespans of D. pulex seen in the lab are shorter than the duration of
their habitats (typically, snowmelt occurs in late February/ early March and the
ponds dry in mid-June, making ponds habitable by D. pulex for about three
months). An exact match is not expected because there is interannual variation
in the duration of ponds, and most individuals are not born on the earliest

possible day, therefore potential lifespan is less than a pond’s average duration.

In contrast, lakes may be a permanent habitat, but habitat quality is not
uniform over the year. Daphnia pulicaria population density is often greatly
reduced in summer and under ice in winter due to low resource availability in
the former and the combination of low resources and very cold temperature in
the latter. Spring and fall are predictable periods of high algal (resource)
availability, and consequently Daphnia population growth is also high
(Hutchinson 1967; Sommer 1989). Individual daphnids who can survive from
spring to fall (or vice-versa) may have a considerable advantage over genotypes
who rely on diapause to get through summer and winter. Diapause is
disadvantageous because these eggs depend on imperfect environmental cues to
decide when to leave diapause. The lifespans we saw in the lab suggest that
substantial numbers of D. pulicaria may live long enough to cross the seasonal
troughs of habitat quality, because the average temperature and resource level
they experience in the field is lower than lab conditions. Daphnid lifespans have
been shown to lengthen in colder temperatures (MacArthur and Baillie 1929;
Ingle 1933) and under reduced resources (Ingle, et a1. 1937). When resources

increase in spring and fall, a population of D. pulicaria will be composed mainly

37

of large adults that persisted from the last good season. These adults can now
resume a high level of reproduction, which had necessarily declined during the
poor resource period (Leibold and Tessier 1998). Because this ”persistent”
demography will be more rapidly responsive to resource change than a
”diapause” demography, the population can exploit the resource flush better
than a population constrained to hatch from diapausing eggs or even one whose

age-structure is relatively juvenile.

We have argued that the differences in senescence between D. pulex and D.
pulicaria have evolved as a result of differences in their habitats. However, we do
not present data to show that mortality imposed by habitat loss (or shielded by
habitat maintenance) is the direct mechanism responsible for the genetic
differences in senescence. Other differences between the habitats may influence
senescence evolution. For example, temporary habitats force Daphnia
populations to use diapause. Senescence differences could evolve as a correlated
effect of selection on investment in diapause through costs of reproduction. This
selective pathway may contribute to the different senescence patterns expressed
by D. pulex and D. pulicaria in our study because D. pulex produced more
diapause structures than did D. pulicaria (1.2 vs. 0.3 ephippia per female per
lifetime). Size-selective predation may also vary between temporary ponds and
permanent lakes. Because size and age are correlated, size-selective predation
could exert selection on senescence. We do not know whether size-selective
predation systematically varies among our study populations. If predation falls
mainly on large individuals in temporary ponds and on small individuals in
permanent lakes, this could contribute to the evolution of the senescence

differences we found.

38

Tradeofis

What allows rapidly senescing organisms to persist despite the existence
of long-lived genotypes? One elaboration of ETS is based on the hypothesis of
genetic tradeoffs: that poor late-life fitness is genetically linked to good early-life
fitness through genes that are antagonistically pleiotropic (Hamilton 1966; Rose
1991 ; Steams 1992). In this scenario, selection for a good early-life phenotype
will indirectly select for a poor late-life phenotype. Optimality approaches to
senescence inextricably tie senescence to development (Hamilton 1966; Partridge
and Barton 1993, 1996), making ETS also a theory of early-life performance. The
developmental theory of senescence hypothesizes that pleiotropic genes benefit
development rate at the expense of extended adult survival and reproduction
(Zwaan, et a1. 1991, 1995b). The simple prediction is that juvenile performance

should be better in taxa that have faster senescence.

Life history theory focusing on offspring quality also predicts a link
between early- and late-life traits. Winkler and Wallin (1987) modeled the
relationship between total reproductive effort and per-offspring investment.
They predicted that organisms investing heavily in reproduction would have
smaller optimal per-offspring investment. Their model explicitly assumed a
tradeoff between reproduction and adult survival. Therefore, their model also
predicts a positive relation between optimal per-offspring investment and

lifespan.

We demonstrated genetically faster juvenile specific growth in D. pulex
than in D. pulicaria, continuing the developmental theory’s hypothesized

relationship between early performance and senescence. We also showed

39

greater per-offspring investment, measured as neonate mass, in D. pulicaria than
in D. pulex, supporting the Winkler and Wallin (1987) model. Hybrids are
intermediate to the parental species for both of these traits, and for senescence.
D. pulex produced relatively small neonates who rapidly grew large, whereas D.
pulicaria produced large, slowly growing neonates. D. pulex evidently gains its
advantage in r by producing cheap, rapidly growing offspring. This negative
relationship between per-offspring investment and juvenile performance is
contrary to the usual expectation that greater investment in each offspring
produces juveniles who are more fit. However, through joint consideration of
the developmental theory and the per-offspring investment model, this negative
relationship becomes intelligible, because juvenile performance and per-
offspring investment are both mediated by a tradeoff with the senescence of

adults.

In a recent review of empirical research on growth rates, Arendt (1997)
argued that research rarely addresses tradeoffs between juvenile growth rate and
adult fitness (reproduction or maintenance) despite the possibilities that
organisms may be unable to reallocate resources perfectly from growth to adult
fitness or that rapid growth may build a less sturdy body. In other words, rapid
juvenile growth may be advantageous to early adult fitness, yet cause poor late-
age fitness. Arendt (1997) also emphasized that genetic variation in growth rate
is often overlooked on the assumption that there is no fitness advantage to
submaximal growth rates. Our juvenile growth experiment clearly shows that
there is variation in specific growth rate among these physiologically similar
taxa. In combination, our data showing associations between genetic variation in
the packaging of reproductive effort, senescence and juvenile growth suggest
that exploration of the mechanisms linking them will be a strong approach to

40

understanding life history. A large literature discusses optimal life histories
from the separate perspectives of senescence and per-offspring investment, but
these perspectives rarely intersect. Our results suggest that empirical work in
other taxa would benefit from a joint consideration of senescence and per-

offspring effort.

The tradeoffs we discovered would prevent invasion of temporary ponds
by long-lived genotypes because their reduced r places them at a disadvantage
when the pond dries, and genotypes can locally persist only by generating a
diapausing stage. At this critical accounting period, the number of individuals of
high r will be greater than those of low r. Assuming that the genotypes have
equivalent diapause, the long-lived will never catch up because the cycle is reset
every year. This rationale may explain the puzzle of rapid senescence, but it
highlights a more general problem in evolutionary biology: why genotypes with
low r are not displaced by genotypes with high r. Although it is speculative, we
suspect that the inability of short-lived, high-r genotypes to invade lakes may be
related to their inability to persist across habitat quality troughs. If an
individual’s lifespan precludes survival through an extended period of poor
habitat quality, its genotype can persist only through offspring production or
through diapause. Neither strategy is likely to be effective in lakes. Offspring
cannot survive and grow well in comparison to adults during poor resource
conditions, and diapause is complicated by the need for a reliable, accurate cue
of improving habitat conditions. Furthermore, a diapause strategy involves a
time lag, delaying the ability of diapausing genotypes to respond to the habitat
improvement compared to genotypes that persist as surviving adults. In sum,
the ability of long-lived individuals to span summer or fall may outweigh the
advantages of high lab-based r, because high r is genetically linked to other traits,

41

i.e. short lifespan and small offspring, that are disadvantageous during low

resource periods in lakes.

We focused our study on several populations of a single species complex
in order to make detailed life history comparisons under common conditions. At
this time, we do not have a phylogenetic estimate for these populations and
cannot say definitively whether our comparison includes independent
evolutionary divergences or the differences we found may be partly due to
historical constraints. However, the local existence of interspecific hybrids,
which occupy an intermediate position along tradeoffs between r and
senescence, suggests that the species are genetically very similar and historical
constraint on life history evolution may be minimal. Even if our populations
prove to be phylogenetically non-independent, the principal alternative theory
for explaining senescence, that degradation is caused simply by metabolic
exhaustion (Comfort 1979), cannot explain the variation we found. A review of
the rather extensive work on metabolic rate in Daphnia (Peters 1987) found no
evidence for taxonomic differences; metabolism scaled negatively with size.
Therefore, if senescence were ultimately caused by metabolic exhaustion, we
would have seen no variation between the species, or faster senescence in D.
pulicaria, because D. pulicaria adults are on average smaller than D. pulex (Fig. 4;

Hebert 1995).

Our results indicate that there is potential for continued evolution of
senescence in these taxa because we observed non-trivial extant genetic variation
for age-specific fitness and lifespan within gene pools. Moreover, at the
macroevolutionary level, these species are not reproductively isolated, so one

could argue that the variation throughout the species complex is available for

42

further evolution in a single gene pool. However, since we only examined
variation at the level of populations (and only 3 populations per species), we are

unable to generate reliable estimates of the heritability of senescence.

A recent study by Lynch, et a1. (1998) proposes an alternative
interpretation to standing genetic variation of life history traits. By comparing
the genetic variation of life history traits in a natural population of D. pulex to
that created by a laboratory mutation accumulation experiment, they concluded
that much of the natural genetic variation may be due to continually arising,
transient deleterious mutations. Our study differs by examining among-
population variation (rather than within) and by emphasizing late-life traits,
whose mutations will be less transient than the early-life traits studied by Lynch,
et al. (1998). Variation in late-life traits will be less transient than in early-life
traits because selection is less effective at late ages. Furthermore, among our
populations senescence trades off with early life history, a pattern that cannot be

explained by uniformly deleterious mutations.

Prior tests of ETS have included laboratory selection on senescence, which
give a detailed picture of senescence, and broad comparisons of many taxa,
usually addressing only mean lifespan. Our report complements these tests by
extending the understanding of senescence evolution to detailed life history
changes in the face of contrasting patterns of natural selection, and by jointly
addressing mortality and fecundity. We conclude that the pattern of senescence
exhibited by Daphnia pulex—pulicaria is consistent with ETS. This result is limited,
though, in that the selection pressures are historical, and thus cannot be directly
quantified, and that our desire for thorough senescence data limited the number

of populations we could study concurrently. It will be possible to expand our

43

findings, because there are many instances of Daphnia sister species inhabiting
different parts of the habitat duration gradient. Our results additionally suggest
a relationship among senescence, offspring investment and juvenile fitness
(specific growth rate). This raises the questions of what physiological traits
proximately cause the large differences in life history, whether the different life
history traits are affected by the same physiological traits, and what ecological
consequences emerge from that life history variation. A finer understanding of
genotype-environment interactions and of age-specific mortality risk in Daphnia

are the next steps to making the link.

LITERATURE CITED

Abrams, P. A. 1991. The fitness costs of senescence: The evolutionary

importance of events in early adult life. Evol. Ecol. 5: 343-360.

 

. 1993. Does increased mortality favor the evolution of more rapid
senescence? Evolution 49: 1055-1066.

Arendt, J. D. 1997. Adaptive intrinsic growth rates: An integration across taxa.
Q. Rev. Biology 72: 149-177.

Bell, G. 1984. Measuring the cost of reproduction. H. The correlation structure
of the life tables of five freshwater invertebrates. Evolution 38: 314-5

Blarer, A., M. Doebeli and S. C. Steams. 1995. Diagnosing senescence: inferring
evolutionary causes from phenotypic patterns can be misleading. Proc. R.
Soc. Lond. B 262: 305-312.

Carey, J. R., P. Liedo, D. Orozco and J. W. Vaupel. 1992. Slowing of mortality
rates at older ages in large medfly cohorts. Science 258: 457-61.

Charlesworth, B. 1980. Evolution in age-structured populations. Cambridge

Univ. Press, Cambridge.

. 1993. Evolutionary mechanisms of senescence. Genetica 91: 11—19.

 

Clark, A. G. 1987. Senescence and the genetic-correlation hang-up. Am. Nat.
129: 932-940.

Comfort, A. 1979. The biology of senescence. Elsevier, New York.

Crease, T. J., S.-K. Lee, S.-L. Yu, K. Spitze, N. Lehman and M. Lynch. 1997.
Allozyme and mtDNA variation in populations of the D. pulex complex

from both sides of the Rocky Mountains. Heredity 79: 242-251.

45

Curtsinger, J. W., H. Fukui, D. Townsend and J. W. Vaupel. 1992. Failure of the
limited-lifespan paradigm in genetically homogeneous populations of
Drosophila melanogaster. Science 258: 461-463.

Deng, H.-W. 1997. Photoperiodic response of sexual reproduction in the
Daphnia pulex group is reversed in two distinct habitats. Lirnnology and
Oceanography 42: 609-611.

Dudycha, J. L. in press. Microcrustacea as a model of aging. in J. E. Morley, H. J.
Armbrecht and R. M. Coe, eds. The science of geriatric medicine. Johns
Hopkins Univ. Press, Boston

Edney, E. G. and R. W. Gill. 1968. Evolution of senescence and specific
longevity. Nature 220: 281-82.

Fox, G. A. 1993. Failure-time analysis: Emergence, flowering, survivorship and
other waiting times. Pp. 253-289 in S. M. Scheiner and J. Gurevitch, eds.
Design and analysis of ecological experiments. Chapman and Hall, New
York.

Geedey, C. K. 1997. Seasonal clonal succession in Daphnia pulicaria p0pulations.
Diss., Michigan State University.

Geedey, C. K., A. J. Tessier and K. Machledt. 1996. Habitat heterogeneity,
environmental change, and the clonal structure of Daphnia populations.
Functional Ecology 10: 613-621.

Goodman, D. 1982. Optimal life histories, optimal notation and the value of
reproductive value. American Natualist 119: 803-323.

Hamilton, W. D. 1966. The moulding of senescence by natural selection. J.

Theor. Biol. 12: 12-45.

46

Hebert, P. D. N. 1995. The Daphnia of North America: An illustrated fauna,
version 1. CD-ROM. The University of Guelph, Guelph, Ontario.

Holmes, D. J. and S. N. Austad. 1995. The evolution of avian senescence
patterns: implications for understanding primary aging processes. Amer.
Zool. 35: 307-317.

Hutchinson, G. E. 1967. A treatise on limnology, vol H. John Wiley 8: Sons, New
York.

Ingle, L. 1933. Effects of environmental conditions on longevity. Science 78:
511.

Ingle, L., T. R. Wood and A. M. Banta. 1937. A study of longevity, growth,
reproduction and heart rate in Daphnia longispina as influenced by
limitations in quantity of food. J. Exp. Zool. 76: 325-52.

Kalbfleisch, J. D. 8: R. L. Prentice. 1980. The statistical analysis of failure time
data. John Wiley 8: Sons, New York.

Keller, L. and M. Genoud. 1997. Extraordinary lifespans in ants: a test of
evolutionary theories of ageing. Nature 389: 958-960.

Knoechel, R. and L. B. Holtby 1986. Construction and validation of a body-
length based model for the prediction of cladoceran community filtering
rates. Lirnnology and Oceanography 31: 1-16.

Lampert, W. 1987. Feeding and nutrition in Daphnia. In Daphnia, Peters and de
Bernardi, eds. Istituto Italiano di Idrobiologia, Verbania Pallanza.

Lawless, J. F. 1982. Statistical models and methods for lifetime data. Wiley,
New York.

Lee, E. T. 1980. Statistical methods for survival data analysis. Lifetime Learning

Publications, Belmont.

47

Lehman, N. M. E. Pfrender, P. A. Morin, T. J. Crease and M. Lynch. 1995. A
hierarchical molecular phylogeny within the genus Daphnia. Mol. Phyl.
Evol. 4: 395-407.

Leibold, M. A. and A. J. Tessier. 1991. Contrasting patterns of body size for

Daphnia species that segregate by habitat. Oecologia 86: 342-348.

 

. 1997. Habitat partitioning by zooplankton and the structure of lake
ecosystems. Pp. 3-30 in B. streit, T. Staedler and C. J. Lively, eds.

Evolutionary ecology of aquatic invertebrates. Birkhauser Verlag, Basle.

 

. 1998. Experimental compromise and mechanistic approaches to the

evolutionary ecology of interacting Daphnia species. in W. J. Resetarits, Jr.

and J. Bernardo, eds. Experimental ecology. Oxford Univ. Press, New
York.

Lynch, M. 1984. The limits to life history evolution in Daphnia. Evolution 38:
465-82.

. 1989. The life history consequences of resource depression in Daphnia

 

pulex. Ecology 70: 246-256.

Lynch, M., L. Latta, J. Hicks and M. Giorgianni. 1998. Mutation, selection and
the maintenance of life-history variation in a natural population.
Evolution 52: 727-733.

MacArthur, J. W. and W. H. T. Baillie. 1929. Metabolic activity and the duration
of life. 1. Influence of temperature on longevity in Daphnia magna. J. Exp.
Zool. 53: 221-242.

McNamara, J. M. and A. 1. Houston. 1996. State-dependent life histories. Nature
380: 215-221.

Medawar, P. B. 1952. An unsolved problem of biology. H. K. Lewis, London.

48

Mueller, L. D. 1987. Evolution of accelerated senescence in laboratory
populations of Drosophila. Proc. Nat. Acad. Sci. USA 84: 1974-1977.

Partridge, L. and N. H. Barton. 1993. Optimality, mutation and the evolution of
ageing. Nature 362: 305-31.

. 1996. On measuring the rate of ageing. Proc. R. Soc. Lond. B 263: 1365-

 

1371.

Partridge, L. and Fowler, K. 1992. Direct and correlated responses to selection
of age at reproduction in Drosophila melanogaster. Evolution 46: 76-91.

Peters, R. H. 1987. Metabolism in Daphnia. Memorie dell’Istituto di Idrobiologia
45: 193-243.

Pletcher, S. D. and J. W. Curtsinger. 1998. Mortality plateaus and the evolution
of senescence: Why are old-age mortality rates so low? Evolution 52: 454-
464.

Pletcher, S. D., D. Houle, and J. W. Curtsinger. 1998. Age-specific properties of
spontaneous mutations affecting mortality in Drosophila melanogaster.
Genetics 148: 287-303.

Promislow, D. E. L. 1991. Senescence in natural populations of mammals: A

comparative study. Evolution 45: 1869-1887.

Promislow, D. E. L., M. Tatar, A. A. Khazaeli, and J. W. Curtsinger. 1996. Age-
specific patterns of genetic variance in Drosophila melanogaster. 1.
Mortality. Genetics 143: 839-848.

Prothero, J. and K. D. Jurgens. 1987. Scaling of maximum life span in mammals:
A review. Basic Life Sci. 42: 49-74.

Reznick, D. 1993. New model systems for studying the evolutionary biology of

aging: Crustacea. Genetica 91: 79-88.

49

Ricklefs, R. E. 1998. Evolutionary theories of aging: Confirmation of a
fundamental prediction, with implications for the genetic basis and evolution
of life span. Am Nat. 152: 24-44.
Rose, M. R. 1984. Laboratory evolution of postponed senescence in Drosophila
melanogaster. Evolution 38: 1004-1010.

. 1991. Evolutionary Biology of Aging. Oxford Univ. Press, New York.

 

Rose, M. R. and B. Charlesworth. 1980. A test of evolutionary theories of
senescence. Nature 287: 141-142.

. 1981. Genetics of life history in Drosophila melanogaster. H. Exploratory

 

selection experiments. Genetics 97: 187-196.

Sommer, U. 1989. Plankton ecology. Springer, Berlin.

Steams, S. C. 1992. The Evolution of Life Histories. Oxford Univ. Press, Oxford.

Tatar, M., J. R. Carey and J. W. Vaupel. 1993. Long-term cost of reproduction
with and without accelerated senescence in Callosobruchus maculatus:
Analysis of age-specific mortality. Evolution 47: 1302-1312.

Tatar, M., D. E. L. Promislow, A. A. Khazaeli, and J. W. Curtsinger. 1996. Age-
specific patterns of genetic variance in Drosophila melanogaster. H.
Fecundity and its genetic covariance with age-specific mortality. Genetics
143: 849-858.

Tatar, M., D. W. Gray and J. W. Carey. 1997. Altitudinal variation for senescence
in Melanopus grasshoppers. Oecologia 111: 357-364.

Tessier, A. J. and N. Consolatti. 1989. Variation in offspring size in Daphnia and
consequences for individual fitness. Oikos 56: 269-276.

. 1991. Resource quantity and offspring quality in Daphnia. Ecology 72:

 

468-478.

50

Tessier, A. J. and C. E. Goulden. 1987. Cladoceran juvenile growth. Limnology
and Oceanography 32: 680-686.

Tessier, A. J. and J. Welser. 1991. Cladoceran assemblages, seasonal succession
and the importance of a hypolirnnetic refuge. Freshwater Biology 25: 85-
93.

T essier, A. J., A. Young and M. Leibold. 1992. Population dynamics and body-
size selection in Daphnia. Lirnnol. Oceanogr. 37: 1-13.

Williams, G. C. 1957. Pleiotropy, natural selection and the evolution of
senescence. Evolution 11: 398-411.

Winkler, D. W. and K. Wallin. 1987. Offspring size and number: A life history
model linking effort per offspring and total effort. Am. Nat. 129: 708-720.

Zwaan, B., R. Bijlsma and R. F. Hoekstra. 1991. On the developmental theory of
ageing. I. Starvation resistance and longevity in Drosophila melanogaster in
relation to pre-adult breeding conditions. Heredity 66: 29-39.

. 1995a. Direct selection on life span in Drosophila melanogaster. Evolution

 

49: 649-659.

. 1995b. Artificial selection for developmental time in Drosophila

 

melanogaster in relation to the evolution of aging: Direct and correlated

responses. Evolution 49: 635-648.

51

Table 2-1. Values of fitness measures and the timing of fitness contribution in

D. pulex and D. pulicaria. See text for explanation of variables.

 

D. pulex D. pulicaria Hybrid
Population WIS 0L3 MCP WAR PIN LAW WHT WIN
r 0.54 0.45 0.50 0.41 0.41 0.29 0.43 .47
R0 108 58 180 232 218 34 240 245
01, peak age 18.5 29.5 23.5 48.5 44.5 62 34.5 42
(days)
i,25 (days) 29.5 35 34 63 62 54.5 50 51

52

it 01111-
"’r’o‘o‘)’1 1 r t 00

9 lOOOOOOO‘eOO
' O 9

 

 

80 90 100

Cumulative Survivorship
.0 .o
O\ 00 H
2 O
a
B
8
8
8
a
a;

.O
m.

0.2

 

 

0 10 20 30 40 50 60 70 80 90 100

Age (days)

Figure 2-1. Survivorship of Daphnia pulex-pulicaria. Median lifespan is the age
where survivorship equals 0.5. A: 3 replicate populations of D. pulex (open
circles and dotted lines) and 3 of D. pulicaria (closed circles and solid line). B:
Data in top panel pooled by species (dotted and solid lines), and 2 replicate
populations of naturally occurring hybrids (crosses and dashed lines).

53

.o .o .o 9
15 U1 0‘ \1
I I I I I I I l I I I I I I l I I I

9
03
I I I I

 

9
N
I'lll

 

Mortality Hazard

.0
i—l
' I

 

 

O

0 10 20 30 40 50 60 70 80 90 100

Age (days)

Figure 2-2. Mortality hazard function of D. pulex, D. pulicaria and hybrids.
Symbols as in Figure 1. Standard error bars are based on replicate populations.

54

OHNOJrkU'IO\\]CXJ

 

 

 

Per-capita Fecundity

Oi—‘NUJtPUTQVWQ

 

 

Figure 2-3. Per-capita fecundity of D. pulex, D. pulicaria and hybrids. A: 3
replicate populations each of D. pulex and D. pulicaria. B: Data in top panel
averaged by species, and 2 populations of hybrids. Symbols as in Figure 1,
except in B: D. pulex and D. pulicaria have solid lines.

55

 

 

25,
20:
02153
,2 E
c>t,10
si
.5 O
3
3GP
s : 2
25- '
293-42; 388‘X1 B
m E
15j

 

 

 

0 10 20 30 40 50 60 70 80 90 100

Age (days)

Figure 2—4. Reproductive value of Daphnia pulex-pulicaria. Grey symbols and
lines are reproductive value data calculated as described in text. Black curves are
cubic polynomials fit to each population. A: 3 replicate populations each of D.
pulex and D. pulicaria. B: Mean values by species of data in top panel, and 2
populations of hybrids. Symbols as in Figure 1.

56

Intrinsic Value

 

 

 

0 10 20 30 40 50 60 70 80 90 100

Age (clays)

Figure 2-5. Intrinsic value of Daphnia pulex-pulicaria. See text for explanation of
intrinsic value. A: 3 replicate populations each of D. pulex and D. pulicaria. B:
Mean values by species of data in top panel, and 2 populations of hybrids.
Symbols as in Figure 1.

57

 

 

 

 

 

 

 

 

 

0.6 I I I r
o

3 0.5 ‘ 5 0 fl -"
a o
5 J I o
g; 0 4 . -
2:
ii

0.3 q - . u

0 2 l l L 1

° 10 30 50 7o 15 35 55 75

Age of Peak Reproductive Value Age of 25% Intrinsic Value

Figure 2-6. Relationship between early fitness and life history tinting in Daphnia
pulex-pulicaria. Age of peak reproductive value is the age when a cubic equation
fit to reproductive value peaks. Age of 25% intrinsic value is the age when only
25% of the intrinsic value remains. Black circles are D. pulicaria, open circles are
D. pulex and stars are hybrid. Each symbol represents a replicate population.
One standard deviation ellipses are drawn around the data.

58

 

Neonate Mass ( Hg)

 

 

 

 

100
Mass at Maturity ( #8)

Figure 2-7. Neonate and adult mass (in 11g) of clones in the Daphnia pulex-pulicaria

complex. Each symbol represents a separate clone. 90% confidence ellipses are
drawn around each taxon. Dark circles are D. pulicaria, light circles are D. pulex
and stars are hybrid.

59

CHAPTER THREE: A REACTION NORM APPROACH TO THE
EVOLUTIONARY ECOLOGY OF SENESCEN CE.

INTRODUCTION

One goal of evolutionary ecology is to understand how genetic
differences among taxa produce ecological differences in the way taxa respond
to environmental variation. In this paper, I examine how genetic variation
of age-specific fitness components is expressed in different resource
environments under benign and stressful temperatures. Despite the
widespread importance of food and temperature to life history, their joint
consideration is surprisingly uncommon (see Sarma and Rao 1991; Butler and
Burns 1991; Orcutt and Porter 1994 for examples). Furthermore,
investigations of age-specific fitness components often lack an ecological
context (Partridge and Harvey 1988; Reznick 1993). In placing age-specific
fitness into an ecological context, I show how this approach is useful for

understanding the distribution of senescence patterns in nature.

The evolutionary theory of senescence offers an understanding of life
history variation through the effects of extrinsic mortality risk on the strength
of natural selection. Traits appearing early in an organism’s life have a
greater effect on total fitness than those appearing late, causing the late traits
to be less exposed to selection than early traits (Medawar 1952; Williams 1957;

Rose 1991). If an allele is beneficial early in life, but detrimental later in life

60

(antagonistic pleiotropy), selection will favor it despite its deleterious effects
(Williams 1957). Deleterious late life traits can also be evolutionarily
sustained through mutation accumulation (Hamilton 1966; Charlesworth
1980). Because extrinsic mortality shapes age-specific potential contribution to
fitness, variation in extrinsic mortality risk among populations should lead to
the evolution of variation in investment in longevity and potential late-life
reproduction (Williams 1957; Hamilton 1966; Edney and Gill 1968;

Charlesworth 1980).

Closely related taxa that experience different mortality risks provide an
opportunity to study the evolution of senescence in nature (Reznick 1993;
Austad 1993a). By comparing their adult performance under controlled
conditions, one can test the hypothesis that higher mortality risk leads to the
evolution of a faster decline in intrinsic physiological fitness as organisms age
(the rate of senescence). Comparison of populations exposed to different
mortality risks has been used to support the evolutionary theory of
senescence in opossums (Austad 1993b), grasshoppers (Tatar, et a1. 1997), ants
(Keller and Genoud 1997) and Daphnia (Dudycha and Tessier, in press).
When comparing populations, it is useful to measure senescence as declines
in the age-specific probability of survival and age-specific fecundity, because
demographic measures synthesize the physiological manifestations of
senescence and reflect the essential components of evolutionary fitness

(Partridge and Barton 1993, 1996).

61

A problem arises in comparative studies of genetic variation because it
is usually impractical to mimic variable natural conditions precisely, and the
environments inhabited by different genetic lineages are likely to differ.
Rates of senescence measured in a single controlled environment may not
adequately reflect senescence as it occurs in nature. Comparisons between
taxa may be misinterpreted if the chosen environment substantially favors
one taxon, when a different environment could have favored the other
(Schlichting and Pigliucci 1998). A possible solution is to compare life
histories in a variety of environments, spanning a range that reflects natural

environmental variation.

Two environmental variables have been repeatedly shown to
influence senescence of animals: food level and temperature (Finch 1990;
Rose 1991). The extension of lifespan under dietary restriction has long been
known in a wide variety of organisms, including Daphmia (Ingle et al. 1937),
rotifers (Fanestil and Barrows 1965), rodents (McCay et al. 1939, 1956), spiders
(Austad 1989) and others (reviewed in Comfort 1979; Weindruch and
Walford 1988; Rose 1991; Yu 1994). Longer lifespan is also clearly associated
with cooler temperature based on evidence from fruit flies (e. g.,
Hollingsworth 1966; Miquel et al. 1976), mm (MacArthur and Baillie 1929;
Korpelainen 1986), rotifers (Meadow and Barrows 1971), grasshoppers (Tatar

et a1. 1997) and nematodes (Klass 1977). Few of these studies have examined

62

naturally evolved differences in the influence of the environment on
senescence (but see Harrison and Archer 1988; Tatar et a1. 1997) or included
fecundity senescence. Most have been primarily concerned with
understanding the proximate mechanisms of senescence and consequently
give only cursory attention, if any, to natural genetic variation or the

ecological relevance of the responses.

Perhaps the ecological relevance has been overlooked because of an
enduring perception that few wild animals die of old age. While this may be
true, it is not true that senescence occurs in few wild animals; senescence is
widespread and may substantially interact with ecological dynamics (Rose
1991; Promislow 1991; Ricklefs 1998). For example, senescence likely
influences the ability of animals to compete, to avoid predation and to find
mates. Senescence evolution may also be influenced indirectly by ecological
factors. If extrinsic mortality risk covaries with other aspects of the
environment, we might see the evolution of a plastic senescence pattern that
responds to detectable environmental changes. To build a useful bridge
between senescence in the laboratory and the wild, we need to address the
relationship between senescence and traits that determine evolutionary
success. Optimality approaches to theory developed for understanding
declines of late-life performance necessarily incorporate early-life events
(Hamilton 1966; Charlesworth 1980; Partridge and Barton 1993, 1996). This

suggests that r, an evolutionary fitness measure that is very sensitive to early

63

life performance, is a worthwhile component of investigations on the

ecological relevance of senescence.

In this paper, I report on the plasticity of senescence patterns of the
freshwater crustacean Daphnia in response to food and temperature. These
environmental factors vary widely in nature and can strongly influence
daphnid life history (e.g., MacArthur and Baillie 1929; Ingle et a1. 1937;
Korpelainen 1986; Orcutt and Porter 1994; Boersma et al. 1998). In previous
work, Dudycha and Tessier (in press) demonstrated substantial genetic
variation in senescence patterns in the WM species complex in a
single environment. I extend this work by quantifying the senescence of
members of two species complexes in multiple environments. My primary
goal here is to compare the senescence patterns of sister species in
environments representing a natural range of variability. I use these
comparisons to address the influence of environment on evolved differences
in senescence and to ask whether plasticity of senescence substantially alters
our understanding of the evolutionary theory of senescence in nature. I also
ask how consistent the patterns of senescence plasticity are at different levels
of biological differentiation. Secondarily, I am interested in how gross
ecological differences between the two species complexes could explain any

macroevolutionary differences in senescence.

Study System

Four ecologically distinct species of Daphnia were chosen for this study
because their ecology in Michigan is well known. D,_p_ulex and mm
are sister species in the daphnja subgenus; mm (formerly called 2.
WM and Milka (formerly W) are sister species in the
hyalgdaphma subgenus (Hebert 1995). Each pair of sisters form a complex
composed of two parental (nominal) species and naturally occurring hybrid
populations (Taylor and Hebert 1992; Hebert 1995; Colbourne and Hebert
1996). It is therefore difficult to demarcate evolutionarily discrete gene pools
at the species level. Despite a low degree of divergence and common
hybridization within each complex (Lehman et al. 1995; Crease et a1. 1997;
Taylor and Hebert 1992, 1993; Taylor et al. 1996), each parental species differs

in habitat use and seasonal phenology.

Daphniapulex is found in temporary ponds, while 1242;111:3113, is
found in deep lakes (Hebert 1995; Deng 1997). Populations of Mex recruit
from diapausing eggs in spring and undergo a variable period of
parthenogenetic reproduction via immediately developing eggs before
producing diapausing eggs. Anoxia or desiccation kills individuals who
survive to early summer. The length of time any one pond is habitable by D_.
pulex is variable, but all of my study ponds dry by summer. Therefore,

maximum lifespan of Mex is constrained to a few months by the

65

temporary nature of their habitat. In local ponds, water temperature varies
from 6°-28°, with potentially substantial diel fluctuation over the period that
M1223 populations are active. Food levels for we; are typically very
high, but can reach low levels when the population peaks and diapausing egg

production begins (A. J. Tessier, unpublished data).

In contrast, W lives in permanent and environmentally more
stable deep lakes, and typically persist year-round. (Geedey et a1. 1996; Geedey
1997). With no strong abiotic constraints, resources and predators regulate I;
pulicaria populations, generating predictable peaks of habitat quality in spring
and fall (Hutchinson 1967). During much of the year food availability is low
(A. J. Tessier, unpublished data). Deep lakes are a heterogenous thermal
environment for m. In winter, lakes are uniformly cold (<4°), but
in summer, temperate lakes thermally stratify, with upper waters reaching 20-
28° and bottom waters remaining near 10°. Although mm generally
remain in the cool bottom waters during summer, individuals can migrate
into the warm surface waters at night (Leibold and Tessier 1991; personal
obervation). Because {2.421113 and 12,_pJ.111g:_a_r_ia occupy strikingly different
habitats, the species face different extrinsic mortality risk (e.g., due to habitat
ephemerality) and we have documented substantial genetic variation of

senescence (Dudycha and Tessier, in press).

66

Daphmamegdgjag and Manama are ecologically less distinct than
2.2141282111153113. because both species inhabit thermally stratified lakes. 1;

mandatae appears to be perennial (like W Hall 1964; A. J. Tessier,
personal communication 1999) and typically remains in the warm surface
waters of large lakes (Taylor and Hebert 1992) during both day and night. In
contrast, Mantifera is found in smaller lakes, and appears to have a summer
annual phenology. In these lakes, fish planktivory may be higher than that
experienced by W and consequently mm typically displays
strong diel vertical migration (Hall 1964; Dodson 1972; Threlkeld 1979, 1980;
Leibold and Tessier 1991, 1998, personal observation). Thus, both species
experience wide temperature variation, albeit at different temporal scales.
Unlike the W contrast, it is unclear which species faces greater

adult mortality.

The differences in phenology and habitat use mean that the four

species differ in both relative mortality risk and average experienced

temperature. Overall, W experience greater extrinsic

mortality than _D_._palarja (Hall 1964; Threlkeld 1979, 1980; Leibold and
Tessier 1998). This is because W make greater use of the

risky upper water habitat in lakes (high fish predation) and have a smaller
body size, making them more vulnerable to invertebrate predators (Hall et a1.
1970; Dodson 1972; Branstrator 1998). It is unclear whether the periodic

catastrophic mortality risk faced by Male; or the consistently high predation

67

risks faced by MW will result in a higher average mortality
risk. It is clear, however, that W experiences the warmest
average environment because they use the warm epilimnetic habitat
extensively. ILpalez. experiences an intermediate average temperature,
ranging from the cold water of late winter to the warm water of late spring.
Maria has the coldest average temperature, as it typically opts to remain

in the cold hypolimnetic habitat during summer.

METHODS

Life Tables

Life tables were obtained for 16 populations of Daphnia to estimate
genetic variation of life histories. Three replicate populations of each
parental species and two p0pulations of each hybrid type were used. Daphnia
will reproduce clonally in the lab, allowing one to sample the genetic
variation in nature by establishing multiple isogenetic lineages. Populations
of M were chosen to reflect the extremes of the habitat
permanence gradient, and although these are the typical habitats of the
species, they are not a random sample of available populations. Species and
hybrid identity was confirmed with the diagnostic LDH allozyme (Hebert
1995). Populations of W were chosen that showed the
greatest relative abundance of the target species, as determined by the AO

allozyme locus (Taylor and Hebert 1992, 1993). Sister parental species do not

68

coexist in any of the chosen populations. All populations are in SW
Michigan, except one of D. palgx, which is in Michigan’s Upper Peninsula. I
isolated five clones from each population, obtained in early spring Meg;
paljgarja) or early summer (W to maximize the genetic
variation sampled (Lynch 1984; Tessier et al. 1992; Geedey et a1. 1996).
Lineages were established from all clones and acclimated to the lab
environment (20-22°, satiating food) for 2 3 generations prior to life tables

(Tessier and Consolatti 1989, 1991).

Mothers of experirnenal individuals were all raised from birth for
several weeks at a low density (1 / 50 ML) to minimize the role of maternal
effects in offspring trait variation. Life tables started with newborn
individuals ~12 hr old from the third or later clutches. For each population,
30 neonates were followed. Neonates were not measured as individuals, but
rather as 10-member cohorts consisting of 2 individuals of each of the 5 clones
from a population. A few clones did not produce an adequate number of
female neonates at the start of a life table; I compensated for this by using
additional neonates from other clones in as balanced a fashion as possible.
Cohorts were placed in fresh, filtered (1 um) lakewater every 2 d and
maintained on a 16:8 L:D cycle. Water volume was adjusted as animals grew
and died to prevent a cohort from filtering more than 50% of their water
between feedings (increased from 10-300 mL/ animal as they grew; Knoechel

and Holtby 1986). A small amount of cetyl alcohol was placed on the water

69

surface to minimize surface film mortality (Desmarais 1997). Mortality was

recorded daily and fecundity every 2 d until all animals died.

Life history variation was assayed under four different environmental
conditions: 20° or 26° at high or low food (20,000 or 3000 cells/mL
Ankistrodesmus falcatus fed daily), abbreviated 20°H, 26°L, 20°H and 26°L
below. Pilot experiments showed that cohorts died in a few days as juveniles
at 28° or at only 2000 cells / mL / day food (J. L. Dudycha, unpublished data).
The environments represent a benign and stressful temperature and satiating
or minimal food levels. All species studied experience a range of variation
including these temperatures and food levels in the wild, despite differences
in their mean environment (personal observation; A. J. Tessier, unpublished
data). In the most stressful environment, 26°L, one population of12_pahgar1a
failed to reproduce and I was therefore unable to calculate age-specific fitness

measures for it.

Space availability precluded conducting all life tables in all
combinations of food and temperature concurrently. I timed life tables
primarily to prevent any bias in comparisons across populations of one
species complex within any given environment (i.e., all populations of a
species complex were run together for each environment). This design
allows comparison of life history plasticity across taxa without bias.

Secondarily, I wanted to minimize bias in comparing the two species

70

complexes. To do this, life tables for the two complexes were run over
roughly the same time period (Manama: early June 1996 through
March 1997, Wm: mid July-1996 through mid-June 1997).
Temporal variation in neonate quality or lakewater chemistry, although
likely to have minor influence on senescence pattern, limits my ability to
assign causation of differences in life history specifically to temperature or

food because not all environments were used concurrently.
Estimation of Age-Specific Fitness

Reproductive value specifies the contribution of each age class to r, the
rate of population increase, and has been suggested as a useful measure for
senescence studies (Partridge and Barton 1996). However, it discounts age-
specific fitness with the population growth rate (r), creating difficulties in
interpreting it as a measure of individual performance (Dudycha and Tessier,
in press). Nonetheless, it usefully combines mortality and fecundity in a way
that shows age-dependent changes in the potential strength of selection. I
also estimate age-specific fitness as the expected contribution of each age class
to R0, the total lifetime reproduction. This has a more straightforward
relationship to individual performance, because it has no dependence on the
passage of clocktime, a factor extrinsic to the biology of organisms. I refer to

this measure as the intrinsic value (Dudycha and Tessier, in press).

.A /_

The form of the reproductive value equation depends on whether
newborns are defined as age-class 0 or 1. I defined them as age = 1, and

therefore calculated reproductive value as:

r(x-l) 00
e .
“’7
2e ljmj
-x

lx j_

 

11x:

(Goodman 1982). I took Ix (cumulative survivorship) and mx (per-capita daily
fecundity) are taken from the life tables. I used Newton-Raphson
approximation to calculate r. Intrinsic value, ix, is the sum of current and
expected future reproduction as a proportion of the expected lifetime

reproduction of an individual:

j=x i=0

(Dudycha and Tessier, in press). The numerator of this equation describes the
expected fitness of an age-class. Taking this as a proportion of expected
lifetime reproduction (R0) simplifies comparison across taxa that differ in
overall capability by causing intrinsic value to reflect relative changes in age-
specific fitness rather than absolute changes. Because this measure

incorporates both age-specific survival and fecundity, and reflects only the

72

intrinsic attributes of the average individual, it is an index of age-specific

fitness in both the evolutionary and physiological senses.

Statistical Analyses

Two types of regression model were used to test for differences between
parental species in their adult survivorship patterns. I used an accelerated
failure-time (AFT) model to compare differences in the timing of mortality
senescence. This model assumes a factor (here, taxon) affects failure time
(lifespan) multiplicatively, shifting the timing of hazardous periods
(Kalbfleisch and Prentice 1980; Fox 1993). I specified a gamma distribution in
the AFT model because its flexibility allows, but does not require, the
theoretical expectation of monotonically increasing hazard (Kalbfleisch and
Prentice 1980; Fox 1993). Alternative underlying distributions yielded

qualitatively concurrent results that are not reported here.

Second, I used a proportional-hazards (PH) model to test for differences
among taxa in the magnitude of mortality hazard increase. This is
biologically interpreted as testing for differences in the rate of mortality
senescence. PH models assume that a factor affects hazard functions
multiplicatively, changing the magnitude of hazard in a given period
(Kalbfleisch and Prentice 1980; Fox 1993). Juvenile mortality was excluded

from all analyses (AFT and PH) by estimating maturation age of each species

73

in each environment and excluding lifespans shorter than that age. All
survivorship analyses were run in SAS v. 6.09 with the Lifereg and PHreg

procedures.

Repeated-measures ANOVA was used as a split-plot, univariate test for
differences among taxa in their patterns of age-specific fitness. This approach
accounts for the correlation between repeated measurements of a singel
experimental unit. A significant time x taxon interaction indicates
differences in age-specific changes of fitness and is evidence of differences in
senescence if horizontal (age) compression of one taxon’s profile relative to
the other’s is graphically evident. Statistical tests for the interaction of a factor
with time in r-m ANOVA draw degrees of freedom from both the number of
replicate experimental units (here, populations) and from the number of
times a trait is measured (von Ende 1993). Therefore, the reported degrees of
freedom may seem overly large, but they are appropriate because they reflect

the high resolution in time at which I measured traits.

Twice I excluded a population from statistical comparisons between
species in a particular environment. One was the [Lpaligarjapopulation that
did not reproduce at 26°L. Because it did not reproduce, I could not calculate
fecundity, intrinsic value or reproductive value. I did not exclude it from
survivorship analysis. In statistical comparisons of W at

20°H, one population of W was excluded as an outlier; its

74

survivorship and fecundity were much worse than all other populations of
the species complex. Although it is possible that this is a true response this
environment, it is more likely a result of an undetected problem at the time
the life table commenced. Because inclusion of this population renders some
otherwise statistically different comparisons insignificant, I have included

this population in all graphs so the reader can see its influence.

Explicit comparisons of life history plasticity within- and between-
species complexes were performed on the population growth rate (r), and the
persistence of fitness into late life. ”Fitness persistence” was defined as the
age when only 25% of intrinsic value remained (i,25), an age sufficiently late
in life to reflect most of the divergence among populations, yet minimally
affected by any tails of very old ages. Response of r and i,25 were tested in an
ANOVA model that treated food, temperature and species complex as fixed
effects, and taxon nested within species complex as a random effect with GLM
in SAS v. 6.09. However, the test of an effect of taxon nested within complex
was done with the Sheffé model, rather than the SAS model because I am
interested in estimating genetic variances due to different taxonomic levels
(Fry 1992). When significant genetic variation was indicated, the proportion
of variation due to genetic differences within- versus between- species
complexes was estimated via maximum likelihood (SAS v. 6.09, proc

VARCOMP).

75

Hybrid populations were not included in tests comparing species’
survival, reproductive value or intrinsic value within a given species
complex because neither their mortality risks nor their genetics are
understood sufficiently to predict their senescence relative to their parents.
Hybrid populations were, however, included in comparisons between species
complexes and in analyses of plastic responses to environmental variation as

a population representative of their complex.

RESULTS

Differences between Max and Lpalarra life histories were
moderately robust across environments. At 20°, Max mortality rates
accelerated substantially earlier and faster than in W. These effects
were either minor or statistically insignificant at 26°. (Figure 1, Table 1). Age-
specific fitness decreased faster in m than in magma in all but the
most stressful environment (26°L), regardless of whether age-specific
contribution to fitness was measured with respect to R0 (intrinsic value; Table
2, Figure 2) or r (reproductive value; Table 2, Figure 3). In no environment
was there evidence that Lpaljgarra senesced faster that M25.
Temperature mainly affected mortality schedules, whereas food level mainly
affected fecundity. Overall, maria exhibited a much more plastic
mortality schedule than Aprils; (Figure 1), causing strong differences

between species in senescence at 20°, while only a slight or no difference at

76

26°. The species achieved similar maximal reproductive values at 20° (Figure
3) with a similar effect of food level. However, at 26° Mia; reproductive
value was higher and more sensitive to food level, reflecting its greater early

fecundity.

Differences between W and Qfinfifara were less
pronounced and more variable than those between Max and ILpaljgarja.
Comparisons of adult mortality are complicated by substantial juvenile
mortality in low food environments, particularly at 26°. However, there was
evidence that adult mortality accelerates later and more slowly at 26°H and

20°L in 12mm than in Mata; whereas the reverse was true at

20°H (Table 1, Figure 4). Intrinsic value followed a pattern similar to
mortality with loss occurring more slowly in Managua than M
at 26°H and 20°L, but more rapidly at 20°H (Figure 5, Table 2). The reversed
pattern at 20°H is dependent on exclusion of the apparent outlier; when it is
included, no differences between these species are statistically significant at
20°H. Survivorship, and consequently, intrinsic value of Madame;
dgrrflfara were most strongly influenced by temperature, with differences
between species evident in three environments (Figure 5). For m
reproductive value varied only mildly with age in all environments, causing
a slow loss of age-specific fitness with respect to r (Figure 6). Sharper declines

were evident in mafia: (Table 2), driven by higher fecundity peaks

(26°H, 26°L, 20°H) or earlier mortality (20°L), however confidence is low in

the differences at low food.

ANOVA showed that temperature and food level affected life histories
in markedly different ways. Food level strongly influenced r in all species,
but had little impact on the index of fitness persistence (i,25); temperature had
the reverse effects (Figure 7). Temperature also influenced the degree of
differentiation in fitness persistence among the populations, with greater
differentiation evident at the benign temperature. These patterns were
strikingly similar in both species complexes, so they were analyzed jointly,

pooling the effects of environment across taxonomy.

Food level was the only factor that had a substantial and significant
influence on r (p < 0.0001, F = 249.2, df = 1, 4.29; Figure 7). Temperature had a
slight effect on r (p = 0.0178, F = 13.7, df = 1, 4.36; Figure 7), but neither species
complex (p = 0.2095, F = 2.19, df = 1, 4.16) nor taxon nested within species
complex (p= 0.4159, F: 1.79, df = 4, 1.68) had any significant effect. There were
no significant interactions among any factors. Driven by food, the total model
explained >80% of the variation in r. The lack of any genetic role in
explaining r variation indicates that comparisons of taxa in other aspects of
life history are not confounded by differences in fitness with respect to the

broad set of environments chosen.

78

Fitness persistence (i,25), in contrast to r, was primarily influenced by
temperature rather than food level, although food effects were significant
(Table 3). Genetic variation was not apparent between complexes but was
strong among taxa within complexes (Table 3). There was also a significant
interaction of temperature with taxa within complex. This was caused by a
larger portion of variation in fitness persistence being due to genetic variation
within complexes at 20° (54.4% of total variation) than under thermal stress
(29.1%). At neither temperature was any of the variation in fitness
persistence due to genetic differences between the two species complexes. N 0
other interactions were significant, although a marginally significant food by
temperature interaction (Table 3) may be contributing to the environmental

effects on fitness persistence.

DISCUSSION

The expected pattern of accelerated senescence under higher
temperature and food level was evident in all four species of QaLhma. This
is consistent with prior studies of these environmental variables on Daphnia
longevity (MacArthur and Baillie 1929; Ingle 1933; Ingle et a1. 1937;
Korpelainen 1986). By studying them jointly, we see that the plastic response
of senescence to food was curtailed under thermal stress. The response of
senescence had little relationship to the response of population growth rate, r

(Figure 7). Across species, r was similar and more responsive to food than

79

temperature. Despite these broad patterns, clear differences between species

emerge from taxonomic comparisons made under different conditions.

Differences were not strong in comparisons between the two species
complexes, the highest taxonomic level studied. None of the genetic
variation in senescence, as indicated by fitness persistence (ix25), could be
attributed to differences between the complexes. However, there was
substantial genetic variation of senescence pattern among taxa within
complex. Although this was true for both complexes, it is apparent that I;
W harbors a wider range (Figures 2, 5). Because Mandates:
dgntjfgra has a smaller range of senescence variation, and its range occurs
roughly midway between 12.47am and Maria, no inter-complex
difference is apparent when their mean senescence is compared. The rank
ordering of slowest senescence in D_p_u_ljcarja, moderate senescence in D,
mandgjagfigmjfera and fast senescence in [Lpulax suggests that senescence
rate does not correspond to ancestry at the species level. Instead, this ranking
roughly corresponds with known ecological variation in the mortality risk
these species face, with D._p_uljgar;ia inhabiting a stable, low-predation
environment, WW facing substantial vertebrate and
invertebrate predation, and m facing catastrophic mortality a few

months after their environment becomes habitable.

80

There was also striking evidence of genetic variation in phenotypic
plasticity among taxa within complexes, but no obvious difference between
complexes (Table 3). Specifically, the effect of temperature on senescence
patterns was strong in both complexes while the effect of food was modest.
Temperature also affected the magnitude of genetic variation within species
complexes, which accounts for more than half of the variation at 20°, but less
than a third at 26° (Figure 7). This means that the phenotypic variation of
senescence on which selection could act is greatest under benign conditions.
Conversely, it suggests that evolved differences in senescence are reduced
under stressful conditions. Though the pattern was the same across species,
the response to temperature was most pronounced in W.
Thus, there are two broad associations of genetic differences in senescence and
the plasticity of senescence. First, variation in both the degree and plasticity of
senescence is pronounced at the taxa-within—complex level, but is not found
between complexes. Second, genetic variation of both senescence and the

plasticity of senescence is greater within MW than within Q

Senescence in Maria

The comparisons of life history between ILpJLle; and D_.pJ.11jgarja

suggest that in the natural environment of ponds and lakes, magma can

extend lifespan and reproduction to greater ages than D. palgx. Under the

81

benign temperature (20°) ILpule; senesced faster than _D._prrljgarja, regardless
of food level or how age-specific fitness was measured. However, at the
stressful temperature (26°), senescence was overall accelerated relative to the
benign temperature and interspecific differences were reduced. In the field,
the temperature animals experience will often be lower than those studied
here. It is unlikely that at lower temperatures, We; would exhibit such
extraordinary survivorship plasticity to generate a slower senescence rate
than D._paji_cari_a. It is similarly unlikely that lower temperatures would
produce substantially faster senescence in Mania, given that this would
be contrary to much that is known about senescence in ectotherms (Comfort
1979; Rose 1991). It seems safe to conclude that, despite some substantial
genetic-environment interactions, the species from a temporary habitat
senesces faster than the species from a permanent habitat, as predicted by the

evolutionary theory of senescence.

The comparison of 124mb; with maria at 26°H demonstrates that
qualitatively different results can be obtained with intrinsic value and
reproductive value, two different methods of combining mortality and
fecundity to measure senescence. At 26°H, the intrinsic value of Me;
appeared to decline more slowly than Marja (Figure 2); though slight,
the difference in rates was significant (Table 2). However, reproductive value
declined more rapidly in mm, as a result of declining from a higher

maximum reproductive value over approximately the same lifespan as Q

82

 

paljrarja (Figure 3); this difference in rates was also significant (Table 2). It is
therefore especially important that a clear specification of the biological
phenomenon of interest is made to draw accurate conclusions about
senescence evolution. The interspecific differences in intrinsic value appear
to be due to a higher level of plasticity in Ill—[211119343 to extend fitness into
late life, whereas interspecific differences in reproductive value appear to be

due to a higher level of plasticity in m to increase fecundity early in life.

Senescence in Lmendotaefientitera

The comparison of life history in W is quite
different than in Wm. This is expected, because the differences
in ecological niches between W and Menfifera are not as large.
Only under ideal conditions, abundant food and benign temperature, were
the survivorships and fecundities of the two species strongly different. Even
so, differences in mortality rate acceleration (timing and magnitude) were
generally significant, and differences in age-specific fitness declines were
significant when resources were abundant. Small but significant differences
between the parent species were a result of the low differentiation among

replicate populations within species.

Interpretation of the mortality patterns in Meagaegfiemifera is

hampered under scarce resources due to poor juvenile survivorship in both

83

species. In part, this may be due to plasticity of maturation time, which
lengthened when food was low. This response was not evident in Mag;
pularja, and implies that resources play different roles in juvenile success of
the two complexes. Adult mortality patterns were clearer with high food
because juvenile mortality was low. In general, the species showed similar
levels of survival plasticity, except that a greater response to temperature was

seen in W at high food than in Mentifera (Figure 4).

Considering all environments, the data suggest that the rate and
timing of senescence have been decoupled in W213. Because
most work on the evolution of senescence has addressed only the rate of
senescence, it is unclear how common decoupling of rate and timing is. I;
mendgjae had faster mortality rate accelerations, but the timing of mortality
rate acceleration was later than in Memflera (Figure 4, Table 1). Declines of
age-specific fitness as measured by reproductive value were also significantly
different between Marlene and Mentifera with high food. Because D_.
dentifera never achieved fecundity as high as those of mm
reproductive value in Mentifera was always low, and consequently declined
slowly. However, it was Lmeaglojae whose reproductive value typically
peaked later (Figure 6). Because faster and earlier senescence do not appear to
co-occur in this complex, this indicates that these are distinct aspects of

senescence.

Patterns of intrinsic value decline were significantly different between
the species at 20°, but the differences were opposite at the two food levels.
The beginning of intrinsic value decline was similar between species, but the
relative rates reversed depending on resource abundance (Figure 5). This
reversal appears to be caused by a greater tendency for Mendatae to respond
to food by changing its investment in late life survival (Figure 4). The greater
plasticity of mamas; is also evident in reproductive value (Figure 6),
likely resulting from a greater responsiveness of fecundity. Although the low
degree of within-species variation allowed detection of significant differences

at 26°H, the small degree of difference may be of little import.

Environmental effects on senescence

Examining the relationship between r, a fitness measure that is highly
sensitive to early-life events and insensitive to late-life, and an organism’s
ability to continue contributing to fitness as it ages is one approach to viewing
the general structure of a life history. Food level had a strong positive effect
on r through its effect on fecundity. However, it had a much smaller effect on
fitness persistence (ix25). The reverse is true of temperature (Figure 7).

Benign temperature led to extended fitness persistence relative to stressful
temperature, but it had only a minor effect on r. Explanations of the genetic
mechanisms causing the evolution of senescence assume, for simplicity, that

the environment has no effect on the expression of deleterious mutations or

85

pleiotropies (Williams 1957; Hamilton 1966, Charlesworth 1980). However, it
is clear that the environment does have an effect, in these data and in reports

from other organisms (Comfort 1979; Finch 1990; Rose 1991).

Temperature also has an interesting impact on the expression of
among-population genetic variation in r versus fitness persistence. At a
stressful temperature, variation of fitness persistence was low relative to a
benign temperature, suggesting that the impact selection could have is
reduced under temperature stress. However, the stressful temperature also
revealed stronger variation in r than the benign temperature, suggesting that
evolutionary change will be rapid. The benign temperature allows expression
of differential fitness persistence, while a high resource level elevates
population growth rate, so even small variation can be selected on quickly.
Thus selection should be most efficient in the most benign environment, and

the life history relatively optimized for such an environment.

Collectively, the data reported here suggest that evolutionary
differentiation of senescence in Daphnia has been recent and may be actively
evolving. Differences in senescence were minimal across the deepest
evolutionary split, but were strong within species complexes that are
incompletely diverged. Genetic variation in senescence is apparently
distributed at finer levels of phylogenetic relationship than subgenus. In

general, differentiation of senescence patterns among taxa is more related to

86

ecological differences among habitats than to the span of time since
evolutionary divergence. The pond-lake contrast of 12:—131W is
ecologically much greater than the lake-lake contrast of W
denrjfera, and we find a correspondingly larger contrast in overall senescence
and plasticity of life history in mill-11193113- This pattern of strong
variation at the tips of phylogeny suggests that senescence is a trait that is

actively evolving in the wild.

87

 

LITERATURE CITED
Austad, S. N. 1989. Life extension by dietary restriction in the bowl and doily
spider, Ermflneljapyramjgja. Experimental Gerontology 24:83-92.
Austad, S. N. 1993a. The comparative perspective and choice of animal
models in aging research. Aging Clinical and Experimental Research 5:
259-267.
Austad, S. N. 1993b. Retarded senescence in an insular population of Virginia

Opossums (nglelphjuirgmjana). Journal of Zoology 229: 695-708.

Branstrator, D. K. 1998. Predicting diet composition from body length in the
zooplankton predator WM. Lirnnology and Oceanography
43: 530-535.

Boersma, M., P. Spaak and L. DeMeester. 1998. Predator-mediated plasticity
in morphology, life history, and behavior of Daphaja: The uncoupling
of responses. American Naturalist 152: 237-248.

Butler, M. I. and C. W. Burns. 1991. The influence of temperature and
resource level on the fecundity of a predatory planktonic mite, Ejgna
e;igua Viets. Oecologia 88: 220-227.

Charlesworth, B. 1980. Evolution in age-structured populations. Cambridge
Univ. Press, Cambridge.

Colbourne, J. K. and P. D. N. Hebert. 1996. The systematics of North
American Daphnia (Crustacea: Anomopoda): a molecular phylogenetic
approach. Philosophical Transactions of the Royal Society of London

Series B 351:349-360.

88

Comfort, A. 1979. The biology of senescence. Elsevier, New York.

Crease, T. J., S.-K. Lee, S.-L. Yu, K. Spitze, N. Lehman and M. Lynch. 1997.
Allozyme and mtDNA variation in populations of the Me; complex
from both sides of the Rocky Mountains. Heredity 79: 242-251.

Deng, H.-W. 1997. Photoperiodic response of sexual reproduction in the
Daphniapale; group is reversed in two distinct habitats. Limnology
and Oceanography 42: 609-611.

Desmarais, K. H. 1997. Keeping Daphnja out of the surface film with cetyl
alcohol. Journal of Plankton Research. 19: 149-154.

Dodson, S. I. 1972. Mortality in a population of W. Ecology
53:1011-1023.

Dudycha, J. L and A. J. Tessier. in press. Natural genetic variation of lifespan,
reproduction and juvenile growth in Daphnia. Evolution.

Edney, E. G. and R. W. Gill. 1968. Evolution of senescence and specific
longevity. Nature 220: 281-82.

Fanestil, D. D. and C. H. Barrows. 1965. Aging in the Rotifer. Journal of
Gerontology 20:462-469.

Finch, C. E. 1990. Longevity, senescence and the genome. University of
Chicago Press, Chicago.

Fox, G. A. 1993. Failure-time analysis: Emergence, flowering, survivorship
and other waiting times. Pp. 253-289 in S. M. Scheiner and J. Gurevitch,
eds. Design and analysis of ecological experiments. Chapman and Hall,

New York.

89

Fry, J. D. 1992. The mixed-model analysis of variance applied to quantitative
genetics: Biological meaning of the parameters. Evolution 46:540-550.

Geedey, C. K. 1997. Seasonal clonal succession in DEW
populations. Diss., Michigan State University.

Geedey, C. K., A. J. Tessier and K. Machledt. 1996. Habitat heterogeneity,
environmental change, and the clonal structure of Daphnia
populations. Functional Ecology 10: 613-621.

Goodman, D. 1982. Optimal life histories, optimal notation and the value of
reproductive value. American Naturalist 119: 803-323.

Hall, D. J. 1964. An experimental approach to the dynamics of a natural
population of Daphnia galeatameadgtae. Ecology 45: 94-112.

Hall, D. J., W. E. Cooper and E. E. Werner. 1970. An experimental approach
to the production dynamics and structure of freshwater animal
communities. Limnology and Oceanography 15: 839-928.

Hamilton, W. D. 1966. The moulding of senescence by natural selection.
Journal of Theoretical Biology 12: 12-45.

Harrison, D. E. and J. R. Archer. 1988. Natural selection for extended
longevity from food restriction. Growth Development and Aging 52:
207-211.

Hebert, P. D. N. 1995. The Daphnia of North America: An illustrated fauna,
version 1. CD-ROM. The University of Guelph, Guelph, Ontario.

Hollingsworth, M. J. 1966. Temperature and the rate of ageing in Drosophila

subobscura. Experimental Gerontology 1: 259-267.

90

Hutchinson, G. E. 1967. A treatise on limnology, vol. II. John Wiley 8: Sons,
New York.

Ingle, L. 1933. Effects of environmental conditions on longevity. Science 78:
511. I

Ingle, L., T. R. Wood and A. M. Banta. 1937. A study of longevity, growth,
reproduction and heart rate in W as influenced by
limitations in quantity of food. Journal of Experimental Zoology 76:
325-52.

Kalbfleisch, J. D. 8: R. L. Prentice. 1980. The statistical analysis of failure time
data. John Wiley 8: Sons, New York.

Keller, L. and M. Genoud. 1997. Extraordinary lifespans in ants: a test of
evolutionary theories of ageing. Nature 389: 958-960.

Klass, M. R. 1977. Aging in the nematode Qaeaerhahdjtiaflegane: major
biological and environmental factors influencing lifespan. Mechanisms
of Ageing and Development 6:413-429.

Knoechel, R. and L. B. Holtby 1986. Construction and validation of a body-
length based model for the prediction of cladoceran community filtering
rates. Limnology and Oceanography 31: 1-16.

Korpelainen, H. 1986. The effects of temperature and photoperiod on life

history parameters of Daphaiamagna (Crustacea: Cladocera).

Freshwater Biology 16: 615-620.

91

Lehman, N. M. E. Pfrender, P. A. Morin, T. J. Crease and M. Lynch. 1995. A
hierarchical molecular phylogeny within the genus Daphnia. Molecular
Phylogenetics and Evolution 4: 395-407.

Leibold, M. A. and A. J. Tessier. 1991. Contrasting patterns of body size for

Daphnia species that segregate by habitat. Oecologia 86: 342-348.

 

. 1998. Experimental compromise and mechanistic approaches to the
evolutionary ecology of interacting Daphnia species. in W. J. Resetarits,
Jr. and J. Bernardo, eds. Experimental ecology. Oxford Univ. Press, New
York.

Lynch, M. 1984. The limits to life history evolution in Qapkmja. Evolution
38: 465-82.

MacArthur, J. W. and W. H. T. Baillie. 1929. Metabolic activity and the
duration of life. 1. Influence of temperature on longevity in Daphnia
magna. Journal of Experimental Zoology 53: 221-242.

McCay, C. M., L. A. Maynard, G. Sperling and L. L. Barnes. 1939. Retarded
growth, lifespan, ultimate body size and age changes in the albino rat
after feeding diets restricted in calories. Journal of Nutrition 18:1-13.

McCay, C. M., F. Pope and W. Lunsford. 1956. Experimental prolongation of
the life Span. Bulletin of the New York Academy of Medicine 32:91-101.

Meadow, N. D. and C. H. Barrows. 1971. Studies on aging in a bdelloid
rotifer. II. The effects of various environmental conditions and
maternal age on longevity and fecundity. Journal of Gerontology 26:302-

309.

92

Medawar, P. B. 1952. An unsolved problem of biology. H. K. Lewis, London.

Miquel, J. P. R. Lundgren, K. G. Bensch and H. Atlan. 1976. Effects of
temperature on the life span, vitality and fine structure of Dmsaphjla
melanagaater. Mechanisms of Ageing and Development. 5: 347-370.

Orcutt, JD. and Porter, KG. 1994. The synergistic effects of temperature and
food concentration on life history parameters of Daphnia. Oecologia 63:
300-306.

Partridge, L. and N. H. Barton. 1993. Optimality, mutation and the evolution
of ageing. Nature 362: 305-31.

1996. On measuring the rate of ageing. Proceedings of the Royal

 

Society of London Series B 263: 1365-1371.

Partridge, L and P. H. Harvey. 1988. The ecological context of life history
evolution. Science 241:119-1455.

Promislow, D. E. L. 1991. Senescence in natural populations of mammals: A
comparative study. Evolution 45: 1869-1887.

Reznick, D. 1993. New model systems for studying the evolutionary biology
of aging: Crustacea. Genetica 91:79-88.

Ricklefs, R. E. 1998. Evolutionary theories of aging: Confirmation of a
fundamental prediction, with implications for the genetic basis and
evolution of life span. American Naturalist 152: 24-44.

Rose, M. R. 1991. Evolutionary Biology of Aging. Oxford Univ. Press, New

York.

93

Sarma, S. S. S. and T. R. Rao. 1991. The combined effects of food and
temperature on the life-history parameters of W Muller
(Rotifera). Internationale Revue der Gesamten Hybrobiologie. 76: 225-
239.

Schlichting, C. D. and M. Pigliucci. 1998. Phenotypic evolution: a reaction
norm perspective. Sinauer, Sunderland MA, USA.

Tatar, M., D. W. Gray and J. W. Carey. 1997. Altitudinal variation for
senescence in Mejanapae grasshoppers. Oecologia 111: 357-364.

Taylor, D. J and P. D. N. Hebert. 1992. Daphnjagaleajamendgtae as a cryptic
species complex with interspecific hybrids. Limnology and
Oceanography 37:658-665.

Taylor, D. J. and P. D. N. Hebert. 1993. Habitat-dependent hybrid parentage
and differential introgression between neighboringly sympatric Daphnia
species. Proceedings of the National Academy of Sciences of the USA 90:
7079-7083.

Taylor, D. J., P. D. N. Hebert and J. K. Colbourne. 1996. Phylogenetics and
evolution of the Daphnia lengiepina group (Crustacea) based on 128
rDNA sequence and allozyme variation. Molecular Phylogenetics and
Evolution 5: 495-510.

Tessier, A. J. and N. Consolatti. 1989. Variation in offspring size in Daphnia

and consequences for individual fitness. Oikos 56: 269-276.

 

. 1991. Resource quantity and offspring quality in Daphnia. Ecology

72: 468-478.

94

Tessier, A. J., A. Young and M. Leibold. 1992. Population dynamics and body-
size selection in Daphnia. Lirnnol. Oceanogr. 37: 1-13.
Threlkeld, S. T. 1979. The midsummer dynamics of two Daphnia species in

Wintergreen Lake, Michigan. Ecology 60:165-179.

 

. 1980. Habitat selection and population growth of two Cladocerans in
seasonal environments. in W. C. Kerfoot, ed. Evolution and ecology of
zooplankton communities. University Press of New England, Hanover,
NH, USA.

von Ende, C. N. 1993. Repeated-measures analysis: Growth and other time-

dependent measures. Pp. 113-137 in S. M. Scheiner and J. Gurevitch, eds.
Design and analysis of ecological experiments. Chapman and Hall, New
York.

Weindruch, R. and R. L. Walford. 1988. The retardation of aging and disease
by dietary restriction. Charles C. Thomas, Springfield, IL, USA.

Williams, G. C. 1957. Pleiotropy, natural selection and the evolution of
senescence. Evolution 11: 398-411.

Yu, B. P., ed. 1994. Modulation of aging processes by dietary restriction. CRC

Press, Boca Raton, FL, USA.

95

Table 3-1. Differences in mortality rate acceleration between nominal species

of Daphnja. PH refers to the proportional hazards model, testing whether the

rate of mortality senescence differs. AFT refers to the accelerated failure-times

model, testing whether the timing of senescence differs. For all tests df = 1.

D. l _ l' .
26° High Food
26° Low Food
20° High Food

20° Low Food

26° High Food
26° Low Food
20° High Food

20° Low Food

96

 

PH AFT

(rate) (timing)
P X2 P X2

==fi=========

.9039 0.015 .8264 0.04
.0190 5.5 .0966 2.8
.0001 57.6 .0001 566.2
.0001 82.3 .0001 548.4
.0002 14.2 .0001 24.1
.5813 0.30 .0001 40,115.l
.0001 23.4 .0001 2896.9
.0076 7.1 .0076 7.1

Table 3-2. Repeated measures ANOVA of intrinsic value and reproductive
value in Daphnia. Differences in rates of senescence are indicated by

significant time x species interactions, so only the interaction effect is listed.

Intrinsic Value Reproductive value
p F df p F df

—_
- i i
26°H < 0.0001 4.11 1, 18 < 0.0001 16.23 1, 17
26°L 0.1638 1.40 1, 19 0.0386 0.33 1, l8
20°H < 0.0001 23.24 1, 46 < 0.0001 8.81 1, 45
20°L < 0.0001 3.99 l, 76 0.0308 3.74 l, 75
D. - if r
26°H < 0.0001 3.72 l, 25 < 0.0001 3.98 1, 24
26°L 0.2689 1.21 1, 19 0.6443 0.47 l, 18
20°H < 0.0001 23.76 1, 38 < 0.0001 24.44 1, 37

20°L < 0.0001 6.38 1, 43 0.2898 6.28 1, 42

 

97

Table 3-3. Tests of genetic variation in the plasticity of the age when only 25% of

intrinsic value remains (ix25); for the full model r2 = 0.91.

 

Source df F p

Species complex 1, 4.0 1.03 0.37
Temperature 1, 4.0 22.32 0 . O 09 1 * *
Food level 1, 4.1 10.00 0 . 0332*
Taxon (complex) 4, 39 12.29 <0 . 0001* * *
Complex x temperature 1, 4.0 1.70 O . 267
Complex x food level 1, 4.1 0.89 0.40
Complex x food x temperature 1, 4.2 1.43 0 .29
Temperature x food level 1, 4.2 6.28 O . 0634
Taxon (complex) x temperature 4,4 12.8 0 . 0149*
Taxon (complex) x food level 4,4 2.61 O . 19
Taxon (complex) x food x temp 4, 39 0.79 O . 54

98

26° High

I I I l l I l I l I I I l I I l l I l l I I l L l I

 

26° Low

 

L L I I l l l I l l l I l l l I l l l I l l l I

   
  
 

20° High

1 l l l J l l I l l I I

Cumulative Survivorship

20° Low

  

 

0 20 40 60 80 100 120 140 160

Age (days)

Figure 3-1. Survivorship of mm m four environments. Solid

lines are replicate populations of m dotted lines are D_p_ul_ex_.
Dashed vertical lines indicate age at maturity.

99

26° High

 

 

 

Intrinsic Value

 

I 1 1 4 L 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 J 1 J
26° Low
I l l 41 I l I I I I I I l I I I I l I I l I I
20° High
0 n l 1 1 l 1 1 Il'hg-ua 1 1 l 1 L I 1 Ia 1 I 1 1 1 I 1 1

 

  

 

 

- "“0 l'l"
1 1 1 I L L. .1 l L I' n

0 20 40 60 80 100 120 140

Figure 2. Intrinsic value of Wm. Solid lines show replicate
populations of mm dotted lines are D_pulex. At 26°, one population
of D_p_ule; 1s virtually hidden behind another.

100

26° High

 

I I l J l I 1 L L I l l I I I I l I I I L I l I

 
    

   

 

25
20 26° Low
15
10

'~“i“.
5 ,.~.x.““ ‘32,

I l l I l I l I I I l I I l l I l I I I I I
20° High

Reproductive Value

 

 

0 20 40 60 80 100 120 140

Age (days)

Figure 3. Reproductive value of W2 Solid lines show replicate
populations of Marja, dotted lines are W. Black curves are cubic
polynomials fit to the data via least-squares regression.

101

 

Cumulative Survivorship

 

 

 

Figure 4. Survivorship of W in four environments.
Solid lines are replicate populations of mm dotted lines are I;
We. Dashed vertical lines indicate age at maturity. The extreme
population of We is excluded from statistical comparison (see text
for explanation)

102

msic Value

Intr

  
  
   
   

26° High

JIIIIIIIIIIIIIIIIIIIII

26° Low

IIIIIIIIIIIIIIIIIIIIIIIII

20° High

l-unlllllll

 

 

'.. ‘‘‘‘‘
.1IIIIlIIIIIIIIIIIIIIII"I"L.II-I

20° Low

10 20 30 40 50 60 70 80 90

Age (days)

Figure 3-5. Intrinsic value of WW. Solid lines show
replicate populations of mm dotted lines are We. The
extreme population of W is excluded from statistical comparison
(see text for explanation)

103

26° High

   

“\“IIIUI'IIUII-I-r,’
“\“nihuuu,” -{

   
  

 

IIIIIIIIIIIJIIIIIIIII

26° Low

:. “‘ullln‘l’

.I‘
\‘ fl“

.5)
“\W-

   

"HI,” ’6

  
 
 
  

IIIILIIIIIIIIIIIIIIIIIIIII

”mu-“umnlumumu 20° ngh
I "hr.

\‘uhllm‘m"'M""”H'u """
Q“-
«3..-

    
 
 
  

Reproductive Value

   

v . ‘
I lllluluu,”

III:
‘\ I I I l l l I I I I II'II"I'H!IIIIIIII I

 

Figure 3-6. Reproductive value of D. menthae-demifera. Solid lines show
replicate populations of D. deng’fera, dotted lines are D. mendgtae. Black
curves are cubic polynomials fit to the data via least-squares regression. The
extreme population of D. mendgtae is excluded from statistical comparison
(see text for explanation)

104

55:28:28 some “8 Saw v5.9a £266 3m momma? 5:33on Hobo vumucfim

.mcocflsmom 3.5%: 95 wfingufi .5597. was 5388 862mm 508 «o mcoufismom 33o acmficegcm sumo
8m flood 32 Ba 33:0 €08 :wE 3w 33m .oom was 309.8% x83 .03.. 98 £089? >20 .mucmficegcw
592082 Bow 5 8: 32 35 mmem wo mucofimuwm was BS 530% Canaan—mom umumgmm NA” 95me

A933 m?

ow on ow om ov om om o_
no-
0 I 1o-
kakmatmfilmfiuONwtmE .Q l mdn
I Nd-
I no-
I od
I Nd
I md
I to
I md
ed

 

q u _ _ d —

 

 

 

£33 33

cm SI cc om ow om om

 

 

utuoazmfitzm .Q

 

410.

Nd
md
v6
I md

 

 

(,1 ) 91123 mmoxg

105

CHAPTER FOUR: MORTALITY RISK AND THE ECOLOGICAL
DISTRIBUTION OF SENESCENCE VARIATION

Introduction

Ecology is classically described as the study of the distribution and
abundance of organisms. Ecologists have long argued that how organisms
respond to their environment is critical to understanding their distribution and
abundance. Ecology is fundamentally joined with evolutionary biology in a
feedback loop because selection is determined by ecological interactions and the
traits through which an organism interacts with its environment are determined
by the evolutionary forces its ancestors experienced. This basic link is at the root
of a view that exploring the distribution of traits in the environment is an
illuminating approach to understanding why particular distribution patterns of
organisms exist (Schoener 1986; Chapin, et a1. 1997; Tessier, et al. in press). In
this approach, researchers ask if environmental heterogeneity determines the

distribution of traits across nature, and if so, how?

Life history encompasses survival and reproduction, the ultimate traits
through which all other traits are filtered by natural selection. Because life
history is a summary outcome of all of an organism’s interactions, it can not
directly reveal how an organism interacts with its environment. Instead it
reflects the net effect of positive and negative interactions. Futhermore, life
history is constrained by evolutionary history and tradeoffs, preventing an

organism from having a life history that is ideally responsive to ecological

106

variation. No one life history can be optimal in all situations. An enormous
diversity of life histories is distributed throughout nature and ecological factors
determine whether a particular life history will succeed or fail through natural
selection. Patterns of ecological heterogeneity generate variation in the traits
favored by selection and therefore influence the patterns in which life histories

are distributed.

Senescence is not traditionally approached as a life history trait crucial to
understanding the relationship between the environment and the distribution of
organisms. Conversely, most scientists seeking to understand senescence have
not sought ecological explanations for why it is highly variable among taxa.
However, senescence is potentially useful for understanding how organisms
perform in their environments because as a component of integrated life histories
it may affect overall constraints on life history optimization. It is also possible
that senescence has widespread, if subtle, effects on ecological interactions, by

lowering an individual’s capabilities.

The basic conundrum of senescence is that it is a deleterious, nearly
ubiquitous trait, and yet natural selection fails to eliminate it. Throughout this
paper I treat senescence as the degradation of an organism’s state (i.e., the ability
to function) that accompanies advancing age. This means that senescence is a
decline in age-specific fitness from the physiological perspective, and
consequently, the Darwinian perspective as well. In life history statistics,
senescence appears as age-specific declines in fecundity and the conditional

probability of survival.

107

Why has natural selection failed to eliminate this trait that is directly
deleterious to Darwinian fitness? Or, in a simplified view, why doesn’t
everything live longer than it does? The obvious answer, that there are costs to
prolonging physiological fitness, is unsatisfactory because cost alone can not
explain why there is so much variation in senescence. The evolutionary theory of
senescence offers an explanation because it predicts a causal relationship
between mortality risk and senescence (Williams 1957). The mortality risks
organisms face are highly variable in space, time and scale and therefore lead to

variation of senscence.

The evolutionary theory of senescence is based on the notion that the force
of natural selection declines as organisms age (Medawar 1952). This decline
occurs because as organisms age, an increasing amount of their reproductive
potential is in their past. This occurs even in an imaginary non-senescing
population that is exposed to extrinsic mortality risks. Traits expressed at
specific ages can only influence the realization of future reproductive potential.
As time passes and future reproductive potential decreases, logically the force of
natural selection will also decrease because genes causing deleterious traits at
late ages will have been passed on to the next generation when the individual

was younger.

The theory leads to a direct link between the extrinsic mortality risk (i.e.,
the risk of death due to causes other than senescence) organisms face and the

degree of senescence they will evolve towards (Williams 1957). As extrinsic

108

mortality risk varies, so does the potential benefit of investing in maintaining a
body so that fitness persists into later life. In an environment with a high risk of
extrinsic mortality, there is little utility in late-life investment because organisms
will likely die before a payoff can be realized. This argument assumes that there
are indeed costs to investment in late-life, otherwise even the slightest chance
that an individual may continue surviving is worth betting on. The basic
prediction is that there will be a negative relationship between extrinsic mortality

risk and investment in late-life performance.

A handful of studies have explored the mortality risk hypothesis of the
evolutionary theory of senescence, generally using coarse descriptors of
mortality risk in conjunction with estimates of mean lifespan. The ability to fly, a
trait that presumably improves predator avoidance, is associated with longer
maximum lifespan in mammals (Prothero and Jurgens 1987) and birds (Holmes
and Austad 1995). A comparable rationale was used by Keller & Genoud (1997),
who attributed the relatively long lifespans of eusocial queen ants to their

protection from extrinsic mortality by workers.

Three studies are noteworthy because they sought to measure senescence
based directly on changes of age-specific life history. Tatar, et al. (1997) predicted
that potential lifespan of grasshoppers would be related to season duration and
quanified age-specific intrinsic mortality rates in male grasshoppers from
populations at different elevations. They showed that, when raised in a common
environment, low elevation grasshoppers (i.e., long season) live longer and have

slower increases in age-specific intrinsic mortality than those from high

109

elevation. In a similar vein, Dudycha and Tessier (1999) demonstrated a
relationship between habitat permanence and senescence of Daphnia, with
populations in temporary ponds senescing more rapidly than those from
permanent lakes. Ricklefs (1998) applied a three-parameter Weibull model to
mortality schedules of birds and mammals to estimate initial adult mortality,
assumed to be representative of the extrinsic mortality risk, and showed this is
related to the rate at which mortality increased later in life. None of these studies

actually measured the mortality risk animals face in the wild.

Objective of this study

Here I report on a study designed to test the hypothesized relationship
between extrinsic mortality risk and senescence by directly measuring the
mortality risk natural populations experience and quantifying genetic variation
of senescence among those populations. Dudycha and Tessier (1999)
documented substantial genetic variation of senescence in the Daphnia pulex-
pulicaria species complex of freshwater microcrustaceans. They showed that the
genetic variation roughly corresponded to the catastrophic mortality risk that
populations in temporary ponds face or the low mortality risk known from local
permanent lakes (Leibold and Tessier 1998). I expand upon this work by
measuring population dynamics, birth rates and death rates of D. pulex-pulicaria
in permanent and temporary waters. 1 measured these properties throughout a
year because there is potentially much seasonal variation in mortality risk.

Additionally, I constructed a phylogenetic tree from mitochondrial DNA to

110

evaluate whether selection was enforcing a match between populations’ life

history characteristics and their ecological conditions.

The model system

Daphnia are an ideal model system for the study of evolutionary ecology
because their natural dynamics are simple to follow, and genetic variation of
traits can be easily quantified. D. pulex and D. pulicaria are common
microcrustaceans, whose status as independent species is uncertain because
interspecific hybridization is common. A hierarchical phylogeny based on DNA
failed to resolve them as discrete species (Lehman, et a1. 1995). Others have
found only weak support for their separation (Crease, et al. 1996; Colbourne, et
a1. 1998). Systematics work historically regarded them as one species (Brooks
1957) and they are generally indistinguishable by morphology (Hebert 1995).
However, their nominal labels have remained useful to ecologists and they are
typically applied on the basis of a diagnostic allozyme (LDH) or habitat (Hebert
1995, Deng 1997). I will continue to use both species names as a shorthand for
the ecological conditions populations experience, despite new evidence that the
specific populations studied here also fail to resolve into two distinct genetic

species (see Local Phylogeny, below).

Daphnia pulex is typically found in temporary ponds, while D. pulicaria is
found in deep lakes (Hebert 1995). Variation in pond size, local hydrology and
annual weather patterns create variation in the length of time ponds are

habitable by D. pulex, but all of my study ponds repeatedly dry by summer.

111

 

Hence, maximum lifespan of D. pulex is constrained to a few months by the
temporary nature of their habitat. In contrast, D. pulicaria lives in a permanent
and environmentally more stable habitat, and can persist year-round in deep
lakes (Geedey, et a1. 1996; Geedey 1997). In the absence of any strong abiotic
constraints, resources and predators regulate D. pulicaria population dynamics.
Planktivorous fish are the most important predator, but D. pulicaria can
effectively avoid them in deep water (Tessier and Welser 1991). Populations of
D. pulicaria display low birth and death rates even at the time of peak
planktivory (Leibold and Tessier 1998).

Genetic variation of senescence

Trait measurement.— I quantified genetic variation of senescence by
constructing mortality and fecundity schedules on replicate populations of D.
pulex and D. pulicaria. To combine mortality and fecundity into a single measure
of age-specific fitness, I calculated the ratio of current and expected future
fecundity to the expected lifetime fecundity (Dudycha and Tessier 1999). This
measure, hereafter called intrinsic value, gives the proportionate age-specific
contribution to R0. Intrinsic value is a useful measure of senescence because it is
uncomplicated by age-dependent tradeoffs between mortality and fecundity and
it reasonably represents the physiological state of an average individual

(Dudycha and Tessier 1999).

112

In this report, I augment the data from six populations reported in
Dudycha and Tessier (1999) with life history information on an additional three
populations of D. pulex and four of D. pulicaria. Most populations were used in
both the estimation of life history traits and field demography measurements;
however, one D. pulicaria population (Deep Lake) and one of D. pulex (Mormon
Creek Pond) were logistically difficult to sample and omitted from demographic
measurements. Clones from one D. pulex population (W oodfrog) were not

available when life tables were constructed.

Populations were chosen to represent the extremes of the habitat
permanence gradient. This did not affect the choice of D. pulex populations, but
D. pulicaria populations were chosen based on data indicating they were
perennial populations (Geedey 1997). All populations are in SW Michigan,
except one of D. pulex, which is in Michigan’s Upper Peninsula. Daphnia will
reproduce in the lab by ameiotic parthenogenesis, allowing the capture of natural
genetic variation by randomly sampling a population and establishing clonal
lines. I isolated five randomly chosen clones from each population. Five clones
will not completely represent the variation present within a population, but will
reduce the chance of being misled by a single clone. Lineages were acclimated to
lab conditions (20-22°, satiating food) for .>_ 3 generations prior to life history trait
measurement (Tessier and Consolatti 1989, 1991). Routine monitoring and

microscopic examination (at 460x) verified that lineages were free of parasites.

Mothers of experimental animals were raised from birth for several weeks

with satiating food at low density (1/50 mL) to minimize variation due to

113

maternal effects. Life tables were run in two blocks (including 6 or 7
populations, respectively), starting each with neonates (~12 hr old) from the
mothers’ third or later clutches. For each population, 3 or 4 cohorts consisting of
2 neonates from each of the 5 clones were set up, for a total of 30 or 40
individuals per population. Minor deviations from this ideal occurred in a few
populations because some clones produced only a few female neonates on the
starting day. The life tables were therfore constructed in a nested design, with
cohorts (the lowest level of measurement) nested in population nested in species.
Juvenile suvivorship was near perfect in all populations, with no bias between
species. Animals were fed a satiating food level (20,000 cells / mL Ankistrodesmus
falcatus daily), and raised in growth chambers at 20° on a 16:8 L:D cycle, and
transferred to fresh, filtered lakewater every other day. As animals grew and
died, water volume was adjusted so animals could not filter more than 50% of
the water between feedings (Knoechel and Holtby 1986). Survivorship was
recorded daily and fecundity every 2 d until all animals died. Females who
produced diapausing eggs were temporarily excluded from estimates of m,
because information to convert investment in diapause into an equivalent
investment in immediate reproduction is inadequate (Lynch 1989; Dudycha &
Tessier 1999).

Data analysis.— Throughout the statistical analyses, I treated the data
from Dudycha and Tessier (1999) and the new populations as two separate
blocks. The experimental blocks were not measured concurrently, and it is
possible that seasonal differences in water chemistry or laboratory differences in

maternal quality affected the experimental animals’ life histories. The first block

114

(30 individuals/ population) included Warner, Pine, Lawrence (D. pulicaria),
Wisdom, Otislake III, and Mormon Creek Pond (D. pulex). The additional
populations in block two (40 individuals / population) were Hamilton, Three
Lakes H, Lake XVI, Deep (D. pulicaria), POVI, Bittem, and Roughwood (D. pulex).

It is common practice in studies of senescence to focus on mortality
despite the potential for tradeoffs with fecundity. Separating the analysis of
mortality and fecundity provides insight into the pathways through which
senescence affects evolutionary fitness. To determine if a degradation in survival

ability was apparent, I pooled the data for each ecologically-defined species and

fit it to a Weibull model. In a Weibull model, age-specific mortality (11) at age x

is defined 11x = Mfix)?‘1 (Lee 1980). If y > 1, the hazard function increases with

age, indicating a declining ability to survive that is typically interpreted as

senescence. A scales the model to a baseline rate of mortality, but is not indicative

of senescence.

I applied two types of regression model to the mortality data to test for
interspecific differences in the degradation of survival ability. With an
accelerated failure-time (AFT) model, I tested the timing of decreasing survival
ability. (Kalbfleisch and Prentice 1980; Fox 1993). AFT models require one to
specify an underlying distribution; I chose a gamma distribution because it is a
flexible general distribution that permits, but does not force, the expectation of
monotonically increasing age-specific mortality. I also ran tests using Weibull

and exponential distributions, and found the results to be robust regardless of

115

distribution. With a proportional-hazards (PH) model, I tested for differences
among taxa in the magnitude of mortality hazard increase, which can be

interpreted as the rate of senescence. Juvenile mortality was excluded from these

analyses.

I used univariate repeated-measures AN OVA to test for differences

between species in the pattern of age-specific fecundity and in the shapes of their

intrinsic value curves. A significant time x species interaction indicates

differences in the rates of change within fecundity profiles. In r-m ANOVA, the
degrees of freedom for interactions between time and the between-subjects factor
(here, species) are drawn from both replication of the experimental unit (here,
population) and the number of repeated measurements made. R-m AN OVA
assumes that the data are circular, meaning that the variance of the differences in
trait value between any two ages is homogeneous. Although the tests are

moderately robust to violations of this assumption, F-statistics can be inflated for

the within-subjects factor (time) and its interactions (time x species) (von Ende

1993). When the assumption of circularity was not met, I adjusted the F-test
degrees of freedom by the Huynh-Feldt epsilon (von Ende 1993). All life history

analyses were run in SAS v. 6.09 with Lifereg, PHreg or GLM.

Life history results.— Life history results confirmed the differences
reported in Dudycha and Tessier (1999). Lifespan was substantially shorter in D.
pulex than in D. pulicaria (Figure 1). Both species had increasing age-specific

mortality hazards. The Weibull shape parameter was notably larger in D. pulex (y

116

= 4.7; 95% CI = 4.3 - 5.3) than in D. pulicaria (y = 2.3; 95% CI = 2.0 - 2.5), indicating
that in D. pulex survival declines faster than in D. pulicaria. Both of these
parameter values are somewhat lower than those reported for the analysis of
only block one (Dudycha & Tessier 1999), reflecting the greater genetic variation
in the pooled data. As in the earlier analysis, D. pulex had both an earlier period

of mortality senescence (AFT model; p < 0.0001, x2 = 97.04, df = 1) and a faster
rate of mortality senescence (PH model; p < 0.0001, 12 = 49.02, df = 1). Fecundity

also differed strongly between the two species (Figure 2). Fecundity peaked

earlier and declined faster in D. pulex than in D. pulicaria (time x species

interaction; F = 4.75, p < 0.0001, df = 1, 48, s = 0.54), but did not differ between

blocks (p = 0.25, F = 1.49, df = 1,9). Intrinsic value also declined faster in D. pulex

than in D. pulicaria (time x species interaction; F = 40.7, p < 0.0001, df = 1, 48, s =

0.10; Figure 3).

Field Demography

Methods.— In order to estimate extrinsic death rates, demographic
parameters were measured in six replicate populations each of Daphnia pulicaria
and D. pulex from October 1997 through September 1998. The general procedure
was to measure change in population density to estimate instantaneous
population growth rates (r) and clutch size, egg age distribution and temperature
to estimate instantaneous birth rates (b), similar to Tessier, et a1. (1992). Egg age

distribution was determined based on developmental stages described in

117

 

 

Threlkeld (1979) and pictured in Esslova (1959). By incorporating information
about the egg age distribution, many assumptions in calculating birth rates from
clutch size and temperature are eliminated and a more accurate estimate of b can
be obtained (Bottrell, et al. 1976; Threlkeld 1979; Rigler and Downing 1984).
Once r and b have been estimated, the instataneous death rate (d) is taken from
the equation r = b + d. Because I used the egg age distribution in birth rate
calculations, estimates of d produced this way are primarily subject to sampling
error through measurement of r (DeMott 1980). For this reason I adopted
sampling protocols for population density that are spatially integrative and

highly repeatable, minimizing sampling variation in r.

Different sampling protocols for lakes and ponds were dictated by habitat

differences. In the six lakes, samples were taken with vertical hauls of an 80um

mesh, 11.5 cm diameter Birge plankton net. On each sampling date, three
replicate samples were taken. Each sample consisted of three vertical hauls
made at different locations (> 10m apart) in the deepest area of the lake and then
pooled. No attempt was made to precisely duplicate locations on different dates
other than returning to the deepest area of the lake. Lakes were sampled to
obtain population densities throughout the year, and estimates of birth rate
during focal periods (late October, late March, late May and early September)
representative of the major changes in D. pulicaria dynamics (Hu and Tessier
1995;Thre1keld 1979). For each focal period, samples were taken on three dates
separated by the egg development time indicated by the water’s average
temperature. Birth rate was also estimated in mid December (under ice-cover) in

four populations.

118

In the six ponds, three replicate spatially pooled samples were taken with
a 3-L pitcher. For each sample, I plunged the pitcher into the water at 10 - 20
locations, moving through the pond without backtracking into disturbed areas.

At each location, I collected three pitcherfuls of water, passing each through a

separate 80um mesh screen. At successive locations, I varied both the depth

from which the water was taken and the order in which the screens were used.
This procedure resulted in three samples composed of 30 - 60 L of water taken
from the same locations, but minimized the effects of spatial structure in Daphnia
distribution or behavioral response to disturbance. In some instances, water
levels were too low to take three complete samples and fewer or smaller samples
were taken. Ponds were monitored for animals starting when they first had
water in them, which varied from November to late February. Once D. pulex
appeared, samples were taken continuously, spaced by the egg development

time, until ponds dried.

Samples were either preserved with cold sugar-formalin (Prepas 1978)
immediately or placed on ice for preservation in the lab. On dates when birth
rate was estimated, egg data on at least 100 randomly chosen adult females (if
available) from one sample was recorded within 36 hours of preservation.
Numbers of adult females, juvenile females and males were counted later.
Usually a 5 - 10% subsample was counted, but if animals were rare, the entire

sample was counted.

There can be substantial vertical layering to the spatial distribution of D.

pulicaria in lakes, therefore it is inappropriate to report densities based on water

119

 

 

volume. Here, I report densities based on the surface area sampled (312 cmz),
assuming that horizontal structure is low and its effect is minimized by the
sampling design. In contrast, I report densities in ponds based on volume.
Although this is problematic for absolute comparisons of the population size, I

am primarily interested in the dynamics of population growth.

Results.— Populations were much more dynamic in temporary ponds
than in permanent lakes (Figure 4). D. pulex populations expanded and
contracted over four orders of magnitude in less than. five months, whereas some
D. pulicaria populations fluctuated at most by three orders of magnitude over a
year. The stability of D. pulicaria produces population growth rates that are
generally near zero. D. pulex, in contrast, shows extreme growth rates, flipping
from +0.3 to -0.4 per day in the span of a month (Figure 5). Birth rates in D.
pulicaria were relatively low, with spring and fall peaks remaining below 0.5
neonates per adult female per day in most cases. In D. pulex, most populations

had a birth rate peak in excess of 2.5 neonates per female per day (Figure 6).

The differences in population growth and birth rates between the habitats
resulted in striking differences in death rates. Death rate in D. pulicaria was both
lower and more constant than in D. pulex (Figure 7). In a few populations,
negative death rates were observed. In zooplankton, this is not surprising
because individuals hatched out of the sedimentary egg bank can contribute to
poulation gowth. Such individuals cause underestimation of death rates, and
negative death rates have been used to estimate the minimum hatching rate of

zooplankton from the egg bank. The negative death rate observed in D. pulicaria

120

in late September and in D. pulex in late March occur at times when hatching is

expected to be highest for these species.

In the D. pulicaria populations I study, I assume that the hatching rate is
negligible relative to overall population dynamics. This assumption is
reasonable because these populations have a perennial phenology, and are
therefore predicted to invest little in diapause. The prediction has not been well-
tested, but I rarely observed D. pulicaria females producing diapause eggs, and
on those occasions it included only a small portion of the adult females. D. pulex,
in contrast, must invest in diapause to persist through the dry period. In
temporary ponds, genotypes that hatch late suffer a demographic cost relative to
genotypes that hatch early, leading to a narrow hatching window. I assumed
that hatching was essentially complete by the time population densities were
sufficiently large to estimate birth rate, and should therefore minimally affect my
death rate estimates. However, failure of this assumption means that D. pulex
death rates, and consequently the difference between species, was

underestimated.
Relationship between Demography and Life History

As a test of whether senescence is a function of extrinsic mortality, I
examined the relationship between mean per capita death rate and an index of
the fitness persistence into late life (i,25, the age when only 25% of intrinsic value
remains). I summarized the mortality measurements by calculating weighed
mean death rates over the year in each population, with each death rate estimate

for a population weighted by population density at the time of measurement.

121

There was a strong negative relationship between mortality risk and the
index of fitness persistence (Figure 8). Linear regression showed that the
relationship is significantly negative (p = 0.0012, F = 21.79, df = 1,9), but a power
relationship had a better fit to the data (linear R2 = 0.68, power R2 = 0.95).

Local Phylogeny

DNA sequencing—In order to better understand the relationships among
the replicate populations, I estimated their phylogenetic relationships based on a
mitochondrial DNA sequence. I generally followed the protocol described in
Lehman, et a1. (1995). When possible, I used up to three clones per population to
better evaluate whether migration was homogenizing the metapopulation. DNA
was extracted by autoclaving fresh or frozen individuals in 250 ul 10% chelating
resin for 15 min (Walsh et a1. 1991). The supernatant from these extractions was
used in the polymerase chain reaction (PCR) to amplify the rapidly-evolving
control region. I used the primers specified in Lehman, et a1. (1995) in a nested
PCR to produce DNA fragments roughly 730 bases long. PCR reactions were
performed with 10 pmol of each primer, 0.2 units of Taq DNA polymerase and
10 ul extraction supernatant or 2 11] PCR product in 50 ul reaction volumes.
External primers were thermocycled for 92° 10 min + (92° 1 min, 50° 1 min, 72° 1
min) X 40 + 72° 10 min. Internal primers were amplified with a similar program,
but denatured at 94° and annealed at 53°. The product of the internal primers
was quantitated on a 1% agarose gel with Boehringer-Mannheim molecular

weight marker XIV. This product was then purified with a Gene Clean” kit (Bio

122

101) and sequenced at Michigan State University’ 3 automated squencing facility.
Sequencing reactions normally included >100 ng purified PCR product and 20
pmol primer. The sequences were then base-checked and aligned by eye for

analysis.

Gene tree construction.— I constructed a tree based on 551 bp of sequence
starting at the 5’ end of the amplified fragment using PAUP" 4.0b2a (Swofford
1998). My goal was to evaluate whether ancestry was entirely confounded with
habitat differences, thus I am interested in whether the ecologically-defined
species are strictly monophyletic. I do not argue that I can recover the true
phylogenetic relationships among the study populations. Rather, I seek to
understand whether there is evidence that selection repeatedly causes only
certain genotypes to be successful in the different habitats. This may occur
through repeated evolutionary divergence, clonal selection, migration between
habitats with subsequent introgression, or long-distance dispersal into the local

area.

I constructed trees using parsimony, phenetic and maximum likelihood
approaches. For the parsimony approach, I performed heuristic searches on 1000
bootstrapped pseudoreplicates of the data and created a 50% majority-rule
consensus tree, using D. melanica as an outgroup (Colbourne and Hebert 1996;
unpublished sequence provided by N. Lehman). I treated gaps as a fifth
character state, which produced 45 equally weighted parsimony-informative
characters. In each pseudoreplicate, tree bisection-reconnection branch swapping

with steepest descent in effect was performed on 25 starting trees generated by

123

random stepwise addition. Character states were optimized via accelerated
transformation. This produced a tree with treelength = 117, CI = 0.73 and R1 =
0.90 (Figure 9). Phenetic analysis was performed by neighbor-joining (NJ) and
the unweighted pair-group method (UPGMA) under the same parameters as
parsimony except that gaps were treated as missing data. 50°/o majority-rule
consensus trees of 1000 bootstrap pseudoreplicates from N] and UPGMA were
topologically similar to Figure 9. Maximum likelihood (ML) analysis used the
empirical nucleotide frequencies, assumed all sites to evolve at the same rate and
did not enforce a clock. For the ML, 1 used only 10 replicate addition sequences.
This produced three trees of maximum likelihood, whose consensus is identical

to Figure 9.

All methods strongly supported a division of the populations into two
clades, one containing only D. pulicaria genotypes and one including a mixture of
D. pulex and D. pulicaria (Figure 9). This rejects the possibility that ancestry and
habitat perfectly covary. Furthermore, multiple clones from single populations
never had identical sequences, but generally grouped together. This indicates
that populations are not simply representatives of one panmictic population.
However, some notable exceptions to this (Three Lakes II and Lawrence Lake

clones are found in both major clades) suggests migration is occurring.

Discussion

The data reported here clearly support the hypothesis that investment in

late life performance is negatively related to extrinsic mortality risk. D. pulex

124

inhabits temporary ponds, resulting in death rates that are both high and
seasonally variable. D. pulicaria, in contrast, inhabits permanent lakes that result
in a much more stable population dynamics and low death rates. Logically, if
maintaining physiological fitness incurs costs that are equal between these taxa,
D. pulex should invest relatively little in maintenance compared to D. pulicaria
because its expected returns are lower. Using demographic life history traits to
integrate physiological traits, the data shows that D. pulex does invest less in
sustaining fitness in late life. This is evident in both intrinsic survival and

fecundity, and also in a combined measure of age-specific fitness.

Although the major differences occur between nominal species associated
with strikingly different habitats, there is good evidence that that nominal
taxonomy does not adquately reflect the genetic history of this species complex.
Thus, the labels applied to D. pulex and D. pulicaria are more appropriately
thought of as defining an ecological rather than a genetic difference. This means
that the relationship is not simply one of two species that happen to differ in
both their habitat occupied and life history. Populations of D. pulex-pulicaria
genetically vary in their senescence characteristics and rapidly senescing
populations occupy habitats with high extrinsic mortality risk. This suggests
that the distribution of senescence variation is directly influenced by ecological

variation.

One apparent discrepancy is that life expectancies in the field are notably
shorter than median lifespans in the lab. In D. pulex, instantaneous mortality

rates average ~0.25 per day, yielding a life expectancy of about 4 days. However,

125

their median lab lifespan is 20-40 days. A similar pattern is true of D. pulicaria,
with a typical instantaneous mortality rate of ~0.1 per day, a life expectancy of 10
days, but a median lab lifespan of 30-90 days. Environmental differences
between the field and the lab are unlikely to have caused this discrepancy
through phenotypic plasticity. Temperature and resource level are the
environmental factors likely to have the strongest influence on senescence
phenotype. Temperature has been shown to be more important in determining
lifespan than resources for D. pulex-pulicaria (Dudycha, in review) and
conducting the laboratory measurements of lifespan at average field
temperatures (i.e., colder) would simply enlarge the difference between field life
expectancy and lab median lifespan. This difference suggests that death due to
intrinsic senescent deterioration, i.e., dying of old age, is not a major factor in the
wild. Instead, mortality in the wild is primarily caused by extrinsic factors such
as predation, starvation, disease or anoxia. Hence, the relationship between
death rates in the field and senescence is not caused by senescence determining

the field death rates.

Juvenile mortality is present in both habitat types (Dudycha, unpubl. data)
but was essentially non-existent in the laboratory. This could partially account
for the differences in field life expectancy and median lifespan. Indeed, field life
expectancy barely appoaches adulthood in D. pulicaria, and is well within

juvenile range for D. pulex.

Juvenile mortality is not just important for comparing life expectancy and

median lifespan; it potentially alters the predicted effect of selection on

126

senescence. As presented, the current analysis does not decompose mortality
risk into juvenile and adult mortality risk. This can be important for the
evolution of senescence because the predicted relationship between mortality
risk and senescence holds true only if juvenile mortality is nil, is constant across
compared groups, or is positively related to adult mortality risk. The reason for
these restrictions is that juvenile mortality risk, by increasing the evoluitonary
value of individuals who have already survived to maturity, will drive selection
for increased preservation of adulthood (Abrams 1991) Consequently, a
limitation of this report is that juvenile mortality, generally associated with
invertebrate predation or starvation, is likely to be important for both D. pulex
and D. pulicaria. It will be important to confirm that the age-structure of
mortality risk does not reverse predictions for D. pulex-pulicaria senescence.
Indications are that adult mortality is higher in D. pulex than D. pulicaria, because
few juveniles are present during much of D. pulex phenology, but it is not clear

how juvenile mortality differs.

Few sources of extrinsic mortality are likely to have equal effects on
juveniles and adults. The importance of the balance between juvenile and adult
mortality risk, and of variation in the balance among populations, suggests that
an understanding of what mechanisms cause Daphnia mortality along the pond-
lake gradient will be necessary to understand senescence evolution fully.
Predation, starvation, disease and abiotic conditions are likely to be the main
factors, but can operate differently in ponds and lakes. No information is
available on the comparative risk of disease-related death either among habitats

or across ages, so it will not be discussed further.

127

Predation can be roughly divided into that by vertebrates (planktivorous
fish, newts) and invertebrates. Vertebrates are known to preferentially prey on
large (=old, Daphnia have indeterminant growth) individuals and can be thought
to select for accelerated senescence. However, local populations of D. pulicaria
are known to avoid fish behaviorally and seasonally, minimizing their risk
exposure. During my demographic monitoring, I observed only one newt in one
pond, although I observed them to be locally abundant in shallow permanent
waters. In lakes, the major invertebrate predator on D. pulicaria is the predatory
rnidge Chaoborus, which is known to preferentially prey on small individuals.
Chaoborus also avoids fish predation, therefore it spatially co-occurs with D.
pulicaria and can strongly influence D. pulicaria dynamics. Temporary ponds
harbor a wide range of invertebrate predators, mostly insects, that are known to
prey on Daphnia. However, it is unclear which, if any, exert important predation
pressure on D. pulex. In sum, it is unclear whether the age-structure of predation
will act to reinforce or countermand selection on senescence based on total

mortality rates.

Starvation is probably more important to juvenile mortality than to adult
mortality. In the lakes I studied, it has been shown that resources peak in the
spring and fall (similar to the birth rate peaks, Figure 5) but that during most of
the warm months resources are low due to competition from other zooplankton.
Winter resource availability is poorly known, but will be low if snow cover
prevents light from penetrating the ice. In contrast, ponds have a high resource

availability until after D. pulex populations have peaked. At that time, resources

128

may decline to levels where starvation is possible, but D. pulex populations are
overwhelmingly adult. The among-habitat pattern of starvation potential,
combined with an expectation that juveniles are more susceptible, indicates that
age-dependent starvation will act to reinforce selection on senescence based on

total mortality rates.

Dessication and anoxia are likely to be the most important abiotic causes
of mortality. Dessication is only relevant in the ponds and is indiscrirninant with
respect to age. At the time ponds dry, there are few juveniles in the population
because reproduction has switched to producing diapausing eggs. Therefore, the
actual influence of dessication is primarily on adult death, but this influence may
be small overall because p0pulations decline prior to drying. Anoxia could play
an important role in both habitat types. In lakes, oxygen is depleted in bottom
waters between spring and fall due to decomposition combined with a lack of
mixing with surface waters. Because D. pulicaria use the bottom waters as a
vertebrate predation refuge in the daytime, their safe habitat constricts
seasonally. I expect that suffocation is a more significant risk for juveniles,
because hemoglobin production is costly and adults can aquire more resources
per unit time. However, juveniles may not be exposed to this risk, since they
have a relatively low risk of fish predation and may choose to remain
continuously in the upper waters (where higher resources and temperature
result in faster development). In ponds, anoxia is temporally structured, but not
spatially structured. Oxygen levels in ponds can have strong diel fluctuations
due to daily cycles of photosynthesis, respiration and decomposition combined

with a seasonal progression towards anoxia as the canopy closes and insolation

129

is reduced. These patterns make it unclear whether anoxia-related death will

reinforce or counteract selection on senescence based on total mortality rates.

One surprising aspect of the mortality dynamics is that most D. pulex
populations experience a decline in mortality risk towards the end of the season
well before the ponds dry. This contrasts with the expectation that pond-drying
causes mortality en masse at the end of the season. Two other major
demographic shifts are occurring substantially before pond-drying: total density
typically declines, and females shift reproduction to diapausing eggs. The shift
in reproduction produces a large reduction of birth rates. Consequently, juvenile
density falls. A reduction of juveniles would reduce total mortality risk if that
mortality disproportionately fell on juveniles. Adult death rates could also be
dropping if the proportion of adults that have grown large enough to enter a size

refuge from invertebrate predation is increasing.

The potential for a relationship between longevity (duration of
reproductive lifespan) and diapause exists from both ecological and evolutionary
considerations and deserves focused attention. In Daphnia, diapause and
longevity are alternative mechanisms of dispersing one’s offspring through time.
These are fundamentally similar approaches to temporal dispersal, but their
usefulness is likely to depend on the scale, degree and predictability of
environmental variation. It would be interesting to examine the relative
importance of these dispersal mechanisms along a gradient of habitat quality
variation that ranges from near-term extreme fluctuation to long-term moderate

changes. Diapause and longevity may trade off in Daphnia, because diapause is

130

not the proximate mechanism producing long life. Such a tradeoff is akin to
tradeoffs between reproduction and survival, but alters the picture because it
introduces a second mode of reproduction. Life history evolution may therefore
be balancing three fundamental fitness components: survival, immediate
reproduction and diapause investment. Experimental approaches to life history
optimization under this scenario may require theoretical development of
multidimensional tradeoffs, rather than examining a collection of pairwise

tradeoffs.

131

LITERATURE CITED

Abrams, P. A. 1991. The fitness costs of senescence: The evolutionary
importance of events in early adult life. Evol. Ecol. 5: 343-360.

Abrams, P. A. 1993. Does increased mortality favor the evolution of more rapid
senescence? Evolution 49: 1055-1066.

Bottrell, H. H., A. Duncan, Z. M. Gliwicz,E. Grygierek, A. Herzig, A. Hillbricht-
Ilkowska, H. Kurasawa, P. Larsson and T. Weglenska. 1976. A review of
some problems in zooplankton production studies. Norwegian Journal of
Zoology 24: 419-456.

Brooks, J. L. 1957. The systematics of North American Daphnia. Connecticut
Academy of Arts and Sciences. Hartford, CT.

Chapin III, F. S., et a1. 1997. Biotic control over the functioning of ecosystems.
Science 227: 500-504.

Colbourne, J. K., T. J. Crease, L. J. Weider, P. D. N. Hebert, F. DuFresne and A.
Hobaek. 1998. Phylogenetics and evolution of a circumarctic species
complex (Cladocera: Daphnia pulex). Biological Journal of the Linnean Society
65: 347-365.

Colbourne, J. K. and P. D. N. Hebert. 1996. The systematics of North American
Daphnia (Crustacea: Anomopoda): a molecular phylogenetic approach.
Philosophical Transactions of the Royal Society of London Series B 351:349-
360.

Crease, T. J., S.-K. Lee, S.-L. Yu, K. Spitze, N. Lehman and M. Lynch. 1997.

Allozyme and mtDNA variation in populations of the D. pulex complex from

both sides of the Rocky Mountains. Heredity 79: 242-251.

132

 

DeMott, W. R. 1980. An analysis of the precision of birth and death rate I
estimates for egg-bearing zooplankters. Pp 337-345 in W. C. Kerfoot, ed.
Evolution and Ecology of Zooplankton Communities. University Press of New
England, Hanover, NH.

Deng, H.-W. 1997. Photoperiodic response of sexual reproduction in the
Daphnia pulex group is reversed in two distinct habitats. Limnology and
Oceanography 42: 609-611.

Dudycha, J. L. and A. J. Tessier. 1999. Natural genetic variation of lifespan,
reproduction and juvenile growth in Daphnia. Evolution, in press.

Esslova, M. 1959. Embryonic development of parthenogenic eggs of Daphnia
pulex. Vestn. Cesk. Spol. Zool. 23: 80-88.

Fox, G. A. 1993. Failure-time analysis: Emergence, flowering, survivorship and
other waiting times. Pp. 253-289 in S. M. Scheiner and J. Gurevitch, eds.
Design and analysis of ecological experiments. Chapman and Hall, New
York.

Geedey, C. K. 1997. Seasonal clonal succession in Daphnia pulicaria populations.
Diss., Michigan State University.

Geedey, C. K., A. J. Tessier and K. Machledt. 1996. Habitat heterogeneity,
environmental change, and the clonal structure of Daphnia populations.
Functional Ecology 10: 613—621.

Hamilton, W. D. 1966. The moulding of senescence by natural selection. J.
Theor. Biol. 12: 12-45.

Hebert, P. D. N. 1995. The Daphnia of North America: An illustrated fauna,
version 1. CD-ROM. The University of Guelph, Guelph, Ontario.

133

 

Holmes, D. J. and S. N. Austad. 1995. The evolution of avian senescence
patterns: implications for understanding primary aging processes. Amer.
Zool. 35: 307-317.

Hu, S. S. and A. J. Tessier. 1995. Seasonal succession and the strength of intra-
and interspecific competition in a Daphnia assemblage. Ecology 76: 2278-
2294.

Kalbfleisch, J. D. & R. L. Prentice. 1980. The statistical analysis of failure time
data. John Wiley & Sons, New York.

Keller, L. and M. Genoud. 1997. Extraordinary lifespans in ants: a test of
evolutionary theories of ageing. Nature 389: 958-960.

Knoechel, R. and L. B. Holtby 1986. Construction and validation of a body-
length based model for the prediction of cladoceran community filtering
rates. Limnology and Oceanography 31: 1-16.

Lee, E. T. 1980. Statistical methods for survival data analysis. Lifetime Learning
Publications, Belmont, CA.

Lehman, N. M. E. Pfrender, P. A. Morin, T. J. Crease and M. Lynch. 1995. A
hierarchical molecular phylogeny within the genus Daphnia. Mol. Phyl. Evol.
4: 395-407.

Leibold, M. A. and A. J. Tessier. 1998. Experimental compromise and
mechanistic approaches to the evolutionary ecology of interacting Daphnia
species. in W. J. Resetarits, Jr. and J. Bernardo, eds. Experimental ecology.
Oxford Univ. Press, New York.

Lynch, M. 1989. The life history consequences of resource depression in Daphnia
pulex. Ecology 70: 246-256.

Medawar, P. B. 1952. An unsolved problem of biology. H. K. Lewis, London.

134

Prepas, E. 1978. Sugar-frosted Daphnia: Improved fixation technique for
Cladocera. Limnology and Oceanography 23: 557-559.

Promislow, D. E. L. 1991. Senescence in natural populations of mammals: A
comparative study. Evolution 45: 1869-1887.

Prothero, J. and K. D. Jurgens. 1987. Scaling of maximum life span in mammals:
A review. Basic Life Sci. 42: 49-74.

Reznick, D. 1993. New model systems for studying the evolutionary biology of
aging: Crustacea. Genetica 91: 79-88.

Ricklefs, R. E. 1998. Evolutionary theories of aging: Confirmation of a
fundamental prediction, with implications for the genetic basis and evolution
of life span. Am Nat. 152: 24-44.

Rigler, F. H. and J. A. Downing. 1984. The calculation of secondary productivity.
Pp. 19-58 in J. A. Downing and F. H. Rigler, eds. A manual on methods for
the assesment of secondary productivity in fresh waters. Blackwell.

Schoener, T. W. 1986. Mechanistic approaches to community ecology: A new
reductionism? American Zoologist 26: 81-106.

Swofford, D. L. 1998. PAUP". Phylogenetic analysis using parsimony (*and
other methods). Version 4. Sinauer Associates, Sunderland MA.

Tatar, M., D. W. Gray and J. W. Carey. 1997. Altitudinal variation for senescence
in Melanopus grasshoppers. Oecologia 111: 357-364.

Tessier, A. J. and J. Welser. 1991. Cladoceran assemblages, seasonal succession
and the importance of a hypolimnetic refuge. Freshwater Biology 25: 85-93.

Tessier, A. J. and N. Consolatti. 1989. Variation in offspring size in Daphnia and

consequences for individual fitness. Oikos 56: 269-276.

135

Tessier, A. J. and N. Consolatti. 1991. Resource quantity and offspring quality in
Daphnia. Ecology 72: 468-478.

Tessier, A. J., A. Young and M. Leibold. 1992. Population dynamics and body-
size selection in Daphnia. Limnol. Oceanogr. 37: 1-13.

Tessier, A. J., M. A. Leibold and J. Tsao. in press. A fundamental trade-off in
resource exploitation by Daphnia and consequences to plankton communities.
Ecology.

Threlkeld, S. T. 1979. Estimating cladoceran birth rates: The importance of egg
mortality and the egg age distribution. Limnol. Oceanogr. 24: 601-612.

von Ende, C. N. 1993. Repeated-measures analysis: Growth and other time-
dependent measures. Pp. 113-137 in S. M. Scheiner and J. Gurevitch, eds.
Design and analysis of ecological experiments. Chapman and Hall, New
York.

Walsh, P. S., D. A. Metzger and R. Higuchi.1991. Chelex 100 as a medium for
simple extraction of DNA for PCR-based typing from forensic material.
BioTechniques 10: 506-513.

Williams, G. C. 1957. Pleiotropy, natural selection and the evolution of

senescence. Evolution 11: 398-411.

136

 

 

._I

       
 

e an I r o o
. I o 4 o
»- ,-}JJJ.°J£°.°.I

    

0.9 3
0.8
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2
0.1

D. pulicaria

  

 

 

0

Cumulative Survivorship
O
8
8
e
8
§

Figure 4-1. Survivorship of 12 populations of D. pulex-pulicaria. Open symbols
and dotted lines are populations from temporary ponds, closed symbols and
solid lines are populations from permanent lakes.

137

I—I t—|
VB 0\ on O N

Per-capita fecundity

 

O

 

Figure 4-2. Fecundity of 12 populations of D. pulex-pulicaria. Symbols are as in
figure 1. Fecundity measurements are per 2-day period, smoothed over three
periods.

138

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1

Intrinsic Value

    
     

I“ '
A L ‘
f n
n 1.
_ ‘ {VIM/19,9117:

..-.¢I“4‘A'A’:H NNNNNN

0 20 40 60 80 100

 

  

 

Figure 4-3. Intrinsic value of 12 populations of D. pulex-pulicaria. Symbols are as
in figure 1. Intrinsic value is the sum of current + expected future fecundity
scaled by the expected lifetime fecundity.

139

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

A
3 100000 E
O E
-l-' .
0) 10000 .
'—‘ s
0% :
4.: 1000 E
H E
Q) .
9.. 100 E
V E
a? 10
CD 5
g :
D 1 l l I
1000 :
a: . D l .
. on ex .=.
.42 100 E is a
I—I 5 g - :
I—I : 5
8d 10 E 5.5:: :1;
5: = .3 3; 3!
m 0.1 “‘Wz: 1:1:
001 :.-“‘:“1““-‘ 44‘. 4;: 51': "il
o’t o‘? 61° 61°
C" N» ‘77» 3"»
0° \‘5 Y3 \

Figure 4-4. Population density of D. pulex-pulicaria over one year. Top: density
of six D. pulicaria populations (permanent lakes) measured per unit surface area.
Bottom: density of six D. pulex populations (temporary ponds) measured per
unit volume. The scale of absolute numbers differs between panels, but the scale

of change in density is identical. Symbols are as in figure 1.

140

 

 

$399.0
Npoxoo

55
NO

-0.4

pulation r (day-1)

Po
,5
CD

 

.9

0,,
06,0
‘2»

Figure 4-5. Comparison of population growth rates in D. pulex-pulicaria over one
year. Symbols are as in figure 1.

141

II:

  

A35
H
:'>\ 3
(G
“d 2.5
v
3 2
(U
H 1.5
5 1
H
u-I
m 0.5
0
6‘ o. 6'? 61° 82° 03° 0"?
9‘ (7‘ ~°N\ ‘3’} N,‘ $3,} C"
0 0‘” s? °~ \5 V9 0"

Figure 4-6. Instantaneous birth rates in D. pulex-pulicaria over one year. Symbols
are as in figure 5.

142

Death rate (day-1)

 

Figure 4-7. Instantaneous death rates in D. pulex-pulicaria over one year.
Symbols are as in figure 5.

143

 

 

IX25

 

 

 

 

20
0.0 0.5 1.0 1.5 2.0

DEATHRISK

Figure 4-8. Relationship between ecological mortality risk and senescence. The
x-axis is the average instantaneous death rate over one year, weighted by
population density. The y-axis is the age when 25% of intrinsic value remains, an
index fitness persistence into late life. Boxplots show the data range for D. pulex
(temporary ponds; circles) or D. pulicaria (permanent lakes; stars).

144

90 am *
—3-'DEEA -*.
62 0558 *-

701

 

100 Pms *-
12 PING *-

 

 

 

 

 

 

 

 

53 i WFA

 

 

 

100 ,
10 66
2

96 *
Al was

WARE *’

 

POVID

 

 

POVIF

 

 

‘5 MELANICA

 

—tchango

Figure 49. Gene tree of 25 clones of D. pulex-pulicaria. Clones are identified by
source population and a letter identifying unique isolates. This tree (length 117;
CI = 0.73; R1 = 0.90) is a 50°/o majority rule consensus of 1000 bootstrap
pseudoreplicates searched heuristically for the most parsimonious trees. Large
numbers are bootstrap percentages, small numbers are branch lengths. Clone
labels with an asterisk are nominally classified as D. pulicaria, all others are D.
pulex. See text for details.

145

 

 

1|lllli‘ll111111111111lllllllfill
3 1293 02092 9281